diff --git a/.gitignore b/.gitignore index c542b51..d414712 100644 --- a/.gitignore +++ b/.gitignore @@ -6,4 +6,4 @@ lwip-1.3.0.tar.gz pciutils-2.2.9.tar.bz2 zlib-1.2.3.tar.gz polarssl-1.1.4-gpl.tgz -/xen-4.13.1.tar.gz +/xen-4.13.3.tar.gz diff --git a/sources b/sources index 0fe92bd..cc9d5b9 100644 --- a/sources +++ b/sources @@ -4,4 +4,4 @@ SHA512 (newlib-1.16.0.tar.gz) = 40eb96bbc6736a16b6399e0cdb73e853d0d90b685c967e77 SHA512 (zlib-1.2.3.tar.gz) = 021b958fcd0d346c4ba761bcf0cc40f3522de6186cf5a0a6ea34a70504ce9622b1c2626fce40675bc8282cf5f5ade18473656abc38050f72f5d6480507a2106e SHA512 (polarssl-1.1.4-gpl.tgz) = 88da614e4d3f4409c4fd3bb3e44c7587ba051e3fed4e33d526069a67e8180212e1ea22da984656f50e290049f60ddca65383e5983c0f8884f648d71f698303ad SHA512 (pciutils-2.2.9.tar.bz2) = 2b3d98d027e46d8c08037366dde6f0781ca03c610ef2b380984639e4ef39899ed8d8b8e4cd9c9dc54df101279b95879bd66bfd4d04ad07fef41e847ea7ae32b5 -SHA512 (xen-4.13.1.tar.gz) = b56d20704155d98d803496cba83eb928e0f986a750831cd5600fc88d0ae772fe1456571654375054043d2da8daca255cc98385ebf08b1b1a75ecf7f4b7a0ee90 +SHA512 (xen-4.13.3.tar.gz) = 622127d824b9c49b57282a887fb404e0bad05ff60bccade82e4e0e9b5ad975ff9aa1fba83392e6d8379e9a15340e8ae9785c0913eb11027816e4600432eea6b6 diff --git a/xen.canonicalize.patch b/xen.canonicalize.patch index 43ccf02..0130355 100644 --- a/xen.canonicalize.patch +++ b/xen.canonicalize.patch @@ -38,17 +38,17 @@ return get_node(conn, ctx, *canonical_name, perm); } ---- xen-4.9.0-rc1.2/tools/xenstore/xenstored_core.h.orig 2017-04-12 16:18:57.000000000 +0100 -+++ xen-4.9.0-rc1.2/tools/xenstore/xenstored_core.h 2017-04-13 21:20:29.146368478 +0100 -@@ -148,7 +148,7 @@ +--- xen-4.13.3/tools/xenstore/xenstored_core.h.orig 2021-03-22 16:57:42.000000000 +0000 ++++ xen-4.13.3/tools/xenstore/xenstored_core.h 2021-03-29 19:56:36.642394283 +0100 +@@ -153,7 +153,7 @@ void send_ack(struct connection *conn, enum xsd_sockmsg_type type); /* Canonicalize this path if possible. */ -char *canonicalize(struct connection *conn, const void *ctx, const char *node); +char *xenstore_canonicalize(struct connection *conn, const void *ctx, const char *node); - /* Write a node to the tdb data base. */ - int write_node_raw(struct connection *conn, TDB_DATA *key, struct node *node); + /* Get access permissions. */ + enum xs_perm_type perm_for_conn(struct connection *conn, --- xen-4.8.0/tools/console/testsuite/console-dom0.c.orig 2016-12-05 12:03:27.000000000 +0000 +++ xen-4.8.0/tools/console/testsuite/console-dom0.c 2017-02-26 21:52:24.554678631 +0000 @@ -18,7 +18,7 @@ diff --git a/xen.gcc10.fixes.patch b/xen.gcc10.fixes.patch deleted file mode 100644 index 3159bd7..0000000 --- a/xen.gcc10.fixes.patch +++ /dev/null @@ -1,65 +0,0 @@ ---- xen-4.13.0/tools/xenstore/utils.h.orig 2019-12-17 14:23:09.000000000 +0000 -+++ xen-4.13.0/tools/xenstore/utils.h 2020-01-21 21:13:05.108957447 +0000 -@@ -24,7 +24,7 @@ - void barf(const char *fmt, ...) __attribute__((noreturn)); - void barf_perror(const char *fmt, ...) __attribute__((noreturn)); - --void (*xprintf)(const char *fmt, ...); -+extern void (*xprintf)(const char *fmt, ...); - - #define eprintf(_fmt, _args...) xprintf("[ERR] %s" _fmt, __FUNCTION__, ##_args) - ---- xen-4.13.0/tools/xenstore/xenstored_core.h.orig 2020-01-21 21:15:19.243931307 +0000 -+++ xen-4.13.0/tools/xenstore/xenstored_core.h 2020-01-21 21:38:35.340617819 +0000 -@@ -204,7 +204,7 @@ - /* Open a pipe for signal handling */ - void init_pipe(int reopen_log_pipe[2]); - --xengnttab_handle **xgt_handle; -+extern xengnttab_handle **xgt_handle; - - int remember_string(struct hashtable *hash, const char *str); - ---- xen-4.13.0/tools/libxl/libxlu_pci.c.orig 2019-12-17 14:23:09.000000000 +0000 -+++ xen-4.13.0/tools/libxl/libxlu_pci.c 2020-01-21 21:56:26.812212916 +0000 -@@ -48,7 +48,7 @@ - int xlu_pci_parse_bdf(XLU_Config *cfg, libxl_device_pci *pcidev, const char *str) - { - unsigned state = STATE_DOMAIN; -- unsigned dom, bus, dev, func, vslot = 0; -+ unsigned dom = 0, bus = 0, dev = 0, func = 0, vslot = 0; - char *buf2, *tok, *ptr, *end, *optkey = NULL; - - if ( NULL == (buf2 = ptr = strdup(str)) ) ---- xen-4.13.0/tools/libxl/libxl_utils.c.orig 2019-12-17 14:23:09.000000000 +0000 -+++ xen-4.13.0/tools/libxl/libxl_utils.c 2020-01-21 22:34:52.096300774 +0000 -@@ -1259,7 +1259,7 @@ - } - memset(un, 0, sizeof(struct sockaddr_un)); - un->sun_family = AF_UNIX; -- strncpy(un->sun_path, path, sizeof(un->sun_path)); -+ strncpy(un->sun_path, path, sizeof(un->sun_path)-1); - return 0; - } - ---- xen-4.13.0/tools/debugger/kdd/kdd.h.orig 2019-12-17 14:23:09.000000000 +0000 -+++ xen-4.13.0/tools/debugger/kdd/kdd.h 2020-01-21 23:35:55.458605582 +0000 -@@ -323,7 +323,7 @@ - kdd_msg msg; - kdd_reg reg; - kdd_stc stc; -- uint8_t payload[0]; -+ uint8_t payload[65535]; - }; - } PACKED kdd_pkt; - ---- xen-4.13.0/tools/xenpmd/Makefile.orig 2019-12-17 14:23:09.000000000 +0000 -+++ xen-4.13.0/tools/xenpmd/Makefile 2020-01-22 22:13:16.564873608 +0000 -@@ -3,6 +3,7 @@ - - CFLAGS += -Werror - CFLAGS += $(CFLAGS_libxenstore) -+CFLAGS += -Wno-error=format-truncation - - LDLIBS += $(LDLIBS_libxenstore) - diff --git a/xen.hypervisor.config b/xen.hypervisor.config index cf6aa02..4f2bb09 100644 --- a/xen.hypervisor.config +++ b/xen.hypervisor.config @@ -103,8 +103,10 @@ CONFIG_HARDEN_BRANCH_PREDICTOR=y CONFIG_ARM64_ERRATUM_827319=y CONFIG_ARM64_ERRATUM_824069=y CONFIG_ARM64_ERRATUM_819472=y +CONFIG_ARM64_ERRATUM_843419=y CONFIG_ARM64_ERRATUM_832075=y CONFIG_ARM64_ERRATUM_834220=y +CONFIG_ARM_ERRATUM_858921=y CONFIG_ARM64_HARDEN_BRANCH_PREDICTOR=y CONFIG_ALL_PLAT=y # CONFIG_QEMU is not set diff --git a/xen.ocaml.4.10.patch b/xen.ocaml.4.10.patch index a24ffdd..f6692d0 100644 --- a/xen.ocaml.4.10.patch +++ b/xen.ocaml.4.10.patch @@ -13,7 +13,7 @@ ret = xc_vcpu_getcontext(_H(xch), _D(domid), Int_val(cpu), &ctxt); context = caml_alloc_string(sizeof(ctxt)); -- memcpy(String_val(context), (char *) &ctxt.c, sizeof(ctxt.c)); +- memcpy((char *) String_val(context), &ctxt.c, sizeof(ctxt.c)); + memcpy((char *) Bp_val(context), (char *) &ctxt.c, sizeof(ctxt.c)); CAMLreturn(context); @@ -22,7 +22,7 @@ conring_size = size; ring = caml_alloc_string(count); -- memcpy(String_val(ring), str, count); +- memcpy((char *) String_val(ring), str, count); + memcpy((char *) Bp_val(ring), str, count); free(str); @@ -45,33 +45,13 @@ r = xc_cpuid_set(_H(xch), _D(domid), c_input, (const char **)c_config, out_config); ---- xen-4.13.0/tools/ocaml/libs/xb/xs_ring_stubs.c.orig 2019-12-17 14:23:09.000000000 +0000 -+++ xen-4.13.0/tools/ocaml/libs/xb/xs_ring_stubs.c 2020-01-21 23:51:35.473330934 +0000 -@@ -44,7 +44,7 @@ - CAMLlocal1(ml_result); - - struct mmap_interface *interface = GET_C_STRUCT(ml_interface); -- char *buffer = String_val(ml_buffer); -+ char *buffer = (char *) Bp_val(ml_buffer); - int len = Int_val(ml_len); - int result; - -@@ -103,7 +103,7 @@ - CAMLlocal1(ml_result); - - struct mmap_interface *interface = GET_C_STRUCT(ml_interface); -- char *buffer = String_val(ml_buffer); -+ char *buffer = (char *) Bp_val(ml_buffer); - int len = Int_val(ml_len); - int result; - --- xen-4.13.0/tools/ocaml/libs/xb/xenbus_stubs.c.orig 2019-12-17 14:23:09.000000000 +0000 +++ xen-4.13.0/tools/ocaml/libs/xb/xenbus_stubs.c 2020-01-22 00:04:09.443168991 +0000 @@ -65,7 +65,7 @@ }; ret = caml_alloc_string(sizeof(struct xsd_sockmsg)); -- memcpy(String_val(ret), &xsd, sizeof(struct xsd_sockmsg)); +- memcpy((char *) String_val(ret), &xsd, sizeof(struct xsd_sockmsg)); + memcpy((char *) Bp_val(ret), &xsd, sizeof(struct xsd_sockmsg)); CAMLreturn(ret); diff --git a/xen.spec b/xen.spec index 60b3c61..7a5d044 100644 --- a/xen.spec +++ b/xen.spec @@ -57,8 +57,8 @@ Summary: Xen is a virtual machine monitor Name: xen -Version: 4.13.1 -Release: 4%{?dist} +Version: 4.13.3 +Release: 1%{?dist} License: GPLv2+ and LGPLv2+ and BSD URL: http://xen.org/ Source0: https://downloads.xenproject.org/release/xen/%{version}/xen-%{version}.tar.gz @@ -113,21 +113,8 @@ Patch40: xen.drop.brctl.patch Patch41: xen.python.env.patch Patch42: xen.gcc9.fixes.patch Patch44: xen.ocaml.4.10.patch -Patch45: xen.gcc10.fixes.patch -Patch46: xsa320-4.13-1.patch -Patch47: xsa320-4.13-2.patch -Patch48: xsa317.patch -Patch49: xsa319.patch -Patch50: xsa328-4.13-1.patch -Patch51: xsa328-4.13-2.patch -Patch52: xsa321-4.13-1.patch -Patch53: xsa321-4.13-2.patch -Patch54: xsa321-4.13-3.patch -Patch55: xsa321-4.13-4.patch -Patch56: xsa321-4.13-5.patch -Patch57: xsa321-4.13-6.patch -Patch58: xsa321-4.13-7.patch -Patch59: xsa327.patch +Patch65: zstd-dom0.patch +Patch100: xsa363.patch %if %build_qemutrad @@ -333,21 +320,8 @@ manage Xen virtual machines. %patch41 -p1 %patch42 -p1 %patch44 -p1 -%patch45 -p1 -%patch46 -p1 -%patch47 -p1 -%patch48 -p1 -%patch49 -p1 -%patch50 -p1 -%patch51 -p1 -%patch52 -p1 -%patch53 -p1 -%patch54 -p1 -%patch55 -p1 -%patch56 -p1 -%patch57 -p1 -%patch58 -p1 -%patch59 -p1 +%patch65 -p1 +%patch100 -p1 # qemu-xen-traditional patches pushd tools/qemu-xen-traditional @@ -940,6 +914,97 @@ fi %endif %changelog +* Mon Mar 29 2021 Michael Young - 4.13.3-1 +- update to 4.13.3 + remove patches now included or superceded upstream + adjust xen.hypervisor.config + +* Thu Mar 18 2021 Michael Young - 4.13.2-8 +- HVM soft-reset crashes toolstack [XSA-368, CVE-2021-28687] (#1940610) + +* Wed Feb 17 2021 Michael Young - 4.13.2-7 +- Linux: display frontend "be-alloc" mode is unsupported (comment only) + [XSA-363, CVE-2021-26934] (#1929549) +- arm: The cache may not be cleaned for newly allocated scrubbed pages + [XSA-364, CVE-2021-26933] (#1929547) + +* Thu Jan 21 2021 Michael Young - 4.13.2-6 +- IRQ vector leak on x86 [XSA-360] + +* Wed Dec 16 2020 Michael Young - 4.13.2-5 +- xenstore watch notifications lacking permission checks [XSA-115, + CVE-2020-29480] (#1908091) +- Xenstore: new domains inheriting existing node permissions [XSA-322, + CVE-2020-29481] (#1908095) +- Xenstore: wrong path length check [XSA-323, CVE-2020-29482] (#1908096) +- Xenstore: guests can crash xenstored via watchs [XSA-324, CVE-2020-29484] + (#1908088) +- Xenstore: guests can disturb domain cleanup [XSA-325, CVE-2020-29483] + (#1908087) +- oxenstored memory leak in reset_watches [XSA-330, CVE-2020-29485] + (#1908000) +- undue recursion in x86 HVM context switch code [XSA-348, CVE-2020-29566] + (#1908085) +- oxenstored: node ownership can be changed by unprivileged clients + [XSA-352, CVE-2020-29486] (#1908003) +- oxenstored: permissions not checked on root node [XSA-353, CVE-2020-29479] + (#1908002) +- FIFO event channels control block related ordering [XSA-358, + CVE-2020-29570] (#1907931) +- FIFO event channels control structure ordering [XSA-359, CVE-2020-29571] + (#1908089) + +* Tue Nov 24 2020 Michael Young - 4.13.2-4 +- stack corruption from XSA-346 change [XSA-355] + +* Mon Nov 23 2020 Michael Young - 4.13.2-3 +- support zstd compressed kernels (dom0 only) based on linux kernel code + +* Tue Nov 10 2020 Michael Young - 4.13.2-2 +- Information leak via power sidechannel [XSA-351, CVE-2020-28368] + (#1897146) + +* Tue Nov 03 2020 Michael Young - 4.13.2-1 +- update to 4.13.2 + remove patches now included or superceded upstream + +* Wed Oct 28 2020 Michael Young - 4.13.1-8 +- x86 PV guest INVLPG-like flushes may leave stale TLB entries + [XSA-286, CVE-2020-27674] (#1891092) + +* Tue Oct 20 2020 Michael Young - 4.13.1-7 +- x86: Race condition in Xen mapping code [XSA-345, CVE-2020-27672] + (#1891097) +- undue deferral of IOMMU TLB flushes [XSA-346, CVE-2020-27671] + (#1891093) +- unsafe AMD IOMMU page table updates [XSA-347 CVE-2020-27670] + (#1891088) + +* Tue Sep 22 2020 Michael Young - 4.13.1-6 +- x86 pv: Crash when handling guest access to MSR_MISC_ENABLE [XSA-333, + CVE-2020-25602] (#1881619) +- Missing unlock in XENMEM_acquire_resource error path [XSA-334, + CVE-2020-25598] (#1881616) +- race when migrating timers between x86 HVM vCPU-s [XSA-336, + CVE-2020-25604] (#1881618) +- PCI passthrough code reading back hardware registers [XSA-337, + CVE-2020-25595] (#1881587) +- once valid event channels may not turn invalid [XSA-338, CVE-2020-25597] + (#1881588) +- x86 pv guest kernel DoS via SYSENTER [XSA-339, CVE-2020-25596] + (#1881617) +- Missing memory barriers when accessing/allocating an event channel [XSA-340, + CVE-2020-25603] (#1881583) +- out of bounds event channels available to 32-bit x86 domains [XSA-342, + CVE-2020-25600] (#1881582) +- races with evtchn_reset() [XSA-343, CVE-2020-25599] (#1881581) +- lack of preemption in evtchn_reset() / evtchn_destroy() [XSA-344, + CVE-2020-25601] (#1881586) + +* Tue Aug 25 2020 Michael Young - 4.13.1-5 +- QEMU: usb: out-of-bounds r/w access issue [XSA-335, CVE-2020-14364] + (#1871850) + * Tue Jul 07 2020 Michael Young - 4.13.1-4 - incorrect error handling in event channel port allocation leads to DoS [XSA-317, CVE-2020-15566] (#1854465) diff --git a/xsa317.patch b/xsa317.patch deleted file mode 100644 index 20e2c64..0000000 --- a/xsa317.patch +++ /dev/null @@ -1,50 +0,0 @@ -From aeb46e92f915f19a61d5a8a1f4b696793f64e6fb Mon Sep 17 00:00:00 2001 -From: Julien Grall -Date: Thu, 19 Mar 2020 13:17:31 +0000 -Subject: [PATCH] xen/common: event_channel: Don't ignore error in - get_free_port() - -Currently, get_free_port() is assuming that the port has been allocated -when evtchn_allocate_port() is not return -EBUSY. - -However, the function may return an error when: - - We exhausted all the event channels. This can happen if the limit - configured by the administrator for the guest ('max_event_channels' - in xl cfg) is higher than the ABI used by the guest. For instance, - if the guest is using 2L, the limit should not be higher than 4095. - - We cannot allocate memory (e.g Xen has not more memory). - -Users of get_free_port() (such as EVTCHNOP_alloc_unbound) will validly -assuming the port was valid and will next call evtchn_from_port(). This -will result to a crash as the memory backing the event channel structure -is not present. - -Fixes: 368ae9a05fe ("xen/pvshim: forward evtchn ops between L0 Xen and L2 DomU") -Signed-off-by: Julien Grall -Reviewed-by: Jan Beulich ---- - xen/common/event_channel.c | 8 ++++---- - 1 file changed, 4 insertions(+), 4 deletions(-) - -diff --git a/xen/common/event_channel.c b/xen/common/event_channel.c -index e86e2bfab0..a8d182b584 100644 ---- a/xen/common/event_channel.c -+++ b/xen/common/event_channel.c -@@ -195,10 +195,10 @@ static int get_free_port(struct domain *d) - { - int rc = evtchn_allocate_port(d, port); - -- if ( rc == -EBUSY ) -- continue; -- -- return port; -+ if ( rc == 0 ) -+ return port; -+ else if ( rc != -EBUSY ) -+ return rc; - } - - return -ENOSPC; --- -2.17.1 - diff --git a/xsa319.patch b/xsa319.patch deleted file mode 100644 index 769443c..0000000 --- a/xsa319.patch +++ /dev/null @@ -1,27 +0,0 @@ -From: Jan Beulich -Subject: x86/shadow: correct an inverted conditional in dirty VRAM tracking - -This originally was "mfn_x(mfn) == INVALID_MFN". Make it like this -again, taking the opportunity to also drop the unnecessary nearby -braces. - -This is XSA-319. - -Fixes: 246a5a3377c2 ("xen: Use a typesafe to define INVALID_MFN") -Signed-off-by: Jan Beulich -Reviewed-by: Andrew Cooper - ---- a/xen/arch/x86/mm/shadow/common.c -+++ b/xen/arch/x86/mm/shadow/common.c -@@ -3252,10 +3252,8 @@ int shadow_track_dirty_vram(struct domai - int dirty = 0; - paddr_t sl1ma = dirty_vram->sl1ma[i]; - -- if ( !mfn_eq(mfn, INVALID_MFN) ) -- { -+ if ( mfn_eq(mfn, INVALID_MFN) ) - dirty = 1; -- } - else - { - page = mfn_to_page(mfn); diff --git a/xsa320-4.13-1.patch b/xsa320-4.13-1.patch deleted file mode 100644 index 09eb8ea..0000000 --- a/xsa320-4.13-1.patch +++ /dev/null @@ -1,117 +0,0 @@ -From: Andrew Cooper -Subject: x86/spec-ctrl: CPUID/MSR definitions for Special Register Buffer Data Sampling - -This is part of XSA-320 / CVE-2020-0543 - -Signed-off-by: Andrew Cooper -Reviewed-by: Jan Beulich -Acked-by: Wei Liu - -diff --git a/docs/misc/xen-command-line.pandoc b/docs/misc/xen-command-line.pandoc -index 1d9d816622..9268454297 100644 ---- a/docs/misc/xen-command-line.pandoc -+++ b/docs/misc/xen-command-line.pandoc -@@ -483,10 +483,10 @@ accounting for hardware capabilities as enumerated via CPUID. - - Currently accepted: - --The Speculation Control hardware features `md-clear`, `ibrsb`, `stibp`, `ibpb`, --`l1d-flush` and `ssbd` are used by default if available and applicable. They can --be ignored, e.g. `no-ibrsb`, at which point Xen won't use them itself, and --won't offer them to guests. -+The Speculation Control hardware features `srbds-ctrl`, `md-clear`, `ibrsb`, -+`stibp`, `ibpb`, `l1d-flush` and `ssbd` are used by default if available and -+applicable. They can be ignored, e.g. `no-ibrsb`, at which point Xen won't -+use them itself, and won't offer them to guests. - - ### cpuid_mask_cpu - > `= fam_0f_rev_[cdefg] | fam_10_rev_[bc] | fam_11_rev_b` -diff --git a/tools/libxl/libxl_cpuid.c b/tools/libxl/libxl_cpuid.c -index 6cea4227ba..a78f08b927 100644 ---- a/tools/libxl/libxl_cpuid.c -+++ b/tools/libxl/libxl_cpuid.c -@@ -213,6 +213,7 @@ int libxl_cpuid_parse_config(libxl_cpuid_policy_list *cpuid, const char* str) - - {"avx512-4vnniw",0x00000007, 0, CPUID_REG_EDX, 2, 1}, - {"avx512-4fmaps",0x00000007, 0, CPUID_REG_EDX, 3, 1}, -+ {"srbds-ctrl", 0x00000007, 0, CPUID_REG_EDX, 9, 1}, - {"md-clear", 0x00000007, 0, CPUID_REG_EDX, 10, 1}, - {"cet-ibt", 0x00000007, 0, CPUID_REG_EDX, 20, 1}, - {"ibrsb", 0x00000007, 0, CPUID_REG_EDX, 26, 1}, -diff --git a/tools/misc/xen-cpuid.c b/tools/misc/xen-cpuid.c -index 603e1d65fd..a09440813b 100644 ---- a/tools/misc/xen-cpuid.c -+++ b/tools/misc/xen-cpuid.c -@@ -157,6 +157,7 @@ static const char *const str_7d0[32] = - [ 2] = "avx512_4vnniw", [ 3] = "avx512_4fmaps", - [ 4] = "fsrm", - -+ /* 8 */ [ 9] = "srbds-ctrl", - [10] = "md-clear", - /* 12 */ [13] = "tsx-force-abort", - -diff --git a/xen/arch/x86/msr.c b/xen/arch/x86/msr.c -index 4b12103482..0cded3c0ad 100644 ---- a/xen/arch/x86/msr.c -+++ b/xen/arch/x86/msr.c -@@ -134,6 +134,7 @@ int guest_rdmsr(struct vcpu *v, uint32_t msr, uint64_t *val) - /* Write-only */ - case MSR_TSX_FORCE_ABORT: - case MSR_TSX_CTRL: -+ case MSR_MCU_OPT_CTRL: - case MSR_U_CET: - case MSR_S_CET: - case MSR_PL0_SSP ... MSR_INTERRUPT_SSP_TABLE: -@@ -288,6 +289,7 @@ int guest_wrmsr(struct vcpu *v, uint32_t msr, uint64_t val) - /* Read-only */ - case MSR_TSX_FORCE_ABORT: - case MSR_TSX_CTRL: -+ case MSR_MCU_OPT_CTRL: - case MSR_U_CET: - case MSR_S_CET: - case MSR_PL0_SSP ... MSR_INTERRUPT_SSP_TABLE: -diff --git a/xen/arch/x86/spec_ctrl.c b/xen/arch/x86/spec_ctrl.c -index 6656c44aec..5fc1c6827e 100644 ---- a/xen/arch/x86/spec_ctrl.c -+++ b/xen/arch/x86/spec_ctrl.c -@@ -312,12 +312,13 @@ static void __init print_details(enum ind_thunk thunk, uint64_t caps) - printk("Speculative mitigation facilities:\n"); - - /* Hardware features which pertain to speculative mitigations. */ -- printk(" Hardware features:%s%s%s%s%s%s%s%s%s%s%s%s%s%s\n", -+ printk(" Hardware features:%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s\n", - (_7d0 & cpufeat_mask(X86_FEATURE_IBRSB)) ? " IBRS/IBPB" : "", - (_7d0 & cpufeat_mask(X86_FEATURE_STIBP)) ? " STIBP" : "", - (_7d0 & cpufeat_mask(X86_FEATURE_L1D_FLUSH)) ? " L1D_FLUSH" : "", - (_7d0 & cpufeat_mask(X86_FEATURE_SSBD)) ? " SSBD" : "", - (_7d0 & cpufeat_mask(X86_FEATURE_MD_CLEAR)) ? " MD_CLEAR" : "", -+ (_7d0 & cpufeat_mask(X86_FEATURE_SRBDS_CTRL)) ? " SRBDS_CTRL" : "", - (e8b & cpufeat_mask(X86_FEATURE_IBPB)) ? " IBPB" : "", - (caps & ARCH_CAPS_IBRS_ALL) ? " IBRS_ALL" : "", - (caps & ARCH_CAPS_RDCL_NO) ? " RDCL_NO" : "", -diff --git a/xen/include/asm-x86/msr-index.h b/xen/include/asm-x86/msr-index.h -index 7693c4a71a..91994669e1 100644 ---- a/xen/include/asm-x86/msr-index.h -+++ b/xen/include/asm-x86/msr-index.h -@@ -179,6 +179,9 @@ - #define MSR_IA32_VMX_TRUE_ENTRY_CTLS 0x490 - #define MSR_IA32_VMX_VMFUNC 0x491 - -+#define MSR_MCU_OPT_CTRL 0x00000123 -+#define MCU_OPT_CTRL_RNGDS_MITG_DIS (_AC(1, ULL) << 0) -+ - #define MSR_U_CET 0x000006a0 - #define MSR_S_CET 0x000006a2 - #define MSR_PL0_SSP 0x000006a4 -diff --git a/xen/include/public/arch-x86/cpufeatureset.h b/xen/include/public/arch-x86/cpufeatureset.h -index 2835688f1c..a2482c3627 100644 ---- a/xen/include/public/arch-x86/cpufeatureset.h -+++ b/xen/include/public/arch-x86/cpufeatureset.h -@@ -252,6 +252,7 @@ XEN_CPUFEATURE(IBPB, 8*32+12) /*A IBPB support only (no IBRS, used by - /* Intel-defined CPU features, CPUID level 0x00000007:0.edx, word 9 */ - XEN_CPUFEATURE(AVX512_4VNNIW, 9*32+ 2) /*A AVX512 Neural Network Instructions */ - XEN_CPUFEATURE(AVX512_4FMAPS, 9*32+ 3) /*A AVX512 Multiply Accumulation Single Precision */ -+XEN_CPUFEATURE(SRBDS_CTRL, 9*32+ 9) /* MSR_MCU_OPT_CTRL and RNGDS_MITG_DIS. */ - XEN_CPUFEATURE(MD_CLEAR, 9*32+10) /*A VERW clears microarchitectural buffers */ - XEN_CPUFEATURE(TSX_FORCE_ABORT, 9*32+13) /* MSR_TSX_FORCE_ABORT.RTM_ABORT */ - XEN_CPUFEATURE(CET_IBT, 9*32+20) /* CET - Indirect Branch Tracking */ diff --git a/xsa320-4.13-2.patch b/xsa320-4.13-2.patch deleted file mode 100644 index 8a8080a..0000000 --- a/xsa320-4.13-2.patch +++ /dev/null @@ -1,179 +0,0 @@ -From: Andrew Cooper -Subject: x86/spec-ctrl: Mitigate the Special Register Buffer Data Sampling sidechannel - -See patch documentation and comments. - -This is part of XSA-320 / CVE-2020-0543 - -Signed-off-by: Andrew Cooper -Reviewed-by: Jan Beulich - -diff --git a/docs/misc/xen-command-line.pandoc b/docs/misc/xen-command-line.pandoc -index 9268454297..c780312531 100644 ---- a/docs/misc/xen-command-line.pandoc -+++ b/docs/misc/xen-command-line.pandoc -@@ -1991,7 +1991,7 @@ By default SSBD will be mitigated at runtime (i.e `ssbd=runtime`). - ### spec-ctrl (x86) - > `= List of [ , xen=, {pv,hvm,msr-sc,rsb,md-clear}=, - > bti-thunk=retpoline|lfence|jmp, {ibrs,ibpb,ssbd,eager-fpu, --> l1d-flush,branch-harden}= ]` -+> l1d-flush,branch-harden,srb-lock}= ]` - - Controls for speculative execution sidechannel mitigations. By default, Xen - will pick the most appropriate mitigations based on compiled in support, -@@ -2068,6 +2068,12 @@ If Xen is compiled with `CONFIG_SPECULATIVE_HARDEN_BRANCH`, the - speculation barriers to protect selected conditional branches. By default, - Xen will enable this mitigation. - -+On hardware supporting SRBDS_CTRL, the `srb-lock=` option can be used to force -+or prevent Xen from protect the Special Register Buffer from leaking stale -+data. By default, Xen will enable this mitigation, except on parts where MDS -+is fixed and TAA is fixed/mitigated (in which case, there is believed to be no -+way for an attacker to obtain the stale data). -+ - ### sync_console - > `= ` - -diff --git a/xen/arch/x86/acpi/power.c b/xen/arch/x86/acpi/power.c -index feb0f6ce20..75c6e34164 100644 ---- a/xen/arch/x86/acpi/power.c -+++ b/xen/arch/x86/acpi/power.c -@@ -295,6 +295,9 @@ static int enter_state(u32 state) - ci->spec_ctrl_flags |= (default_spec_ctrl_flags & SCF_ist_wrmsr); - spec_ctrl_exit_idle(ci); - -+ if ( boot_cpu_has(X86_FEATURE_SRBDS_CTRL) ) -+ wrmsrl(MSR_MCU_OPT_CTRL, default_xen_mcu_opt_ctrl); -+ - done: - spin_debug_enable(); - local_irq_restore(flags); -diff --git a/xen/arch/x86/smpboot.c b/xen/arch/x86/smpboot.c -index dc8fdac1a1..b1e51b3aff 100644 ---- a/xen/arch/x86/smpboot.c -+++ b/xen/arch/x86/smpboot.c -@@ -361,12 +361,14 @@ void start_secondary(void *unused) - microcode_update_one(false); - - /* -- * If MSR_SPEC_CTRL is available, apply Xen's default setting and discard -- * any firmware settings. Note: MSR_SPEC_CTRL may only become available -- * after loading microcode. -+ * If any speculative control MSRs are available, apply Xen's default -+ * settings. Note: These MSRs may only become available after loading -+ * microcode. - */ - if ( boot_cpu_has(X86_FEATURE_IBRSB) ) - wrmsrl(MSR_SPEC_CTRL, default_xen_spec_ctrl); -+ if ( boot_cpu_has(X86_FEATURE_SRBDS_CTRL) ) -+ wrmsrl(MSR_MCU_OPT_CTRL, default_xen_mcu_opt_ctrl); - - tsx_init(); /* Needs microcode. May change HLE/RTM feature bits. */ - -diff --git a/xen/arch/x86/spec_ctrl.c b/xen/arch/x86/spec_ctrl.c -index 5fc1c6827e..33343062a7 100644 ---- a/xen/arch/x86/spec_ctrl.c -+++ b/xen/arch/x86/spec_ctrl.c -@@ -65,6 +65,9 @@ static unsigned int __initdata l1d_maxphysaddr; - static bool __initdata cpu_has_bug_msbds_only; /* => minimal HT impact. */ - static bool __initdata cpu_has_bug_mds; /* Any other M{LP,SB,FB}DS combination. */ - -+static int8_t __initdata opt_srb_lock = -1; -+uint64_t __read_mostly default_xen_mcu_opt_ctrl; -+ - static int __init parse_spec_ctrl(const char *s) - { - const char *ss; -@@ -112,6 +115,7 @@ static int __init parse_spec_ctrl(const char *s) - opt_ssbd = false; - opt_l1d_flush = 0; - opt_branch_harden = false; -+ opt_srb_lock = 0; - } - else if ( val > 0 ) - rc = -EINVAL; -@@ -178,6 +182,8 @@ static int __init parse_spec_ctrl(const char *s) - opt_l1d_flush = val; - else if ( (val = parse_boolean("branch-harden", s, ss)) >= 0 ) - opt_branch_harden = val; -+ else if ( (val = parse_boolean("srb-lock", s, ss)) >= 0 ) -+ opt_srb_lock = val; - else - rc = -EINVAL; - -@@ -341,7 +347,7 @@ static void __init print_details(enum ind_thunk thunk, uint64_t caps) - "\n"); - - /* Settings for Xen's protection, irrespective of guests. */ -- printk(" Xen settings: BTI-Thunk %s, SPEC_CTRL: %s%s%s, Other:%s%s%s%s\n", -+ printk(" Xen settings: BTI-Thunk %s, SPEC_CTRL: %s%s%s, Other:%s%s%s%s%s\n", - thunk == THUNK_NONE ? "N/A" : - thunk == THUNK_RETPOLINE ? "RETPOLINE" : - thunk == THUNK_LFENCE ? "LFENCE" : -@@ -352,6 +358,8 @@ static void __init print_details(enum ind_thunk thunk, uint64_t caps) - (default_xen_spec_ctrl & SPEC_CTRL_SSBD) ? " SSBD+" : " SSBD-", - !(caps & ARCH_CAPS_TSX_CTRL) ? "" : - (opt_tsx & 1) ? " TSX+" : " TSX-", -+ !boot_cpu_has(X86_FEATURE_SRBDS_CTRL) ? "" : -+ opt_srb_lock ? " SRB_LOCK+" : " SRB_LOCK-", - opt_ibpb ? " IBPB" : "", - opt_l1d_flush ? " L1D_FLUSH" : "", - opt_md_clear_pv || opt_md_clear_hvm ? " VERW" : "", -@@ -1149,6 +1157,34 @@ void __init init_speculation_mitigations(void) - tsx_init(); - } - -+ /* Calculate suitable defaults for MSR_MCU_OPT_CTRL */ -+ if ( boot_cpu_has(X86_FEATURE_SRBDS_CTRL) ) -+ { -+ uint64_t val; -+ -+ rdmsrl(MSR_MCU_OPT_CTRL, val); -+ -+ /* -+ * On some SRBDS-affected hardware, it may be safe to relax srb-lock -+ * by default. -+ * -+ * On parts which enumerate MDS_NO and not TAA_NO, TSX is the only way -+ * to access the Fill Buffer. If TSX isn't available (inc. SKU -+ * reasons on some models), or TSX is explicitly disabled, then there -+ * is no need for the extra overhead to protect RDRAND/RDSEED. -+ */ -+ if ( opt_srb_lock == -1 && -+ (caps & (ARCH_CAPS_MDS_NO|ARCH_CAPS_TAA_NO)) == ARCH_CAPS_MDS_NO && -+ (!cpu_has_hle || ((caps & ARCH_CAPS_TSX_CTRL) && opt_tsx == 0)) ) -+ opt_srb_lock = 0; -+ -+ val &= ~MCU_OPT_CTRL_RNGDS_MITG_DIS; -+ if ( !opt_srb_lock ) -+ val |= MCU_OPT_CTRL_RNGDS_MITG_DIS; -+ -+ default_xen_mcu_opt_ctrl = val; -+ } -+ - print_details(thunk, caps); - - /* -@@ -1180,6 +1216,9 @@ void __init init_speculation_mitigations(void) - - wrmsrl(MSR_SPEC_CTRL, bsp_delay_spec_ctrl ? 0 : default_xen_spec_ctrl); - } -+ -+ if ( boot_cpu_has(X86_FEATURE_SRBDS_CTRL) ) -+ wrmsrl(MSR_MCU_OPT_CTRL, default_xen_mcu_opt_ctrl); - } - - static void __init __maybe_unused build_assertions(void) -diff --git a/xen/include/asm-x86/spec_ctrl.h b/xen/include/asm-x86/spec_ctrl.h -index 9caecddfec..b252bb8631 100644 ---- a/xen/include/asm-x86/spec_ctrl.h -+++ b/xen/include/asm-x86/spec_ctrl.h -@@ -54,6 +54,8 @@ extern int8_t opt_pv_l1tf_hwdom, opt_pv_l1tf_domu; - */ - extern paddr_t l1tf_addr_mask, l1tf_safe_maddr; - -+extern uint64_t default_xen_mcu_opt_ctrl; -+ - static inline void init_shadow_spec_ctrl_state(void) - { - struct cpu_info *info = get_cpu_info(); diff --git a/xsa321-4.13-1.patch b/xsa321-4.13-1.patch deleted file mode 100644 index 9a08ab2..0000000 --- a/xsa321-4.13-1.patch +++ /dev/null @@ -1,31 +0,0 @@ -From: Jan Beulich -Subject: vtd: improve IOMMU TLB flush - -Do not limit PSI flushes to order 0 pages, in order to avoid doing a -full TLB flush if the passed in page has an order greater than 0 and -is aligned. Should increase the performance of IOMMU TLB flushes when -dealing with page orders greater than 0. - -This is part of XSA-321. - -Signed-off-by: Jan Beulich - ---- a/xen/drivers/passthrough/vtd/iommu.c -+++ b/xen/drivers/passthrough/vtd/iommu.c -@@ -570,13 +570,14 @@ static int __must_check iommu_flush_iotl - if ( iommu_domid == -1 ) - continue; - -- if ( page_count != 1 || dfn_eq(dfn, INVALID_DFN) ) -+ if ( !page_count || (page_count & (page_count - 1)) || -+ dfn_eq(dfn, INVALID_DFN) || !IS_ALIGNED(dfn_x(dfn), page_count) ) - rc = iommu_flush_iotlb_dsi(iommu, iommu_domid, - 0, flush_dev_iotlb); - else - rc = iommu_flush_iotlb_psi(iommu, iommu_domid, - dfn_to_daddr(dfn), -- PAGE_ORDER_4K, -+ get_order_from_pages(page_count), - !dma_old_pte_present, - flush_dev_iotlb); - diff --git a/xsa321-4.13-2.patch b/xsa321-4.13-2.patch deleted file mode 100644 index 1e48615..0000000 --- a/xsa321-4.13-2.patch +++ /dev/null @@ -1,175 +0,0 @@ -From: -Subject: vtd: prune (and rename) cache flush functions - -Rename __iommu_flush_cache to iommu_sync_cache and remove -iommu_flush_cache_page. Also remove the iommu_flush_cache_entry -wrapper and just use iommu_sync_cache instead. Note the _entry suffix -was meaningless as the wrapper was already taking a size parameter in -bytes. While there also constify the addr parameter. - -No functional change intended. - -This is part of XSA-321. - -Reviewed-by: Jan Beulich - ---- a/xen/drivers/passthrough/vtd/extern.h -+++ b/xen/drivers/passthrough/vtd/extern.h -@@ -43,8 +43,7 @@ void disable_qinval(struct vtd_iommu *io - int enable_intremap(struct vtd_iommu *iommu, int eim); - void disable_intremap(struct vtd_iommu *iommu); - --void iommu_flush_cache_entry(void *addr, unsigned int size); --void iommu_flush_cache_page(void *addr, unsigned long npages); -+void iommu_sync_cache(const void *addr, unsigned int size); - int iommu_alloc(struct acpi_drhd_unit *drhd); - void iommu_free(struct acpi_drhd_unit *drhd); - ---- a/xen/drivers/passthrough/vtd/intremap.c -+++ b/xen/drivers/passthrough/vtd/intremap.c -@@ -230,7 +230,7 @@ static void free_remap_entry(struct vtd_ - iremap_entries, iremap_entry); - - update_irte(iommu, iremap_entry, &new_ire, false); -- iommu_flush_cache_entry(iremap_entry, sizeof(*iremap_entry)); -+ iommu_sync_cache(iremap_entry, sizeof(*iremap_entry)); - iommu_flush_iec_index(iommu, 0, index); - - unmap_vtd_domain_page(iremap_entries); -@@ -406,7 +406,7 @@ static int ioapic_rte_to_remap_entry(str - } - - update_irte(iommu, iremap_entry, &new_ire, !init); -- iommu_flush_cache_entry(iremap_entry, sizeof(*iremap_entry)); -+ iommu_sync_cache(iremap_entry, sizeof(*iremap_entry)); - iommu_flush_iec_index(iommu, 0, index); - - unmap_vtd_domain_page(iremap_entries); -@@ -695,7 +695,7 @@ static int msi_msg_to_remap_entry( - update_irte(iommu, iremap_entry, &new_ire, msi_desc->irte_initialized); - msi_desc->irte_initialized = true; - -- iommu_flush_cache_entry(iremap_entry, sizeof(*iremap_entry)); -+ iommu_sync_cache(iremap_entry, sizeof(*iremap_entry)); - iommu_flush_iec_index(iommu, 0, index); - - unmap_vtd_domain_page(iremap_entries); ---- a/xen/drivers/passthrough/vtd/iommu.c -+++ b/xen/drivers/passthrough/vtd/iommu.c -@@ -140,7 +140,8 @@ static int context_get_domain_id(struct - } - - static int iommus_incoherent; --static void __iommu_flush_cache(void *addr, unsigned int size) -+ -+void iommu_sync_cache(const void *addr, unsigned int size) - { - int i; - static unsigned int clflush_size = 0; -@@ -155,16 +156,6 @@ static void __iommu_flush_cache(void *ad - cacheline_flush((char *)addr + i); - } - --void iommu_flush_cache_entry(void *addr, unsigned int size) --{ -- __iommu_flush_cache(addr, size); --} -- --void iommu_flush_cache_page(void *addr, unsigned long npages) --{ -- __iommu_flush_cache(addr, PAGE_SIZE * npages); --} -- - /* Allocate page table, return its machine address */ - uint64_t alloc_pgtable_maddr(unsigned long npages, nodeid_t node) - { -@@ -183,7 +174,7 @@ uint64_t alloc_pgtable_maddr(unsigned lo - vaddr = __map_domain_page(cur_pg); - memset(vaddr, 0, PAGE_SIZE); - -- iommu_flush_cache_page(vaddr, 1); -+ iommu_sync_cache(vaddr, PAGE_SIZE); - unmap_domain_page(vaddr); - cur_pg++; - } -@@ -216,7 +207,7 @@ static u64 bus_to_context_maddr(struct v - } - set_root_value(*root, maddr); - set_root_present(*root); -- iommu_flush_cache_entry(root, sizeof(struct root_entry)); -+ iommu_sync_cache(root, sizeof(struct root_entry)); - } - maddr = (u64) get_context_addr(*root); - unmap_vtd_domain_page(root_entries); -@@ -263,7 +254,7 @@ static u64 addr_to_dma_page_maddr(struct - */ - dma_set_pte_readable(*pte); - dma_set_pte_writable(*pte); -- iommu_flush_cache_entry(pte, sizeof(struct dma_pte)); -+ iommu_sync_cache(pte, sizeof(struct dma_pte)); - } - - if ( level == 2 ) -@@ -640,7 +631,7 @@ static int __must_check dma_pte_clear_on - *flush_flags |= IOMMU_FLUSHF_modified; - - spin_unlock(&hd->arch.mapping_lock); -- iommu_flush_cache_entry(pte, sizeof(struct dma_pte)); -+ iommu_sync_cache(pte, sizeof(struct dma_pte)); - - unmap_vtd_domain_page(page); - -@@ -679,7 +670,7 @@ static void iommu_free_page_table(struct - iommu_free_pagetable(dma_pte_addr(*pte), next_level); - - dma_clear_pte(*pte); -- iommu_flush_cache_entry(pte, sizeof(struct dma_pte)); -+ iommu_sync_cache(pte, sizeof(struct dma_pte)); - } - - unmap_vtd_domain_page(pt_vaddr); -@@ -1400,7 +1391,7 @@ int domain_context_mapping_one( - context_set_address_width(*context, agaw); - context_set_fault_enable(*context); - context_set_present(*context); -- iommu_flush_cache_entry(context, sizeof(struct context_entry)); -+ iommu_sync_cache(context, sizeof(struct context_entry)); - spin_unlock(&iommu->lock); - - /* Context entry was previously non-present (with domid 0). */ -@@ -1564,7 +1555,7 @@ int domain_context_unmap_one( - - context_clear_present(*context); - context_clear_entry(*context); -- iommu_flush_cache_entry(context, sizeof(struct context_entry)); -+ iommu_sync_cache(context, sizeof(struct context_entry)); - - iommu_domid= domain_iommu_domid(domain, iommu); - if ( iommu_domid == -1 ) -@@ -1791,7 +1782,7 @@ static int __must_check intel_iommu_map_ - - *pte = new; - -- iommu_flush_cache_entry(pte, sizeof(struct dma_pte)); -+ iommu_sync_cache(pte, sizeof(struct dma_pte)); - spin_unlock(&hd->arch.mapping_lock); - unmap_vtd_domain_page(page); - -@@ -1866,7 +1857,7 @@ int iommu_pte_flush(struct domain *d, ui - int iommu_domid; - int rc = 0; - -- iommu_flush_cache_entry(pte, sizeof(struct dma_pte)); -+ iommu_sync_cache(pte, sizeof(struct dma_pte)); - - for_each_drhd_unit ( drhd ) - { -@@ -2724,7 +2715,7 @@ static int __init intel_iommu_quarantine - dma_set_pte_addr(*pte, maddr); - dma_set_pte_readable(*pte); - } -- iommu_flush_cache_page(parent, 1); -+ iommu_sync_cache(parent, PAGE_SIZE); - - unmap_vtd_domain_page(parent); - parent = map_vtd_domain_page(maddr); diff --git a/xsa321-4.13-3.patch b/xsa321-4.13-3.patch deleted file mode 100644 index c141c4b..0000000 --- a/xsa321-4.13-3.patch +++ /dev/null @@ -1,82 +0,0 @@ -From: -Subject: x86/iommu: introduce a cache sync hook - -The hook is only implemented for VT-d and it uses the already existing -iommu_sync_cache function present in VT-d code. The new hook is -added so that the cache can be flushed by code outside of VT-d when -using shared page tables. - -Note that alloc_pgtable_maddr must use the now locally defined -sync_cache function, because IOMMU ops are not yet setup the first -time the function gets called during IOMMU initialization. - -No functional change intended. - -This is part of XSA-321. - -Reviewed-by: Jan Beulich - ---- a/xen/drivers/passthrough/vtd/extern.h -+++ b/xen/drivers/passthrough/vtd/extern.h -@@ -43,7 +43,6 @@ void disable_qinval(struct vtd_iommu *io - int enable_intremap(struct vtd_iommu *iommu, int eim); - void disable_intremap(struct vtd_iommu *iommu); - --void iommu_sync_cache(const void *addr, unsigned int size); - int iommu_alloc(struct acpi_drhd_unit *drhd); - void iommu_free(struct acpi_drhd_unit *drhd); - ---- a/xen/drivers/passthrough/vtd/iommu.c -+++ b/xen/drivers/passthrough/vtd/iommu.c -@@ -141,7 +141,7 @@ static int context_get_domain_id(struct - - static int iommus_incoherent; - --void iommu_sync_cache(const void *addr, unsigned int size) -+static void sync_cache(const void *addr, unsigned int size) - { - int i; - static unsigned int clflush_size = 0; -@@ -174,7 +174,7 @@ uint64_t alloc_pgtable_maddr(unsigned lo - vaddr = __map_domain_page(cur_pg); - memset(vaddr, 0, PAGE_SIZE); - -- iommu_sync_cache(vaddr, PAGE_SIZE); -+ sync_cache(vaddr, PAGE_SIZE); - unmap_domain_page(vaddr); - cur_pg++; - } -@@ -2763,6 +2763,7 @@ const struct iommu_ops __initconstrel in - .iotlb_flush_all = iommu_flush_iotlb_all, - .get_reserved_device_memory = intel_iommu_get_reserved_device_memory, - .dump_p2m_table = vtd_dump_p2m_table, -+ .sync_cache = sync_cache, - }; - - const struct iommu_init_ops __initconstrel intel_iommu_init_ops = { ---- a/xen/include/asm-x86/iommu.h -+++ b/xen/include/asm-x86/iommu.h -@@ -121,6 +121,13 @@ extern bool untrusted_msi; - int pi_update_irte(const struct pi_desc *pi_desc, const struct pirq *pirq, - const uint8_t gvec); - -+#define iommu_sync_cache(addr, size) ({ \ -+ const struct iommu_ops *ops = iommu_get_ops(); \ -+ \ -+ if ( ops->sync_cache ) \ -+ iommu_vcall(ops, sync_cache, addr, size); \ -+}) -+ - #endif /* !__ARCH_X86_IOMMU_H__ */ - /* - * Local variables: ---- a/xen/include/xen/iommu.h -+++ b/xen/include/xen/iommu.h -@@ -250,6 +250,7 @@ struct iommu_ops { - int (*setup_hpet_msi)(struct msi_desc *); - - int (*adjust_irq_affinities)(void); -+ void (*sync_cache)(const void *addr, unsigned int size); - #endif /* CONFIG_X86 */ - - int __must_check (*suspend)(void); diff --git a/xsa321-4.13-4.patch b/xsa321-4.13-4.patch deleted file mode 100644 index 62bbcc7..0000000 --- a/xsa321-4.13-4.patch +++ /dev/null @@ -1,36 +0,0 @@ -From: -Subject: vtd: don't assume addresses are aligned in sync_cache - -Current code in sync_cache assume that the address passed in is -aligned to a cache line size. Fix the code to support passing in -arbitrary addresses not necessarily aligned to a cache line size. - -This is part of XSA-321. - -Reviewed-by: Jan Beulich - ---- a/xen/drivers/passthrough/vtd/iommu.c -+++ b/xen/drivers/passthrough/vtd/iommu.c -@@ -143,8 +143,8 @@ static int iommus_incoherent; - - static void sync_cache(const void *addr, unsigned int size) - { -- int i; -- static unsigned int clflush_size = 0; -+ static unsigned long clflush_size = 0; -+ const void *end = addr + size; - - if ( !iommus_incoherent ) - return; -@@ -152,8 +152,9 @@ static void sync_cache(const void *addr, - if ( clflush_size == 0 ) - clflush_size = get_cache_line_size(); - -- for ( i = 0; i < size; i += clflush_size ) -- cacheline_flush((char *)addr + i); -+ addr -= (unsigned long)addr & (clflush_size - 1); -+ for ( ; addr < end; addr += clflush_size ) -+ cacheline_flush((char *)addr); - } - - /* Allocate page table, return its machine address */ diff --git a/xsa321-4.13-5.patch b/xsa321-4.13-5.patch deleted file mode 100644 index 60cfe6c..0000000 --- a/xsa321-4.13-5.patch +++ /dev/null @@ -1,24 +0,0 @@ -From: -Subject: x86/alternative: introduce alternative_2 - -It's based on alternative_io_2 without inputs or outputs but with an -added memory clobber. - -This is part of XSA-321. - -Acked-by: Jan Beulich - ---- a/xen/include/asm-x86/alternative.h -+++ b/xen/include/asm-x86/alternative.h -@@ -114,6 +114,11 @@ extern void alternative_branches(void); - #define alternative(oldinstr, newinstr, feature) \ - asm volatile (ALTERNATIVE(oldinstr, newinstr, feature) : : : "memory") - -+#define alternative_2(oldinstr, newinstr1, feature1, newinstr2, feature2) \ -+ asm volatile (ALTERNATIVE_2(oldinstr, newinstr1, feature1, \ -+ newinstr2, feature2) \ -+ : : : "memory") -+ - /* - * Alternative inline assembly with input. - * diff --git a/xsa321-4.13-6.patch b/xsa321-4.13-6.patch deleted file mode 100644 index 4c5c5ab..0000000 --- a/xsa321-4.13-6.patch +++ /dev/null @@ -1,91 +0,0 @@ -From: -Subject: vtd: optimize CPU cache sync - -Some VT-d IOMMUs are non-coherent, which requires a cache write back -in order for the changes made by the CPU to be visible to the IOMMU. -This cache write back was unconditionally done using clflush, but there are -other more efficient instructions to do so, hence implement support -for them using the alternative framework. - -This is part of XSA-321. - -Reviewed-by: Jan Beulich - ---- a/xen/drivers/passthrough/vtd/extern.h -+++ b/xen/drivers/passthrough/vtd/extern.h -@@ -68,7 +68,6 @@ int __must_check qinval_device_iotlb_syn - u16 did, u16 size, u64 addr); - - unsigned int get_cache_line_size(void); --void cacheline_flush(char *); - void flush_all_cache(void); - - uint64_t alloc_pgtable_maddr(unsigned long npages, nodeid_t node); ---- a/xen/drivers/passthrough/vtd/iommu.c -+++ b/xen/drivers/passthrough/vtd/iommu.c -@@ -31,6 +31,7 @@ - #include - #include - #include -+#include - #include - #include - #include -@@ -154,7 +155,42 @@ static void sync_cache(const void *addr, - - addr -= (unsigned long)addr & (clflush_size - 1); - for ( ; addr < end; addr += clflush_size ) -- cacheline_flush((char *)addr); -+/* -+ * The arguments to a macro must not include preprocessor directives. Doing so -+ * results in undefined behavior, so we have to create some defines here in -+ * order to avoid it. -+ */ -+#if defined(HAVE_AS_CLWB) -+# define CLWB_ENCODING "clwb %[p]" -+#elif defined(HAVE_AS_XSAVEOPT) -+# define CLWB_ENCODING "data16 xsaveopt %[p]" /* clwb */ -+#else -+# define CLWB_ENCODING ".byte 0x66, 0x0f, 0xae, 0x30" /* clwb (%%rax) */ -+#endif -+ -+#define BASE_INPUT(addr) [p] "m" (*(const char *)(addr)) -+#if defined(HAVE_AS_CLWB) || defined(HAVE_AS_XSAVEOPT) -+# define INPUT BASE_INPUT -+#else -+# define INPUT(addr) "a" (addr), BASE_INPUT(addr) -+#endif -+ /* -+ * Note regarding the use of NOP_DS_PREFIX: it's faster to do a clflush -+ * + prefix than a clflush + nop, and hence the prefix is added instead -+ * of letting the alternative framework fill the gap by appending nops. -+ */ -+ alternative_io_2(".byte " __stringify(NOP_DS_PREFIX) "; clflush %[p]", -+ "data16 clflush %[p]", /* clflushopt */ -+ X86_FEATURE_CLFLUSHOPT, -+ CLWB_ENCODING, -+ X86_FEATURE_CLWB, /* no outputs */, -+ INPUT(addr)); -+#undef INPUT -+#undef BASE_INPUT -+#undef CLWB_ENCODING -+ -+ alternative_2("", "sfence", X86_FEATURE_CLFLUSHOPT, -+ "sfence", X86_FEATURE_CLWB); - } - - /* Allocate page table, return its machine address */ ---- a/xen/drivers/passthrough/vtd/x86/vtd.c -+++ b/xen/drivers/passthrough/vtd/x86/vtd.c -@@ -51,11 +51,6 @@ unsigned int get_cache_line_size(void) - return ((cpuid_ebx(1) >> 8) & 0xff) * 8; - } - --void cacheline_flush(char * addr) --{ -- clflush(addr); --} -- - void flush_all_cache() - { - wbinvd(); diff --git a/xsa321-4.13-7.patch b/xsa321-4.13-7.patch deleted file mode 100644 index 0bd018f..0000000 --- a/xsa321-4.13-7.patch +++ /dev/null @@ -1,153 +0,0 @@ -From: -Subject: x86/ept: flush cache when modifying PTEs and sharing page tables - -Modifications made to the page tables by EPT code need to be written -to memory when the page tables are shared with the IOMMU, as Intel -IOMMUs can be non-coherent and thus require changes to be written to -memory in order to be visible to the IOMMU. - -In order to achieve this make sure data is written back to memory -after writing an EPT entry when the recalc bit is not set in -atomic_write_ept_entry. If such bit is set, the entry will be -adjusted and atomic_write_ept_entry will be called a second time -without the recalc bit set. Note that when splitting a super page the -new tables resulting of the split should also be written back. - -Failure to do so can allow devices behind the IOMMU access to the -stale super page, or cause coherency issues as changes made by the -processor to the page tables are not visible to the IOMMU. - -This allows to remove the VT-d specific iommu_pte_flush helper, since -the cache write back is now performed by atomic_write_ept_entry, and -hence iommu_iotlb_flush can be used to flush the IOMMU TLB. The newly -used method (iommu_iotlb_flush) can result in less flushes, since it -might sometimes be called rightly with 0 flags, in which case it -becomes a no-op. - -This is part of XSA-321. - -Reviewed-by: Jan Beulich - ---- a/xen/arch/x86/mm/p2m-ept.c -+++ b/xen/arch/x86/mm/p2m-ept.c -@@ -58,6 +58,19 @@ static int atomic_write_ept_entry(struct - - write_atomic(&entryptr->epte, new.epte); - -+ /* -+ * The recalc field on the EPT is used to signal either that a -+ * recalculation of the EMT field is required (which doesn't effect the -+ * IOMMU), or a type change. Type changes can only be between ram_rw, -+ * logdirty and ioreq_server: changes to/from logdirty won't work well with -+ * an IOMMU anyway, as IOMMU #PFs are not synchronous and will lead to -+ * aborts, and changes to/from ioreq_server are already fully flushed -+ * before returning to guest context (see -+ * XEN_DMOP_map_mem_type_to_ioreq_server). -+ */ -+ if ( !new.recalc && iommu_use_hap_pt(p2m->domain) ) -+ iommu_sync_cache(entryptr, sizeof(*entryptr)); -+ - return 0; - } - -@@ -278,6 +291,9 @@ static bool_t ept_split_super_page(struc - break; - } - -+ if ( iommu_use_hap_pt(p2m->domain) ) -+ iommu_sync_cache(table, EPT_PAGETABLE_ENTRIES * sizeof(ept_entry_t)); -+ - unmap_domain_page(table); - - /* Even failed we should install the newly allocated ept page. */ -@@ -337,6 +353,9 @@ static int ept_next_level(struct p2m_dom - if ( !next ) - return GUEST_TABLE_MAP_FAILED; - -+ if ( iommu_use_hap_pt(p2m->domain) ) -+ iommu_sync_cache(next, EPT_PAGETABLE_ENTRIES * sizeof(ept_entry_t)); -+ - rc = atomic_write_ept_entry(p2m, ept_entry, e, next_level); - ASSERT(rc == 0); - } -@@ -821,7 +840,10 @@ out: - need_modify_vtd_table ) - { - if ( iommu_use_hap_pt(d) ) -- rc = iommu_pte_flush(d, gfn, &ept_entry->epte, order, vtd_pte_present); -+ rc = iommu_iotlb_flush(d, _dfn(gfn), (1u << order), -+ (iommu_flags ? IOMMU_FLUSHF_added : 0) | -+ (vtd_pte_present ? IOMMU_FLUSHF_modified -+ : 0)); - else if ( need_iommu_pt_sync(d) ) - rc = iommu_flags ? - iommu_legacy_map(d, _dfn(gfn), mfn, order, iommu_flags) : ---- a/xen/drivers/passthrough/vtd/iommu.c -+++ b/xen/drivers/passthrough/vtd/iommu.c -@@ -1884,53 +1884,6 @@ static int intel_iommu_lookup_page(struc - return 0; - } - --int iommu_pte_flush(struct domain *d, uint64_t dfn, uint64_t *pte, -- int order, int present) --{ -- struct acpi_drhd_unit *drhd; -- struct vtd_iommu *iommu = NULL; -- struct domain_iommu *hd = dom_iommu(d); -- bool_t flush_dev_iotlb; -- int iommu_domid; -- int rc = 0; -- -- iommu_sync_cache(pte, sizeof(struct dma_pte)); -- -- for_each_drhd_unit ( drhd ) -- { -- iommu = drhd->iommu; -- if ( !test_bit(iommu->index, &hd->arch.iommu_bitmap) ) -- continue; -- -- flush_dev_iotlb = !!find_ats_dev_drhd(iommu); -- iommu_domid= domain_iommu_domid(d, iommu); -- if ( iommu_domid == -1 ) -- continue; -- -- rc = iommu_flush_iotlb_psi(iommu, iommu_domid, -- __dfn_to_daddr(dfn), -- order, !present, flush_dev_iotlb); -- if ( rc > 0 ) -- { -- iommu_flush_write_buffer(iommu); -- rc = 0; -- } -- } -- -- if ( unlikely(rc) ) -- { -- if ( !d->is_shutting_down && printk_ratelimit() ) -- printk(XENLOG_ERR VTDPREFIX -- " d%d: IOMMU pages flush failed: %d\n", -- d->domain_id, rc); -- -- if ( !is_hardware_domain(d) ) -- domain_crash(d); -- } -- -- return rc; --} -- - static int __init vtd_ept_page_compatible(struct vtd_iommu *iommu) - { - u64 ept_cap, vtd_cap = iommu->cap; ---- a/xen/include/asm-x86/iommu.h -+++ b/xen/include/asm-x86/iommu.h -@@ -97,10 +97,6 @@ static inline int iommu_adjust_irq_affin - : 0; - } - --/* While VT-d specific, this must get declared in a generic header. */ --int __must_check iommu_pte_flush(struct domain *d, u64 gfn, u64 *pte, -- int order, int present); -- - static inline bool iommu_supports_x2apic(void) - { - return iommu_init_ops && iommu_init_ops->supports_x2apic diff --git a/xsa327.patch b/xsa327.patch deleted file mode 100644 index 0541cfa..0000000 --- a/xsa327.patch +++ /dev/null @@ -1,63 +0,0 @@ -From 030300ebbb86c40c12db038714479d746167c767 Mon Sep 17 00:00:00 2001 -From: Julien Grall -Date: Tue, 26 May 2020 18:31:33 +0100 -Subject: [PATCH] xen: Check the alignment of the offset pased via - VCPUOP_register_vcpu_info - -Currently a guest is able to register any guest physical address to use -for the vcpu_info structure as long as the structure can fits in the -rest of the frame. - -This means a guest can provide an address that is not aligned to the -natural alignment of the structure. - -On Arm 32-bit, unaligned access are completely forbidden by the -hypervisor. This will result to a data abort which is fatal. - -On Arm 64-bit, unaligned access are only forbidden when used for atomic -access. As the structure contains fields (such as evtchn_pending_self) -that are updated using atomic operations, any unaligned access will be -fatal as well. - -While the misalignment is only fatal on Arm, a generic check is added -as an x86 guest shouldn't sensibly pass an unaligned address (this -would result to a split lock). - -This is XSA-327. - -Reported-by: Julien Grall -Signed-off-by: Julien Grall -Reviewed-by: Andrew Cooper -Reviewed-by: Stefano Stabellini ---- - xen/common/domain.c | 10 ++++++++++ - 1 file changed, 10 insertions(+) - -diff --git a/xen/common/domain.c b/xen/common/domain.c -index 7cc9526139a6..e9be05f1d05f 100644 ---- a/xen/common/domain.c -+++ b/xen/common/domain.c -@@ -1227,10 +1227,20 @@ int map_vcpu_info(struct vcpu *v, unsigned long gfn, unsigned offset) - void *mapping; - vcpu_info_t *new_info; - struct page_info *page; -+ unsigned int align; - - if ( offset > (PAGE_SIZE - sizeof(vcpu_info_t)) ) - return -EINVAL; - -+#ifdef CONFIG_COMPAT -+ if ( has_32bit_shinfo(d) ) -+ align = alignof(new_info->compat); -+ else -+#endif -+ align = alignof(*new_info); -+ if ( offset & (align - 1) ) -+ return -EINVAL; -+ - if ( !mfn_eq(v->vcpu_info_mfn, INVALID_MFN) ) - return -EINVAL; - --- -2.17.1 - diff --git a/xsa328-4.13-1.patch b/xsa328-4.13-1.patch deleted file mode 100644 index 56e48de..0000000 --- a/xsa328-4.13-1.patch +++ /dev/null @@ -1,118 +0,0 @@ -From: Jan Beulich -Subject: x86/EPT: ept_set_middle_entry() related adjustments - -ept_split_super_page() wants to further modify the newly allocated -table, so have ept_set_middle_entry() return the mapped pointer rather -than tearing it down and then getting re-established right again. - -Similarly ept_next_level() wants to hand back a mapped pointer of -the next level page, so re-use the one established by -ept_set_middle_entry() in case that path was taken. - -Pull the setting of suppress_ve ahead of insertion into the higher level -table, and don't have ept_split_super_page() set the field a 2nd time. - -This is part of XSA-328. - -Signed-off-by: Jan Beulich - ---- a/xen/arch/x86/mm/p2m-ept.c -+++ b/xen/arch/x86/mm/p2m-ept.c -@@ -187,8 +187,9 @@ static void ept_p2m_type_to_flags(struct - #define GUEST_TABLE_SUPER_PAGE 2 - #define GUEST_TABLE_POD_PAGE 3 - --/* Fill in middle levels of ept table */ --static int ept_set_middle_entry(struct p2m_domain *p2m, ept_entry_t *ept_entry) -+/* Fill in middle level of ept table; return pointer to mapped new table. */ -+static ept_entry_t *ept_set_middle_entry(struct p2m_domain *p2m, -+ ept_entry_t *ept_entry) - { - mfn_t mfn; - ept_entry_t *table; -@@ -196,7 +197,12 @@ static int ept_set_middle_entry(struct p - - mfn = p2m_alloc_ptp(p2m, 0); - if ( mfn_eq(mfn, INVALID_MFN) ) -- return 0; -+ return NULL; -+ -+ table = map_domain_page(mfn); -+ -+ for ( i = 0; i < EPT_PAGETABLE_ENTRIES; i++ ) -+ table[i].suppress_ve = 1; - - ept_entry->epte = 0; - ept_entry->mfn = mfn_x(mfn); -@@ -208,14 +214,7 @@ static int ept_set_middle_entry(struct p - - ept_entry->suppress_ve = 1; - -- table = map_domain_page(mfn); -- -- for ( i = 0; i < EPT_PAGETABLE_ENTRIES; i++ ) -- table[i].suppress_ve = 1; -- -- unmap_domain_page(table); -- -- return 1; -+ return table; - } - - /* free ept sub tree behind an entry */ -@@ -253,10 +252,10 @@ static bool_t ept_split_super_page(struc - - ASSERT(is_epte_superpage(ept_entry)); - -- if ( !ept_set_middle_entry(p2m, &new_ept) ) -+ table = ept_set_middle_entry(p2m, &new_ept); -+ if ( !table ) - return 0; - -- table = map_domain_page(_mfn(new_ept.mfn)); - trunk = 1UL << ((level - 1) * EPT_TABLE_ORDER); - - for ( i = 0; i < EPT_PAGETABLE_ENTRIES; i++ ) -@@ -267,7 +266,6 @@ static bool_t ept_split_super_page(struc - epte->sp = (level > 1); - epte->mfn += i * trunk; - epte->snp = is_iommu_enabled(p2m->domain) && iommu_snoop; -- epte->suppress_ve = 1; - - ept_p2m_type_to_flags(p2m, epte, epte->sa_p2mt, epte->access); - -@@ -306,8 +304,7 @@ static int ept_next_level(struct p2m_dom - ept_entry_t **table, unsigned long *gfn_remainder, - int next_level) - { -- unsigned long mfn; -- ept_entry_t *ept_entry, e; -+ ept_entry_t *ept_entry, *next = NULL, e; - u32 shift, index; - - shift = next_level * EPT_TABLE_ORDER; -@@ -332,19 +329,17 @@ static int ept_next_level(struct p2m_dom - if ( read_only ) - return GUEST_TABLE_MAP_FAILED; - -- if ( !ept_set_middle_entry(p2m, ept_entry) ) -+ next = ept_set_middle_entry(p2m, ept_entry); -+ if ( !next ) - return GUEST_TABLE_MAP_FAILED; -- else -- e = atomic_read_ept_entry(ept_entry); /* Refresh */ -+ /* e is now stale and hence may not be used anymore below. */ - } -- - /* The only time sp would be set here is if we had hit a superpage */ -- if ( is_epte_superpage(&e) ) -+ else if ( is_epte_superpage(&e) ) - return GUEST_TABLE_SUPER_PAGE; - -- mfn = e.mfn; - unmap_domain_page(*table); -- *table = map_domain_page(_mfn(mfn)); -+ *table = next ?: map_domain_page(_mfn(e.mfn)); - *gfn_remainder &= (1UL << shift) - 1; - return GUEST_TABLE_NORMAL_PAGE; - } diff --git a/xsa328-4.13-2.patch b/xsa328-4.13-2.patch deleted file mode 100644 index c4f437f..0000000 --- a/xsa328-4.13-2.patch +++ /dev/null @@ -1,48 +0,0 @@ -From: -Subject: x86/ept: atomically modify entries in ept_next_level - -ept_next_level was passing a live PTE pointer to ept_set_middle_entry, -which was then modified without taking into account that the PTE could -be part of a live EPT table. This wasn't a security issue because the -pages returned by p2m_alloc_ptp are zeroed, so adding such an entry -before actually initializing it didn't allow a guest to access -physical memory addresses it wasn't supposed to access. - -This is part of XSA-328. - -Reviewed-by: Jan Beulich - ---- a/xen/arch/x86/mm/p2m-ept.c -+++ b/xen/arch/x86/mm/p2m-ept.c -@@ -307,6 +307,8 @@ static int ept_next_level(struct p2m_dom - ept_entry_t *ept_entry, *next = NULL, e; - u32 shift, index; - -+ ASSERT(next_level); -+ - shift = next_level * EPT_TABLE_ORDER; - - index = *gfn_remainder >> shift; -@@ -323,16 +325,20 @@ static int ept_next_level(struct p2m_dom - - if ( !is_epte_present(&e) ) - { -+ int rc; -+ - if ( e.sa_p2mt == p2m_populate_on_demand ) - return GUEST_TABLE_POD_PAGE; - - if ( read_only ) - return GUEST_TABLE_MAP_FAILED; - -- next = ept_set_middle_entry(p2m, ept_entry); -+ next = ept_set_middle_entry(p2m, &e); - if ( !next ) - return GUEST_TABLE_MAP_FAILED; -- /* e is now stale and hence may not be used anymore below. */ -+ -+ rc = atomic_write_ept_entry(p2m, ept_entry, e, next_level); -+ ASSERT(rc == 0); - } - /* The only time sp would be set here is if we had hit a superpage */ - else if ( is_epte_superpage(&e) ) diff --git a/xsa363.patch b/xsa363.patch new file mode 100644 index 0000000..c8a3de3 --- /dev/null +++ b/xsa363.patch @@ -0,0 +1,22 @@ +From: Jan Beulich +Subject: SUPPORT.md: PV display frontend is unsupported in "backend allocation" mode + +This wasn't meant to be supported, but wasn't stated this way. + +This is XSA-363. + +Reported-by: Jan Belich +Signed-off-by: Jan Beulich + +--- a/SUPPORT.md ++++ b/SUPPORT.md +@@ -414,7 +414,8 @@ Guest-side driver capable of speaking th + + Guest-side driver capable of speaking the Xen PV display protocol + +- Status, Linux: Supported ++ Status, Linux: Supported (outside of "backend allocation" mode) ++ Status, Linux: Experimental (in "backend allocation" mode) + + ### PV Console (frontend) + diff --git a/zstd-dom0.patch b/zstd-dom0.patch new file mode 100644 index 0000000..57b7f76 --- /dev/null +++ b/zstd-dom0.patch @@ -0,0 +1,9214 @@ +diff --git a/xen/common/Makefile b/xen/common/Makefile +index d109f279a4..5ba09f04ac 100644 +--- a/xen/common/Makefile ++++ b/xen/common/Makefile +@@ -59,7 +59,7 @@ obj-bin-y += warning.init.o + obj-$(CONFIG_XENOPROF) += xenoprof.o + obj-y += xmalloc_tlsf.o + +-obj-bin-$(CONFIG_X86) += $(foreach n,decompress bunzip2 unxz unlzma lzo unlzo unlz4 earlycpio,$(n).init.o) ++obj-bin-$(CONFIG_X86) += $(foreach n,decompress bunzip2 unxz unlzma lzo unlzo unlz4 unzstd earlycpio,$(n).init.o) + + obj-$(CONFIG_COMPAT) += $(addprefix compat/,domain.o kernel.o memory.o multicall.o xlat.o) + +diff --git a/xen/common/decompress.c b/xen/common/decompress.c +index 9d6e0c4ab0..0da27b0ab6 100644 +--- a/xen/common/decompress.c ++++ b/xen/common/decompress.c +@@ -31,5 +31,8 @@ int __init decompress(void *inbuf, unsigned int len, void *outbuf) + if ( len >= 2 && !memcmp(inbuf, "\x02\x21", 2) ) + return unlz4(inbuf, len, NULL, NULL, outbuf, NULL, error); + ++ if ( len >= 4 && !memcmp(inbuf, "\050\265\057\375", 4) ) ++ return unzstd(inbuf, len, NULL, NULL, outbuf, NULL, error); ++ + return 1; + } +diff --git a/xen/common/unzstd.c b/xen/common/unzstd.c +new file mode 100644 +index 0000000000..a2c382fddc +--- /dev/null ++++ b/xen/common/unzstd.c +@@ -0,0 +1,332 @@ ++/* ++ * Important notes about in-place decompression ++ * ++ * At least on x86, the kernel is decompressed in place: the compressed data ++ * is placed to the end of the output buffer, and the decompressor overwrites ++ * most of the compressed data. There must be enough safety margin to ++ * guarantee that the write position is always behind the read position. ++ * ++ * The safety margin for ZSTD with a 128 KB block size is calculated below. ++ * Note that the margin with ZSTD is bigger than with GZIP or XZ! ++ * ++ * The worst case for in-place decompression is that the beginning of ++ * the file is compressed extremely well, and the rest of the file is ++ * uncompressible. Thus, we must look for worst-case expansion when the ++ * compressor is encoding uncompressible data. ++ * ++ * The structure of the .zst file in case of a compresed kernel is as follows. ++ * Maximum sizes (as bytes) of the fields are in parenthesis. ++ * ++ * Frame Header: (18) ++ * Blocks: (N) ++ * Checksum: (4) ++ * ++ * The frame header and checksum overhead is at most 22 bytes. ++ * ++ * ZSTD stores the data in blocks. Each block has a header whose size is ++ * a 3 bytes. After the block header, there is up to 128 KB of payload. ++ * The maximum uncompressed size of the payload is 128 KB. The minimum ++ * uncompressed size of the payload is never less than the payload size ++ * (excluding the block header). ++ * ++ * The assumption, that the uncompressed size of the payload is never ++ * smaller than the payload itself, is valid only when talking about ++ * the payload as a whole. It is possible that the payload has parts where ++ * the decompressor consumes more input than it produces output. Calculating ++ * the worst case for this would be tricky. Instead of trying to do that, ++ * let's simply make sure that the decompressor never overwrites any bytes ++ * of the payload which it is currently reading. ++ * ++ * Now we have enough information to calculate the safety margin. We need ++ * - 22 bytes for the .zst file format headers; ++ * - 3 bytes per every 128 KiB of uncompressed size (one block header per ++ * block); and ++ * - 128 KiB (biggest possible zstd block size) to make sure that the ++ * decompressor never overwrites anything from the block it is currently ++ * reading. ++ * ++ * We get the following formula: ++ * ++ * safety_margin = 22 + uncompressed_size * 3 / 131072 + 131072 ++ * <= 22 + (uncompressed_size >> 15) + 131072 ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License version 2 as ++ * published by the Free Software Foundation. ++ */ ++ ++/* ++ * Preboot environments #include "path/to/decompress_unzstd.c". ++ * All of the source files we depend on must be #included. ++ * zstd's only source dependeny is xxhash, which has no source ++ * dependencies. ++ * ++ * When UNZSTD_PREBOOT is defined we declare __decompress(), which is ++ * used for kernel decompression, instead of unzstd(). ++ * ++ * Define __DISABLE_EXPORTS in preboot environments to prevent symbols ++ * from xxhash and zstd from being exported by the EXPORT_SYMBOL macro. ++ */ ++ ++#include "decompress.h" ++#include "xxhash.c" ++#include "zstd/entropy_common.c" ++#include "zstd/fse_decompress.c" ++#include "zstd/huf_decompress.c" ++#include "zstd/zstd_common.c" ++#include "zstd/decompress.c" ++ ++#include ++ ++/* 128MB is the maximum window size supported by zstd. */ ++#define ZSTD_WINDOWSIZE_MAX (1 << ZSTD_WINDOWLOG_MAX) ++/* ++ * Size of the input and output buffers in multi-call mode. ++ * Pick a larger size because it isn't used during kernel decompression, ++ * since that is single pass, and we have to allocate a large buffer for ++ * zstd's window anyway. The larger size speeds up initramfs decompression. ++ */ ++#define ZSTD_IOBUF_SIZE (1 << 17) ++ ++static int INIT handle_zstd_error(size_t ret, void (*error)(const char *x)) ++{ ++ const int err = ZSTD_getErrorCode(ret); ++ ++ if (!ZSTD_isError(ret)) ++ return 0; ++ ++ switch (err) { ++ case ZSTD_error_memory_allocation: ++ error("ZSTD decompressor ran out of memory"); ++ break; ++ case ZSTD_error_prefix_unknown: ++ error("Input is not in the ZSTD format (wrong magic bytes)"); ++ break; ++ case ZSTD_error_dstSize_tooSmall: ++ case ZSTD_error_corruption_detected: ++ case ZSTD_error_checksum_wrong: ++ error("ZSTD-compressed data is corrupt"); ++ break; ++ default: ++ error("ZSTD-compressed data is probably corrupt"); ++ break; ++ } ++ return -1; ++} ++ ++/* ++ * Handle the case where we have the entire input and output in one segment. ++ * We can allocate less memory (no circular buffer for the sliding window), ++ * and avoid some memcpy() calls. ++ */ ++static int INIT decompress_single(const u8 *in_buf, unsigned int in_len, u8 *out_buf, ++ long out_len, unsigned int *in_pos, ++ void (*error)(const char *x)) ++{ ++ const size_t wksp_size = ZSTD_DCtxWorkspaceBound(); ++ void *wksp = large_malloc(wksp_size); ++ ZSTD_DCtx *dctx = ZSTD_initDCtx(wksp, wksp_size); ++ int err; ++ size_t ret; ++ ++ if (dctx == NULL) { ++ error("Out of memory while allocating ZSTD_DCtx"); ++ err = -1; ++ goto out; ++ } ++ /* ++ * Find out how large the frame actually is, there may be junk at ++ * the end of the frame that ZSTD_decompressDCtx() can't handle. ++ */ ++ ret = ZSTD_findFrameCompressedSize(in_buf, in_len); ++ err = handle_zstd_error(ret, error); ++ if (err) ++ goto out; ++ in_len = (long)ret; ++ ++ ret = ZSTD_decompressDCtx(dctx, out_buf, out_len, in_buf, in_len); ++ err = handle_zstd_error(ret, error); ++ if (err) ++ goto out; ++ ++ if (in_pos != NULL) ++ *in_pos = in_len; ++ ++ err = 0; ++out: ++ if (wksp != NULL) ++ large_free(wksp); ++ return err; ++} ++ ++static int INIT __unzstd(unsigned char *in_buf, unsigned int in_len, ++ int (*fill)(void*, unsigned int), ++ int (*flush)(void*, unsigned int), ++ unsigned char *out_buf, long out_len, ++ unsigned int *in_pos, ++ void (*error)(const char *x)) ++{ ++ ZSTD_inBuffer in; ++ ZSTD_outBuffer out; ++ ZSTD_frameParams params; ++ void *in_allocated = NULL; ++ void *out_allocated = NULL; ++ void *wksp = NULL; ++ size_t wksp_size; ++ ZSTD_DStream *dstream; ++ int err; ++ size_t ret; ++ ++ if (out_len == 0) ++ out_len = INT_MAX; /* no limit */ ++ ++ if (fill == NULL && flush == NULL) ++ /* ++ * We can decompress faster and with less memory when we have a ++ * single chunk. ++ */ ++ return decompress_single(in_buf, in_len, out_buf, out_len, ++ in_pos, error); ++ ++ /* ++ * If in_buf is not provided, we must be using fill(), so allocate ++ * a large enough buffer. If it is provided, it must be at least ++ * ZSTD_IOBUF_SIZE large. ++ */ ++ if (in_buf == NULL) { ++ in_allocated = large_malloc(ZSTD_IOBUF_SIZE); ++ if (in_allocated == NULL) { ++ error("Out of memory while allocating input buffer"); ++ err = -1; ++ goto out; ++ } ++ in_buf = in_allocated; ++ in_len = 0; ++ } ++ /* Read the first chunk, since we need to decode the frame header. */ ++ if (fill != NULL) ++ in_len = fill(in_buf, ZSTD_IOBUF_SIZE); ++ if (in_len < 0) { ++ error("ZSTD-compressed data is truncated"); ++ err = -1; ++ goto out; ++ } ++ /* Set the first non-empty input buffer. */ ++ in.src = in_buf; ++ in.pos = 0; ++ in.size = in_len; ++ /* Allocate the output buffer if we are using flush(). */ ++ if (flush != NULL) { ++ out_allocated = large_malloc(ZSTD_IOBUF_SIZE); ++ if (out_allocated == NULL) { ++ error("Out of memory while allocating output buffer"); ++ err = -1; ++ goto out; ++ } ++ out_buf = out_allocated; ++ out_len = ZSTD_IOBUF_SIZE; ++ } ++ /* Set the output buffer. */ ++ out.dst = out_buf; ++ out.pos = 0; ++ out.size = out_len; ++ ++ /* ++ * We need to know the window size to allocate the ZSTD_DStream. ++ * Since we are streaming, we need to allocate a buffer for the sliding ++ * window. The window size varies from 1 KB to ZSTD_WINDOWSIZE_MAX ++ * (8 MB), so it is important to use the actual value so as not to ++ * waste memory when it is smaller. ++ */ ++ ret = ZSTD_getFrameParams(¶ms, in.src, in.size); ++ err = handle_zstd_error(ret, error); ++ if (err) ++ goto out; ++ if (ret != 0) { ++ error("ZSTD-compressed data has an incomplete frame header"); ++ err = -1; ++ goto out; ++ } ++ if (params.windowSize > ZSTD_WINDOWSIZE_MAX) { ++ error("ZSTD-compressed data has too large a window size"); ++ err = -1; ++ goto out; ++ } ++ ++ /* ++ * Allocate the ZSTD_DStream now that we know how much memory is ++ * required. ++ */ ++ wksp_size = ZSTD_DStreamWorkspaceBound(params.windowSize); ++ wksp = large_malloc(wksp_size); ++ dstream = ZSTD_initDStream(params.windowSize, wksp, wksp_size); ++ if (dstream == NULL) { ++ error("Out of memory while allocating ZSTD_DStream"); ++ err = -1; ++ goto out; ++ } ++ ++ /* ++ * Decompression loop: ++ * Read more data if necessary (error if no more data can be read). ++ * Call the decompression function, which returns 0 when finished. ++ * Flush any data produced if using flush(). ++ */ ++ if (in_pos != NULL) ++ *in_pos = 0; ++ do { ++ /* ++ * If we need to reload data, either we have fill() and can ++ * try to get more data, or we don't and the input is truncated. ++ */ ++ if (in.pos == in.size) { ++ if (in_pos != NULL) ++ *in_pos += in.pos; ++ in_len = fill ? fill(in_buf, ZSTD_IOBUF_SIZE) : -1; ++ if (in_len < 0) { ++ error("ZSTD-compressed data is truncated"); ++ err = -1; ++ goto out; ++ } ++ in.pos = 0; ++ in.size = in_len; ++ } ++ /* Returns zero when the frame is complete. */ ++ ret = ZSTD_decompressStream(dstream, &out, &in); ++ err = handle_zstd_error(ret, error); ++ if (err) ++ goto out; ++ /* Flush all of the data produced if using flush(). */ ++ if (flush != NULL && out.pos > 0) { ++ if (out.pos != flush(out.dst, out.pos)) { ++ error("Failed to flush()"); ++ err = -1; ++ goto out; ++ } ++ out.pos = 0; ++ } ++ } while (ret != 0); ++ ++ if (in_pos != NULL) ++ *in_pos += in.pos; ++ ++ err = 0; ++out: ++ if (in_allocated != NULL) ++ large_free(in_allocated); ++ if (out_allocated != NULL) ++ large_free(out_allocated); ++ if (wksp != NULL) ++ large_free(wksp); ++ return err; ++} ++ ++STATIC int INIT unzstd(unsigned char *buf, unsigned int len, ++ int (*fill)(void*, unsigned int), ++ int (*flush)(void*, unsigned int), ++ unsigned char *out_buf, ++ unsigned int *pos, ++ void (*error)(const char *x)) ++{ ++ return __unzstd(buf, len, fill, flush, out_buf, 0, pos, error); ++} +diff --git a/xen/common/xxhash.c b/xen/common/xxhash.c +new file mode 100644 +index 0000000000..3ab3e01859 +--- /dev/null ++++ b/xen/common/xxhash.c +@@ -0,0 +1,484 @@ ++/* ++ * xxHash - Extremely Fast Hash algorithm ++ * Copyright (C) 2012-2016, Yann Collet. ++ * ++ * BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions are ++ * met: ++ * ++ * * Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * * Redistributions in binary form must reproduce the above ++ * copyright notice, this list of conditions and the following disclaimer ++ * in the documentation and/or other materials provided with the ++ * distribution. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS ++ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT ++ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR ++ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT ++ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, ++ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT ++ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE ++ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++ * ++ * This program is free software; you can redistribute it and/or modify it under ++ * the terms of the GNU General Public License version 2 as published by the ++ * Free Software Foundation. This program is dual-licensed; you may select ++ * either version 2 of the GNU General Public License ("GPL") or BSD license ++ * ("BSD"). ++ * ++ * You can contact the author at: ++ * - xxHash homepage: https://cyan4973.github.io/xxHash/ ++ * - xxHash source repository: https://github.com/Cyan4973/xxHash ++ */ ++ ++#include ++#include ++#include ++#include "zstd/private.h" ++ ++/*-************************************* ++ * Macros ++ **************************************/ ++#define xxh_rotl32(x, r) ((x << r) | (x >> (32 - r))) ++#define xxh_rotl64(x, r) ((x << r) | (x >> (64 - r))) ++ ++#ifdef __LITTLE_ENDIAN ++# define XXH_CPU_LITTLE_ENDIAN 1 ++#else ++# define XXH_CPU_LITTLE_ENDIAN 0 ++#endif ++ ++/*-************************************* ++ * Constants ++ **************************************/ ++static const uint32_t PRIME32_1 = 2654435761U; ++static const uint32_t PRIME32_2 = 2246822519U; ++static const uint32_t PRIME32_3 = 3266489917U; ++static const uint32_t PRIME32_4 = 668265263U; ++static const uint32_t PRIME32_5 = 374761393U; ++ ++static const uint64_t PRIME64_1 = 11400714785074694791ULL; ++static const uint64_t PRIME64_2 = 14029467366897019727ULL; ++static const uint64_t PRIME64_3 = 1609587929392839161ULL; ++static const uint64_t PRIME64_4 = 9650029242287828579ULL; ++static const uint64_t PRIME64_5 = 2870177450012600261ULL; ++ ++/*-************************** ++ * Utils ++ ***************************/ ++void INIT xxh32_copy_state(struct xxh32_state *dst, const struct xxh32_state *src) ++{ ++ memcpy(dst, src, sizeof(*dst)); ++} ++ ++void INIT xxh64_copy_state(struct xxh64_state *dst, const struct xxh64_state *src) ++{ ++ memcpy(dst, src, sizeof(*dst)); ++} ++ ++/*-*************************** ++ * Simple Hash Functions ++ ****************************/ ++static uint32_t INIT xxh32_round(uint32_t seed, const uint32_t input) ++{ ++ seed += input * PRIME32_2; ++ seed = xxh_rotl32(seed, 13); ++ seed *= PRIME32_1; ++ return seed; ++} ++ ++uint32_t INIT xxh32(const void *input, const size_t len, const uint32_t seed) ++{ ++ const uint8_t *p = (const uint8_t *)input; ++ const uint8_t *b_end = p + len; ++ uint32_t h32; ++ ++ if (len >= 16) { ++ const uint8_t *const limit = b_end - 16; ++ uint32_t v1 = seed + PRIME32_1 + PRIME32_2; ++ uint32_t v2 = seed + PRIME32_2; ++ uint32_t v3 = seed + 0; ++ uint32_t v4 = seed - PRIME32_1; ++ ++ do { ++ v1 = xxh32_round(v1, get_unaligned_le32(p)); ++ p += 4; ++ v2 = xxh32_round(v2, get_unaligned_le32(p)); ++ p += 4; ++ v3 = xxh32_round(v3, get_unaligned_le32(p)); ++ p += 4; ++ v4 = xxh32_round(v4, get_unaligned_le32(p)); ++ p += 4; ++ } while (p <= limit); ++ ++ h32 = xxh_rotl32(v1, 1) + xxh_rotl32(v2, 7) + ++ xxh_rotl32(v3, 12) + xxh_rotl32(v4, 18); ++ } else { ++ h32 = seed + PRIME32_5; ++ } ++ ++ h32 += (uint32_t)len; ++ ++ while (p + 4 <= b_end) { ++ h32 += get_unaligned_le32(p) * PRIME32_3; ++ h32 = xxh_rotl32(h32, 17) * PRIME32_4; ++ p += 4; ++ } ++ ++ while (p < b_end) { ++ h32 += (*p) * PRIME32_5; ++ h32 = xxh_rotl32(h32, 11) * PRIME32_1; ++ p++; ++ } ++ ++ h32 ^= h32 >> 15; ++ h32 *= PRIME32_2; ++ h32 ^= h32 >> 13; ++ h32 *= PRIME32_3; ++ h32 ^= h32 >> 16; ++ ++ return h32; ++} ++ ++static uint64_t INIT xxh64_round(uint64_t acc, const uint64_t input) ++{ ++ acc += input * PRIME64_2; ++ acc = xxh_rotl64(acc, 31); ++ acc *= PRIME64_1; ++ return acc; ++} ++ ++static uint64_t INIT xxh64_merge_round(uint64_t acc, uint64_t val) ++{ ++ val = xxh64_round(0, val); ++ acc ^= val; ++ acc = acc * PRIME64_1 + PRIME64_4; ++ return acc; ++} ++ ++uint64_t INIT xxh64(const void *input, const size_t len, const uint64_t seed) ++{ ++ const uint8_t *p = (const uint8_t *)input; ++ const uint8_t *const b_end = p + len; ++ uint64_t h64; ++ ++ if (len >= 32) { ++ const uint8_t *const limit = b_end - 32; ++ uint64_t v1 = seed + PRIME64_1 + PRIME64_2; ++ uint64_t v2 = seed + PRIME64_2; ++ uint64_t v3 = seed + 0; ++ uint64_t v4 = seed - PRIME64_1; ++ ++ do { ++ v1 = xxh64_round(v1, get_unaligned_le64(p)); ++ p += 8; ++ v2 = xxh64_round(v2, get_unaligned_le64(p)); ++ p += 8; ++ v3 = xxh64_round(v3, get_unaligned_le64(p)); ++ p += 8; ++ v4 = xxh64_round(v4, get_unaligned_le64(p)); ++ p += 8; ++ } while (p <= limit); ++ ++ h64 = xxh_rotl64(v1, 1) + xxh_rotl64(v2, 7) + ++ xxh_rotl64(v3, 12) + xxh_rotl64(v4, 18); ++ h64 = xxh64_merge_round(h64, v1); ++ h64 = xxh64_merge_round(h64, v2); ++ h64 = xxh64_merge_round(h64, v3); ++ h64 = xxh64_merge_round(h64, v4); ++ ++ } else { ++ h64 = seed + PRIME64_5; ++ } ++ ++ h64 += (uint64_t)len; ++ ++ while (p + 8 <= b_end) { ++ const uint64_t k1 = xxh64_round(0, get_unaligned_le64(p)); ++ ++ h64 ^= k1; ++ h64 = xxh_rotl64(h64, 27) * PRIME64_1 + PRIME64_4; ++ p += 8; ++ } ++ ++ if (p + 4 <= b_end) { ++ h64 ^= (uint64_t)(get_unaligned_le32(p)) * PRIME64_1; ++ h64 = xxh_rotl64(h64, 23) * PRIME64_2 + PRIME64_3; ++ p += 4; ++ } ++ ++ while (p < b_end) { ++ h64 ^= (*p) * PRIME64_5; ++ h64 = xxh_rotl64(h64, 11) * PRIME64_1; ++ p++; ++ } ++ ++ h64 ^= h64 >> 33; ++ h64 *= PRIME64_2; ++ h64 ^= h64 >> 29; ++ h64 *= PRIME64_3; ++ h64 ^= h64 >> 32; ++ ++ return h64; ++} ++ ++/*-************************************************** ++ * Advanced Hash Functions ++ ***************************************************/ ++void INIT xxh32_reset(struct xxh32_state *statePtr, const uint32_t seed) ++{ ++ /* use a local state for memcpy() to avoid strict-aliasing warnings */ ++ struct xxh32_state state; ++ ++ memset(&state, 0, sizeof(state)); ++ state.v1 = seed + PRIME32_1 + PRIME32_2; ++ state.v2 = seed + PRIME32_2; ++ state.v3 = seed + 0; ++ state.v4 = seed - PRIME32_1; ++ memcpy(statePtr, &state, sizeof(state)); ++} ++ ++void INIT xxh64_reset(struct xxh64_state *statePtr, const uint64_t seed) ++{ ++ /* use a local state for memcpy() to avoid strict-aliasing warnings */ ++ struct xxh64_state state; ++ ++ memset(&state, 0, sizeof(state)); ++ state.v1 = seed + PRIME64_1 + PRIME64_2; ++ state.v2 = seed + PRIME64_2; ++ state.v3 = seed + 0; ++ state.v4 = seed - PRIME64_1; ++ memcpy(statePtr, &state, sizeof(state)); ++} ++ ++int INIT xxh32_update(struct xxh32_state *state, const void *input, const size_t len) ++{ ++ const uint8_t *p = (const uint8_t *)input; ++ const uint8_t *const b_end = p + len; ++ ++ if (input == NULL) ++ return -EINVAL; ++ ++ state->total_len_32 += (uint32_t)len; ++ state->large_len |= (len >= 16) | (state->total_len_32 >= 16); ++ ++ if (state->memsize + len < 16) { /* fill in tmp buffer */ ++ memcpy((uint8_t *)(state->mem32) + state->memsize, input, len); ++ state->memsize += (uint32_t)len; ++ return 0; ++ } ++ ++ if (state->memsize) { /* some data left from previous update */ ++ const uint32_t *p32 = state->mem32; ++ ++ memcpy((uint8_t *)(state->mem32) + state->memsize, input, ++ 16 - state->memsize); ++ ++ state->v1 = xxh32_round(state->v1, get_unaligned_le32(p32)); ++ p32++; ++ state->v2 = xxh32_round(state->v2, get_unaligned_le32(p32)); ++ p32++; ++ state->v3 = xxh32_round(state->v3, get_unaligned_le32(p32)); ++ p32++; ++ state->v4 = xxh32_round(state->v4, get_unaligned_le32(p32)); ++ p32++; ++ ++ p += 16-state->memsize; ++ state->memsize = 0; ++ } ++ ++ if (p <= b_end - 16) { ++ const uint8_t *const limit = b_end - 16; ++ uint32_t v1 = state->v1; ++ uint32_t v2 = state->v2; ++ uint32_t v3 = state->v3; ++ uint32_t v4 = state->v4; ++ ++ do { ++ v1 = xxh32_round(v1, get_unaligned_le32(p)); ++ p += 4; ++ v2 = xxh32_round(v2, get_unaligned_le32(p)); ++ p += 4; ++ v3 = xxh32_round(v3, get_unaligned_le32(p)); ++ p += 4; ++ v4 = xxh32_round(v4, get_unaligned_le32(p)); ++ p += 4; ++ } while (p <= limit); ++ ++ state->v1 = v1; ++ state->v2 = v2; ++ state->v3 = v3; ++ state->v4 = v4; ++ } ++ ++ if (p < b_end) { ++ memcpy(state->mem32, p, (size_t)(b_end-p)); ++ state->memsize = (uint32_t)(b_end-p); ++ } ++ ++ return 0; ++} ++ ++uint32_t INIT xxh32_digest(const struct xxh32_state *state) ++{ ++ const uint8_t *p = (const uint8_t *)state->mem32; ++ const uint8_t *const b_end = (const uint8_t *)(state->mem32) + ++ state->memsize; ++ uint32_t h32; ++ ++ if (state->large_len) { ++ h32 = xxh_rotl32(state->v1, 1) + xxh_rotl32(state->v2, 7) + ++ xxh_rotl32(state->v3, 12) + xxh_rotl32(state->v4, 18); ++ } else { ++ h32 = state->v3 /* == seed */ + PRIME32_5; ++ } ++ ++ h32 += state->total_len_32; ++ ++ while (p + 4 <= b_end) { ++ h32 += get_unaligned_le32(p) * PRIME32_3; ++ h32 = xxh_rotl32(h32, 17) * PRIME32_4; ++ p += 4; ++ } ++ ++ while (p < b_end) { ++ h32 += (*p) * PRIME32_5; ++ h32 = xxh_rotl32(h32, 11) * PRIME32_1; ++ p++; ++ } ++ ++ h32 ^= h32 >> 15; ++ h32 *= PRIME32_2; ++ h32 ^= h32 >> 13; ++ h32 *= PRIME32_3; ++ h32 ^= h32 >> 16; ++ ++ return h32; ++} ++ ++int INIT xxh64_update(struct xxh64_state *state, const void *input, const size_t len) ++{ ++ const uint8_t *p = (const uint8_t *)input; ++ const uint8_t *const b_end = p + len; ++ ++ if (input == NULL) ++ return -EINVAL; ++ ++ state->total_len += len; ++ ++ if (state->memsize + len < 32) { /* fill in tmp buffer */ ++ memcpy(((uint8_t *)state->mem64) + state->memsize, input, len); ++ state->memsize += (uint32_t)len; ++ return 0; ++ } ++ ++ if (state->memsize) { /* tmp buffer is full */ ++ uint64_t *p64 = state->mem64; ++ ++ memcpy(((uint8_t *)p64) + state->memsize, input, ++ 32 - state->memsize); ++ ++ state->v1 = xxh64_round(state->v1, get_unaligned_le64(p64)); ++ p64++; ++ state->v2 = xxh64_round(state->v2, get_unaligned_le64(p64)); ++ p64++; ++ state->v3 = xxh64_round(state->v3, get_unaligned_le64(p64)); ++ p64++; ++ state->v4 = xxh64_round(state->v4, get_unaligned_le64(p64)); ++ ++ p += 32 - state->memsize; ++ state->memsize = 0; ++ } ++ ++ if (p + 32 <= b_end) { ++ const uint8_t *const limit = b_end - 32; ++ uint64_t v1 = state->v1; ++ uint64_t v2 = state->v2; ++ uint64_t v3 = state->v3; ++ uint64_t v4 = state->v4; ++ ++ do { ++ v1 = xxh64_round(v1, get_unaligned_le64(p)); ++ p += 8; ++ v2 = xxh64_round(v2, get_unaligned_le64(p)); ++ p += 8; ++ v3 = xxh64_round(v3, get_unaligned_le64(p)); ++ p += 8; ++ v4 = xxh64_round(v4, get_unaligned_le64(p)); ++ p += 8; ++ } while (p <= limit); ++ ++ state->v1 = v1; ++ state->v2 = v2; ++ state->v3 = v3; ++ state->v4 = v4; ++ } ++ ++ if (p < b_end) { ++ memcpy(state->mem64, p, (size_t)(b_end-p)); ++ state->memsize = (uint32_t)(b_end - p); ++ } ++ ++ return 0; ++} ++ ++uint64_t INIT xxh64_digest(const struct xxh64_state *state) ++{ ++ const uint8_t *p = (const uint8_t *)state->mem64; ++ const uint8_t *const b_end = (const uint8_t *)state->mem64 + ++ state->memsize; ++ uint64_t h64; ++ ++ if (state->total_len >= 32) { ++ const uint64_t v1 = state->v1; ++ const uint64_t v2 = state->v2; ++ const uint64_t v3 = state->v3; ++ const uint64_t v4 = state->v4; ++ ++ h64 = xxh_rotl64(v1, 1) + xxh_rotl64(v2, 7) + ++ xxh_rotl64(v3, 12) + xxh_rotl64(v4, 18); ++ h64 = xxh64_merge_round(h64, v1); ++ h64 = xxh64_merge_round(h64, v2); ++ h64 = xxh64_merge_round(h64, v3); ++ h64 = xxh64_merge_round(h64, v4); ++ } else { ++ h64 = state->v3 + PRIME64_5; ++ } ++ ++ h64 += (uint64_t)state->total_len; ++ ++ while (p + 8 <= b_end) { ++ const uint64_t k1 = xxh64_round(0, get_unaligned_le64(p)); ++ ++ h64 ^= k1; ++ h64 = xxh_rotl64(h64, 27) * PRIME64_1 + PRIME64_4; ++ p += 8; ++ } ++ ++ if (p + 4 <= b_end) { ++ h64 ^= (uint64_t)(get_unaligned_le32(p)) * PRIME64_1; ++ h64 = xxh_rotl64(h64, 23) * PRIME64_2 + PRIME64_3; ++ p += 4; ++ } ++ ++ while (p < b_end) { ++ h64 ^= (*p) * PRIME64_5; ++ h64 = xxh_rotl64(h64, 11) * PRIME64_1; ++ p++; ++ } ++ ++ h64 ^= h64 >> 33; ++ h64 *= PRIME64_2; ++ h64 ^= h64 >> 29; ++ h64 *= PRIME64_3; ++ h64 ^= h64 >> 32; ++ ++ return h64; ++} +diff --git a/xen/common/zstd/bitstream.h b/xen/common/zstd/bitstream.h +new file mode 100644 +index 0000000000..3a49784d5c +--- /dev/null ++++ b/xen/common/zstd/bitstream.h +@@ -0,0 +1,379 @@ ++/* ++ * bitstream ++ * Part of FSE library ++ * header file (to include) ++ * Copyright (C) 2013-2016, Yann Collet. ++ * ++ * BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions are ++ * met: ++ * ++ * * Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * * Redistributions in binary form must reproduce the above ++ * copyright notice, this list of conditions and the following disclaimer ++ * in the documentation and/or other materials provided with the ++ * distribution. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS ++ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT ++ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR ++ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT ++ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, ++ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT ++ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE ++ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++ * ++ * This program is free software; you can redistribute it and/or modify it under ++ * the terms of the GNU General Public License version 2 as published by the ++ * Free Software Foundation. This program is dual-licensed; you may select ++ * either version 2 of the GNU General Public License ("GPL") or BSD license ++ * ("BSD"). ++ * ++ * You can contact the author at : ++ * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy ++ */ ++#ifndef BITSTREAM_H_MODULE ++#define BITSTREAM_H_MODULE ++ ++/* ++* This API consists of small unitary functions, which must be inlined for best performance. ++* Since link-time-optimization is not available for all compilers, ++* these functions are defined into a .h to be included. ++*/ ++ ++/*-**************************************** ++* Dependencies ++******************************************/ ++#include "error_private.h" /* error codes and messages */ ++#include "mem.h" /* unaligned access routines */ ++ ++/*========================================= ++* Target specific ++=========================================*/ ++#define STREAM_ACCUMULATOR_MIN_32 25 ++#define STREAM_ACCUMULATOR_MIN_64 57 ++#define STREAM_ACCUMULATOR_MIN ((U32)(ZSTD_32bits() ? STREAM_ACCUMULATOR_MIN_32 : STREAM_ACCUMULATOR_MIN_64)) ++ ++/*-****************************************** ++* bitStream encoding API (write forward) ++********************************************/ ++/* bitStream can mix input from multiple sources. ++* A critical property of these streams is that they encode and decode in **reverse** direction. ++* So the first bit sequence you add will be the last to be read, like a LIFO stack. ++*/ ++typedef struct { ++ size_t bitContainer; ++ int bitPos; ++ char *startPtr; ++ char *ptr; ++ char *endPtr; ++} BIT_CStream_t; ++ ++ZSTD_STATIC size_t BIT_initCStream(BIT_CStream_t *bitC, void *dstBuffer, size_t dstCapacity); ++ZSTD_STATIC void BIT_addBits(BIT_CStream_t *bitC, size_t value, unsigned nbBits); ++ZSTD_STATIC void BIT_flushBits(BIT_CStream_t *bitC); ++ZSTD_STATIC size_t BIT_closeCStream(BIT_CStream_t *bitC); ++ ++/* Start with initCStream, providing the size of buffer to write into. ++* bitStream will never write outside of this buffer. ++* `dstCapacity` must be >= sizeof(bitD->bitContainer), otherwise @return will be an error code. ++* ++* bits are first added to a local register. ++* Local register is size_t, hence 64-bits on 64-bits systems, or 32-bits on 32-bits systems. ++* Writing data into memory is an explicit operation, performed by the flushBits function. ++* Hence keep track how many bits are potentially stored into local register to avoid register overflow. ++* After a flushBits, a maximum of 7 bits might still be stored into local register. ++* ++* Avoid storing elements of more than 24 bits if you want compatibility with 32-bits bitstream readers. ++* ++* Last operation is to close the bitStream. ++* The function returns the final size of CStream in bytes. ++* If data couldn't fit into `dstBuffer`, it will return a 0 ( == not storable) ++*/ ++ ++/*-******************************************** ++* bitStream decoding API (read backward) ++**********************************************/ ++typedef struct { ++ size_t bitContainer; ++ unsigned bitsConsumed; ++ const char *ptr; ++ const char *start; ++} BIT_DStream_t; ++ ++typedef enum { ++ BIT_DStream_unfinished = 0, ++ BIT_DStream_endOfBuffer = 1, ++ BIT_DStream_completed = 2, ++ BIT_DStream_overflow = 3 ++} BIT_DStream_status; /* result of BIT_reloadDStream() */ ++/* 1,2,4,8 would be better for bitmap combinations, but slows down performance a bit ... :( */ ++ ++ZSTD_STATIC size_t BIT_initDStream(BIT_DStream_t *bitD, const void *srcBuffer, size_t srcSize); ++ZSTD_STATIC size_t BIT_readBits(BIT_DStream_t *bitD, unsigned nbBits); ++ZSTD_STATIC BIT_DStream_status BIT_reloadDStream(BIT_DStream_t *bitD); ++ZSTD_STATIC unsigned BIT_endOfDStream(const BIT_DStream_t *bitD); ++ ++/* Start by invoking BIT_initDStream(). ++* A chunk of the bitStream is then stored into a local register. ++* Local register size is 64-bits on 64-bits systems, 32-bits on 32-bits systems (size_t). ++* You can then retrieve bitFields stored into the local register, **in reverse order**. ++* Local register is explicitly reloaded from memory by the BIT_reloadDStream() method. ++* A reload guarantee a minimum of ((8*sizeof(bitD->bitContainer))-7) bits when its result is BIT_DStream_unfinished. ++* Otherwise, it can be less than that, so proceed accordingly. ++* Checking if DStream has reached its end can be performed with BIT_endOfDStream(). ++*/ ++ ++/*-**************************************** ++* unsafe API ++******************************************/ ++ZSTD_STATIC void BIT_addBitsFast(BIT_CStream_t *bitC, size_t value, unsigned nbBits); ++/* faster, but works only if value is "clean", meaning all high bits above nbBits are 0 */ ++ ++ZSTD_STATIC void BIT_flushBitsFast(BIT_CStream_t *bitC); ++/* unsafe version; does not check buffer overflow */ ++ ++ZSTD_STATIC size_t BIT_readBitsFast(BIT_DStream_t *bitD, unsigned nbBits); ++/* faster, but works only if nbBits >= 1 */ ++ ++/*-************************************************************** ++* Internal functions ++****************************************************************/ ++ZSTD_STATIC unsigned BIT_highbit32(register U32 val) { return 31 - __builtin_clz(val); } ++ ++/*===== Local Constants =====*/ ++static const unsigned BIT_mask[] = {0, 1, 3, 7, 0xF, 0x1F, 0x3F, 0x7F, 0xFF, ++ 0x1FF, 0x3FF, 0x7FF, 0xFFF, 0x1FFF, 0x3FFF, 0x7FFF, 0xFFFF, 0x1FFFF, ++ 0x3FFFF, 0x7FFFF, 0xFFFFF, 0x1FFFFF, 0x3FFFFF, 0x7FFFFF, 0xFFFFFF, 0x1FFFFFF, 0x3FFFFFF}; /* up to 26 bits */ ++ ++/*-************************************************************** ++* bitStream encoding ++****************************************************************/ ++/*! BIT_initCStream() : ++ * `dstCapacity` must be > sizeof(void*) ++ * @return : 0 if success, ++ otherwise an error code (can be tested using ERR_isError() ) */ ++ZSTD_STATIC size_t BIT_initCStream(BIT_CStream_t *bitC, void *startPtr, size_t dstCapacity) ++{ ++ bitC->bitContainer = 0; ++ bitC->bitPos = 0; ++ bitC->startPtr = (char *)startPtr; ++ bitC->ptr = bitC->startPtr; ++ bitC->endPtr = bitC->startPtr + dstCapacity - sizeof(bitC->ptr); ++ if (dstCapacity <= sizeof(bitC->ptr)) ++ return ERROR(dstSize_tooSmall); ++ return 0; ++} ++ ++/*! BIT_addBits() : ++ can add up to 26 bits into `bitC`. ++ Does not check for register overflow ! */ ++ZSTD_STATIC void BIT_addBits(BIT_CStream_t *bitC, size_t value, unsigned nbBits) ++{ ++ bitC->bitContainer |= (value & BIT_mask[nbBits]) << bitC->bitPos; ++ bitC->bitPos += nbBits; ++} ++ ++/*! BIT_addBitsFast() : ++ * works only if `value` is _clean_, meaning all high bits above nbBits are 0 */ ++ZSTD_STATIC void BIT_addBitsFast(BIT_CStream_t *bitC, size_t value, unsigned nbBits) ++{ ++ bitC->bitContainer |= value << bitC->bitPos; ++ bitC->bitPos += nbBits; ++} ++ ++/*! BIT_flushBitsFast() : ++ * unsafe version; does not check buffer overflow */ ++ZSTD_STATIC void BIT_flushBitsFast(BIT_CStream_t *bitC) ++{ ++ size_t const nbBytes = bitC->bitPos >> 3; ++ ZSTD_writeLEST(bitC->ptr, bitC->bitContainer); ++ bitC->ptr += nbBytes; ++ bitC->bitPos &= 7; ++ bitC->bitContainer >>= nbBytes * 8; /* if bitPos >= sizeof(bitContainer)*8 --> undefined behavior */ ++} ++ ++/*! BIT_flushBits() : ++ * safe version; check for buffer overflow, and prevents it. ++ * note : does not signal buffer overflow. This will be revealed later on using BIT_closeCStream() */ ++ZSTD_STATIC void BIT_flushBits(BIT_CStream_t *bitC) ++{ ++ size_t const nbBytes = bitC->bitPos >> 3; ++ ZSTD_writeLEST(bitC->ptr, bitC->bitContainer); ++ bitC->ptr += nbBytes; ++ if (bitC->ptr > bitC->endPtr) ++ bitC->ptr = bitC->endPtr; ++ bitC->bitPos &= 7; ++ bitC->bitContainer >>= nbBytes * 8; /* if bitPos >= sizeof(bitContainer)*8 --> undefined behavior */ ++} ++ ++/*! BIT_closeCStream() : ++ * @return : size of CStream, in bytes, ++ or 0 if it could not fit into dstBuffer */ ++ZSTD_STATIC size_t BIT_closeCStream(BIT_CStream_t *bitC) ++{ ++ BIT_addBitsFast(bitC, 1, 1); /* endMark */ ++ BIT_flushBits(bitC); ++ ++ if (bitC->ptr >= bitC->endPtr) ++ return 0; /* doesn't fit within authorized budget : cancel */ ++ ++ return (bitC->ptr - bitC->startPtr) + (bitC->bitPos > 0); ++} ++ ++/*-******************************************************** ++* bitStream decoding ++**********************************************************/ ++/*! BIT_initDStream() : ++* Initialize a BIT_DStream_t. ++* `bitD` : a pointer to an already allocated BIT_DStream_t structure. ++* `srcSize` must be the *exact* size of the bitStream, in bytes. ++* @return : size of stream (== srcSize) or an errorCode if a problem is detected ++*/ ++ZSTD_STATIC size_t BIT_initDStream(BIT_DStream_t *bitD, const void *srcBuffer, size_t srcSize) ++{ ++ if (srcSize < 1) { ++ memset(bitD, 0, sizeof(*bitD)); ++ return ERROR(srcSize_wrong); ++ } ++ ++ if (srcSize >= sizeof(bitD->bitContainer)) { /* normal case */ ++ bitD->start = (const char *)srcBuffer; ++ bitD->ptr = (const char *)srcBuffer + srcSize - sizeof(bitD->bitContainer); ++ bitD->bitContainer = ZSTD_readLEST(bitD->ptr); ++ { ++ BYTE const lastByte = ((const BYTE *)srcBuffer)[srcSize - 1]; ++ bitD->bitsConsumed = lastByte ? 8 - BIT_highbit32(lastByte) : 0; /* ensures bitsConsumed is always set */ ++ if (lastByte == 0) ++ return ERROR(GENERIC); /* endMark not present */ ++ } ++ } else { ++ bitD->start = (const char *)srcBuffer; ++ bitD->ptr = bitD->start; ++ bitD->bitContainer = *(const BYTE *)(bitD->start); ++ switch (srcSize) { ++ case 7: bitD->bitContainer += (size_t)(((const BYTE *)(srcBuffer))[6]) << (sizeof(bitD->bitContainer) * 8 - 16); ++ /* fall through */ ++ case 6: bitD->bitContainer += (size_t)(((const BYTE *)(srcBuffer))[5]) << (sizeof(bitD->bitContainer) * 8 - 24); ++ /* fall through */ ++ case 5: bitD->bitContainer += (size_t)(((const BYTE *)(srcBuffer))[4]) << (sizeof(bitD->bitContainer) * 8 - 32); ++ /* fall through */ ++ case 4: bitD->bitContainer += (size_t)(((const BYTE *)(srcBuffer))[3]) << 24; ++ /* fall through */ ++ case 3: bitD->bitContainer += (size_t)(((const BYTE *)(srcBuffer))[2]) << 16; ++ /* fall through */ ++ case 2: bitD->bitContainer += (size_t)(((const BYTE *)(srcBuffer))[1]) << 8; ++ default:; ++ } ++ { ++ BYTE const lastByte = ((const BYTE *)srcBuffer)[srcSize - 1]; ++ bitD->bitsConsumed = lastByte ? 8 - BIT_highbit32(lastByte) : 0; ++ if (lastByte == 0) ++ return ERROR(GENERIC); /* endMark not present */ ++ } ++ bitD->bitsConsumed += (U32)(sizeof(bitD->bitContainer) - srcSize) * 8; ++ } ++ ++ return srcSize; ++} ++ ++ZSTD_STATIC size_t BIT_getUpperBits(size_t bitContainer, U32 const start) { return bitContainer >> start; } ++ ++ZSTD_STATIC size_t BIT_getMiddleBits(size_t bitContainer, U32 const start, U32 const nbBits) { return (bitContainer >> start) & BIT_mask[nbBits]; } ++ ++ZSTD_STATIC size_t BIT_getLowerBits(size_t bitContainer, U32 const nbBits) { return bitContainer & BIT_mask[nbBits]; } ++ ++/*! BIT_lookBits() : ++ * Provides next n bits from local register. ++ * local register is not modified. ++ * On 32-bits, maxNbBits==24. ++ * On 64-bits, maxNbBits==56. ++ * @return : value extracted ++ */ ++ZSTD_STATIC size_t BIT_lookBits(const BIT_DStream_t *bitD, U32 nbBits) ++{ ++ U32 const bitMask = sizeof(bitD->bitContainer) * 8 - 1; ++ return ((bitD->bitContainer << (bitD->bitsConsumed & bitMask)) >> 1) >> ((bitMask - nbBits) & bitMask); ++} ++ ++/*! BIT_lookBitsFast() : ++* unsafe version; only works only if nbBits >= 1 */ ++ZSTD_STATIC size_t BIT_lookBitsFast(const BIT_DStream_t *bitD, U32 nbBits) ++{ ++ U32 const bitMask = sizeof(bitD->bitContainer) * 8 - 1; ++ return (bitD->bitContainer << (bitD->bitsConsumed & bitMask)) >> (((bitMask + 1) - nbBits) & bitMask); ++} ++ ++ZSTD_STATIC void BIT_skipBits(BIT_DStream_t *bitD, U32 nbBits) { bitD->bitsConsumed += nbBits; } ++ ++/*! BIT_readBits() : ++ * Read (consume) next n bits from local register and update. ++ * Pay attention to not read more than nbBits contained into local register. ++ * @return : extracted value. ++ */ ++ZSTD_STATIC size_t BIT_readBits(BIT_DStream_t *bitD, U32 nbBits) ++{ ++ size_t const value = BIT_lookBits(bitD, nbBits); ++ BIT_skipBits(bitD, nbBits); ++ return value; ++} ++ ++/*! BIT_readBitsFast() : ++* unsafe version; only works only if nbBits >= 1 */ ++ZSTD_STATIC size_t BIT_readBitsFast(BIT_DStream_t *bitD, U32 nbBits) ++{ ++ size_t const value = BIT_lookBitsFast(bitD, nbBits); ++ BIT_skipBits(bitD, nbBits); ++ return value; ++} ++ ++/*! BIT_reloadDStream() : ++* Refill `bitD` from buffer previously set in BIT_initDStream() . ++* This function is safe, it guarantees it will not read beyond src buffer. ++* @return : status of `BIT_DStream_t` internal register. ++ if status == BIT_DStream_unfinished, internal register is filled with >= (sizeof(bitD->bitContainer)*8 - 7) bits */ ++ZSTD_STATIC BIT_DStream_status BIT_reloadDStream(BIT_DStream_t *bitD) ++{ ++ if (bitD->bitsConsumed > (sizeof(bitD->bitContainer) * 8)) /* should not happen => corruption detected */ ++ return BIT_DStream_overflow; ++ ++ if (bitD->ptr >= bitD->start + sizeof(bitD->bitContainer)) { ++ bitD->ptr -= bitD->bitsConsumed >> 3; ++ bitD->bitsConsumed &= 7; ++ bitD->bitContainer = ZSTD_readLEST(bitD->ptr); ++ return BIT_DStream_unfinished; ++ } ++ if (bitD->ptr == bitD->start) { ++ if (bitD->bitsConsumed < sizeof(bitD->bitContainer) * 8) ++ return BIT_DStream_endOfBuffer; ++ return BIT_DStream_completed; ++ } ++ { ++ U32 nbBytes = bitD->bitsConsumed >> 3; ++ BIT_DStream_status result = BIT_DStream_unfinished; ++ if (bitD->ptr - nbBytes < bitD->start) { ++ nbBytes = (U32)(bitD->ptr - bitD->start); /* ptr > start */ ++ result = BIT_DStream_endOfBuffer; ++ } ++ bitD->ptr -= nbBytes; ++ bitD->bitsConsumed -= nbBytes * 8; ++ bitD->bitContainer = ZSTD_readLEST(bitD->ptr); /* reminder : srcSize > sizeof(bitD) */ ++ return result; ++ } ++} ++ ++/*! BIT_endOfDStream() : ++* @return Tells if DStream has exactly reached its end (all bits consumed). ++*/ ++ZSTD_STATIC unsigned BIT_endOfDStream(const BIT_DStream_t *DStream) ++{ ++ return ((DStream->ptr == DStream->start) && (DStream->bitsConsumed == sizeof(DStream->bitContainer) * 8)); ++} ++ ++#endif /* BITSTREAM_H_MODULE */ +diff --git a/xen/common/zstd/decompress.c b/xen/common/zstd/decompress.c +new file mode 100644 +index 0000000000..8e627d881a +--- /dev/null ++++ b/xen/common/zstd/decompress.c +@@ -0,0 +1,2489 @@ ++/** ++ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc. ++ * All rights reserved. ++ * ++ * This source code is licensed under the BSD-style license found in the ++ * LICENSE file in the root directory of https://github.com/facebook/zstd. ++ * An additional grant of patent rights can be found in the PATENTS file in the ++ * same directory. ++ * ++ * This program is free software; you can redistribute it and/or modify it under ++ * the terms of the GNU General Public License version 2 as published by the ++ * Free Software Foundation. This program is dual-licensed; you may select ++ * either version 2 of the GNU General Public License ("GPL") or BSD license ++ * ("BSD"). ++ */ ++ ++/* *************************************************************** ++* Tuning parameters ++*****************************************************************/ ++/*! ++* MAXWINDOWSIZE_DEFAULT : ++* maximum window size accepted by DStream, by default. ++* Frames requiring more memory will be rejected. ++*/ ++#ifndef ZSTD_MAXWINDOWSIZE_DEFAULT ++#define ZSTD_MAXWINDOWSIZE_DEFAULT ((1 << ZSTD_WINDOWLOG_MAX) + 1) /* defined within zstd.h */ ++#endif ++ ++/*-******************************************************* ++* Dependencies ++*********************************************************/ ++#include "fse.h" ++#include "huf.h" ++#include "mem.h" /* low level memory routines */ ++#include "zstd_internal.h" ++#include /* memcpy, memmove, memset */ ++ ++#define ZSTD_PREFETCH(ptr) __builtin_prefetch(ptr, 0, 0) ++ ++/*-************************************* ++* Macros ++***************************************/ ++#define ZSTD_isError ERR_isError /* for inlining */ ++#define FSE_isError ERR_isError ++#define HUF_isError ERR_isError ++ ++/*_******************************************************* ++* Memory operations ++**********************************************************/ ++static void INIT ZSTD_copy4(void *dst, const void *src) { memcpy(dst, src, 4); } ++ ++/*-************************************************************* ++* Context management ++***************************************************************/ ++typedef enum { ++ ZSTDds_getFrameHeaderSize, ++ ZSTDds_decodeFrameHeader, ++ ZSTDds_decodeBlockHeader, ++ ZSTDds_decompressBlock, ++ ZSTDds_decompressLastBlock, ++ ZSTDds_checkChecksum, ++ ZSTDds_decodeSkippableHeader, ++ ZSTDds_skipFrame ++} ZSTD_dStage; ++ ++typedef struct { ++ FSE_DTable LLTable[FSE_DTABLE_SIZE_U32(LLFSELog)]; ++ FSE_DTable OFTable[FSE_DTABLE_SIZE_U32(OffFSELog)]; ++ FSE_DTable MLTable[FSE_DTABLE_SIZE_U32(MLFSELog)]; ++ HUF_DTable hufTable[HUF_DTABLE_SIZE(HufLog)]; /* can accommodate HUF_decompress4X */ ++ U64 workspace[HUF_DECOMPRESS_WORKSPACE_SIZE_U32 / 2]; ++ U32 rep[ZSTD_REP_NUM]; ++} ZSTD_entropyTables_t; ++ ++struct ZSTD_DCtx_s { ++ const FSE_DTable *LLTptr; ++ const FSE_DTable *MLTptr; ++ const FSE_DTable *OFTptr; ++ const HUF_DTable *HUFptr; ++ ZSTD_entropyTables_t entropy; ++ const void *previousDstEnd; /* detect continuity */ ++ const void *base; /* start of curr segment */ ++ const void *vBase; /* virtual start of previous segment if it was just before curr one */ ++ const void *dictEnd; /* end of previous segment */ ++ size_t expected; ++ ZSTD_frameParams fParams; ++ blockType_e bType; /* used in ZSTD_decompressContinue(), to transfer blockType between header decoding and block decoding stages */ ++ ZSTD_dStage stage; ++ U32 litEntropy; ++ U32 fseEntropy; ++ struct xxh64_state xxhState; ++ size_t headerSize; ++ U32 dictID; ++ const BYTE *litPtr; ++ ZSTD_customMem customMem; ++ size_t litSize; ++ size_t rleSize; ++ BYTE litBuffer[ZSTD_BLOCKSIZE_ABSOLUTEMAX + WILDCOPY_OVERLENGTH]; ++ BYTE headerBuffer[ZSTD_FRAMEHEADERSIZE_MAX]; ++}; /* typedef'd to ZSTD_DCtx within "zstd.h" */ ++ ++size_t INIT ZSTD_DCtxWorkspaceBound(void) { return ZSTD_ALIGN(sizeof(ZSTD_stack)) + ZSTD_ALIGN(sizeof(ZSTD_DCtx)); } ++ ++size_t INIT ZSTD_decompressBegin(ZSTD_DCtx *dctx) ++{ ++ dctx->expected = ZSTD_frameHeaderSize_prefix; ++ dctx->stage = ZSTDds_getFrameHeaderSize; ++ dctx->previousDstEnd = NULL; ++ dctx->base = NULL; ++ dctx->vBase = NULL; ++ dctx->dictEnd = NULL; ++ dctx->entropy.hufTable[0] = (HUF_DTable)((HufLog)*0x1000001); /* cover both little and big endian */ ++ dctx->litEntropy = dctx->fseEntropy = 0; ++ dctx->dictID = 0; ++ ZSTD_STATIC_ASSERT(sizeof(dctx->entropy.rep) == sizeof(repStartValue)); ++ memcpy(dctx->entropy.rep, repStartValue, sizeof(repStartValue)); /* initial repcodes */ ++ dctx->LLTptr = dctx->entropy.LLTable; ++ dctx->MLTptr = dctx->entropy.MLTable; ++ dctx->OFTptr = dctx->entropy.OFTable; ++ dctx->HUFptr = dctx->entropy.hufTable; ++ return 0; ++} ++ ++ZSTD_DCtx INIT *ZSTD_createDCtx_advanced(ZSTD_customMem customMem) ++{ ++ ZSTD_DCtx *dctx; ++ ++ if (!customMem.customAlloc || !customMem.customFree) ++ return NULL; ++ ++ dctx = (ZSTD_DCtx *)ZSTD_malloc(sizeof(ZSTD_DCtx), customMem); ++ if (!dctx) ++ return NULL; ++ memcpy(&dctx->customMem, &customMem, sizeof(customMem)); ++ ZSTD_decompressBegin(dctx); ++ return dctx; ++} ++ ++ZSTD_DCtx INIT *ZSTD_initDCtx(void *workspace, size_t workspaceSize) ++{ ++ ZSTD_customMem const stackMem = ZSTD_initStack(workspace, workspaceSize); ++ return ZSTD_createDCtx_advanced(stackMem); ++} ++ ++size_t INIT ZSTD_freeDCtx(ZSTD_DCtx *dctx) ++{ ++ if (dctx == NULL) ++ return 0; /* support free on NULL */ ++ ZSTD_free(dctx, dctx->customMem); ++ return 0; /* reserved as a potential error code in the future */ ++} ++ ++void INIT ZSTD_copyDCtx(ZSTD_DCtx *dstDCtx, const ZSTD_DCtx *srcDCtx) ++{ ++ size_t const workSpaceSize = (ZSTD_BLOCKSIZE_ABSOLUTEMAX + WILDCOPY_OVERLENGTH) + ZSTD_frameHeaderSize_max; ++ memcpy(dstDCtx, srcDCtx, sizeof(ZSTD_DCtx) - workSpaceSize); /* no need to copy workspace */ ++} ++ ++static void INIT ZSTD_refDDict(ZSTD_DCtx *dstDCtx, const ZSTD_DDict *ddict); ++ ++/*-************************************************************* ++* Decompression section ++***************************************************************/ ++ ++/*! ZSTD_isFrame() : ++ * Tells if the content of `buffer` starts with a valid Frame Identifier. ++ * Note : Frame Identifier is 4 bytes. If `size < 4`, @return will always be 0. ++ * Note 2 : Legacy Frame Identifiers are considered valid only if Legacy Support is enabled. ++ * Note 3 : Skippable Frame Identifiers are considered valid. */ ++unsigned INIT ZSTD_isFrame(const void *buffer, size_t size) ++{ ++ if (size < 4) ++ return 0; ++ { ++ U32 const magic = ZSTD_readLE32(buffer); ++ if (magic == ZSTD_MAGICNUMBER) ++ return 1; ++ if ((magic & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) ++ return 1; ++ } ++ return 0; ++} ++ ++/** ZSTD_frameHeaderSize() : ++* srcSize must be >= ZSTD_frameHeaderSize_prefix. ++* @return : size of the Frame Header */ ++static size_t INIT ZSTD_frameHeaderSize(const void *src, size_t srcSize) ++{ ++ if (srcSize < ZSTD_frameHeaderSize_prefix) ++ return ERROR(srcSize_wrong); ++ { ++ BYTE const fhd = ((const BYTE *)src)[4]; ++ U32 const dictID = fhd & 3; ++ U32 const singleSegment = (fhd >> 5) & 1; ++ U32 const fcsId = fhd >> 6; ++ return ZSTD_frameHeaderSize_prefix + !singleSegment + ZSTD_did_fieldSize[dictID] + ZSTD_fcs_fieldSize[fcsId] + (singleSegment && !fcsId); ++ } ++} ++ ++/** ZSTD_getFrameParams() : ++* decode Frame Header, or require larger `srcSize`. ++* @return : 0, `fparamsPtr` is correctly filled, ++* >0, `srcSize` is too small, result is expected `srcSize`, ++* or an error code, which can be tested using ZSTD_isError() */ ++size_t INIT ZSTD_getFrameParams(ZSTD_frameParams *fparamsPtr, const void *src, size_t srcSize) ++{ ++ const BYTE *ip = (const BYTE *)src; ++ ++ if (srcSize < ZSTD_frameHeaderSize_prefix) ++ return ZSTD_frameHeaderSize_prefix; ++ if (ZSTD_readLE32(src) != ZSTD_MAGICNUMBER) { ++ if ((ZSTD_readLE32(src) & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) { ++ if (srcSize < ZSTD_skippableHeaderSize) ++ return ZSTD_skippableHeaderSize; /* magic number + skippable frame length */ ++ memset(fparamsPtr, 0, sizeof(*fparamsPtr)); ++ fparamsPtr->frameContentSize = ZSTD_readLE32((const char *)src + 4); ++ fparamsPtr->windowSize = 0; /* windowSize==0 means a frame is skippable */ ++ return 0; ++ } ++ return ERROR(prefix_unknown); ++ } ++ ++ /* ensure there is enough `srcSize` to fully read/decode frame header */ ++ { ++ size_t const fhsize = ZSTD_frameHeaderSize(src, srcSize); ++ if (srcSize < fhsize) ++ return fhsize; ++ } ++ ++ { ++ BYTE const fhdByte = ip[4]; ++ size_t pos = 5; ++ U32 const dictIDSizeCode = fhdByte & 3; ++ U32 const checksumFlag = (fhdByte >> 2) & 1; ++ U32 const singleSegment = (fhdByte >> 5) & 1; ++ U32 const fcsID = fhdByte >> 6; ++ U32 const windowSizeMax = 1U << ZSTD_WINDOWLOG_MAX; ++ U32 windowSize = 0; ++ U32 dictID = 0; ++ U64 frameContentSize = 0; ++ if ((fhdByte & 0x08) != 0) ++ return ERROR(frameParameter_unsupported); /* reserved bits, which must be zero */ ++ if (!singleSegment) { ++ BYTE const wlByte = ip[pos++]; ++ U32 const windowLog = (wlByte >> 3) + ZSTD_WINDOWLOG_ABSOLUTEMIN; ++ if (windowLog > ZSTD_WINDOWLOG_MAX) ++ return ERROR(frameParameter_windowTooLarge); /* avoids issue with 1 << windowLog */ ++ windowSize = (1U << windowLog); ++ windowSize += (windowSize >> 3) * (wlByte & 7); ++ } ++ ++ switch (dictIDSizeCode) { ++ default: /* impossible */ ++ case 0: break; ++ case 1: ++ dictID = ip[pos]; ++ pos++; ++ break; ++ case 2: ++ dictID = ZSTD_readLE16(ip + pos); ++ pos += 2; ++ break; ++ case 3: ++ dictID = ZSTD_readLE32(ip + pos); ++ pos += 4; ++ break; ++ } ++ switch (fcsID) { ++ default: /* impossible */ ++ case 0: ++ if (singleSegment) ++ frameContentSize = ip[pos]; ++ break; ++ case 1: frameContentSize = ZSTD_readLE16(ip + pos) + 256; break; ++ case 2: frameContentSize = ZSTD_readLE32(ip + pos); break; ++ case 3: frameContentSize = ZSTD_readLE64(ip + pos); break; ++ } ++ if (!windowSize) ++ windowSize = (U32)frameContentSize; ++ if (windowSize > windowSizeMax) ++ return ERROR(frameParameter_windowTooLarge); ++ fparamsPtr->frameContentSize = frameContentSize; ++ fparamsPtr->windowSize = windowSize; ++ fparamsPtr->dictID = dictID; ++ fparamsPtr->checksumFlag = checksumFlag; ++ } ++ return 0; ++} ++ ++/** ZSTD_getFrameContentSize() : ++* compatible with legacy mode ++* @return : decompressed size of the single frame pointed to be `src` if known, otherwise ++* - ZSTD_CONTENTSIZE_UNKNOWN if the size cannot be determined ++* - ZSTD_CONTENTSIZE_ERROR if an error occurred (e.g. invalid magic number, srcSize too small) */ ++unsigned long long INIT ZSTD_getFrameContentSize(const void *src, size_t srcSize) ++{ ++ { ++ ZSTD_frameParams fParams; ++ if (ZSTD_getFrameParams(&fParams, src, srcSize) != 0) ++ return ZSTD_CONTENTSIZE_ERROR; ++ if (fParams.windowSize == 0) { ++ /* Either skippable or empty frame, size == 0 either way */ ++ return 0; ++ } else if (fParams.frameContentSize != 0) { ++ return fParams.frameContentSize; ++ } else { ++ return ZSTD_CONTENTSIZE_UNKNOWN; ++ } ++ } ++} ++ ++/** ZSTD_findDecompressedSize() : ++ * compatible with legacy mode ++ * `srcSize` must be the exact length of some number of ZSTD compressed and/or ++ * skippable frames ++ * @return : decompressed size of the frames contained */ ++unsigned long long INIT ZSTD_findDecompressedSize(const void *src, size_t srcSize) ++{ ++ { ++ unsigned long long totalDstSize = 0; ++ while (srcSize >= ZSTD_frameHeaderSize_prefix) { ++ const U32 magicNumber = ZSTD_readLE32(src); ++ ++ if ((magicNumber & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) { ++ size_t skippableSize; ++ if (srcSize < ZSTD_skippableHeaderSize) ++ return ERROR(srcSize_wrong); ++ skippableSize = ZSTD_readLE32((const BYTE *)src + 4) + ZSTD_skippableHeaderSize; ++ if (srcSize < skippableSize) { ++ return ZSTD_CONTENTSIZE_ERROR; ++ } ++ ++ src = (const BYTE *)src + skippableSize; ++ srcSize -= skippableSize; ++ continue; ++ } ++ ++ { ++ unsigned long long const ret = ZSTD_getFrameContentSize(src, srcSize); ++ if (ret >= ZSTD_CONTENTSIZE_ERROR) ++ return ret; ++ ++ /* check for overflow */ ++ if (totalDstSize + ret < totalDstSize) ++ return ZSTD_CONTENTSIZE_ERROR; ++ totalDstSize += ret; ++ } ++ { ++ size_t const frameSrcSize = ZSTD_findFrameCompressedSize(src, srcSize); ++ if (ZSTD_isError(frameSrcSize)) { ++ return ZSTD_CONTENTSIZE_ERROR; ++ } ++ ++ src = (const BYTE *)src + frameSrcSize; ++ srcSize -= frameSrcSize; ++ } ++ } ++ ++ if (srcSize) { ++ return ZSTD_CONTENTSIZE_ERROR; ++ } ++ ++ return totalDstSize; ++ } ++} ++ ++/** ZSTD_decodeFrameHeader() : ++* `headerSize` must be the size provided by ZSTD_frameHeaderSize(). ++* @return : 0 if success, or an error code, which can be tested using ZSTD_isError() */ ++static size_t INIT ZSTD_decodeFrameHeader(ZSTD_DCtx *dctx, const void *src, size_t headerSize) ++{ ++ size_t const result = ZSTD_getFrameParams(&(dctx->fParams), src, headerSize); ++ if (ZSTD_isError(result)) ++ return result; /* invalid header */ ++ if (result > 0) ++ return ERROR(srcSize_wrong); /* headerSize too small */ ++ if (dctx->fParams.dictID && (dctx->dictID != dctx->fParams.dictID)) ++ return ERROR(dictionary_wrong); ++ if (dctx->fParams.checksumFlag) ++ xxh64_reset(&dctx->xxhState, 0); ++ return 0; ++} ++ ++typedef struct { ++ blockType_e blockType; ++ U32 lastBlock; ++ U32 origSize; ++} blockProperties_t; ++ ++/*! ZSTD_getcBlockSize() : ++* Provides the size of compressed block from block header `src` */ ++size_t INIT ZSTD_getcBlockSize(const void *src, size_t srcSize, blockProperties_t *bpPtr) ++{ ++ if (srcSize < ZSTD_blockHeaderSize) ++ return ERROR(srcSize_wrong); ++ { ++ U32 const cBlockHeader = ZSTD_readLE24(src); ++ U32 const cSize = cBlockHeader >> 3; ++ bpPtr->lastBlock = cBlockHeader & 1; ++ bpPtr->blockType = (blockType_e)((cBlockHeader >> 1) & 3); ++ bpPtr->origSize = cSize; /* only useful for RLE */ ++ if (bpPtr->blockType == bt_rle) ++ return 1; ++ if (bpPtr->blockType == bt_reserved) ++ return ERROR(corruption_detected); ++ return cSize; ++ } ++} ++ ++static size_t INIT ZSTD_copyRawBlock(void *dst, size_t dstCapacity, const void *src, size_t srcSize) ++{ ++ if (srcSize > dstCapacity) ++ return ERROR(dstSize_tooSmall); ++ memcpy(dst, src, srcSize); ++ return srcSize; ++} ++ ++static size_t INIT ZSTD_setRleBlock(void *dst, size_t dstCapacity, const void *src, size_t srcSize, size_t regenSize) ++{ ++ if (srcSize != 1) ++ return ERROR(srcSize_wrong); ++ if (regenSize > dstCapacity) ++ return ERROR(dstSize_tooSmall); ++ memset(dst, *(const BYTE *)src, regenSize); ++ return regenSize; ++} ++ ++/*! ZSTD_decodeLiteralsBlock() : ++ @return : nb of bytes read from src (< srcSize ) */ ++size_t INIT ZSTD_decodeLiteralsBlock(ZSTD_DCtx *dctx, const void *src, size_t srcSize) /* note : srcSize < BLOCKSIZE */ ++{ ++ if (srcSize < MIN_CBLOCK_SIZE) ++ return ERROR(corruption_detected); ++ ++ { ++ const BYTE *const istart = (const BYTE *)src; ++ symbolEncodingType_e const litEncType = (symbolEncodingType_e)(istart[0] & 3); ++ ++ switch (litEncType) { ++ case set_repeat: ++ if (dctx->litEntropy == 0) ++ return ERROR(dictionary_corrupted); ++ /* fall through */ ++ case set_compressed: ++ if (srcSize < 5) ++ return ERROR(corruption_detected); /* srcSize >= MIN_CBLOCK_SIZE == 3; here we need up to 5 for case 3 */ ++ { ++ size_t lhSize, litSize, litCSize; ++ U32 singleStream = 0; ++ U32 const lhlCode = (istart[0] >> 2) & 3; ++ U32 const lhc = ZSTD_readLE32(istart); ++ switch (lhlCode) { ++ case 0: ++ case 1: ++ default: /* note : default is impossible, since lhlCode into [0..3] */ ++ /* 2 - 2 - 10 - 10 */ ++ singleStream = !lhlCode; ++ lhSize = 3; ++ litSize = (lhc >> 4) & 0x3FF; ++ litCSize = (lhc >> 14) & 0x3FF; ++ break; ++ case 2: ++ /* 2 - 2 - 14 - 14 */ ++ lhSize = 4; ++ litSize = (lhc >> 4) & 0x3FFF; ++ litCSize = lhc >> 18; ++ break; ++ case 3: ++ /* 2 - 2 - 18 - 18 */ ++ lhSize = 5; ++ litSize = (lhc >> 4) & 0x3FFFF; ++ litCSize = (lhc >> 22) + (istart[4] << 10); ++ break; ++ } ++ if (litSize > ZSTD_BLOCKSIZE_ABSOLUTEMAX) ++ return ERROR(corruption_detected); ++ if (litCSize + lhSize > srcSize) ++ return ERROR(corruption_detected); ++ ++ if (HUF_isError( ++ (litEncType == set_repeat) ++ ? (singleStream ? HUF_decompress1X_usingDTable(dctx->litBuffer, litSize, istart + lhSize, litCSize, dctx->HUFptr) ++ : HUF_decompress4X_usingDTable(dctx->litBuffer, litSize, istart + lhSize, litCSize, dctx->HUFptr)) ++ : (singleStream ++ ? HUF_decompress1X2_DCtx_wksp(dctx->entropy.hufTable, dctx->litBuffer, litSize, istart + lhSize, litCSize, ++ dctx->entropy.workspace, sizeof(dctx->entropy.workspace)) ++ : HUF_decompress4X_hufOnly_wksp(dctx->entropy.hufTable, dctx->litBuffer, litSize, istart + lhSize, litCSize, ++ dctx->entropy.workspace, sizeof(dctx->entropy.workspace))))) ++ return ERROR(corruption_detected); ++ ++ dctx->litPtr = dctx->litBuffer; ++ dctx->litSize = litSize; ++ dctx->litEntropy = 1; ++ if (litEncType == set_compressed) ++ dctx->HUFptr = dctx->entropy.hufTable; ++ memset(dctx->litBuffer + dctx->litSize, 0, WILDCOPY_OVERLENGTH); ++ return litCSize + lhSize; ++ } ++ ++ case set_basic: { ++ size_t litSize, lhSize; ++ U32 const lhlCode = ((istart[0]) >> 2) & 3; ++ switch (lhlCode) { ++ case 0: ++ case 2: ++ default: /* note : default is impossible, since lhlCode into [0..3] */ ++ lhSize = 1; ++ litSize = istart[0] >> 3; ++ break; ++ case 1: ++ lhSize = 2; ++ litSize = ZSTD_readLE16(istart) >> 4; ++ break; ++ case 3: ++ lhSize = 3; ++ litSize = ZSTD_readLE24(istart) >> 4; ++ break; ++ } ++ ++ if (lhSize + litSize + WILDCOPY_OVERLENGTH > srcSize) { /* risk reading beyond src buffer with wildcopy */ ++ if (litSize + lhSize > srcSize) ++ return ERROR(corruption_detected); ++ memcpy(dctx->litBuffer, istart + lhSize, litSize); ++ dctx->litPtr = dctx->litBuffer; ++ dctx->litSize = litSize; ++ memset(dctx->litBuffer + dctx->litSize, 0, WILDCOPY_OVERLENGTH); ++ return lhSize + litSize; ++ } ++ /* direct reference into compressed stream */ ++ dctx->litPtr = istart + lhSize; ++ dctx->litSize = litSize; ++ return lhSize + litSize; ++ } ++ ++ case set_rle: { ++ U32 const lhlCode = ((istart[0]) >> 2) & 3; ++ size_t litSize, lhSize; ++ switch (lhlCode) { ++ case 0: ++ case 2: ++ default: /* note : default is impossible, since lhlCode into [0..3] */ ++ lhSize = 1; ++ litSize = istart[0] >> 3; ++ break; ++ case 1: ++ lhSize = 2; ++ litSize = ZSTD_readLE16(istart) >> 4; ++ break; ++ case 3: ++ lhSize = 3; ++ litSize = ZSTD_readLE24(istart) >> 4; ++ if (srcSize < 4) ++ return ERROR(corruption_detected); /* srcSize >= MIN_CBLOCK_SIZE == 3; here we need lhSize+1 = 4 */ ++ break; ++ } ++ if (litSize > ZSTD_BLOCKSIZE_ABSOLUTEMAX) ++ return ERROR(corruption_detected); ++ memset(dctx->litBuffer, istart[lhSize], litSize + WILDCOPY_OVERLENGTH); ++ dctx->litPtr = dctx->litBuffer; ++ dctx->litSize = litSize; ++ return lhSize + 1; ++ } ++ default: ++ return ERROR(corruption_detected); /* impossible */ ++ } ++ } ++} ++ ++typedef union { ++ FSE_decode_t realData; ++ U32 alignedBy4; ++} FSE_decode_t4; ++ ++static const FSE_decode_t4 LL_defaultDTable[(1 << LL_DEFAULTNORMLOG) + 1] = { ++ {{LL_DEFAULTNORMLOG, 1, 1}}, /* header : tableLog, fastMode, fastMode */ ++ {{0, 0, 4}}, /* 0 : base, symbol, bits */ ++ {{16, 0, 4}}, ++ {{32, 1, 5}}, ++ {{0, 3, 5}}, ++ {{0, 4, 5}}, ++ {{0, 6, 5}}, ++ {{0, 7, 5}}, ++ {{0, 9, 5}}, ++ {{0, 10, 5}}, ++ {{0, 12, 5}}, ++ {{0, 14, 6}}, ++ {{0, 16, 5}}, ++ {{0, 18, 5}}, ++ {{0, 19, 5}}, ++ {{0, 21, 5}}, ++ {{0, 22, 5}}, ++ {{0, 24, 5}}, ++ {{32, 25, 5}}, ++ {{0, 26, 5}}, ++ {{0, 27, 6}}, ++ {{0, 29, 6}}, ++ {{0, 31, 6}}, ++ {{32, 0, 4}}, ++ {{0, 1, 4}}, ++ {{0, 2, 5}}, ++ {{32, 4, 5}}, ++ {{0, 5, 5}}, ++ {{32, 7, 5}}, ++ {{0, 8, 5}}, ++ {{32, 10, 5}}, ++ {{0, 11, 5}}, ++ {{0, 13, 6}}, ++ {{32, 16, 5}}, ++ {{0, 17, 5}}, ++ {{32, 19, 5}}, ++ {{0, 20, 5}}, ++ {{32, 22, 5}}, ++ {{0, 23, 5}}, ++ {{0, 25, 4}}, ++ {{16, 25, 4}}, ++ {{32, 26, 5}}, ++ {{0, 28, 6}}, ++ {{0, 30, 6}}, ++ {{48, 0, 4}}, ++ {{16, 1, 4}}, ++ {{32, 2, 5}}, ++ {{32, 3, 5}}, ++ {{32, 5, 5}}, ++ {{32, 6, 5}}, ++ {{32, 8, 5}}, ++ {{32, 9, 5}}, ++ {{32, 11, 5}}, ++ {{32, 12, 5}}, ++ {{0, 15, 6}}, ++ {{32, 17, 5}}, ++ {{32, 18, 5}}, ++ {{32, 20, 5}}, ++ {{32, 21, 5}}, ++ {{32, 23, 5}}, ++ {{32, 24, 5}}, ++ {{0, 35, 6}}, ++ {{0, 34, 6}}, ++ {{0, 33, 6}}, ++ {{0, 32, 6}}, ++}; /* LL_defaultDTable */ ++ ++static const FSE_decode_t4 ML_defaultDTable[(1 << ML_DEFAULTNORMLOG) + 1] = { ++ {{ML_DEFAULTNORMLOG, 1, 1}}, /* header : tableLog, fastMode, fastMode */ ++ {{0, 0, 6}}, /* 0 : base, symbol, bits */ ++ {{0, 1, 4}}, ++ {{32, 2, 5}}, ++ {{0, 3, 5}}, ++ {{0, 5, 5}}, ++ {{0, 6, 5}}, ++ {{0, 8, 5}}, ++ {{0, 10, 6}}, ++ {{0, 13, 6}}, ++ {{0, 16, 6}}, ++ {{0, 19, 6}}, ++ {{0, 22, 6}}, ++ {{0, 25, 6}}, ++ {{0, 28, 6}}, ++ {{0, 31, 6}}, ++ {{0, 33, 6}}, ++ {{0, 35, 6}}, ++ {{0, 37, 6}}, ++ {{0, 39, 6}}, ++ {{0, 41, 6}}, ++ {{0, 43, 6}}, ++ {{0, 45, 6}}, ++ {{16, 1, 4}}, ++ {{0, 2, 4}}, ++ {{32, 3, 5}}, ++ {{0, 4, 5}}, ++ {{32, 6, 5}}, ++ {{0, 7, 5}}, ++ {{0, 9, 6}}, ++ {{0, 12, 6}}, ++ {{0, 15, 6}}, ++ {{0, 18, 6}}, ++ {{0, 21, 6}}, ++ {{0, 24, 6}}, ++ {{0, 27, 6}}, ++ {{0, 30, 6}}, ++ {{0, 32, 6}}, ++ {{0, 34, 6}}, ++ {{0, 36, 6}}, ++ {{0, 38, 6}}, ++ {{0, 40, 6}}, ++ {{0, 42, 6}}, ++ {{0, 44, 6}}, ++ {{32, 1, 4}}, ++ {{48, 1, 4}}, ++ {{16, 2, 4}}, ++ {{32, 4, 5}}, ++ {{32, 5, 5}}, ++ {{32, 7, 5}}, ++ {{32, 8, 5}}, ++ {{0, 11, 6}}, ++ {{0, 14, 6}}, ++ {{0, 17, 6}}, ++ {{0, 20, 6}}, ++ {{0, 23, 6}}, ++ {{0, 26, 6}}, ++ {{0, 29, 6}}, ++ {{0, 52, 6}}, ++ {{0, 51, 6}}, ++ {{0, 50, 6}}, ++ {{0, 49, 6}}, ++ {{0, 48, 6}}, ++ {{0, 47, 6}}, ++ {{0, 46, 6}}, ++}; /* ML_defaultDTable */ ++ ++static const FSE_decode_t4 OF_defaultDTable[(1 << OF_DEFAULTNORMLOG) + 1] = { ++ {{OF_DEFAULTNORMLOG, 1, 1}}, /* header : tableLog, fastMode, fastMode */ ++ {{0, 0, 5}}, /* 0 : base, symbol, bits */ ++ {{0, 6, 4}}, ++ {{0, 9, 5}}, ++ {{0, 15, 5}}, ++ {{0, 21, 5}}, ++ {{0, 3, 5}}, ++ {{0, 7, 4}}, ++ {{0, 12, 5}}, ++ {{0, 18, 5}}, ++ {{0, 23, 5}}, ++ {{0, 5, 5}}, ++ {{0, 8, 4}}, ++ {{0, 14, 5}}, ++ {{0, 20, 5}}, ++ {{0, 2, 5}}, ++ {{16, 7, 4}}, ++ {{0, 11, 5}}, ++ {{0, 17, 5}}, ++ {{0, 22, 5}}, ++ {{0, 4, 5}}, ++ {{16, 8, 4}}, ++ {{0, 13, 5}}, ++ {{0, 19, 5}}, ++ {{0, 1, 5}}, ++ {{16, 6, 4}}, ++ {{0, 10, 5}}, ++ {{0, 16, 5}}, ++ {{0, 28, 5}}, ++ {{0, 27, 5}}, ++ {{0, 26, 5}}, ++ {{0, 25, 5}}, ++ {{0, 24, 5}}, ++}; /* OF_defaultDTable */ ++ ++/*! ZSTD_buildSeqTable() : ++ @return : nb bytes read from src, ++ or an error code if it fails, testable with ZSTD_isError() ++*/ ++static size_t INIT ZSTD_buildSeqTable(FSE_DTable *DTableSpace, const FSE_DTable **DTablePtr, symbolEncodingType_e type, U32 max, U32 maxLog, const void *src, ++ size_t srcSize, const FSE_decode_t4 *defaultTable, U32 flagRepeatTable, void *workspace, size_t workspaceSize) ++{ ++ const void *const tmpPtr = defaultTable; /* bypass strict aliasing */ ++ switch (type) { ++ case set_rle: ++ if (!srcSize) ++ return ERROR(srcSize_wrong); ++ if ((*(const BYTE *)src) > max) ++ return ERROR(corruption_detected); ++ FSE_buildDTable_rle(DTableSpace, *(const BYTE *)src); ++ *DTablePtr = DTableSpace; ++ return 1; ++ case set_basic: *DTablePtr = (const FSE_DTable *)tmpPtr; return 0; ++ case set_repeat: ++ if (!flagRepeatTable) ++ return ERROR(corruption_detected); ++ return 0; ++ default: /* impossible */ ++ case set_compressed: { ++ U32 tableLog; ++ S16 *norm = (S16 *)workspace; ++ size_t const spaceUsed32 = ALIGN(sizeof(S16) * (MaxSeq + 1), sizeof(U32)) >> 2; ++ ++ if ((spaceUsed32 << 2) > workspaceSize) ++ return ERROR(GENERIC); ++ workspace = (U32 *)workspace + spaceUsed32; ++ workspaceSize -= (spaceUsed32 << 2); ++ { ++ size_t const headerSize = FSE_readNCount(norm, &max, &tableLog, src, srcSize); ++ if (FSE_isError(headerSize)) ++ return ERROR(corruption_detected); ++ if (tableLog > maxLog) ++ return ERROR(corruption_detected); ++ FSE_buildDTable_wksp(DTableSpace, norm, max, tableLog, workspace, workspaceSize); ++ *DTablePtr = DTableSpace; ++ return headerSize; ++ } ++ } ++ } ++} ++ ++size_t INIT ZSTD_decodeSeqHeaders(ZSTD_DCtx *dctx, int *nbSeqPtr, const void *src, size_t srcSize) ++{ ++ const BYTE *const istart = (const BYTE *const)src; ++ const BYTE *const iend = istart + srcSize; ++ const BYTE *ip = istart; ++ ++ /* check */ ++ if (srcSize < MIN_SEQUENCES_SIZE) ++ return ERROR(srcSize_wrong); ++ ++ /* SeqHead */ ++ { ++ int nbSeq = *ip++; ++ if (!nbSeq) { ++ *nbSeqPtr = 0; ++ return 1; ++ } ++ if (nbSeq > 0x7F) { ++ if (nbSeq == 0xFF) { ++ if (ip + 2 > iend) ++ return ERROR(srcSize_wrong); ++ nbSeq = ZSTD_readLE16(ip) + LONGNBSEQ, ip += 2; ++ } else { ++ if (ip >= iend) ++ return ERROR(srcSize_wrong); ++ nbSeq = ((nbSeq - 0x80) << 8) + *ip++; ++ } ++ } ++ *nbSeqPtr = nbSeq; ++ } ++ ++ /* FSE table descriptors */ ++ if (ip + 4 > iend) ++ return ERROR(srcSize_wrong); /* minimum possible size */ ++ { ++ symbolEncodingType_e const LLtype = (symbolEncodingType_e)(*ip >> 6); ++ symbolEncodingType_e const OFtype = (symbolEncodingType_e)((*ip >> 4) & 3); ++ symbolEncodingType_e const MLtype = (symbolEncodingType_e)((*ip >> 2) & 3); ++ ip++; ++ ++ /* Build DTables */ ++ { ++ size_t const llhSize = ZSTD_buildSeqTable(dctx->entropy.LLTable, &dctx->LLTptr, LLtype, MaxLL, LLFSELog, ip, iend - ip, ++ LL_defaultDTable, dctx->fseEntropy, dctx->entropy.workspace, sizeof(dctx->entropy.workspace)); ++ if (ZSTD_isError(llhSize)) ++ return ERROR(corruption_detected); ++ ip += llhSize; ++ } ++ { ++ size_t const ofhSize = ZSTD_buildSeqTable(dctx->entropy.OFTable, &dctx->OFTptr, OFtype, MaxOff, OffFSELog, ip, iend - ip, ++ OF_defaultDTable, dctx->fseEntropy, dctx->entropy.workspace, sizeof(dctx->entropy.workspace)); ++ if (ZSTD_isError(ofhSize)) ++ return ERROR(corruption_detected); ++ ip += ofhSize; ++ } ++ { ++ size_t const mlhSize = ZSTD_buildSeqTable(dctx->entropy.MLTable, &dctx->MLTptr, MLtype, MaxML, MLFSELog, ip, iend - ip, ++ ML_defaultDTable, dctx->fseEntropy, dctx->entropy.workspace, sizeof(dctx->entropy.workspace)); ++ if (ZSTD_isError(mlhSize)) ++ return ERROR(corruption_detected); ++ ip += mlhSize; ++ } ++ } ++ ++ return ip - istart; ++} ++ ++typedef struct { ++ size_t litLength; ++ size_t matchLength; ++ size_t offset; ++ const BYTE *match; ++} seq_t; ++ ++typedef struct { ++ BIT_DStream_t DStream; ++ FSE_DState_t stateLL; ++ FSE_DState_t stateOffb; ++ FSE_DState_t stateML; ++ size_t prevOffset[ZSTD_REP_NUM]; ++ const BYTE *base; ++ size_t pos; ++ uPtrDiff gotoDict; ++} seqState_t; ++ ++FORCE_NOINLINE ++size_t INIT ZSTD_execSequenceLast7(BYTE *op, BYTE *const oend, seq_t sequence, const BYTE **litPtr, const BYTE *const litLimit, const BYTE *const base, ++ const BYTE *const vBase, const BYTE *const dictEnd) ++{ ++ BYTE *const oLitEnd = op + sequence.litLength; ++ size_t const sequenceLength = sequence.litLength + sequence.matchLength; ++ BYTE *const oMatchEnd = op + sequenceLength; /* risk : address space overflow (32-bits) */ ++ BYTE *const oend_w = oend - WILDCOPY_OVERLENGTH; ++ const BYTE *const iLitEnd = *litPtr + sequence.litLength; ++ const BYTE *match = oLitEnd - sequence.offset; ++ ++ /* check */ ++ if (oMatchEnd > oend) ++ return ERROR(dstSize_tooSmall); /* last match must start at a minimum distance of WILDCOPY_OVERLENGTH from oend */ ++ if (iLitEnd > litLimit) ++ return ERROR(corruption_detected); /* over-read beyond lit buffer */ ++ if (oLitEnd <= oend_w) ++ return ERROR(GENERIC); /* Precondition */ ++ ++ /* copy literals */ ++ if (op < oend_w) { ++ ZSTD_wildcopy(op, *litPtr, oend_w - op); ++ *litPtr += oend_w - op; ++ op = oend_w; ++ } ++ while (op < oLitEnd) ++ *op++ = *(*litPtr)++; ++ ++ /* copy Match */ ++ if (sequence.offset > (size_t)(oLitEnd - base)) { ++ /* offset beyond prefix */ ++ if (sequence.offset > (size_t)(oLitEnd - vBase)) ++ return ERROR(corruption_detected); ++ match = dictEnd - (base - match); ++ if (match + sequence.matchLength <= dictEnd) { ++ memmove(oLitEnd, match, sequence.matchLength); ++ return sequenceLength; ++ } ++ /* span extDict & currPrefixSegment */ ++ { ++ size_t const length1 = dictEnd - match; ++ memmove(oLitEnd, match, length1); ++ op = oLitEnd + length1; ++ sequence.matchLength -= length1; ++ match = base; ++ } ++ } ++ while (op < oMatchEnd) ++ *op++ = *match++; ++ return sequenceLength; ++} ++ ++static seq_t INIT ZSTD_decodeSequence(seqState_t *seqState) ++{ ++ seq_t seq; ++ ++ U32 const llCode = FSE_peekSymbol(&seqState->stateLL); ++ U32 const mlCode = FSE_peekSymbol(&seqState->stateML); ++ U32 const ofCode = FSE_peekSymbol(&seqState->stateOffb); /* <= maxOff, by table construction */ ++ ++ U32 const llBits = LL_bits[llCode]; ++ U32 const mlBits = ML_bits[mlCode]; ++ U32 const ofBits = ofCode; ++ U32 const totalBits = llBits + mlBits + ofBits; ++ ++ static const U32 LL_base[MaxLL + 1] = {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 18, ++ 20, 22, 24, 28, 32, 40, 48, 64, 0x80, 0x100, 0x200, 0x400, 0x800, 0x1000, 0x2000, 0x4000, 0x8000, 0x10000}; ++ ++ static const U32 ML_base[MaxML + 1] = {3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, ++ 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 37, 39, 41, ++ 43, 47, 51, 59, 67, 83, 99, 0x83, 0x103, 0x203, 0x403, 0x803, 0x1003, 0x2003, 0x4003, 0x8003, 0x10003}; ++ ++ static const U32 OF_base[MaxOff + 1] = {0, 1, 1, 5, 0xD, 0x1D, 0x3D, 0x7D, 0xFD, 0x1FD, ++ 0x3FD, 0x7FD, 0xFFD, 0x1FFD, 0x3FFD, 0x7FFD, 0xFFFD, 0x1FFFD, 0x3FFFD, 0x7FFFD, ++ 0xFFFFD, 0x1FFFFD, 0x3FFFFD, 0x7FFFFD, 0xFFFFFD, 0x1FFFFFD, 0x3FFFFFD, 0x7FFFFFD, 0xFFFFFFD}; ++ ++ /* sequence */ ++ { ++ size_t offset; ++ if (!ofCode) ++ offset = 0; ++ else { ++ offset = OF_base[ofCode] + BIT_readBitsFast(&seqState->DStream, ofBits); /* <= (ZSTD_WINDOWLOG_MAX-1) bits */ ++ if (ZSTD_32bits()) ++ BIT_reloadDStream(&seqState->DStream); ++ } ++ ++ if (ofCode <= 1) { ++ offset += (llCode == 0); ++ if (offset) { ++ size_t temp = (offset == 3) ? seqState->prevOffset[0] - 1 : seqState->prevOffset[offset]; ++ temp += !temp; /* 0 is not valid; input is corrupted; force offset to 1 */ ++ if (offset != 1) ++ seqState->prevOffset[2] = seqState->prevOffset[1]; ++ seqState->prevOffset[1] = seqState->prevOffset[0]; ++ seqState->prevOffset[0] = offset = temp; ++ } else { ++ offset = seqState->prevOffset[0]; ++ } ++ } else { ++ seqState->prevOffset[2] = seqState->prevOffset[1]; ++ seqState->prevOffset[1] = seqState->prevOffset[0]; ++ seqState->prevOffset[0] = offset; ++ } ++ seq.offset = offset; ++ } ++ ++ seq.matchLength = ML_base[mlCode] + ((mlCode > 31) ? BIT_readBitsFast(&seqState->DStream, mlBits) : 0); /* <= 16 bits */ ++ if (ZSTD_32bits() && (mlBits + llBits > 24)) ++ BIT_reloadDStream(&seqState->DStream); ++ ++ seq.litLength = LL_base[llCode] + ((llCode > 15) ? BIT_readBitsFast(&seqState->DStream, llBits) : 0); /* <= 16 bits */ ++ if (ZSTD_32bits() || (totalBits > 64 - 7 - (LLFSELog + MLFSELog + OffFSELog))) ++ BIT_reloadDStream(&seqState->DStream); ++ ++ /* ANS state update */ ++ FSE_updateState(&seqState->stateLL, &seqState->DStream); /* <= 9 bits */ ++ FSE_updateState(&seqState->stateML, &seqState->DStream); /* <= 9 bits */ ++ if (ZSTD_32bits()) ++ BIT_reloadDStream(&seqState->DStream); /* <= 18 bits */ ++ FSE_updateState(&seqState->stateOffb, &seqState->DStream); /* <= 8 bits */ ++ ++ seq.match = NULL; ++ ++ return seq; ++} ++ ++FORCE_INLINE ++size_t ZSTD_execSequence(BYTE *op, BYTE *const oend, seq_t sequence, const BYTE **litPtr, const BYTE *const litLimit, const BYTE *const base, ++ const BYTE *const vBase, const BYTE *const dictEnd) ++{ ++ BYTE *const oLitEnd = op + sequence.litLength; ++ size_t const sequenceLength = sequence.litLength + sequence.matchLength; ++ BYTE *const oMatchEnd = op + sequenceLength; /* risk : address space overflow (32-bits) */ ++ BYTE *const oend_w = oend - WILDCOPY_OVERLENGTH; ++ const BYTE *const iLitEnd = *litPtr + sequence.litLength; ++ const BYTE *match = oLitEnd - sequence.offset; ++ ++ /* check */ ++ if (oMatchEnd > oend) ++ return ERROR(dstSize_tooSmall); /* last match must start at a minimum distance of WILDCOPY_OVERLENGTH from oend */ ++ if (iLitEnd > litLimit) ++ return ERROR(corruption_detected); /* over-read beyond lit buffer */ ++ if (oLitEnd > oend_w) ++ return ZSTD_execSequenceLast7(op, oend, sequence, litPtr, litLimit, base, vBase, dictEnd); ++ ++ /* copy Literals */ ++ ZSTD_copy8(op, *litPtr); ++ if (sequence.litLength > 8) ++ ZSTD_wildcopy(op + 8, (*litPtr) + 8, ++ sequence.litLength - 8); /* note : since oLitEnd <= oend-WILDCOPY_OVERLENGTH, no risk of overwrite beyond oend */ ++ op = oLitEnd; ++ *litPtr = iLitEnd; /* update for next sequence */ ++ ++ /* copy Match */ ++ if (sequence.offset > (size_t)(oLitEnd - base)) { ++ /* offset beyond prefix */ ++ if (sequence.offset > (size_t)(oLitEnd - vBase)) ++ return ERROR(corruption_detected); ++ match = dictEnd + (match - base); ++ if (match + sequence.matchLength <= dictEnd) { ++ memmove(oLitEnd, match, sequence.matchLength); ++ return sequenceLength; ++ } ++ /* span extDict & currPrefixSegment */ ++ { ++ size_t const length1 = dictEnd - match; ++ memmove(oLitEnd, match, length1); ++ op = oLitEnd + length1; ++ sequence.matchLength -= length1; ++ match = base; ++ if (op > oend_w || sequence.matchLength < MINMATCH) { ++ U32 i; ++ for (i = 0; i < sequence.matchLength; ++i) ++ op[i] = match[i]; ++ return sequenceLength; ++ } ++ } ++ } ++ /* Requirement: op <= oend_w && sequence.matchLength >= MINMATCH */ ++ ++ /* match within prefix */ ++ if (sequence.offset < 8) { ++ /* close range match, overlap */ ++ static const U32 dec32table[] = {0, 1, 2, 1, 4, 4, 4, 4}; /* added */ ++ static const int dec64table[] = {8, 8, 8, 7, 8, 9, 10, 11}; /* subtracted */ ++ int const sub2 = dec64table[sequence.offset]; ++ op[0] = match[0]; ++ op[1] = match[1]; ++ op[2] = match[2]; ++ op[3] = match[3]; ++ match += dec32table[sequence.offset]; ++ ZSTD_copy4(op + 4, match); ++ match -= sub2; ++ } else { ++ ZSTD_copy8(op, match); ++ } ++ op += 8; ++ match += 8; ++ ++ if (oMatchEnd > oend - (16 - MINMATCH)) { ++ if (op < oend_w) { ++ ZSTD_wildcopy(op, match, oend_w - op); ++ match += oend_w - op; ++ op = oend_w; ++ } ++ while (op < oMatchEnd) ++ *op++ = *match++; ++ } else { ++ ZSTD_wildcopy(op, match, (ptrdiff_t)sequence.matchLength - 8); /* works even if matchLength < 8 */ ++ } ++ return sequenceLength; ++} ++ ++static size_t INIT ZSTD_decompressSequences(ZSTD_DCtx *dctx, void *dst, size_t maxDstSize, const void *seqStart, size_t seqSize) ++{ ++ const BYTE *ip = (const BYTE *)seqStart; ++ const BYTE *const iend = ip + seqSize; ++ BYTE *const ostart = (BYTE * const)dst; ++ BYTE *const oend = ostart + maxDstSize; ++ BYTE *op = ostart; ++ const BYTE *litPtr = dctx->litPtr; ++ const BYTE *const litEnd = litPtr + dctx->litSize; ++ const BYTE *const base = (const BYTE *)(dctx->base); ++ const BYTE *const vBase = (const BYTE *)(dctx->vBase); ++ const BYTE *const dictEnd = (const BYTE *)(dctx->dictEnd); ++ int nbSeq; ++ ++ /* Build Decoding Tables */ ++ { ++ size_t const seqHSize = ZSTD_decodeSeqHeaders(dctx, &nbSeq, ip, seqSize); ++ if (ZSTD_isError(seqHSize)) ++ return seqHSize; ++ ip += seqHSize; ++ } ++ ++ /* Regen sequences */ ++ if (nbSeq) { ++ seqState_t seqState; ++ dctx->fseEntropy = 1; ++ { ++ U32 i; ++ for (i = 0; i < ZSTD_REP_NUM; i++) ++ seqState.prevOffset[i] = dctx->entropy.rep[i]; ++ } ++ CHECK_E(BIT_initDStream(&seqState.DStream, ip, iend - ip), corruption_detected); ++ FSE_initDState(&seqState.stateLL, &seqState.DStream, dctx->LLTptr); ++ FSE_initDState(&seqState.stateOffb, &seqState.DStream, dctx->OFTptr); ++ FSE_initDState(&seqState.stateML, &seqState.DStream, dctx->MLTptr); ++ ++ for (; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && nbSeq;) { ++ nbSeq--; ++ { ++ seq_t const sequence = ZSTD_decodeSequence(&seqState); ++ size_t const oneSeqSize = ZSTD_execSequence(op, oend, sequence, &litPtr, litEnd, base, vBase, dictEnd); ++ if (ZSTD_isError(oneSeqSize)) ++ return oneSeqSize; ++ op += oneSeqSize; ++ } ++ } ++ ++ /* check if reached exact end */ ++ if (nbSeq) ++ return ERROR(corruption_detected); ++ /* save reps for next block */ ++ { ++ U32 i; ++ for (i = 0; i < ZSTD_REP_NUM; i++) ++ dctx->entropy.rep[i] = (U32)(seqState.prevOffset[i]); ++ } ++ } ++ ++ /* last literal segment */ ++ { ++ size_t const lastLLSize = litEnd - litPtr; ++ if (lastLLSize > (size_t)(oend - op)) ++ return ERROR(dstSize_tooSmall); ++ memcpy(op, litPtr, lastLLSize); ++ op += lastLLSize; ++ } ++ ++ return op - ostart; ++} ++ ++FORCE_INLINE seq_t INIT ZSTD_decodeSequenceLong_generic(seqState_t *seqState, int const longOffsets) ++{ ++ seq_t seq; ++ ++ U32 const llCode = FSE_peekSymbol(&seqState->stateLL); ++ U32 const mlCode = FSE_peekSymbol(&seqState->stateML); ++ U32 const ofCode = FSE_peekSymbol(&seqState->stateOffb); /* <= maxOff, by table construction */ ++ ++ U32 const llBits = LL_bits[llCode]; ++ U32 const mlBits = ML_bits[mlCode]; ++ U32 const ofBits = ofCode; ++ U32 const totalBits = llBits + mlBits + ofBits; ++ ++ static const U32 LL_base[MaxLL + 1] = {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 18, ++ 20, 22, 24, 28, 32, 40, 48, 64, 0x80, 0x100, 0x200, 0x400, 0x800, 0x1000, 0x2000, 0x4000, 0x8000, 0x10000}; ++ ++ static const U32 ML_base[MaxML + 1] = {3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, ++ 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 37, 39, 41, ++ 43, 47, 51, 59, 67, 83, 99, 0x83, 0x103, 0x203, 0x403, 0x803, 0x1003, 0x2003, 0x4003, 0x8003, 0x10003}; ++ ++ static const U32 OF_base[MaxOff + 1] = {0, 1, 1, 5, 0xD, 0x1D, 0x3D, 0x7D, 0xFD, 0x1FD, ++ 0x3FD, 0x7FD, 0xFFD, 0x1FFD, 0x3FFD, 0x7FFD, 0xFFFD, 0x1FFFD, 0x3FFFD, 0x7FFFD, ++ 0xFFFFD, 0x1FFFFD, 0x3FFFFD, 0x7FFFFD, 0xFFFFFD, 0x1FFFFFD, 0x3FFFFFD, 0x7FFFFFD, 0xFFFFFFD}; ++ ++ /* sequence */ ++ { ++ size_t offset; ++ if (!ofCode) ++ offset = 0; ++ else { ++ if (longOffsets) { ++ int const extraBits = ofBits - MIN(ofBits, STREAM_ACCUMULATOR_MIN); ++ offset = OF_base[ofCode] + (BIT_readBitsFast(&seqState->DStream, ofBits - extraBits) << extraBits); ++ if (ZSTD_32bits() || extraBits) ++ BIT_reloadDStream(&seqState->DStream); ++ if (extraBits) ++ offset += BIT_readBitsFast(&seqState->DStream, extraBits); ++ } else { ++ offset = OF_base[ofCode] + BIT_readBitsFast(&seqState->DStream, ofBits); /* <= (ZSTD_WINDOWLOG_MAX-1) bits */ ++ if (ZSTD_32bits()) ++ BIT_reloadDStream(&seqState->DStream); ++ } ++ } ++ ++ if (ofCode <= 1) { ++ offset += (llCode == 0); ++ if (offset) { ++ size_t temp = (offset == 3) ? seqState->prevOffset[0] - 1 : seqState->prevOffset[offset]; ++ temp += !temp; /* 0 is not valid; input is corrupted; force offset to 1 */ ++ if (offset != 1) ++ seqState->prevOffset[2] = seqState->prevOffset[1]; ++ seqState->prevOffset[1] = seqState->prevOffset[0]; ++ seqState->prevOffset[0] = offset = temp; ++ } else { ++ offset = seqState->prevOffset[0]; ++ } ++ } else { ++ seqState->prevOffset[2] = seqState->prevOffset[1]; ++ seqState->prevOffset[1] = seqState->prevOffset[0]; ++ seqState->prevOffset[0] = offset; ++ } ++ seq.offset = offset; ++ } ++ ++ seq.matchLength = ML_base[mlCode] + ((mlCode > 31) ? BIT_readBitsFast(&seqState->DStream, mlBits) : 0); /* <= 16 bits */ ++ if (ZSTD_32bits() && (mlBits + llBits > 24)) ++ BIT_reloadDStream(&seqState->DStream); ++ ++ seq.litLength = LL_base[llCode] + ((llCode > 15) ? BIT_readBitsFast(&seqState->DStream, llBits) : 0); /* <= 16 bits */ ++ if (ZSTD_32bits() || (totalBits > 64 - 7 - (LLFSELog + MLFSELog + OffFSELog))) ++ BIT_reloadDStream(&seqState->DStream); ++ ++ { ++ size_t const pos = seqState->pos + seq.litLength; ++ seq.match = seqState->base + pos - seq.offset; /* single memory segment */ ++ if (seq.offset > pos) ++ seq.match += seqState->gotoDict; /* separate memory segment */ ++ seqState->pos = pos + seq.matchLength; ++ } ++ ++ /* ANS state update */ ++ FSE_updateState(&seqState->stateLL, &seqState->DStream); /* <= 9 bits */ ++ FSE_updateState(&seqState->stateML, &seqState->DStream); /* <= 9 bits */ ++ if (ZSTD_32bits()) ++ BIT_reloadDStream(&seqState->DStream); /* <= 18 bits */ ++ FSE_updateState(&seqState->stateOffb, &seqState->DStream); /* <= 8 bits */ ++ ++ return seq; ++} ++ ++static seq_t INIT ZSTD_decodeSequenceLong(seqState_t *seqState, unsigned const windowSize) ++{ ++ if (ZSTD_highbit32(windowSize) > STREAM_ACCUMULATOR_MIN) { ++ return ZSTD_decodeSequenceLong_generic(seqState, 1); ++ } else { ++ return ZSTD_decodeSequenceLong_generic(seqState, 0); ++ } ++} ++ ++FORCE_INLINE ++size_t ZSTD_execSequenceLong(BYTE *op, BYTE *const oend, seq_t sequence, const BYTE **litPtr, const BYTE *const litLimit, const BYTE *const base, ++ const BYTE *const vBase, const BYTE *const dictEnd) ++{ ++ BYTE *const oLitEnd = op + sequence.litLength; ++ size_t const sequenceLength = sequence.litLength + sequence.matchLength; ++ BYTE *const oMatchEnd = op + sequenceLength; /* risk : address space overflow (32-bits) */ ++ BYTE *const oend_w = oend - WILDCOPY_OVERLENGTH; ++ const BYTE *const iLitEnd = *litPtr + sequence.litLength; ++ const BYTE *match = sequence.match; ++ ++ /* check */ ++ if (oMatchEnd > oend) ++ return ERROR(dstSize_tooSmall); /* last match must start at a minimum distance of WILDCOPY_OVERLENGTH from oend */ ++ if (iLitEnd > litLimit) ++ return ERROR(corruption_detected); /* over-read beyond lit buffer */ ++ if (oLitEnd > oend_w) ++ return ZSTD_execSequenceLast7(op, oend, sequence, litPtr, litLimit, base, vBase, dictEnd); ++ ++ /* copy Literals */ ++ ZSTD_copy8(op, *litPtr); ++ if (sequence.litLength > 8) ++ ZSTD_wildcopy(op + 8, (*litPtr) + 8, ++ sequence.litLength - 8); /* note : since oLitEnd <= oend-WILDCOPY_OVERLENGTH, no risk of overwrite beyond oend */ ++ op = oLitEnd; ++ *litPtr = iLitEnd; /* update for next sequence */ ++ ++ /* copy Match */ ++ if (sequence.offset > (size_t)(oLitEnd - base)) { ++ /* offset beyond prefix */ ++ if (sequence.offset > (size_t)(oLitEnd - vBase)) ++ return ERROR(corruption_detected); ++ if (match + sequence.matchLength <= dictEnd) { ++ memmove(oLitEnd, match, sequence.matchLength); ++ return sequenceLength; ++ } ++ /* span extDict & currPrefixSegment */ ++ { ++ size_t const length1 = dictEnd - match; ++ memmove(oLitEnd, match, length1); ++ op = oLitEnd + length1; ++ sequence.matchLength -= length1; ++ match = base; ++ if (op > oend_w || sequence.matchLength < MINMATCH) { ++ U32 i; ++ for (i = 0; i < sequence.matchLength; ++i) ++ op[i] = match[i]; ++ return sequenceLength; ++ } ++ } ++ } ++ /* Requirement: op <= oend_w && sequence.matchLength >= MINMATCH */ ++ ++ /* match within prefix */ ++ if (sequence.offset < 8) { ++ /* close range match, overlap */ ++ static const U32 dec32table[] = {0, 1, 2, 1, 4, 4, 4, 4}; /* added */ ++ static const int dec64table[] = {8, 8, 8, 7, 8, 9, 10, 11}; /* subtracted */ ++ int const sub2 = dec64table[sequence.offset]; ++ op[0] = match[0]; ++ op[1] = match[1]; ++ op[2] = match[2]; ++ op[3] = match[3]; ++ match += dec32table[sequence.offset]; ++ ZSTD_copy4(op + 4, match); ++ match -= sub2; ++ } else { ++ ZSTD_copy8(op, match); ++ } ++ op += 8; ++ match += 8; ++ ++ if (oMatchEnd > oend - (16 - MINMATCH)) { ++ if (op < oend_w) { ++ ZSTD_wildcopy(op, match, oend_w - op); ++ match += oend_w - op; ++ op = oend_w; ++ } ++ while (op < oMatchEnd) ++ *op++ = *match++; ++ } else { ++ ZSTD_wildcopy(op, match, (ptrdiff_t)sequence.matchLength - 8); /* works even if matchLength < 8 */ ++ } ++ return sequenceLength; ++} ++ ++static size_t INIT ZSTD_decompressSequencesLong(ZSTD_DCtx *dctx, void *dst, size_t maxDstSize, const void *seqStart, size_t seqSize) ++{ ++ const BYTE *ip = (const BYTE *)seqStart; ++ const BYTE *const iend = ip + seqSize; ++ BYTE *const ostart = (BYTE * const)dst; ++ BYTE *const oend = ostart + maxDstSize; ++ BYTE *op = ostart; ++ const BYTE *litPtr = dctx->litPtr; ++ const BYTE *const litEnd = litPtr + dctx->litSize; ++ const BYTE *const base = (const BYTE *)(dctx->base); ++ const BYTE *const vBase = (const BYTE *)(dctx->vBase); ++ const BYTE *const dictEnd = (const BYTE *)(dctx->dictEnd); ++ unsigned const windowSize = dctx->fParams.windowSize; ++ int nbSeq; ++ ++ /* Build Decoding Tables */ ++ { ++ size_t const seqHSize = ZSTD_decodeSeqHeaders(dctx, &nbSeq, ip, seqSize); ++ if (ZSTD_isError(seqHSize)) ++ return seqHSize; ++ ip += seqHSize; ++ } ++ ++ /* Regen sequences */ ++ if (nbSeq) { ++#define STORED_SEQS 4 ++#define STOSEQ_MASK (STORED_SEQS - 1) ++#define ADVANCED_SEQS 4 ++ seq_t *sequences = (seq_t *)dctx->entropy.workspace; ++ int const seqAdvance = MIN(nbSeq, ADVANCED_SEQS); ++ seqState_t seqState; ++ int seqNb; ++ ZSTD_STATIC_ASSERT(sizeof(dctx->entropy.workspace) >= sizeof(seq_t) * STORED_SEQS); ++ dctx->fseEntropy = 1; ++ { ++ U32 i; ++ for (i = 0; i < ZSTD_REP_NUM; i++) ++ seqState.prevOffset[i] = dctx->entropy.rep[i]; ++ } ++ seqState.base = base; ++ seqState.pos = (size_t)(op - base); ++ seqState.gotoDict = (uPtrDiff)dictEnd - (uPtrDiff)base; /* cast to avoid undefined behaviour */ ++ CHECK_E(BIT_initDStream(&seqState.DStream, ip, iend - ip), corruption_detected); ++ FSE_initDState(&seqState.stateLL, &seqState.DStream, dctx->LLTptr); ++ FSE_initDState(&seqState.stateOffb, &seqState.DStream, dctx->OFTptr); ++ FSE_initDState(&seqState.stateML, &seqState.DStream, dctx->MLTptr); ++ ++ /* prepare in advance */ ++ for (seqNb = 0; (BIT_reloadDStream(&seqState.DStream) <= BIT_DStream_completed) && seqNb < seqAdvance; seqNb++) { ++ sequences[seqNb] = ZSTD_decodeSequenceLong(&seqState, windowSize); ++ } ++ if (seqNb < seqAdvance) ++ return ERROR(corruption_detected); ++ ++ /* decode and decompress */ ++ for (; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && seqNb < nbSeq; seqNb++) { ++ seq_t const sequence = ZSTD_decodeSequenceLong(&seqState, windowSize); ++ size_t const oneSeqSize = ++ ZSTD_execSequenceLong(op, oend, sequences[(seqNb - ADVANCED_SEQS) & STOSEQ_MASK], &litPtr, litEnd, base, vBase, dictEnd); ++ if (ZSTD_isError(oneSeqSize)) ++ return oneSeqSize; ++ ZSTD_PREFETCH(sequence.match); ++ sequences[seqNb & STOSEQ_MASK] = sequence; ++ op += oneSeqSize; ++ } ++ if (seqNb < nbSeq) ++ return ERROR(corruption_detected); ++ ++ /* finish queue */ ++ seqNb -= seqAdvance; ++ for (; seqNb < nbSeq; seqNb++) { ++ size_t const oneSeqSize = ZSTD_execSequenceLong(op, oend, sequences[seqNb & STOSEQ_MASK], &litPtr, litEnd, base, vBase, dictEnd); ++ if (ZSTD_isError(oneSeqSize)) ++ return oneSeqSize; ++ op += oneSeqSize; ++ } ++ ++ /* save reps for next block */ ++ { ++ U32 i; ++ for (i = 0; i < ZSTD_REP_NUM; i++) ++ dctx->entropy.rep[i] = (U32)(seqState.prevOffset[i]); ++ } ++ } ++ ++ /* last literal segment */ ++ { ++ size_t const lastLLSize = litEnd - litPtr; ++ if (lastLLSize > (size_t)(oend - op)) ++ return ERROR(dstSize_tooSmall); ++ memcpy(op, litPtr, lastLLSize); ++ op += lastLLSize; ++ } ++ ++ return op - ostart; ++} ++ ++static size_t INIT ZSTD_decompressBlock_internal(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize) ++{ /* blockType == blockCompressed */ ++ const BYTE *ip = (const BYTE *)src; ++ ++ if (srcSize >= ZSTD_BLOCKSIZE_ABSOLUTEMAX) ++ return ERROR(srcSize_wrong); ++ ++ /* Decode literals section */ ++ { ++ size_t const litCSize = ZSTD_decodeLiteralsBlock(dctx, src, srcSize); ++ if (ZSTD_isError(litCSize)) ++ return litCSize; ++ ip += litCSize; ++ srcSize -= litCSize; ++ } ++ if (sizeof(size_t) > 4) /* do not enable prefetching on 32-bits x86, as it's performance detrimental */ ++ /* likely because of register pressure */ ++ /* if that's the correct cause, then 32-bits ARM should be affected differently */ ++ /* it would be good to test this on ARM real hardware, to see if prefetch version improves speed */ ++ if (dctx->fParams.windowSize > (1 << 23)) ++ return ZSTD_decompressSequencesLong(dctx, dst, dstCapacity, ip, srcSize); ++ return ZSTD_decompressSequences(dctx, dst, dstCapacity, ip, srcSize); ++} ++ ++static void INIT ZSTD_checkContinuity(ZSTD_DCtx *dctx, const void *dst) ++{ ++ if (dst != dctx->previousDstEnd) { /* not contiguous */ ++ dctx->dictEnd = dctx->previousDstEnd; ++ dctx->vBase = (const char *)dst - ((const char *)(dctx->previousDstEnd) - (const char *)(dctx->base)); ++ dctx->base = dst; ++ dctx->previousDstEnd = dst; ++ } ++} ++ ++size_t INIT ZSTD_decompressBlock(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize) ++{ ++ size_t dSize; ++ ZSTD_checkContinuity(dctx, dst); ++ dSize = ZSTD_decompressBlock_internal(dctx, dst, dstCapacity, src, srcSize); ++ dctx->previousDstEnd = (char *)dst + dSize; ++ return dSize; ++} ++ ++/** ZSTD_insertBlock() : ++ insert `src` block into `dctx` history. Useful to track uncompressed blocks. */ ++size_t INIT ZSTD_insertBlock(ZSTD_DCtx *dctx, const void *blockStart, size_t blockSize) ++{ ++ ZSTD_checkContinuity(dctx, blockStart); ++ dctx->previousDstEnd = (const char *)blockStart + blockSize; ++ return blockSize; ++} ++ ++size_t INIT ZSTD_generateNxBytes(void *dst, size_t dstCapacity, BYTE byte, size_t length) ++{ ++ if (length > dstCapacity) ++ return ERROR(dstSize_tooSmall); ++ memset(dst, byte, length); ++ return length; ++} ++ ++/** ZSTD_findFrameCompressedSize() : ++ * compatible with legacy mode ++ * `src` must point to the start of a ZSTD frame, ZSTD legacy frame, or skippable frame ++ * `srcSize` must be at least as large as the frame contained ++ * @return : the compressed size of the frame starting at `src` */ ++size_t INIT ZSTD_findFrameCompressedSize(const void *src, size_t srcSize) ++{ ++ if (srcSize >= ZSTD_skippableHeaderSize && (ZSTD_readLE32(src) & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) { ++ return ZSTD_skippableHeaderSize + ZSTD_readLE32((const BYTE *)src + 4); ++ } else { ++ const BYTE *ip = (const BYTE *)src; ++ const BYTE *const ipstart = ip; ++ size_t remainingSize = srcSize; ++ ZSTD_frameParams fParams; ++ ++ size_t const headerSize = ZSTD_frameHeaderSize(ip, remainingSize); ++ if (ZSTD_isError(headerSize)) ++ return headerSize; ++ ++ /* Frame Header */ ++ { ++ size_t const ret = ZSTD_getFrameParams(&fParams, ip, remainingSize); ++ if (ZSTD_isError(ret)) ++ return ret; ++ if (ret > 0) ++ return ERROR(srcSize_wrong); ++ } ++ ++ ip += headerSize; ++ remainingSize -= headerSize; ++ ++ /* Loop on each block */ ++ while (1) { ++ blockProperties_t blockProperties; ++ size_t const cBlockSize = ZSTD_getcBlockSize(ip, remainingSize, &blockProperties); ++ if (ZSTD_isError(cBlockSize)) ++ return cBlockSize; ++ ++ if (ZSTD_blockHeaderSize + cBlockSize > remainingSize) ++ return ERROR(srcSize_wrong); ++ ++ ip += ZSTD_blockHeaderSize + cBlockSize; ++ remainingSize -= ZSTD_blockHeaderSize + cBlockSize; ++ ++ if (blockProperties.lastBlock) ++ break; ++ } ++ ++ if (fParams.checksumFlag) { /* Frame content checksum */ ++ if (remainingSize < 4) ++ return ERROR(srcSize_wrong); ++ ip += 4; ++ remainingSize -= 4; ++ } ++ ++ return ip - ipstart; ++ } ++} ++ ++/*! ZSTD_decompressFrame() : ++* @dctx must be properly initialized */ ++static size_t INIT ZSTD_decompressFrame(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, const void **srcPtr, size_t *srcSizePtr) ++{ ++ const BYTE *ip = (const BYTE *)(*srcPtr); ++ BYTE *const ostart = (BYTE * const)dst; ++ BYTE *const oend = ostart + dstCapacity; ++ BYTE *op = ostart; ++ size_t remainingSize = *srcSizePtr; ++ ++ /* check */ ++ if (remainingSize < ZSTD_frameHeaderSize_min + ZSTD_blockHeaderSize) ++ return ERROR(srcSize_wrong); ++ ++ /* Frame Header */ ++ { ++ size_t const frameHeaderSize = ZSTD_frameHeaderSize(ip, ZSTD_frameHeaderSize_prefix); ++ if (ZSTD_isError(frameHeaderSize)) ++ return frameHeaderSize; ++ if (remainingSize < frameHeaderSize + ZSTD_blockHeaderSize) ++ return ERROR(srcSize_wrong); ++ CHECK_F(ZSTD_decodeFrameHeader(dctx, ip, frameHeaderSize)); ++ ip += frameHeaderSize; ++ remainingSize -= frameHeaderSize; ++ } ++ ++ /* Loop on each block */ ++ while (1) { ++ size_t decodedSize; ++ blockProperties_t blockProperties; ++ size_t const cBlockSize = ZSTD_getcBlockSize(ip, remainingSize, &blockProperties); ++ if (ZSTD_isError(cBlockSize)) ++ return cBlockSize; ++ ++ ip += ZSTD_blockHeaderSize; ++ remainingSize -= ZSTD_blockHeaderSize; ++ if (cBlockSize > remainingSize) ++ return ERROR(srcSize_wrong); ++ ++ switch (blockProperties.blockType) { ++ case bt_compressed: decodedSize = ZSTD_decompressBlock_internal(dctx, op, oend - op, ip, cBlockSize); break; ++ case bt_raw: decodedSize = ZSTD_copyRawBlock(op, oend - op, ip, cBlockSize); break; ++ case bt_rle: decodedSize = ZSTD_generateNxBytes(op, oend - op, *ip, blockProperties.origSize); break; ++ case bt_reserved: ++ default: return ERROR(corruption_detected); ++ } ++ ++ if (ZSTD_isError(decodedSize)) ++ return decodedSize; ++ if (dctx->fParams.checksumFlag) ++ xxh64_update(&dctx->xxhState, op, decodedSize); ++ op += decodedSize; ++ ip += cBlockSize; ++ remainingSize -= cBlockSize; ++ if (blockProperties.lastBlock) ++ break; ++ } ++ ++ if (dctx->fParams.checksumFlag) { /* Frame content checksum verification */ ++ U32 const checkCalc = (U32)xxh64_digest(&dctx->xxhState); ++ U32 checkRead; ++ if (remainingSize < 4) ++ return ERROR(checksum_wrong); ++ checkRead = ZSTD_readLE32(ip); ++ if (checkRead != checkCalc) ++ return ERROR(checksum_wrong); ++ ip += 4; ++ remainingSize -= 4; ++ } ++ ++ /* Allow caller to get size read */ ++ *srcPtr = ip; ++ *srcSizePtr = remainingSize; ++ return op - ostart; ++} ++ ++static const void INIT *ZSTD_DDictDictContent(const ZSTD_DDict *ddict); ++static size_t INIT ZSTD_DDictDictSize(const ZSTD_DDict *ddict); ++ ++static size_t INIT ZSTD_decompressMultiFrame(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize, const void *dict, size_t dictSize, ++ const ZSTD_DDict *ddict) ++{ ++ void *const dststart = dst; ++ ++ if (ddict) { ++ if (dict) { ++ /* programmer error, these two cases should be mutually exclusive */ ++ return ERROR(GENERIC); ++ } ++ ++ dict = ZSTD_DDictDictContent(ddict); ++ dictSize = ZSTD_DDictDictSize(ddict); ++ } ++ ++ while (srcSize >= ZSTD_frameHeaderSize_prefix) { ++ U32 magicNumber; ++ ++ magicNumber = ZSTD_readLE32(src); ++ if (magicNumber != ZSTD_MAGICNUMBER) { ++ if ((magicNumber & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) { ++ size_t skippableSize; ++ if (srcSize < ZSTD_skippableHeaderSize) ++ return ERROR(srcSize_wrong); ++ skippableSize = ZSTD_readLE32((const BYTE *)src + 4) + ZSTD_skippableHeaderSize; ++ if (srcSize < skippableSize) { ++ return ERROR(srcSize_wrong); ++ } ++ ++ src = (const BYTE *)src + skippableSize; ++ srcSize -= skippableSize; ++ continue; ++ } else { ++ return ERROR(prefix_unknown); ++ } ++ } ++ ++ if (ddict) { ++ /* we were called from ZSTD_decompress_usingDDict */ ++ ZSTD_refDDict(dctx, ddict); ++ } else { ++ /* this will initialize correctly with no dict if dict == NULL, so ++ * use this in all cases but ddict */ ++ CHECK_F(ZSTD_decompressBegin_usingDict(dctx, dict, dictSize)); ++ } ++ ZSTD_checkContinuity(dctx, dst); ++ ++ { ++ const size_t res = ZSTD_decompressFrame(dctx, dst, dstCapacity, &src, &srcSize); ++ if (ZSTD_isError(res)) ++ return res; ++ /* don't need to bounds check this, ZSTD_decompressFrame will have ++ * already */ ++ dst = (BYTE *)dst + res; ++ dstCapacity -= res; ++ } ++ } ++ ++ if (srcSize) ++ return ERROR(srcSize_wrong); /* input not entirely consumed */ ++ ++ return (BYTE *)dst - (BYTE *)dststart; ++} ++ ++size_t INIT ZSTD_decompress_usingDict(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize, const void *dict, size_t dictSize) ++{ ++ return ZSTD_decompressMultiFrame(dctx, dst, dstCapacity, src, srcSize, dict, dictSize, NULL); ++} ++ ++size_t INIT ZSTD_decompressDCtx(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize) ++{ ++ return ZSTD_decompress_usingDict(dctx, dst, dstCapacity, src, srcSize, NULL, 0); ++} ++ ++/*-************************************** ++* Advanced Streaming Decompression API ++* Bufferless and synchronous ++****************************************/ ++size_t INIT ZSTD_nextSrcSizeToDecompress(ZSTD_DCtx *dctx) { return dctx->expected; } ++ ++ZSTD_nextInputType_e INIT ZSTD_nextInputType(ZSTD_DCtx *dctx) ++{ ++ switch (dctx->stage) { ++ default: /* should not happen */ ++ case ZSTDds_getFrameHeaderSize: ++ case ZSTDds_decodeFrameHeader: return ZSTDnit_frameHeader; ++ case ZSTDds_decodeBlockHeader: return ZSTDnit_blockHeader; ++ case ZSTDds_decompressBlock: return ZSTDnit_block; ++ case ZSTDds_decompressLastBlock: return ZSTDnit_lastBlock; ++ case ZSTDds_checkChecksum: return ZSTDnit_checksum; ++ case ZSTDds_decodeSkippableHeader: ++ case ZSTDds_skipFrame: return ZSTDnit_skippableFrame; ++ } ++} ++ ++int INIT ZSTD_isSkipFrame(ZSTD_DCtx *dctx) { return dctx->stage == ZSTDds_skipFrame; } /* for zbuff */ ++ ++/** ZSTD_decompressContinue() : ++* @return : nb of bytes generated into `dst` (necessarily <= `dstCapacity) ++* or an error code, which can be tested using ZSTD_isError() */ ++size_t INIT ZSTD_decompressContinue(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize) ++{ ++ /* Sanity check */ ++ if (srcSize != dctx->expected) ++ return ERROR(srcSize_wrong); ++ if (dstCapacity) ++ ZSTD_checkContinuity(dctx, dst); ++ ++ switch (dctx->stage) { ++ case ZSTDds_getFrameHeaderSize: ++ if (srcSize != ZSTD_frameHeaderSize_prefix) ++ return ERROR(srcSize_wrong); /* impossible */ ++ if ((ZSTD_readLE32(src) & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) { /* skippable frame */ ++ memcpy(dctx->headerBuffer, src, ZSTD_frameHeaderSize_prefix); ++ dctx->expected = ZSTD_skippableHeaderSize - ZSTD_frameHeaderSize_prefix; /* magic number + skippable frame length */ ++ dctx->stage = ZSTDds_decodeSkippableHeader; ++ return 0; ++ } ++ dctx->headerSize = ZSTD_frameHeaderSize(src, ZSTD_frameHeaderSize_prefix); ++ if (ZSTD_isError(dctx->headerSize)) ++ return dctx->headerSize; ++ memcpy(dctx->headerBuffer, src, ZSTD_frameHeaderSize_prefix); ++ if (dctx->headerSize > ZSTD_frameHeaderSize_prefix) { ++ dctx->expected = dctx->headerSize - ZSTD_frameHeaderSize_prefix; ++ dctx->stage = ZSTDds_decodeFrameHeader; ++ return 0; ++ } ++ dctx->expected = 0; /* not necessary to copy more */ ++ /* fall through */ ++ ++ case ZSTDds_decodeFrameHeader: ++ memcpy(dctx->headerBuffer + ZSTD_frameHeaderSize_prefix, src, dctx->expected); ++ CHECK_F(ZSTD_decodeFrameHeader(dctx, dctx->headerBuffer, dctx->headerSize)); ++ dctx->expected = ZSTD_blockHeaderSize; ++ dctx->stage = ZSTDds_decodeBlockHeader; ++ return 0; ++ ++ case ZSTDds_decodeBlockHeader: { ++ blockProperties_t bp; ++ size_t const cBlockSize = ZSTD_getcBlockSize(src, ZSTD_blockHeaderSize, &bp); ++ if (ZSTD_isError(cBlockSize)) ++ return cBlockSize; ++ dctx->expected = cBlockSize; ++ dctx->bType = bp.blockType; ++ dctx->rleSize = bp.origSize; ++ if (cBlockSize) { ++ dctx->stage = bp.lastBlock ? ZSTDds_decompressLastBlock : ZSTDds_decompressBlock; ++ return 0; ++ } ++ /* empty block */ ++ if (bp.lastBlock) { ++ if (dctx->fParams.checksumFlag) { ++ dctx->expected = 4; ++ dctx->stage = ZSTDds_checkChecksum; ++ } else { ++ dctx->expected = 0; /* end of frame */ ++ dctx->stage = ZSTDds_getFrameHeaderSize; ++ } ++ } else { ++ dctx->expected = 3; /* go directly to next header */ ++ dctx->stage = ZSTDds_decodeBlockHeader; ++ } ++ return 0; ++ } ++ case ZSTDds_decompressLastBlock: ++ case ZSTDds_decompressBlock: { ++ size_t rSize; ++ switch (dctx->bType) { ++ case bt_compressed: rSize = ZSTD_decompressBlock_internal(dctx, dst, dstCapacity, src, srcSize); break; ++ case bt_raw: rSize = ZSTD_copyRawBlock(dst, dstCapacity, src, srcSize); break; ++ case bt_rle: rSize = ZSTD_setRleBlock(dst, dstCapacity, src, srcSize, dctx->rleSize); break; ++ case bt_reserved: /* should never happen */ ++ default: return ERROR(corruption_detected); ++ } ++ if (ZSTD_isError(rSize)) ++ return rSize; ++ if (dctx->fParams.checksumFlag) ++ xxh64_update(&dctx->xxhState, dst, rSize); ++ ++ if (dctx->stage == ZSTDds_decompressLastBlock) { /* end of frame */ ++ if (dctx->fParams.checksumFlag) { /* another round for frame checksum */ ++ dctx->expected = 4; ++ dctx->stage = ZSTDds_checkChecksum; ++ } else { ++ dctx->expected = 0; /* ends here */ ++ dctx->stage = ZSTDds_getFrameHeaderSize; ++ } ++ } else { ++ dctx->stage = ZSTDds_decodeBlockHeader; ++ dctx->expected = ZSTD_blockHeaderSize; ++ dctx->previousDstEnd = (char *)dst + rSize; ++ } ++ return rSize; ++ } ++ case ZSTDds_checkChecksum: { ++ U32 const h32 = (U32)xxh64_digest(&dctx->xxhState); ++ U32 const check32 = ZSTD_readLE32(src); /* srcSize == 4, guaranteed by dctx->expected */ ++ if (check32 != h32) ++ return ERROR(checksum_wrong); ++ dctx->expected = 0; ++ dctx->stage = ZSTDds_getFrameHeaderSize; ++ return 0; ++ } ++ case ZSTDds_decodeSkippableHeader: { ++ memcpy(dctx->headerBuffer + ZSTD_frameHeaderSize_prefix, src, dctx->expected); ++ dctx->expected = ZSTD_readLE32(dctx->headerBuffer + 4); ++ dctx->stage = ZSTDds_skipFrame; ++ return 0; ++ } ++ case ZSTDds_skipFrame: { ++ dctx->expected = 0; ++ dctx->stage = ZSTDds_getFrameHeaderSize; ++ return 0; ++ } ++ default: ++ return ERROR(GENERIC); /* impossible */ ++ } ++} ++ ++static size_t INIT ZSTD_refDictContent(ZSTD_DCtx *dctx, const void *dict, size_t dictSize) ++{ ++ dctx->dictEnd = dctx->previousDstEnd; ++ dctx->vBase = (const char *)dict - ((const char *)(dctx->previousDstEnd) - (const char *)(dctx->base)); ++ dctx->base = dict; ++ dctx->previousDstEnd = (const char *)dict + dictSize; ++ return 0; ++} ++ ++/* ZSTD_loadEntropy() : ++ * dict : must point at beginning of a valid zstd dictionary ++ * @return : size of entropy tables read */ ++static size_t INIT ZSTD_loadEntropy(ZSTD_entropyTables_t *entropy, const void *const dict, size_t const dictSize) ++{ ++ const BYTE *dictPtr = (const BYTE *)dict; ++ const BYTE *const dictEnd = dictPtr + dictSize; ++ ++ if (dictSize <= 8) ++ return ERROR(dictionary_corrupted); ++ dictPtr += 8; /* skip header = magic + dictID */ ++ ++ { ++ size_t const hSize = HUF_readDTableX4_wksp(entropy->hufTable, dictPtr, dictEnd - dictPtr, entropy->workspace, sizeof(entropy->workspace)); ++ if (HUF_isError(hSize)) ++ return ERROR(dictionary_corrupted); ++ dictPtr += hSize; ++ } ++ ++ { ++ short offcodeNCount[MaxOff + 1]; ++ U32 offcodeMaxValue = MaxOff, offcodeLog; ++ size_t const offcodeHeaderSize = FSE_readNCount(offcodeNCount, &offcodeMaxValue, &offcodeLog, dictPtr, dictEnd - dictPtr); ++ if (FSE_isError(offcodeHeaderSize)) ++ return ERROR(dictionary_corrupted); ++ if (offcodeLog > OffFSELog) ++ return ERROR(dictionary_corrupted); ++ CHECK_E(FSE_buildDTable_wksp(entropy->OFTable, offcodeNCount, offcodeMaxValue, offcodeLog, entropy->workspace, sizeof(entropy->workspace)), dictionary_corrupted); ++ dictPtr += offcodeHeaderSize; ++ } ++ ++ { ++ short matchlengthNCount[MaxML + 1]; ++ unsigned matchlengthMaxValue = MaxML, matchlengthLog; ++ size_t const matchlengthHeaderSize = FSE_readNCount(matchlengthNCount, &matchlengthMaxValue, &matchlengthLog, dictPtr, dictEnd - dictPtr); ++ if (FSE_isError(matchlengthHeaderSize)) ++ return ERROR(dictionary_corrupted); ++ if (matchlengthLog > MLFSELog) ++ return ERROR(dictionary_corrupted); ++ CHECK_E(FSE_buildDTable_wksp(entropy->MLTable, matchlengthNCount, matchlengthMaxValue, matchlengthLog, entropy->workspace, sizeof(entropy->workspace)), dictionary_corrupted); ++ dictPtr += matchlengthHeaderSize; ++ } ++ ++ { ++ short litlengthNCount[MaxLL + 1]; ++ unsigned litlengthMaxValue = MaxLL, litlengthLog; ++ size_t const litlengthHeaderSize = FSE_readNCount(litlengthNCount, &litlengthMaxValue, &litlengthLog, dictPtr, dictEnd - dictPtr); ++ if (FSE_isError(litlengthHeaderSize)) ++ return ERROR(dictionary_corrupted); ++ if (litlengthLog > LLFSELog) ++ return ERROR(dictionary_corrupted); ++ CHECK_E(FSE_buildDTable_wksp(entropy->LLTable, litlengthNCount, litlengthMaxValue, litlengthLog, entropy->workspace, sizeof(entropy->workspace)), dictionary_corrupted); ++ dictPtr += litlengthHeaderSize; ++ } ++ ++ if (dictPtr + 12 > dictEnd) ++ return ERROR(dictionary_corrupted); ++ { ++ int i; ++ size_t const dictContentSize = (size_t)(dictEnd - (dictPtr + 12)); ++ for (i = 0; i < 3; i++) { ++ U32 const rep = ZSTD_readLE32(dictPtr); ++ dictPtr += 4; ++ if (rep == 0 || rep >= dictContentSize) ++ return ERROR(dictionary_corrupted); ++ entropy->rep[i] = rep; ++ } ++ } ++ ++ return dictPtr - (const BYTE *)dict; ++} ++ ++static size_t INIT ZSTD_decompress_insertDictionary(ZSTD_DCtx *dctx, const void *dict, size_t dictSize) ++{ ++ if (dictSize < 8) ++ return ZSTD_refDictContent(dctx, dict, dictSize); ++ { ++ U32 const magic = ZSTD_readLE32(dict); ++ if (magic != ZSTD_DICT_MAGIC) { ++ return ZSTD_refDictContent(dctx, dict, dictSize); /* pure content mode */ ++ } ++ } ++ dctx->dictID = ZSTD_readLE32((const char *)dict + 4); ++ ++ /* load entropy tables */ ++ { ++ size_t const eSize = ZSTD_loadEntropy(&dctx->entropy, dict, dictSize); ++ if (ZSTD_isError(eSize)) ++ return ERROR(dictionary_corrupted); ++ dict = (const char *)dict + eSize; ++ dictSize -= eSize; ++ } ++ dctx->litEntropy = dctx->fseEntropy = 1; ++ ++ /* reference dictionary content */ ++ return ZSTD_refDictContent(dctx, dict, dictSize); ++} ++ ++size_t INIT ZSTD_decompressBegin_usingDict(ZSTD_DCtx *dctx, const void *dict, size_t dictSize) ++{ ++ CHECK_F(ZSTD_decompressBegin(dctx)); ++ if (dict && dictSize) ++ CHECK_E(ZSTD_decompress_insertDictionary(dctx, dict, dictSize), dictionary_corrupted); ++ return 0; ++} ++ ++/* ====== ZSTD_DDict ====== */ ++ ++struct ZSTD_DDict_s { ++ void *dictBuffer; ++ const void *dictContent; ++ size_t dictSize; ++ ZSTD_entropyTables_t entropy; ++ U32 dictID; ++ U32 entropyPresent; ++ ZSTD_customMem cMem; ++}; /* typedef'd to ZSTD_DDict within "zstd.h" */ ++ ++size_t INIT ZSTD_DDictWorkspaceBound(void) { return ZSTD_ALIGN(sizeof(ZSTD_stack)) + ZSTD_ALIGN(sizeof(ZSTD_DDict)); } ++ ++static const void INIT *ZSTD_DDictDictContent(const ZSTD_DDict *ddict) { return ddict->dictContent; } ++ ++static size_t INIT ZSTD_DDictDictSize(const ZSTD_DDict *ddict) { return ddict->dictSize; } ++ ++static void INIT ZSTD_refDDict(ZSTD_DCtx *dstDCtx, const ZSTD_DDict *ddict) ++{ ++ ZSTD_decompressBegin(dstDCtx); /* init */ ++ if (ddict) { /* support refDDict on NULL */ ++ dstDCtx->dictID = ddict->dictID; ++ dstDCtx->base = ddict->dictContent; ++ dstDCtx->vBase = ddict->dictContent; ++ dstDCtx->dictEnd = (const BYTE *)ddict->dictContent + ddict->dictSize; ++ dstDCtx->previousDstEnd = dstDCtx->dictEnd; ++ if (ddict->entropyPresent) { ++ dstDCtx->litEntropy = 1; ++ dstDCtx->fseEntropy = 1; ++ dstDCtx->LLTptr = ddict->entropy.LLTable; ++ dstDCtx->MLTptr = ddict->entropy.MLTable; ++ dstDCtx->OFTptr = ddict->entropy.OFTable; ++ dstDCtx->HUFptr = ddict->entropy.hufTable; ++ dstDCtx->entropy.rep[0] = ddict->entropy.rep[0]; ++ dstDCtx->entropy.rep[1] = ddict->entropy.rep[1]; ++ dstDCtx->entropy.rep[2] = ddict->entropy.rep[2]; ++ } else { ++ dstDCtx->litEntropy = 0; ++ dstDCtx->fseEntropy = 0; ++ } ++ } ++} ++ ++static size_t INIT ZSTD_loadEntropy_inDDict(ZSTD_DDict *ddict) ++{ ++ ddict->dictID = 0; ++ ddict->entropyPresent = 0; ++ if (ddict->dictSize < 8) ++ return 0; ++ { ++ U32 const magic = ZSTD_readLE32(ddict->dictContent); ++ if (magic != ZSTD_DICT_MAGIC) ++ return 0; /* pure content mode */ ++ } ++ ddict->dictID = ZSTD_readLE32((const char *)ddict->dictContent + 4); ++ ++ /* load entropy tables */ ++ CHECK_E(ZSTD_loadEntropy(&ddict->entropy, ddict->dictContent, ddict->dictSize), dictionary_corrupted); ++ ddict->entropyPresent = 1; ++ return 0; ++} ++ ++static ZSTD_DDict INIT *ZSTD_createDDict_advanced(const void *dict, size_t dictSize, unsigned byReference, ZSTD_customMem customMem) ++{ ++ if (!customMem.customAlloc || !customMem.customFree) ++ return NULL; ++ ++ { ++ ZSTD_DDict *const ddict = (ZSTD_DDict *)ZSTD_malloc(sizeof(ZSTD_DDict), customMem); ++ if (!ddict) ++ return NULL; ++ ddict->cMem = customMem; ++ ++ if ((byReference) || (!dict) || (!dictSize)) { ++ ddict->dictBuffer = NULL; ++ ddict->dictContent = dict; ++ } else { ++ void *const internalBuffer = ZSTD_malloc(dictSize, customMem); ++ if (!internalBuffer) { ++ ZSTD_freeDDict(ddict); ++ return NULL; ++ } ++ memcpy(internalBuffer, dict, dictSize); ++ ddict->dictBuffer = internalBuffer; ++ ddict->dictContent = internalBuffer; ++ } ++ ddict->dictSize = dictSize; ++ ddict->entropy.hufTable[0] = (HUF_DTable)((HufLog)*0x1000001); /* cover both little and big endian */ ++ /* parse dictionary content */ ++ { ++ size_t const errorCode = ZSTD_loadEntropy_inDDict(ddict); ++ if (ZSTD_isError(errorCode)) { ++ ZSTD_freeDDict(ddict); ++ return NULL; ++ } ++ } ++ ++ return ddict; ++ } ++} ++ ++/*! ZSTD_initDDict() : ++* Create a digested dictionary, to start decompression without startup delay. ++* `dict` content is copied inside DDict. ++* Consequently, `dict` can be released after `ZSTD_DDict` creation */ ++ZSTD_DDict INIT *ZSTD_initDDict(const void *dict, size_t dictSize, void *workspace, size_t workspaceSize) ++{ ++ ZSTD_customMem const stackMem = ZSTD_initStack(workspace, workspaceSize); ++ return ZSTD_createDDict_advanced(dict, dictSize, 1, stackMem); ++} ++ ++size_t INIT ZSTD_freeDDict(ZSTD_DDict *ddict) ++{ ++ if (ddict == NULL) ++ return 0; /* support free on NULL */ ++ { ++ ZSTD_customMem const cMem = ddict->cMem; ++ ZSTD_free(ddict->dictBuffer, cMem); ++ ZSTD_free(ddict, cMem); ++ return 0; ++ } ++} ++ ++/*! ZSTD_getDictID_fromDict() : ++ * Provides the dictID stored within dictionary. ++ * if @return == 0, the dictionary is not conformant with Zstandard specification. ++ * It can still be loaded, but as a content-only dictionary. */ ++unsigned INIT ZSTD_getDictID_fromDict(const void *dict, size_t dictSize) ++{ ++ if (dictSize < 8) ++ return 0; ++ if (ZSTD_readLE32(dict) != ZSTD_DICT_MAGIC) ++ return 0; ++ return ZSTD_readLE32((const char *)dict + 4); ++} ++ ++/*! ZSTD_getDictID_fromDDict() : ++ * Provides the dictID of the dictionary loaded into `ddict`. ++ * If @return == 0, the dictionary is not conformant to Zstandard specification, or empty. ++ * Non-conformant dictionaries can still be loaded, but as content-only dictionaries. */ ++unsigned INIT ZSTD_getDictID_fromDDict(const ZSTD_DDict *ddict) ++{ ++ if (ddict == NULL) ++ return 0; ++ return ZSTD_getDictID_fromDict(ddict->dictContent, ddict->dictSize); ++} ++ ++/*! ZSTD_getDictID_fromFrame() : ++ * Provides the dictID required to decompressed the frame stored within `src`. ++ * If @return == 0, the dictID could not be decoded. ++ * This could for one of the following reasons : ++ * - The frame does not require a dictionary to be decoded (most common case). ++ * - The frame was built with dictID intentionally removed. Whatever dictionary is necessary is a hidden information. ++ * Note : this use case also happens when using a non-conformant dictionary. ++ * - `srcSize` is too small, and as a result, the frame header could not be decoded (only possible if `srcSize < ZSTD_FRAMEHEADERSIZE_MAX`). ++ * - This is not a Zstandard frame. ++ * When identifying the exact failure cause, it's possible to used ZSTD_getFrameParams(), which will provide a more precise error code. */ ++unsigned INIT ZSTD_getDictID_fromFrame(const void *src, size_t srcSize) ++{ ++ ZSTD_frameParams zfp = {0, 0, 0, 0}; ++ size_t const hError = ZSTD_getFrameParams(&zfp, src, srcSize); ++ if (ZSTD_isError(hError)) ++ return 0; ++ return zfp.dictID; ++} ++ ++/*! ZSTD_decompress_usingDDict() : ++* Decompression using a pre-digested Dictionary ++* Use dictionary without significant overhead. */ ++size_t INIT ZSTD_decompress_usingDDict(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize, const ZSTD_DDict *ddict) ++{ ++ /* pass content and size in case legacy frames are encountered */ ++ return ZSTD_decompressMultiFrame(dctx, dst, dstCapacity, src, srcSize, NULL, 0, ddict); ++} ++ ++/*===================================== ++* Streaming decompression ++*====================================*/ ++ ++typedef enum { zdss_init, zdss_loadHeader, zdss_read, zdss_load, zdss_flush } ZSTD_dStreamStage; ++ ++/* *** Resource management *** */ ++struct ZSTD_DStream_s { ++ ZSTD_DCtx *dctx; ++ ZSTD_DDict *ddictLocal; ++ const ZSTD_DDict *ddict; ++ ZSTD_frameParams fParams; ++ ZSTD_dStreamStage stage; ++ char *inBuff; ++ size_t inBuffSize; ++ size_t inPos; ++ size_t maxWindowSize; ++ char *outBuff; ++ size_t outBuffSize; ++ size_t outStart; ++ size_t outEnd; ++ size_t blockSize; ++ BYTE headerBuffer[ZSTD_FRAMEHEADERSIZE_MAX]; /* tmp buffer to store frame header */ ++ size_t lhSize; ++ ZSTD_customMem customMem; ++ void *legacyContext; ++ U32 previousLegacyVersion; ++ U32 legacyVersion; ++ U32 hostageByte; ++}; /* typedef'd to ZSTD_DStream within "zstd.h" */ ++ ++size_t INIT ZSTD_DStreamWorkspaceBound(size_t maxWindowSize) ++{ ++ size_t const blockSize = MIN(maxWindowSize, ZSTD_BLOCKSIZE_ABSOLUTEMAX); ++ size_t const inBuffSize = blockSize; ++ size_t const outBuffSize = maxWindowSize + blockSize + WILDCOPY_OVERLENGTH * 2; ++ return ZSTD_DCtxWorkspaceBound() + ZSTD_ALIGN(sizeof(ZSTD_DStream)) + ZSTD_ALIGN(inBuffSize) + ZSTD_ALIGN(outBuffSize); ++} ++ ++static ZSTD_DStream INIT *ZSTD_createDStream_advanced(ZSTD_customMem customMem) ++{ ++ ZSTD_DStream *zds; ++ ++ if (!customMem.customAlloc || !customMem.customFree) ++ return NULL; ++ ++ zds = (ZSTD_DStream *)ZSTD_malloc(sizeof(ZSTD_DStream), customMem); ++ if (zds == NULL) ++ return NULL; ++ memset(zds, 0, sizeof(ZSTD_DStream)); ++ memcpy(&zds->customMem, &customMem, sizeof(ZSTD_customMem)); ++ zds->dctx = ZSTD_createDCtx_advanced(customMem); ++ if (zds->dctx == NULL) { ++ ZSTD_freeDStream(zds); ++ return NULL; ++ } ++ zds->stage = zdss_init; ++ zds->maxWindowSize = ZSTD_MAXWINDOWSIZE_DEFAULT; ++ return zds; ++} ++ ++ZSTD_DStream INIT *ZSTD_initDStream(size_t maxWindowSize, void *workspace, size_t workspaceSize) ++{ ++ ZSTD_customMem const stackMem = ZSTD_initStack(workspace, workspaceSize); ++ ZSTD_DStream *zds = ZSTD_createDStream_advanced(stackMem); ++ if (!zds) { ++ return NULL; ++ } ++ ++ zds->maxWindowSize = maxWindowSize; ++ zds->stage = zdss_loadHeader; ++ zds->lhSize = zds->inPos = zds->outStart = zds->outEnd = 0; ++ ZSTD_freeDDict(zds->ddictLocal); ++ zds->ddictLocal = NULL; ++ zds->ddict = zds->ddictLocal; ++ zds->legacyVersion = 0; ++ zds->hostageByte = 0; ++ ++ { ++ size_t const blockSize = MIN(zds->maxWindowSize, ZSTD_BLOCKSIZE_ABSOLUTEMAX); ++ size_t const neededOutSize = zds->maxWindowSize + blockSize + WILDCOPY_OVERLENGTH * 2; ++ ++ zds->inBuff = (char *)ZSTD_malloc(blockSize, zds->customMem); ++ zds->inBuffSize = blockSize; ++ zds->outBuff = (char *)ZSTD_malloc(neededOutSize, zds->customMem); ++ zds->outBuffSize = neededOutSize; ++ if (zds->inBuff == NULL || zds->outBuff == NULL) { ++ ZSTD_freeDStream(zds); ++ return NULL; ++ } ++ } ++ return zds; ++} ++ ++ZSTD_DStream INIT *ZSTD_initDStream_usingDDict(size_t maxWindowSize, const ZSTD_DDict *ddict, void *workspace, size_t workspaceSize) ++{ ++ ZSTD_DStream *zds = ZSTD_initDStream(maxWindowSize, workspace, workspaceSize); ++ if (zds) { ++ zds->ddict = ddict; ++ } ++ return zds; ++} ++ ++size_t INIT ZSTD_freeDStream(ZSTD_DStream *zds) ++{ ++ if (zds == NULL) ++ return 0; /* support free on null */ ++ { ++ ZSTD_customMem const cMem = zds->customMem; ++ ZSTD_freeDCtx(zds->dctx); ++ zds->dctx = NULL; ++ ZSTD_freeDDict(zds->ddictLocal); ++ zds->ddictLocal = NULL; ++ ZSTD_free(zds->inBuff, cMem); ++ zds->inBuff = NULL; ++ ZSTD_free(zds->outBuff, cMem); ++ zds->outBuff = NULL; ++ ZSTD_free(zds, cMem); ++ return 0; ++ } ++} ++ ++/* *** Initialization *** */ ++ ++size_t INIT ZSTD_DStreamInSize(void) { return ZSTD_BLOCKSIZE_ABSOLUTEMAX + ZSTD_blockHeaderSize; } ++size_t INIT ZSTD_DStreamOutSize(void) { return ZSTD_BLOCKSIZE_ABSOLUTEMAX; } ++ ++size_t INIT ZSTD_resetDStream(ZSTD_DStream *zds) ++{ ++ zds->stage = zdss_loadHeader; ++ zds->lhSize = zds->inPos = zds->outStart = zds->outEnd = 0; ++ zds->legacyVersion = 0; ++ zds->hostageByte = 0; ++ return ZSTD_frameHeaderSize_prefix; ++} ++ ++/* ***** Decompression ***** */ ++ ++ZSTD_STATIC size_t INIT ZSTD_limitCopy(void *dst, size_t dstCapacity, const void *src, size_t srcSize) ++{ ++ size_t const length = MIN(dstCapacity, srcSize); ++ memcpy(dst, src, length); ++ return length; ++} ++ ++size_t INIT ZSTD_decompressStream(ZSTD_DStream *zds, ZSTD_outBuffer *output, ZSTD_inBuffer *input) ++{ ++ const char *const istart = (const char *)(input->src) + input->pos; ++ const char *const iend = (const char *)(input->src) + input->size; ++ const char *ip = istart; ++ char *const ostart = (char *)(output->dst) + output->pos; ++ char *const oend = (char *)(output->dst) + output->size; ++ char *op = ostart; ++ U32 someMoreWork = 1; ++ ++ while (someMoreWork) { ++ switch (zds->stage) { ++ case zdss_init: ++ ZSTD_resetDStream(zds); /* transparent reset on starting decoding a new frame */ ++ /* fall through */ ++ ++ case zdss_loadHeader: { ++ size_t const hSize = ZSTD_getFrameParams(&zds->fParams, zds->headerBuffer, zds->lhSize); ++ if (ZSTD_isError(hSize)) ++ return hSize; ++ if (hSize != 0) { /* need more input */ ++ size_t const toLoad = hSize - zds->lhSize; /* if hSize!=0, hSize > zds->lhSize */ ++ if (toLoad > (size_t)(iend - ip)) { /* not enough input to load full header */ ++ memcpy(zds->headerBuffer + zds->lhSize, ip, iend - ip); ++ zds->lhSize += iend - ip; ++ input->pos = input->size; ++ return (MAX(ZSTD_frameHeaderSize_min, hSize) - zds->lhSize) + ++ ZSTD_blockHeaderSize; /* remaining header bytes + next block header */ ++ } ++ memcpy(zds->headerBuffer + zds->lhSize, ip, toLoad); ++ zds->lhSize = hSize; ++ ip += toLoad; ++ break; ++ } ++ ++ /* check for single-pass mode opportunity */ ++ if (zds->fParams.frameContentSize && zds->fParams.windowSize /* skippable frame if == 0 */ ++ && (U64)(size_t)(oend - op) >= zds->fParams.frameContentSize) { ++ size_t const cSize = ZSTD_findFrameCompressedSize(istart, iend - istart); ++ if (cSize <= (size_t)(iend - istart)) { ++ size_t const decompressedSize = ZSTD_decompress_usingDDict(zds->dctx, op, oend - op, istart, cSize, zds->ddict); ++ if (ZSTD_isError(decompressedSize)) ++ return decompressedSize; ++ ip = istart + cSize; ++ op += decompressedSize; ++ zds->dctx->expected = 0; ++ zds->stage = zdss_init; ++ someMoreWork = 0; ++ break; ++ } ++ } ++ ++ /* Consume header */ ++ ZSTD_refDDict(zds->dctx, zds->ddict); ++ { ++ size_t const h1Size = ZSTD_nextSrcSizeToDecompress(zds->dctx); /* == ZSTD_frameHeaderSize_prefix */ ++ CHECK_F(ZSTD_decompressContinue(zds->dctx, NULL, 0, zds->headerBuffer, h1Size)); ++ { ++ size_t const h2Size = ZSTD_nextSrcSizeToDecompress(zds->dctx); ++ CHECK_F(ZSTD_decompressContinue(zds->dctx, NULL, 0, zds->headerBuffer + h1Size, h2Size)); ++ } ++ } ++ ++ zds->fParams.windowSize = MAX(zds->fParams.windowSize, 1U << ZSTD_WINDOWLOG_ABSOLUTEMIN); ++ if (zds->fParams.windowSize > zds->maxWindowSize) ++ return ERROR(frameParameter_windowTooLarge); ++ ++ /* Buffers are preallocated, but double check */ ++ { ++ size_t const blockSize = MIN(zds->maxWindowSize, ZSTD_BLOCKSIZE_ABSOLUTEMAX); ++ size_t const neededOutSize = zds->maxWindowSize + blockSize + WILDCOPY_OVERLENGTH * 2; ++ if (zds->inBuffSize < blockSize) { ++ return ERROR(GENERIC); ++ } ++ if (zds->outBuffSize < neededOutSize) { ++ return ERROR(GENERIC); ++ } ++ zds->blockSize = blockSize; ++ } ++ zds->stage = zdss_read; ++ } ++ /* fall through */ ++ ++ case zdss_read: { ++ size_t const neededInSize = ZSTD_nextSrcSizeToDecompress(zds->dctx); ++ if (neededInSize == 0) { /* end of frame */ ++ zds->stage = zdss_init; ++ someMoreWork = 0; ++ break; ++ } ++ if ((size_t)(iend - ip) >= neededInSize) { /* decode directly from src */ ++ const int isSkipFrame = ZSTD_isSkipFrame(zds->dctx); ++ size_t const decodedSize = ZSTD_decompressContinue(zds->dctx, zds->outBuff + zds->outStart, ++ (isSkipFrame ? 0 : zds->outBuffSize - zds->outStart), ip, neededInSize); ++ if (ZSTD_isError(decodedSize)) ++ return decodedSize; ++ ip += neededInSize; ++ if (!decodedSize && !isSkipFrame) ++ break; /* this was just a header */ ++ zds->outEnd = zds->outStart + decodedSize; ++ zds->stage = zdss_flush; ++ break; ++ } ++ if (ip == iend) { ++ someMoreWork = 0; ++ break; ++ } /* no more input */ ++ zds->stage = zdss_load; ++ /* pass-through */ ++ } ++ /* fall through */ ++ ++ case zdss_load: { ++ size_t const neededInSize = ZSTD_nextSrcSizeToDecompress(zds->dctx); ++ size_t const toLoad = neededInSize - zds->inPos; /* should always be <= remaining space within inBuff */ ++ size_t loadedSize; ++ if (toLoad > zds->inBuffSize - zds->inPos) ++ return ERROR(corruption_detected); /* should never happen */ ++ loadedSize = ZSTD_limitCopy(zds->inBuff + zds->inPos, toLoad, ip, iend - ip); ++ ip += loadedSize; ++ zds->inPos += loadedSize; ++ if (loadedSize < toLoad) { ++ someMoreWork = 0; ++ break; ++ } /* not enough input, wait for more */ ++ ++ /* decode loaded input */ ++ { ++ const int isSkipFrame = ZSTD_isSkipFrame(zds->dctx); ++ size_t const decodedSize = ZSTD_decompressContinue(zds->dctx, zds->outBuff + zds->outStart, zds->outBuffSize - zds->outStart, ++ zds->inBuff, neededInSize); ++ if (ZSTD_isError(decodedSize)) ++ return decodedSize; ++ zds->inPos = 0; /* input is consumed */ ++ if (!decodedSize && !isSkipFrame) { ++ zds->stage = zdss_read; ++ break; ++ } /* this was just a header */ ++ zds->outEnd = zds->outStart + decodedSize; ++ zds->stage = zdss_flush; ++ /* pass-through */ ++ } ++ } ++ /* fall through */ ++ ++ case zdss_flush: { ++ size_t const toFlushSize = zds->outEnd - zds->outStart; ++ size_t const flushedSize = ZSTD_limitCopy(op, oend - op, zds->outBuff + zds->outStart, toFlushSize); ++ op += flushedSize; ++ zds->outStart += flushedSize; ++ if (flushedSize == toFlushSize) { /* flush completed */ ++ zds->stage = zdss_read; ++ if (zds->outStart + zds->blockSize > zds->outBuffSize) ++ zds->outStart = zds->outEnd = 0; ++ break; ++ } ++ /* cannot complete flush */ ++ someMoreWork = 0; ++ break; ++ } ++ default: ++ return ERROR(GENERIC); /* impossible */ ++ } ++ } ++ ++ /* result */ ++ input->pos += (size_t)(ip - istart); ++ output->pos += (size_t)(op - ostart); ++ { ++ size_t nextSrcSizeHint = ZSTD_nextSrcSizeToDecompress(zds->dctx); ++ if (!nextSrcSizeHint) { /* frame fully decoded */ ++ if (zds->outEnd == zds->outStart) { /* output fully flushed */ ++ if (zds->hostageByte) { ++ if (input->pos >= input->size) { ++ zds->stage = zdss_read; ++ return 1; ++ } /* can't release hostage (not present) */ ++ input->pos++; /* release hostage */ ++ } ++ return 0; ++ } ++ if (!zds->hostageByte) { /* output not fully flushed; keep last byte as hostage; will be released when all output is flushed */ ++ input->pos--; /* note : pos > 0, otherwise, impossible to finish reading last block */ ++ zds->hostageByte = 1; ++ } ++ return 1; ++ } ++ nextSrcSizeHint += ZSTD_blockHeaderSize * (ZSTD_nextInputType(zds->dctx) == ZSTDnit_block); /* preload header of next block */ ++ if (zds->inPos > nextSrcSizeHint) ++ return ERROR(GENERIC); /* should never happen */ ++ nextSrcSizeHint -= zds->inPos; /* already loaded*/ ++ return nextSrcSizeHint; ++ } ++} +diff --git a/xen/common/zstd/entropy_common.c b/xen/common/zstd/entropy_common.c +new file mode 100644 +index 0000000000..bcdb57982b +--- /dev/null ++++ b/xen/common/zstd/entropy_common.c +@@ -0,0 +1,243 @@ ++/* ++ * Common functions of New Generation Entropy library ++ * Copyright (C) 2016, Yann Collet. ++ * ++ * BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions are ++ * met: ++ * ++ * * Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * * Redistributions in binary form must reproduce the above ++ * copyright notice, this list of conditions and the following disclaimer ++ * in the documentation and/or other materials provided with the ++ * distribution. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS ++ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT ++ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR ++ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT ++ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, ++ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT ++ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE ++ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++ * ++ * This program is free software; you can redistribute it and/or modify it under ++ * the terms of the GNU General Public License version 2 as published by the ++ * Free Software Foundation. This program is dual-licensed; you may select ++ * either version 2 of the GNU General Public License ("GPL") or BSD license ++ * ("BSD"). ++ * ++ * You can contact the author at : ++ * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy ++ */ ++ ++/* ************************************* ++* Dependencies ++***************************************/ ++#include "error_private.h" /* ERR_*, ERROR */ ++#include "fse.h" ++#include "huf.h" ++#include "mem.h" ++ ++/*=== Version ===*/ ++unsigned INIT FSE_versionNumber(void) { return FSE_VERSION_NUMBER; } ++ ++/*=== Error Management ===*/ ++unsigned INIT FSE_isError(size_t code) { return ERR_isError(code); } ++ ++unsigned INIT HUF_isError(size_t code) { return ERR_isError(code); } ++ ++/*-************************************************************** ++* FSE NCount encoding-decoding ++****************************************************************/ ++size_t INIT FSE_readNCount(short *normalizedCounter, unsigned *maxSVPtr, unsigned *tableLogPtr, const void *headerBuffer, size_t hbSize) ++{ ++ const BYTE *const istart = (const BYTE *)headerBuffer; ++ const BYTE *const iend = istart + hbSize; ++ const BYTE *ip = istart; ++ int nbBits; ++ int remaining; ++ int threshold; ++ U32 bitStream; ++ int bitCount; ++ unsigned charnum = 0; ++ int previous0 = 0; ++ ++ if (hbSize < 4) ++ return ERROR(srcSize_wrong); ++ bitStream = ZSTD_readLE32(ip); ++ nbBits = (bitStream & 0xF) + FSE_MIN_TABLELOG; /* extract tableLog */ ++ if (nbBits > FSE_TABLELOG_ABSOLUTE_MAX) ++ return ERROR(tableLog_tooLarge); ++ bitStream >>= 4; ++ bitCount = 4; ++ *tableLogPtr = nbBits; ++ remaining = (1 << nbBits) + 1; ++ threshold = 1 << nbBits; ++ nbBits++; ++ ++ while ((remaining > 1) & (charnum <= *maxSVPtr)) { ++ if (previous0) { ++ unsigned n0 = charnum; ++ while ((bitStream & 0xFFFF) == 0xFFFF) { ++ n0 += 24; ++ if (ip < iend - 5) { ++ ip += 2; ++ bitStream = ZSTD_readLE32(ip) >> bitCount; ++ } else { ++ bitStream >>= 16; ++ bitCount += 16; ++ } ++ } ++ while ((bitStream & 3) == 3) { ++ n0 += 3; ++ bitStream >>= 2; ++ bitCount += 2; ++ } ++ n0 += bitStream & 3; ++ bitCount += 2; ++ if (n0 > *maxSVPtr) ++ return ERROR(maxSymbolValue_tooSmall); ++ while (charnum < n0) ++ normalizedCounter[charnum++] = 0; ++ if ((ip <= iend - 7) || (ip + (bitCount >> 3) <= iend - 4)) { ++ ip += bitCount >> 3; ++ bitCount &= 7; ++ bitStream = ZSTD_readLE32(ip) >> bitCount; ++ } else { ++ bitStream >>= 2; ++ } ++ } ++ { ++ int const max = (2 * threshold - 1) - remaining; ++ int count; ++ ++ if ((bitStream & (threshold - 1)) < (U32)max) { ++ count = bitStream & (threshold - 1); ++ bitCount += nbBits - 1; ++ } else { ++ count = bitStream & (2 * threshold - 1); ++ if (count >= threshold) ++ count -= max; ++ bitCount += nbBits; ++ } ++ ++ count--; /* extra accuracy */ ++ remaining -= count < 0 ? -count : count; /* -1 means +1 */ ++ normalizedCounter[charnum++] = (short)count; ++ previous0 = !count; ++ while (remaining < threshold) { ++ nbBits--; ++ threshold >>= 1; ++ } ++ ++ if ((ip <= iend - 7) || (ip + (bitCount >> 3) <= iend - 4)) { ++ ip += bitCount >> 3; ++ bitCount &= 7; ++ } else { ++ bitCount -= (int)(8 * (iend - 4 - ip)); ++ ip = iend - 4; ++ } ++ bitStream = ZSTD_readLE32(ip) >> (bitCount & 31); ++ } ++ } /* while ((remaining>1) & (charnum<=*maxSVPtr)) */ ++ if (remaining != 1) ++ return ERROR(corruption_detected); ++ if (bitCount > 32) ++ return ERROR(corruption_detected); ++ *maxSVPtr = charnum - 1; ++ ++ ip += (bitCount + 7) >> 3; ++ return ip - istart; ++} ++ ++/*! HUF_readStats() : ++ Read compact Huffman tree, saved by HUF_writeCTable(). ++ `huffWeight` is destination buffer. ++ `rankStats` is assumed to be a table of at least HUF_TABLELOG_MAX U32. ++ @return : size read from `src` , or an error Code . ++ Note : Needed by HUF_readCTable() and HUF_readDTableX?() . ++*/ ++size_t INIT HUF_readStats_wksp(BYTE *huffWeight, size_t hwSize, U32 *rankStats, U32 *nbSymbolsPtr, U32 *tableLogPtr, const void *src, size_t srcSize, void *workspace, size_t workspaceSize) ++{ ++ U32 weightTotal; ++ const BYTE *ip = (const BYTE *)src; ++ size_t iSize; ++ size_t oSize; ++ ++ if (!srcSize) ++ return ERROR(srcSize_wrong); ++ iSize = ip[0]; ++ /* memset(huffWeight, 0, hwSize); */ /* is not necessary, even though some analyzer complain ... */ ++ ++ if (iSize >= 128) { /* special header */ ++ oSize = iSize - 127; ++ iSize = ((oSize + 1) / 2); ++ if (iSize + 1 > srcSize) ++ return ERROR(srcSize_wrong); ++ if (oSize >= hwSize) ++ return ERROR(corruption_detected); ++ ip += 1; ++ { ++ U32 n; ++ for (n = 0; n < oSize; n += 2) { ++ huffWeight[n] = ip[n / 2] >> 4; ++ huffWeight[n + 1] = ip[n / 2] & 15; ++ } ++ } ++ } else { /* header compressed with FSE (normal case) */ ++ if (iSize + 1 > srcSize) ++ return ERROR(srcSize_wrong); ++ oSize = FSE_decompress_wksp(huffWeight, hwSize - 1, ip + 1, iSize, 6, workspace, workspaceSize); /* max (hwSize-1) values decoded, as last one is implied */ ++ if (FSE_isError(oSize)) ++ return oSize; ++ } ++ ++ /* collect weight stats */ ++ memset(rankStats, 0, (HUF_TABLELOG_MAX + 1) * sizeof(U32)); ++ weightTotal = 0; ++ { ++ U32 n; ++ for (n = 0; n < oSize; n++) { ++ if (huffWeight[n] >= HUF_TABLELOG_MAX) ++ return ERROR(corruption_detected); ++ rankStats[huffWeight[n]]++; ++ weightTotal += (1 << huffWeight[n]) >> 1; ++ } ++ } ++ if (weightTotal == 0) ++ return ERROR(corruption_detected); ++ ++ /* get last non-null symbol weight (implied, total must be 2^n) */ ++ { ++ U32 const tableLog = BIT_highbit32(weightTotal) + 1; ++ if (tableLog > HUF_TABLELOG_MAX) ++ return ERROR(corruption_detected); ++ *tableLogPtr = tableLog; ++ /* determine last weight */ ++ { ++ U32 const total = 1 << tableLog; ++ U32 const rest = total - weightTotal; ++ U32 const verif = 1 << BIT_highbit32(rest); ++ U32 const lastWeight = BIT_highbit32(rest) + 1; ++ if (verif != rest) ++ return ERROR(corruption_detected); /* last value must be a clean power of 2 */ ++ huffWeight[oSize] = (BYTE)lastWeight; ++ rankStats[lastWeight]++; ++ } ++ } ++ ++ /* check tree construction validity */ ++ if ((rankStats[1] < 2) || (rankStats[1] & 1)) ++ return ERROR(corruption_detected); /* by construction : at least 2 elts of rank 1, must be even */ ++ ++ /* results */ ++ *nbSymbolsPtr = (U32)(oSize + 1); ++ return iSize + 1; ++} +diff --git a/xen/common/zstd/error_private.h b/xen/common/zstd/error_private.h +new file mode 100644 +index 0000000000..ecbfe51dfb +--- /dev/null ++++ b/xen/common/zstd/error_private.h +@@ -0,0 +1,53 @@ ++/** ++ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc. ++ * All rights reserved. ++ * ++ * This source code is licensed under the BSD-style license found in the ++ * LICENSE file in the root directory of https://github.com/facebook/zstd. ++ * An additional grant of patent rights can be found in the PATENTS file in the ++ * same directory. ++ * ++ * This program is free software; you can redistribute it and/or modify it under ++ * the terms of the GNU General Public License version 2 as published by the ++ * Free Software Foundation. This program is dual-licensed; you may select ++ * either version 2 of the GNU General Public License ("GPL") or BSD license ++ * ("BSD"). ++ */ ++ ++/* Note : this module is expected to remain private, do not expose it */ ++ ++#ifndef ERROR_H_MODULE ++#define ERROR_H_MODULE ++ ++/* **************************************** ++* Dependencies ++******************************************/ ++#include /* size_t */ ++#include /* enum list */ ++ ++/* **************************************** ++* Compiler-specific ++******************************************/ ++#define ERR_STATIC static __attribute__((unused)) ++ ++/*-**************************************** ++* Customization (error_public.h) ++******************************************/ ++typedef ZSTD_ErrorCode ERR_enum; ++#define PREFIX(name) ZSTD_error_##name ++ ++/*-**************************************** ++* Error codes handling ++******************************************/ ++#define ERROR(name) ((size_t)-PREFIX(name)) ++ ++ERR_STATIC unsigned ERR_isError(size_t code) { return (code > ERROR(maxCode)); } ++ ++ERR_STATIC ERR_enum ERR_getErrorCode(size_t code) ++{ ++ if (!ERR_isError(code)) ++ return (ERR_enum)0; ++ return (ERR_enum)(0 - code); ++} ++ ++#endif /* ERROR_H_MODULE */ +diff --git a/xen/common/zstd/fse.h b/xen/common/zstd/fse.h +new file mode 100644 +index 0000000000..b86717c34d +--- /dev/null ++++ b/xen/common/zstd/fse.h +@@ -0,0 +1,575 @@ ++/* ++ * FSE : Finite State Entropy codec ++ * Public Prototypes declaration ++ * Copyright (C) 2013-2016, Yann Collet. ++ * ++ * BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions are ++ * met: ++ * ++ * * Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * * Redistributions in binary form must reproduce the above ++ * copyright notice, this list of conditions and the following disclaimer ++ * in the documentation and/or other materials provided with the ++ * distribution. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS ++ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT ++ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR ++ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT ++ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, ++ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT ++ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE ++ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++ * ++ * This program is free software; you can redistribute it and/or modify it under ++ * the terms of the GNU General Public License version 2 as published by the ++ * Free Software Foundation. This program is dual-licensed; you may select ++ * either version 2 of the GNU General Public License ("GPL") or BSD license ++ * ("BSD"). ++ * ++ * You can contact the author at : ++ * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy ++ */ ++#ifndef FSE_H ++#define FSE_H ++ ++/*-***************************************** ++* Dependencies ++******************************************/ ++#include /* size_t, ptrdiff_t */ ++ ++/*-***************************************** ++* FSE_PUBLIC_API : control library symbols visibility ++******************************************/ ++#define FSE_PUBLIC_API ++ ++/*------ Version ------*/ ++#define FSE_VERSION_MAJOR 0 ++#define FSE_VERSION_MINOR 9 ++#define FSE_VERSION_RELEASE 0 ++ ++#define FSE_LIB_VERSION FSE_VERSION_MAJOR.FSE_VERSION_MINOR.FSE_VERSION_RELEASE ++#define FSE_QUOTE(str) #str ++#define FSE_EXPAND_AND_QUOTE(str) FSE_QUOTE(str) ++#define FSE_VERSION_STRING FSE_EXPAND_AND_QUOTE(FSE_LIB_VERSION) ++ ++#define FSE_VERSION_NUMBER (FSE_VERSION_MAJOR * 100 * 100 + FSE_VERSION_MINOR * 100 + FSE_VERSION_RELEASE) ++FSE_PUBLIC_API unsigned FSE_versionNumber(void); /**< library version number; to be used when checking dll version */ ++ ++/*-***************************************** ++* Tool functions ++******************************************/ ++FSE_PUBLIC_API size_t FSE_compressBound(size_t size); /* maximum compressed size */ ++ ++/* Error Management */ ++FSE_PUBLIC_API unsigned FSE_isError(size_t code); /* tells if a return value is an error code */ ++ ++/*-***************************************** ++* FSE detailed API ++******************************************/ ++/*! ++FSE_compress() does the following: ++1. count symbol occurrence from source[] into table count[] ++2. normalize counters so that sum(count[]) == Power_of_2 (2^tableLog) ++3. save normalized counters to memory buffer using writeNCount() ++4. build encoding table 'CTable' from normalized counters ++5. encode the data stream using encoding table 'CTable' ++ ++FSE_decompress() does the following: ++1. read normalized counters with readNCount() ++2. build decoding table 'DTable' from normalized counters ++3. decode the data stream using decoding table 'DTable' ++ ++The following API allows targeting specific sub-functions for advanced tasks. ++For example, it's possible to compress several blocks using the same 'CTable', ++or to save and provide normalized distribution using external method. ++*/ ++ ++/* *** COMPRESSION *** */ ++/*! FSE_optimalTableLog(): ++ dynamically downsize 'tableLog' when conditions are met. ++ It saves CPU time, by using smaller tables, while preserving or even improving compression ratio. ++ @return : recommended tableLog (necessarily <= 'maxTableLog') */ ++FSE_PUBLIC_API unsigned FSE_optimalTableLog(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue); ++ ++/*! FSE_normalizeCount(): ++ normalize counts so that sum(count[]) == Power_of_2 (2^tableLog) ++ 'normalizedCounter' is a table of short, of minimum size (maxSymbolValue+1). ++ @return : tableLog, ++ or an errorCode, which can be tested using FSE_isError() */ ++FSE_PUBLIC_API size_t FSE_normalizeCount(short *normalizedCounter, unsigned tableLog, const unsigned *count, size_t srcSize, unsigned maxSymbolValue); ++ ++/*! FSE_NCountWriteBound(): ++ Provides the maximum possible size of an FSE normalized table, given 'maxSymbolValue' and 'tableLog'. ++ Typically useful for allocation purpose. */ ++FSE_PUBLIC_API size_t FSE_NCountWriteBound(unsigned maxSymbolValue, unsigned tableLog); ++ ++/*! FSE_writeNCount(): ++ Compactly save 'normalizedCounter' into 'buffer'. ++ @return : size of the compressed table, ++ or an errorCode, which can be tested using FSE_isError(). */ ++FSE_PUBLIC_API size_t FSE_writeNCount(void *buffer, size_t bufferSize, const short *normalizedCounter, unsigned maxSymbolValue, unsigned tableLog); ++ ++/*! Constructor and Destructor of FSE_CTable. ++ Note that FSE_CTable size depends on 'tableLog' and 'maxSymbolValue' */ ++typedef unsigned FSE_CTable; /* don't allocate that. It's only meant to be more restrictive than void* */ ++ ++/*! FSE_compress_usingCTable(): ++ Compress `src` using `ct` into `dst` which must be already allocated. ++ @return : size of compressed data (<= `dstCapacity`), ++ or 0 if compressed data could not fit into `dst`, ++ or an errorCode, which can be tested using FSE_isError() */ ++FSE_PUBLIC_API size_t FSE_compress_usingCTable(void *dst, size_t dstCapacity, const void *src, size_t srcSize, const FSE_CTable *ct); ++ ++/*! ++Tutorial : ++---------- ++The first step is to count all symbols. FSE_count() does this job very fast. ++Result will be saved into 'count', a table of unsigned int, which must be already allocated, and have 'maxSymbolValuePtr[0]+1' cells. ++'src' is a table of bytes of size 'srcSize'. All values within 'src' MUST be <= maxSymbolValuePtr[0] ++maxSymbolValuePtr[0] will be updated, with its real value (necessarily <= original value) ++FSE_count() will return the number of occurrence of the most frequent symbol. ++This can be used to know if there is a single symbol within 'src', and to quickly evaluate its compressibility. ++If there is an error, the function will return an ErrorCode (which can be tested using FSE_isError()). ++ ++The next step is to normalize the frequencies. ++FSE_normalizeCount() will ensure that sum of frequencies is == 2 ^'tableLog'. ++It also guarantees a minimum of 1 to any Symbol with frequency >= 1. ++You can use 'tableLog'==0 to mean "use default tableLog value". ++If you are unsure of which tableLog value to use, you can ask FSE_optimalTableLog(), ++which will provide the optimal valid tableLog given sourceSize, maxSymbolValue, and a user-defined maximum (0 means "default"). ++ ++The result of FSE_normalizeCount() will be saved into a table, ++called 'normalizedCounter', which is a table of signed short. ++'normalizedCounter' must be already allocated, and have at least 'maxSymbolValue+1' cells. ++The return value is tableLog if everything proceeded as expected. ++It is 0 if there is a single symbol within distribution. ++If there is an error (ex: invalid tableLog value), the function will return an ErrorCode (which can be tested using FSE_isError()). ++ ++'normalizedCounter' can be saved in a compact manner to a memory area using FSE_writeNCount(). ++'buffer' must be already allocated. ++For guaranteed success, buffer size must be at least FSE_headerBound(). ++The result of the function is the number of bytes written into 'buffer'. ++If there is an error, the function will return an ErrorCode (which can be tested using FSE_isError(); ex : buffer size too small). ++ ++'normalizedCounter' can then be used to create the compression table 'CTable'. ++The space required by 'CTable' must be already allocated, using FSE_createCTable(). ++You can then use FSE_buildCTable() to fill 'CTable'. ++If there is an error, both functions will return an ErrorCode (which can be tested using FSE_isError()). ++ ++'CTable' can then be used to compress 'src', with FSE_compress_usingCTable(). ++Similar to FSE_count(), the convention is that 'src' is assumed to be a table of char of size 'srcSize' ++The function returns the size of compressed data (without header), necessarily <= `dstCapacity`. ++If it returns '0', compressed data could not fit into 'dst'. ++If there is an error, the function will return an ErrorCode (which can be tested using FSE_isError()). ++*/ ++ ++/* *** DECOMPRESSION *** */ ++ ++/*! FSE_readNCount(): ++ Read compactly saved 'normalizedCounter' from 'rBuffer'. ++ @return : size read from 'rBuffer', ++ or an errorCode, which can be tested using FSE_isError(). ++ maxSymbolValuePtr[0] and tableLogPtr[0] will also be updated with their respective values */ ++FSE_PUBLIC_API size_t FSE_readNCount(short *normalizedCounter, unsigned *maxSymbolValuePtr, unsigned *tableLogPtr, const void *rBuffer, size_t rBuffSize); ++ ++/*! Constructor and Destructor of FSE_DTable. ++ Note that its size depends on 'tableLog' */ ++typedef unsigned FSE_DTable; /* don't allocate that. It's just a way to be more restrictive than void* */ ++ ++/*! FSE_buildDTable(): ++ Builds 'dt', which must be already allocated, using FSE_createDTable(). ++ return : 0, or an errorCode, which can be tested using FSE_isError() */ ++FSE_PUBLIC_API size_t FSE_buildDTable_wksp(FSE_DTable *dt, const short *normalizedCounter, unsigned maxSymbolValue, unsigned tableLog, void *workspace, size_t workspaceSize); ++ ++/*! FSE_decompress_usingDTable(): ++ Decompress compressed source `cSrc` of size `cSrcSize` using `dt` ++ into `dst` which must be already allocated. ++ @return : size of regenerated data (necessarily <= `dstCapacity`), ++ or an errorCode, which can be tested using FSE_isError() */ ++FSE_PUBLIC_API size_t FSE_decompress_usingDTable(void *dst, size_t dstCapacity, const void *cSrc, size_t cSrcSize, const FSE_DTable *dt); ++ ++/*! ++Tutorial : ++---------- ++(Note : these functions only decompress FSE-compressed blocks. ++ If block is uncompressed, use memcpy() instead ++ If block is a single repeated byte, use memset() instead ) ++ ++The first step is to obtain the normalized frequencies of symbols. ++This can be performed by FSE_readNCount() if it was saved using FSE_writeNCount(). ++'normalizedCounter' must be already allocated, and have at least 'maxSymbolValuePtr[0]+1' cells of signed short. ++In practice, that means it's necessary to know 'maxSymbolValue' beforehand, ++or size the table to handle worst case situations (typically 256). ++FSE_readNCount() will provide 'tableLog' and 'maxSymbolValue'. ++The result of FSE_readNCount() is the number of bytes read from 'rBuffer'. ++Note that 'rBufferSize' must be at least 4 bytes, even if useful information is less than that. ++If there is an error, the function will return an error code, which can be tested using FSE_isError(). ++ ++The next step is to build the decompression tables 'FSE_DTable' from 'normalizedCounter'. ++This is performed by the function FSE_buildDTable(). ++The space required by 'FSE_DTable' must be already allocated using FSE_createDTable(). ++If there is an error, the function will return an error code, which can be tested using FSE_isError(). ++ ++`FSE_DTable` can then be used to decompress `cSrc`, with FSE_decompress_usingDTable(). ++`cSrcSize` must be strictly correct, otherwise decompression will fail. ++FSE_decompress_usingDTable() result will tell how many bytes were regenerated (<=`dstCapacity`). ++If there is an error, the function will return an error code, which can be tested using FSE_isError(). (ex: dst buffer too small) ++*/ ++ ++/* *** Dependency *** */ ++#include "bitstream.h" ++ ++/* ***************************************** ++* Static allocation ++*******************************************/ ++/* FSE buffer bounds */ ++#define FSE_NCOUNTBOUND 512 ++#define FSE_BLOCKBOUND(size) (size + (size >> 7)) ++#define FSE_COMPRESSBOUND(size) (FSE_NCOUNTBOUND + FSE_BLOCKBOUND(size)) /* Macro version, useful for static allocation */ ++ ++/* It is possible to statically allocate FSE CTable/DTable as a table of FSE_CTable/FSE_DTable using below macros */ ++#define FSE_CTABLE_SIZE_U32(maxTableLog, maxSymbolValue) (1 + (1 << (maxTableLog - 1)) + ((maxSymbolValue + 1) * 2)) ++#define FSE_DTABLE_SIZE_U32(maxTableLog) (1 + (1 << maxTableLog)) ++ ++/* ***************************************** ++* FSE advanced API ++*******************************************/ ++/* FSE_count_wksp() : ++ * Same as FSE_count(), but using an externally provided scratch buffer. ++ * `workSpace` size must be table of >= `1024` unsigned ++ */ ++size_t FSE_count_wksp(unsigned *count, unsigned *maxSymbolValuePtr, const void *source, size_t sourceSize, unsigned *workSpace); ++ ++/* FSE_countFast_wksp() : ++ * Same as FSE_countFast(), but using an externally provided scratch buffer. ++ * `workSpace` must be a table of minimum `1024` unsigned ++ */ ++size_t FSE_countFast_wksp(unsigned *count, unsigned *maxSymbolValuePtr, const void *src, size_t srcSize, unsigned *workSpace); ++ ++/*! FSE_count_simple ++ * Same as FSE_countFast(), but does not use any additional memory (not even on stack). ++ * This function is unsafe, and will segfault if any value within `src` is `> *maxSymbolValuePtr` (presuming it's also the size of `count`). ++*/ ++size_t FSE_count_simple(unsigned *count, unsigned *maxSymbolValuePtr, const void *src, size_t srcSize); ++ ++unsigned FSE_optimalTableLog_internal(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue, unsigned minus); ++/**< same as FSE_optimalTableLog(), which used `minus==2` */ ++ ++size_t FSE_buildCTable_raw(FSE_CTable *ct, unsigned nbBits); ++/**< build a fake FSE_CTable, designed for a flat distribution, where each symbol uses nbBits */ ++ ++size_t FSE_buildCTable_rle(FSE_CTable *ct, unsigned char symbolValue); ++/**< build a fake FSE_CTable, designed to compress always the same symbolValue */ ++ ++/* FSE_buildCTable_wksp() : ++ * Same as FSE_buildCTable(), but using an externally allocated scratch buffer (`workSpace`). ++ * `wkspSize` must be >= `(1<= BIT_DStream_completed ++ ++When it's done, verify decompression is fully completed, by checking both DStream and the relevant states. ++Checking if DStream has reached its end is performed by : ++ BIT_endOfDStream(&DStream); ++Check also the states. There might be some symbols left there, if some high probability ones (>50%) are possible. ++ FSE_endOfDState(&DState); ++*/ ++ ++/* ***************************************** ++* FSE unsafe API ++*******************************************/ ++static unsigned char FSE_decodeSymbolFast(FSE_DState_t *DStatePtr, BIT_DStream_t *bitD); ++/* faster, but works only if nbBits is always >= 1 (otherwise, result will be corrupted) */ ++ ++/* ***************************************** ++* Implementation of inlined functions ++*******************************************/ ++typedef struct { ++ int deltaFindState; ++ U32 deltaNbBits; ++} FSE_symbolCompressionTransform; /* total 8 bytes */ ++ ++ZSTD_STATIC void FSE_initCState(FSE_CState_t *statePtr, const FSE_CTable *ct) ++{ ++ const void *ptr = ct; ++ const U16 *u16ptr = (const U16 *)ptr; ++ const U32 tableLog = ZSTD_read16(ptr); ++ statePtr->value = (ptrdiff_t)1 << tableLog; ++ statePtr->stateTable = u16ptr + 2; ++ statePtr->symbolTT = ((const U32 *)ct + 1 + (tableLog ? (1 << (tableLog - 1)) : 1)); ++ statePtr->stateLog = tableLog; ++} ++ ++/*! FSE_initCState2() : ++* Same as FSE_initCState(), but the first symbol to include (which will be the last to be read) ++* uses the smallest state value possible, saving the cost of this symbol */ ++ZSTD_STATIC void FSE_initCState2(FSE_CState_t *statePtr, const FSE_CTable *ct, U32 symbol) ++{ ++ FSE_initCState(statePtr, ct); ++ { ++ const FSE_symbolCompressionTransform symbolTT = ((const FSE_symbolCompressionTransform *)(statePtr->symbolTT))[symbol]; ++ const U16 *stateTable = (const U16 *)(statePtr->stateTable); ++ U32 nbBitsOut = (U32)((symbolTT.deltaNbBits + (1 << 15)) >> 16); ++ statePtr->value = (nbBitsOut << 16) - symbolTT.deltaNbBits; ++ statePtr->value = stateTable[(statePtr->value >> nbBitsOut) + symbolTT.deltaFindState]; ++ } ++} ++ ++ZSTD_STATIC void FSE_encodeSymbol(BIT_CStream_t *bitC, FSE_CState_t *statePtr, U32 symbol) ++{ ++ const FSE_symbolCompressionTransform symbolTT = ((const FSE_symbolCompressionTransform *)(statePtr->symbolTT))[symbol]; ++ const U16 *const stateTable = (const U16 *)(statePtr->stateTable); ++ U32 nbBitsOut = (U32)((statePtr->value + symbolTT.deltaNbBits) >> 16); ++ BIT_addBits(bitC, statePtr->value, nbBitsOut); ++ statePtr->value = stateTable[(statePtr->value >> nbBitsOut) + symbolTT.deltaFindState]; ++} ++ ++ZSTD_STATIC void FSE_flushCState(BIT_CStream_t *bitC, const FSE_CState_t *statePtr) ++{ ++ BIT_addBits(bitC, statePtr->value, statePtr->stateLog); ++ BIT_flushBits(bitC); ++} ++ ++/* ====== Decompression ====== */ ++ ++typedef struct { ++ U16 tableLog; ++ U16 fastMode; ++} FSE_DTableHeader; /* sizeof U32 */ ++ ++typedef struct { ++ unsigned short newState; ++ unsigned char symbol; ++ unsigned char nbBits; ++} FSE_decode_t; /* size == U32 */ ++ ++ZSTD_STATIC void FSE_initDState(FSE_DState_t *DStatePtr, BIT_DStream_t *bitD, const FSE_DTable *dt) ++{ ++ const void *ptr = dt; ++ const FSE_DTableHeader *const DTableH = (const FSE_DTableHeader *)ptr; ++ DStatePtr->state = BIT_readBits(bitD, DTableH->tableLog); ++ BIT_reloadDStream(bitD); ++ DStatePtr->table = dt + 1; ++} ++ ++ZSTD_STATIC BYTE FSE_peekSymbol(const FSE_DState_t *DStatePtr) ++{ ++ FSE_decode_t const DInfo = ((const FSE_decode_t *)(DStatePtr->table))[DStatePtr->state]; ++ return DInfo.symbol; ++} ++ ++ZSTD_STATIC void FSE_updateState(FSE_DState_t *DStatePtr, BIT_DStream_t *bitD) ++{ ++ FSE_decode_t const DInfo = ((const FSE_decode_t *)(DStatePtr->table))[DStatePtr->state]; ++ U32 const nbBits = DInfo.nbBits; ++ size_t const lowBits = BIT_readBits(bitD, nbBits); ++ DStatePtr->state = DInfo.newState + lowBits; ++} ++ ++ZSTD_STATIC BYTE FSE_decodeSymbol(FSE_DState_t *DStatePtr, BIT_DStream_t *bitD) ++{ ++ FSE_decode_t const DInfo = ((const FSE_decode_t *)(DStatePtr->table))[DStatePtr->state]; ++ U32 const nbBits = DInfo.nbBits; ++ BYTE const symbol = DInfo.symbol; ++ size_t const lowBits = BIT_readBits(bitD, nbBits); ++ ++ DStatePtr->state = DInfo.newState + lowBits; ++ return symbol; ++} ++ ++/*! FSE_decodeSymbolFast() : ++ unsafe, only works if no symbol has a probability > 50% */ ++ZSTD_STATIC BYTE FSE_decodeSymbolFast(FSE_DState_t *DStatePtr, BIT_DStream_t *bitD) ++{ ++ FSE_decode_t const DInfo = ((const FSE_decode_t *)(DStatePtr->table))[DStatePtr->state]; ++ U32 const nbBits = DInfo.nbBits; ++ BYTE const symbol = DInfo.symbol; ++ size_t const lowBits = BIT_readBitsFast(bitD, nbBits); ++ ++ DStatePtr->state = DInfo.newState + lowBits; ++ return symbol; ++} ++ ++ZSTD_STATIC unsigned FSE_endOfDState(const FSE_DState_t *DStatePtr) { return DStatePtr->state == 0; } ++ ++/* ************************************************************** ++* Tuning parameters ++****************************************************************/ ++/*!MEMORY_USAGE : ++* Memory usage formula : N->2^N Bytes (examples : 10 -> 1KB; 12 -> 4KB ; 16 -> 64KB; 20 -> 1MB; etc.) ++* Increasing memory usage improves compression ratio ++* Reduced memory usage can improve speed, due to cache effect ++* Recommended max value is 14, for 16KB, which nicely fits into Intel x86 L1 cache */ ++#ifndef FSE_MAX_MEMORY_USAGE ++#define FSE_MAX_MEMORY_USAGE 14 ++#endif ++#ifndef FSE_DEFAULT_MEMORY_USAGE ++#define FSE_DEFAULT_MEMORY_USAGE 13 ++#endif ++ ++/*!FSE_MAX_SYMBOL_VALUE : ++* Maximum symbol value authorized. ++* Required for proper stack allocation */ ++#ifndef FSE_MAX_SYMBOL_VALUE ++#define FSE_MAX_SYMBOL_VALUE 255 ++#endif ++ ++/* ************************************************************** ++* template functions type & suffix ++****************************************************************/ ++#define FSE_FUNCTION_TYPE BYTE ++#define FSE_FUNCTION_EXTENSION ++#define FSE_DECODE_TYPE FSE_decode_t ++ ++/* *************************************************************** ++* Constants ++*****************************************************************/ ++#define FSE_MAX_TABLELOG (FSE_MAX_MEMORY_USAGE - 2) ++#define FSE_MAX_TABLESIZE (1U << FSE_MAX_TABLELOG) ++#define FSE_MAXTABLESIZE_MASK (FSE_MAX_TABLESIZE - 1) ++#define FSE_DEFAULT_TABLELOG (FSE_DEFAULT_MEMORY_USAGE - 2) ++#define FSE_MIN_TABLELOG 5 ++ ++#define FSE_TABLELOG_ABSOLUTE_MAX 15 ++#if FSE_MAX_TABLELOG > FSE_TABLELOG_ABSOLUTE_MAX ++#error "FSE_MAX_TABLELOG > FSE_TABLELOG_ABSOLUTE_MAX is not supported" ++#endif ++ ++#define FSE_TABLESTEP(tableSize) ((tableSize >> 1) + (tableSize >> 3) + 3) ++ ++#endif /* FSE_H */ +diff --git a/xen/common/zstd/fse_decompress.c b/xen/common/zstd/fse_decompress.c +new file mode 100644 +index 0000000000..041a5a1f0a +--- /dev/null ++++ b/xen/common/zstd/fse_decompress.c +@@ -0,0 +1,323 @@ ++/* ++ * FSE : Finite State Entropy decoder ++ * Copyright (C) 2013-2015, Yann Collet. ++ * ++ * BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions are ++ * met: ++ * ++ * * Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * * Redistributions in binary form must reproduce the above ++ * copyright notice, this list of conditions and the following disclaimer ++ * in the documentation and/or other materials provided with the ++ * distribution. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS ++ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT ++ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR ++ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT ++ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, ++ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT ++ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE ++ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++ * ++ * This program is free software; you can redistribute it and/or modify it under ++ * the terms of the GNU General Public License version 2 as published by the ++ * Free Software Foundation. This program is dual-licensed; you may select ++ * either version 2 of the GNU General Public License ("GPL") or BSD license ++ * ("BSD"). ++ * ++ * You can contact the author at : ++ * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy ++ */ ++ ++/* ************************************************************** ++* Compiler specifics ++****************************************************************/ ++#define FORCE_INLINE static always_inline ++ ++/* ************************************************************** ++* Includes ++****************************************************************/ ++#include "bitstream.h" ++#include "fse.h" ++#include "zstd_internal.h" ++#include /* memcpy, memset */ ++ ++/* ************************************************************** ++* Error Management ++****************************************************************/ ++#define FSE_isError ERR_isError ++#define FSE_STATIC_ASSERT(c) \ ++ { \ ++ enum { FSE_static_assert = 1 / (int)(!!(c)) }; \ ++ } /* use only *after* variable declarations */ ++ ++/* ************************************************************** ++* Templates ++****************************************************************/ ++/* ++ designed to be included ++ for type-specific functions (template emulation in C) ++ Objective is to write these functions only once, for improved maintenance ++*/ ++ ++/* safety checks */ ++#ifndef FSE_FUNCTION_EXTENSION ++#error "FSE_FUNCTION_EXTENSION must be defined" ++#endif ++#ifndef FSE_FUNCTION_TYPE ++#error "FSE_FUNCTION_TYPE must be defined" ++#endif ++ ++/* Function names */ ++#define FSE_CAT(X, Y) X##Y ++#define FSE_FUNCTION_NAME(X, Y) FSE_CAT(X, Y) ++#define FSE_TYPE_NAME(X, Y) FSE_CAT(X, Y) ++ ++/* Function templates */ ++ ++size_t INIT FSE_buildDTable_wksp(FSE_DTable *dt, const short *normalizedCounter, unsigned maxSymbolValue, unsigned tableLog, void *workspace, size_t workspaceSize) ++{ ++ void *const tdPtr = dt + 1; /* because *dt is unsigned, 32-bits aligned on 32-bits */ ++ FSE_DECODE_TYPE *const tableDecode = (FSE_DECODE_TYPE *)(tdPtr); ++ U16 *symbolNext = (U16 *)workspace; ++ ++ U32 const maxSV1 = maxSymbolValue + 1; ++ U32 const tableSize = 1 << tableLog; ++ U32 highThreshold = tableSize - 1; ++ ++ /* Sanity Checks */ ++ if (workspaceSize < sizeof(U16) * (FSE_MAX_SYMBOL_VALUE + 1)) ++ return ERROR(tableLog_tooLarge); ++ if (maxSymbolValue > FSE_MAX_SYMBOL_VALUE) ++ return ERROR(maxSymbolValue_tooLarge); ++ if (tableLog > FSE_MAX_TABLELOG) ++ return ERROR(tableLog_tooLarge); ++ ++ /* Init, lay down lowprob symbols */ ++ { ++ FSE_DTableHeader DTableH; ++ DTableH.tableLog = (U16)tableLog; ++ DTableH.fastMode = 1; ++ { ++ S16 const largeLimit = (S16)(1 << (tableLog - 1)); ++ U32 s; ++ for (s = 0; s < maxSV1; s++) { ++ if (normalizedCounter[s] == -1) { ++ tableDecode[highThreshold--].symbol = (FSE_FUNCTION_TYPE)s; ++ symbolNext[s] = 1; ++ } else { ++ if (normalizedCounter[s] >= largeLimit) ++ DTableH.fastMode = 0; ++ symbolNext[s] = normalizedCounter[s]; ++ } ++ } ++ } ++ memcpy(dt, &DTableH, sizeof(DTableH)); ++ } ++ ++ /* Spread symbols */ ++ { ++ U32 const tableMask = tableSize - 1; ++ U32 const step = FSE_TABLESTEP(tableSize); ++ U32 s, position = 0; ++ for (s = 0; s < maxSV1; s++) { ++ int i; ++ for (i = 0; i < normalizedCounter[s]; i++) { ++ tableDecode[position].symbol = (FSE_FUNCTION_TYPE)s; ++ position = (position + step) & tableMask; ++ while (position > highThreshold) ++ position = (position + step) & tableMask; /* lowprob area */ ++ } ++ } ++ if (position != 0) ++ return ERROR(GENERIC); /* position must reach all cells once, otherwise normalizedCounter is incorrect */ ++ } ++ ++ /* Build Decoding table */ ++ { ++ U32 u; ++ for (u = 0; u < tableSize; u++) { ++ FSE_FUNCTION_TYPE const symbol = (FSE_FUNCTION_TYPE)(tableDecode[u].symbol); ++ U16 nextState = symbolNext[symbol]++; ++ tableDecode[u].nbBits = (BYTE)(tableLog - BIT_highbit32((U32)nextState)); ++ tableDecode[u].newState = (U16)((nextState << tableDecode[u].nbBits) - tableSize); ++ } ++ } ++ ++ return 0; ++} ++ ++/*-******************************************************* ++* Decompression (Byte symbols) ++*********************************************************/ ++size_t INIT FSE_buildDTable_rle(FSE_DTable *dt, BYTE symbolValue) ++{ ++ void *ptr = dt; ++ FSE_DTableHeader *const DTableH = (FSE_DTableHeader *)ptr; ++ void *dPtr = dt + 1; ++ FSE_decode_t *const cell = (FSE_decode_t *)dPtr; ++ ++ DTableH->tableLog = 0; ++ DTableH->fastMode = 0; ++ ++ cell->newState = 0; ++ cell->symbol = symbolValue; ++ cell->nbBits = 0; ++ ++ return 0; ++} ++ ++size_t INIT FSE_buildDTable_raw(FSE_DTable *dt, unsigned nbBits) ++{ ++ void *ptr = dt; ++ FSE_DTableHeader *const DTableH = (FSE_DTableHeader *)ptr; ++ void *dPtr = dt + 1; ++ FSE_decode_t *const dinfo = (FSE_decode_t *)dPtr; ++ const unsigned tableSize = 1 << nbBits; ++ const unsigned tableMask = tableSize - 1; ++ const unsigned maxSV1 = tableMask + 1; ++ unsigned s; ++ ++ /* Sanity checks */ ++ if (nbBits < 1) ++ return ERROR(GENERIC); /* min size */ ++ ++ /* Build Decoding Table */ ++ DTableH->tableLog = (U16)nbBits; ++ DTableH->fastMode = 1; ++ for (s = 0; s < maxSV1; s++) { ++ dinfo[s].newState = 0; ++ dinfo[s].symbol = (BYTE)s; ++ dinfo[s].nbBits = (BYTE)nbBits; ++ } ++ ++ return 0; ++} ++ ++FORCE_INLINE size_t FSE_decompress_usingDTable_generic(void *dst, size_t maxDstSize, const void *cSrc, size_t cSrcSize, const FSE_DTable *dt, ++ const unsigned fast) ++{ ++ BYTE *const ostart = (BYTE *)dst; ++ BYTE *op = ostart; ++ BYTE *const omax = op + maxDstSize; ++ BYTE *const olimit = omax - 3; ++ ++ BIT_DStream_t bitD; ++ FSE_DState_t state1; ++ FSE_DState_t state2; ++ ++ /* Init */ ++ CHECK_F(BIT_initDStream(&bitD, cSrc, cSrcSize)); ++ ++ FSE_initDState(&state1, &bitD, dt); ++ FSE_initDState(&state2, &bitD, dt); ++ ++#define FSE_GETSYMBOL(statePtr) fast ? FSE_decodeSymbolFast(statePtr, &bitD) : FSE_decodeSymbol(statePtr, &bitD) ++ ++ /* 4 symbols per loop */ ++ for (; (BIT_reloadDStream(&bitD) == BIT_DStream_unfinished) & (op < olimit); op += 4) { ++ op[0] = FSE_GETSYMBOL(&state1); ++ ++ if (FSE_MAX_TABLELOG * 2 + 7 > sizeof(bitD.bitContainer) * 8) /* This test must be static */ ++ BIT_reloadDStream(&bitD); ++ ++ op[1] = FSE_GETSYMBOL(&state2); ++ ++ if (FSE_MAX_TABLELOG * 4 + 7 > sizeof(bitD.bitContainer) * 8) /* This test must be static */ ++ { ++ if (BIT_reloadDStream(&bitD) > BIT_DStream_unfinished) { ++ op += 2; ++ break; ++ } ++ } ++ ++ op[2] = FSE_GETSYMBOL(&state1); ++ ++ if (FSE_MAX_TABLELOG * 2 + 7 > sizeof(bitD.bitContainer) * 8) /* This test must be static */ ++ BIT_reloadDStream(&bitD); ++ ++ op[3] = FSE_GETSYMBOL(&state2); ++ } ++ ++ /* tail */ ++ /* note : BIT_reloadDStream(&bitD) >= FSE_DStream_partiallyFilled; Ends at exactly BIT_DStream_completed */ ++ while (1) { ++ if (op > (omax - 2)) ++ return ERROR(dstSize_tooSmall); ++ *op++ = FSE_GETSYMBOL(&state1); ++ if (BIT_reloadDStream(&bitD) == BIT_DStream_overflow) { ++ *op++ = FSE_GETSYMBOL(&state2); ++ break; ++ } ++ ++ if (op > (omax - 2)) ++ return ERROR(dstSize_tooSmall); ++ *op++ = FSE_GETSYMBOL(&state2); ++ if (BIT_reloadDStream(&bitD) == BIT_DStream_overflow) { ++ *op++ = FSE_GETSYMBOL(&state1); ++ break; ++ } ++ } ++ ++ return op - ostart; ++} ++ ++size_t INIT FSE_decompress_usingDTable(void *dst, size_t originalSize, const void *cSrc, size_t cSrcSize, const FSE_DTable *dt) ++{ ++ const void *ptr = dt; ++ const FSE_DTableHeader *DTableH = (const FSE_DTableHeader *)ptr; ++ const U32 fastMode = DTableH->fastMode; ++ ++ /* select fast mode (static) */ ++ if (fastMode) ++ return FSE_decompress_usingDTable_generic(dst, originalSize, cSrc, cSrcSize, dt, 1); ++ return FSE_decompress_usingDTable_generic(dst, originalSize, cSrc, cSrcSize, dt, 0); ++} ++ ++size_t INIT FSE_decompress_wksp(void *dst, size_t dstCapacity, const void *cSrc, size_t cSrcSize, unsigned maxLog, void *workspace, size_t workspaceSize) ++{ ++ const BYTE *const istart = (const BYTE *)cSrc; ++ const BYTE *ip = istart; ++ unsigned tableLog; ++ unsigned maxSymbolValue = FSE_MAX_SYMBOL_VALUE; ++ size_t NCountLength; ++ ++ FSE_DTable *dt; ++ short *counting; ++ size_t spaceUsed32 = 0; ++ ++ FSE_STATIC_ASSERT(sizeof(FSE_DTable) == sizeof(U32)); ++ ++ dt = (FSE_DTable *)((U32 *)workspace + spaceUsed32); ++ spaceUsed32 += FSE_DTABLE_SIZE_U32(maxLog); ++ counting = (short *)((U32 *)workspace + spaceUsed32); ++ spaceUsed32 += ALIGN(sizeof(short) * (FSE_MAX_SYMBOL_VALUE + 1), sizeof(U32)) >> 2; ++ ++ if ((spaceUsed32 << 2) > workspaceSize) ++ return ERROR(tableLog_tooLarge); ++ workspace = (U32 *)workspace + spaceUsed32; ++ workspaceSize -= (spaceUsed32 << 2); ++ ++ /* normal FSE decoding mode */ ++ NCountLength = FSE_readNCount(counting, &maxSymbolValue, &tableLog, istart, cSrcSize); ++ if (FSE_isError(NCountLength)) ++ return NCountLength; ++ // if (NCountLength >= cSrcSize) return ERROR(srcSize_wrong); /* too small input size; supposed to be already checked in NCountLength, only remaining ++ // case : NCountLength==cSrcSize */ ++ if (tableLog > maxLog) ++ return ERROR(tableLog_tooLarge); ++ ip += NCountLength; ++ cSrcSize -= NCountLength; ++ ++ CHECK_F(FSE_buildDTable_wksp(dt, counting, maxSymbolValue, tableLog, workspace, workspaceSize)); ++ ++ return FSE_decompress_usingDTable(dst, dstCapacity, ip, cSrcSize, dt); /* always return, even if it is an error code */ ++} +diff --git a/xen/common/zstd/huf.h b/xen/common/zstd/huf.h +new file mode 100644 +index 0000000000..a9d522c7bb +--- /dev/null ++++ b/xen/common/zstd/huf.h +@@ -0,0 +1,212 @@ ++/* ++ * Huffman coder, part of New Generation Entropy library ++ * header file ++ * Copyright (C) 2013-2016, Yann Collet. ++ * ++ * BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions are ++ * met: ++ * ++ * * Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * * Redistributions in binary form must reproduce the above ++ * copyright notice, this list of conditions and the following disclaimer ++ * in the documentation and/or other materials provided with the ++ * distribution. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS ++ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT ++ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR ++ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT ++ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, ++ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT ++ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE ++ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++ * ++ * This program is free software; you can redistribute it and/or modify it under ++ * the terms of the GNU General Public License version 2 as published by the ++ * Free Software Foundation. This program is dual-licensed; you may select ++ * either version 2 of the GNU General Public License ("GPL") or BSD license ++ * ("BSD"). ++ * ++ * You can contact the author at : ++ * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy ++ */ ++#ifndef HUF_H_298734234 ++#define HUF_H_298734234 ++ ++/* *** Dependencies *** */ ++#include /* size_t */ ++ ++/* *** Tool functions *** */ ++#define HUF_BLOCKSIZE_MAX (128 * 1024) /**< maximum input size for a single block compressed with HUF_compress */ ++size_t HUF_compressBound(size_t size); /**< maximum compressed size (worst case) */ ++ ++/* Error Management */ ++unsigned HUF_isError(size_t code); /**< tells if a return value is an error code */ ++ ++/* *** Advanced function *** */ ++ ++/** HUF_compress4X_wksp() : ++* Same as HUF_compress2(), but uses externally allocated `workSpace`, which must be a table of >= 1024 unsigned */ ++size_t HUF_compress4X_wksp(void *dst, size_t dstSize, const void *src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog, void *workSpace, ++ size_t wkspSize); /**< `workSpace` must be a table of at least HUF_COMPRESS_WORKSPACE_SIZE_U32 unsigned */ ++ ++/* *** Dependencies *** */ ++#include "mem.h" /* U32 */ ++ ++/* *** Constants *** */ ++#define HUF_TABLELOG_MAX 12 /* max configured tableLog (for static allocation); can be modified up to HUF_ABSOLUTEMAX_TABLELOG */ ++#define HUF_TABLELOG_DEFAULT 11 /* tableLog by default, when not specified */ ++#define HUF_SYMBOLVALUE_MAX 255 ++ ++#define HUF_TABLELOG_ABSOLUTEMAX 15 /* absolute limit of HUF_MAX_TABLELOG. Beyond that value, code does not work */ ++#if (HUF_TABLELOG_MAX > HUF_TABLELOG_ABSOLUTEMAX) ++#error "HUF_TABLELOG_MAX is too large !" ++#endif ++ ++/* **************************************** ++* Static allocation ++******************************************/ ++/* HUF buffer bounds */ ++#define HUF_CTABLEBOUND 129 ++#define HUF_BLOCKBOUND(size) (size + (size >> 8) + 8) /* only true if incompressible pre-filtered with fast heuristic */ ++#define HUF_COMPRESSBOUND(size) (HUF_CTABLEBOUND + HUF_BLOCKBOUND(size)) /* Macro version, useful for static allocation */ ++ ++/* static allocation of HUF's Compression Table */ ++#define HUF_CREATE_STATIC_CTABLE(name, maxSymbolValue) \ ++ U32 name##hb[maxSymbolValue + 1]; \ ++ void *name##hv = &(name##hb); \ ++ HUF_CElt *name = (HUF_CElt *)(name##hv) /* no final ; */ ++ ++/* static allocation of HUF's DTable */ ++typedef U32 HUF_DTable; ++#define HUF_DTABLE_SIZE(maxTableLog) (1 + (1 << (maxTableLog))) ++#define HUF_CREATE_STATIC_DTABLEX2(DTable, maxTableLog) HUF_DTable DTable[HUF_DTABLE_SIZE((maxTableLog)-1)] = {((U32)((maxTableLog)-1) * 0x01000001)} ++#define HUF_CREATE_STATIC_DTABLEX4(DTable, maxTableLog) HUF_DTable DTable[HUF_DTABLE_SIZE(maxTableLog)] = {((U32)(maxTableLog)*0x01000001)} ++ ++/* The workspace must have alignment at least 4 and be at least this large */ ++#define HUF_COMPRESS_WORKSPACE_SIZE (6 << 10) ++#define HUF_COMPRESS_WORKSPACE_SIZE_U32 (HUF_COMPRESS_WORKSPACE_SIZE / sizeof(U32)) ++ ++/* The workspace must have alignment at least 4 and be at least this large */ ++#define HUF_DECOMPRESS_WORKSPACE_SIZE (3 << 10) ++#define HUF_DECOMPRESS_WORKSPACE_SIZE_U32 (HUF_DECOMPRESS_WORKSPACE_SIZE / sizeof(U32)) ++ ++/* **************************************** ++* Advanced decompression functions ++******************************************/ ++size_t HUF_decompress4X_DCtx_wksp(HUF_DTable *dctx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace, size_t workspaceSize); /**< decodes RLE and uncompressed */ ++size_t HUF_decompress4X_hufOnly_wksp(HUF_DTable *dctx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace, ++ size_t workspaceSize); /**< considers RLE and uncompressed as errors */ ++size_t HUF_decompress4X2_DCtx_wksp(HUF_DTable *dctx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace, ++ size_t workspaceSize); /**< single-symbol decoder */ ++size_t HUF_decompress4X4_DCtx_wksp(HUF_DTable *dctx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace, ++ size_t workspaceSize); /**< double-symbols decoder */ ++ ++/* **************************************** ++* HUF detailed API ++******************************************/ ++/*! ++HUF_compress() does the following: ++1. count symbol occurrence from source[] into table count[] using FSE_count() ++2. (optional) refine tableLog using HUF_optimalTableLog() ++3. build Huffman table from count using HUF_buildCTable() ++4. save Huffman table to memory buffer using HUF_writeCTable_wksp() ++5. encode the data stream using HUF_compress4X_usingCTable() ++ ++The following API allows targeting specific sub-functions for advanced tasks. ++For example, it's possible to compress several blocks using the same 'CTable', ++or to save and regenerate 'CTable' using external methods. ++*/ ++/* FSE_count() : find it within "fse.h" */ ++unsigned HUF_optimalTableLog(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue); ++typedef struct HUF_CElt_s HUF_CElt; /* incomplete type */ ++size_t HUF_writeCTable_wksp(void *dst, size_t maxDstSize, const HUF_CElt *CTable, unsigned maxSymbolValue, unsigned huffLog, void *workspace, size_t workspaceSize); ++size_t HUF_compress4X_usingCTable(void *dst, size_t dstSize, const void *src, size_t srcSize, const HUF_CElt *CTable); ++ ++typedef enum { ++ HUF_repeat_none, /**< Cannot use the previous table */ ++ HUF_repeat_check, /**< Can use the previous table but it must be checked. Note : The previous table must have been constructed by HUF_compress{1, ++ 4}X_repeat */ ++ HUF_repeat_valid /**< Can use the previous table and it is asumed to be valid */ ++} HUF_repeat; ++/** HUF_compress4X_repeat() : ++* Same as HUF_compress4X_wksp(), but considers using hufTable if *repeat != HUF_repeat_none. ++* If it uses hufTable it does not modify hufTable or repeat. ++* If it doesn't, it sets *repeat = HUF_repeat_none, and it sets hufTable to the table used. ++* If preferRepeat then the old table will always be used if valid. */ ++size_t HUF_compress4X_repeat(void *dst, size_t dstSize, const void *src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog, void *workSpace, ++ size_t wkspSize, HUF_CElt *hufTable, HUF_repeat *repeat, ++ int preferRepeat); /**< `workSpace` must be a table of at least HUF_COMPRESS_WORKSPACE_SIZE_U32 unsigned */ ++ ++/** HUF_buildCTable_wksp() : ++ * Same as HUF_buildCTable(), but using externally allocated scratch buffer. ++ * `workSpace` must be aligned on 4-bytes boundaries, and be at least as large as a table of 1024 unsigned. ++ */ ++size_t HUF_buildCTable_wksp(HUF_CElt *tree, const U32 *count, U32 maxSymbolValue, U32 maxNbBits, void *workSpace, size_t wkspSize); ++ ++/*! HUF_readStats() : ++ Read compact Huffman tree, saved by HUF_writeCTable(). ++ `huffWeight` is destination buffer. ++ @return : size read from `src` , or an error Code . ++ Note : Needed by HUF_readCTable() and HUF_readDTableXn() . */ ++size_t HUF_readStats_wksp(BYTE *huffWeight, size_t hwSize, U32 *rankStats, U32 *nbSymbolsPtr, U32 *tableLogPtr, const void *src, size_t srcSize, ++ void *workspace, size_t workspaceSize); ++ ++/** HUF_readCTable() : ++* Loading a CTable saved with HUF_writeCTable() */ ++size_t HUF_readCTable_wksp(HUF_CElt *CTable, unsigned maxSymbolValue, const void *src, size_t srcSize, void *workspace, size_t workspaceSize); ++ ++/* ++HUF_decompress() does the following: ++1. select the decompression algorithm (X2, X4) based on pre-computed heuristics ++2. build Huffman table from save, using HUF_readDTableXn() ++3. decode 1 or 4 segments in parallel using HUF_decompressSXn_usingDTable ++*/ ++ ++/** HUF_selectDecoder() : ++* Tells which decoder is likely to decode faster, ++* based on a set of pre-determined metrics. ++* @return : 0==HUF_decompress4X2, 1==HUF_decompress4X4 . ++* Assumption : 0 < cSrcSize < dstSize <= 128 KB */ ++U32 HUF_selectDecoder(size_t dstSize, size_t cSrcSize); ++ ++size_t HUF_readDTableX2_wksp(HUF_DTable *DTable, const void *src, size_t srcSize, void *workspace, size_t workspaceSize); ++size_t HUF_readDTableX4_wksp(HUF_DTable *DTable, const void *src, size_t srcSize, void *workspace, size_t workspaceSize); ++ ++size_t HUF_decompress4X_usingDTable(void *dst, size_t maxDstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable); ++size_t HUF_decompress4X2_usingDTable(void *dst, size_t maxDstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable); ++size_t HUF_decompress4X4_usingDTable(void *dst, size_t maxDstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable); ++ ++/* single stream variants */ ++ ++size_t HUF_compress1X_wksp(void *dst, size_t dstSize, const void *src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog, void *workSpace, ++ size_t wkspSize); /**< `workSpace` must be a table of at least HUF_COMPRESS_WORKSPACE_SIZE_U32 unsigned */ ++size_t HUF_compress1X_usingCTable(void *dst, size_t dstSize, const void *src, size_t srcSize, const HUF_CElt *CTable); ++/** HUF_compress1X_repeat() : ++* Same as HUF_compress1X_wksp(), but considers using hufTable if *repeat != HUF_repeat_none. ++* If it uses hufTable it does not modify hufTable or repeat. ++* If it doesn't, it sets *repeat = HUF_repeat_none, and it sets hufTable to the table used. ++* If preferRepeat then the old table will always be used if valid. */ ++size_t HUF_compress1X_repeat(void *dst, size_t dstSize, const void *src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog, void *workSpace, ++ size_t wkspSize, HUF_CElt *hufTable, HUF_repeat *repeat, ++ int preferRepeat); /**< `workSpace` must be a table of at least HUF_COMPRESS_WORKSPACE_SIZE_U32 unsigned */ ++ ++size_t HUF_decompress1X_DCtx_wksp(HUF_DTable *dctx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace, size_t workspaceSize); ++size_t HUF_decompress1X2_DCtx_wksp(HUF_DTable *dctx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace, ++ size_t workspaceSize); /**< single-symbol decoder */ ++size_t HUF_decompress1X4_DCtx_wksp(HUF_DTable *dctx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace, ++ size_t workspaceSize); /**< double-symbols decoder */ ++ ++size_t HUF_decompress1X_usingDTable(void *dst, size_t maxDstSize, const void *cSrc, size_t cSrcSize, ++ const HUF_DTable *DTable); /**< automatic selection of sing or double symbol decoder, based on DTable */ ++size_t HUF_decompress1X2_usingDTable(void *dst, size_t maxDstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable); ++size_t HUF_decompress1X4_usingDTable(void *dst, size_t maxDstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable); ++ ++#endif /* HUF_H_298734234 */ +diff --git a/xen/common/zstd/huf_decompress.c b/xen/common/zstd/huf_decompress.c +new file mode 100644 +index 0000000000..f79603a12f +--- /dev/null ++++ b/xen/common/zstd/huf_decompress.c +@@ -0,0 +1,958 @@ ++/* ++ * Huffman decoder, part of New Generation Entropy library ++ * Copyright (C) 2013-2016, Yann Collet. ++ * ++ * BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions are ++ * met: ++ * ++ * * Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * * Redistributions in binary form must reproduce the above ++ * copyright notice, this list of conditions and the following disclaimer ++ * in the documentation and/or other materials provided with the ++ * distribution. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS ++ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT ++ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR ++ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT ++ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, ++ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT ++ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE ++ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++ * ++ * This program is free software; you can redistribute it and/or modify it under ++ * the terms of the GNU General Public License version 2 as published by the ++ * Free Software Foundation. This program is dual-licensed; you may select ++ * either version 2 of the GNU General Public License ("GPL") or BSD license ++ * ("BSD"). ++ * ++ * You can contact the author at : ++ * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy ++ */ ++ ++/* ************************************************************** ++* Compiler specifics ++****************************************************************/ ++#define FORCE_INLINE static always_inline ++ ++/* ************************************************************** ++* Dependencies ++****************************************************************/ ++#include "bitstream.h" /* BIT_* */ ++#include "fse.h" /* header compression */ ++#include "huf.h" ++#include /* memcpy, memset */ ++ ++/* ************************************************************** ++* Error Management ++****************************************************************/ ++#define HUF_STATIC_ASSERT(c) \ ++ { \ ++ enum { HUF_static_assert = 1 / (int)(!!(c)) }; \ ++ } /* use only *after* variable declarations */ ++ ++/*-***************************/ ++/* generic DTableDesc */ ++/*-***************************/ ++ ++typedef struct { ++ BYTE maxTableLog; ++ BYTE tableType; ++ BYTE tableLog; ++ BYTE reserved; ++} DTableDesc; ++ ++static DTableDesc INIT HUF_getDTableDesc(const HUF_DTable *table) ++{ ++ DTableDesc dtd; ++ memcpy(&dtd, table, sizeof(dtd)); ++ return dtd; ++} ++ ++/*-***************************/ ++/* single-symbol decoding */ ++/*-***************************/ ++ ++typedef struct { ++ BYTE byte; ++ BYTE nbBits; ++} HUF_DEltX2; /* single-symbol decoding */ ++ ++size_t INIT HUF_readDTableX2_wksp(HUF_DTable *DTable, const void *src, size_t srcSize, void *workspace, size_t workspaceSize) ++{ ++ U32 tableLog = 0; ++ U32 nbSymbols = 0; ++ size_t iSize; ++ void *const dtPtr = DTable + 1; ++ HUF_DEltX2 *const dt = (HUF_DEltX2 *)dtPtr; ++ ++ U32 *rankVal; ++ BYTE *huffWeight; ++ size_t spaceUsed32 = 0; ++ ++ rankVal = (U32 *)workspace + spaceUsed32; ++ spaceUsed32 += HUF_TABLELOG_ABSOLUTEMAX + 1; ++ huffWeight = (BYTE *)((U32 *)workspace + spaceUsed32); ++ spaceUsed32 += ALIGN(HUF_SYMBOLVALUE_MAX + 1, sizeof(U32)) >> 2; ++ ++ if ((spaceUsed32 << 2) > workspaceSize) ++ return ERROR(tableLog_tooLarge); ++ workspace = (U32 *)workspace + spaceUsed32; ++ workspaceSize -= (spaceUsed32 << 2); ++ ++ HUF_STATIC_ASSERT(sizeof(DTableDesc) == sizeof(HUF_DTable)); ++ /* memset(huffWeight, 0, sizeof(huffWeight)); */ /* is not necessary, even though some analyzer complain ... */ ++ ++ iSize = HUF_readStats_wksp(huffWeight, HUF_SYMBOLVALUE_MAX + 1, rankVal, &nbSymbols, &tableLog, src, srcSize, workspace, workspaceSize); ++ if (HUF_isError(iSize)) ++ return iSize; ++ ++ /* Table header */ ++ { ++ DTableDesc dtd = HUF_getDTableDesc(DTable); ++ if (tableLog > (U32)(dtd.maxTableLog + 1)) ++ return ERROR(tableLog_tooLarge); /* DTable too small, Huffman tree cannot fit in */ ++ dtd.tableType = 0; ++ dtd.tableLog = (BYTE)tableLog; ++ memcpy(DTable, &dtd, sizeof(dtd)); ++ } ++ ++ /* Calculate starting value for each rank */ ++ { ++ U32 n, nextRankStart = 0; ++ for (n = 1; n < tableLog + 1; n++) { ++ U32 const curr = nextRankStart; ++ nextRankStart += (rankVal[n] << (n - 1)); ++ rankVal[n] = curr; ++ } ++ } ++ ++ /* fill DTable */ ++ { ++ U32 n; ++ for (n = 0; n < nbSymbols; n++) { ++ U32 const w = huffWeight[n]; ++ U32 const length = (1 << w) >> 1; ++ U32 u; ++ HUF_DEltX2 D; ++ D.byte = (BYTE)n; ++ D.nbBits = (BYTE)(tableLog + 1 - w); ++ for (u = rankVal[w]; u < rankVal[w] + length; u++) ++ dt[u] = D; ++ rankVal[w] += length; ++ } ++ } ++ ++ return iSize; ++} ++ ++static BYTE INIT HUF_decodeSymbolX2(BIT_DStream_t *Dstream, const HUF_DEltX2 *dt, const U32 dtLog) ++{ ++ size_t const val = BIT_lookBitsFast(Dstream, dtLog); /* note : dtLog >= 1 */ ++ BYTE const c = dt[val].byte; ++ BIT_skipBits(Dstream, dt[val].nbBits); ++ return c; ++} ++ ++#define HUF_DECODE_SYMBOLX2_0(ptr, DStreamPtr) *ptr++ = HUF_decodeSymbolX2(DStreamPtr, dt, dtLog) ++ ++#define HUF_DECODE_SYMBOLX2_1(ptr, DStreamPtr) \ ++ if (ZSTD_64bits() || (HUF_TABLELOG_MAX <= 12)) \ ++ HUF_DECODE_SYMBOLX2_0(ptr, DStreamPtr) ++ ++#define HUF_DECODE_SYMBOLX2_2(ptr, DStreamPtr) \ ++ if (ZSTD_64bits()) \ ++ HUF_DECODE_SYMBOLX2_0(ptr, DStreamPtr) ++ ++FORCE_INLINE size_t HUF_decodeStreamX2(BYTE *p, BIT_DStream_t *const bitDPtr, BYTE *const pEnd, const HUF_DEltX2 *const dt, const U32 dtLog) ++{ ++ BYTE *const pStart = p; ++ ++ /* up to 4 symbols at a time */ ++ while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) && (p <= pEnd - 4)) { ++ HUF_DECODE_SYMBOLX2_2(p, bitDPtr); ++ HUF_DECODE_SYMBOLX2_1(p, bitDPtr); ++ HUF_DECODE_SYMBOLX2_2(p, bitDPtr); ++ HUF_DECODE_SYMBOLX2_0(p, bitDPtr); ++ } ++ ++ /* closer to the end */ ++ while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) && (p < pEnd)) ++ HUF_DECODE_SYMBOLX2_0(p, bitDPtr); ++ ++ /* no more data to retrieve from bitstream, hence no need to reload */ ++ while (p < pEnd) ++ HUF_DECODE_SYMBOLX2_0(p, bitDPtr); ++ ++ return pEnd - pStart; ++} ++ ++static size_t INIT HUF_decompress1X2_usingDTable_internal(void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable) ++{ ++ BYTE *op = (BYTE *)dst; ++ BYTE *const oend = op + dstSize; ++ const void *dtPtr = DTable + 1; ++ const HUF_DEltX2 *const dt = (const HUF_DEltX2 *)dtPtr; ++ BIT_DStream_t bitD; ++ DTableDesc const dtd = HUF_getDTableDesc(DTable); ++ U32 const dtLog = dtd.tableLog; ++ ++ { ++ size_t const errorCode = BIT_initDStream(&bitD, cSrc, cSrcSize); ++ if (HUF_isError(errorCode)) ++ return errorCode; ++ } ++ ++ HUF_decodeStreamX2(op, &bitD, oend, dt, dtLog); ++ ++ /* check */ ++ if (!BIT_endOfDStream(&bitD)) ++ return ERROR(corruption_detected); ++ ++ return dstSize; ++} ++ ++size_t INIT HUF_decompress1X2_usingDTable(void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable) ++{ ++ DTableDesc dtd = HUF_getDTableDesc(DTable); ++ if (dtd.tableType != 0) ++ return ERROR(GENERIC); ++ return HUF_decompress1X2_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable); ++} ++ ++size_t INIT HUF_decompress1X2_DCtx_wksp(HUF_DTable *DCtx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace, size_t workspaceSize) ++{ ++ const BYTE *ip = (const BYTE *)cSrc; ++ ++ size_t const hSize = HUF_readDTableX2_wksp(DCtx, cSrc, cSrcSize, workspace, workspaceSize); ++ if (HUF_isError(hSize)) ++ return hSize; ++ if (hSize >= cSrcSize) ++ return ERROR(srcSize_wrong); ++ ip += hSize; ++ cSrcSize -= hSize; ++ ++ return HUF_decompress1X2_usingDTable_internal(dst, dstSize, ip, cSrcSize, DCtx); ++} ++ ++static size_t INIT HUF_decompress4X2_usingDTable_internal(void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable) ++{ ++ /* Check */ ++ if (cSrcSize < 10) ++ return ERROR(corruption_detected); /* strict minimum : jump table + 1 byte per stream */ ++ ++ { ++ const BYTE *const istart = (const BYTE *)cSrc; ++ BYTE *const ostart = (BYTE *)dst; ++ BYTE *const oend = ostart + dstSize; ++ const void *const dtPtr = DTable + 1; ++ const HUF_DEltX2 *const dt = (const HUF_DEltX2 *)dtPtr; ++ ++ /* Init */ ++ BIT_DStream_t bitD1; ++ BIT_DStream_t bitD2; ++ BIT_DStream_t bitD3; ++ BIT_DStream_t bitD4; ++ size_t const length1 = ZSTD_readLE16(istart); ++ size_t const length2 = ZSTD_readLE16(istart + 2); ++ size_t const length3 = ZSTD_readLE16(istart + 4); ++ size_t const length4 = cSrcSize - (length1 + length2 + length3 + 6); ++ const BYTE *const istart1 = istart + 6; /* jumpTable */ ++ const BYTE *const istart2 = istart1 + length1; ++ const BYTE *const istart3 = istart2 + length2; ++ const BYTE *const istart4 = istart3 + length3; ++ const size_t segmentSize = (dstSize + 3) / 4; ++ BYTE *const opStart2 = ostart + segmentSize; ++ BYTE *const opStart3 = opStart2 + segmentSize; ++ BYTE *const opStart4 = opStart3 + segmentSize; ++ BYTE *op1 = ostart; ++ BYTE *op2 = opStart2; ++ BYTE *op3 = opStart3; ++ BYTE *op4 = opStart4; ++ U32 endSignal; ++ DTableDesc const dtd = HUF_getDTableDesc(DTable); ++ U32 const dtLog = dtd.tableLog; ++ ++ if (length4 > cSrcSize) ++ return ERROR(corruption_detected); /* overflow */ ++ { ++ size_t const errorCode = BIT_initDStream(&bitD1, istart1, length1); ++ if (HUF_isError(errorCode)) ++ return errorCode; ++ } ++ { ++ size_t const errorCode = BIT_initDStream(&bitD2, istart2, length2); ++ if (HUF_isError(errorCode)) ++ return errorCode; ++ } ++ { ++ size_t const errorCode = BIT_initDStream(&bitD3, istart3, length3); ++ if (HUF_isError(errorCode)) ++ return errorCode; ++ } ++ { ++ size_t const errorCode = BIT_initDStream(&bitD4, istart4, length4); ++ if (HUF_isError(errorCode)) ++ return errorCode; ++ } ++ ++ /* 16-32 symbols per loop (4-8 symbols per stream) */ ++ endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4); ++ for (; (endSignal == BIT_DStream_unfinished) && (op4 < (oend - 7));) { ++ HUF_DECODE_SYMBOLX2_2(op1, &bitD1); ++ HUF_DECODE_SYMBOLX2_2(op2, &bitD2); ++ HUF_DECODE_SYMBOLX2_2(op3, &bitD3); ++ HUF_DECODE_SYMBOLX2_2(op4, &bitD4); ++ HUF_DECODE_SYMBOLX2_1(op1, &bitD1); ++ HUF_DECODE_SYMBOLX2_1(op2, &bitD2); ++ HUF_DECODE_SYMBOLX2_1(op3, &bitD3); ++ HUF_DECODE_SYMBOLX2_1(op4, &bitD4); ++ HUF_DECODE_SYMBOLX2_2(op1, &bitD1); ++ HUF_DECODE_SYMBOLX2_2(op2, &bitD2); ++ HUF_DECODE_SYMBOLX2_2(op3, &bitD3); ++ HUF_DECODE_SYMBOLX2_2(op4, &bitD4); ++ HUF_DECODE_SYMBOLX2_0(op1, &bitD1); ++ HUF_DECODE_SYMBOLX2_0(op2, &bitD2); ++ HUF_DECODE_SYMBOLX2_0(op3, &bitD3); ++ HUF_DECODE_SYMBOLX2_0(op4, &bitD4); ++ endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4); ++ } ++ ++ /* check corruption */ ++ if (op1 > opStart2) ++ return ERROR(corruption_detected); ++ if (op2 > opStart3) ++ return ERROR(corruption_detected); ++ if (op3 > opStart4) ++ return ERROR(corruption_detected); ++ /* note : op4 supposed already verified within main loop */ ++ ++ /* finish bitStreams one by one */ ++ HUF_decodeStreamX2(op1, &bitD1, opStart2, dt, dtLog); ++ HUF_decodeStreamX2(op2, &bitD2, opStart3, dt, dtLog); ++ HUF_decodeStreamX2(op3, &bitD3, opStart4, dt, dtLog); ++ HUF_decodeStreamX2(op4, &bitD4, oend, dt, dtLog); ++ ++ /* check */ ++ endSignal = BIT_endOfDStream(&bitD1) & BIT_endOfDStream(&bitD2) & BIT_endOfDStream(&bitD3) & BIT_endOfDStream(&bitD4); ++ if (!endSignal) ++ return ERROR(corruption_detected); ++ ++ /* decoded size */ ++ return dstSize; ++ } ++} ++ ++size_t INIT HUF_decompress4X2_usingDTable(void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable) ++{ ++ DTableDesc dtd = HUF_getDTableDesc(DTable); ++ if (dtd.tableType != 0) ++ return ERROR(GENERIC); ++ return HUF_decompress4X2_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable); ++} ++ ++size_t INIT HUF_decompress4X2_DCtx_wksp(HUF_DTable *dctx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace, size_t workspaceSize) ++{ ++ const BYTE *ip = (const BYTE *)cSrc; ++ ++ size_t const hSize = HUF_readDTableX2_wksp(dctx, cSrc, cSrcSize, workspace, workspaceSize); ++ if (HUF_isError(hSize)) ++ return hSize; ++ if (hSize >= cSrcSize) ++ return ERROR(srcSize_wrong); ++ ip += hSize; ++ cSrcSize -= hSize; ++ ++ return HUF_decompress4X2_usingDTable_internal(dst, dstSize, ip, cSrcSize, dctx); ++} ++ ++/* *************************/ ++/* double-symbols decoding */ ++/* *************************/ ++typedef struct { ++ U16 sequence; ++ BYTE nbBits; ++ BYTE length; ++} HUF_DEltX4; /* double-symbols decoding */ ++ ++typedef struct { ++ BYTE symbol; ++ BYTE weight; ++} sortedSymbol_t; ++ ++/* HUF_fillDTableX4Level2() : ++ * `rankValOrigin` must be a table of at least (HUF_TABLELOG_MAX + 1) U32 */ ++static void INIT HUF_fillDTableX4Level2(HUF_DEltX4 *DTable, U32 sizeLog, const U32 consumed, const U32 *rankValOrigin, const int minWeight, ++ const sortedSymbol_t *sortedSymbols, const U32 sortedListSize, U32 nbBitsBaseline, U16 baseSeq) ++{ ++ HUF_DEltX4 DElt; ++ U32 rankVal[HUF_TABLELOG_MAX + 1]; ++ ++ /* get pre-calculated rankVal */ ++ memcpy(rankVal, rankValOrigin, sizeof(rankVal)); ++ ++ /* fill skipped values */ ++ if (minWeight > 1) { ++ U32 i, skipSize = rankVal[minWeight]; ++ ZSTD_writeLE16(&(DElt.sequence), baseSeq); ++ DElt.nbBits = (BYTE)(consumed); ++ DElt.length = 1; ++ for (i = 0; i < skipSize; i++) ++ DTable[i] = DElt; ++ } ++ ++ /* fill DTable */ ++ { ++ U32 s; ++ for (s = 0; s < sortedListSize; s++) { /* note : sortedSymbols already skipped */ ++ const U32 symbol = sortedSymbols[s].symbol; ++ const U32 weight = sortedSymbols[s].weight; ++ const U32 nbBits = nbBitsBaseline - weight; ++ const U32 length = 1 << (sizeLog - nbBits); ++ const U32 start = rankVal[weight]; ++ U32 i = start; ++ const U32 end = start + length; ++ ++ ZSTD_writeLE16(&(DElt.sequence), (U16)(baseSeq + (symbol << 8))); ++ DElt.nbBits = (BYTE)(nbBits + consumed); ++ DElt.length = 2; ++ do { ++ DTable[i++] = DElt; ++ } while (i < end); /* since length >= 1 */ ++ ++ rankVal[weight] += length; ++ } ++ } ++} ++ ++typedef U32 rankVal_t[HUF_TABLELOG_MAX][HUF_TABLELOG_MAX + 1]; ++typedef U32 rankValCol_t[HUF_TABLELOG_MAX + 1]; ++ ++static void INIT HUF_fillDTableX4(HUF_DEltX4 *DTable, const U32 targetLog, const sortedSymbol_t *sortedList, const U32 sortedListSize, const U32 *rankStart, ++ rankVal_t rankValOrigin, const U32 maxWeight, const U32 nbBitsBaseline) ++{ ++ U32 rankVal[HUF_TABLELOG_MAX + 1]; ++ const int scaleLog = nbBitsBaseline - targetLog; /* note : targetLog >= srcLog, hence scaleLog <= 1 */ ++ const U32 minBits = nbBitsBaseline - maxWeight; ++ U32 s; ++ ++ memcpy(rankVal, rankValOrigin, sizeof(rankVal)); ++ ++ /* fill DTable */ ++ for (s = 0; s < sortedListSize; s++) { ++ const U16 symbol = sortedList[s].symbol; ++ const U32 weight = sortedList[s].weight; ++ const U32 nbBits = nbBitsBaseline - weight; ++ const U32 start = rankVal[weight]; ++ const U32 length = 1 << (targetLog - nbBits); ++ ++ if (targetLog - nbBits >= minBits) { /* enough room for a second symbol */ ++ U32 sortedRank; ++ int minWeight = nbBits + scaleLog; ++ if (minWeight < 1) ++ minWeight = 1; ++ sortedRank = rankStart[minWeight]; ++ HUF_fillDTableX4Level2(DTable + start, targetLog - nbBits, nbBits, rankValOrigin[nbBits], minWeight, sortedList + sortedRank, ++ sortedListSize - sortedRank, nbBitsBaseline, symbol); ++ } else { ++ HUF_DEltX4 DElt; ++ ZSTD_writeLE16(&(DElt.sequence), symbol); ++ DElt.nbBits = (BYTE)(nbBits); ++ DElt.length = 1; ++ { ++ U32 const end = start + length; ++ U32 u; ++ for (u = start; u < end; u++) ++ DTable[u] = DElt; ++ } ++ } ++ rankVal[weight] += length; ++ } ++} ++ ++size_t INIT HUF_readDTableX4_wksp(HUF_DTable *DTable, const void *src, size_t srcSize, void *workspace, size_t workspaceSize) ++{ ++ U32 tableLog, maxW, sizeOfSort, nbSymbols; ++ DTableDesc dtd = HUF_getDTableDesc(DTable); ++ U32 const maxTableLog = dtd.maxTableLog; ++ size_t iSize; ++ void *dtPtr = DTable + 1; /* force compiler to avoid strict-aliasing */ ++ HUF_DEltX4 *const dt = (HUF_DEltX4 *)dtPtr; ++ U32 *rankStart; ++ ++ rankValCol_t *rankVal; ++ U32 *rankStats; ++ U32 *rankStart0; ++ sortedSymbol_t *sortedSymbol; ++ BYTE *weightList; ++ size_t spaceUsed32 = 0; ++ ++ HUF_STATIC_ASSERT((sizeof(rankValCol_t) & 3) == 0); ++ ++ rankVal = (rankValCol_t *)((U32 *)workspace + spaceUsed32); ++ spaceUsed32 += (sizeof(rankValCol_t) * HUF_TABLELOG_MAX) >> 2; ++ rankStats = (U32 *)workspace + spaceUsed32; ++ spaceUsed32 += HUF_TABLELOG_MAX + 1; ++ rankStart0 = (U32 *)workspace + spaceUsed32; ++ spaceUsed32 += HUF_TABLELOG_MAX + 2; ++ sortedSymbol = (sortedSymbol_t *)((U32 *)workspace + spaceUsed32); ++ spaceUsed32 += ALIGN(sizeof(sortedSymbol_t) * (HUF_SYMBOLVALUE_MAX + 1), sizeof(U32)) >> 2; ++ weightList = (BYTE *)((U32 *)workspace + spaceUsed32); ++ spaceUsed32 += ALIGN(HUF_SYMBOLVALUE_MAX + 1, sizeof(U32)) >> 2; ++ ++ if ((spaceUsed32 << 2) > workspaceSize) ++ return ERROR(tableLog_tooLarge); ++ workspace = (U32 *)workspace + spaceUsed32; ++ workspaceSize -= (spaceUsed32 << 2); ++ ++ rankStart = rankStart0 + 1; ++ memset(rankStats, 0, sizeof(U32) * (2 * HUF_TABLELOG_MAX + 2 + 1)); ++ ++ HUF_STATIC_ASSERT(sizeof(HUF_DEltX4) == sizeof(HUF_DTable)); /* if compiler fails here, assertion is wrong */ ++ if (maxTableLog > HUF_TABLELOG_MAX) ++ return ERROR(tableLog_tooLarge); ++ /* memset(weightList, 0, sizeof(weightList)); */ /* is not necessary, even though some analyzer complain ... */ ++ ++ iSize = HUF_readStats_wksp(weightList, HUF_SYMBOLVALUE_MAX + 1, rankStats, &nbSymbols, &tableLog, src, srcSize, workspace, workspaceSize); ++ if (HUF_isError(iSize)) ++ return iSize; ++ ++ /* check result */ ++ if (tableLog > maxTableLog) ++ return ERROR(tableLog_tooLarge); /* DTable can't fit code depth */ ++ ++ /* find maxWeight */ ++ for (maxW = tableLog; rankStats[maxW] == 0; maxW--) { ++ } /* necessarily finds a solution before 0 */ ++ ++ /* Get start index of each weight */ ++ { ++ U32 w, nextRankStart = 0; ++ for (w = 1; w < maxW + 1; w++) { ++ U32 curr = nextRankStart; ++ nextRankStart += rankStats[w]; ++ rankStart[w] = curr; ++ } ++ rankStart[0] = nextRankStart; /* put all 0w symbols at the end of sorted list*/ ++ sizeOfSort = nextRankStart; ++ } ++ ++ /* sort symbols by weight */ ++ { ++ U32 s; ++ for (s = 0; s < nbSymbols; s++) { ++ U32 const w = weightList[s]; ++ U32 const r = rankStart[w]++; ++ sortedSymbol[r].symbol = (BYTE)s; ++ sortedSymbol[r].weight = (BYTE)w; ++ } ++ rankStart[0] = 0; /* forget 0w symbols; this is beginning of weight(1) */ ++ } ++ ++ /* Build rankVal */ ++ { ++ U32 *const rankVal0 = rankVal[0]; ++ { ++ int const rescale = (maxTableLog - tableLog) - 1; /* tableLog <= maxTableLog */ ++ U32 nextRankVal = 0; ++ U32 w; ++ for (w = 1; w < maxW + 1; w++) { ++ U32 curr = nextRankVal; ++ nextRankVal += rankStats[w] << (w + rescale); ++ rankVal0[w] = curr; ++ } ++ } ++ { ++ U32 const minBits = tableLog + 1 - maxW; ++ U32 consumed; ++ for (consumed = minBits; consumed < maxTableLog - minBits + 1; consumed++) { ++ U32 *const rankValPtr = rankVal[consumed]; ++ U32 w; ++ for (w = 1; w < maxW + 1; w++) { ++ rankValPtr[w] = rankVal0[w] >> consumed; ++ } ++ } ++ } ++ } ++ ++ HUF_fillDTableX4(dt, maxTableLog, sortedSymbol, sizeOfSort, rankStart0, rankVal, maxW, tableLog + 1); ++ ++ dtd.tableLog = (BYTE)maxTableLog; ++ dtd.tableType = 1; ++ memcpy(DTable, &dtd, sizeof(dtd)); ++ return iSize; ++} ++ ++static U32 INIT HUF_decodeSymbolX4(void *op, BIT_DStream_t *DStream, const HUF_DEltX4 *dt, const U32 dtLog) ++{ ++ size_t const val = BIT_lookBitsFast(DStream, dtLog); /* note : dtLog >= 1 */ ++ memcpy(op, dt + val, 2); ++ BIT_skipBits(DStream, dt[val].nbBits); ++ return dt[val].length; ++} ++ ++static U32 INIT HUF_decodeLastSymbolX4(void *op, BIT_DStream_t *DStream, const HUF_DEltX4 *dt, const U32 dtLog) ++{ ++ size_t const val = BIT_lookBitsFast(DStream, dtLog); /* note : dtLog >= 1 */ ++ memcpy(op, dt + val, 1); ++ if (dt[val].length == 1) ++ BIT_skipBits(DStream, dt[val].nbBits); ++ else { ++ if (DStream->bitsConsumed < (sizeof(DStream->bitContainer) * 8)) { ++ BIT_skipBits(DStream, dt[val].nbBits); ++ if (DStream->bitsConsumed > (sizeof(DStream->bitContainer) * 8)) ++ /* ugly hack; works only because it's the last symbol. Note : can't easily extract nbBits from just this symbol */ ++ DStream->bitsConsumed = (sizeof(DStream->bitContainer) * 8); ++ } ++ } ++ return 1; ++} ++ ++#define HUF_DECODE_SYMBOLX4_0(ptr, DStreamPtr) ptr += HUF_decodeSymbolX4(ptr, DStreamPtr, dt, dtLog) ++ ++#define HUF_DECODE_SYMBOLX4_1(ptr, DStreamPtr) \ ++ if (ZSTD_64bits() || (HUF_TABLELOG_MAX <= 12)) \ ++ ptr += HUF_decodeSymbolX4(ptr, DStreamPtr, dt, dtLog) ++ ++#define HUF_DECODE_SYMBOLX4_2(ptr, DStreamPtr) \ ++ if (ZSTD_64bits()) \ ++ ptr += HUF_decodeSymbolX4(ptr, DStreamPtr, dt, dtLog) ++ ++FORCE_INLINE size_t HUF_decodeStreamX4(BYTE *p, BIT_DStream_t *bitDPtr, BYTE *const pEnd, const HUF_DEltX4 *const dt, const U32 dtLog) ++{ ++ BYTE *const pStart = p; ++ ++ /* up to 8 symbols at a time */ ++ while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) & (p < pEnd - (sizeof(bitDPtr->bitContainer) - 1))) { ++ HUF_DECODE_SYMBOLX4_2(p, bitDPtr); ++ HUF_DECODE_SYMBOLX4_1(p, bitDPtr); ++ HUF_DECODE_SYMBOLX4_2(p, bitDPtr); ++ HUF_DECODE_SYMBOLX4_0(p, bitDPtr); ++ } ++ ++ /* closer to end : up to 2 symbols at a time */ ++ while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) & (p <= pEnd - 2)) ++ HUF_DECODE_SYMBOLX4_0(p, bitDPtr); ++ ++ while (p <= pEnd - 2) ++ HUF_DECODE_SYMBOLX4_0(p, bitDPtr); /* no need to reload : reached the end of DStream */ ++ ++ if (p < pEnd) ++ p += HUF_decodeLastSymbolX4(p, bitDPtr, dt, dtLog); ++ ++ return p - pStart; ++} ++ ++static size_t INIT HUF_decompress1X4_usingDTable_internal(void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable) ++{ ++ BIT_DStream_t bitD; ++ ++ /* Init */ ++ { ++ size_t const errorCode = BIT_initDStream(&bitD, cSrc, cSrcSize); ++ if (HUF_isError(errorCode)) ++ return errorCode; ++ } ++ ++ /* decode */ ++ { ++ BYTE *const ostart = (BYTE *)dst; ++ BYTE *const oend = ostart + dstSize; ++ const void *const dtPtr = DTable + 1; /* force compiler to not use strict-aliasing */ ++ const HUF_DEltX4 *const dt = (const HUF_DEltX4 *)dtPtr; ++ DTableDesc const dtd = HUF_getDTableDesc(DTable); ++ HUF_decodeStreamX4(ostart, &bitD, oend, dt, dtd.tableLog); ++ } ++ ++ /* check */ ++ if (!BIT_endOfDStream(&bitD)) ++ return ERROR(corruption_detected); ++ ++ /* decoded size */ ++ return dstSize; ++} ++ ++size_t INIT HUF_decompress1X4_usingDTable(void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable) ++{ ++ DTableDesc dtd = HUF_getDTableDesc(DTable); ++ if (dtd.tableType != 1) ++ return ERROR(GENERIC); ++ return HUF_decompress1X4_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable); ++} ++ ++size_t INIT HUF_decompress1X4_DCtx_wksp(HUF_DTable *DCtx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace, size_t workspaceSize) ++{ ++ const BYTE *ip = (const BYTE *)cSrc; ++ ++ size_t const hSize = HUF_readDTableX4_wksp(DCtx, cSrc, cSrcSize, workspace, workspaceSize); ++ if (HUF_isError(hSize)) ++ return hSize; ++ if (hSize >= cSrcSize) ++ return ERROR(srcSize_wrong); ++ ip += hSize; ++ cSrcSize -= hSize; ++ ++ return HUF_decompress1X4_usingDTable_internal(dst, dstSize, ip, cSrcSize, DCtx); ++} ++ ++static size_t INIT HUF_decompress4X4_usingDTable_internal(void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable) ++{ ++ if (cSrcSize < 10) ++ return ERROR(corruption_detected); /* strict minimum : jump table + 1 byte per stream */ ++ ++ { ++ const BYTE *const istart = (const BYTE *)cSrc; ++ BYTE *const ostart = (BYTE *)dst; ++ BYTE *const oend = ostart + dstSize; ++ const void *const dtPtr = DTable + 1; ++ const HUF_DEltX4 *const dt = (const HUF_DEltX4 *)dtPtr; ++ ++ /* Init */ ++ BIT_DStream_t bitD1; ++ BIT_DStream_t bitD2; ++ BIT_DStream_t bitD3; ++ BIT_DStream_t bitD4; ++ size_t const length1 = ZSTD_readLE16(istart); ++ size_t const length2 = ZSTD_readLE16(istart + 2); ++ size_t const length3 = ZSTD_readLE16(istart + 4); ++ size_t const length4 = cSrcSize - (length1 + length2 + length3 + 6); ++ const BYTE *const istart1 = istart + 6; /* jumpTable */ ++ const BYTE *const istart2 = istart1 + length1; ++ const BYTE *const istart3 = istart2 + length2; ++ const BYTE *const istart4 = istart3 + length3; ++ size_t const segmentSize = (dstSize + 3) / 4; ++ BYTE *const opStart2 = ostart + segmentSize; ++ BYTE *const opStart3 = opStart2 + segmentSize; ++ BYTE *const opStart4 = opStart3 + segmentSize; ++ BYTE *op1 = ostart; ++ BYTE *op2 = opStart2; ++ BYTE *op3 = opStart3; ++ BYTE *op4 = opStart4; ++ U32 endSignal; ++ DTableDesc const dtd = HUF_getDTableDesc(DTable); ++ U32 const dtLog = dtd.tableLog; ++ ++ if (length4 > cSrcSize) ++ return ERROR(corruption_detected); /* overflow */ ++ { ++ size_t const errorCode = BIT_initDStream(&bitD1, istart1, length1); ++ if (HUF_isError(errorCode)) ++ return errorCode; ++ } ++ { ++ size_t const errorCode = BIT_initDStream(&bitD2, istart2, length2); ++ if (HUF_isError(errorCode)) ++ return errorCode; ++ } ++ { ++ size_t const errorCode = BIT_initDStream(&bitD3, istart3, length3); ++ if (HUF_isError(errorCode)) ++ return errorCode; ++ } ++ { ++ size_t const errorCode = BIT_initDStream(&bitD4, istart4, length4); ++ if (HUF_isError(errorCode)) ++ return errorCode; ++ } ++ ++ /* 16-32 symbols per loop (4-8 symbols per stream) */ ++ endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4); ++ for (; (endSignal == BIT_DStream_unfinished) & (op4 < (oend - (sizeof(bitD4.bitContainer) - 1)));) { ++ HUF_DECODE_SYMBOLX4_2(op1, &bitD1); ++ HUF_DECODE_SYMBOLX4_2(op2, &bitD2); ++ HUF_DECODE_SYMBOLX4_2(op3, &bitD3); ++ HUF_DECODE_SYMBOLX4_2(op4, &bitD4); ++ HUF_DECODE_SYMBOLX4_1(op1, &bitD1); ++ HUF_DECODE_SYMBOLX4_1(op2, &bitD2); ++ HUF_DECODE_SYMBOLX4_1(op3, &bitD3); ++ HUF_DECODE_SYMBOLX4_1(op4, &bitD4); ++ HUF_DECODE_SYMBOLX4_2(op1, &bitD1); ++ HUF_DECODE_SYMBOLX4_2(op2, &bitD2); ++ HUF_DECODE_SYMBOLX4_2(op3, &bitD3); ++ HUF_DECODE_SYMBOLX4_2(op4, &bitD4); ++ HUF_DECODE_SYMBOLX4_0(op1, &bitD1); ++ HUF_DECODE_SYMBOLX4_0(op2, &bitD2); ++ HUF_DECODE_SYMBOLX4_0(op3, &bitD3); ++ HUF_DECODE_SYMBOLX4_0(op4, &bitD4); ++ ++ endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4); ++ } ++ ++ /* check corruption */ ++ if (op1 > opStart2) ++ return ERROR(corruption_detected); ++ if (op2 > opStart3) ++ return ERROR(corruption_detected); ++ if (op3 > opStart4) ++ return ERROR(corruption_detected); ++ /* note : op4 already verified within main loop */ ++ ++ /* finish bitStreams one by one */ ++ HUF_decodeStreamX4(op1, &bitD1, opStart2, dt, dtLog); ++ HUF_decodeStreamX4(op2, &bitD2, opStart3, dt, dtLog); ++ HUF_decodeStreamX4(op3, &bitD3, opStart4, dt, dtLog); ++ HUF_decodeStreamX4(op4, &bitD4, oend, dt, dtLog); ++ ++ /* check */ ++ { ++ U32 const endCheck = BIT_endOfDStream(&bitD1) & BIT_endOfDStream(&bitD2) & BIT_endOfDStream(&bitD3) & BIT_endOfDStream(&bitD4); ++ if (!endCheck) ++ return ERROR(corruption_detected); ++ } ++ ++ /* decoded size */ ++ return dstSize; ++ } ++} ++ ++size_t INIT HUF_decompress4X4_usingDTable(void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable) ++{ ++ DTableDesc dtd = HUF_getDTableDesc(DTable); ++ if (dtd.tableType != 1) ++ return ERROR(GENERIC); ++ return HUF_decompress4X4_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable); ++} ++ ++size_t INIT HUF_decompress4X4_DCtx_wksp(HUF_DTable *dctx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace, size_t workspaceSize) ++{ ++ const BYTE *ip = (const BYTE *)cSrc; ++ ++ size_t hSize = HUF_readDTableX4_wksp(dctx, cSrc, cSrcSize, workspace, workspaceSize); ++ if (HUF_isError(hSize)) ++ return hSize; ++ if (hSize >= cSrcSize) ++ return ERROR(srcSize_wrong); ++ ip += hSize; ++ cSrcSize -= hSize; ++ ++ return HUF_decompress4X4_usingDTable_internal(dst, dstSize, ip, cSrcSize, dctx); ++} ++ ++/* ********************************/ ++/* Generic decompression selector */ ++/* ********************************/ ++ ++size_t INIT HUF_decompress1X_usingDTable(void *dst, size_t maxDstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable) ++{ ++ DTableDesc const dtd = HUF_getDTableDesc(DTable); ++ return dtd.tableType ? HUF_decompress1X4_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable) ++ : HUF_decompress1X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable); ++} ++ ++size_t INIT HUF_decompress4X_usingDTable(void *dst, size_t maxDstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable) ++{ ++ DTableDesc const dtd = HUF_getDTableDesc(DTable); ++ return dtd.tableType ? HUF_decompress4X4_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable) ++ : HUF_decompress4X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable); ++} ++ ++typedef struct { ++ U32 tableTime; ++ U32 decode256Time; ++} algo_time_t; ++static const algo_time_t algoTime[16 /* Quantization */][3 /* single, double, quad */] = { ++ /* single, double, quad */ ++ {{0, 0}, {1, 1}, {2, 2}}, /* Q==0 : impossible */ ++ {{0, 0}, {1, 1}, {2, 2}}, /* Q==1 : impossible */ ++ {{38, 130}, {1313, 74}, {2151, 38}}, /* Q == 2 : 12-18% */ ++ {{448, 128}, {1353, 74}, {2238, 41}}, /* Q == 3 : 18-25% */ ++ {{556, 128}, {1353, 74}, {2238, 47}}, /* Q == 4 : 25-32% */ ++ {{714, 128}, {1418, 74}, {2436, 53}}, /* Q == 5 : 32-38% */ ++ {{883, 128}, {1437, 74}, {2464, 61}}, /* Q == 6 : 38-44% */ ++ {{897, 128}, {1515, 75}, {2622, 68}}, /* Q == 7 : 44-50% */ ++ {{926, 128}, {1613, 75}, {2730, 75}}, /* Q == 8 : 50-56% */ ++ {{947, 128}, {1729, 77}, {3359, 77}}, /* Q == 9 : 56-62% */ ++ {{1107, 128}, {2083, 81}, {4006, 84}}, /* Q ==10 : 62-69% */ ++ {{1177, 128}, {2379, 87}, {4785, 88}}, /* Q ==11 : 69-75% */ ++ {{1242, 128}, {2415, 93}, {5155, 84}}, /* Q ==12 : 75-81% */ ++ {{1349, 128}, {2644, 106}, {5260, 106}}, /* Q ==13 : 81-87% */ ++ {{1455, 128}, {2422, 124}, {4174, 124}}, /* Q ==14 : 87-93% */ ++ {{722, 128}, {1891, 145}, {1936, 146}}, /* Q ==15 : 93-99% */ ++}; ++ ++/** HUF_selectDecoder() : ++* Tells which decoder is likely to decode faster, ++* based on a set of pre-determined metrics. ++* @return : 0==HUF_decompress4X2, 1==HUF_decompress4X4 . ++* Assumption : 0 < cSrcSize < dstSize <= 128 KB */ ++U32 INIT HUF_selectDecoder(size_t dstSize, size_t cSrcSize) ++{ ++ /* decoder timing evaluation */ ++ U32 const Q = (U32)(cSrcSize * 16 / dstSize); /* Q < 16 since dstSize > cSrcSize */ ++ U32 const D256 = (U32)(dstSize >> 8); ++ U32 const DTime0 = algoTime[Q][0].tableTime + (algoTime[Q][0].decode256Time * D256); ++ U32 DTime1 = algoTime[Q][1].tableTime + (algoTime[Q][1].decode256Time * D256); ++ DTime1 += DTime1 >> 3; /* advantage to algorithm using less memory, for cache eviction */ ++ ++ return DTime1 < DTime0; ++} ++ ++typedef size_t (*decompressionAlgo)(void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize); ++ ++size_t INIT HUF_decompress4X_DCtx_wksp(HUF_DTable *dctx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace, size_t workspaceSize) ++{ ++ /* validation checks */ ++ if (dstSize == 0) ++ return ERROR(dstSize_tooSmall); ++ if (cSrcSize > dstSize) ++ return ERROR(corruption_detected); /* invalid */ ++ if (cSrcSize == dstSize) { ++ memcpy(dst, cSrc, dstSize); ++ return dstSize; ++ } /* not compressed */ ++ if (cSrcSize == 1) { ++ memset(dst, *(const BYTE *)cSrc, dstSize); ++ return dstSize; ++ } /* RLE */ ++ ++ { ++ U32 const algoNb = HUF_selectDecoder(dstSize, cSrcSize); ++ return algoNb ? HUF_decompress4X4_DCtx_wksp(dctx, dst, dstSize, cSrc, cSrcSize, workspace, workspaceSize) ++ : HUF_decompress4X2_DCtx_wksp(dctx, dst, dstSize, cSrc, cSrcSize, workspace, workspaceSize); ++ } ++} ++ ++size_t INIT HUF_decompress4X_hufOnly_wksp(HUF_DTable *dctx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace, size_t workspaceSize) ++{ ++ /* validation checks */ ++ if (dstSize == 0) ++ return ERROR(dstSize_tooSmall); ++ if ((cSrcSize >= dstSize) || (cSrcSize <= 1)) ++ return ERROR(corruption_detected); /* invalid */ ++ ++ { ++ U32 const algoNb = HUF_selectDecoder(dstSize, cSrcSize); ++ return algoNb ? HUF_decompress4X4_DCtx_wksp(dctx, dst, dstSize, cSrc, cSrcSize, workspace, workspaceSize) ++ : HUF_decompress4X2_DCtx_wksp(dctx, dst, dstSize, cSrc, cSrcSize, workspace, workspaceSize); ++ } ++} ++ ++size_t INIT HUF_decompress1X_DCtx_wksp(HUF_DTable *dctx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace, size_t workspaceSize) ++{ ++ /* validation checks */ ++ if (dstSize == 0) ++ return ERROR(dstSize_tooSmall); ++ if (cSrcSize > dstSize) ++ return ERROR(corruption_detected); /* invalid */ ++ if (cSrcSize == dstSize) { ++ memcpy(dst, cSrc, dstSize); ++ return dstSize; ++ } /* not compressed */ ++ if (cSrcSize == 1) { ++ memset(dst, *(const BYTE *)cSrc, dstSize); ++ return dstSize; ++ } /* RLE */ ++ ++ { ++ U32 const algoNb = HUF_selectDecoder(dstSize, cSrcSize); ++ return algoNb ? HUF_decompress1X4_DCtx_wksp(dctx, dst, dstSize, cSrc, cSrcSize, workspace, workspaceSize) ++ : HUF_decompress1X2_DCtx_wksp(dctx, dst, dstSize, cSrc, cSrcSize, workspace, workspaceSize); ++ } ++} +diff --git a/xen/common/zstd/mem.h b/xen/common/zstd/mem.h +new file mode 100644 +index 0000000000..d2fa444687 +--- /dev/null ++++ b/xen/common/zstd/mem.h +@@ -0,0 +1,151 @@ ++/** ++ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc. ++ * All rights reserved. ++ * ++ * This source code is licensed under the BSD-style license found in the ++ * LICENSE file in the root directory of https://github.com/facebook/zstd. ++ * An additional grant of patent rights can be found in the PATENTS file in the ++ * same directory. ++ * ++ * This program is free software; you can redistribute it and/or modify it under ++ * the terms of the GNU General Public License version 2 as published by the ++ * Free Software Foundation. This program is dual-licensed; you may select ++ * either version 2 of the GNU General Public License ("GPL") or BSD license ++ * ("BSD"). ++ */ ++ ++#ifndef MEM_H_MODULE ++#define MEM_H_MODULE ++ ++/*-**************************************** ++* Dependencies ++******************************************/ ++#include /* memcpy */ ++#include /* size_t, ptrdiff_t */ ++#include "private.h" ++ ++/*-**************************************** ++* Compiler specifics ++******************************************/ ++#define ZSTD_STATIC static inline ++ ++/*-************************************************************** ++* Basic Types ++*****************************************************************/ ++typedef uint8_t BYTE; ++typedef uint16_t U16; ++typedef int16_t S16; ++typedef uint32_t U32; ++typedef int32_t S32; ++typedef uint64_t U64; ++typedef int64_t S64; ++typedef ptrdiff_t iPtrDiff; ++typedef uintptr_t uPtrDiff; ++ ++/*-************************************************************** ++* Memory I/O ++*****************************************************************/ ++ZSTD_STATIC unsigned ZSTD_32bits(void) { return sizeof(size_t) == 4; } ++ZSTD_STATIC unsigned ZSTD_64bits(void) { return sizeof(size_t) == 8; } ++ ++#if defined(__LITTLE_ENDIAN) ++#define ZSTD_LITTLE_ENDIAN 1 ++#else ++#define ZSTD_LITTLE_ENDIAN 0 ++#endif ++ ++ZSTD_STATIC unsigned ZSTD_isLittleEndian(void) { return ZSTD_LITTLE_ENDIAN; } ++ ++ZSTD_STATIC U16 ZSTD_read16(const void *memPtr) { return get_unaligned((const U16 *)memPtr); } ++ ++ZSTD_STATIC U32 ZSTD_read32(const void *memPtr) { return get_unaligned((const U32 *)memPtr); } ++ ++ZSTD_STATIC U64 ZSTD_read64(const void *memPtr) { return get_unaligned((const U64 *)memPtr); } ++ ++ZSTD_STATIC size_t ZSTD_readST(const void *memPtr) { return get_unaligned((const size_t *)memPtr); } ++ ++ZSTD_STATIC void ZSTD_write16(void *memPtr, U16 value) { put_unaligned(value, (U16 *)memPtr); } ++ ++ZSTD_STATIC void ZSTD_write32(void *memPtr, U32 value) { put_unaligned(value, (U32 *)memPtr); } ++ ++ZSTD_STATIC void ZSTD_write64(void *memPtr, U64 value) { put_unaligned(value, (U64 *)memPtr); } ++ ++/*=== Little endian r/w ===*/ ++ ++ZSTD_STATIC U16 ZSTD_readLE16(const void *memPtr) { return get_unaligned_le16(memPtr); } ++ ++ZSTD_STATIC void ZSTD_writeLE16(void *memPtr, U16 val) { put_unaligned_le16(val, memPtr); } ++ ++ZSTD_STATIC U32 ZSTD_readLE24(const void *memPtr) { return ZSTD_readLE16(memPtr) + (((const BYTE *)memPtr)[2] << 16); } ++ ++ZSTD_STATIC void ZSTD_writeLE24(void *memPtr, U32 val) ++{ ++ ZSTD_writeLE16(memPtr, (U16)val); ++ ((BYTE *)memPtr)[2] = (BYTE)(val >> 16); ++} ++ ++ZSTD_STATIC U32 ZSTD_readLE32(const void *memPtr) { return get_unaligned_le32(memPtr); } ++ ++ZSTD_STATIC void ZSTD_writeLE32(void *memPtr, U32 val32) { put_unaligned_le32(val32, memPtr); } ++ ++ZSTD_STATIC U64 ZSTD_readLE64(const void *memPtr) { return get_unaligned_le64(memPtr); } ++ ++ZSTD_STATIC void ZSTD_writeLE64(void *memPtr, U64 val64) { put_unaligned_le64(val64, memPtr); } ++ ++ZSTD_STATIC size_t ZSTD_readLEST(const void *memPtr) ++{ ++ if (ZSTD_32bits()) ++ return (size_t)ZSTD_readLE32(memPtr); ++ else ++ return (size_t)ZSTD_readLE64(memPtr); ++} ++ ++ZSTD_STATIC void ZSTD_writeLEST(void *memPtr, size_t val) ++{ ++ if (ZSTD_32bits()) ++ ZSTD_writeLE32(memPtr, (U32)val); ++ else ++ ZSTD_writeLE64(memPtr, (U64)val); ++} ++ ++/*=== Big endian r/w ===*/ ++ ++ZSTD_STATIC U32 ZSTD_readBE32(const void *memPtr) { return get_unaligned_be32(memPtr); } ++ ++ZSTD_STATIC void ZSTD_writeBE32(void *memPtr, U32 val32) { put_unaligned_be32(val32, memPtr); } ++ ++ZSTD_STATIC U64 ZSTD_readBE64(const void *memPtr) { return get_unaligned_be64(memPtr); } ++ ++ZSTD_STATIC void ZSTD_writeBE64(void *memPtr, U64 val64) { put_unaligned_be64(val64, memPtr); } ++ ++ZSTD_STATIC size_t ZSTD_readBEST(const void *memPtr) ++{ ++ if (ZSTD_32bits()) ++ return (size_t)ZSTD_readBE32(memPtr); ++ else ++ return (size_t)ZSTD_readBE64(memPtr); ++} ++ ++ZSTD_STATIC void ZSTD_writeBEST(void *memPtr, size_t val) ++{ ++ if (ZSTD_32bits()) ++ ZSTD_writeBE32(memPtr, (U32)val); ++ else ++ ZSTD_writeBE64(memPtr, (U64)val); ++} ++ ++/* function safe only for comparisons */ ++ZSTD_STATIC U32 ZSTD_readMINMATCH(const void *memPtr, U32 length) ++{ ++ switch (length) { ++ default: ++ case 4: return ZSTD_read32(memPtr); ++ case 3: ++ if (ZSTD_isLittleEndian()) ++ return ZSTD_read32(memPtr) << 8; ++ else ++ return ZSTD_read32(memPtr) >> 8; ++ } ++} ++ ++#endif /* MEM_H_MODULE */ +diff --git a/xen/common/zstd/private.h b/xen/common/zstd/private.h +new file mode 100644 +index 0000000000..fac4d3c095 +--- /dev/null ++++ b/xen/common/zstd/private.h +@@ -0,0 +1,105 @@ ++#ifndef ZSTD_PRIVATE_H ++#define ZSTD_PRIVATE_H ++ ++#include ++#include ++#include ++ ++typedef ssize_t __attribute__((__mode__(__pointer__))) ptrdiff_t; ++ ++/* from kernel include/linux/unaligned/access_ok.h */ ++ ++static always_inline u16 get_unaligned_le16(const void *p) ++{ ++ return le16_to_cpup((__le16 *)p); ++} ++ ++static always_inline u32 get_unaligned_le32(const void *p) ++{ ++ return le32_to_cpup((__le32 *)p); ++} ++ ++static always_inline u64 get_unaligned_le64(const void *p) ++{ ++ return le64_to_cpup((__le64 *)p); ++} ++ ++static always_inline u32 get_unaligned_be32(const void *p) ++{ ++ return be32_to_cpup((__be32 *)p); ++} ++ ++static always_inline u64 get_unaligned_be64(const void *p) ++{ ++ return be64_to_cpup((__be64 *)p); ++} ++ ++static always_inline void put_unaligned_le16(u16 val, void *p) ++{ ++ *((__le16 *)p) = cpu_to_le16(val); ++} ++ ++static always_inline void put_unaligned_le32(u32 val, void *p) ++{ ++ *((__le32 *)p) = cpu_to_le32(val); ++} ++ ++static always_inline void put_unaligned_le64(u64 val, void *p) ++{ ++ *((__le64 *)p) = cpu_to_le64(val); ++} ++ ++static always_inline void put_unaligned_be32(u32 val, void *p) ++{ ++ *((__be32 *)p) = cpu_to_be32(val); ++} ++ ++static always_inline void put_unaligned_be64(u64 val, void *p) ++{ ++ *((__be64 *)p) = cpu_to_be64(val); ++} ++ ++ ++/* from kernel include/asm-generic/unaligned.h with linux/unaligned/generic.h ++ assuming little endian */ ++ ++extern void __bad_unaligned_access_size(void); ++ ++#define get_unaligned(ptr) ((__force typeof(*(ptr)))({ \ ++ __builtin_choose_expr(sizeof(*(ptr)) == 1, *(ptr), \ ++ __builtin_choose_expr(sizeof(*(ptr)) == 2, get_unaligned_le16((ptr)), \ ++ __builtin_choose_expr(sizeof(*(ptr)) == 4, get_unaligned_le32((ptr)), \ ++ __builtin_choose_expr(sizeof(*(ptr)) == 8, get_unaligned_le64((ptr)), \ ++ __bad_unaligned_access_size())))); \ ++ })) ++ ++#define put_unaligned(val, ptr) ({ \ ++ void *__gu_p = (ptr); \ ++ switch (sizeof(*(ptr))) { \ ++ case 1: \ ++ *(u8 *)__gu_p = (__force u8)(val); \ ++ break; \ ++ case 2: \ ++ put_unaligned_le16((__force u16)(val), __gu_p); \ ++ break; \ ++ case 4: \ ++ put_unaligned_le32((__force u32)(val), __gu_p); \ ++ break; \ ++ case 8: \ ++ put_unaligned_le64((__force u64)(val), __gu_p); \ ++ break; \ ++ default: \ ++ __bad_unaligned_access_size(); \ ++ break; \ ++ } \ ++ (void)0; }) ++ ++ ++/* from kernel linux/kernel.h and uapi/linux/kernel.h */ ++ ++#define __ALIGN_KERNEL(x, a) __ALIGN_KERNEL_MASK(x, (typeof(x))(a) - 1) ++#define __ALIGN_KERNEL_MASK(x, mask) (((x) + (mask)) & ~(mask)) ++#define ALIGN(x, a) __ALIGN_KERNEL((x), (a)) ++#define PTR_ALIGN(p, a) ((typeof(p))ALIGN((unsigned long)(p), (a))) ++ ++#endif /* ZSTD_PRIVATE_H */ +diff --git a/xen/common/zstd/zstd_common.c b/xen/common/zstd/zstd_common.c +new file mode 100644 +index 0000000000..1b13903538 +--- /dev/null ++++ b/xen/common/zstd/zstd_common.c +@@ -0,0 +1,74 @@ ++/** ++ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc. ++ * All rights reserved. ++ * ++ * This source code is licensed under the BSD-style license found in the ++ * LICENSE file in the root directory of https://github.com/facebook/zstd. ++ * An additional grant of patent rights can be found in the PATENTS file in the ++ * same directory. ++ * ++ * This program is free software; you can redistribute it and/or modify it under ++ * the terms of the GNU General Public License version 2 as published by the ++ * Free Software Foundation. This program is dual-licensed; you may select ++ * either version 2 of the GNU General Public License ("GPL") or BSD license ++ * ("BSD"). ++ */ ++ ++/*-************************************* ++* Dependencies ++***************************************/ ++#include "error_private.h" ++#include "zstd_internal.h" /* declaration of ZSTD_isError, ZSTD_getErrorName, ZSTD_getErrorCode, ZSTD_getErrorString, ZSTD_versionNumber */ ++ ++/*=************************************************************** ++* Custom allocator ++****************************************************************/ ++ ++#define stack_push(stack, size) \ ++ ({ \ ++ void *const ptr = ZSTD_PTR_ALIGN((stack)->ptr); \ ++ (stack)->ptr = (char *)ptr + (size); \ ++ (stack)->ptr <= (stack)->end ? ptr : NULL; \ ++ }) ++ ++ZSTD_customMem INIT ZSTD_initStack(void *workspace, size_t workspaceSize) ++{ ++ ZSTD_customMem stackMem = {ZSTD_stackAlloc, ZSTD_stackFree, workspace}; ++ ZSTD_stack *stack = (ZSTD_stack *)workspace; ++ /* Verify preconditions */ ++ if (!workspace || workspaceSize < sizeof(ZSTD_stack) || workspace != ZSTD_PTR_ALIGN(workspace)) { ++ ZSTD_customMem error = {NULL, NULL, NULL}; ++ return error; ++ } ++ /* Initialize the stack */ ++ stack->ptr = workspace; ++ stack->end = (char *)workspace + workspaceSize; ++ stack_push(stack, sizeof(ZSTD_stack)); ++ return stackMem; ++} ++ ++void INIT *ZSTD_stackAllocAll(void *opaque, size_t *size) ++{ ++ ZSTD_stack *stack = (ZSTD_stack *)opaque; ++ *size = (BYTE const *)stack->end - (BYTE *)ZSTD_PTR_ALIGN(stack->ptr); ++ return stack_push(stack, *size); ++} ++ ++void INIT *ZSTD_stackAlloc(void *opaque, size_t size) ++{ ++ ZSTD_stack *stack = (ZSTD_stack *)opaque; ++ return stack_push(stack, size); ++} ++void INIT ZSTD_stackFree(void *opaque, void *address) ++{ ++ (void)opaque; ++ (void)address; ++} ++ ++void INIT *ZSTD_malloc(size_t size, ZSTD_customMem customMem) { return customMem.customAlloc(customMem.opaque, size); } ++ ++void INIT ZSTD_free(void *ptr, ZSTD_customMem customMem) ++{ ++ if (ptr != NULL) ++ customMem.customFree(customMem.opaque, ptr); ++} +diff --git a/xen/common/zstd/zstd_internal.h b/xen/common/zstd/zstd_internal.h +new file mode 100644 +index 0000000000..1b13840c44 +--- /dev/null ++++ b/xen/common/zstd/zstd_internal.h +@@ -0,0 +1,265 @@ ++/** ++ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc. ++ * All rights reserved. ++ * ++ * This source code is licensed under the BSD-style license found in the ++ * LICENSE file in the root directory of https://github.com/facebook/zstd. ++ * An additional grant of patent rights can be found in the PATENTS file in the ++ * same directory. ++ * ++ * This program is free software; you can redistribute it and/or modify it under ++ * the terms of the GNU General Public License version 2 as published by the ++ * Free Software Foundation. This program is dual-licensed; you may select ++ * either version 2 of the GNU General Public License ("GPL") or BSD license ++ * ("BSD"). ++ */ ++ ++#ifndef ZSTD_CCOMMON_H_MODULE ++#define ZSTD_CCOMMON_H_MODULE ++ ++/*-******************************************************* ++* Compiler specifics ++*********************************************************/ ++#define FORCE_INLINE static always_inline ++#define FORCE_NOINLINE static noinline ++ ++/*-************************************* ++* Dependencies ++***************************************/ ++#include "error_private.h" ++#include "mem.h" ++#include ++#include ++ ++/*-************************************* ++* shared macros ++***************************************/ ++#define CHECK_F(f) \ ++ { \ ++ size_t const errcod = f; \ ++ if (ERR_isError(errcod)) \ ++ return errcod; \ ++ } /* check and Forward error code */ ++#define CHECK_E(f, e) \ ++ { \ ++ size_t const errcod = f; \ ++ if (ERR_isError(errcod)) \ ++ return ERROR(e); \ ++ } /* check and send Error code */ ++#define ZSTD_STATIC_ASSERT(c) \ ++ { \ ++ enum { ZSTD_static_assert = 1 / (int)(!!(c)) }; \ ++ } ++ ++/*-************************************* ++* Common constants ++***************************************/ ++#define ZSTD_OPT_NUM (1 << 12) ++#define ZSTD_DICT_MAGIC 0xEC30A437 /* v0.7+ */ ++ ++#define ZSTD_REP_NUM 3 /* number of repcodes */ ++#define ZSTD_REP_CHECK (ZSTD_REP_NUM) /* number of repcodes to check by the optimal parser */ ++#define ZSTD_REP_MOVE (ZSTD_REP_NUM - 1) ++#define ZSTD_REP_MOVE_OPT (ZSTD_REP_NUM) ++static const U32 repStartValue[ZSTD_REP_NUM] = {1, 4, 8}; ++ ++#define BIT7 128 ++#define BIT6 64 ++#define BIT5 32 ++#define BIT4 16 ++#define BIT1 2 ++#define BIT0 1 ++ ++#define ZSTD_WINDOWLOG_ABSOLUTEMIN 10 ++static const size_t ZSTD_fcs_fieldSize[4] = {0, 2, 4, 8}; ++static const size_t ZSTD_did_fieldSize[4] = {0, 1, 2, 4}; ++ ++#define ZSTD_BLOCKHEADERSIZE 3 /* C standard doesn't allow `static const` variable to be init using another `static const` variable */ ++static const size_t ZSTD_blockHeaderSize = ZSTD_BLOCKHEADERSIZE; ++typedef enum { bt_raw, bt_rle, bt_compressed, bt_reserved } blockType_e; ++ ++#define MIN_SEQUENCES_SIZE 1 /* nbSeq==0 */ ++#define MIN_CBLOCK_SIZE (1 /*litCSize*/ + 1 /* RLE or RAW */ + MIN_SEQUENCES_SIZE /* nbSeq==0 */) /* for a non-null block */ ++ ++#define HufLog 12 ++typedef enum { set_basic, set_rle, set_compressed, set_repeat } symbolEncodingType_e; ++ ++#define LONGNBSEQ 0x7F00 ++ ++#define MINMATCH 3 ++#define EQUAL_READ32 4 ++ ++#define Litbits 8 ++#define MaxLit ((1 << Litbits) - 1) ++#define MaxML 52 ++#define MaxLL 35 ++#define MaxOff 28 ++#define MaxSeq MAX(MaxLL, MaxML) /* Assumption : MaxOff < MaxLL,MaxML */ ++#define MLFSELog 9 ++#define LLFSELog 9 ++#define OffFSELog 8 ++ ++static const U32 LL_bits[MaxLL + 1] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 3, 3, 4, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16}; ++static const S16 LL_defaultNorm[MaxLL + 1] = {4, 3, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 2, 1, 1, 1, 1, 1, -1, -1, -1, -1}; ++#define LL_DEFAULTNORMLOG 6 /* for static allocation */ ++static const U32 LL_defaultNormLog = LL_DEFAULTNORMLOG; ++ ++static const U32 ML_bits[MaxML + 1] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ++ 0, 0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 3, 3, 4, 4, 5, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16}; ++static const S16 ML_defaultNorm[MaxML + 1] = {1, 4, 3, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ++ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1}; ++#define ML_DEFAULTNORMLOG 6 /* for static allocation */ ++static const U32 ML_defaultNormLog = ML_DEFAULTNORMLOG; ++ ++static const S16 OF_defaultNorm[MaxOff + 1] = {1, 1, 1, 1, 1, 1, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1}; ++#define OF_DEFAULTNORMLOG 5 /* for static allocation */ ++static const U32 OF_defaultNormLog = OF_DEFAULTNORMLOG; ++ ++/*-******************************************* ++* Shared functions to include for inlining ++*********************************************/ ++ZSTD_STATIC void ZSTD_copy8(void *dst, const void *src) { ++ /* ++ * zstd relies heavily on gcc being able to analyze and inline this ++ * memcpy() call, since it is called in a tight loop. Preboot mode ++ * is compiled in freestanding mode, which stops gcc from analyzing ++ * memcpy(). Use __builtin_memcpy() to tell gcc to analyze this as a ++ * regular memcpy(). ++ */ ++ __builtin_memcpy(dst, src, 8); ++} ++/*! ZSTD_wildcopy() : ++* custom version of memcpy(), can copy up to 7 bytes too many (8 bytes if length==0) */ ++#define WILDCOPY_OVERLENGTH 8 ++ZSTD_STATIC void ZSTD_wildcopy(void *dst, const void *src, ptrdiff_t length) ++{ ++ const BYTE* ip = (const BYTE*)src; ++ BYTE* op = (BYTE*)dst; ++ BYTE* const oend = op + length; ++#if defined(GCC_VERSION) && GCC_VERSION >= 70000 && GCC_VERSION < 70200 ++ /* ++ * Work around https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81388. ++ * Avoid the bad case where the loop only runs once by handling the ++ * special case separately. This doesn't trigger the bug because it ++ * doesn't involve pointer/integer overflow. ++ */ ++ if (length <= 8) ++ return ZSTD_copy8(dst, src); ++#endif ++ do { ++ ZSTD_copy8(op, ip); ++ op += 8; ++ ip += 8; ++ } while (op < oend); ++} ++ ++/*-******************************************* ++* Private interfaces ++*********************************************/ ++typedef struct ZSTD_stats_s ZSTD_stats_t; ++ ++typedef struct { ++ U32 off; ++ U32 len; ++} ZSTD_match_t; ++ ++typedef struct { ++ U32 price; ++ U32 off; ++ U32 mlen; ++ U32 litlen; ++ U32 rep[ZSTD_REP_NUM]; ++} ZSTD_optimal_t; ++ ++typedef struct seqDef_s { ++ U32 offset; ++ U16 litLength; ++ U16 matchLength; ++} seqDef; ++ ++typedef struct { ++ seqDef *sequencesStart; ++ seqDef *sequences; ++ BYTE *litStart; ++ BYTE *lit; ++ BYTE *llCode; ++ BYTE *mlCode; ++ BYTE *ofCode; ++ U32 longLengthID; /* 0 == no longLength; 1 == Lit.longLength; 2 == Match.longLength; */ ++ U32 longLengthPos; ++ /* opt */ ++ ZSTD_optimal_t *priceTable; ++ ZSTD_match_t *matchTable; ++ U32 *matchLengthFreq; ++ U32 *litLengthFreq; ++ U32 *litFreq; ++ U32 *offCodeFreq; ++ U32 matchLengthSum; ++ U32 matchSum; ++ U32 litLengthSum; ++ U32 litSum; ++ U32 offCodeSum; ++ U32 log2matchLengthSum; ++ U32 log2matchSum; ++ U32 log2litLengthSum; ++ U32 log2litSum; ++ U32 log2offCodeSum; ++ U32 factor; ++ U32 staticPrices; ++ U32 cachedPrice; ++ U32 cachedLitLength; ++ const BYTE *cachedLiterals; ++} seqStore_t; ++ ++const seqStore_t *ZSTD_getSeqStore(const ZSTD_CCtx *ctx); ++void ZSTD_seqToCodes(const seqStore_t *seqStorePtr); ++int ZSTD_isSkipFrame(ZSTD_DCtx *dctx); ++ ++/*= Custom memory allocation functions */ ++typedef void *(*ZSTD_allocFunction)(void *opaque, size_t size); ++typedef void (*ZSTD_freeFunction)(void *opaque, void *address); ++typedef struct { ++ ZSTD_allocFunction customAlloc; ++ ZSTD_freeFunction customFree; ++ void *opaque; ++} ZSTD_customMem; ++ ++void *ZSTD_malloc(size_t size, ZSTD_customMem customMem); ++void ZSTD_free(void *ptr, ZSTD_customMem customMem); ++ ++/*====== stack allocation ======*/ ++ ++typedef struct { ++ void *ptr; ++ const void *end; ++} ZSTD_stack; ++ ++#define ZSTD_ALIGN(x) ALIGN(x, sizeof(size_t)) ++#define ZSTD_PTR_ALIGN(p) PTR_ALIGN(p, sizeof(size_t)) ++ ++ZSTD_customMem ZSTD_initStack(void *workspace, size_t workspaceSize); ++ ++void *ZSTD_stackAllocAll(void *opaque, size_t *size); ++void *ZSTD_stackAlloc(void *opaque, size_t size); ++void ZSTD_stackFree(void *opaque, void *address); ++ ++/*====== common function ======*/ ++ ++ZSTD_STATIC U32 ZSTD_highbit32(U32 val) { return 31 - __builtin_clz(val); } ++ ++/* hidden functions */ ++ ++/* ZSTD_invalidateRepCodes() : ++ * ensures next compression will not use repcodes from previous block. ++ * Note : only works with regular variant; ++ * do not use with extDict variant ! */ ++void ZSTD_invalidateRepCodes(ZSTD_CCtx *cctx); ++ ++size_t ZSTD_freeCCtx(ZSTD_CCtx *cctx); ++size_t ZSTD_freeDCtx(ZSTD_DCtx *dctx); ++size_t ZSTD_freeCDict(ZSTD_CDict *cdict); ++size_t ZSTD_freeDDict(ZSTD_DDict *cdict); ++size_t ZSTD_freeCStream(ZSTD_CStream *zcs); ++size_t ZSTD_freeDStream(ZSTD_DStream *zds); ++ ++#endif /* ZSTD_CCOMMON_H_MODULE */ +diff --git a/xen/common/zstd/zstd_opt.h b/xen/common/zstd/zstd_opt.h +new file mode 100644 +index 0000000000..55e1b4cba8 +--- /dev/null ++++ b/xen/common/zstd/zstd_opt.h +@@ -0,0 +1,1014 @@ ++/** ++ * Copyright (c) 2016-present, Przemyslaw Skibinski, Yann Collet, Facebook, Inc. ++ * All rights reserved. ++ * ++ * This source code is licensed under the BSD-style license found in the ++ * LICENSE file in the root directory of https://github.com/facebook/zstd. ++ * An additional grant of patent rights can be found in the PATENTS file in the ++ * same directory. ++ * ++ * This program is free software; you can redistribute it and/or modify it under ++ * the terms of the GNU General Public License version 2 as published by the ++ * Free Software Foundation. This program is dual-licensed; you may select ++ * either version 2 of the GNU General Public License ("GPL") or BSD license ++ * ("BSD"). ++ */ ++ ++/* Note : this file is intended to be included within zstd_compress.c */ ++ ++#ifndef ZSTD_OPT_H_91842398743 ++#define ZSTD_OPT_H_91842398743 ++ ++#define ZSTD_LITFREQ_ADD 2 ++#define ZSTD_FREQ_DIV 4 ++#define ZSTD_MAX_PRICE (1 << 30) ++ ++/*-************************************* ++* Price functions for optimal parser ++***************************************/ ++FORCE_INLINE void ZSTD_setLog2Prices(seqStore_t *ssPtr) ++{ ++ ssPtr->log2matchLengthSum = ZSTD_highbit32(ssPtr->matchLengthSum + 1); ++ ssPtr->log2litLengthSum = ZSTD_highbit32(ssPtr->litLengthSum + 1); ++ ssPtr->log2litSum = ZSTD_highbit32(ssPtr->litSum + 1); ++ ssPtr->log2offCodeSum = ZSTD_highbit32(ssPtr->offCodeSum + 1); ++ ssPtr->factor = 1 + ((ssPtr->litSum >> 5) / ssPtr->litLengthSum) + ((ssPtr->litSum << 1) / (ssPtr->litSum + ssPtr->matchSum)); ++} ++ ++ZSTD_STATIC void ZSTD_rescaleFreqs(seqStore_t *ssPtr, const BYTE *src, size_t srcSize) ++{ ++ unsigned u; ++ ++ ssPtr->cachedLiterals = NULL; ++ ssPtr->cachedPrice = ssPtr->cachedLitLength = 0; ++ ssPtr->staticPrices = 0; ++ ++ if (ssPtr->litLengthSum == 0) { ++ if (srcSize <= 1024) ++ ssPtr->staticPrices = 1; ++ ++ for (u = 0; u <= MaxLit; u++) ++ ssPtr->litFreq[u] = 0; ++ for (u = 0; u < srcSize; u++) ++ ssPtr->litFreq[src[u]]++; ++ ++ ssPtr->litSum = 0; ++ ssPtr->litLengthSum = MaxLL + 1; ++ ssPtr->matchLengthSum = MaxML + 1; ++ ssPtr->offCodeSum = (MaxOff + 1); ++ ssPtr->matchSum = (ZSTD_LITFREQ_ADD << Litbits); ++ ++ for (u = 0; u <= MaxLit; u++) { ++ ssPtr->litFreq[u] = 1 + (ssPtr->litFreq[u] >> ZSTD_FREQ_DIV); ++ ssPtr->litSum += ssPtr->litFreq[u]; ++ } ++ for (u = 0; u <= MaxLL; u++) ++ ssPtr->litLengthFreq[u] = 1; ++ for (u = 0; u <= MaxML; u++) ++ ssPtr->matchLengthFreq[u] = 1; ++ for (u = 0; u <= MaxOff; u++) ++ ssPtr->offCodeFreq[u] = 1; ++ } else { ++ ssPtr->matchLengthSum = 0; ++ ssPtr->litLengthSum = 0; ++ ssPtr->offCodeSum = 0; ++ ssPtr->matchSum = 0; ++ ssPtr->litSum = 0; ++ ++ for (u = 0; u <= MaxLit; u++) { ++ ssPtr->litFreq[u] = 1 + (ssPtr->litFreq[u] >> (ZSTD_FREQ_DIV + 1)); ++ ssPtr->litSum += ssPtr->litFreq[u]; ++ } ++ for (u = 0; u <= MaxLL; u++) { ++ ssPtr->litLengthFreq[u] = 1 + (ssPtr->litLengthFreq[u] >> (ZSTD_FREQ_DIV + 1)); ++ ssPtr->litLengthSum += ssPtr->litLengthFreq[u]; ++ } ++ for (u = 0; u <= MaxML; u++) { ++ ssPtr->matchLengthFreq[u] = 1 + (ssPtr->matchLengthFreq[u] >> ZSTD_FREQ_DIV); ++ ssPtr->matchLengthSum += ssPtr->matchLengthFreq[u]; ++ ssPtr->matchSum += ssPtr->matchLengthFreq[u] * (u + 3); ++ } ++ ssPtr->matchSum *= ZSTD_LITFREQ_ADD; ++ for (u = 0; u <= MaxOff; u++) { ++ ssPtr->offCodeFreq[u] = 1 + (ssPtr->offCodeFreq[u] >> ZSTD_FREQ_DIV); ++ ssPtr->offCodeSum += ssPtr->offCodeFreq[u]; ++ } ++ } ++ ++ ZSTD_setLog2Prices(ssPtr); ++} ++ ++FORCE_INLINE U32 ZSTD_getLiteralPrice(seqStore_t *ssPtr, U32 litLength, const BYTE *literals) ++{ ++ U32 price, u; ++ ++ if (ssPtr->staticPrices) ++ return ZSTD_highbit32((U32)litLength + 1) + (litLength * 6); ++ ++ if (litLength == 0) ++ return ssPtr->log2litLengthSum - ZSTD_highbit32(ssPtr->litLengthFreq[0] + 1); ++ ++ /* literals */ ++ if (ssPtr->cachedLiterals == literals) { ++ U32 const additional = litLength - ssPtr->cachedLitLength; ++ const BYTE *literals2 = ssPtr->cachedLiterals + ssPtr->cachedLitLength; ++ price = ssPtr->cachedPrice + additional * ssPtr->log2litSum; ++ for (u = 0; u < additional; u++) ++ price -= ZSTD_highbit32(ssPtr->litFreq[literals2[u]] + 1); ++ ssPtr->cachedPrice = price; ++ ssPtr->cachedLitLength = litLength; ++ } else { ++ price = litLength * ssPtr->log2litSum; ++ for (u = 0; u < litLength; u++) ++ price -= ZSTD_highbit32(ssPtr->litFreq[literals[u]] + 1); ++ ++ if (litLength >= 12) { ++ ssPtr->cachedLiterals = literals; ++ ssPtr->cachedPrice = price; ++ ssPtr->cachedLitLength = litLength; ++ } ++ } ++ ++ /* literal Length */ ++ { ++ const BYTE LL_deltaCode = 19; ++ const BYTE llCode = (litLength > 63) ? (BYTE)ZSTD_highbit32(litLength) + LL_deltaCode : LL_Code[litLength]; ++ price += LL_bits[llCode] + ssPtr->log2litLengthSum - ZSTD_highbit32(ssPtr->litLengthFreq[llCode] + 1); ++ } ++ ++ return price; ++} ++ ++FORCE_INLINE U32 ZSTD_getPrice(seqStore_t *seqStorePtr, U32 litLength, const BYTE *literals, U32 offset, U32 matchLength, const int ultra) ++{ ++ /* offset */ ++ U32 price; ++ BYTE const offCode = (BYTE)ZSTD_highbit32(offset + 1); ++ ++ if (seqStorePtr->staticPrices) ++ return ZSTD_getLiteralPrice(seqStorePtr, litLength, literals) + ZSTD_highbit32((U32)matchLength + 1) + 16 + offCode; ++ ++ price = offCode + seqStorePtr->log2offCodeSum - ZSTD_highbit32(seqStorePtr->offCodeFreq[offCode] + 1); ++ if (!ultra && offCode >= 20) ++ price += (offCode - 19) * 2; ++ ++ /* match Length */ ++ { ++ const BYTE ML_deltaCode = 36; ++ const BYTE mlCode = (matchLength > 127) ? (BYTE)ZSTD_highbit32(matchLength) + ML_deltaCode : ML_Code[matchLength]; ++ price += ML_bits[mlCode] + seqStorePtr->log2matchLengthSum - ZSTD_highbit32(seqStorePtr->matchLengthFreq[mlCode] + 1); ++ } ++ ++ return price + ZSTD_getLiteralPrice(seqStorePtr, litLength, literals) + seqStorePtr->factor; ++} ++ ++ZSTD_STATIC void ZSTD_updatePrice(seqStore_t *seqStorePtr, U32 litLength, const BYTE *literals, U32 offset, U32 matchLength) ++{ ++ U32 u; ++ ++ /* literals */ ++ seqStorePtr->litSum += litLength * ZSTD_LITFREQ_ADD; ++ for (u = 0; u < litLength; u++) ++ seqStorePtr->litFreq[literals[u]] += ZSTD_LITFREQ_ADD; ++ ++ /* literal Length */ ++ { ++ const BYTE LL_deltaCode = 19; ++ const BYTE llCode = (litLength > 63) ? (BYTE)ZSTD_highbit32(litLength) + LL_deltaCode : LL_Code[litLength]; ++ seqStorePtr->litLengthFreq[llCode]++; ++ seqStorePtr->litLengthSum++; ++ } ++ ++ /* match offset */ ++ { ++ BYTE const offCode = (BYTE)ZSTD_highbit32(offset + 1); ++ seqStorePtr->offCodeSum++; ++ seqStorePtr->offCodeFreq[offCode]++; ++ } ++ ++ /* match Length */ ++ { ++ const BYTE ML_deltaCode = 36; ++ const BYTE mlCode = (matchLength > 127) ? (BYTE)ZSTD_highbit32(matchLength) + ML_deltaCode : ML_Code[matchLength]; ++ seqStorePtr->matchLengthFreq[mlCode]++; ++ seqStorePtr->matchLengthSum++; ++ } ++ ++ ZSTD_setLog2Prices(seqStorePtr); ++} ++ ++#define SET_PRICE(pos, mlen_, offset_, litlen_, price_) \ ++ { \ ++ while (last_pos < pos) { \ ++ opt[last_pos + 1].price = ZSTD_MAX_PRICE; \ ++ last_pos++; \ ++ } \ ++ opt[pos].mlen = mlen_; \ ++ opt[pos].off = offset_; \ ++ opt[pos].litlen = litlen_; \ ++ opt[pos].price = price_; \ ++ } ++ ++/* Update hashTable3 up to ip (excluded) ++ Assumption : always within prefix (i.e. not within extDict) */ ++FORCE_INLINE ++U32 ZSTD_insertAndFindFirstIndexHash3(ZSTD_CCtx *zc, const BYTE *ip) ++{ ++ U32 *const hashTable3 = zc->hashTable3; ++ U32 const hashLog3 = zc->hashLog3; ++ const BYTE *const base = zc->base; ++ U32 idx = zc->nextToUpdate3; ++ const U32 target = zc->nextToUpdate3 = (U32)(ip - base); ++ const size_t hash3 = ZSTD_hash3Ptr(ip, hashLog3); ++ ++ while (idx < target) { ++ hashTable3[ZSTD_hash3Ptr(base + idx, hashLog3)] = idx; ++ idx++; ++ } ++ ++ return hashTable3[hash3]; ++} ++ ++/*-************************************* ++* Binary Tree search ++***************************************/ ++static U32 ZSTD_insertBtAndGetAllMatches(ZSTD_CCtx *zc, const BYTE *const ip, const BYTE *const iLimit, U32 nbCompares, const U32 mls, U32 extDict, ++ ZSTD_match_t *matches, const U32 minMatchLen) ++{ ++ const BYTE *const base = zc->base; ++ const U32 curr = (U32)(ip - base); ++ const U32 hashLog = zc->params.cParams.hashLog; ++ const size_t h = ZSTD_hashPtr(ip, hashLog, mls); ++ U32 *const hashTable = zc->hashTable; ++ U32 matchIndex = hashTable[h]; ++ U32 *const bt = zc->chainTable; ++ const U32 btLog = zc->params.cParams.chainLog - 1; ++ const U32 btMask = (1U << btLog) - 1; ++ size_t commonLengthSmaller = 0, commonLengthLarger = 0; ++ const BYTE *const dictBase = zc->dictBase; ++ const U32 dictLimit = zc->dictLimit; ++ const BYTE *const dictEnd = dictBase + dictLimit; ++ const BYTE *const prefixStart = base + dictLimit; ++ const U32 btLow = btMask >= curr ? 0 : curr - btMask; ++ const U32 windowLow = zc->lowLimit; ++ U32 *smallerPtr = bt + 2 * (curr & btMask); ++ U32 *largerPtr = bt + 2 * (curr & btMask) + 1; ++ U32 matchEndIdx = curr + 8; ++ U32 dummy32; /* to be nullified at the end */ ++ U32 mnum = 0; ++ ++ const U32 minMatch = (mls == 3) ? 3 : 4; ++ size_t bestLength = minMatchLen - 1; ++ ++ if (minMatch == 3) { /* HC3 match finder */ ++ U32 const matchIndex3 = ZSTD_insertAndFindFirstIndexHash3(zc, ip); ++ if (matchIndex3 > windowLow && (curr - matchIndex3 < (1 << 18))) { ++ const BYTE *match; ++ size_t currMl = 0; ++ if ((!extDict) || matchIndex3 >= dictLimit) { ++ match = base + matchIndex3; ++ if (match[bestLength] == ip[bestLength]) ++ currMl = ZSTD_count(ip, match, iLimit); ++ } else { ++ match = dictBase + matchIndex3; ++ if (ZSTD_readMINMATCH(match, MINMATCH) == ++ ZSTD_readMINMATCH(ip, MINMATCH)) /* assumption : matchIndex3 <= dictLimit-4 (by table construction) */ ++ currMl = ZSTD_count_2segments(ip + MINMATCH, match + MINMATCH, iLimit, dictEnd, prefixStart) + MINMATCH; ++ } ++ ++ /* save best solution */ ++ if (currMl > bestLength) { ++ bestLength = currMl; ++ matches[mnum].off = ZSTD_REP_MOVE_OPT + curr - matchIndex3; ++ matches[mnum].len = (U32)currMl; ++ mnum++; ++ if (currMl > ZSTD_OPT_NUM) ++ goto update; ++ if (ip + currMl == iLimit) ++ goto update; /* best possible, and avoid read overflow*/ ++ } ++ } ++ } ++ ++ hashTable[h] = curr; /* Update Hash Table */ ++ ++ while (nbCompares-- && (matchIndex > windowLow)) { ++ U32 *nextPtr = bt + 2 * (matchIndex & btMask); ++ size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger); /* guaranteed minimum nb of common bytes */ ++ const BYTE *match; ++ ++ if ((!extDict) || (matchIndex + matchLength >= dictLimit)) { ++ match = base + matchIndex; ++ if (match[matchLength] == ip[matchLength]) { ++ matchLength += ZSTD_count(ip + matchLength + 1, match + matchLength + 1, iLimit) + 1; ++ } ++ } else { ++ match = dictBase + matchIndex; ++ matchLength += ZSTD_count_2segments(ip + matchLength, match + matchLength, iLimit, dictEnd, prefixStart); ++ if (matchIndex + matchLength >= dictLimit) ++ match = base + matchIndex; /* to prepare for next usage of match[matchLength] */ ++ } ++ ++ if (matchLength > bestLength) { ++ if (matchLength > matchEndIdx - matchIndex) ++ matchEndIdx = matchIndex + (U32)matchLength; ++ bestLength = matchLength; ++ matches[mnum].off = ZSTD_REP_MOVE_OPT + curr - matchIndex; ++ matches[mnum].len = (U32)matchLength; ++ mnum++; ++ if (matchLength > ZSTD_OPT_NUM) ++ break; ++ if (ip + matchLength == iLimit) /* equal : no way to know if inf or sup */ ++ break; /* drop, to guarantee consistency (miss a little bit of compression) */ ++ } ++ ++ if (match[matchLength] < ip[matchLength]) { ++ /* match is smaller than curr */ ++ *smallerPtr = matchIndex; /* update smaller idx */ ++ commonLengthSmaller = matchLength; /* all smaller will now have at least this guaranteed common length */ ++ if (matchIndex <= btLow) { ++ smallerPtr = &dummy32; ++ break; ++ } /* beyond tree size, stop the search */ ++ smallerPtr = nextPtr + 1; /* new "smaller" => larger of match */ ++ matchIndex = nextPtr[1]; /* new matchIndex larger than previous (closer to curr) */ ++ } else { ++ /* match is larger than curr */ ++ *largerPtr = matchIndex; ++ commonLengthLarger = matchLength; ++ if (matchIndex <= btLow) { ++ largerPtr = &dummy32; ++ break; ++ } /* beyond tree size, stop the search */ ++ largerPtr = nextPtr; ++ matchIndex = nextPtr[0]; ++ } ++ } ++ ++ *smallerPtr = *largerPtr = 0; ++ ++update: ++ zc->nextToUpdate = (matchEndIdx > curr + 8) ? matchEndIdx - 8 : curr + 1; ++ return mnum; ++} ++ ++/** Tree updater, providing best match */ ++static U32 ZSTD_BtGetAllMatches(ZSTD_CCtx *zc, const BYTE *const ip, const BYTE *const iLimit, const U32 maxNbAttempts, const U32 mls, ZSTD_match_t *matches, ++ const U32 minMatchLen) ++{ ++ if (ip < zc->base + zc->nextToUpdate) ++ return 0; /* skipped area */ ++ ZSTD_updateTree(zc, ip, iLimit, maxNbAttempts, mls); ++ return ZSTD_insertBtAndGetAllMatches(zc, ip, iLimit, maxNbAttempts, mls, 0, matches, minMatchLen); ++} ++ ++static U32 ZSTD_BtGetAllMatches_selectMLS(ZSTD_CCtx *zc, /* Index table will be updated */ ++ const BYTE *ip, const BYTE *const iHighLimit, const U32 maxNbAttempts, const U32 matchLengthSearch, ++ ZSTD_match_t *matches, const U32 minMatchLen) ++{ ++ switch (matchLengthSearch) { ++ case 3: return ZSTD_BtGetAllMatches(zc, ip, iHighLimit, maxNbAttempts, 3, matches, minMatchLen); ++ default: ++ case 4: return ZSTD_BtGetAllMatches(zc, ip, iHighLimit, maxNbAttempts, 4, matches, minMatchLen); ++ case 5: return ZSTD_BtGetAllMatches(zc, ip, iHighLimit, maxNbAttempts, 5, matches, minMatchLen); ++ case 7: ++ case 6: return ZSTD_BtGetAllMatches(zc, ip, iHighLimit, maxNbAttempts, 6, matches, minMatchLen); ++ } ++} ++ ++/** Tree updater, providing best match */ ++static U32 ZSTD_BtGetAllMatches_extDict(ZSTD_CCtx *zc, const BYTE *const ip, const BYTE *const iLimit, const U32 maxNbAttempts, const U32 mls, ++ ZSTD_match_t *matches, const U32 minMatchLen) ++{ ++ if (ip < zc->base + zc->nextToUpdate) ++ return 0; /* skipped area */ ++ ZSTD_updateTree_extDict(zc, ip, iLimit, maxNbAttempts, mls); ++ return ZSTD_insertBtAndGetAllMatches(zc, ip, iLimit, maxNbAttempts, mls, 1, matches, minMatchLen); ++} ++ ++static U32 ZSTD_BtGetAllMatches_selectMLS_extDict(ZSTD_CCtx *zc, /* Index table will be updated */ ++ const BYTE *ip, const BYTE *const iHighLimit, const U32 maxNbAttempts, const U32 matchLengthSearch, ++ ZSTD_match_t *matches, const U32 minMatchLen) ++{ ++ switch (matchLengthSearch) { ++ case 3: return ZSTD_BtGetAllMatches_extDict(zc, ip, iHighLimit, maxNbAttempts, 3, matches, minMatchLen); ++ default: ++ case 4: return ZSTD_BtGetAllMatches_extDict(zc, ip, iHighLimit, maxNbAttempts, 4, matches, minMatchLen); ++ case 5: return ZSTD_BtGetAllMatches_extDict(zc, ip, iHighLimit, maxNbAttempts, 5, matches, minMatchLen); ++ case 7: ++ case 6: return ZSTD_BtGetAllMatches_extDict(zc, ip, iHighLimit, maxNbAttempts, 6, matches, minMatchLen); ++ } ++} ++ ++/*-******************************* ++* Optimal parser ++*********************************/ ++FORCE_INLINE ++void ZSTD_compressBlock_opt_generic(ZSTD_CCtx *ctx, const void *src, size_t srcSize, const int ultra) ++{ ++ seqStore_t *seqStorePtr = &(ctx->seqStore); ++ const BYTE *const istart = (const BYTE *)src; ++ const BYTE *ip = istart; ++ const BYTE *anchor = istart; ++ const BYTE *const iend = istart + srcSize; ++ const BYTE *const ilimit = iend - 8; ++ const BYTE *const base = ctx->base; ++ const BYTE *const prefixStart = base + ctx->dictLimit; ++ ++ const U32 maxSearches = 1U << ctx->params.cParams.searchLog; ++ const U32 sufficient_len = ctx->params.cParams.targetLength; ++ const U32 mls = ctx->params.cParams.searchLength; ++ const U32 minMatch = (ctx->params.cParams.searchLength == 3) ? 3 : 4; ++ ++ ZSTD_optimal_t *opt = seqStorePtr->priceTable; ++ ZSTD_match_t *matches = seqStorePtr->matchTable; ++ const BYTE *inr; ++ U32 offset, rep[ZSTD_REP_NUM]; ++ ++ /* init */ ++ ctx->nextToUpdate3 = ctx->nextToUpdate; ++ ZSTD_rescaleFreqs(seqStorePtr, (const BYTE *)src, srcSize); ++ ip += (ip == prefixStart); ++ { ++ U32 i; ++ for (i = 0; i < ZSTD_REP_NUM; i++) ++ rep[i] = ctx->rep[i]; ++ } ++ ++ /* Match Loop */ ++ while (ip < ilimit) { ++ U32 cur, match_num, last_pos, litlen, price; ++ U32 u, mlen, best_mlen, best_off, litLength; ++ memset(opt, 0, sizeof(ZSTD_optimal_t)); ++ last_pos = 0; ++ litlen = (U32)(ip - anchor); ++ ++ /* check repCode */ ++ { ++ U32 i, last_i = ZSTD_REP_CHECK + (ip == anchor); ++ for (i = (ip == anchor); i < last_i; i++) { ++ const S32 repCur = (i == ZSTD_REP_MOVE_OPT) ? (rep[0] - 1) : rep[i]; ++ if ((repCur > 0) && (repCur < (S32)(ip - prefixStart)) && ++ (ZSTD_readMINMATCH(ip, minMatch) == ZSTD_readMINMATCH(ip - repCur, minMatch))) { ++ mlen = (U32)ZSTD_count(ip + minMatch, ip + minMatch - repCur, iend) + minMatch; ++ if (mlen > sufficient_len || mlen >= ZSTD_OPT_NUM) { ++ best_mlen = mlen; ++ best_off = i; ++ cur = 0; ++ last_pos = 1; ++ goto _storeSequence; ++ } ++ best_off = i - (ip == anchor); ++ do { ++ price = ZSTD_getPrice(seqStorePtr, litlen, anchor, best_off, mlen - MINMATCH, ultra); ++ if (mlen > last_pos || price < opt[mlen].price) ++ SET_PRICE(mlen, mlen, i, litlen, price); /* note : macro modifies last_pos */ ++ mlen--; ++ } while (mlen >= minMatch); ++ } ++ } ++ } ++ ++ match_num = ZSTD_BtGetAllMatches_selectMLS(ctx, ip, iend, maxSearches, mls, matches, minMatch); ++ ++ if (!last_pos && !match_num) { ++ ip++; ++ continue; ++ } ++ ++ if (match_num && (matches[match_num - 1].len > sufficient_len || matches[match_num - 1].len >= ZSTD_OPT_NUM)) { ++ best_mlen = matches[match_num - 1].len; ++ best_off = matches[match_num - 1].off; ++ cur = 0; ++ last_pos = 1; ++ goto _storeSequence; ++ } ++ ++ /* set prices using matches at position = 0 */ ++ best_mlen = (last_pos) ? last_pos : minMatch; ++ for (u = 0; u < match_num; u++) { ++ mlen = (u > 0) ? matches[u - 1].len + 1 : best_mlen; ++ best_mlen = matches[u].len; ++ while (mlen <= best_mlen) { ++ price = ZSTD_getPrice(seqStorePtr, litlen, anchor, matches[u].off - 1, mlen - MINMATCH, ultra); ++ if (mlen > last_pos || price < opt[mlen].price) ++ SET_PRICE(mlen, mlen, matches[u].off, litlen, price); /* note : macro modifies last_pos */ ++ mlen++; ++ } ++ } ++ ++ if (last_pos < minMatch) { ++ ip++; ++ continue; ++ } ++ ++ /* initialize opt[0] */ ++ { ++ U32 i; ++ for (i = 0; i < ZSTD_REP_NUM; i++) ++ opt[0].rep[i] = rep[i]; ++ } ++ opt[0].mlen = 1; ++ opt[0].litlen = litlen; ++ ++ /* check further positions */ ++ for (cur = 1; cur <= last_pos; cur++) { ++ inr = ip + cur; ++ ++ if (opt[cur - 1].mlen == 1) { ++ litlen = opt[cur - 1].litlen + 1; ++ if (cur > litlen) { ++ price = opt[cur - litlen].price + ZSTD_getLiteralPrice(seqStorePtr, litlen, inr - litlen); ++ } else ++ price = ZSTD_getLiteralPrice(seqStorePtr, litlen, anchor); ++ } else { ++ litlen = 1; ++ price = opt[cur - 1].price + ZSTD_getLiteralPrice(seqStorePtr, litlen, inr - 1); ++ } ++ ++ if (cur > last_pos || price <= opt[cur].price) ++ SET_PRICE(cur, 1, 0, litlen, price); ++ ++ if (cur == last_pos) ++ break; ++ ++ if (inr > ilimit) /* last match must start at a minimum distance of 8 from oend */ ++ continue; ++ ++ mlen = opt[cur].mlen; ++ if (opt[cur].off > ZSTD_REP_MOVE_OPT) { ++ opt[cur].rep[2] = opt[cur - mlen].rep[1]; ++ opt[cur].rep[1] = opt[cur - mlen].rep[0]; ++ opt[cur].rep[0] = opt[cur].off - ZSTD_REP_MOVE_OPT; ++ } else { ++ opt[cur].rep[2] = (opt[cur].off > 1) ? opt[cur - mlen].rep[1] : opt[cur - mlen].rep[2]; ++ opt[cur].rep[1] = (opt[cur].off > 0) ? opt[cur - mlen].rep[0] : opt[cur - mlen].rep[1]; ++ opt[cur].rep[0] = ++ ((opt[cur].off == ZSTD_REP_MOVE_OPT) && (mlen != 1)) ? (opt[cur - mlen].rep[0] - 1) : (opt[cur - mlen].rep[opt[cur].off]); ++ } ++ ++ best_mlen = minMatch; ++ { ++ U32 i, last_i = ZSTD_REP_CHECK + (mlen != 1); ++ for (i = (opt[cur].mlen != 1); i < last_i; i++) { /* check rep */ ++ const S32 repCur = (i == ZSTD_REP_MOVE_OPT) ? (opt[cur].rep[0] - 1) : opt[cur].rep[i]; ++ if ((repCur > 0) && (repCur < (S32)(inr - prefixStart)) && ++ (ZSTD_readMINMATCH(inr, minMatch) == ZSTD_readMINMATCH(inr - repCur, minMatch))) { ++ mlen = (U32)ZSTD_count(inr + minMatch, inr + minMatch - repCur, iend) + minMatch; ++ ++ if (mlen > sufficient_len || cur + mlen >= ZSTD_OPT_NUM) { ++ best_mlen = mlen; ++ best_off = i; ++ last_pos = cur + 1; ++ goto _storeSequence; ++ } ++ ++ best_off = i - (opt[cur].mlen != 1); ++ if (mlen > best_mlen) ++ best_mlen = mlen; ++ ++ do { ++ if (opt[cur].mlen == 1) { ++ litlen = opt[cur].litlen; ++ if (cur > litlen) { ++ price = opt[cur - litlen].price + ZSTD_getPrice(seqStorePtr, litlen, inr - litlen, ++ best_off, mlen - MINMATCH, ultra); ++ } else ++ price = ZSTD_getPrice(seqStorePtr, litlen, anchor, best_off, mlen - MINMATCH, ultra); ++ } else { ++ litlen = 0; ++ price = opt[cur].price + ZSTD_getPrice(seqStorePtr, 0, NULL, best_off, mlen - MINMATCH, ultra); ++ } ++ ++ if (cur + mlen > last_pos || price <= opt[cur + mlen].price) ++ SET_PRICE(cur + mlen, mlen, i, litlen, price); ++ mlen--; ++ } while (mlen >= minMatch); ++ } ++ } ++ } ++ ++ match_num = ZSTD_BtGetAllMatches_selectMLS(ctx, inr, iend, maxSearches, mls, matches, best_mlen); ++ ++ if (match_num > 0 && (matches[match_num - 1].len > sufficient_len || cur + matches[match_num - 1].len >= ZSTD_OPT_NUM)) { ++ best_mlen = matches[match_num - 1].len; ++ best_off = matches[match_num - 1].off; ++ last_pos = cur + 1; ++ goto _storeSequence; ++ } ++ ++ /* set prices using matches at position = cur */ ++ for (u = 0; u < match_num; u++) { ++ mlen = (u > 0) ? matches[u - 1].len + 1 : best_mlen; ++ best_mlen = matches[u].len; ++ ++ while (mlen <= best_mlen) { ++ if (opt[cur].mlen == 1) { ++ litlen = opt[cur].litlen; ++ if (cur > litlen) ++ price = opt[cur - litlen].price + ZSTD_getPrice(seqStorePtr, litlen, ip + cur - litlen, ++ matches[u].off - 1, mlen - MINMATCH, ultra); ++ else ++ price = ZSTD_getPrice(seqStorePtr, litlen, anchor, matches[u].off - 1, mlen - MINMATCH, ultra); ++ } else { ++ litlen = 0; ++ price = opt[cur].price + ZSTD_getPrice(seqStorePtr, 0, NULL, matches[u].off - 1, mlen - MINMATCH, ultra); ++ } ++ ++ if (cur + mlen > last_pos || (price < opt[cur + mlen].price)) ++ SET_PRICE(cur + mlen, mlen, matches[u].off, litlen, price); ++ ++ mlen++; ++ } ++ } ++ } ++ ++ best_mlen = opt[last_pos].mlen; ++ best_off = opt[last_pos].off; ++ cur = last_pos - best_mlen; ++ ++ /* store sequence */ ++_storeSequence: /* cur, last_pos, best_mlen, best_off have to be set */ ++ opt[0].mlen = 1; ++ ++ while (1) { ++ mlen = opt[cur].mlen; ++ offset = opt[cur].off; ++ opt[cur].mlen = best_mlen; ++ opt[cur].off = best_off; ++ best_mlen = mlen; ++ best_off = offset; ++ if (mlen > cur) ++ break; ++ cur -= mlen; ++ } ++ ++ for (u = 0; u <= last_pos;) { ++ u += opt[u].mlen; ++ } ++ ++ for (cur = 0; cur < last_pos;) { ++ mlen = opt[cur].mlen; ++ if (mlen == 1) { ++ ip++; ++ cur++; ++ continue; ++ } ++ offset = opt[cur].off; ++ cur += mlen; ++ litLength = (U32)(ip - anchor); ++ ++ if (offset > ZSTD_REP_MOVE_OPT) { ++ rep[2] = rep[1]; ++ rep[1] = rep[0]; ++ rep[0] = offset - ZSTD_REP_MOVE_OPT; ++ offset--; ++ } else { ++ if (offset != 0) { ++ best_off = (offset == ZSTD_REP_MOVE_OPT) ? (rep[0] - 1) : (rep[offset]); ++ if (offset != 1) ++ rep[2] = rep[1]; ++ rep[1] = rep[0]; ++ rep[0] = best_off; ++ } ++ if (litLength == 0) ++ offset--; ++ } ++ ++ ZSTD_updatePrice(seqStorePtr, litLength, anchor, offset, mlen - MINMATCH); ++ ZSTD_storeSeq(seqStorePtr, litLength, anchor, offset, mlen - MINMATCH); ++ anchor = ip = ip + mlen; ++ } ++ } /* for (cur=0; cur < last_pos; ) */ ++ ++ /* Save reps for next block */ ++ { ++ int i; ++ for (i = 0; i < ZSTD_REP_NUM; i++) ++ ctx->repToConfirm[i] = rep[i]; ++ } ++ ++ /* Last Literals */ ++ { ++ size_t const lastLLSize = iend - anchor; ++ memcpy(seqStorePtr->lit, anchor, lastLLSize); ++ seqStorePtr->lit += lastLLSize; ++ } ++} ++ ++FORCE_INLINE ++void ZSTD_compressBlock_opt_extDict_generic(ZSTD_CCtx *ctx, const void *src, size_t srcSize, const int ultra) ++{ ++ seqStore_t *seqStorePtr = &(ctx->seqStore); ++ const BYTE *const istart = (const BYTE *)src; ++ const BYTE *ip = istart; ++ const BYTE *anchor = istart; ++ const BYTE *const iend = istart + srcSize; ++ const BYTE *const ilimit = iend - 8; ++ const BYTE *const base = ctx->base; ++ const U32 lowestIndex = ctx->lowLimit; ++ const U32 dictLimit = ctx->dictLimit; ++ const BYTE *const prefixStart = base + dictLimit; ++ const BYTE *const dictBase = ctx->dictBase; ++ const BYTE *const dictEnd = dictBase + dictLimit; ++ ++ const U32 maxSearches = 1U << ctx->params.cParams.searchLog; ++ const U32 sufficient_len = ctx->params.cParams.targetLength; ++ const U32 mls = ctx->params.cParams.searchLength; ++ const U32 minMatch = (ctx->params.cParams.searchLength == 3) ? 3 : 4; ++ ++ ZSTD_optimal_t *opt = seqStorePtr->priceTable; ++ ZSTD_match_t *matches = seqStorePtr->matchTable; ++ const BYTE *inr; ++ ++ /* init */ ++ U32 offset, rep[ZSTD_REP_NUM]; ++ { ++ U32 i; ++ for (i = 0; i < ZSTD_REP_NUM; i++) ++ rep[i] = ctx->rep[i]; ++ } ++ ++ ctx->nextToUpdate3 = ctx->nextToUpdate; ++ ZSTD_rescaleFreqs(seqStorePtr, (const BYTE *)src, srcSize); ++ ip += (ip == prefixStart); ++ ++ /* Match Loop */ ++ while (ip < ilimit) { ++ U32 cur, match_num, last_pos, litlen, price; ++ U32 u, mlen, best_mlen, best_off, litLength; ++ U32 curr = (U32)(ip - base); ++ memset(opt, 0, sizeof(ZSTD_optimal_t)); ++ last_pos = 0; ++ opt[0].litlen = (U32)(ip - anchor); ++ ++ /* check repCode */ ++ { ++ U32 i, last_i = ZSTD_REP_CHECK + (ip == anchor); ++ for (i = (ip == anchor); i < last_i; i++) { ++ const S32 repCur = (i == ZSTD_REP_MOVE_OPT) ? (rep[0] - 1) : rep[i]; ++ const U32 repIndex = (U32)(curr - repCur); ++ const BYTE *const repBase = repIndex < dictLimit ? dictBase : base; ++ const BYTE *const repMatch = repBase + repIndex; ++ if ((repCur > 0 && repCur <= (S32)curr) && ++ (((U32)((dictLimit - 1) - repIndex) >= 3) & (repIndex > lowestIndex)) /* intentional overflow */ ++ && (ZSTD_readMINMATCH(ip, minMatch) == ZSTD_readMINMATCH(repMatch, minMatch))) { ++ /* repcode detected we should take it */ ++ const BYTE *const repEnd = repIndex < dictLimit ? dictEnd : iend; ++ mlen = (U32)ZSTD_count_2segments(ip + minMatch, repMatch + minMatch, iend, repEnd, prefixStart) + minMatch; ++ ++ if (mlen > sufficient_len || mlen >= ZSTD_OPT_NUM) { ++ best_mlen = mlen; ++ best_off = i; ++ cur = 0; ++ last_pos = 1; ++ goto _storeSequence; ++ } ++ ++ best_off = i - (ip == anchor); ++ litlen = opt[0].litlen; ++ do { ++ price = ZSTD_getPrice(seqStorePtr, litlen, anchor, best_off, mlen - MINMATCH, ultra); ++ if (mlen > last_pos || price < opt[mlen].price) ++ SET_PRICE(mlen, mlen, i, litlen, price); /* note : macro modifies last_pos */ ++ mlen--; ++ } while (mlen >= minMatch); ++ } ++ } ++ } ++ ++ match_num = ZSTD_BtGetAllMatches_selectMLS_extDict(ctx, ip, iend, maxSearches, mls, matches, minMatch); /* first search (depth 0) */ ++ ++ if (!last_pos && !match_num) { ++ ip++; ++ continue; ++ } ++ ++ { ++ U32 i; ++ for (i = 0; i < ZSTD_REP_NUM; i++) ++ opt[0].rep[i] = rep[i]; ++ } ++ opt[0].mlen = 1; ++ ++ if (match_num && (matches[match_num - 1].len > sufficient_len || matches[match_num - 1].len >= ZSTD_OPT_NUM)) { ++ best_mlen = matches[match_num - 1].len; ++ best_off = matches[match_num - 1].off; ++ cur = 0; ++ last_pos = 1; ++ goto _storeSequence; ++ } ++ ++ best_mlen = (last_pos) ? last_pos : minMatch; ++ ++ /* set prices using matches at position = 0 */ ++ for (u = 0; u < match_num; u++) { ++ mlen = (u > 0) ? matches[u - 1].len + 1 : best_mlen; ++ best_mlen = matches[u].len; ++ litlen = opt[0].litlen; ++ while (mlen <= best_mlen) { ++ price = ZSTD_getPrice(seqStorePtr, litlen, anchor, matches[u].off - 1, mlen - MINMATCH, ultra); ++ if (mlen > last_pos || price < opt[mlen].price) ++ SET_PRICE(mlen, mlen, matches[u].off, litlen, price); ++ mlen++; ++ } ++ } ++ ++ if (last_pos < minMatch) { ++ ip++; ++ continue; ++ } ++ ++ /* check further positions */ ++ for (cur = 1; cur <= last_pos; cur++) { ++ inr = ip + cur; ++ ++ if (opt[cur - 1].mlen == 1) { ++ litlen = opt[cur - 1].litlen + 1; ++ if (cur > litlen) { ++ price = opt[cur - litlen].price + ZSTD_getLiteralPrice(seqStorePtr, litlen, inr - litlen); ++ } else ++ price = ZSTD_getLiteralPrice(seqStorePtr, litlen, anchor); ++ } else { ++ litlen = 1; ++ price = opt[cur - 1].price + ZSTD_getLiteralPrice(seqStorePtr, litlen, inr - 1); ++ } ++ ++ if (cur > last_pos || price <= opt[cur].price) ++ SET_PRICE(cur, 1, 0, litlen, price); ++ ++ if (cur == last_pos) ++ break; ++ ++ if (inr > ilimit) /* last match must start at a minimum distance of 8 from oend */ ++ continue; ++ ++ mlen = opt[cur].mlen; ++ if (opt[cur].off > ZSTD_REP_MOVE_OPT) { ++ opt[cur].rep[2] = opt[cur - mlen].rep[1]; ++ opt[cur].rep[1] = opt[cur - mlen].rep[0]; ++ opt[cur].rep[0] = opt[cur].off - ZSTD_REP_MOVE_OPT; ++ } else { ++ opt[cur].rep[2] = (opt[cur].off > 1) ? opt[cur - mlen].rep[1] : opt[cur - mlen].rep[2]; ++ opt[cur].rep[1] = (opt[cur].off > 0) ? opt[cur - mlen].rep[0] : opt[cur - mlen].rep[1]; ++ opt[cur].rep[0] = ++ ((opt[cur].off == ZSTD_REP_MOVE_OPT) && (mlen != 1)) ? (opt[cur - mlen].rep[0] - 1) : (opt[cur - mlen].rep[opt[cur].off]); ++ } ++ ++ best_mlen = minMatch; ++ { ++ U32 i, last_i = ZSTD_REP_CHECK + (mlen != 1); ++ for (i = (mlen != 1); i < last_i; i++) { ++ const S32 repCur = (i == ZSTD_REP_MOVE_OPT) ? (opt[cur].rep[0] - 1) : opt[cur].rep[i]; ++ const U32 repIndex = (U32)(curr + cur - repCur); ++ const BYTE *const repBase = repIndex < dictLimit ? dictBase : base; ++ const BYTE *const repMatch = repBase + repIndex; ++ if ((repCur > 0 && repCur <= (S32)(curr + cur)) && ++ (((U32)((dictLimit - 1) - repIndex) >= 3) & (repIndex > lowestIndex)) /* intentional overflow */ ++ && (ZSTD_readMINMATCH(inr, minMatch) == ZSTD_readMINMATCH(repMatch, minMatch))) { ++ /* repcode detected */ ++ const BYTE *const repEnd = repIndex < dictLimit ? dictEnd : iend; ++ mlen = (U32)ZSTD_count_2segments(inr + minMatch, repMatch + minMatch, iend, repEnd, prefixStart) + minMatch; ++ ++ if (mlen > sufficient_len || cur + mlen >= ZSTD_OPT_NUM) { ++ best_mlen = mlen; ++ best_off = i; ++ last_pos = cur + 1; ++ goto _storeSequence; ++ } ++ ++ best_off = i - (opt[cur].mlen != 1); ++ if (mlen > best_mlen) ++ best_mlen = mlen; ++ ++ do { ++ if (opt[cur].mlen == 1) { ++ litlen = opt[cur].litlen; ++ if (cur > litlen) { ++ price = opt[cur - litlen].price + ZSTD_getPrice(seqStorePtr, litlen, inr - litlen, ++ best_off, mlen - MINMATCH, ultra); ++ } else ++ price = ZSTD_getPrice(seqStorePtr, litlen, anchor, best_off, mlen - MINMATCH, ultra); ++ } else { ++ litlen = 0; ++ price = opt[cur].price + ZSTD_getPrice(seqStorePtr, 0, NULL, best_off, mlen - MINMATCH, ultra); ++ } ++ ++ if (cur + mlen > last_pos || price <= opt[cur + mlen].price) ++ SET_PRICE(cur + mlen, mlen, i, litlen, price); ++ mlen--; ++ } while (mlen >= minMatch); ++ } ++ } ++ } ++ ++ match_num = ZSTD_BtGetAllMatches_selectMLS_extDict(ctx, inr, iend, maxSearches, mls, matches, minMatch); ++ ++ if (match_num > 0 && (matches[match_num - 1].len > sufficient_len || cur + matches[match_num - 1].len >= ZSTD_OPT_NUM)) { ++ best_mlen = matches[match_num - 1].len; ++ best_off = matches[match_num - 1].off; ++ last_pos = cur + 1; ++ goto _storeSequence; ++ } ++ ++ /* set prices using matches at position = cur */ ++ for (u = 0; u < match_num; u++) { ++ mlen = (u > 0) ? matches[u - 1].len + 1 : best_mlen; ++ best_mlen = matches[u].len; ++ ++ while (mlen <= best_mlen) { ++ if (opt[cur].mlen == 1) { ++ litlen = opt[cur].litlen; ++ if (cur > litlen) ++ price = opt[cur - litlen].price + ZSTD_getPrice(seqStorePtr, litlen, ip + cur - litlen, ++ matches[u].off - 1, mlen - MINMATCH, ultra); ++ else ++ price = ZSTD_getPrice(seqStorePtr, litlen, anchor, matches[u].off - 1, mlen - MINMATCH, ultra); ++ } else { ++ litlen = 0; ++ price = opt[cur].price + ZSTD_getPrice(seqStorePtr, 0, NULL, matches[u].off - 1, mlen - MINMATCH, ultra); ++ } ++ ++ if (cur + mlen > last_pos || (price < opt[cur + mlen].price)) ++ SET_PRICE(cur + mlen, mlen, matches[u].off, litlen, price); ++ ++ mlen++; ++ } ++ } ++ } /* for (cur = 1; cur <= last_pos; cur++) */ ++ ++ best_mlen = opt[last_pos].mlen; ++ best_off = opt[last_pos].off; ++ cur = last_pos - best_mlen; ++ ++ /* store sequence */ ++_storeSequence: /* cur, last_pos, best_mlen, best_off have to be set */ ++ opt[0].mlen = 1; ++ ++ while (1) { ++ mlen = opt[cur].mlen; ++ offset = opt[cur].off; ++ opt[cur].mlen = best_mlen; ++ opt[cur].off = best_off; ++ best_mlen = mlen; ++ best_off = offset; ++ if (mlen > cur) ++ break; ++ cur -= mlen; ++ } ++ ++ for (u = 0; u <= last_pos;) { ++ u += opt[u].mlen; ++ } ++ ++ for (cur = 0; cur < last_pos;) { ++ mlen = opt[cur].mlen; ++ if (mlen == 1) { ++ ip++; ++ cur++; ++ continue; ++ } ++ offset = opt[cur].off; ++ cur += mlen; ++ litLength = (U32)(ip - anchor); ++ ++ if (offset > ZSTD_REP_MOVE_OPT) { ++ rep[2] = rep[1]; ++ rep[1] = rep[0]; ++ rep[0] = offset - ZSTD_REP_MOVE_OPT; ++ offset--; ++ } else { ++ if (offset != 0) { ++ best_off = (offset == ZSTD_REP_MOVE_OPT) ? (rep[0] - 1) : (rep[offset]); ++ if (offset != 1) ++ rep[2] = rep[1]; ++ rep[1] = rep[0]; ++ rep[0] = best_off; ++ } ++ ++ if (litLength == 0) ++ offset--; ++ } ++ ++ ZSTD_updatePrice(seqStorePtr, litLength, anchor, offset, mlen - MINMATCH); ++ ZSTD_storeSeq(seqStorePtr, litLength, anchor, offset, mlen - MINMATCH); ++ anchor = ip = ip + mlen; ++ } ++ } /* for (cur=0; cur < last_pos; ) */ ++ ++ /* Save reps for next block */ ++ { ++ int i; ++ for (i = 0; i < ZSTD_REP_NUM; i++) ++ ctx->repToConfirm[i] = rep[i]; ++ } ++ ++ /* Last Literals */ ++ { ++ size_t lastLLSize = iend - anchor; ++ memcpy(seqStorePtr->lit, anchor, lastLLSize); ++ seqStorePtr->lit += lastLLSize; ++ } ++} ++ ++#endif /* ZSTD_OPT_H_91842398743 */ +diff --git a/xen/include/xen/decompress.h b/xen/include/xen/decompress.h +index b2955faa4b..f5bc17f2b6 100644 +--- a/xen/include/xen/decompress.h ++++ b/xen/include/xen/decompress.h +@@ -31,7 +31,7 @@ typedef int decompress_fn(unsigned char *inbuf, unsigned int len, + * dependent). + */ + +-decompress_fn bunzip2, unxz, unlzma, unlzo, unlz4; ++decompress_fn bunzip2, unxz, unlzma, unlzo, unlz4, unzstd; + + int decompress(void *inbuf, unsigned int len, void *outbuf); + +diff --git a/xen/include/xen/xxhash.h b/xen/include/xen/xxhash.h +new file mode 100644 +index 0000000000..13ddc616d1 +--- /dev/null ++++ b/xen/include/xen/xxhash.h +@@ -0,0 +1,259 @@ ++/* ++ * xxHash - Extremely Fast Hash algorithm ++ * Copyright (C) 2012-2016, Yann Collet. ++ * ++ * BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions are ++ * met: ++ * ++ * * Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * * Redistributions in binary form must reproduce the above ++ * copyright notice, this list of conditions and the following disclaimer ++ * in the documentation and/or other materials provided with the ++ * distribution. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS ++ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT ++ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR ++ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT ++ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, ++ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT ++ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE ++ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++ * ++ * This program is free software; you can redistribute it and/or modify it under ++ * the terms of the GNU General Public License version 2 as published by the ++ * Free Software Foundation. This program is dual-licensed; you may select ++ * either version 2 of the GNU General Public License ("GPL") or BSD license ++ * ("BSD"). ++ * ++ * You can contact the author at: ++ * - xxHash homepage: https://cyan4973.github.io/xxHash/ ++ * - xxHash source repository: https://github.com/Cyan4973/xxHash ++ */ ++ ++/* ++ * Notice extracted from xxHash homepage: ++ * ++ * xxHash is an extremely fast Hash algorithm, running at RAM speed limits. ++ * It also successfully passes all tests from the SMHasher suite. ++ * ++ * Comparison (single thread, Windows Seven 32 bits, using SMHasher on a Core 2 ++ * Duo @3GHz) ++ * ++ * Name Speed Q.Score Author ++ * xxHash 5.4 GB/s 10 ++ * CrapWow 3.2 GB/s 2 Andrew ++ * MumurHash 3a 2.7 GB/s 10 Austin Appleby ++ * SpookyHash 2.0 GB/s 10 Bob Jenkins ++ * SBox 1.4 GB/s 9 Bret Mulvey ++ * Lookup3 1.2 GB/s 9 Bob Jenkins ++ * SuperFastHash 1.2 GB/s 1 Paul Hsieh ++ * CityHash64 1.05 GB/s 10 Pike & Alakuijala ++ * FNV 0.55 GB/s 5 Fowler, Noll, Vo ++ * CRC32 0.43 GB/s 9 ++ * MD5-32 0.33 GB/s 10 Ronald L. Rivest ++ * SHA1-32 0.28 GB/s 10 ++ * ++ * Q.Score is a measure of quality of the hash function. ++ * It depends on successfully passing SMHasher test set. ++ * 10 is a perfect score. ++ * ++ * A 64-bits version, named xxh64 offers much better speed, ++ * but for 64-bits applications only. ++ * Name Speed on 64 bits Speed on 32 bits ++ * xxh64 13.8 GB/s 1.9 GB/s ++ * xxh32 6.8 GB/s 6.0 GB/s ++ */ ++ ++#ifndef XXHASH_H ++#define XXHASH_H ++ ++#include ++ ++/*-**************************** ++ * Simple Hash Functions ++ *****************************/ ++ ++/** ++ * xxh32() - calculate the 32-bit hash of the input with a given seed. ++ * ++ * @input: The data to hash. ++ * @length: The length of the data to hash. ++ * @seed: The seed can be used to alter the result predictably. ++ * ++ * Speed on Core 2 Duo @ 3 GHz (single thread, SMHasher benchmark) : 5.4 GB/s ++ * ++ * Return: The 32-bit hash of the data. ++ */ ++uint32_t xxh32(const void *input, size_t length, uint32_t seed); ++ ++/** ++ * xxh64() - calculate the 64-bit hash of the input with a given seed. ++ * ++ * @input: The data to hash. ++ * @length: The length of the data to hash. ++ * @seed: The seed can be used to alter the result predictably. ++ * ++ * This function runs 2x faster on 64-bit systems, but slower on 32-bit systems. ++ * ++ * Return: The 64-bit hash of the data. ++ */ ++uint64_t xxh64(const void *input, size_t length, uint64_t seed); ++ ++/** ++ * xxhash() - calculate wordsize hash of the input with a given seed ++ * @input: The data to hash. ++ * @length: The length of the data to hash. ++ * @seed: The seed can be used to alter the result predictably. ++ * ++ * If the hash does not need to be comparable between machines with ++ * different word sizes, this function will call whichever of xxh32() ++ * or xxh64() is faster. ++ * ++ * Return: wordsize hash of the data. ++ */ ++ ++static inline unsigned long xxhash(const void *input, size_t length, ++ uint64_t seed) ++{ ++#if BITS_PER_LONG == 64 ++ return xxh64(input, length, seed); ++#else ++ return xxh32(input, length, seed); ++#endif ++} ++ ++/*-**************************** ++ * Streaming Hash Functions ++ *****************************/ ++ ++/* ++ * These definitions are only meant to allow allocation of XXH state ++ * statically, on stack, or in a struct for example. ++ * Do not use members directly. ++ */ ++ ++/** ++ * struct xxh32_state - private xxh32 state, do not use members directly ++ */ ++struct xxh32_state { ++ uint32_t total_len_32; ++ uint32_t large_len; ++ uint32_t v1; ++ uint32_t v2; ++ uint32_t v3; ++ uint32_t v4; ++ uint32_t mem32[4]; ++ uint32_t memsize; ++}; ++ ++/** ++ * struct xxh32_state - private xxh64 state, do not use members directly ++ */ ++struct xxh64_state { ++ uint64_t total_len; ++ uint64_t v1; ++ uint64_t v2; ++ uint64_t v3; ++ uint64_t v4; ++ uint64_t mem64[4]; ++ uint32_t memsize; ++}; ++ ++/** ++ * xxh32_reset() - reset the xxh32 state to start a new hashing operation ++ * ++ * @state: The xxh32 state to reset. ++ * @seed: Initialize the hash state with this seed. ++ * ++ * Call this function on any xxh32_state to prepare for a new hashing operation. ++ */ ++void xxh32_reset(struct xxh32_state *state, uint32_t seed); ++ ++/** ++ * xxh32_update() - hash the data given and update the xxh32 state ++ * ++ * @state: The xxh32 state to update. ++ * @input: The data to hash. ++ * @length: The length of the data to hash. ++ * ++ * After calling xxh32_reset() call xxh32_update() as many times as necessary. ++ * ++ * Return: Zero on success, otherwise an error code. ++ */ ++int xxh32_update(struct xxh32_state *state, const void *input, size_t length); ++ ++/** ++ * xxh32_digest() - produce the current xxh32 hash ++ * ++ * @state: Produce the current xxh32 hash of this state. ++ * ++ * A hash value can be produced at any time. It is still possible to continue ++ * inserting input into the hash state after a call to xxh32_digest(), and ++ * generate new hashes later on, by calling xxh32_digest() again. ++ * ++ * Return: The xxh32 hash stored in the state. ++ */ ++uint32_t xxh32_digest(const struct xxh32_state *state); ++ ++/** ++ * xxh64_reset() - reset the xxh64 state to start a new hashing operation ++ * ++ * @state: The xxh64 state to reset. ++ * @seed: Initialize the hash state with this seed. ++ */ ++void xxh64_reset(struct xxh64_state *state, uint64_t seed); ++ ++/** ++ * xxh64_update() - hash the data given and update the xxh64 state ++ * @state: The xxh64 state to update. ++ * @input: The data to hash. ++ * @length: The length of the data to hash. ++ * ++ * After calling xxh64_reset() call xxh64_update() as many times as necessary. ++ * ++ * Return: Zero on success, otherwise an error code. ++ */ ++int xxh64_update(struct xxh64_state *state, const void *input, size_t length); ++ ++/** ++ * xxh64_digest() - produce the current xxh64 hash ++ * ++ * @state: Produce the current xxh64 hash of this state. ++ * ++ * A hash value can be produced at any time. It is still possible to continue ++ * inserting input into the hash state after a call to xxh64_digest(), and ++ * generate new hashes later on, by calling xxh64_digest() again. ++ * ++ * Return: The xxh64 hash stored in the state. ++ */ ++uint64_t xxh64_digest(const struct xxh64_state *state); ++ ++/*-************************** ++ * Utils ++ ***************************/ ++ ++/** ++ * xxh32_copy_state() - copy the source state into the destination state ++ * ++ * @src: The source xxh32 state. ++ * @dst: The destination xxh32 state. ++ */ ++void xxh32_copy_state(struct xxh32_state *dst, const struct xxh32_state *src); ++ ++/** ++ * xxh64_copy_state() - copy the source state into the destination state ++ * ++ * @src: The source xxh64 state. ++ * @dst: The destination xxh64 state. ++ */ ++void xxh64_copy_state(struct xxh64_state *dst, const struct xxh64_state *src); ++ ++#endif /* XXHASH_H */ +diff --git a/xen/include/xen/zstd.h b/xen/include/xen/zstd.h +new file mode 100644 +index 0000000000..eb33582a18 +--- /dev/null ++++ b/xen/include/xen/zstd.h +@@ -0,0 +1,1157 @@ ++/* ++ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc. ++ * All rights reserved. ++ * ++ * This source code is licensed under the BSD-style license found in the ++ * LICENSE file in the root directory of https://github.com/facebook/zstd. ++ * An additional grant of patent rights can be found in the PATENTS file in the ++ * same directory. ++ * ++ * This program is free software; you can redistribute it and/or modify it under ++ * the terms of the GNU General Public License version 2 as published by the ++ * Free Software Foundation. This program is dual-licensed; you may select ++ * either version 2 of the GNU General Public License ("GPL") or BSD license ++ * ("BSD"). ++ */ ++ ++#ifndef ZSTD_H ++#define ZSTD_H ++ ++/* ====== Dependency ======*/ ++#include /* size_t */ ++ ++ ++/*-***************************************************************************** ++ * Introduction ++ * ++ * zstd, short for Zstandard, is a fast lossless compression algorithm, ++ * targeting real-time compression scenarios at zlib-level and better ++ * compression ratios. The zstd compression library provides in-memory ++ * compression and decompression functions. The library supports compression ++ * levels from 1 up to ZSTD_maxCLevel() which is 22. Levels >= 20, labeled ++ * ultra, should be used with caution, as they require more memory. ++ * Compression can be done in: ++ * - a single step, reusing a context (described as Explicit memory management) ++ * - unbounded multiple steps (described as Streaming compression) ++ * The compression ratio achievable on small data can be highly improved using ++ * compression with a dictionary in: ++ * - a single step (described as Simple dictionary API) ++ * - a single step, reusing a dictionary (described as Fast dictionary API) ++ ******************************************************************************/ ++ ++/*====== Helper functions ======*/ ++ ++/** ++ * enum ZSTD_ErrorCode - zstd error codes ++ * ++ * Functions that return size_t can be checked for errors using ZSTD_isError() ++ * and the ZSTD_ErrorCode can be extracted using ZSTD_getErrorCode(). ++ */ ++typedef enum { ++ ZSTD_error_no_error, ++ ZSTD_error_GENERIC, ++ ZSTD_error_prefix_unknown, ++ ZSTD_error_version_unsupported, ++ ZSTD_error_parameter_unknown, ++ ZSTD_error_frameParameter_unsupported, ++ ZSTD_error_frameParameter_unsupportedBy32bits, ++ ZSTD_error_frameParameter_windowTooLarge, ++ ZSTD_error_compressionParameter_unsupported, ++ ZSTD_error_init_missing, ++ ZSTD_error_memory_allocation, ++ ZSTD_error_stage_wrong, ++ ZSTD_error_dstSize_tooSmall, ++ ZSTD_error_srcSize_wrong, ++ ZSTD_error_corruption_detected, ++ ZSTD_error_checksum_wrong, ++ ZSTD_error_tableLog_tooLarge, ++ ZSTD_error_maxSymbolValue_tooLarge, ++ ZSTD_error_maxSymbolValue_tooSmall, ++ ZSTD_error_dictionary_corrupted, ++ ZSTD_error_dictionary_wrong, ++ ZSTD_error_dictionaryCreation_failed, ++ ZSTD_error_maxCode ++} ZSTD_ErrorCode; ++ ++/** ++ * ZSTD_maxCLevel() - maximum compression level available ++ * ++ * Return: Maximum compression level available. ++ */ ++int ZSTD_maxCLevel(void); ++/** ++ * ZSTD_compressBound() - maximum compressed size in worst case scenario ++ * @srcSize: The size of the data to compress. ++ * ++ * Return: The maximum compressed size in the worst case scenario. ++ */ ++size_t ZSTD_compressBound(size_t srcSize); ++/** ++ * ZSTD_isError() - tells if a size_t function result is an error code ++ * @code: The function result to check for error. ++ * ++ * Return: Non-zero iff the code is an error. ++ */ ++static __attribute__((unused)) unsigned int ZSTD_isError(size_t code) ++{ ++ return code > (size_t)-ZSTD_error_maxCode; ++} ++/** ++ * ZSTD_getErrorCode() - translates an error function result to a ZSTD_ErrorCode ++ * @functionResult: The result of a function for which ZSTD_isError() is true. ++ * ++ * Return: The ZSTD_ErrorCode corresponding to the functionResult or 0 ++ * if the functionResult isn't an error. ++ */ ++static __attribute__((unused)) ZSTD_ErrorCode ZSTD_getErrorCode( ++ size_t functionResult) ++{ ++ if (!ZSTD_isError(functionResult)) ++ return (ZSTD_ErrorCode)0; ++ return (ZSTD_ErrorCode)(0 - functionResult); ++} ++ ++/** ++ * enum ZSTD_strategy - zstd compression search strategy ++ * ++ * From faster to stronger. ++ */ ++typedef enum { ++ ZSTD_fast, ++ ZSTD_dfast, ++ ZSTD_greedy, ++ ZSTD_lazy, ++ ZSTD_lazy2, ++ ZSTD_btlazy2, ++ ZSTD_btopt, ++ ZSTD_btopt2 ++} ZSTD_strategy; ++ ++/** ++ * struct ZSTD_compressionParameters - zstd compression parameters ++ * @windowLog: Log of the largest match distance. Larger means more ++ * compression, and more memory needed during decompression. ++ * @chainLog: Fully searched segment. Larger means more compression, slower, ++ * and more memory (useless for fast). ++ * @hashLog: Dispatch table. Larger means more compression, ++ * slower, and more memory. ++ * @searchLog: Number of searches. Larger means more compression and slower. ++ * @searchLength: Match length searched. Larger means faster decompression, ++ * sometimes less compression. ++ * @targetLength: Acceptable match size for optimal parser (only). Larger means ++ * more compression, and slower. ++ * @strategy: The zstd compression strategy. ++ */ ++typedef struct { ++ unsigned int windowLog; ++ unsigned int chainLog; ++ unsigned int hashLog; ++ unsigned int searchLog; ++ unsigned int searchLength; ++ unsigned int targetLength; ++ ZSTD_strategy strategy; ++} ZSTD_compressionParameters; ++ ++/** ++ * struct ZSTD_frameParameters - zstd frame parameters ++ * @contentSizeFlag: Controls whether content size will be present in the frame ++ * header (when known). ++ * @checksumFlag: Controls whether a 32-bit checksum is generated at the end ++ * of the frame for error detection. ++ * @noDictIDFlag: Controls whether dictID will be saved into the frame header ++ * when using dictionary compression. ++ * ++ * The default value is all fields set to 0. ++ */ ++typedef struct { ++ unsigned int contentSizeFlag; ++ unsigned int checksumFlag; ++ unsigned int noDictIDFlag; ++} ZSTD_frameParameters; ++ ++/** ++ * struct ZSTD_parameters - zstd parameters ++ * @cParams: The compression parameters. ++ * @fParams: The frame parameters. ++ */ ++typedef struct { ++ ZSTD_compressionParameters cParams; ++ ZSTD_frameParameters fParams; ++} ZSTD_parameters; ++ ++/** ++ * ZSTD_getCParams() - returns ZSTD_compressionParameters for selected level ++ * @compressionLevel: The compression level from 1 to ZSTD_maxCLevel(). ++ * @estimatedSrcSize: The estimated source size to compress or 0 if unknown. ++ * @dictSize: The dictionary size or 0 if a dictionary isn't being used. ++ * ++ * Return: The selected ZSTD_compressionParameters. ++ */ ++ZSTD_compressionParameters ZSTD_getCParams(int compressionLevel, ++ unsigned long long estimatedSrcSize, size_t dictSize); ++ ++/** ++ * ZSTD_getParams() - returns ZSTD_parameters for selected level ++ * @compressionLevel: The compression level from 1 to ZSTD_maxCLevel(). ++ * @estimatedSrcSize: The estimated source size to compress or 0 if unknown. ++ * @dictSize: The dictionary size or 0 if a dictionary isn't being used. ++ * ++ * The same as ZSTD_getCParams() except also selects the default frame ++ * parameters (all zero). ++ * ++ * Return: The selected ZSTD_parameters. ++ */ ++ZSTD_parameters ZSTD_getParams(int compressionLevel, ++ unsigned long long estimatedSrcSize, size_t dictSize); ++ ++/*-************************************* ++ * Explicit memory management ++ **************************************/ ++ ++/** ++ * ZSTD_CCtxWorkspaceBound() - amount of memory needed to initialize a ZSTD_CCtx ++ * @cParams: The compression parameters to be used for compression. ++ * ++ * If multiple compression parameters might be used, the caller must call ++ * ZSTD_CCtxWorkspaceBound() for each set of parameters and use the maximum ++ * size. ++ * ++ * Return: A lower bound on the size of the workspace that is passed to ++ * ZSTD_initCCtx(). ++ */ ++size_t ZSTD_CCtxWorkspaceBound(ZSTD_compressionParameters cParams); ++ ++/** ++ * struct ZSTD_CCtx - the zstd compression context ++ * ++ * When compressing many times it is recommended to allocate a context just once ++ * and reuse it for each successive compression operation. ++ */ ++typedef struct ZSTD_CCtx_s ZSTD_CCtx; ++/** ++ * ZSTD_initCCtx() - initialize a zstd compression context ++ * @workspace: The workspace to emplace the context into. It must outlive ++ * the returned context. ++ * @workspaceSize: The size of workspace. Use ZSTD_CCtxWorkspaceBound() to ++ * determine how large the workspace must be. ++ * ++ * Return: A compression context emplaced into workspace. ++ */ ++ZSTD_CCtx *ZSTD_initCCtx(void *workspace, size_t workspaceSize); ++ ++/** ++ * ZSTD_compressCCtx() - compress src into dst ++ * @ctx: The context. Must have been initialized with a workspace at ++ * least as large as ZSTD_CCtxWorkspaceBound(params.cParams). ++ * @dst: The buffer to compress src into. ++ * @dstCapacity: The size of the destination buffer. May be any size, but ++ * ZSTD_compressBound(srcSize) is guaranteed to be large enough. ++ * @src: The data to compress. ++ * @srcSize: The size of the data to compress. ++ * @params: The parameters to use for compression. See ZSTD_getParams(). ++ * ++ * Return: The compressed size or an error, which can be checked using ++ * ZSTD_isError(). ++ */ ++size_t ZSTD_compressCCtx(ZSTD_CCtx *ctx, void *dst, size_t dstCapacity, ++ const void *src, size_t srcSize, ZSTD_parameters params); ++ ++/** ++ * ZSTD_DCtxWorkspaceBound() - amount of memory needed to initialize a ZSTD_DCtx ++ * ++ * Return: A lower bound on the size of the workspace that is passed to ++ * ZSTD_initDCtx(). ++ */ ++size_t ZSTD_DCtxWorkspaceBound(void); ++ ++/** ++ * struct ZSTD_DCtx - the zstd decompression context ++ * ++ * When decompressing many times it is recommended to allocate a context just ++ * once and reuse it for each successive decompression operation. ++ */ ++typedef struct ZSTD_DCtx_s ZSTD_DCtx; ++/** ++ * ZSTD_initDCtx() - initialize a zstd decompression context ++ * @workspace: The workspace to emplace the context into. It must outlive ++ * the returned context. ++ * @workspaceSize: The size of workspace. Use ZSTD_DCtxWorkspaceBound() to ++ * determine how large the workspace must be. ++ * ++ * Return: A decompression context emplaced into workspace. ++ */ ++ZSTD_DCtx *ZSTD_initDCtx(void *workspace, size_t workspaceSize); ++ ++/** ++ * ZSTD_decompressDCtx() - decompress zstd compressed src into dst ++ * @ctx: The decompression context. ++ * @dst: The buffer to decompress src into. ++ * @dstCapacity: The size of the destination buffer. Must be at least as large ++ * as the decompressed size. If the caller cannot upper bound the ++ * decompressed size, then it's better to use the streaming API. ++ * @src: The zstd compressed data to decompress. Multiple concatenated ++ * frames and skippable frames are allowed. ++ * @srcSize: The exact size of the data to decompress. ++ * ++ * Return: The decompressed size or an error, which can be checked using ++ * ZSTD_isError(). ++ */ ++size_t ZSTD_decompressDCtx(ZSTD_DCtx *ctx, void *dst, size_t dstCapacity, ++ const void *src, size_t srcSize); ++ ++/*-************************ ++ * Simple dictionary API ++ **************************/ ++ ++/** ++ * ZSTD_compress_usingDict() - compress src into dst using a dictionary ++ * @ctx: The context. Must have been initialized with a workspace at ++ * least as large as ZSTD_CCtxWorkspaceBound(params.cParams). ++ * @dst: The buffer to compress src into. ++ * @dstCapacity: The size of the destination buffer. May be any size, but ++ * ZSTD_compressBound(srcSize) is guaranteed to be large enough. ++ * @src: The data to compress. ++ * @srcSize: The size of the data to compress. ++ * @dict: The dictionary to use for compression. ++ * @dictSize: The size of the dictionary. ++ * @params: The parameters to use for compression. See ZSTD_getParams(). ++ * ++ * Compression using a predefined dictionary. The same dictionary must be used ++ * during decompression. ++ * ++ * Return: The compressed size or an error, which can be checked using ++ * ZSTD_isError(). ++ */ ++size_t ZSTD_compress_usingDict(ZSTD_CCtx *ctx, void *dst, size_t dstCapacity, ++ const void *src, size_t srcSize, const void *dict, size_t dictSize, ++ ZSTD_parameters params); ++ ++/** ++ * ZSTD_decompress_usingDict() - decompress src into dst using a dictionary ++ * @ctx: The decompression context. ++ * @dst: The buffer to decompress src into. ++ * @dstCapacity: The size of the destination buffer. Must be at least as large ++ * as the decompressed size. If the caller cannot upper bound the ++ * decompressed size, then it's better to use the streaming API. ++ * @src: The zstd compressed data to decompress. Multiple concatenated ++ * frames and skippable frames are allowed. ++ * @srcSize: The exact size of the data to decompress. ++ * @dict: The dictionary to use for decompression. The same dictionary ++ * must've been used to compress the data. ++ * @dictSize: The size of the dictionary. ++ * ++ * Return: The decompressed size or an error, which can be checked using ++ * ZSTD_isError(). ++ */ ++size_t ZSTD_decompress_usingDict(ZSTD_DCtx *ctx, void *dst, size_t dstCapacity, ++ const void *src, size_t srcSize, const void *dict, size_t dictSize); ++ ++/*-************************** ++ * Fast dictionary API ++ ***************************/ ++ ++/** ++ * ZSTD_CDictWorkspaceBound() - memory needed to initialize a ZSTD_CDict ++ * @cParams: The compression parameters to be used for compression. ++ * ++ * Return: A lower bound on the size of the workspace that is passed to ++ * ZSTD_initCDict(). ++ */ ++size_t ZSTD_CDictWorkspaceBound(ZSTD_compressionParameters cParams); ++ ++/** ++ * struct ZSTD_CDict - a digested dictionary to be used for compression ++ */ ++typedef struct ZSTD_CDict_s ZSTD_CDict; ++ ++/** ++ * ZSTD_initCDict() - initialize a digested dictionary for compression ++ * @dictBuffer: The dictionary to digest. The buffer is referenced by the ++ * ZSTD_CDict so it must outlive the returned ZSTD_CDict. ++ * @dictSize: The size of the dictionary. ++ * @params: The parameters to use for compression. See ZSTD_getParams(). ++ * @workspace: The workspace. It must outlive the returned ZSTD_CDict. ++ * @workspaceSize: The workspace size. Must be at least ++ * ZSTD_CDictWorkspaceBound(params.cParams). ++ * ++ * When compressing multiple messages / blocks with the same dictionary it is ++ * recommended to load it just once. The ZSTD_CDict merely references the ++ * dictBuffer, so it must outlive the returned ZSTD_CDict. ++ * ++ * Return: The digested dictionary emplaced into workspace. ++ */ ++ZSTD_CDict *ZSTD_initCDict(const void *dictBuffer, size_t dictSize, ++ ZSTD_parameters params, void *workspace, size_t workspaceSize); ++ ++/** ++ * ZSTD_compress_usingCDict() - compress src into dst using a ZSTD_CDict ++ * @ctx: The context. Must have been initialized with a workspace at ++ * least as large as ZSTD_CCtxWorkspaceBound(cParams) where ++ * cParams are the compression parameters used to initialize the ++ * cdict. ++ * @dst: The buffer to compress src into. ++ * @dstCapacity: The size of the destination buffer. May be any size, but ++ * ZSTD_compressBound(srcSize) is guaranteed to be large enough. ++ * @src: The data to compress. ++ * @srcSize: The size of the data to compress. ++ * @cdict: The digested dictionary to use for compression. ++ * @params: The parameters to use for compression. See ZSTD_getParams(). ++ * ++ * Compression using a digested dictionary. The same dictionary must be used ++ * during decompression. ++ * ++ * Return: The compressed size or an error, which can be checked using ++ * ZSTD_isError(). ++ */ ++size_t ZSTD_compress_usingCDict(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity, ++ const void *src, size_t srcSize, const ZSTD_CDict *cdict); ++ ++ ++/** ++ * ZSTD_DDictWorkspaceBound() - memory needed to initialize a ZSTD_DDict ++ * ++ * Return: A lower bound on the size of the workspace that is passed to ++ * ZSTD_initDDict(). ++ */ ++size_t ZSTD_DDictWorkspaceBound(void); ++ ++/** ++ * struct ZSTD_DDict - a digested dictionary to be used for decompression ++ */ ++typedef struct ZSTD_DDict_s ZSTD_DDict; ++ ++/** ++ * ZSTD_initDDict() - initialize a digested dictionary for decompression ++ * @dictBuffer: The dictionary to digest. The buffer is referenced by the ++ * ZSTD_DDict so it must outlive the returned ZSTD_DDict. ++ * @dictSize: The size of the dictionary. ++ * @workspace: The workspace. It must outlive the returned ZSTD_DDict. ++ * @workspaceSize: The workspace size. Must be at least ++ * ZSTD_DDictWorkspaceBound(). ++ * ++ * When decompressing multiple messages / blocks with the same dictionary it is ++ * recommended to load it just once. The ZSTD_DDict merely references the ++ * dictBuffer, so it must outlive the returned ZSTD_DDict. ++ * ++ * Return: The digested dictionary emplaced into workspace. ++ */ ++ZSTD_DDict *ZSTD_initDDict(const void *dictBuffer, size_t dictSize, ++ void *workspace, size_t workspaceSize); ++ ++/** ++ * ZSTD_decompress_usingDDict() - decompress src into dst using a ZSTD_DDict ++ * @ctx: The decompression context. ++ * @dst: The buffer to decompress src into. ++ * @dstCapacity: The size of the destination buffer. Must be at least as large ++ * as the decompressed size. If the caller cannot upper bound the ++ * decompressed size, then it's better to use the streaming API. ++ * @src: The zstd compressed data to decompress. Multiple concatenated ++ * frames and skippable frames are allowed. ++ * @srcSize: The exact size of the data to decompress. ++ * @ddict: The digested dictionary to use for decompression. The same ++ * dictionary must've been used to compress the data. ++ * ++ * Return: The decompressed size or an error, which can be checked using ++ * ZSTD_isError(). ++ */ ++size_t ZSTD_decompress_usingDDict(ZSTD_DCtx *dctx, void *dst, ++ size_t dstCapacity, const void *src, size_t srcSize, ++ const ZSTD_DDict *ddict); ++ ++ ++/*-************************** ++ * Streaming ++ ***************************/ ++ ++/** ++ * struct ZSTD_inBuffer - input buffer for streaming ++ * @src: Start of the input buffer. ++ * @size: Size of the input buffer. ++ * @pos: Position where reading stopped. Will be updated. ++ * Necessarily 0 <= pos <= size. ++ */ ++typedef struct ZSTD_inBuffer_s { ++ const void *src; ++ size_t size; ++ size_t pos; ++} ZSTD_inBuffer; ++ ++/** ++ * struct ZSTD_outBuffer - output buffer for streaming ++ * @dst: Start of the output buffer. ++ * @size: Size of the output buffer. ++ * @pos: Position where writing stopped. Will be updated. ++ * Necessarily 0 <= pos <= size. ++ */ ++typedef struct ZSTD_outBuffer_s { ++ void *dst; ++ size_t size; ++ size_t pos; ++} ZSTD_outBuffer; ++ ++ ++ ++/*-***************************************************************************** ++ * Streaming compression - HowTo ++ * ++ * A ZSTD_CStream object is required to track streaming operation. ++ * Use ZSTD_initCStream() to initialize a ZSTD_CStream object. ++ * ZSTD_CStream objects can be reused multiple times on consecutive compression ++ * operations. It is recommended to re-use ZSTD_CStream in situations where many ++ * streaming operations will be achieved consecutively. Use one separate ++ * ZSTD_CStream per thread for parallel execution. ++ * ++ * Use ZSTD_compressStream() repetitively to consume input stream. ++ * The function will automatically update both `pos` fields. ++ * Note that it may not consume the entire input, in which case `pos < size`, ++ * and it's up to the caller to present again remaining data. ++ * It returns a hint for the preferred number of bytes to use as an input for ++ * the next function call. ++ * ++ * At any moment, it's possible to flush whatever data remains within internal ++ * buffer, using ZSTD_flushStream(). `output->pos` will be updated. There might ++ * still be some content left within the internal buffer if `output->size` is ++ * too small. It returns the number of bytes left in the internal buffer and ++ * must be called until it returns 0. ++ * ++ * ZSTD_endStream() instructs to finish a frame. It will perform a flush and ++ * write frame epilogue. The epilogue is required for decoders to consider a ++ * frame completed. Similar to ZSTD_flushStream(), it may not be able to flush ++ * the full content if `output->size` is too small. In which case, call again ++ * ZSTD_endStream() to complete the flush. It returns the number of bytes left ++ * in the internal buffer and must be called until it returns 0. ++ ******************************************************************************/ ++ ++/** ++ * ZSTD_CStreamWorkspaceBound() - memory needed to initialize a ZSTD_CStream ++ * @cParams: The compression parameters to be used for compression. ++ * ++ * Return: A lower bound on the size of the workspace that is passed to ++ * ZSTD_initCStream() and ZSTD_initCStream_usingCDict(). ++ */ ++size_t ZSTD_CStreamWorkspaceBound(ZSTD_compressionParameters cParams); ++ ++/** ++ * struct ZSTD_CStream - the zstd streaming compression context ++ */ ++typedef struct ZSTD_CStream_s ZSTD_CStream; ++ ++/*===== ZSTD_CStream management functions =====*/ ++/** ++ * ZSTD_initCStream() - initialize a zstd streaming compression context ++ * @params: The zstd compression parameters. ++ * @pledgedSrcSize: If params.fParams.contentSizeFlag == 1 then the caller must ++ * pass the source size (zero means empty source). Otherwise, ++ * the caller may optionally pass the source size, or zero if ++ * unknown. ++ * @workspace: The workspace to emplace the context into. It must outlive ++ * the returned context. ++ * @workspaceSize: The size of workspace. ++ * Use ZSTD_CStreamWorkspaceBound(params.cParams) to determine ++ * how large the workspace must be. ++ * ++ * Return: The zstd streaming compression context. ++ */ ++ZSTD_CStream *ZSTD_initCStream(ZSTD_parameters params, ++ unsigned long long pledgedSrcSize, void *workspace, ++ size_t workspaceSize); ++ ++/** ++ * ZSTD_initCStream_usingCDict() - initialize a streaming compression context ++ * @cdict: The digested dictionary to use for compression. ++ * @pledgedSrcSize: Optionally the source size, or zero if unknown. ++ * @workspace: The workspace to emplace the context into. It must outlive ++ * the returned context. ++ * @workspaceSize: The size of workspace. Call ZSTD_CStreamWorkspaceBound() ++ * with the cParams used to initialize the cdict to determine ++ * how large the workspace must be. ++ * ++ * Return: The zstd streaming compression context. ++ */ ++ZSTD_CStream *ZSTD_initCStream_usingCDict(const ZSTD_CDict *cdict, ++ unsigned long long pledgedSrcSize, void *workspace, ++ size_t workspaceSize); ++ ++/*===== Streaming compression functions =====*/ ++/** ++ * ZSTD_resetCStream() - reset the context using parameters from creation ++ * @zcs: The zstd streaming compression context to reset. ++ * @pledgedSrcSize: Optionally the source size, or zero if unknown. ++ * ++ * Resets the context using the parameters from creation. Skips dictionary ++ * loading, since it can be reused. If `pledgedSrcSize` is non-zero the frame ++ * content size is always written into the frame header. ++ * ++ * Return: Zero or an error, which can be checked using ZSTD_isError(). ++ */ ++size_t ZSTD_resetCStream(ZSTD_CStream *zcs, unsigned long long pledgedSrcSize); ++/** ++ * ZSTD_compressStream() - streaming compress some of input into output ++ * @zcs: The zstd streaming compression context. ++ * @output: Destination buffer. `output->pos` is updated to indicate how much ++ * compressed data was written. ++ * @input: Source buffer. `input->pos` is updated to indicate how much data was ++ * read. Note that it may not consume the entire input, in which case ++ * `input->pos < input->size`, and it's up to the caller to present ++ * remaining data again. ++ * ++ * The `input` and `output` buffers may be any size. Guaranteed to make some ++ * forward progress if `input` and `output` are not empty. ++ * ++ * Return: A hint for the number of bytes to use as the input for the next ++ * function call or an error, which can be checked using ++ * ZSTD_isError(). ++ */ ++size_t ZSTD_compressStream(ZSTD_CStream *zcs, ZSTD_outBuffer *output, ++ ZSTD_inBuffer *input); ++/** ++ * ZSTD_flushStream() - flush internal buffers into output ++ * @zcs: The zstd streaming compression context. ++ * @output: Destination buffer. `output->pos` is updated to indicate how much ++ * compressed data was written. ++ * ++ * ZSTD_flushStream() must be called until it returns 0, meaning all the data ++ * has been flushed. Since ZSTD_flushStream() causes a block to be ended, ++ * calling it too often will degrade the compression ratio. ++ * ++ * Return: The number of bytes still present within internal buffers or an ++ * error, which can be checked using ZSTD_isError(). ++ */ ++size_t ZSTD_flushStream(ZSTD_CStream *zcs, ZSTD_outBuffer *output); ++/** ++ * ZSTD_endStream() - flush internal buffers into output and end the frame ++ * @zcs: The zstd streaming compression context. ++ * @output: Destination buffer. `output->pos` is updated to indicate how much ++ * compressed data was written. ++ * ++ * ZSTD_endStream() must be called until it returns 0, meaning all the data has ++ * been flushed and the frame epilogue has been written. ++ * ++ * Return: The number of bytes still present within internal buffers or an ++ * error, which can be checked using ZSTD_isError(). ++ */ ++size_t ZSTD_endStream(ZSTD_CStream *zcs, ZSTD_outBuffer *output); ++ ++/** ++ * ZSTD_CStreamInSize() - recommended size for the input buffer ++ * ++ * Return: The recommended size for the input buffer. ++ */ ++size_t ZSTD_CStreamInSize(void); ++/** ++ * ZSTD_CStreamOutSize() - recommended size for the output buffer ++ * ++ * When the output buffer is at least this large, it is guaranteed to be large ++ * enough to flush at least one complete compressed block. ++ * ++ * Return: The recommended size for the output buffer. ++ */ ++size_t ZSTD_CStreamOutSize(void); ++ ++ ++ ++/*-***************************************************************************** ++ * Streaming decompression - HowTo ++ * ++ * A ZSTD_DStream object is required to track streaming operations. ++ * Use ZSTD_initDStream() to initialize a ZSTD_DStream object. ++ * ZSTD_DStream objects can be re-used multiple times. ++ * ++ * Use ZSTD_decompressStream() repetitively to consume your input. ++ * The function will update both `pos` fields. ++ * If `input->pos < input->size`, some input has not been consumed. ++ * It's up to the caller to present again remaining data. ++ * If `output->pos < output->size`, decoder has flushed everything it could. ++ * Returns 0 iff a frame is completely decoded and fully flushed. ++ * Otherwise it returns a suggested next input size that will never load more ++ * than the current frame. ++ ******************************************************************************/ ++ ++/** ++ * ZSTD_DStreamWorkspaceBound() - memory needed to initialize a ZSTD_DStream ++ * @maxWindowSize: The maximum window size allowed for compressed frames. ++ * ++ * Return: A lower bound on the size of the workspace that is passed to ++ * ZSTD_initDStream() and ZSTD_initDStream_usingDDict(). ++ */ ++size_t ZSTD_DStreamWorkspaceBound(size_t maxWindowSize); ++ ++/** ++ * struct ZSTD_DStream - the zstd streaming decompression context ++ */ ++typedef struct ZSTD_DStream_s ZSTD_DStream; ++/*===== ZSTD_DStream management functions =====*/ ++/** ++ * ZSTD_initDStream() - initialize a zstd streaming decompression context ++ * @maxWindowSize: The maximum window size allowed for compressed frames. ++ * @workspace: The workspace to emplace the context into. It must outlive ++ * the returned context. ++ * @workspaceSize: The size of workspace. ++ * Use ZSTD_DStreamWorkspaceBound(maxWindowSize) to determine ++ * how large the workspace must be. ++ * ++ * Return: The zstd streaming decompression context. ++ */ ++ZSTD_DStream *ZSTD_initDStream(size_t maxWindowSize, void *workspace, ++ size_t workspaceSize); ++/** ++ * ZSTD_initDStream_usingDDict() - initialize streaming decompression context ++ * @maxWindowSize: The maximum window size allowed for compressed frames. ++ * @ddict: The digested dictionary to use for decompression. ++ * @workspace: The workspace to emplace the context into. It must outlive ++ * the returned context. ++ * @workspaceSize: The size of workspace. ++ * Use ZSTD_DStreamWorkspaceBound(maxWindowSize) to determine ++ * how large the workspace must be. ++ * ++ * Return: The zstd streaming decompression context. ++ */ ++ZSTD_DStream *ZSTD_initDStream_usingDDict(size_t maxWindowSize, ++ const ZSTD_DDict *ddict, void *workspace, size_t workspaceSize); ++ ++/*===== Streaming decompression functions =====*/ ++/** ++ * ZSTD_resetDStream() - reset the context using parameters from creation ++ * @zds: The zstd streaming decompression context to reset. ++ * ++ * Resets the context using the parameters from creation. Skips dictionary ++ * loading, since it can be reused. ++ * ++ * Return: Zero or an error, which can be checked using ZSTD_isError(). ++ */ ++size_t ZSTD_resetDStream(ZSTD_DStream *zds); ++/** ++ * ZSTD_decompressStream() - streaming decompress some of input into output ++ * @zds: The zstd streaming decompression context. ++ * @output: Destination buffer. `output.pos` is updated to indicate how much ++ * decompressed data was written. ++ * @input: Source buffer. `input.pos` is updated to indicate how much data was ++ * read. Note that it may not consume the entire input, in which case ++ * `input.pos < input.size`, and it's up to the caller to present ++ * remaining data again. ++ * ++ * The `input` and `output` buffers may be any size. Guaranteed to make some ++ * forward progress if `input` and `output` are not empty. ++ * ZSTD_decompressStream() will not consume the last byte of the frame until ++ * the entire frame is flushed. ++ * ++ * Return: Returns 0 iff a frame is completely decoded and fully flushed. ++ * Otherwise returns a hint for the number of bytes to use as the input ++ * for the next function call or an error, which can be checked using ++ * ZSTD_isError(). The size hint will never load more than the frame. ++ */ ++size_t ZSTD_decompressStream(ZSTD_DStream *zds, ZSTD_outBuffer *output, ++ ZSTD_inBuffer *input); ++ ++/** ++ * ZSTD_DStreamInSize() - recommended size for the input buffer ++ * ++ * Return: The recommended size for the input buffer. ++ */ ++size_t ZSTD_DStreamInSize(void); ++/** ++ * ZSTD_DStreamOutSize() - recommended size for the output buffer ++ * ++ * When the output buffer is at least this large, it is guaranteed to be large ++ * enough to flush at least one complete decompressed block. ++ * ++ * Return: The recommended size for the output buffer. ++ */ ++size_t ZSTD_DStreamOutSize(void); ++ ++ ++/* --- Constants ---*/ ++#define ZSTD_MAGICNUMBER 0xFD2FB528 /* >= v0.8.0 */ ++#define ZSTD_MAGIC_SKIPPABLE_START 0x184D2A50U ++ ++#define ZSTD_CONTENTSIZE_UNKNOWN (0ULL - 1) ++#define ZSTD_CONTENTSIZE_ERROR (0ULL - 2) ++ ++#define ZSTD_WINDOWLOG_MAX_32 27 ++#define ZSTD_WINDOWLOG_MAX_64 27 ++#define ZSTD_WINDOWLOG_MAX \ ++ ((unsigned int)(sizeof(size_t) == 4 \ ++ ? ZSTD_WINDOWLOG_MAX_32 \ ++ : ZSTD_WINDOWLOG_MAX_64)) ++#define ZSTD_WINDOWLOG_MIN 10 ++#define ZSTD_HASHLOG_MAX ZSTD_WINDOWLOG_MAX ++#define ZSTD_HASHLOG_MIN 6 ++#define ZSTD_CHAINLOG_MAX (ZSTD_WINDOWLOG_MAX+1) ++#define ZSTD_CHAINLOG_MIN ZSTD_HASHLOG_MIN ++#define ZSTD_HASHLOG3_MAX 17 ++#define ZSTD_SEARCHLOG_MAX (ZSTD_WINDOWLOG_MAX-1) ++#define ZSTD_SEARCHLOG_MIN 1 ++/* only for ZSTD_fast, other strategies are limited to 6 */ ++#define ZSTD_SEARCHLENGTH_MAX 7 ++/* only for ZSTD_btopt, other strategies are limited to 4 */ ++#define ZSTD_SEARCHLENGTH_MIN 3 ++#define ZSTD_TARGETLENGTH_MIN 4 ++#define ZSTD_TARGETLENGTH_MAX 999 ++ ++/* for static allocation */ ++#define ZSTD_FRAMEHEADERSIZE_MAX 18 ++#define ZSTD_FRAMEHEADERSIZE_MIN 6 ++static const size_t ZSTD_frameHeaderSize_prefix = 5; ++static const size_t ZSTD_frameHeaderSize_min = ZSTD_FRAMEHEADERSIZE_MIN; ++static const size_t ZSTD_frameHeaderSize_max = ZSTD_FRAMEHEADERSIZE_MAX; ++/* magic number + skippable frame length */ ++static const size_t ZSTD_skippableHeaderSize = 8; ++ ++ ++/*-************************************* ++ * Compressed size functions ++ **************************************/ ++ ++/** ++ * ZSTD_findFrameCompressedSize() - returns the size of a compressed frame ++ * @src: Source buffer. It should point to the start of a zstd encoded frame ++ * or a skippable frame. ++ * @srcSize: The size of the source buffer. It must be at least as large as the ++ * size of the frame. ++ * ++ * Return: The compressed size of the frame pointed to by `src` or an error, ++ * which can be check with ZSTD_isError(). ++ * Suitable to pass to ZSTD_decompress() or similar functions. ++ */ ++size_t ZSTD_findFrameCompressedSize(const void *src, size_t srcSize); ++ ++/*-************************************* ++ * Decompressed size functions ++ **************************************/ ++/** ++ * ZSTD_getFrameContentSize() - returns the content size in a zstd frame header ++ * @src: It should point to the start of a zstd encoded frame. ++ * @srcSize: The size of the source buffer. It must be at least as large as the ++ * frame header. `ZSTD_frameHeaderSize_max` is always large enough. ++ * ++ * Return: The frame content size stored in the frame header if known. ++ * `ZSTD_CONTENTSIZE_UNKNOWN` if the content size isn't stored in the ++ * frame header. `ZSTD_CONTENTSIZE_ERROR` on invalid input. ++ */ ++unsigned long long ZSTD_getFrameContentSize(const void *src, size_t srcSize); ++ ++/** ++ * ZSTD_findDecompressedSize() - returns decompressed size of a series of frames ++ * @src: It should point to the start of a series of zstd encoded and/or ++ * skippable frames. ++ * @srcSize: The exact size of the series of frames. ++ * ++ * If any zstd encoded frame in the series doesn't have the frame content size ++ * set, `ZSTD_CONTENTSIZE_UNKNOWN` is returned. But frame content size is always ++ * set when using ZSTD_compress(). The decompressed size can be very large. ++ * If the source is untrusted, the decompressed size could be wrong or ++ * intentionally modified. Always ensure the result fits within the ++ * application's authorized limits. ZSTD_findDecompressedSize() handles multiple ++ * frames, and so it must traverse the input to read each frame header. This is ++ * efficient as most of the data is skipped, however it does mean that all frame ++ * data must be present and valid. ++ * ++ * Return: Decompressed size of all the data contained in the frames if known. ++ * `ZSTD_CONTENTSIZE_UNKNOWN` if the decompressed size is unknown. ++ * `ZSTD_CONTENTSIZE_ERROR` if an error occurred. ++ */ ++unsigned long long ZSTD_findDecompressedSize(const void *src, size_t srcSize); ++ ++/*-************************************* ++ * Advanced compression functions ++ **************************************/ ++/** ++ * ZSTD_checkCParams() - ensure parameter values remain within authorized range ++ * @cParams: The zstd compression parameters. ++ * ++ * Return: Zero or an error, which can be checked using ZSTD_isError(). ++ */ ++size_t ZSTD_checkCParams(ZSTD_compressionParameters cParams); ++ ++/** ++ * ZSTD_adjustCParams() - optimize parameters for a given srcSize and dictSize ++ * @srcSize: Optionally the estimated source size, or zero if unknown. ++ * @dictSize: Optionally the estimated dictionary size, or zero if unknown. ++ * ++ * Return: The optimized parameters. ++ */ ++ZSTD_compressionParameters ZSTD_adjustCParams( ++ ZSTD_compressionParameters cParams, unsigned long long srcSize, ++ size_t dictSize); ++ ++/*--- Advanced decompression functions ---*/ ++ ++/** ++ * ZSTD_isFrame() - returns true iff the buffer starts with a valid frame ++ * @buffer: The source buffer to check. ++ * @size: The size of the source buffer, must be at least 4 bytes. ++ * ++ * Return: True iff the buffer starts with a zstd or skippable frame identifier. ++ */ ++unsigned int ZSTD_isFrame(const void *buffer, size_t size); ++ ++/** ++ * ZSTD_getDictID_fromDict() - returns the dictionary id stored in a dictionary ++ * @dict: The dictionary buffer. ++ * @dictSize: The size of the dictionary buffer. ++ * ++ * Return: The dictionary id stored within the dictionary or 0 if the ++ * dictionary is not a zstd dictionary. If it returns 0 the ++ * dictionary can still be loaded as a content-only dictionary. ++ */ ++unsigned int ZSTD_getDictID_fromDict(const void *dict, size_t dictSize); ++ ++/** ++ * ZSTD_getDictID_fromDDict() - returns the dictionary id stored in a ZSTD_DDict ++ * @ddict: The ddict to find the id of. ++ * ++ * Return: The dictionary id stored within `ddict` or 0 if the dictionary is not ++ * a zstd dictionary. If it returns 0 `ddict` will be loaded as a ++ * content-only dictionary. ++ */ ++unsigned int ZSTD_getDictID_fromDDict(const ZSTD_DDict *ddict); ++ ++/** ++ * ZSTD_getDictID_fromFrame() - returns the dictionary id stored in a zstd frame ++ * @src: Source buffer. It must be a zstd encoded frame. ++ * @srcSize: The size of the source buffer. It must be at least as large as the ++ * frame header. `ZSTD_frameHeaderSize_max` is always large enough. ++ * ++ * Return: The dictionary id required to decompress the frame stored within ++ * `src` or 0 if the dictionary id could not be decoded. It can return ++ * 0 if the frame does not require a dictionary, the dictionary id ++ * wasn't stored in the frame, `src` is not a zstd frame, or `srcSize` ++ * is too small. ++ */ ++unsigned int ZSTD_getDictID_fromFrame(const void *src, size_t srcSize); ++ ++/** ++ * struct ZSTD_frameParams - zstd frame parameters stored in the frame header ++ * @frameContentSize: The frame content size, or 0 if not present. ++ * @windowSize: The window size, or 0 if the frame is a skippable frame. ++ * @dictID: The dictionary id, or 0 if not present. ++ * @checksumFlag: Whether a checksum was used. ++ */ ++typedef struct { ++ unsigned long long frameContentSize; ++ unsigned int windowSize; ++ unsigned int dictID; ++ unsigned int checksumFlag; ++} ZSTD_frameParams; ++ ++/** ++ * ZSTD_getFrameParams() - extracts parameters from a zstd or skippable frame ++ * @fparamsPtr: On success the frame parameters are written here. ++ * @src: The source buffer. It must point to a zstd or skippable frame. ++ * @srcSize: The size of the source buffer. `ZSTD_frameHeaderSize_max` is ++ * always large enough to succeed. ++ * ++ * Return: 0 on success. If more data is required it returns how many bytes ++ * must be provided to make forward progress. Otherwise it returns ++ * an error, which can be checked using ZSTD_isError(). ++ */ ++size_t ZSTD_getFrameParams(ZSTD_frameParams *fparamsPtr, const void *src, ++ size_t srcSize); ++ ++/*-***************************************************************************** ++ * Buffer-less and synchronous inner streaming functions ++ * ++ * This is an advanced API, giving full control over buffer management, for ++ * users which need direct control over memory. ++ * But it's also a complex one, with many restrictions (documented below). ++ * Prefer using normal streaming API for an easier experience ++ ******************************************************************************/ ++ ++/*-***************************************************************************** ++ * Buffer-less streaming compression (synchronous mode) ++ * ++ * A ZSTD_CCtx object is required to track streaming operations. ++ * Use ZSTD_initCCtx() to initialize a context. ++ * ZSTD_CCtx object can be re-used multiple times within successive compression ++ * operations. ++ * ++ * Start by initializing a context. ++ * Use ZSTD_compressBegin(), or ZSTD_compressBegin_usingDict() for dictionary ++ * compression, ++ * or ZSTD_compressBegin_advanced(), for finer parameter control. ++ * It's also possible to duplicate a reference context which has already been ++ * initialized, using ZSTD_copyCCtx() ++ * ++ * Then, consume your input using ZSTD_compressContinue(). ++ * There are some important considerations to keep in mind when using this ++ * advanced function : ++ * - ZSTD_compressContinue() has no internal buffer. It uses externally provided ++ * buffer only. ++ * - Interface is synchronous : input is consumed entirely and produce 1+ ++ * (or more) compressed blocks. ++ * - Caller must ensure there is enough space in `dst` to store compressed data ++ * under worst case scenario. Worst case evaluation is provided by ++ * ZSTD_compressBound(). ++ * ZSTD_compressContinue() doesn't guarantee recover after a failed ++ * compression. ++ * - ZSTD_compressContinue() presumes prior input ***is still accessible and ++ * unmodified*** (up to maximum distance size, see WindowLog). ++ * It remembers all previous contiguous blocks, plus one separated memory ++ * segment (which can itself consists of multiple contiguous blocks) ++ * - ZSTD_compressContinue() detects that prior input has been overwritten when ++ * `src` buffer overlaps. In which case, it will "discard" the relevant memory ++ * section from its history. ++ * ++ * Finish a frame with ZSTD_compressEnd(), which will write the last block(s) ++ * and optional checksum. It's possible to use srcSize==0, in which case, it ++ * will write a final empty block to end the frame. Without last block mark, ++ * frames will be considered unfinished (corrupted) by decoders. ++ * ++ * `ZSTD_CCtx` object can be re-used (ZSTD_compressBegin()) to compress some new ++ * frame. ++ ******************************************************************************/ ++ ++/*===== Buffer-less streaming compression functions =====*/ ++size_t ZSTD_compressBegin(ZSTD_CCtx *cctx, int compressionLevel); ++size_t ZSTD_compressBegin_usingDict(ZSTD_CCtx *cctx, const void *dict, ++ size_t dictSize, int compressionLevel); ++size_t ZSTD_compressBegin_advanced(ZSTD_CCtx *cctx, const void *dict, ++ size_t dictSize, ZSTD_parameters params, ++ unsigned long long pledgedSrcSize); ++size_t ZSTD_copyCCtx(ZSTD_CCtx *cctx, const ZSTD_CCtx *preparedCCtx, ++ unsigned long long pledgedSrcSize); ++size_t ZSTD_compressBegin_usingCDict(ZSTD_CCtx *cctx, const ZSTD_CDict *cdict, ++ unsigned long long pledgedSrcSize); ++size_t ZSTD_compressContinue(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity, ++ const void *src, size_t srcSize); ++size_t ZSTD_compressEnd(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity, ++ const void *src, size_t srcSize); ++ ++ ++ ++/*-***************************************************************************** ++ * Buffer-less streaming decompression (synchronous mode) ++ * ++ * A ZSTD_DCtx object is required to track streaming operations. ++ * Use ZSTD_initDCtx() to initialize a context. ++ * A ZSTD_DCtx object can be re-used multiple times. ++ * ++ * First typical operation is to retrieve frame parameters, using ++ * ZSTD_getFrameParams(). It fills a ZSTD_frameParams structure which provide ++ * important information to correctly decode the frame, such as the minimum ++ * rolling buffer size to allocate to decompress data (`windowSize`), and the ++ * dictionary ID used. ++ * Note: content size is optional, it may not be present. 0 means unknown. ++ * Note that these values could be wrong, either because of data malformation, ++ * or because an attacker is spoofing deliberate false information. As a ++ * consequence, check that values remain within valid application range, ++ * especially `windowSize`, before allocation. Each application can set its own ++ * limit, depending on local restrictions. For extended interoperability, it is ++ * recommended to support at least 8 MB. ++ * Frame parameters are extracted from the beginning of the compressed frame. ++ * Data fragment must be large enough to ensure successful decoding, typically ++ * `ZSTD_frameHeaderSize_max` bytes. ++ * Result: 0: successful decoding, the `ZSTD_frameParams` structure is filled. ++ * >0: `srcSize` is too small, provide at least this many bytes. ++ * errorCode, which can be tested using ZSTD_isError(). ++ * ++ * Start decompression, with ZSTD_decompressBegin() or ++ * ZSTD_decompressBegin_usingDict(). Alternatively, you can copy a prepared ++ * context, using ZSTD_copyDCtx(). ++ * ++ * Then use ZSTD_nextSrcSizeToDecompress() and ZSTD_decompressContinue() ++ * alternatively. ++ * ZSTD_nextSrcSizeToDecompress() tells how many bytes to provide as 'srcSize' ++ * to ZSTD_decompressContinue(). ++ * ZSTD_decompressContinue() requires this _exact_ amount of bytes, or it will ++ * fail. ++ * ++ * The result of ZSTD_decompressContinue() is the number of bytes regenerated ++ * within 'dst' (necessarily <= dstCapacity). It can be zero, which is not an ++ * error; it just means ZSTD_decompressContinue() has decoded some metadata ++ * item. It can also be an error code, which can be tested with ZSTD_isError(). ++ * ++ * ZSTD_decompressContinue() needs previous data blocks during decompression, up ++ * to `windowSize`. They should preferably be located contiguously, prior to ++ * current block. Alternatively, a round buffer of sufficient size is also ++ * possible. Sufficient size is determined by frame parameters. ++ * ZSTD_decompressContinue() is very sensitive to contiguity, if 2 blocks don't ++ * follow each other, make sure that either the compressor breaks contiguity at ++ * the same place, or that previous contiguous segment is large enough to ++ * properly handle maximum back-reference. ++ * ++ * A frame is fully decoded when ZSTD_nextSrcSizeToDecompress() returns zero. ++ * Context can then be reset to start a new decompression. ++ * ++ * Note: it's possible to know if next input to present is a header or a block, ++ * using ZSTD_nextInputType(). This information is not required to properly ++ * decode a frame. ++ * ++ * == Special case: skippable frames == ++ * ++ * Skippable frames allow integration of user-defined data into a flow of ++ * concatenated frames. Skippable frames will be ignored (skipped) by a ++ * decompressor. The format of skippable frames is as follows: ++ * a) Skippable frame ID - 4 Bytes, Little endian format, any value from ++ * 0x184D2A50 to 0x184D2A5F ++ * b) Frame Size - 4 Bytes, Little endian format, unsigned 32-bits ++ * c) Frame Content - any content (User Data) of length equal to Frame Size ++ * For skippable frames ZSTD_decompressContinue() always returns 0. ++ * For skippable frames ZSTD_getFrameParams() returns fparamsPtr->windowLog==0 ++ * what means that a frame is skippable. ++ * Note: If fparamsPtr->frameContentSize==0, it is ambiguous: the frame might ++ * actually be a zstd encoded frame with no content. For purposes of ++ * decompression, it is valid in both cases to skip the frame using ++ * ZSTD_findFrameCompressedSize() to find its size in bytes. ++ * It also returns frame size as fparamsPtr->frameContentSize. ++ ******************************************************************************/ ++ ++/*===== Buffer-less streaming decompression functions =====*/ ++size_t ZSTD_decompressBegin(ZSTD_DCtx *dctx); ++size_t ZSTD_decompressBegin_usingDict(ZSTD_DCtx *dctx, const void *dict, ++ size_t dictSize); ++void ZSTD_copyDCtx(ZSTD_DCtx *dctx, const ZSTD_DCtx *preparedDCtx); ++size_t ZSTD_nextSrcSizeToDecompress(ZSTD_DCtx *dctx); ++size_t ZSTD_decompressContinue(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, ++ const void *src, size_t srcSize); ++typedef enum { ++ ZSTDnit_frameHeader, ++ ZSTDnit_blockHeader, ++ ZSTDnit_block, ++ ZSTDnit_lastBlock, ++ ZSTDnit_checksum, ++ ZSTDnit_skippableFrame ++} ZSTD_nextInputType_e; ++ZSTD_nextInputType_e ZSTD_nextInputType(ZSTD_DCtx *dctx); ++ ++/*-***************************************************************************** ++ * Block functions ++ * ++ * Block functions produce and decode raw zstd blocks, without frame metadata. ++ * Frame metadata cost is typically ~18 bytes, which can be non-negligible for ++ * very small blocks (< 100 bytes). User will have to take in charge required ++ * information to regenerate data, such as compressed and content sizes. ++ * ++ * A few rules to respect: ++ * - Compressing and decompressing require a context structure ++ * + Use ZSTD_initCCtx() and ZSTD_initDCtx() ++ * - It is necessary to init context before starting ++ * + compression : ZSTD_compressBegin() ++ * + decompression : ZSTD_decompressBegin() ++ * + variants _usingDict() are also allowed ++ * + copyCCtx() and copyDCtx() work too ++ * - Block size is limited, it must be <= ZSTD_getBlockSizeMax() ++ * + If you need to compress more, cut data into multiple blocks ++ * + Consider using the regular ZSTD_compress() instead, as frame metadata ++ * costs become negligible when source size is large. ++ * - When a block is considered not compressible enough, ZSTD_compressBlock() ++ * result will be zero. In which case, nothing is produced into `dst`. ++ * + User must test for such outcome and deal directly with uncompressed data ++ * + ZSTD_decompressBlock() doesn't accept uncompressed data as input!!! ++ * + In case of multiple successive blocks, decoder must be informed of ++ * uncompressed block existence to follow proper history. Use ++ * ZSTD_insertBlock() in such a case. ++ ******************************************************************************/ ++ ++/* Define for static allocation */ ++#define ZSTD_BLOCKSIZE_ABSOLUTEMAX (128 * 1024) ++/*===== Raw zstd block functions =====*/ ++size_t ZSTD_getBlockSizeMax(ZSTD_CCtx *cctx); ++size_t ZSTD_compressBlock(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity, ++ const void *src, size_t srcSize); ++size_t ZSTD_decompressBlock(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, ++ const void *src, size_t srcSize); ++size_t ZSTD_insertBlock(ZSTD_DCtx *dctx, const void *blockStart, ++ size_t blockSize); ++ ++#endif /* ZSTD_H */