From: Nick Terrell <terrelln@fb.com>
Date: Fri, 20 Sep 2019 23:39:27 +0000 (-0700)
Subject: HINT_INLINE ZSTD_storeSeq()
X-Git-Tag: v1.4.4~1^2~42^2~1
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=5dc0a1d65921821eddfcf0a16ef4a3d79be0ba3f;p=thirdparty%2Fzstd.git

HINT_INLINE ZSTD_storeSeq()

Clang on Mac wasn't inlining `ZSTD_storeSeq()` in level 1, which was
causing a 5% performance regression. This fixes it.
---

diff --git a/lib/compress/zstd_compress_internal.h b/lib/compress/zstd_compress_internal.h
index e80686cca..42660e1e0 100644
--- a/lib/compress/zstd_compress_internal.h
+++ b/lib/compress/zstd_compress_internal.h
@@ -361,7 +361,7 @@ static void ZSTD_safecopyLiterals(BYTE* op, BYTE const* ip, BYTE const* const ie
  *  `mlBase` : matchLength - MINMATCH
  *  Allowed to overread literals up to litLimit.
 */
-MEM_STATIC void ZSTD_storeSeq(seqStore_t* seqStorePtr, size_t litLength, const BYTE* literals, const BYTE* litLimit, U32 offCode, size_t mlBase)
+HINT_INLINE void ZSTD_storeSeq(seqStore_t* seqStorePtr, size_t litLength, const BYTE* literals, const BYTE* litLimit, U32 offCode, size_t mlBase)
 {
     BYTE const* const litLimit_w = litLimit - WILDCOPY_OVERLENGTH;
     BYTE const* const litEnd = literals + litLength;