From 37706a677c09a1051f8b02361928c475bd094e67 Mon Sep 17 00:00:00 2001
From: Yann Collet <cyan@fb.com>
Date: Fri, 25 Oct 2024 20:29:15 -0700
Subject: [PATCH] added a test

test both that the new parameter works as intended,
and that the over-split protection works as intended
---
 lib/compress/zstd_compress.c |  7 ++--
 lib/compress/zstd_preSplit.c |  1 +
 tests/fuzzer.c               | 63 ++++++++++++++++++++++++++++++++++++
 3 files changed, 69 insertions(+), 2 deletions(-)

diff --git a/lib/compress/zstd_compress.c b/lib/compress/zstd_compress.c
index fbc11349e..ca9e5edff 100644
--- a/lib/compress/zstd_compress.c
+++ b/lib/compress/zstd_compress.c
@@ -2114,7 +2114,7 @@ static size_t ZSTD_resetCCtx_internal(ZSTD_CCtx* zc,
 {
     ZSTD_cwksp* const ws = &zc->workspace;
     DEBUGLOG(4, "ZSTD_resetCCtx_internal: pledgedSrcSize=%u, wlog=%u, useRowMatchFinder=%d useBlockSplitter=%d",
-                (U32)pledgedSrcSize, params->cParams.windowLog, (int)params->useRowMatchFinder, (int)params->useBlockSplitter);
+                (U32)pledgedSrcSize, params->cParams.windowLog, (int)params->useRowMatchFinder, (int)params->postBlockSplitter);
     assert(!ZSTD_isError(ZSTD_checkCParams(params->cParams)));
 
     zc->isFirstBlock = 1;
@@ -4520,7 +4520,10 @@ static size_t ZSTD_optimalBlockSize(ZSTD_CCtx* cctx, const void* src, size_t src
      * require verified savings to allow pre-splitting.
      * Note: as a consequence, the first full block is not split.
      */
-    if (savings < 3) return 128 KB;
+    if (savings < 3) {
+        DEBUGLOG(6, "don't attempt splitting: savings (%lli) too low", savings);
+        return 128 KB;
+    }
     /* apply @splitLevel, or use default value (which depends on @strat).
      * note that splitting heuristic is still conditioned by @savings >= 3,
      * so the first block will not reach this code path */
diff --git a/lib/compress/zstd_preSplit.c b/lib/compress/zstd_preSplit.c
index d25773950..d820c20ac 100644
--- a/lib/compress/zstd_preSplit.c
+++ b/lib/compress/zstd_preSplit.c
@@ -229,6 +229,7 @@ size_t ZSTD_splitBlock(const void* blockStart, size_t blockSize,
                     int level,
                     void* workspace, size_t wkspSize)
 {
+    DEBUGLOG(6, "ZSTD_splitBlock (level=%i)", level);
     assert(0<=level && level<=4);
     if (level == 0)
         return ZSTD_splitBlock_fromBorders(blockStart, blockSize, workspace, wkspSize);
diff --git a/tests/fuzzer.c b/tests/fuzzer.c
index f5a894354..e32c0130d 100644
--- a/tests/fuzzer.c
+++ b/tests/fuzzer.c
@@ -559,6 +559,67 @@ static void test_setCParams(unsigned tnb)
     DISPLAYLEVEL(3, "OK \n");
 }
 
+static void test_blockSplitter_incompressibleExpansionProtection(unsigned testNb, unsigned seed)
+{
+    DISPLAYLEVEL(3, "test%3i : Check block splitter doesn't oversplit incompressible data (seed %u): ", testNb, seed);
+    {   ZSTD_CCtx* cctx = ZSTD_createCCtx();
+        size_t const srcSize = 256 * 1024; /* needs to be at least 2 blocks */
+        void* incompressible = malloc(srcSize);
+        size_t const dstCapacity = ZSTD_compressBound(srcSize);
+        void* cBuffer = malloc(dstCapacity);
+        size_t const chunkSize = 8 KB;
+        size_t const nbChunks = srcSize / chunkSize;
+        size_t chunkNb, cSizeNoSplit, cSizeWithSplit;
+        assert(cctx != NULL);
+        assert(incompressible != NULL);
+        assert(cBuffer != NULL);
+
+        /* let's fill input with random noise (incompressible) */
+        RDG_genBuffer(incompressible, srcSize, 0.0, 0.0, seed);
+        DISPLAYLEVEL(4, "(hash: %llx) ", XXH64(incompressible, srcSize, 0));
+
+        /* this pattern targets the fastest _byChunk variant's sampling (level 3).
+         * manually checked that, without the @savings protection, it would over-split.
+         */
+        for (chunkNb=0; chunkNb<nbChunks; chunkNb++) {
+            BYTE* const p = (BYTE*)incompressible + chunkNb * chunkSize;
+            size_t const samplingRate = 43;
+            int addOrRemove = chunkNb % 2;
+            size_t n;
+            for (n=0; n<chunkSize; n+=samplingRate) {
+                if (addOrRemove) {
+                    p[n] &= 0x80;
+                } else {
+                    p[n] |= 0x80;
+                }
+            }
+        }
+
+        /* run first without splitting */
+        ZSTD_CCtx_setParameter(cctx, ZSTD_c_blockSplitter_level, 1 /* no split */);
+        cSizeNoSplit = ZSTD_compress2(cctx, cBuffer, dstCapacity, incompressible, srcSize);
+
+        /* run with sample43 splitter, check it's still the same */
+        ZSTD_CCtx_setParameter(cctx, ZSTD_c_blockSplitter_level, 3 /* sample43, fastest _byChunk variant */);
+        cSizeWithSplit = ZSTD_compress2(cctx, cBuffer, dstCapacity, incompressible, srcSize);
+
+        if (cSizeWithSplit != cSizeNoSplit) {
+            DISPLAYLEVEL(1, "invalid compressed size: cSizeWithSplit %u != %u cSizeNoSplit \n",
+                    (unsigned)cSizeWithSplit, (unsigned)cSizeNoSplit);
+            abort();
+        }
+        DISPLAYLEVEL(4, "compressed size: cSizeWithSplit %u == %u cSizeNoSplit : ",
+                (unsigned)cSizeWithSplit, (unsigned)cSizeNoSplit);
+
+        free(incompressible);
+        free(cBuffer);
+        ZSTD_freeCCtx(cctx);
+    }
+    DISPLAYLEVEL(3, "OK \n");
+}
+
+/* ============================================================= */
+
 static int basicUnitTests(U32 const seed, double compressibility)
 {
     size_t const CNBuffSize = 5 MB;
@@ -1374,6 +1435,8 @@ static int basicUnitTests(U32 const seed, double compressibility)
     }
     DISPLAYLEVEL(3, "OK \n");
 
+    test_blockSplitter_incompressibleExpansionProtection(testNb++, seed);
+
     DISPLAYLEVEL(3, "test%3d : superblock uncompressible data: too many nocompress superblocks : ", testNb++);
     {
         ZSTD_CCtx* const cctx = ZSTD_createCCtx();