git/list[1] front-page[2] threads[3] people[4] search[5] about
 

[PATCH v2 1/6] midx-write: only load initialized packs

From
Derrick Stolee via GitGitGadget <gitgitgadget@gmail.com>
Date
Aug 30, 2025, 21:23 UTC
Message-ID
<e02a444315acbc638a3d31279c10a936f0adb7b4.1756589007.git.gitgitgadget@gmail.com>
In-Reply-To
<pull.1965.v2.git.1756589007.gitgitgadget@gmail.com>
From: Derrick Stolee <stolee@gmail.com>

The fill_packs_from_midx() method was refactored in fcb2205b77 (midx: implement support for writing incremental MIDX chains, 2024-08-06) to allow for preferred packfiles and incremental multi-pack-indexes. However, this led to some conditions that can cause improperly initialized memory in the context's list of packfiles.

The conditions caring about the preferred pack name or the incremental flag are currently necessary to load a packfile. But the context is still being populated with pack_info structs based on the packfile array for the existing multi-pack-index even if prepare_midx_pack() isn't called.

Add a new test that breaks under --stress when compiled with SANITIZE=address. The chosen number of 100 packfiles was selected to get the --stress output to fail about 50% of the time, while 50 packfiles could not get a failure in most --stress runs.

The test case is marked as EXPENSIVE not only because of the number of packfiles it creates, but because some CI environments were reporting errors during the test that I could not reproduce, specifically around being unable to open the packfiles or their pack-indexes.

When it fails under SANITIZE=address, it provides the following error:
AddressSanitizer:DEADLYSIGNAL
=================================================================
==3263517==ERROR: AddressSanitizer: SEGV on unknown address 0x000000000027
==3263517==The signal is caused by a READ memory access.
==3263517==Hint: address points to the zero page.
    #0 0x562d5d82d1fb in close_pack_windows packfile.c:299
    #1 0x562d5d82d3ab in close_pack packfile.c:354
    #2 0x562d5d7bfdb4 in write_midx_internal midx-write.c:1490
    #3 0x562d5d7c7aec in midx_repack midx-write.c:1795
    #4 0x562d5d46fff6 in cmd_multi_pack_index builtin/multi-pack-index.c:305
    ...

This failure stack trace is disconnected from the real fix because the bad pointers are accessed later when closing the packfiles from the context.

There are a few different aspects to this fix that are worth noting:
 1. We return to the previous behavior of fill_packs_from_midx to not
    rely on the incremental flag or existence of a preferred pack.
 2. The behavior to scan all layers of an incremental midx is kept, so
    this is not a full revert of the change.
 3. We skip allocating more room in the pack_info array if the pack
    fails prepare_midx_pack().
 4. The method has always returned 0 for success and 1 for failure, but
    the condition checking for error added a check for a negative result
    for failure, so that is now updated.
 5. The call to open_pack_index() is removed, but this is needed later
    in the case of a preferred pack. That call is moved to immediately
    before its result is needed (checking for the object count).
Signed-off-by: Derrick Stolee <stolee@gmail.com>
---
 midx-write.c                | 46 +++++++++++++++----------------------
 t/t5319-multi-pack-index.sh | 17 ++++++++++++++
 2 files changed, 36 insertions(+), 27 deletions(-)
diff --git a/midx-write.c b/midx-write.c
index a0aceab5e0..070a7f61f4 100644
--- a/midx-write.c
+++ b/midx-write.c
@@ -920,8 +920,7 @@ static struct multi_pack_index *lookup_multi_pack_index(struct repository *r,
 	return get_multi_pack_index(source);
 }
 
-static int fill_packs_from_midx(struct write_midx_context *ctx,
-				const char *preferred_pack_name, uint32_t flags)
+static int fill_packs_from_midx(struct write_midx_context *ctx)
 {
 	struct multi_pack_index *m;
 
@@ -929,30 +928,11 @@ static int fill_packs_from_midx(struct write_midx_context *ctx,
 		uint32_t i;
 
 		for (i = 0; i < m->num_packs; i++) {
-			ALLOC_GROW(ctx->info, ctx->nr + 1, ctx->alloc);
-
-			/*
-			 * If generating a reverse index, need to have
-			 * packed_git's loaded to compare their
-			 * mtimes and object count.
-			 *
-			 * If a preferred pack is specified, need to
-			 * have packed_git's loaded to ensure the chosen
-			 * preferred pack has a non-zero object count.
-			 */
-			if (flags & MIDX_WRITE_REV_INDEX ||
-			    preferred_pack_name) {
-				if (prepare_midx_pack(ctx->repo, m,
-						      m->num_packs_in_base + i)) {
-					error(_("could not load pack"));
-					return 1;
-				}
-
-				if (open_pack_index(m->packs[i]))
-					die(_("could not open index for %s"),
-					    m->packs[i]->pack_name);
-			}
+			if (prepare_midx_pack(ctx->repo, m,
+					      m->num_packs_in_base + i))
+				return error(_("could not load pack"));
 
+			ALLOC_GROW(ctx->info, ctx->nr + 1, ctx->alloc);
 			fill_pack_info(&ctx->info[ctx->nr++], m->packs[i],
 				       m->pack_names[i],
 				       m->num_packs_in_base + i);
@@ -1123,8 +1103,7 @@ static int write_midx_internal(struct repository *r, const char *object_dir,
 			ctx.num_multi_pack_indexes_before++;
 			m = m->base_midx;
 		}
-	} else if (ctx.m && fill_packs_from_midx(&ctx, preferred_pack_name,
-						 flags) < 0) {
+	} else if (ctx.m && fill_packs_from_midx(&ctx)) {
 		goto cleanup;
 	}
 
@@ -1186,6 +1165,13 @@ static int write_midx_internal(struct repository *r, const char *object_dir,
 		struct packed_git *oldest = ctx.info[ctx.preferred_pack_idx].p;
 		ctx.preferred_pack_idx = 0;
 
+		/*
+		 * Attempt opening the pack index to populate num_objects.
+		 * Ignore failiures as they can be expected and are not
+		 * fatal during this selection time.
+		 */
+		open_pack_index(oldest);
+
 		if (packs_to_drop && packs_to_drop->nr)
 			BUG("cannot write a MIDX bitmap during expiration");
 
@@ -1200,6 +1186,7 @@ static int write_midx_internal(struct repository *r, const char *object_dir,
 
 			if (!oldest->num_objects || p->mtime < oldest->mtime) {
 				oldest = p;
+				open_pack_index(oldest);
 				ctx.preferred_pack_idx = i;
 			}
 		}
@@ -1223,6 +1210,11 @@ static int write_midx_internal(struct repository *r, const char *object_dir,
 
 	if (ctx.preferred_pack_idx > -1) {
 		struct packed_git *preferred = ctx.info[ctx.preferred_pack_idx].p;
+
+		if (open_pack_index(preferred))
+			die(_("failed to open preferred pack %s"),
+			    ctx.info[ctx.preferred_pack_idx].pack_name);
+
 		if (!preferred->num_objects) {
 			error(_("cannot select preferred pack %s with no objects"),
 			      preferred->pack_name);
diff --git a/t/t5319-multi-pack-index.sh b/t/t5319-multi-pack-index.sh
index bd75dea950..49705c62a2 100755
--- a/t/t5319-multi-pack-index.sh
+++ b/t/t5319-multi-pack-index.sh
@@ -989,6 +989,23 @@ test_expect_success 'repack --batch-size=0 repacks everything' '
 	)
 '
 
+test_expect_success EXPENSIVE 'repack/expire with many packs' '
+	cp -r dup many &&
+	(
+		cd many &&
+
+		for i in $(test_seq 1 100)
+		do
+			test_commit extra$i &&
+			git maintenance run --task=loose-objects || return 1
+		done &&
+
+		git multi-pack-index write &&
+		git multi-pack-index repack &&
+		git multi-pack-index expire
+	)
+'
+
 test_expect_success 'repack --batch-size=<large> repacks everything' '
 	(
 		cd dup2 &&
-- 
gitgitgadget
Previous: Derrick Stolee via GitGitGadgetNext: Patrick Steinhardt
Message 21 of 47 in “midx-write: fix segfault and do several cleanups”
  1. 0/5 midx-write: fix segfault and do several cleanupsDerrick Stolee via GitGitGadget, Aug 28, 2025
  2. 1/5 midx-write: only load initialized packsDerrick Stolee via GitGitGadget, Aug 28, 2025
  3. Junio C HamanoAug 28, 2025
  4. Taylor BlauAug 29, 2025
  5. Derrick StoleeAug 30, 2025
  6. 2/5 midx-write: put failing response value backDerrick Stolee via GitGitGadget, Aug 28, 2025
  7. Junio C HamanoAug 28, 2025
  8. Taylor BlauAug 29, 2025
  9. 3/5 midx-write: use cleanup when incremental midx failsDerrick Stolee via GitGitGadget, Aug 28, 2025
  10. Junio C HamanoAug 28, 2025
  11. Taylor BlauAug 29, 2025
  12. Derrick StoleeAug 30, 2025
  13. 4/5 midx-write: use uint32_t for preferred_pack_idxDerrick Stolee via GitGitGadget, Aug 28, 2025
  14. Junio C HamanoAug 28, 2025
  15. Taylor BlauAug 29, 2025
  16. 5/5 midx-write: reenable signed comparison errorsDerrick Stolee via GitGitGadget, Aug 28, 2025
  17. Junio C HamanoAug 28, 2025
  18. Taylor BlauAug 29, 2025
  19. Taylor BlauAug 29, 2025
  20. 0/6 midx-write: fix segfault and do several cleanupsDerrick Stolee via GitGitGadget, Aug 30, 2025
  21. 1/6 midx-write: only load initialized packsDerrick Stolee via GitGitGadget, Aug 30, 2025
  22. Patrick SteinhardtSep 3, 2025
  23. Derrick StoleeSep 5, 2025
  24. 2/6 midx-write: put failing response value backDerrick Stolee via GitGitGadget, Aug 30, 2025
  25. Patrick SteinhardtSep 3, 2025
  26. Derrick StoleeSep 5, 2025
  27. 3/6 midx-write: use cleanup when incremental midx failsDerrick Stolee via GitGitGadget, Aug 30, 2025
  28. Patrick SteinhardtSep 3, 2025
  29. 4/6 midx-write: use uint32_t for preferred_pack_idxDerrick Stolee via GitGitGadget, Aug 30, 2025
  30. Patrick SteinhardtSep 3, 2025
  31. Derrick StoleeSep 5, 2025
  32. 5/6 midx-write: reenable signed comparison errorsDerrick Stolee via GitGitGadget, Aug 30, 2025
  33. Patrick SteinhardtSep 3, 2025
  34. 6/6 midx-write: simplify error casesDerrick Stolee via GitGitGadget, Aug 30, 2025
  35. Patrick SteinhardtSep 3, 2025
  36. Junio C HamanoSep 3, 2025
  37. 0/6 midx-write: fix segfault and do several cleanupsDerrick Stolee via GitGitGadget, Sep 5, 2025
  38. 1/6 midx-write: only load initialized packsDerrick Stolee via GitGitGadget, Sep 5, 2025
  39. 2/6 midx-write: put failing response value backDerrick Stolee via GitGitGadget, Sep 5, 2025
  40. 3/6 midx-write: use cleanup when incremental midx failsDerrick Stolee via GitGitGadget, Sep 5, 2025
  41. 4/6 midx-write: use uint32_t for preferred_pack_idxDerrick Stolee via GitGitGadget, Sep 5, 2025
  42. 5/6 midx-write: reenable signed comparison errorsDerrick Stolee via GitGitGadget, Sep 5, 2025
  43. 6/6 midx-write: simplify error casesDerrick Stolee via GitGitGadget, Sep 5, 2025
  44. Junio C HamanoSep 5, 2025
  45. Derrick StoleeSep 5, 2025
  46. Taylor BlauSep 11, 2025
  47. Junio C HamanoSep 11, 2025

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.