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

[PATCH v2 1/8] midx-write.c: tolerate `--preferred-pack` without bitmaps

From
Taylor Blau <me@ttaylorr.com>
Date
May 29, 2024, 22:55 UTC
Message-ID
<ad268bd264e34f02f9c32e31db696c7acc1efc2b.1717023301.git.me@ttaylorr.com>
In-Reply-To
<cover.1717023301.git.me@ttaylorr.com>

When passing a preferred pack to the MIDX write machinery, we ensure that the given preferred pack is non-empty since 5d3cd09a808 (midx: reject empty `--preferred-pack`'s, 2021-08-31).

However packs are only loaded (via `write_midx_internal()`, though a subsequent patch will refactor this code out to its own function) when the `MIDX_WRITE_REV_INDEX` flag is set.

So if a caller runs:
    $ git multi-pack-index write --preferred-pack=...

with both (a) an existing MIDX, and (b) specifies a pack from that MIDX as the preferred one, without passing `--bitmap`, then the check added in 5d3cd09a808 will result in a segfault.

Note that packs loaded from disk which don't appear in an existing MIDX do not trigger this issue, as those packs are loaded unconditionally. We conditionally load packs from a MIDX since we tolerate MIDXs whose packs do not resolve (i.e., via the MIDX write after removing unreferenced packs via 'git multi-pack-index expire').

In practice, this isn't possible to trigger when running `git multi-pack-index write` from `git repack`, as the latter always passes `--stdin-packs`, which prevents us from loading an existing MIDX, as it forces all packs to be read from disk.

But a future commit in this series will change that behavior to unconditionally load an existing MIDX, even with `--stdin-packs`, making this behavior trigger-able from 'repack' much more easily.

Prevent this from being an issue by removing the segfault altogether by calling `prepare_midx_pack()` on packs loaded from an existing MIDX when either the `MIDX_WRITE_REV_INDEX` flag is set *or* we specified a `--preferred-pack`.

Signed-off-by: Taylor Blau <me@ttaylorr.com>
---
 midx-write.c                |  8 +++++++-
 t/t5319-multi-pack-index.sh | 23 +++++++++++++++++++++++
 2 files changed, 30 insertions(+), 1 deletion(-)
diff --git a/midx-write.c b/midx-write.c
index 2fa120c213..591b79bcbf 100644
--- a/midx-write.c
+++ b/midx-write.c
@@ -935,11 +935,17 @@ static int write_midx_internal(const char *object_dir,
 		for (i = 0; i < ctx.m->num_packs; i++) {
 			ALLOC_GROW(ctx.info, ctx.nr + 1, ctx.alloc);
 
-			if (flags & MIDX_WRITE_REV_INDEX) {
+			if (flags & MIDX_WRITE_REV_INDEX ||
+			    preferred_pack_name) {
 				/*
 				 * 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 (prepare_midx_pack(the_repository, ctx.m, i)) {
 					error(_("could not load pack"));
diff --git a/t/t5319-multi-pack-index.sh b/t/t5319-multi-pack-index.sh
index dd09134db0..10d2a6bf92 100755
--- a/t/t5319-multi-pack-index.sh
+++ b/t/t5319-multi-pack-index.sh
@@ -350,6 +350,29 @@ test_expect_success 'preferred packs must be non-empty' '
 	)
 '
 
+test_expect_success 'preferred pack from existing MIDX without bitmaps' '
+	git init preferred-without-bitmaps &&
+	(
+		cd preferred-without-bitmaps &&
+
+		test_commit one &&
+		pack="$(git pack-objects --all $objdir/pack/pack </dev/null)" &&
+
+		git multi-pack-index write &&
+
+		# make another pack so that the subsequent MIDX write
+		# has something to do
+		test_commit two &&
+		git repack -d &&
+
+		# write a new MIDX without bitmaps reusing the singular
+		# pack from the existing MIDX as the preferred pack in
+		# the new MIDX
+		git multi-pack-index write --preferred-pack=pack-$pack.pack
+	)
+
+'
+
 test_expect_success 'verify multi-pack-index success' '
 	git multi-pack-index verify --object-dir=$objdir
 '
-- 
2.45.1.321.gbcadaf92783
Previous: Taylor BlauNext: Taylor Blau
Message 21 of 34 in “midx-write: miscellaneous clean-ups for incremental MIDXs”
  1. 0/8 midx-write: miscellaneous clean-ups for incremental MIDXsTaylor Blau, May 23, 2024
  2. 1/8 midx-write.c: tolerate `--preferred-pack` without bitmapsTaylor Blau, May 23, 2024
  3. Patrick SteinhardtMay 29, 2024
  4. Taylor BlauMay 29, 2024
  5. 2/8 midx-write.c: reduce argument count for `get_sorted_entries()`Taylor Blau, May 23, 2024
  6. Patrick SteinhardtMay 29, 2024
  7. Taylor BlauMay 29, 2024
  8. 3/8 midx-write.c: pass `start_pack` to `get_sorted_entries()`Taylor Blau, May 23, 2024
  9. Patrick SteinhardtMay 29, 2024
  10. Taylor BlauMay 29, 2024
  11. 4/8 midx-write.c: extract `should_include_pack()`Taylor Blau, May 23, 2024
  12. Patrick SteinhardtMay 29, 2024
  13. Taylor BlauMay 29, 2024
  14. 5/8 midx-write.c: extract `fill_packs_from_midx()`Taylor Blau, May 23, 2024
  15. 6/8 midx-write.c: support reading an existing MIDX with `packs_to_include`Taylor Blau, May 23, 2024
  16. Patrick SteinhardtMay 29, 2024
  17. Taylor BlauMay 29, 2024
  18. 7/8 midx: replace `get_midx_rev_filename()` with a generic helperTaylor Blau, May 23, 2024
  19. 8/8 pack-bitmap.c: reimplement `midx_bitmap_filename()` with helperTaylor Blau, May 23, 2024
  20. 0/8 midx-write: miscellaneous clean-ups for incremental MIDXsTaylor Blau, May 29, 2024
  21. 1/8 midx-write.c: tolerate `--preferred-pack` without bitmapsTaylor Blau, May 29, 2024
  22. 2/8 midx-write.c: reduce argument count for `get_sorted_entries()`Taylor Blau, May 29, 2024
  23. 3/8 midx-write.c: pass `start_pack` to `compute_sorted_entries()`Taylor Blau, May 29, 2024
  24. Jeff KingMay 30, 2024
  25. 4/8 midx-write.c: extract `should_include_pack()`Taylor Blau, May 29, 2024
  26. 5/8 midx-write.c: extract `fill_packs_from_midx()`Taylor Blau, May 29, 2024
  27. 6/8 midx-write.c: support reading an existing MIDX with `packs_to_include`Taylor Blau, May 29, 2024
  28. Jeff KingMay 30, 2024
  29. 7/8 midx: replace `get_midx_rev_filename()` with a generic helperTaylor Blau, May 29, 2024
  30. Jeff KingMay 30, 2024
  31. 8/8 pack-bitmap.c: reimplement `midx_bitmap_filename()` with helperTaylor Blau, May 29, 2024
  32. Jeff KingMay 30, 2024
  33. Taylor BlauMay 30, 2024
  34. Patrick SteinhardtMay 31, 2024

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.