Serialized Commit Graph #2

derrickstolee · 2018-01-20T18:55:14Z

Teach Git to construct and consume a serialized commit graph for faster commit walks. Has ability to extend to include generation numbers.

* jt/binsearch-with-fanout: packfile: refactor hash search with fanout table packfile: remove GIT_DEBUG_LOOKUP log statements

* jk/cached-commit-buffer: revision: drop --show-all option commit: drop uses of get_cached_commit_buffer() Git 2.16.2

The function ce_write_entry() uses a 'self-initialised' variable construct, for the symbol 'saved_namelen', to suppress a gcc '-Wmaybe-uninitialized' warning, given that the warning is a false positive. For the purposes of this discussion, the ce_write_entry() function has three code blocks of interest, that look like so: /* block #1 */ if (ce->ce_flags & CE_STRIP_NAME) { saved_namelen = ce_namelen(ce); ce->ce_namelen = 0; } /* block #2 */ /* * several code blocks that contain, among others, calls * to copy_cache_entry_to_ondisk(ondisk, ce); */ /* block #3 */ if (ce->ce_flags & CE_STRIP_NAME) { ce->ce_namelen = saved_namelen; ce->ce_flags &= ~CE_STRIP_NAME; } The warning implies that gcc thinks it is possible that the first block is not entered, the calls to copy_cache_entry_to_ondisk() could toggle the CE_STRIP_NAME flag on, thereby entering block #3 with saved_namelen unset. However, the copy_cache_entry_to_ondisk() function does not write to ce->ce_flags (it only reads). gcc could easily determine this, since that function is local to this file, but it obviously doesn't. In order to suppress this warning, we make it clear to the reader (human and compiler), that block #3 will only be entered when the first block has been entered, by introducing a new 'stripped_name' boolean variable. We also take the opportunity to change the type of 'saved_namelen' to 'unsigned int' to match ce->ce_namelen. Signed-off-by: Ramsay Jones <ramsay@ramsayjones.plus.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

The hashclose() method behaves very differently depending on the flags parameter. In particular, the file descriptor is not always closed. Perform a simple rename of "hashclose()" to "finalize_hashfile()" in preparation for functional changes. Signed-off-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

If we want to use a hashfile on the temporary file for a lockfile, then we need finalize_hashfile() to fully write the trailing hash but also keep the file descriptor open. Do this by adding a new CSUM_HASH_IN_STREAM flag along with a functional change that checks this flag before writing the checksum to the stream. This differs from previous behavior since it would be written if either CSUM_CLOSE or CSUM_FSYNC is provided. Signed-off-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

Add document specifying the binary format for commit graphs. This format allows for: * New versions. * New hash functions and hash lengths. * Optional extensions. Basic header information is followed by a binary table of contents into "chunks" that include: * An ordered list of commit object IDs. * A 256-entry fanout into that list of OIDs. * A list of metadata for the commits. * A list of "large edges" to enable octopus merges. The format automatically includes two parent positions for every commit. This favors speed over space, since using only one position per commit would cause an extra level of indirection for every merge commit. (Octopus merges suffer from this indirection, but they are very rare.) Signed-off-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

Add Documentation/technical/commit-graph.txt with details of the planned commit graph feature, including future plans. Signed-off-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

Teach git the 'commit-graph' builtin that will be used for writing and reading packed graph files. The current implementation is mostly empty, except for an '--object-dir' option. Signed-off-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

Teach Git to write a commit graph file by checking all packed objects to see if they are commits, then store the file in the given object directory. Helped-by: Jeff King <peff@peff.net> Signed-off-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

Teach git-commit-graph to write graph files. Create new test script to verify this command succeeds without failure. Signed-off-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

Teach git-commit-graph to read commit graph files and summarize their contents. Use the read subcommand to verify the contents of a commit graph file in the tests. Signed-off-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

The commit graph feature is controlled by the new core.commitGraph config setting. This defaults to 0, so the feature is opt-in. The intention of core.commitGraph is that a user can always stop checking for or parsing commit graph files if core.commitGraph=0. Signed-off-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

Teach write_commit_graph() to walk all parents from the commits discovered in packfiles. This prevents gaps given by loose objects or previously-missed packfiles. Also automatically add commits from the existing graph file, if it exists. Signed-off-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

Teach Git to inspect a commit graph file to supply the contents of a struct commit when calling parse_commit_gently(). This implementation satisfies all post-conditions on the struct commit, including loading parents, the root tree, and the commit date. If core.commitGraph is false, then do not check graph files. In test script t5318-commit-graph.sh, add output-matching conditions on read-only graph operations. By loading commits from the graph instead of parsing commit buffers, we save a lot of time on long commit walks. Here are some performance results for a copy of the Linux repository where 'master' has 678,653 reachable commits and is behind 'origin/master' by 59,929 commits. | Command | Before | After | Rel % | |----------------------------------|--------|--------|-------| | log --oneline --topo-order -1000 | 8.31s | 0.94s | -88% | | branch -vv | 1.02s | 0.14s | -86% | | rev-list --all | 5.89s | 1.07s | -81% | | rev-list --all --objects | 66.15s | 58.45s | -11% | Signed-off-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

Teach git-commit-graph to inspect the objects only in a certain list of pack-indexes within the given pack directory. This allows updating the commit graph iteratively. Signed-off-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

In the preceding [1] (pack-objects: move revs out of get_object_list(), 2022-03-22) the "repo_init_revisions()" was moved to cmd_pack_objects() so that it unconditionally took place for all invocations of "git pack-objects". We'd thus start leaking memory, which is easily reproduced in e.g. git.git by feeding e83c516 (Initial revision of "git", the information manager from hell, 2005-04-07) to "git pack-objects"; $ echo e83c516 | ./git pack-objects initial [...] ==19130==ERROR: LeakSanitizer: detected memory leaks Direct leak of 7120 byte(s) in 1 object(s) allocated from: #0 0x455308 in __interceptor_malloc (/home/avar/g/git/git+0x455308) #1 0x75b399 in do_xmalloc /home/avar/g/git/wrapper.c:41:8 #2 0x75b356 in xmalloc /home/avar/g/git/wrapper.c:62:9 #3 0x5d7609 in prep_parse_options /home/avar/g/git/diff.c:5647:2 #4 0x5d415a in repo_diff_setup /home/avar/g/git/diff.c:4621:2 #5 0x6dffbb in repo_init_revisions /home/avar/g/git/revision.c:1853:2 #6 0x4f599d in cmd_pack_objects /home/avar/g/git/builtin/pack-objects.c:3980:2 #7 0x4592ca in run_builtin /home/avar/g/git/git.c:465:11 #8 0x457d81 in handle_builtin /home/avar/g/git/git.c:718:3 #9 0x458ca5 in run_argv /home/avar/g/git/git.c:785:4 #10 0x457b40 in cmd_main /home/avar/g/git/git.c:916:19 #11 0x562259 in main /home/avar/g/git/common-main.c:56:11 #12 0x7fce792ac7ec in __libc_start_main csu/../csu/libc-start.c:332:16 #13 0x4300f9 in _start (/home/avar/g/git/git+0x4300f9) SUMMARY: LeakSanitizer: 7120 byte(s) leaked in 1 allocation(s). Aborted Narrowly fixing that commit would have been easy, just add call repo_init_revisions() right before get_object_list(), which is effectively what was done before that commit. But an unstated constraint when setting it up early is that it was needed for the subsequent [2] (pack-objects: parse --filter directly into revs.filter, 2022-03-22), i.e. we might have a --filter command-line option, and need to either have the "struct rev_info" setup when we encounter that option, or later. Let's just change the control flow so that we'll instead set up the "struct rev_info" only when we need it. Doing so leads to a bit more verbosity, but it's a lot clearer what we're doing and why. An earlier version of this commit[3] went behind opt_parse_list_objects_filter()'s back by faking up a "struct option" before calling it. Let's avoid that and instead create a blessed API for this pattern. We could furthermore combine the two get_object_list() invocations here by having repo_init_revisions() invoked on &pfd.revs, but I think clearly separating the two makes the flow clearer. Likewise redundantly but explicitly (i.e. redundant v.s. a "{ 0 }") "0" to "have_revs" early in cmd_pack_objects(). While we're at it add parentheses around the arguments to the OPT_* macros in in list-objects-filter-options.h, as we need to change those lines anyway. It doesn't matter in this case, but is good general practice. 1. https://lore.kernel.org/git/619b757d98465dbc4995bdc11a5282fbfcbd3daa.1647970119.git.gitgitgadget@gmail.com 2. https://lore.kernel.org/git/97de926904988b89b5663bd4c59c011a1723a8f5.1647970119.git.gitgitgadget@gmail.com 3. https://lore.kernel.org/git/patch-1.1-193534b0f07-20220325T121715Z-avarab@gmail.com/ Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

…checkout` builtin This integrates the `sparse-checkout` builtin with the sparse index. The tricky part here is that we need to partially expand the index when we are modifying the sparse-checkout definition. Note that we modify the pattern list in a careful way: we create a `struct pattern_list` in-memory in `builtin/sparse-checkout.c` then apply those patterns to the index before writing the patterns to the sparse-checkout file. The `update_sparsity()` method does the work to assign the `SKIP_WORKTREE` bit appropriately, but this doesn't work if the files that are within the new sparse-checkout cone are still hidden behind a sparse directory. The new `expand_to_pattern_list()` method does the hard work of expanding the sparse directories that are now within the new patterns. This expands only as far as needed, possibly creating new sparse directory entries. This method does not contract existing files to sparse directories, and a big reason why is because of the check for ignored files as we delete those directories. The `clean_tracked_sparse_directories()` method is called after `update_sparsity()`, but we need to read the `A/B/.gitignore` file (or lack thereof) before we can delete `A/B/`. If we convert to sparse too quickly, then we lose this information and cause a full expansion. Most of the correctness is handled by existing tests in `t1092`, but I add checks for `ensure_not_expanded` in some hopefully interesting cases. As for performance, `git sparse-checkout set` can be slow if it needs to move a lot of files. However, no-op `git sparse-checkout set` (i.e. set the sparse-checkout cone to only include files at root, and do this on repeat) has these performance results on Linux in a monorepo with 2+ million files at `HEAD`: ``` Benchmark #1: baseline Time (mean ± σ): 10.465 s ± 0.018 s [User: 9.885 s, System: 0.573 s] Range (min … max): 10.450 s … 10.497 s 5 runs Benchmark #2: new code Time (mean ± σ): 68.9 ms ± 2.9 ms [User: 45.8 ms, System: 17.1 ms] Range (min … max): 63.4 ms … 74.0 ms 41 runs Summary 'new code' ran 151.89 ± 6.30 times faster than 'baseline' ```

"dst" can legitimately be "0\{40\}" for a creation patch, e.g. when the stat information is stale, but it falls into "look at work tree" case. The original description in b6d8f30 ([PATCH] diff-raw format update take #2., 2005-05-23) forgot that deletion also makes the "dst" 0* SHA-1. Signed-off-by: Philippe Blain <levraiphilippeblain@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

Near the end of the "Raw output format" section, an example shows the output of 'git diff-files' for a tracked file modified on disk but not yet added to the index. However the wording is: <sha1> is shown as all 0's if a file is new on the filesystem and it is out of sync with the index. which is confusing since it can be understood to mean that 'file' is a new, yet untracked file, in which case 'git diff-files' does not care about it at all. When this example was introduced all the way back in c64b9b8 (Reference documentation for the core git commands., 2005-05-05), 'old' and 'new' referred to the two entities being compared, depending on the command being used (diff-index, diff-tree or diff-files - which at the time were diff-cache, diff-tree and show-diff). The wording used at the time was: <new-sha1> is shown as all 0's if new is a file on the filesystem and it is out of sync with the cache. This section was reworked in 81e50ea ([PATCH] The diff-raw format updates., 2005-05-21) and the mention of the meaning of 'new' and 'old' was removed. Then in f73ae1f (Some typos and light editing of various manpages, 2005-10-05), the wording was changed to what it is now. In addition, in b6d8f30 ([PATCH] diff-raw format update take #2., 2005-05-23), the section was further reworked and did not use '<sha1>' anymore, making the example the sole user of this token. Rework the introductory sentence of the example to instead refer to 'sha1 for "dst"', which is what the text description above it uses, and fix the wording so that we do not mention a "new file". While at it, also tweak the wording used in the description of the raw format to explicitely state that all 0's are used for the destination hash if the working tree is out of sync with the index, instead of the more vague "look at worktree". Signed-off-by: Philippe Blain <levraiphilippeblain@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

The two examples in the doc for 'git diff-index' were not updated when the raw output format was changed in 81e50ea ([PATCH] The diff-raw format updates., 2005-05-21) (first example) and in b6d8f30 ([PATCH] diff-raw format update take #2., 2005-05-23) and 7cb6ac1 (diff: diff_aligned_abbrev: remove ellipsis after abbreviated SHA-1 value, 2017-12-03) (second example). Update the output, inventing some characters to complete the source hash in the second example. Also correct the destination mode in the second example, which was wrongly '100664' since the addition of the example in c64b9b8 (Reference documentation for the core git commands., 2005-05-05). Signed-off-by: Philippe Blain <levraiphilippeblain@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

Git uses zlib for its own object store, but calls gzip when creating tgz archives. Add an option to perform the gzip compression for the latter using zlib, without depending on the external gzip binary. Plug it in by making write_block a function pointer and switching to a compressing variant if the filter command has the magic value "git archive gzip". Does that indirection slow down tar creation? Not really, at least not in this test: $ hyperfine -w3 -L rev HEAD,origin/main -p 'git checkout {rev} && make' \ './git -C ../linux archive --format=tar HEAD # {rev}' Benchmark #1: ./git -C ../linux archive --format=tar HEAD # HEAD Time (mean ± σ): 4.044 s ± 0.007 s [User: 3.901 s, System: 0.137 s] Range (min … max): 4.038 s … 4.059 s 10 runs Benchmark #2: ./git -C ../linux archive --format=tar HEAD # origin/main Time (mean ± σ): 4.047 s ± 0.009 s [User: 3.903 s, System: 0.138 s] Range (min … max): 4.038 s … 4.066 s 10 runs How does tgz creation perform? $ hyperfine -w3 -L command 'gzip -cn','git archive gzip' \ './git -c tar.tgz.command="{command}" -C ../linux archive --format=tgz HEAD' Benchmark #1: ./git -c tar.tgz.command="gzip -cn" -C ../linux archive --format=tgz HEAD Time (mean ± σ): 20.404 s ± 0.006 s [User: 23.943 s, System: 0.401 s] Range (min … max): 20.395 s … 20.414 s 10 runs Benchmark #2: ./git -c tar.tgz.command="git archive gzip" -C ../linux archive --format=tgz HEAD Time (mean ± σ): 23.807 s ± 0.023 s [User: 23.655 s, System: 0.145 s] Range (min … max): 23.782 s … 23.857 s 10 runs Summary './git -c tar.tgz.command="gzip -cn" -C ../linux archive --format=tgz HEAD' ran 1.17 ± 0.00 times faster than './git -c tar.tgz.command="git archive gzip" -C ../linux archive --format=tgz HEAD' So the internal implementation takes 17% longer on the Linux repo, but uses 2% less CPU time. That's because the external gzip can run in parallel on its own processor, while the internal one works sequentially and avoids the inter-process communication overhead. What are the benefits? Only an internal sequential implementation can offer this eco mode, and it allows avoiding the gzip(1) requirement. This implementation uses the helper functions from our zlib.c instead of the convenient gz* functions from zlib, because the latter doesn't give the control over the generated gzip header that the next patch requires. Original-patch-by: Rohit Ashiwal <rohit.ashiwal265@gmail.com> Signed-off-by: René Scharfe <l.s.r@web.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>

Since commit fcc07e9 (is_promisor_object(): free tree buffer after parsing, 2021-04-13), we'll always free the buffers attached to a "struct tree" after searching them for promisor links. But there's an important case where we don't want to do so: if somebody else is already using the tree! This can happen during a "rev-list --missing=allow-promisor" traversal in a partial clone that is missing one or more trees or blobs. The backtrace for the free looks like this: #1 free_tree_buffer tree.c:147 #2 add_promisor_object packfile.c:2250 #3 for_each_object_in_pack packfile.c:2190 #4 for_each_packed_object packfile.c:2215 #5 is_promisor_object packfile.c:2272 #6 finish_object__ma builtin/rev-list.c:245 #7 finish_object builtin/rev-list.c:261 #8 show_object builtin/rev-list.c:274 #9 process_blob list-objects.c:63 #10 process_tree_contents list-objects.c:145 #11 process_tree list-objects.c:201 #12 traverse_trees_and_blobs list-objects.c:344 [...] We're in the middle of walking through the entries of a tree object via process_tree_contents(). We see a blob (or it could even be another tree entry) that we don't have, so we call is_promisor_object() to check it. That function loops over all of the objects in the promisor packfile, including the tree we're currently walking. When we're done with it there, we free the tree buffer. But as we return to the walk in process_tree_contents(), it's still holding on to a pointer to that buffer, via its tree_desc iterator, and it accesses the freed memory. Even a trivial use of "--missing=allow-promisor" triggers this problem, as the included test demonstrates (it's just a vanilla --blob:none clone). We can detect this case by only freeing the tree buffer if it was allocated on our behalf. This is a little tricky since that happens inside parse_object(), and it doesn't tell us whether the object was already parsed, or whether it allocated the buffer itself. But by checking for an already-parsed tree beforehand, we can distinguish the two cases. That feels a little hacky, and does incur an extra lookup in the object-hash table. But that cost is fairly minimal compared to actually loading objects (and since we're iterating the whole pack here, we're likely to be loading most objects, rather than reusing cached results). It may also be a good direction for this function in general, as there are other possible optimizations that rely on doing some analysis before parsing: - we could detect blobs and avoid reading their contents; they can't link to other objects, but parse_object() doesn't know that we don't care about checking their hashes. - we could avoid allocating object structs entirely for most objects (since we really only need them in the oidset), which would save some memory. - promisor commits could use the commit-graph rather than loading the object from disk This commit doesn't do any of those optimizations, but I think it argues that this direction is reasonable, rather than relying on parse_object() and trying to teach it to give us more information about whether it parsed. The included test fails reliably under SANITIZE=address just when running "rev-list --missing=allow-promisor". Checking the output isn't strictly necessary to detect the bug, but it seems like a reasonable addition given the general lack of coverage for "allow-promisor" in the test suite. Reported-by: Andrew Olsen <andrew.olsen@koordinates.com> Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>

Fix a memory leak occuring in case of pathspec copy in preload_index. Direct leak of 8 byte(s) in 8 object(s) allocated from: #0 0x7f0a353ead47 in __interceptor_malloc (/usr/lib/gcc/x86_64-pc-linux-gnu/11.3.0/libasan.so.6+0xb5d47) #1 0x55750995e840 in do_xmalloc /home/anthony/src/c/git/wrapper.c:51 #2 0x55750995e840 in xmalloc /home/anthony/src/c/git/wrapper.c:72 #3 0x55750970f824 in copy_pathspec /home/anthony/src/c/git/pathspec.c:684 #4 0x557509717278 in preload_index /home/anthony/src/c/git/preload-index.c:135 #5 0x55750975f21e in refresh_index /home/anthony/src/c/git/read-cache.c:1633 #6 0x55750915b926 in cmd_status builtin/commit.c:1547 #7 0x5575090e1680 in run_builtin /home/anthony/src/c/git/git.c:466 #8 0x5575090e1680 in handle_builtin /home/anthony/src/c/git/git.c:720 #9 0x5575090e284a in run_argv /home/anthony/src/c/git/git.c:787 #10 0x5575090e284a in cmd_main /home/anthony/src/c/git/git.c:920 #11 0x5575090dbf82 in main /home/anthony/src/c/git/common-main.c:56 #12 0x7f0a348230ab (/lib64/libc.so.6+0x290ab) Signed-off-by: Anthony Delannoy <anthony.2lannoy@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

…checkout` builtin This integrates the `sparse-checkout` builtin with the sparse index. The tricky part here is that we need to partially expand the index when we are modifying the sparse-checkout definition. Note that we modify the pattern list in a careful way: we create a `struct pattern_list` in-memory in `builtin/sparse-checkout.c` then apply those patterns to the index before writing the patterns to the sparse-checkout file. The `update_sparsity()` method does the work to assign the `SKIP_WORKTREE` bit appropriately, but this doesn't work if the files that are within the new sparse-checkout cone are still hidden behind a sparse directory. The new `expand_to_pattern_list()` method does the hard work of expanding the sparse directories that are now within the new patterns. This expands only as far as needed, possibly creating new sparse directory entries. This method does not contract existing files to sparse directories, and a big reason why is because of the check for ignored files as we delete those directories. The `clean_tracked_sparse_directories()` method is called after `update_sparsity()`, but we need to read the `A/B/.gitignore` file (or lack thereof) before we can delete `A/B/`. If we convert to sparse too quickly, then we lose this information and cause a full expansion. Most of the correctness is handled by existing tests in `t1092`, but I add checks for `ensure_not_expanded` in some hopefully interesting cases. As for performance, `git sparse-checkout set` can be slow if it needs to move a lot of files. However, no-op `git sparse-checkout set` (i.e. set the sparse-checkout cone to only include files at root, and do this on repeat) has these performance results on Linux in a monorepo with 2+ million files at `HEAD`: ``` Benchmark #1: baseline Time (mean ± σ): 10.465 s ± 0.018 s [User: 9.885 s, System: 0.573 s] Range (min … max): 10.450 s … 10.497 s 5 runs Benchmark #2: new code Time (mean ± σ): 68.9 ms ± 2.9 ms [User: 45.8 ms, System: 17.1 ms] Range (min … max): 63.4 ms … 74.0 ms 41 runs Summary 'new code' ran 151.89 ± 6.30 times faster than 'baseline' ```

In the tmp-objdir api, tmp_objdir_create will create a temporary directory but also register signal handlers responsible for removing the directory's contents and the directory itself. However, the function responsible for recursively removing the contents and directory, remove_dir_recurse() calls opendir(3) and closedir(3). This can be problematic because these functions allocate and free memory, which are not async-signal-safe functions. This can lead to deadlocks. One place we call tmp_objdir_create() is in git-receive-pack, where we create a temporary quarantine directory "incoming". Incoming objects will be written to this directory before they get moved to the object directory. We have observed this code leading to a deadlock: Thread 1 (Thread 0x7f621ba0b200 (LWP 326305)): #0 __lll_lock_wait_private (futex=futex@entry=0x7f621bbf8b80 <main_arena>) at ./lowlevellock.c:35 #1 0x00007f621baa635b in __GI___libc_malloc (bytes=bytes@entry=32816) at malloc.c:3064 #2 0x00007f621bae9f49 in __alloc_dir (statp=0x7fff2ea7ed60, flags=0, close_fd=true, fd=5) at ../sysdeps/posix/opendir.c:118 #3 opendir_tail (fd=5) at ../sysdeps/posix/opendir.c:69 #4 __opendir (name=<optimized out>) at ../sysdeps/posix/opendir.c:92 #5 0x0000557c19c77de1 in remove_dir_recurse () git#6 0x0000557c19d81a4f in remove_tmp_objdir_on_signal () #7 <signal handler called> git#8 _int_malloc (av=av@entry=0x7f621bbf8b80 <main_arena>, bytes=bytes@entry=7160) at malloc.c:4116 git#9 0x00007f621baa62c9 in __GI___libc_malloc (bytes=7160) at malloc.c:3066 git#10 0x00007f621bd1e987 in inflateInit2_ () from /opt/gitlab/embedded/lib/libz.so.1 git#11 0x0000557c19dbe5f4 in git_inflate_init () git#12 0x0000557c19cee02a in unpack_compressed_entry () git#13 0x0000557c19cf08cb in unpack_entry () git#14 0x0000557c19cf0f32 in packed_object_info () git#15 0x0000557c19cd68cd in do_oid_object_info_extended () git#16 0x0000557c19cd6e2b in read_object_file_extended () git#17 0x0000557c19cdec2f in parse_object () git#18 0x0000557c19c34977 in lookup_commit_reference_gently () git#19 0x0000557c19d69309 in mark_uninteresting () git#20 0x0000557c19d2d180 in do_for_each_repo_ref_iterator () git#21 0x0000557c19d21678 in for_each_ref () git#22 0x0000557c19d6a94f in assign_shallow_commits_to_refs () git#23 0x0000557c19bc02b2 in cmd_receive_pack () git#24 0x0000557c19b29fdd in handle_builtin () git#25 0x0000557c19b2a526 in cmd_main () git#26 0x0000557c19b28ea2 in main () Since we can't do the cleanup in a portable and signal-safe way, skip the cleanup when we're handling a signal. This means that when signal handling, the temporary directory may not get cleaned up properly. This is mitigated by b3cecf4 (tmp-objdir: new API for creating temporary writable databases, 2021-12-06) which changed the default name and allows gc to clean up these temporary directories. In the event of a normal exit, we should still be cleaning up via the atexit() handler. Helped-by: Jeff King <peff@peff.net> Signed-off-by: John Cai <johncai86@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

There is an out-of-bounds read possible when parsing gitattributes that have an attribute that is 2^31+1 bytes long. This is caused due to an integer overflow when we assign the result of strlen(3P) to an `int`, where we use the wrapped-around value in a subsequent call to memcpy(3P). The following code reproduces the issue: blob=$(perl -e 'print "a" x 2147483649 . " attr"' | git hash-object -w --stdin) git update-index --add --cacheinfo 100644,$blob,.gitattributes git check-attr --all file AddressSanitizer:DEADLYSIGNAL ================================================================= ==8451==ERROR: AddressSanitizer: SEGV on unknown address 0x7f93efa00800 (pc 0x7f94f1f8f082 bp 0x7ffddb59b3a0 sp 0x7ffddb59ab28 T0) ==8451==The signal is caused by a READ memory access. #0 0x7f94f1f8f082 (/usr/lib/libc.so.6+0x176082) #1 0x7f94f2047d9c in __interceptor_strspn /usr/src/debug/gcc/libsanitizer/sanitizer_common/sanitizer_common_interceptors.inc:752 #2 0x560e190f7f26 in parse_attr_line attr.c:375 #3 0x560e190f9663 in handle_attr_line attr.c:660 #4 0x560e190f9ddd in read_attr_from_index attr.c:769 #5 0x560e190f9f14 in read_attr attr.c:797 #6 0x560e190fa24e in bootstrap_attr_stack attr.c:867 #7 0x560e190fa4a5 in prepare_attr_stack attr.c:902 #8 0x560e190fb5dc in collect_some_attrs attr.c:1097 #9 0x560e190fb93f in git_all_attrs attr.c:1128 #10 0x560e18e6136e in check_attr builtin/check-attr.c:67 #11 0x560e18e61c12 in cmd_check_attr builtin/check-attr.c:183 #12 0x560e18e15993 in run_builtin git.c:466 #13 0x560e18e16397 in handle_builtin git.c:721 #14 0x560e18e16b2b in run_argv git.c:788 #15 0x560e18e17991 in cmd_main git.c:926 #16 0x560e190ae2bd in main common-main.c:57 #17 0x7f94f1e3c28f (/usr/lib/libc.so.6+0x2328f) #18 0x7f94f1e3c349 in __libc_start_main (/usr/lib/libc.so.6+0x23349) #19 0x560e18e110e4 in _start ../sysdeps/x86_64/start.S:115 AddressSanitizer can not provide additional info. SUMMARY: AddressSanitizer: SEGV (/usr/lib/libc.so.6+0x176082) ==8451==ABORTING Fix this bug by converting the variable to a `size_t` instead. Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>

It is possible to trigger an integer overflow when parsing attribute names when there are more than 2^31 of them for a single pattern. This can either lead to us dying due to trying to request too many bytes: blob=$(perl -e 'print "f" . " a=" x 2147483649' | git hash-object -w --stdin) git update-index --add --cacheinfo 100644,$blob,.gitattributes git attr-check --all file ================================================================= ==1022==ERROR: AddressSanitizer: requested allocation size 0xfffffff800000032 (0xfffffff800001038 after adjustments for alignment, red zones etc.) exceeds maximum supported size of 0x10000000000 (thread T0) #0 0x7fd3efabf411 in __interceptor_calloc /usr/src/debug/gcc/libsanitizer/asan/asan_malloc_linux.cpp:77 #1 0x5563a0a1e3d3 in xcalloc wrapper.c:150 #2 0x5563a058d005 in parse_attr_line attr.c:384 #3 0x5563a058e661 in handle_attr_line attr.c:660 #4 0x5563a058eddb in read_attr_from_index attr.c:769 #5 0x5563a058ef12 in read_attr attr.c:797 #6 0x5563a058f24c in bootstrap_attr_stack attr.c:867 #7 0x5563a058f4a3 in prepare_attr_stack attr.c:902 #8 0x5563a05905da in collect_some_attrs attr.c:1097 #9 0x5563a059093d in git_all_attrs attr.c:1128 #10 0x5563a02f636e in check_attr builtin/check-attr.c:67 #11 0x5563a02f6c12 in cmd_check_attr builtin/check-attr.c:183 #12 0x5563a02aa993 in run_builtin git.c:466 #13 0x5563a02ab397 in handle_builtin git.c:721 #14 0x5563a02abb2b in run_argv git.c:788 #15 0x5563a02ac991 in cmd_main git.c:926 #16 0x5563a05432bd in main common-main.c:57 #17 0x7fd3ef82228f (/usr/lib/libc.so.6+0x2328f) ==1022==HINT: if you don't care about these errors you may set allocator_may_return_null=1 SUMMARY: AddressSanitizer: allocation-size-too-big /usr/src/debug/gcc/libsanitizer/asan/asan_malloc_linux.cpp:77 in __interceptor_calloc ==1022==ABORTING Or, much worse, it can lead to an out-of-bounds write because we underallocate and then memcpy(3P) into an array: perl -e ' print "A " . "\rh="x2000000000; print "\rh="x2000000000; print "\rh="x294967294 . "\n" ' >.gitattributes git add .gitattributes git commit -am "evil attributes" $ git clone --quiet /path/to/repo ================================================================= ==15062==ERROR: AddressSanitizer: heap-buffer-overflow on address 0x602000002550 at pc 0x5555559884d5 bp 0x7fffffffbc60 sp 0x7fffffffbc58 WRITE of size 8 at 0x602000002550 thread T0 #0 0x5555559884d4 in parse_attr_line attr.c:393 #1 0x5555559884d4 in handle_attr_line attr.c:660 #2 0x555555988902 in read_attr_from_index attr.c:784 #3 0x555555988902 in read_attr_from_index attr.c:747 #4 0x555555988a1d in read_attr attr.c:800 #5 0x555555989b0c in bootstrap_attr_stack attr.c:882 #6 0x555555989b0c in prepare_attr_stack attr.c:917 #7 0x555555989b0c in collect_some_attrs attr.c:1112 #8 0x55555598b141 in git_check_attr attr.c:1126 #9 0x555555a13004 in convert_attrs convert.c:1311 #10 0x555555a95e04 in checkout_entry_ca entry.c:553 #11 0x555555d58bf6 in checkout_entry entry.h:42 #12 0x555555d58bf6 in check_updates unpack-trees.c:480 #13 0x555555d5eb55 in unpack_trees unpack-trees.c:2040 #14 0x555555785ab7 in checkout builtin/clone.c:724 #15 0x555555785ab7 in cmd_clone builtin/clone.c:1384 #16 0x55555572443c in run_builtin git.c:466 #17 0x55555572443c in handle_builtin git.c:721 #18 0x555555727872 in run_argv git.c:788 #19 0x555555727872 in cmd_main git.c:926 #20 0x555555721fa0 in main common-main.c:57 #21 0x7ffff73f1d09 in __libc_start_main ../csu/libc-start.c:308 #22 0x555555723f39 in _start (git+0x1cff39) 0x602000002552 is located 0 bytes to the right of 2-byte region [0x602000002550,0x602000002552) allocated by thread T0 here: #0 0x7ffff768c037 in __interceptor_calloc ../../../../src/libsanitizer/asan/asan_malloc_linux.cpp:154 #1 0x555555d7fff7 in xcalloc wrapper.c:150 #2 0x55555598815f in parse_attr_line attr.c:384 #3 0x55555598815f in handle_attr_line attr.c:660 #4 0x555555988902 in read_attr_from_index attr.c:784 #5 0x555555988902 in read_attr_from_index attr.c:747 #6 0x555555988a1d in read_attr attr.c:800 #7 0x555555989b0c in bootstrap_attr_stack attr.c:882 #8 0x555555989b0c in prepare_attr_stack attr.c:917 #9 0x555555989b0c in collect_some_attrs attr.c:1112 #10 0x55555598b141 in git_check_attr attr.c:1126 #11 0x555555a13004 in convert_attrs convert.c:1311 #12 0x555555a95e04 in checkout_entry_ca entry.c:553 #13 0x555555d58bf6 in checkout_entry entry.h:42 #14 0x555555d58bf6 in check_updates unpack-trees.c:480 #15 0x555555d5eb55 in unpack_trees unpack-trees.c:2040 #16 0x555555785ab7 in checkout builtin/clone.c:724 #17 0x555555785ab7 in cmd_clone builtin/clone.c:1384 #18 0x55555572443c in run_builtin git.c:466 #19 0x55555572443c in handle_builtin git.c:721 #20 0x555555727872 in run_argv git.c:788 #21 0x555555727872 in cmd_main git.c:926 #22 0x555555721fa0 in main common-main.c:57 #23 0x7ffff73f1d09 in __libc_start_main ../csu/libc-start.c:308 SUMMARY: AddressSanitizer: heap-buffer-overflow attr.c:393 in parse_attr_line Shadow bytes around the buggy address: 0x0c047fff8450: fa fa 00 02 fa fa 00 07 fa fa fd fd fa fa 00 00 0x0c047fff8460: fa fa 02 fa fa fa fd fd fa fa 00 06 fa fa 05 fa 0x0c047fff8470: fa fa fd fd fa fa 00 02 fa fa 06 fa fa fa 05 fa 0x0c047fff8480: fa fa 07 fa fa fa fd fd fa fa 00 01 fa fa 00 02 0x0c047fff8490: fa fa 00 03 fa fa 00 fa fa fa 00 01 fa fa 00 03 =>0x0c047fff84a0: fa fa 00 01 fa fa 00 02 fa fa[02]fa fa fa fa fa 0x0c047fff84b0: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x0c047fff84c0: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x0c047fff84d0: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x0c047fff84e0: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x0c047fff84f0: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa Shadow byte legend (one shadow byte represents 8 application bytes): Addressable: 00 Partially addressable: 01 02 03 04 05 06 07 Heap left redzone: fa Freed heap region: fd Stack left redzone: f1 Stack mid redzone: f2 Stack right redzone: f3 Stack after return: f5 Stack use after scope: f8 Global redzone: f9 Global init order: f6 Poisoned by user: f7 Container overflow: fc Array cookie: ac Intra object redzone: bb ASan internal: fe Left alloca redzone: ca Right alloca redzone: cb Shadow gap: cc ==15062==ABORTING Fix this bug by using `size_t` instead to count the number of attributes so that this value cannot reasonably overflow without running out of memory before already. Reported-by: Markus Vervier <markus.vervier@x41-dsec.de> Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>

When using a padding specifier in the pretty format passed to git-log(1) we need to calculate the string length in several places. These string lengths are stored in `int`s though, which means that these can easily overflow when the input lengths exceeds 2GB. This can ultimately lead to an out-of-bounds write when these are used in a call to memcpy(3P): ==8340==ERROR: AddressSanitizer: heap-buffer-overflow on address 0x7f1ec62f97fe at pc 0x7f2127e5f427 bp 0x7ffd3bd63de0 sp 0x7ffd3bd63588 WRITE of size 1 at 0x7f1ec62f97fe thread T0 #0 0x7f2127e5f426 in __interceptor_memcpy /usr/src/debug/gcc/libsanitizer/sanitizer_common/sanitizer_common_interceptors.inc:827 #1 0x5628e96aa605 in format_and_pad_commit pretty.c:1762 #2 0x5628e96aa7f4 in format_commit_item pretty.c:1801 #3 0x5628e97cdb24 in strbuf_expand strbuf.c:429 #4 0x5628e96ab060 in repo_format_commit_message pretty.c:1869 #5 0x5628e96acd0f in pretty_print_commit pretty.c:2161 #6 0x5628e95a44c8 in show_log log-tree.c:781 #7 0x5628e95a76ba in log_tree_commit log-tree.c:1117 #8 0x5628e922bed5 in cmd_log_walk_no_free builtin/log.c:508 #9 0x5628e922c35b in cmd_log_walk builtin/log.c:549 #10 0x5628e922f1a2 in cmd_log builtin/log.c:883 #11 0x5628e9106993 in run_builtin git.c:466 #12 0x5628e9107397 in handle_builtin git.c:721 #13 0x5628e9107b07 in run_argv git.c:788 #14 0x5628e91088a7 in cmd_main git.c:923 #15 0x5628e939d682 in main common-main.c:57 #16 0x7f2127c3c28f (/usr/lib/libc.so.6+0x2328f) #17 0x7f2127c3c349 in __libc_start_main (/usr/lib/libc.so.6+0x23349) #18 0x5628e91020e4 in _start ../sysdeps/x86_64/start.S:115 0x7f1ec62f97fe is located 2 bytes to the left of 4831838265-byte region [0x7f1ec62f9800,0x7f1fe62f9839) allocated by thread T0 here: #0 0x7f2127ebe7ea in __interceptor_realloc /usr/src/debug/gcc/libsanitizer/asan/asan_malloc_linux.cpp:85 #1 0x5628e98774d4 in xrealloc wrapper.c:136 #2 0x5628e97cb01c in strbuf_grow strbuf.c:99 #3 0x5628e97ccd42 in strbuf_addchars strbuf.c:327 #4 0x5628e96aa55c in format_and_pad_commit pretty.c:1761 #5 0x5628e96aa7f4 in format_commit_item pretty.c:1801 #6 0x5628e97cdb24 in strbuf_expand strbuf.c:429 #7 0x5628e96ab060 in repo_format_commit_message pretty.c:1869 #8 0x5628e96acd0f in pretty_print_commit pretty.c:2161 #9 0x5628e95a44c8 in show_log log-tree.c:781 #10 0x5628e95a76ba in log_tree_commit log-tree.c:1117 #11 0x5628e922bed5 in cmd_log_walk_no_free builtin/log.c:508 #12 0x5628e922c35b in cmd_log_walk builtin/log.c:549 #13 0x5628e922f1a2 in cmd_log builtin/log.c:883 #14 0x5628e9106993 in run_builtin git.c:466 #15 0x5628e9107397 in handle_builtin git.c:721 #16 0x5628e9107b07 in run_argv git.c:788 #17 0x5628e91088a7 in cmd_main git.c:923 #18 0x5628e939d682 in main common-main.c:57 #19 0x7f2127c3c28f (/usr/lib/libc.so.6+0x2328f) #20 0x7f2127c3c349 in __libc_start_main (/usr/lib/libc.so.6+0x23349) #21 0x5628e91020e4 in _start ../sysdeps/x86_64/start.S:115 SUMMARY: AddressSanitizer: heap-buffer-overflow /usr/src/debug/gcc/libsanitizer/sanitizer_common/sanitizer_common_interceptors.inc:827 in __interceptor_memcpy Shadow bytes around the buggy address: 0x0fe458c572a0: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x0fe458c572b0: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x0fe458c572c0: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x0fe458c572d0: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x0fe458c572e0: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa =>0x0fe458c572f0: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa[fa] 0x0fe458c57300: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 0x0fe458c57310: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 0x0fe458c57320: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 0x0fe458c57330: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 0x0fe458c57340: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 Shadow byte legend (one shadow byte represents 8 application bytes): Addressable: 00 Partially addressable: 01 02 03 04 05 06 07 Heap left redzone: fa Freed heap region: fd Stack left redzone: f1 Stack mid redzone: f2 Stack right redzone: f3 Stack after return: f5 Stack use after scope: f8 Global redzone: f9 Global init order: f6 Poisoned by user: f7 Container overflow: fc Array cookie: ac Intra object redzone: bb ASan internal: fe Left alloca redzone: ca Right alloca redzone: cb ==8340==ABORTING The pretty format can also be used in `git archive` operations via the `export-subst` attribute. So this is what in our opinion makes this a critical issue in the context of Git forges which allow to download an archive of user supplied Git repositories. Fix this vulnerability by using `size_t` instead of `int` to track the string lengths. Add tests which detect this vulnerability when Git is compiled with the address sanitizer. Reported-by: Joern Schneeweisz <jschneeweisz@gitlab.com> Original-patch-by: Joern Schneeweisz <jschneeweisz@gitlab.com> Modified-by: Taylor Blau <me@ttalorr.com> Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>

With the `%>>(<N>)` pretty formatter, you can ask git-log(1) et al to steal spaces. To do so we need to look ahead of the next token to see whether there are spaces there. This loop takes into account ANSI sequences that end with an `m`, and if it finds any it will skip them until it finds the first space. While doing so it does not take into account the buffer's limits though and easily does an out-of-bounds read. Add a test that hits this behaviour. While we don't have an easy way to verify this, the test causes the following failure when run with `SANITIZE=address`: ==37941==ERROR: AddressSanitizer: heap-buffer-overflow on address 0x603000000baf at pc 0x55ba6f88e0d0 bp 0x7ffc84c50d20 sp 0x7ffc84c50d10 READ of size 1 at 0x603000000baf thread T0 #0 0x55ba6f88e0cf in format_and_pad_commit pretty.c:1712 #1 0x55ba6f88e7b4 in format_commit_item pretty.c:1801 #2 0x55ba6f9b1ae4 in strbuf_expand strbuf.c:429 #3 0x55ba6f88f020 in repo_format_commit_message pretty.c:1869 #4 0x55ba6f890ccf in pretty_print_commit pretty.c:2161 #5 0x55ba6f7884c8 in show_log log-tree.c:781 #6 0x55ba6f78b6ba in log_tree_commit log-tree.c:1117 #7 0x55ba6f40fed5 in cmd_log_walk_no_free builtin/log.c:508 #8 0x55ba6f41035b in cmd_log_walk builtin/log.c:549 #9 0x55ba6f4131a2 in cmd_log builtin/log.c:883 #10 0x55ba6f2ea993 in run_builtin git.c:466 #11 0x55ba6f2eb397 in handle_builtin git.c:721 #12 0x55ba6f2ebb07 in run_argv git.c:788 #13 0x55ba6f2ec8a7 in cmd_main git.c:923 #14 0x55ba6f581682 in main common-main.c:57 #15 0x7f2d08c3c28f (/usr/lib/libc.so.6+0x2328f) #16 0x7f2d08c3c349 in __libc_start_main (/usr/lib/libc.so.6+0x23349) #17 0x55ba6f2e60e4 in _start ../sysdeps/x86_64/start.S:115 0x603000000baf is located 1 bytes to the left of 24-byte region [0x603000000bb0,0x603000000bc8) allocated by thread T0 here: #0 0x7f2d08ebe7ea in __interceptor_realloc /usr/src/debug/gcc/libsanitizer/asan/asan_malloc_linux.cpp:85 #1 0x55ba6fa5b494 in xrealloc wrapper.c:136 #2 0x55ba6f9aefdc in strbuf_grow strbuf.c:99 #3 0x55ba6f9b0a06 in strbuf_add strbuf.c:298 #4 0x55ba6f9b1a25 in strbuf_expand strbuf.c:418 #5 0x55ba6f88f020 in repo_format_commit_message pretty.c:1869 #6 0x55ba6f890ccf in pretty_print_commit pretty.c:2161 #7 0x55ba6f7884c8 in show_log log-tree.c:781 #8 0x55ba6f78b6ba in log_tree_commit log-tree.c:1117 #9 0x55ba6f40fed5 in cmd_log_walk_no_free builtin/log.c:508 #10 0x55ba6f41035b in cmd_log_walk builtin/log.c:549 #11 0x55ba6f4131a2 in cmd_log builtin/log.c:883 #12 0x55ba6f2ea993 in run_builtin git.c:466 #13 0x55ba6f2eb397 in handle_builtin git.c:721 #14 0x55ba6f2ebb07 in run_argv git.c:788 #15 0x55ba6f2ec8a7 in cmd_main git.c:923 #16 0x55ba6f581682 in main common-main.c:57 #17 0x7f2d08c3c28f (/usr/lib/libc.so.6+0x2328f) #18 0x7f2d08c3c349 in __libc_start_main (/usr/lib/libc.so.6+0x23349) #19 0x55ba6f2e60e4 in _start ../sysdeps/x86_64/start.S:115 SUMMARY: AddressSanitizer: heap-buffer-overflow pretty.c:1712 in format_and_pad_commit Shadow bytes around the buggy address: 0x0c067fff8120: fa fa fd fd fd fa fa fa fd fd fd fa fa fa fd fd 0x0c067fff8130: fd fd fa fa fd fd fd fd fa fa fd fd fd fa fa fa 0x0c067fff8140: fd fd fd fa fa fa fd fd fd fa fa fa fd fd fd fa 0x0c067fff8150: fa fa fd fd fd fd fa fa 00 00 00 fa fa fa fd fd 0x0c067fff8160: fd fa fa fa fd fd fd fa fa fa fd fd fd fa fa fa =>0x0c067fff8170: fd fd fd fa fa[fa]00 00 00 fa fa fa 00 00 00 fa 0x0c067fff8180: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x0c067fff8190: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x0c067fff81a0: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x0c067fff81b0: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x0c067fff81c0: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa Shadow byte legend (one shadow byte represents 8 application bytes): Addressable: 00 Partially addressable: 01 02 03 04 05 06 07 Heap left redzone: fa Freed heap region: fd Stack left redzone: f1 Stack mid redzone: f2 Stack right redzone: f3 Stack after return: f5 Stack use after scope: f8 Global redzone: f9 Global init order: f6 Poisoned by user: f7 Container overflow: fc Array cookie: ac Intra object redzone: bb ASan internal: fe Left alloca redzone: ca Right alloca redzone: cb Luckily enough, this would only cause us to copy the out-of-bounds data into the formatted commit in case we really had an ANSI sequence preceding our buffer. So this bug likely has no security consequences. Fix it regardless by not traversing past the buffer's start. Reported-by: Patrick Steinhardt <ps@pks.im> Reported-by: Eric Sesterhenn <eric.sesterhenn@x41-dsec.de> Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>

An out-of-bounds read can be triggered when parsing an incomplete padding format string passed via `--pretty=format` or in Git archives when files are marked with the `export-subst` gitattribute. This bug exists since we have introduced support for truncating output via the `trunc` keyword a7f01c6 (pretty: support truncating in %>, %< and %><, 2013-04-19). Before this commit, we used to find the end of the formatting string by using strchr(3P). This function returns a `NULL` pointer in case the character in question wasn't found. The subsequent check whether any character was found thus simply checked the returned pointer. After the commit we switched to strcspn(3P) though, which only returns the offset to the first found character or to the trailing NUL byte. As the end pointer is now computed by adding the offset to the start pointer it won't be `NULL` anymore, and as a consequence the check doesn't do anything anymore. The out-of-bounds data that is being read can in fact end up in the formatted string. As a consequence, it is possible to leak memory contents either by calling git-log(1) or via git-archive(1) when any of the archived files is marked with the `export-subst` gitattribute. ==10888==ERROR: AddressSanitizer: heap-buffer-overflow on address 0x602000000398 at pc 0x7f0356047cb2 bp 0x7fff3ffb95d0 sp 0x7fff3ffb8d78 READ of size 1 at 0x602000000398 thread T0 #0 0x7f0356047cb1 in __interceptor_strchrnul /usr/src/debug/gcc/libsanitizer/sanitizer_common/sanitizer_common_interceptors.inc:725 #1 0x563b7cec9a43 in strbuf_expand strbuf.c:417 #2 0x563b7cda7060 in repo_format_commit_message pretty.c:1869 #3 0x563b7cda8d0f in pretty_print_commit pretty.c:2161 #4 0x563b7cca04c8 in show_log log-tree.c:781 #5 0x563b7cca36ba in log_tree_commit log-tree.c:1117 #6 0x563b7c927ed5 in cmd_log_walk_no_free builtin/log.c:508 #7 0x563b7c92835b in cmd_log_walk builtin/log.c:549 #8 0x563b7c92b1a2 in cmd_log builtin/log.c:883 #9 0x563b7c802993 in run_builtin git.c:466 #10 0x563b7c803397 in handle_builtin git.c:721 #11 0x563b7c803b07 in run_argv git.c:788 #12 0x563b7c8048a7 in cmd_main git.c:923 #13 0x563b7ca99682 in main common-main.c:57 #14 0x7f0355e3c28f (/usr/lib/libc.so.6+0x2328f) #15 0x7f0355e3c349 in __libc_start_main (/usr/lib/libc.so.6+0x23349) #16 0x563b7c7fe0e4 in _start ../sysdeps/x86_64/start.S:115 0x602000000398 is located 0 bytes to the right of 8-byte region [0x602000000390,0x602000000398) allocated by thread T0 here: #0 0x7f0356072faa in __interceptor_strdup /usr/src/debug/gcc/libsanitizer/asan/asan_interceptors.cpp:439 #1 0x563b7cf7317c in xstrdup wrapper.c:39 #2 0x563b7cd9a06a in save_user_format pretty.c:40 #3 0x563b7cd9b3e5 in get_commit_format pretty.c:173 #4 0x563b7ce54ea0 in handle_revision_opt revision.c:2456 #5 0x563b7ce597c9 in setup_revisions revision.c:2850 #6 0x563b7c9269e0 in cmd_log_init_finish builtin/log.c:269 #7 0x563b7c927362 in cmd_log_init builtin/log.c:348 #8 0x563b7c92b193 in cmd_log builtin/log.c:882 #9 0x563b7c802993 in run_builtin git.c:466 #10 0x563b7c803397 in handle_builtin git.c:721 #11 0x563b7c803b07 in run_argv git.c:788 #12 0x563b7c8048a7 in cmd_main git.c:923 #13 0x563b7ca99682 in main common-main.c:57 #14 0x7f0355e3c28f (/usr/lib/libc.so.6+0x2328f) #15 0x7f0355e3c349 in __libc_start_main (/usr/lib/libc.so.6+0x23349) #16 0x563b7c7fe0e4 in _start ../sysdeps/x86_64/start.S:115 SUMMARY: AddressSanitizer: heap-buffer-overflow /usr/src/debug/gcc/libsanitizer/sanitizer_common/sanitizer_common_interceptors.inc:725 in __interceptor_strchrnul Shadow bytes around the buggy address: 0x0c047fff8020: fa fa fd fd fa fa 00 06 fa fa 05 fa fa fa fd fd 0x0c047fff8030: fa fa 00 02 fa fa 06 fa fa fa 05 fa fa fa fd fd 0x0c047fff8040: fa fa 00 07 fa fa 03 fa fa fa fd fd fa fa 00 00 0x0c047fff8050: fa fa 00 01 fa fa fd fd fa fa 00 00 fa fa 00 01 0x0c047fff8060: fa fa 00 06 fa fa 00 06 fa fa 05 fa fa fa 05 fa =>0x0c047fff8070: fa fa 00[fa]fa fa fd fa fa fa fd fd fa fa fd fd 0x0c047fff8080: fa fa fd fd fa fa 00 00 fa fa 00 fa fa fa fd fa 0x0c047fff8090: fa fa fd fd fa fa 00 00 fa fa fa fa fa fa fa fa 0x0c047fff80a0: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x0c047fff80b0: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x0c047fff80c0: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa Shadow byte legend (one shadow byte represents 8 application bytes): Addressable: 00 Partially addressable: 01 02 03 04 05 06 07 Heap left redzone: fa Freed heap region: fd Stack left redzone: f1 Stack mid redzone: f2 Stack right redzone: f3 Stack after return: f5 Stack use after scope: f8 Global redzone: f9 Global init order: f6 Poisoned by user: f7 Container overflow: fc Array cookie: ac Intra object redzone: bb ASan internal: fe Left alloca redzone: ca Right alloca redzone: cb ==10888==ABORTING Fix this bug by checking whether `end` points at the trailing NUL byte. Add a test which catches this out-of-bounds read and which demonstrates that we used to write out-of-bounds data into the formatted message. Reported-by: Markus Vervier <markus.vervier@x41-dsec.de> Original-patch-by: Markus Vervier <markus.vervier@x41-dsec.de> Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>

The return type of both `utf8_strwidth()` and `utf8_strnwidth()` is `int`, but we operate on string lengths which are typically of type `size_t`. This means that when the string is longer than `INT_MAX`, we will overflow and thus return a negative result. This can lead to an out-of-bounds write with `--pretty=format:%<1)%B` and a commit message that is 2^31+1 bytes long: ================================================================= ==26009==ERROR: AddressSanitizer: heap-buffer-overflow on address 0x603000001168 at pc 0x7f95c4e5f427 bp 0x7ffd8541c900 sp 0x7ffd8541c0a8 WRITE of size 2147483649 at 0x603000001168 thread T0 #0 0x7f95c4e5f426 in __interceptor_memcpy /usr/src/debug/gcc/libsanitizer/sanitizer_common/sanitizer_common_interceptors.inc:827 #1 0x5612bbb1068c in format_and_pad_commit pretty.c:1763 #2 0x5612bbb1087a in format_commit_item pretty.c:1801 #3 0x5612bbc33bab in strbuf_expand strbuf.c:429 #4 0x5612bbb110e7 in repo_format_commit_message pretty.c:1869 #5 0x5612bbb12d96 in pretty_print_commit pretty.c:2161 #6 0x5612bba0a4d5 in show_log log-tree.c:781 #7 0x5612bba0d6c7 in log_tree_commit log-tree.c:1117 #8 0x5612bb691ed5 in cmd_log_walk_no_free builtin/log.c:508 #9 0x5612bb69235b in cmd_log_walk builtin/log.c:549 #10 0x5612bb6951a2 in cmd_log builtin/log.c:883 #11 0x5612bb56c993 in run_builtin git.c:466 #12 0x5612bb56d397 in handle_builtin git.c:721 #13 0x5612bb56db07 in run_argv git.c:788 #14 0x5612bb56e8a7 in cmd_main git.c:923 #15 0x5612bb803682 in main common-main.c:57 #16 0x7f95c4c3c28f (/usr/lib/libc.so.6+0x2328f) #17 0x7f95c4c3c349 in __libc_start_main (/usr/lib/libc.so.6+0x23349) #18 0x5612bb5680e4 in _start ../sysdeps/x86_64/start.S:115 0x603000001168 is located 0 bytes to the right of 24-byte region [0x603000001150,0x603000001168) allocated by thread T0 here: #0 0x7f95c4ebe7ea in __interceptor_realloc /usr/src/debug/gcc/libsanitizer/asan/asan_malloc_linux.cpp:85 #1 0x5612bbcdd556 in xrealloc wrapper.c:136 #2 0x5612bbc310a3 in strbuf_grow strbuf.c:99 #3 0x5612bbc32acd in strbuf_add strbuf.c:298 #4 0x5612bbc33aec in strbuf_expand strbuf.c:418 #5 0x5612bbb110e7 in repo_format_commit_message pretty.c:1869 #6 0x5612bbb12d96 in pretty_print_commit pretty.c:2161 #7 0x5612bba0a4d5 in show_log log-tree.c:781 #8 0x5612bba0d6c7 in log_tree_commit log-tree.c:1117 #9 0x5612bb691ed5 in cmd_log_walk_no_free builtin/log.c:508 #10 0x5612bb69235b in cmd_log_walk builtin/log.c:549 #11 0x5612bb6951a2 in cmd_log builtin/log.c:883 #12 0x5612bb56c993 in run_builtin git.c:466 #13 0x5612bb56d397 in handle_builtin git.c:721 #14 0x5612bb56db07 in run_argv git.c:788 #15 0x5612bb56e8a7 in cmd_main git.c:923 #16 0x5612bb803682 in main common-main.c:57 #17 0x7f95c4c3c28f (/usr/lib/libc.so.6+0x2328f) SUMMARY: AddressSanitizer: heap-buffer-overflow /usr/src/debug/gcc/libsanitizer/sanitizer_common/sanitizer_common_interceptors.inc:827 in __interceptor_memcpy Shadow bytes around the buggy address: 0x0c067fff81d0: fd fd fd fa fa fa fd fd fd fa fa fa fd fd fd fa 0x0c067fff81e0: fa fa fd fd fd fd fa fa fd fd fd fd fa fa fd fd 0x0c067fff81f0: fd fa fa fa fd fd fd fa fa fa fd fd fd fa fa fa 0x0c067fff8200: fd fd fd fa fa fa fd fd fd fd fa fa 00 00 00 fa 0x0c067fff8210: fa fa fd fd fd fa fa fa fd fd fd fa fa fa fd fd =>0x0c067fff8220: fd fa fa fa fd fd fd fa fa fa 00 00 00[fa]fa fa 0x0c067fff8230: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x0c067fff8240: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x0c067fff8250: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x0c067fff8260: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x0c067fff8270: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa Shadow byte legend (one shadow byte represents 8 application bytes): Addressable: 00 Partially addressable: 01 02 03 04 05 06 07 Heap left redzone: fa Freed heap region: fd Stack left redzone: f1 Stack mid redzone: f2 Stack right redzone: f3 Stack after return: f5 Stack use after scope: f8 Global redzone: f9 Global init order: f6 Poisoned by user: f7 Container overflow: fc Array cookie: ac Intra object redzone: bb ASan internal: fe Left alloca redzone: ca Right alloca redzone: cb ==26009==ABORTING Now the proper fix for this would be to convert both functions to return an `size_t` instead of an `int`. But given that this commit may be part of a security release, let's instead do the minimal viable fix and die in case we see an overflow. Add a test that would have previously caused us to crash. Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>

When t5583-push-branches.sh was originally introduced via 425b4d7 (push: introduce '--branches' option, 2023-05-06), it was not leak-free. In fact, the test did not even run correctly until 022fbb6 (t5583: fix shebang line, 2023-05-12), but after applying that patch, we see a failure at t5583.8: ==2529087==ERROR: LeakSanitizer: detected memory leaks Direct leak of 384 byte(s) in 1 object(s) allocated from: #0 0x7fb536330986 in __interceptor_realloc ../../../../src/libsanitizer/lsan/lsan_interceptors.cpp:98 #1 0x55e07606cbf9 in xrealloc wrapper.c:140 #2 0x55e075fb6cb3 in prio_queue_put prio-queue.c:42 #3 0x55e075ec81cb in get_reachable_subset commit-reach.c:917 #4 0x55e075fe9cce in add_missing_tags remote.c:1518 #5 0x55e075fea1e4 in match_push_refs remote.c:1665 #6 0x55e076050a8e in transport_push transport.c:1378 #7 0x55e075e2eb74 in push_with_options builtin/push.c:401 #8 0x55e075e2edb0 in do_push builtin/push.c:458 #9 0x55e075e2ff7a in cmd_push builtin/push.c:702 #10 0x55e075d8aaf0 in run_builtin git.c:452 #11 0x55e075d8af08 in handle_builtin git.c:706 #12 0x55e075d8b12c in run_argv git.c:770 #13 0x55e075d8b6a0 in cmd_main git.c:905 #14 0x55e075e81f07 in main common-main.c:60 #15 0x7fb5360ab6c9 in __libc_start_call_main ../sysdeps/nptl/libc_start_call_main.h:58 #16 0x7fb5360ab784 in __libc_start_main_impl ../csu/libc-start.c:360 #17 0x55e075d88f40 in _start (git+0x1ff40) (BuildId: 38ad998b85a535e786129979443630d025ec2453) SUMMARY: LeakSanitizer: 384 byte(s) leaked in 1 allocation(s). This leak was addressed independently via 68b5117 (commit-reach: fix memory leak in get_reachable_subset(), 2023-06-03), which makes t5583 leak-free. But t5583 was not in the tree when 68b5117 was written, and the two only met after the latter was merged back in via 693bde4 (Merge branch 'mh/commit-reach-get-reachable-plug-leak', 2023-06-20). At that point, t5583 was leak-free. Let's mark it as such accordingly. Signed-off-by: Taylor Blau <me@ttaylorr.com> Acked-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>

When "update-index --unresolve $path" cannot find the resolve-undo record for the path the user requested to unresolve, it stuffs the blobs from HEAD and MERGE_HEAD to stage #2 and stage #3 as a fallback. For this reason, the operation does not even start unless both "HEAD" and "MERGE_HEAD" exist. This is suboptimal in a few ways: * It does not recreate stage #1. Even though it is a correct design decision not to do so (because it is impossible to recreate in general cases, without knowing how we got there, including what merge strategy was used), it is much less useful not to have that information in the index. * It limits the "unresolve" operation only during a conflicted "git merge" and nothing else. Other operations like "rebase", "cherry-pick", and "switch -m" may result in conflicts, and the user may want to unresolve the conflict that they incorrectly resolved in order to redo the resolution, but the fallback would not kick in. * Most importantly, the entire "unresolve" operation is disabled after a conflicted merge is committed and MERGE_HEAD is removed, even though the index has perfectly usable resolve-undo records. By lazily reading the HEAD and MERGE_HEAD only when we need to go to the fallback codepath, we will allow cases where resolve-undo records are available (which is 100% of the time, unless the user is reading from an index file created by Git more than 10 years ago) to proceed even after a conflicted merge was committed, during other mergy operations that do not use MERGE_HEAD, or after the result of such mergy operations has been committed. Signed-off-by: Junio C Hamano <gitster@pobox.com>

It is tempting to think of "files and directories" of the current directory as valid inputs to the add and set subcommands of git sparse-checkout. However, in non-cone mode, they often aren't and using them as potential completions leads to *many* forms of confusion: Issue #1. It provides the *wrong* files and directories. For git sparse-checkout add we always want to add files and directories not currently in our sparse checkout, which means we want file and directories not currently present in the current working tree. Providing the files and directories currently present is thus always wrong. For git sparse-checkout set we have a similar problem except in the subset of cases where we are trying to narrow our checkout to a strict subset of what we already have. That is not a very common scenario, especially since it often does not even happen to be true for the first use of the command; for years we required users to create a sparse-checkout via git sparse-checkout init git sparse-checkout set <args...> (or use a clone option that did the init step for you at clone time). The init command creates a minimal sparse-checkout with just the top-level directory present, meaning the set command has to be used to expand the checkout. Thus, only in a special and perhaps unusual cases would any of the suggestions from normal file and directory completion be appropriate. Issue #2: Suggesting patterns that lead to warnings is unfriendly. If the user specifies any regular file and omits the leading '/', then the sparse-checkout command will warn the user that their command is problematic and suggest they use a leading slash instead. Issue #3: Completion gets confused by leading '/', and provides wrong paths. Users often want to anchor their patterns to the toplevel of the repository, especially when listing individual files. There are a number of reasons for this, but notably even sparse-checkout encourages them to do so (as noted above). However, if users do so (via adding a leading '/' to their pattern), then bash completion will interpret the leading slash not as a request for a path at the toplevel of the repository, but as a request for a path at the root of the filesytem. That means at best that completion cannot help with such paths, and if it does find any completions, they are almost guaranteed to be wrong. Issue #4: Suggesting invalid patterns from subdirectories is unfriendly. There is no per-directory equivalent to .gitignore with sparse-checkouts. There is only a single worktree-global $GIT_DIR/info/sparse-checkout file. As such, paths to files must be specified relative to the toplevel of a repository. Providing suggestions of paths that are relative to the current working directory, as bash completion defaults to, is wrong when the current working directory is not the worktree toplevel directory. Issue #5: Paths with special characters will be interpreted incorrectly The entries in the sparse-checkout file are patterns, not paths. While most paths also qualify as patterns (though even in such cases it would be better for users to not use them directly but prefix them with a leading '/'), there are a variety of special characters that would need special escaping beyond the normal shell escaping: '*', '?', '\', '[', ']', and any leading '#' or '!'. If completion suggests any such paths, users will likely expect them to be treated as an exact path rather than as a pattern that might match some number of files other than 1. However, despite the first four issues, we can note that _if_ users are using tab completion, then they are probably trying to specify a path in the index. As such, we transform their argument into a top-level-rooted pattern that matches such a file. For example, if they type: git sparse-checkout add Make<TAB> we could "complete" to git sparse-checkout add /Makefile or, if they ran from the Documentation/technical/ subdirectory: git sparse-checkout add m<TAB> we could "complete" it to: git sparse-checkout add /Documentation/technical/multi-pack-index.txt Note in both cases I use "complete" in quotes, because we actually add characters both before and after the argument in question, so we are kind of abusing "bash completions" to be "bash completions AND beginnings". The fifth issue is a bit stickier, especially when you consider that we not only need to deal with escaping issues because of special meanings of patterns in sparse-checkout & gitignore files, but also that we need to consider escaping issues due to ls-files needing to sometimes quote or escape characters, and because the shell needs to escape some characters. The multiple interacting forms of escaping could get ugly; this patch makes no attempt to do so and simply documents that we decided to not deal with those corner cases for now but at least get the common cases right. Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

When reusing objects from a pack, we keep track of a set of one or more `reused_chunk`s, corresponding to sections of one or more object(s) from a source pack that we are reusing. Each chunk contains two pieces of information: - the offset of the first object in the source pack (relative to the beginning of the source pack) - the difference between that offset, and the corresponding offset in the pack we're generating The purpose of keeping track of these is so that we can patch an OFS_DELTAs that cross over a section of the reuse pack that we didn't take. For instance, consider a hypothetical pack as shown below: (chunk #2) __________... / / +--------+---------+-------------------+---------+ ... | <base> | <other> | (unused) | <delta> | ... +--------+---------+-------------------+---------+ \ / \______________/ (chunk #1) Suppose that we are sending objects "base", "other", and "delta", and that the "delta" object is stored as an OFS_DELTA, and that its base is "base". If we don't send any objects in the "(unused)" range, we can't copy the delta'd object directly, since its delta offset includes a range of the pack that we didn't copy, so we have to account for that difference when patching and reassembling the delta. In order to compute this value correctly, we need to know not only where we are in the packfile we're assembling (with `hashfile_total(f)`) but also the position of the first byte of the packfile that we are currently reusing. Currently, this works just fine, since when reusing only a single pack those two values are always identical (because verbatim reuse is the first thing pack-objects does when enabled after writing the pack header). But when reusing multiple packs which have one or more gaps, we'll need to account for these two values diverging. Together, these two allow us to compute the reused chunk's offset difference relative to the start of the reused pack, as desired. Helped-by: Jeff King <peff@peff.net> Signed-off-by: Taylor Blau <me@ttaylorr.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

The t5309 script triggers a racy false positive with SANITIZE=leak on a multi-core system. Running with "--stress --run=6" usually fails within 10 seconds or so for me, complaining with something like: + git index-pack --fix-thin --stdin fatal: REF_DELTA at offset 46 already resolved (duplicate base 01d7713666f4de822776c7622c10f1b07de280dc?) ================================================================= ==3904583==ERROR: LeakSanitizer: detected memory leaks Direct leak of 32 byte(s) in 1 object(s) allocated from: #0 0x7fa790d01986 in __interceptor_realloc ../../../../src/libsanitizer/lsan/lsan_interceptors.cpp:98 #1 0x7fa790add769 in __pthread_getattr_np nptl/pthread_getattr_np.c:180 #2 0x7fa790d117c5 in __sanitizer::GetThreadStackTopAndBottom(bool, unsigned long*, unsigned long*) ../../../../src/libsanitizer/sanitizer_common/sanitizer_linux_libcdep.cpp:150 #3 0x7fa790d11957 in __sanitizer::GetThreadStackAndTls(bool, unsigned long*, unsigned long*, unsigned long*, unsigned long*) ../../../../src/libsanitizer/sanitizer_common/sanitizer_linux_libcdep.cpp:598 #4 0x7fa790d03fe8 in __lsan::ThreadStart(unsigned int, unsigned long long, __sanitizer::ThreadType) ../../../../src/libsanitizer/lsan/lsan_posix.cpp:51 #5 0x7fa790d013fd in __lsan_thread_start_func ../../../../src/libsanitizer/lsan/lsan_interceptors.cpp:440 #6 0x7fa790adc3eb in start_thread nptl/pthread_create.c:444 #7 0x7fa790b5ca5b in clone3 ../sysdeps/unix/sysv/linux/x86_64/clone3.S:81 SUMMARY: LeakSanitizer: 32 byte(s) leaked in 1 allocation(s). Aborted What happens is this: 0. We construct a bogus pack with a duplicate object in it and trigger index-pack. 1. We spawn a bunch of worker threads to resolve deltas (on my system it is 16 threads). 2. One of the threads sees the duplicate object and bails by calling exit(), taking down all of the threads. This is expected and is the point of the test. 3. At the time exit() is called, we may still be spawning threads from the main process via pthread_create(). LSan hooks thread creation to update its book-keeping; it has to know where each thread's stack is (so it can find entry points for reachable memory). So it calls pthread_getattr_np() to get information about the new thread. That may allocate memory that must be freed with a matching call to pthread_attr_destroy(). Probably LSan does that immediately, but if you're unlucky enough, the exit() will happen while it's between those two calls, and the allocated pthread_attr_t appears as a leak. This isn't a real leak. It's not even in our code, but rather in the LSan instrumentation code. So we could just ignore it. But the false positive can cause people to waste time tracking it down. It's possibly something that LSan could protect against (e.g., cover the getattr/destroy pair with a mutex, and then in the final post-exit() check for leaks try to take the same mutex). But I don't know enough about LSan to say if that's a reasonable approach or not (or if my analysis is even completely correct). In the meantime, it's pretty easy to avoid the race by making creation of the worker threads "atomic". That is, we'll spawn all of them before letting any of them start to work. That's easy to do because we already have a work_lock() mutex for handing out that work. If the main process takes it, then all of the threads will immediately block until we've finished spawning and released it. This shouldn't make any practical difference for non-LSan runs. The thread spawning is quick, and could happen before any worker thread gets scheduled anyway. Probably other spots that use threads are subject to the same issues. But since we have to manually insert locking (and since this really is kind of a hack), let's not bother with them unless somebody experiences a similar racy false-positive in practice. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>

Commit d6a8c58 (midx-write.c: support reading an existing MIDX with `packs_to_include`, 2024-05-29) changed the MIDX generation machinery to support reading from an existing MIDX when writing a new one. Unfortunately, the rest of the MIDX generation machinery is not prepared to deal with such a change. For instance, the function responsible for adding to the object ID fanout table from a MIDX source (midx_fanout_add_midx_fanout()) will gladly add objects from an existing MIDX for some fanout level regardless of whether or not those objects came from packs that are to be included in the subsequent MIDX write. This results in broken pseudo-pack object order (leading to incorrect object traversal results) and segmentation faults, like so (generated by running the added test prior to the changes in midx-write.c): #0 0x000055ee31393f47 in midx_pack_order (ctx=0x7ffdde205c70) at midx-write.c:590 #1 0x000055ee31395a69 in write_midx_internal (object_dir=0x55ee32570440 ".git/objects", packs_to_include=0x7ffdde205e20, packs_to_drop=0x0, preferred_pack_name=0x0, refs_snapshot=0x0, flags=15) at midx-write.c:1171 #2 0x000055ee31395f38 in write_midx_file_only (object_dir=0x55ee32570440 ".git/objects", packs_to_include=0x7ffdde205e20, preferred_pack_name=0x0, refs_snapshot=0x0, flags=15) at midx-write.c:1274 [...] In stack frame #0, the code on midx-write.c:590 is using the new pack ID corresponding to some object which was added from the existing MIDX. Importantly, the pack from which that object was selected in the existing MIDX does not appear in the new MIDX as it was excluded via `--stdin-packs`. In this instance, the pack in question had pack ID "1" in the existing MIDX, but since it was excluded from the new MIDX, we never filled in that entry in the pack_perm table, resulting in: (gdb) p *ctx->pack_perm@2 $1 = {0, 1515870810} Which is what causes the segfault above when we try and read: struct pack_info *pack = &ctx->info[ctx->pack_perm[i]]; if (pack->bitmap_pos == BITMAP_POS_UNKNOWN) pack->bitmap_pos = 0; Fundamentally, we should be able to read information from an existing MIDX when generating a new one. But in practice the midx-write.c code assumes that we won't run into issues like the above with incongruent pack IDs, and often makes those assumptions in extremely subtle and fragile ways. Instead, let's avoid reading from an existing MIDX altogether, and stick with the pre-d6a8c58675 implementation. Harden against any regressions in this area by adding a test which demonstrates these issues. Signed-off-by: Taylor Blau <me@ttaylorr.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

When performing multi-pack reuse, reuse_partial_packfile_from_bitmap() is responsible for generating an array of bitmapped_pack structs from which to perform reuse. In the multi-pack case, we loop over the MIDXs packs and copy the result of calling `nth_bitmapped_pack()` to construct the list of reusable paths. But we may also want to do pack-reuse over a single pack, either because we only had one pack to perform reuse over (in the case of single-pack bitmaps), or because we explicitly asked to do single pack reuse even with a MIDX[^1]. When this is the case, the array we generate of reusable packs contains only a single element, which is either (a) the pack attached to the single-pack bitmap, or (b) the MIDX's preferred pack. In 795006f (pack-bitmap: gracefully handle missing BTMP chunks, 2024-04-15), we refactored the reuse_partial_packfile_from_bitmap() function and stopped assigning the pack_int_id field when reusing only the MIDX's preferred pack. This results in an uninitialized read down in try_partial_reuse() like so: ==7474==WARNING: MemorySanitizer: use-of-uninitialized-value #0 0x55c5cd191dde in try_partial_reuse pack-bitmap.c:1887:8 #1 0x55c5cd191dde in reuse_partial_packfile_from_bitmap_1 pack-bitmap.c:2001:8 #2 0x55c5cd191dde in reuse_partial_packfile_from_bitmap pack-bitmap.c:2105:3 #3 0x55c5cce0bd0e in get_object_list_from_bitmap builtin/pack-objects.c:4043:3 #4 0x55c5cce0bd0e in get_object_list builtin/pack-objects.c:4156:27 #5 0x55c5cce0bd0e in cmd_pack_objects builtin/pack-objects.c:4596:3 #6 0x55c5ccc8fac8 in run_builtin git.c:474:11 which happens when try_partial_reuse() tries to call midx_pair_to_pack_pos() when it tries to reject cross-pack deltas. Avoid the uninitialized read by ensuring that the pack_int_id field is set in the single-pack reuse case by setting it to either the MIDX preferred pack's pack_int_id, or '-1', in the case of single-pack bitmaps. In the latter case, we never read the pack_int_id field, so the choice of '-1' is intentional as a "garbage in, garbage out" measure. Guard against further regressions in this area by adding a test which ensures that we do not throw out deltas from the preferred pack as "cross-pack" due to an uninitialized pack_int_id. [^1]: This can happen for a couple of reasons, either because the repository is configured with 'pack.allowPackReuse=(true|single)', or because the MIDX was generated prior to the introduction of the BTMP chunk, which contains information necessary to perform multi-pack reuse. Reported-by: Kyle Lippincott <spectral@google.com> Signed-off-by: Taylor Blau <me@ttaylorr.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

Memory sanitizer (msan) is detecting a use of an uninitialized variable (`size`) in `read_attr_from_index`: ==2268==WARNING: MemorySanitizer: use-of-uninitialized-value #0 0x5651f3416504 in read_attr_from_index git/attr.c:868:11 #1 0x5651f3415530 in read_attr git/attr.c #2 0x5651f3413d74 in bootstrap_attr_stack git/attr.c:968:6 #3 0x5651f3413d74 in prepare_attr_stack git/attr.c:1004:2 #4 0x5651f3413d74 in collect_some_attrs git/attr.c:1199:2 #5 0x5651f3413144 in git_check_attr git/attr.c:1345:2 #6 0x5651f34728da in convert_attrs git/convert.c:1320:2 #7 0x5651f3473425 in would_convert_to_git_filter_fd git/convert.c:1373:2 #8 0x5651f357a35e in index_fd git/object-file.c:2630:34 #9 0x5651f357aa15 in index_path git/object-file.c:2657:7 #10 0x5651f35db9d9 in add_to_index git/read-cache.c:766:7 #11 0x5651f35dc170 in add_file_to_index git/read-cache.c:799:9 #12 0x5651f321f9b2 in add_files git/builtin/add.c:346:7 #13 0x5651f321f9b2 in cmd_add git/builtin/add.c:565:18 #14 0x5651f321d327 in run_builtin git/git.c:474:11 #15 0x5651f321bc9e in handle_builtin git/git.c:729:3 #16 0x5651f321a792 in run_argv git/git.c:793:4 #17 0x5651f321a792 in cmd_main git/git.c:928:19 #18 0x5651f33dde1f in main git/common-main.c:62:11 The issue exists because `size` is an output parameter from `read_blob_data_from_index`, but it's only modified if `read_blob_data_from_index` returns non-NULL. The read of `size` when calling `read_attr_from_buf` unconditionally may read from an uninitialized value. `read_attr_from_buf` checks that `buf` is non-NULL before reading from `size`, but by then it's already too late: the uninitialized read will have happened already. Furthermore, there's no guarantee that the compiler won't reorder things so that it checks `size` before checking `!buf`. Make the call to `read_attr_from_buf` conditional on `buf` being non-NULL, ensuring that `size` is not read if it's never set. Signed-off-by: Kyle Lippincott <spectral@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

Signed-off-by: Junio C Hamano <gitster@pobox.com>

It was recently reported that concurrent reads and writes may cause the reftable backend to segfault. The root cause of this is that we do not properly keep track of reftable readers across reloads. Suppose that you have a reftable iterator and then decide to reload the stack while iterating through the iterator. When the stack has been rewritten since we have created the iterator, then we would end up discarding a subset of readers that may still be in use by the iterator. The consequence is that we now try to reference deallocated memory, which of course segfaults. One way to trigger this is in t5616, where some background maintenance jobs have been leaking from one test into another. This leads to stack traces like the following one: + git -c protocol.version=0 -C pc1 fetch --filter=blob:limit=29999 --refetch origin AddressSanitizer:DEADLYSIGNAL ================================================================= ==657994==ERROR: AddressSanitizer: SEGV on unknown address 0x7fa0f0ec6089 (pc 0x55f23e52ddf9 bp 0x7ffe7bfa1700 sp 0x7ffe7bfa1700 T0) ==657994==The signal is caused by a READ memory access. #0 0x55f23e52ddf9 in get_var_int reftable/record.c:29 #1 0x55f23e53295e in reftable_decode_keylen reftable/record.c:170 #2 0x55f23e532cc0 in reftable_decode_key reftable/record.c:194 #3 0x55f23e54e72e in block_iter_next reftable/block.c:398 #4 0x55f23e5573dc in table_iter_next_in_block reftable/reader.c:240 #5 0x55f23e5573dc in table_iter_next reftable/reader.c:355 #6 0x55f23e5573dc in table_iter_next reftable/reader.c:339 #7 0x55f23e551283 in merged_iter_advance_subiter reftable/merged.c:69 #8 0x55f23e55169e in merged_iter_next_entry reftable/merged.c:123 #9 0x55f23e55169e in merged_iter_next_void reftable/merged.c:172 #10 0x55f23e537625 in reftable_iterator_next_ref reftable/generic.c:175 #11 0x55f23e2cf9c6 in reftable_ref_iterator_advance refs/reftable-backend.c:464 #12 0x55f23e2d996e in ref_iterator_advance refs/iterator.c:13 #13 0x55f23e2d996e in do_for_each_ref_iterator refs/iterator.c:452 #14 0x55f23dca6767 in get_ref_map builtin/fetch.c:623 #15 0x55f23dca6767 in do_fetch builtin/fetch.c:1659 #16 0x55f23dca6767 in fetch_one builtin/fetch.c:2133 #17 0x55f23dca6767 in cmd_fetch builtin/fetch.c:2432 #18 0x55f23dba7764 in run_builtin git.c:484 #19 0x55f23dba7764 in handle_builtin git.c:741 #20 0x55f23dbab61e in run_argv git.c:805 #21 0x55f23dbab61e in cmd_main git.c:1000 #22 0x55f23dba4781 in main common-main.c:64 #23 0x7fa0f063fc89 in __libc_start_call_main ../sysdeps/nptl/libc_start_call_main.h:58 #24 0x7fa0f063fd44 in __libc_start_main_impl ../csu/libc-start.c:360 #25 0x55f23dba6ad0 in _start (git+0xadfad0) (BuildId: 803b2b7f59beb03d7849fb8294a8e2145dd4aa27) While it is somewhat awkward that the maintenance processes survive tests in the first place, it is totally expected that reftables should work alright with concurrent writers. Seemingly they don't. The only underlying resource that we need to care about in this context is the reftable reader, which is responsible for reading a single table from disk. These readers get discarded immediately (unless reused) when calling `reftable_stack_reload()`, which is wrong. We can only close them once we know that there are no iterators using them anymore. Prepare for a fix by converting the reftable readers to be refcounted. Reported-by: Jeff King <peff@peff.net> Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>

An internal customer reported a segfault when running `git sparse-checkout set` with the `index.sparse` config enabled. I was unable to reproduce it locally, but with their help we debugged into the failing process and discovered the following stacktrace: ``` #0 0x00007ff6318fb7b0 in rehash (map=0x3dfb00d0440, newsize=1048576) at hashmap.c:125 #1 0x00007ff6318fbc66 in hashmap_add (map=0x3dfb00d0440, entry=0x3dfb5c58bc8) at hashmap.c:247 #2 0x00007ff631937a70 in hash_index_entry (istate=0x3dfb00d0400, ce=0x3dfb5c58bc8) at name-hash.c:122 #3 0x00007ff631938a2f in add_name_hash (istate=0x3dfb00d0400, ce=0x3dfb5c58bc8) at name-hash.c:638 #4 0x00007ff631a064de in set_index_entry (istate=0x3dfb00d0400, nr=8291, ce=0x3dfb5c58bc8) at sparse-index.c:255 #5 0x00007ff631a06692 in add_path_to_index (oid=0x5ff130, base=0x5ff580, path=0x3dfb4b725da "<redacted>", mode=33188, context=0x5ff570) at sparse-index.c:307 #6 0x00007ff631a3b48c in read_tree_at (r=0x7ff631c026a0 <the_repo>, tree=0x3dfb5b41f60, base=0x5ff580, depth=2, pathspec=0x5ff5a0, fn=0x7ff631a064e5 <add_path_to_index>, context=0x5ff570) at tree.c:46 #7 0x00007ff631a3b60b in read_tree_at (r=0x7ff631c026a0 <the_repo>, tree=0x3dfb5b41e80, base=0x5ff580, depth=1, pathspec=0x5ff5a0, fn=0x7ff631a064e5 <add_path_to_index>, context=0x5ff570) at tree.c:80 #8 0x00007ff631a3b60b in read_tree_at (r=0x7ff631c026a0 <the_repo>, tree=0x3dfb5b41ac8, base=0x5ff580, depth=0, pathspec=0x5ff5a0, fn=0x7ff631a064e5 <add_path_to_index>, context=0x5ff570) at tree.c:80 #9 0x00007ff631a06a95 in expand_index (istate=0x3dfb00d0100, pl=0x0) at sparse-index.c:422 #10 0x00007ff631a06cbd in ensure_full_index (istate=0x3dfb00d0100) at sparse-index.c:456 #11 0x00007ff631990d08 in index_name_stage_pos (istate=0x3dfb00d0100, name=0x3dfb0020080 "algorithm/levenshtein", namelen=21, stage=0, search_mode=EXPAND_SPARSE) at read-cache.c:556 #12 0x00007ff631990d6c in index_name_pos (istate=0x3dfb00d0100, name=0x3dfb0020080 "algorithm/levenshtein", namelen=21) at read-cache.c:566 #13 0x00007ff63180dbb5 in sanitize_paths (argc=185, argv=0x3dfb0030018, prefix=0x0, skip_checks=0) at builtin/sparse-checkout.c:756 #14 0x00007ff63180de50 in sparse_checkout_set (argc=185, argv=0x3dfb0030018, prefix=0x0) at builtin/sparse-checkout.c:860 #15 0x00007ff63180e6c5 in cmd_sparse_checkout (argc=186, argv=0x3dfb0030018, prefix=0x0) at builtin/sparse-checkout.c:1063 #16 0x00007ff6317234cb in run_builtin (p=0x7ff631ad9b38 <commands+2808>, argc=187, argv=0x3dfb0030018) at git.c:548 #17 0x00007ff6317239c0 in handle_builtin (argc=187, argv=0x3dfb0030018) at git.c:808 #18 0x00007ff631723c7d in run_argv (argcp=0x5ffdd0, argv=0x5ffd78) at git.c:877 #19 0x00007ff6317241d1 in cmd_main (argc=187, argv=0x3dfb0030018) at git.c:1017 #20 0x00007ff631838b60 in main (argc=190, argv=0x3dfb0030000) at common-main.c:64 ``` The very bottom of the stack being the `rehash()` method from `hashmap.c` as called within the `name-hash` API made me look at where these hashmaps were being used in the sparse index logic. These were being copied across indexes, which seems dangerous. Indeed, clearing these hashmaps and setting them as not initialized fixes the segfault. The second commit is a response to a test failure that happens in `t1092-sparse-checkout-compatibility.sh` where `git stash pop` starts to fail because the underlying `git checkout-index` process fails due to colliding files. Passing the `-f` flag appears to work, but it's unclear why this name-hash change causes that change in behavior.

derrickstolee force-pushed the graph/upstream branch 3 times, most recently from 085b37f to 9951cbb Compare January 25, 2018 12:33

derrickstolee force-pushed the graph/upstream branch 3 times, most recently from 8222c3b to ed4c97d Compare January 30, 2018 20:57

derrickstolee force-pushed the graph/upstream branch from ed4c97d to 9b7079f Compare February 8, 2018 20:40

derrickstolee force-pushed the graph/upstream branch 2 times, most recently from c3435ea to 07c1ac6 Compare February 27, 2018 01:34

gitster added 2 commits March 13, 2018 13:34

Merge branch 'jt/binsearch-with-fanout' into HEAD

7d4bebf

* jt/binsearch-with-fanout: packfile: refactor hash search with fanout table packfile: remove GIT_DEBUG_LOOKUP log statements

Merge branch 'jk/cached-commit-buffer' into HEAD

2ee13a7

* jk/cached-commit-buffer: revision: drop --show-all option commit: drop uses of get_cached_commit_buffer() Git 2.16.2

derrickstolee force-pushed the graph/upstream branch 3 times, most recently from c66768f to 437cc16 Compare March 19, 2018 21:54

derrickstolee force-pushed the graph/upstream branch from 437cc16 to 2646653 Compare March 26, 2018 15:26

derrickstolee force-pushed the graph/upstream branch from 2646653 to ba1221a Compare April 2, 2018 20:16

derrickstolee added 7 commits April 2, 2018 14:27

graph: add commit graph design document

ae30d7b

Add Documentation/technical/commit-graph.txt with details of the planned commit graph feature, including future plans. Signed-off-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

commit-graph: implement git-commit-graph write

f237c8b

Teach git-commit-graph to write graph files. Create new test script to verify this command succeeds without failure. Signed-off-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

derrickstolee force-pushed the graph/upstream branch from ba1221a to 15b8e28 Compare April 10, 2018 12:56

derrickstolee added 5 commits April 11, 2018 10:43

derrickstolee pushed a commit that referenced this pull request Aug 12, 2024

Post 2.46-rc0 batch #2

04f5a52

Signed-off-by: Junio C Hamano <gitster@pobox.com>

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Serialized Commit Graph #2

Serialized Commit Graph #2

derrickstolee commented Jan 20, 2018

Serialized Commit Graph #2

Serialized Commit Graph #2

Conversation

derrickstolee commented Jan 20, 2018