This allows us to replace includes of cache.h with includes of the much
smaller alloc.h in many places. It does mean that we also need to add
includes of alloc.h in a number of C files.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Support names of any length in base_name_compare() and df_name_compare()
by using size_t for their length parameters. They pass the length on to
memcmp(3), which also takes it as a size_t.
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When the "repo" member was added to "the_index" in [1] the
repo_read_index() was made to populate it, but the unpopulated
"the_index" variable didn't get the same treatment.
Let's do that in initialize_the_repository() when we set it up, and
likewise for all of the current callers initialized an empty "struct
index_state".
This simplifies code that needs to deal with "the_index" or a custom
"struct index_state", we no longer need to second-guess this part of
the "index_state" deep in the stack. A recent example of such
second-guessing is the "istate->repo ? istate->repo : the_repository"
code in [2]. We can now simply use "istate->repo".
We're doing this by making use of the INDEX_STATE_INIT() macro (and
corresponding function) added in [3], which now have mandatory "repo"
arguments.
Because we now call index_state_init() in repository.c's
initialize_the_repository() we don't need to handle the case where we
have a "repo->index" whose "repo" member doesn't match the "repo"
we're setting up, i.e. the "Complete the double-reference" code in
repo_read_index() being altered here. That logic was originally added
in [1], and was working around the lack of what we now have in
initialize_the_repository().
For "fsmonitor-settings.c" we can remove the initialization of a NULL
"r" argument to "the_repository". This was added back in [4], and was
needed at the time for callers that would pass us the "r" from an
"istate->repo". Before this change such a change to
"fsmonitor-settings.c" would segfault all over the test suite (e.g. in
t0002-gitfile.sh).
This change has wider eventual implications for
"fsmonitor-settings.c". The reason the other lazy loading behavior in
it is required (starting with "if (!r->settings.fsmonitor) ..." is
because of the previously passed "r" being "NULL".
I have other local changes on top of this which move its configuration
reading to "prepare_repo_settings()" in "repo-settings.c", as we could
now start to rely on it being called for our "r". But let's leave all
of that for now, and narrowly remove this particular part of the
lazy-loading.
1. 1fd9ae517c (repository: add repo reference to index_state,
2021-01-23)
2. ee1f0c242e (read-cache: add index.skipHash config option,
2023-01-06)
3. 2f6b1eb794 (cache API: add a "INDEX_STATE_INIT" macro/function,
add release_index(), 2023-01-12)
4. 1e0ea5c431 (fsmonitor: config settings are repository-specific,
2022-03-25)
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Acked-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* ab/cache-api-cleanup:
cache API: add a "INDEX_STATE_INIT" macro/function, add release_index()
read-cache.c: refactor set_new_index_sparsity() for subsequent commit
sparse-index API: BUG() out on NULL ensure_full_index()
sparse-index.c: expand_to_path() can assume non-NULL "istate"
builtin/difftool.c: { 0 }-initialize rather than using memset()
Introduce an optional configuration to allow the trailing hash that
protects the index file from bit flipping.
* ds/omit-trailing-hash-in-index:
features: feature.manyFiles implies fast index writes
test-lib-functions: add helper for trailing hash
read-cache: add index.skipHash config option
hashfile: allow skipping the hash function
Hopefully in some not so distant future, we'll get advantages from always
initializing the "repo" member of the "struct index_state". To make
that easier let's introduce an initialization macro & function.
The various ad-hoc initialization of the structure can then be changed
over to it, and we can remove the various "0" assignments in
discard_index() in favor of calling index_state_init() at the end.
While not strictly necessary, let's also change the CALLOC_ARRAY() of
various "struct index_state *" to use an ALLOC_ARRAY() followed by
index_state_init() instead.
We're then adding the release_index() function and converting some
callers (including some of these allocations) over to it if they
either won't need to use their "struct index_state" again, or are just
about to call index_state_init().
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Acked-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Refactor code added to set_new_index_sparsity() in [1] to eliminate
indentation resulting from putting the body of his function within the
"if" block. Let's instead return early if we have no
istate->repo. This trivial change makes the subsequent commit's diff
smaller.
1. 491df5f679 (read-cache: set sparsity when index is new, 2022-05-10)
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Acked-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The recent addition of the index.skipHash config option allows index
writes to speed up by skipping the hash computation for the trailing
checksum. This is particularly critical for repositories with many files
at HEAD, so add this config option to two cases where users in that
scenario may opt-in to such behavior:
1. The feature.manyFiles config option enables some options that are
helpful for repositories with many files at HEAD.
2. 'scalar register' and 'scalar reconfigure' set config options that
optimize for large repositories.
In both of these cases, set index.skipHash=true to gain this
speedup. Add tests that demonstrate the proper way that
index.skipHash=true can override feature.manyFiles=true.
Signed-off-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The previous change allowed skipping the hashing portion of the
hashwrite API, using it instead as a buffered write API. Disabling the
hashwrite can be particularly helpful when the write operation is in a
critical path.
One such critical path is the writing of the index. This operation is so
critical that the sparse index was created specifically to reduce the
size of the index to make these writes (and reads) faster.
This trade-off between file stability at rest and write-time performance
is not easy to balance. The index is an interesting case for a couple
reasons:
1. Writes block users. Writing the index takes place in many user-
blocking foreground operations. The speed improvement directly
impacts their use. Other file formats are typically written in the
background (commit-graph, multi-pack-index) or are super-critical to
correctness (pack-files).
2. Index files are short lived. It is rare that a user leaves an index
for a long time with many staged changes. Outside of staged changes,
the index can be completely destroyed and rewritten with minimal
impact to the user.
Following a similar approach to one used in the microsoft/git fork [1],
add a new config option (index.skipHash) that allows disabling this
hashing during the index write. The cost is that we can no longer
validate the contents for corruption-at-rest using the trailing hash.
[1] 21fed2d914
We load this config from the repository config given by istate->repo,
with a fallback to the_repository if it is not set.
While older Git versions will not recognize the null hash as a special
case, the file format itself is still being met in terms of its
structure. Using this null hash will still allow Git operations to
function across older versions.
The one exception is 'git fsck' which checks the hash of the index file.
This used to be a check on every index read, but was split out to just
the index in a33fc72fe9 (read-cache: force_verify_index_checksum,
2017-04-14) and released first in Git 2.13.0. Document the versions that
relaxed these restrictions, with the optimistic expectation that this
change will be included in Git 2.40.0.
Here, we disable this check if the trailing hash is all zeroes. We add a
warning to the config option that this may cause undesirable behavior
with older Git versions.
As a quick comparison, I tested 'git update-index --force-write' with
and without index.skipHash=true on a copy of the Linux kernel
repository.
Benchmark 1: with hash
Time (mean ± σ): 46.3 ms ± 13.8 ms [User: 34.3 ms, System: 11.9 ms]
Range (min … max): 34.3 ms … 79.1 ms 82 runs
Benchmark 2: without hash
Time (mean ± σ): 26.0 ms ± 7.9 ms [User: 11.8 ms, System: 14.2 ms]
Range (min … max): 16.3 ms … 42.0 ms 69 runs
Summary
'without hash' ran
1.78 ± 0.76 times faster than 'with hash'
These performance benefits are substantial enough to allow users the
ability to opt-in to this feature, even with the potential confusion
with older 'git fsck' versions.
Test this new config option, both at a command-line level and within a
submodule. The confirmation is currently limited to confirm that 'git
fsck' does not complain about the index. Future updates will make this
test more robust.
It is critical that this test is placed before the test_index_version
tests, since those tests obliterate the .git/config file and hence lose
the setting from GIT_TEST_DEFAULT_HASH, if set.
Signed-off-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Various leak fixes.
* ab/various-leak-fixes:
built-ins: use free() not UNLEAK() if trivial, rm dead code
revert: fix parse_options_concat() leak
cherry-pick: free "struct replay_opts" members
rebase: don't leak on "--abort"
connected.c: free the "struct packed_git"
sequencer.c: fix "opts->strategy" leak in read_strategy_opts()
ls-files: fix a --with-tree memory leak
revision API: call graph_clear() in release_revisions()
unpack-file: fix ancient leak in create_temp_file()
built-ins & libs & helpers: add/move destructors, fix leaks
dir.c: free "ident" and "exclude_per_dir" in "struct untracked_cache"
read-cache.c: clear and free "sparse_checkout_patterns"
commit: discard partial cache before (re-)reading it
{reset,merge}: call discard_index() before returning
tests: mark tests as passing with SANITIZE=leak
The "sparse_checkout_patterns" member was added to the "struct
index_state" in 836e25c51b (sparse-checkout: hold pattern list in
index, 2021-03-30), but wasn't added to discard_index(). Let's do
that.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Taylor Blau <me@ttaylorr.com>
The discard_index() function has not returned non-zero since
7a51ed66f6 (Make on-disk index representation separate from in-core
one, 2008-01-14), but we've had various code in-tree still acting as
though that might be the case.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The process for reading the index into memory from disk is to first read its
contents into a single memory-mapped file buffer (type 'char *'), then
sequentially convert each on-disk index entry into a corresponding incore
'cache_entry'. To access the contents of the on-disk entry for processing, a
moving pointer within the memory-mapped file is cast to type 'struct
ondisk_cache_entry *'.
In index v4, the entries in the on-disk index file are written *without*
aligning their first byte to a 4-byte boundary; entries are a variable
length (depending on the entry name and whether or not extended flags are
used). As a result, casting the 'char *' buffer pointer to 'struct
ondisk_cache_entry *' then accessing its contents in a 'SANITIZE=undefined'
build can trigger the following error:
read-cache.c:1886:46: runtime error: member access within misaligned
address <address> for type 'struct ondisk_cache_entry', which requires 4
byte alignment
Avoid this error by reading fields directly from the 'char *' buffer, using
the 'offsetof' individual fields in 'struct ondisk_cache_entry'.
Additionally, add documentation describing why the new approach avoids the
misaligned address error, as well as advice on how to improve the
implementation in the future.
Reported-by: Jeff King <peff@peff.net>
Signed-off-by: Victoria Dye <vdye@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Fixes to sparse index compatibility work for "reset" and "checkout"
commands.
source: <pull.1312.v3.git.1659985672.gitgitgadget@gmail.com>
* vd/sparse-reset-checkout-fixes:
unpack-trees: unpack new trees as sparse directories
cache.h: create 'index_name_pos_sparse()'
oneway_diff: handle removed sparse directories
checkout: fix nested sparse directory diff in sparse index
Fix for a bug that makes write-tree to fail to write out a
non-existent index as a tree, introduced in 2.37.
source: <20220722212232.833188-1-martin.agren@gmail.com>
* tk/untracked-cache-with-uall:
read-cache: make `do_read_index()` always set up `istate->repo`
Add 'index_name_pos_sparse()', which behaves the same as 'index_name_pos()',
except that it does not expand a sparse index to search for an entry inside
a sparse directory.
'index_entry_exists()' was originally implemented in 20ec2d034c (reset: make
sparse-aware (except --mixed), 2021-11-29) as an alternative to
'index_name_pos()' to allow callers to search for an index entry without
expanding a sparse index. However, that particular use case only required
knowing whether the requested entry existed, so 'index_entry_exists()' does
not return the index positioning information provided by 'index_name_pos()'.
This patch implements 'index_name_pos_sparse()' to accommodate callers that
need the positioning information of 'index_name_pos()', but do not want to
expand the index.
Signed-off-by: Victoria Dye <vdye@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
If there is no index file, e.g., because the repository has just been
created, we return zero early (unless `must_exist` makes us die
instead.)
This early return means we do not set up `istate->repo`. With
`core.untrackedCache=true`, the recent e6a653554b ("untracked-cache:
support '--untracked-files=all' if configured", 2022-03-31) will
eventually pass down `istate->repo` as a null pointer to
`repo_config_get_string()`, causing a segmentation fault.
If we do hit this early return, set up `istate->repo` similar to when we
actually read the index.
Reported-by: Joey Hess <id@joeyh.name>
Signed-off-by: Martin Ågren <martin.agren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In 998330ac2e (read-cache: look for shared index files next to the
index, too, 2021-08-26), we added code that allocates memory to store
the base path of a shared index, but we never released that memory.
Reported by Coverity.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Remove redundant copying (with index v3 and older) or possible
over-reading beyond end of mmapped memory (with index v4) has been
corrected.
* zh/read-cache-copy-name-entry-fix:
read-cache.c: reduce unnecessary cache entry name copying
575fa8a3 (read-cache: read data in a hash-independent way,
2019-02-19) added a new code to copy from the on-disk data into the
name member of the in-core cache entry, which is already done
immediately after that in a way that takes prefix-compression into
account.
Remove this code, as it is not just unnecessary, but also can be
reading beyond the on-disk data, when we are copying very long
prefix string from the previous entry.
Signed-off-by: ZheNing Hu <adlternative@gmail.com>
[jc: rewrote the log message with Réne's findings]
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"sparse-checkout" learns to work well with the sparse-index
feature.
* ds/sparse-sparse-checkout:
sparse-checkout: integrate with sparse index
p2000: add test for 'git sparse-checkout [add|set]'
sparse-index: complete partial expansion
sparse-index: partially expand directories
sparse-checkout: --no-sparse-index needs a full index
cache-tree: implement cache_tree_find_path()
sparse-index: introduce partially-sparse indexes
sparse-index: create expand_index()
t1092: stress test 'git sparse-checkout set'
t1092: refactor 'sparse-index contents' test
A future change will present a temporary, in-memory mode where the index
can both contain sparse directory entries but also not be completely
collapsed to the smallest possible sparse directories. This will be
necessary for modifying the sparse-checkout definition while using a
sparse index.
For now, convert the single-bit member 'sparse_index' in 'struct
index_state' to be a an 'enum sparse_index_mode' with three modes:
* INDEX_EXPANDED (0): No sparse directories exist. This is always the
case for repositories that do not use cone-mode sparse-checkout.
* INDEX_COLLAPSED: Sparse directories may exist. Files outside the
sparse-checkout cone are reduced to sparse directory entries whenever
possible.
* INDEX_PARTIALLY_SPARSE: Sparse directories may exist. Some file
entries outside the sparse-checkout cone may exist. Running
convert_to_sparse() may further reduce those files to sparse directory
entries.
The main reason to store this extra information is to allow
convert_to_sparse() to short-circuit when the index is already in
INDEX_EXPANDED mode but to actually do the necessary work when in
INDEX_PARTIALLY_SPARSE mode.
The INDEX_PARTIALLY_SPARSE mode will be used in an upcoming change.
Signed-off-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When the index read in 'do_read_index()' does not exist on-disk, mark the
index "sparse" if the executing command does not require a full index and
sparse index is otherwise enabled.
Some commands (such as 'git stash -u') implicitly create a new index (when
the 'GIT_INDEX_FILE' variable points to a non-existent file) and perform
some operation on it. However, when this index is created, it isn't created
with the same sparsity settings as the repo index. As a result, while these
indexes may be sparse during the operation, they are always expanded before
being written to disk. We can avoid that expansion by defaulting the index
to "sparse", in which case it will only be expanded if the full index is
needed.
Note that the function 'set_new_index_sparsity()' is created despite having
only a single caller because additional callers will be added in a
subsequent patch.
Signed-off-by: Victoria Dye <vdye@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Update the stat info of the moved index entry in 'rename_index_entry_at()'
if the entry is up-to-date with the index. Internally, 'git mv' uses
'rename_index_entry_at()' to move the source index entry to the destination.
However, it directly copies the stat info of the original cache entry, which
will not reflect the 'ctime' of the file renaming operation that happened as
part of the move. If a file is otherwise up-to-date with the index, that
difference in 'ctime' will make the entry appear out-of-date until the next
index-refreshing operation (e.g., 'git status').
Some commands, such as 'git reset', use the cached stat information to
determine whether a file is up-to-date; if this information is incorrect,
the command will fail when it should pass. In order to ensure a moved entry
is evaluated as 'up-to-date' when appropriate, refresh the destination index
entry's stat info in 'git mv' if and only if the file is up-to-date.
Note that the test added in 't7001-mv.sh' requires a "sleep 1" to ensure the
'ctime' of the file creation will be definitively older than the 'ctime' of
the renamed file in 'git mv'.
Reported-by: Maximilian Reichel <reichemn@icloud.com>
Signed-off-by: Victoria Dye <vdye@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Replace core.fsyncObjectFiles with two new configuration variables,
core.fsync and core.fsyncMethod.
* ns/core-fsyncmethod:
core.fsync: documentation and user-friendly aggregate options
core.fsync: new option to harden the index
core.fsync: add configuration parsing
core.fsync: introduce granular fsync control infrastructure
core.fsyncmethod: add writeout-only mode
wrapper: make inclusion of Windows csprng header tightly scoped
Object-file API shuffling.
* ab/object-file-api-updates:
object-file API: pass an enum to read_object_with_reference()
object-file.c: add a literal version of write_object_file_prepare()
object-file API: have hash_object_file() take "enum object_type"
object API: rename hash_object_file_literally() to write_*()
object-file API: split up and simplify check_object_signature()
object API users + docs: check <0, not !0 with check_object_signature()
object API docs: move check_object_signature() docs to cache.h
object API: correct "buf" v.s. "map" mismatch in *.c and *.h
object-file API: have write_object_file() take "enum object_type"
object-file API: add a format_object_header() function
object-file API: return "void", not "int" from hash_object_file()
object-file.c: split up declaration of unrelated variables
This commit introduces the new ability for the user to harden
the index. In the event of a system crash, the index must be
durable for the user to actually find a file that has been added
to the repo and then deleted from the working tree.
We use the presence of the COMMIT_LOCK flag and absence of the
alternate_index_output as a proxy for determining whether we're
updating the persistent index of the repo or some temporary
index. We don't sync these temporary indexes.
Signed-off-by: Neeraj Singh <neerajsi@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This commit introduces the infrastructure for the core.fsync
configuration knob. The repository components we want to sync
are identified by flags so that we can turn on or off syncing
for specific components.
If core.fsyncObjectFiles is set and the core.fsync configuration
also includes FSYNC_COMPONENT_LOOSE_OBJECT, we will fsync any
loose objects. This picks the strictest data integrity behavior
if core.fsync and core.fsyncObjectFiles are set to conflicting values.
This change introduces the currently unused fsync_component
helper, which will be used by a later patch that adds fsyncing to
the refs backend.
Actual configuration and documentation of the fsync components
list are in other patches in the series to separate review of
the underlying mechanism from the policy of how it's configured.
Helped-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Neeraj Singh <neerajsi@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Change the write_object_file() function to take an "enum object_type"
instead of a "const char *type". Its callers either passed
{commit,tree,blob,tag}_type and can pass the corresponding OBJ_* type
instead, or were hardcoding strings like "blob".
This avoids the back & forth fragility where the callers of
write_object_file() would have the enum type, and convert it
themselves via type_name(). We do have to now do that conversion
ourselves before calling write_object_file_prepare(), but those
codepaths will be similarly adjusted in subsequent commits.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git update-index", "git checkout-index", and "git clean" are
taught to work better with the sparse checkout feature.
* vd/sparse-clean-etc:
update-index: reduce scope of index expansion in do_reupdate
update-index: integrate with sparse index
update-index: add tests for sparse-checkout compatibility
checkout-index: integrate with sparse index
checkout-index: add --ignore-skip-worktree-bits option
checkout-index: expand sparse checkout compatibility tests
clean: integrate with sparse index
reset: reorder wildcard pathspec conditions
reset: fix validation in sparse index test
Mark in various places in the code that the sparse index and the
split index features are mutually incompatible.
* js/sparse-vs-split-index:
split-index: it really is incompatible with the sparse index
t1091: disable split index
sparse-index: sparse index is disallowed when split index is active
More "config-based hooks".
* ab/config-based-hooks-2:
run-command: remove old run_hook_{le,ve}() hook API
receive-pack: convert push-to-checkout hook to hook.h
read-cache: convert post-index-change to use hook.h
commit: convert {pre-commit,prepare-commit-msg} hook to hook.h
git-p4: use 'git hook' to run hooks
send-email: use 'git hook run' for 'sendemail-validate'
git hook run: add an --ignore-missing flag
hooks: convert worktree 'post-checkout' hook to hook library
hooks: convert non-worktree 'post-checkout' hook to hook library
merge: convert post-merge to use hook.h
am: convert applypatch-msg to use hook.h
rebase: convert pre-rebase to use hook.h
hook API: add a run_hooks_l() wrapper
am: convert {pre,post}-applypatch to use hook.h
gc: use hook library for pre-auto-gc hook
hook API: add a run_hooks() wrapper
hook: add 'run' subcommand
... at least for now. So let's error out if we are even trying to
initialize the split index when the index is sparse, or when trying to
write the split index extension for a sparse index.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Reviewed-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Enable use of the sparse index with `update-index`. Most variations of
`update-index` work without explicitly expanding the index or making any
other updates in or outside of `update-index.c`.
The one usage requiring additional changes is `--cacheinfo`; if a file
inside a sparse directory was specified, the index would not be expanded
until after the cache tree is invalidated, leading to a mismatch between the
index and cache tree. This scenario is handled by rearranging
`add_index_entry_with_check`, allowing `index_name_stage_pos` to expand the
index *before* attempting to invalidate the relevant cache tree path,
avoiding cache tree/index corruption.
Signed-off-by: Victoria Dye <vdye@github.com>
Reviewed-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Move the post-index-change hook away from run-command.h to and over to
the new hook.h library.
This removes the last direct user of "run_hook_ve()" outside of
run-command.c ("run_hook_le()" still uses it). So we can make the
function static now. A subsequent commit will remove this code
entirely when "run_hook_le()" itself goes away.
Signed-off-by: Emily Shaffer <emilyshaffer@google.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Acked-by: Emily Shaffer <emilyshaffer@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
'git update-index --refresh' and '--really-refresh' should force writing
of the index file if racy timestamps have been encountered, as
'git status' already does [1].
Note that calling 'git update-index --refresh' still does not guarantee
that there will be no more racy timestamps afterwards (the same holds
true for 'git status'):
- calling 'git update-index --refresh' immediately after touching and
adding a file may still leave racy timestamps if all three operations
occur within the racy-tolerance (usually 1 second unless USE_NSEC has
been defined)
- calling 'git update-index --refresh' for timestamps which are set into
the future will leave them racy
To guarantee that such racy timestamps will be resolved would require to
wait until the system clock has passed beyond these timestamps and only
then write the index file. Especially for future timestamps, this does
not seem feasible because of possibly long delays/hangs.
[1] https://lore.kernel.org/git/d3dd805c-7c1d-30a9-6574-a7bfcb7fc013@syntevo.com/
Signed-off-by: Marc Strapetz <marc.strapetz@syntevo.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Various operating modes of "git reset" have been made to work
better with the sparse index.
* vd/sparse-reset:
unpack-trees: improve performance of next_cache_entry
reset: make --mixed sparse-aware
reset: make sparse-aware (except --mixed)
reset: integrate with sparse index
reset: expand test coverage for sparse checkouts
sparse-index: update command for expand/collapse test
reset: preserve skip-worktree bit in mixed reset
reset: rename is_missing to !is_in_reset_tree
Ensure that the sparseness of the in-core index matches the
index.sparse configuration specified by the repository immediately
after the on-disk index file is read.
* vd/sparse-sparsity-fix-on-read:
sparse-index: update do_read_index to ensure correct sparsity
sparse-index: add ensure_correct_sparsity function
sparse-index: avoid unnecessary cache tree clearing
test-read-cache.c: prepare_repo_settings after config init
Remove `ensure_full_index` guard on `prime_cache_tree` and update
`prime_cache_tree_rec` to correctly reconstruct sparse directory entries in
the cache tree. While processing a tree's entries, `prime_cache_tree_rec`
must determine whether a directory entry is sparse or not by searching for
it in the index (*without* expanding the index). If a matching sparse
directory index entry is found, no subtrees are added to the cache tree
entry and the entry count is set to 1 (representing the sparse directory
itself). Otherwise, the tree is assumed to not be sparse and its subtrees
are recursively added to the cache tree.
Helped-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Victoria Dye <vdye@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Unless `command_requires_full_index` forces index expansion, ensure in-core
index sparsity matches config settings on read by calling
`ensure_correct_sparsity`. This makes the behavior of the in-core index more
consistent between different methods of updating sparsity: manually changing
the `index.sparse` config setting vs. executing
`git sparse-checkout --[no-]sparse-index init`
Although index sparsity is normally updated with `git sparse-checkout init`,
ensuring correct sparsity after a manual `index.sparse` change has some
practical benefits:
1. It allows for command-by-command sparsity toggling with
`-c index.sparse=<true|false>`, e.g. when troubleshooting issues with the
sparse index.
2. It prevents users from experiencing abnormal slowness after setting
`index.sparse` to `true` due to use of a full index in all commands until
the on-disk index is updated.
Helped-by: Junio C Hamano <gitster@pobox.com>
Co-authored-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Victoria Dye <vdye@github.com>
Reviewed-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Recent sparse-index work broke safety against attempts to add paths
with trailing slashes to the index, which has been corrected.
* rs/make-verify-path-really-verify-again:
read-cache: let verify_path() reject trailing dir separators again
read-cache: add verify_path_internal()
t3905: show failure to ignore sub-repo
When the option --dry-run/-n is given, "git add" doesn't change the
index, but still writes out new object files. Only hash the latter
without writing instead to make the run as dry as possible.
Use this opportunity to also make the hash_flags variable unsigned,
to match the index_path() parameter it is used as.
Reported-by: git.mexon@spamgourmet.com
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Test updates.
* sg/test-split-index-fix:
read-cache: fix GIT_TEST_SPLIT_INDEX
tests: disable GIT_TEST_SPLIT_INDEX for sparse index tests
read-cache: look for shared index files next to the index, too
t1600-index: disable GIT_TEST_SPLIT_INDEX
t1600-index: don't run git commands upstream of a pipe
t1600-index: remove unnecessary redirection
6e773527b6 (sparse-index: convert from full to sparse, 2021-03-30) made
verify_path() accept trailing directory separators for directories,
which is necessary for sparse directory entries. This clemency causes
"git stash" to stumble over sub-repositories, though, and there may be
more unintended side-effects.
Avoid them by restoring the old verify_path() behavior and accepting
trailing directory separators only in places that are supposed to handle
sparse directory entries.
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Turn verify_path() into an internal function that distinguishes between
valid paths and those with trailing directory separators and rename it
to verify_path_internal(). Provide a wrapper with the old behavior
under the old name. No functional change intended. The new function
will be used in the next patch.
Suggested-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Code cleanup.
* ab/repo-settings-cleanup:
repository.h: don't use a mix of int and bitfields
repo-settings.c: simplify the setup
read-cache & fetch-negotiator: check "enum" values in switch()
environment.c: remove test-specific "ignore_untracked..." variable
wrapper.c: add x{un,}setenv(), and use xsetenv() in environment.c
Simplify the setup code in repo-settings.c in various ways, making the
code shorter, easier to read, and requiring fewer hacks to do the same
thing as it did before:
Since 7211b9e753 (repo-settings: consolidate some config settings,
2019-08-13) we have memset() the whole "settings" structure to -1 in
prepare_repo_settings(), and subsequently relied on the -1 value.
Most of the fields did not need to be initialized to -1, and because
we were doing that we had the enum labels "UNTRACKED_CACHE_UNSET" and
"FETCH_NEGOTIATION_UNSET" purely to reflect the resulting state
created this memset() in prepare_repo_settings(). No other code used
or relied on them, more on that below.
For the rest most of the subsequent "are we -1, then read xyz" can
simply be removed by re-arranging what we read first. E.g. when
setting the "index.version" setting we should have first read
"feature.experimental", so that it (and "feature.manyfiles") can
provide a default for our "index.version".
Instead the code setting it, added when "feature.manyFiles"[1] was
created, was using the UPDATE_DEFAULT_BOOL() macro added in an earlier
commit[2]. That macro is now gone, since it was only needed for this
pattern of reading things in the wrong order.
This also fixes an (admittedly obscure) logic error where we'd
conflate an explicit "-1" value in the config with our own earlier
memset() -1.
We can also remove the UPDATE_DEFAULT_BOOL() wrapper added in
[3]. Using it is redundant to simply using the return value from
repo_config_get_bool(), which is non-zero if the provided key exists
in the config.
Details on edge cases relating to the memset() to -1, continued from
"more on that below" above:
* UNTRACKED_CACHE_KEEP:
In [4] the "unset" and "keep" handling for core.untrackedCache was
consolidated. But it while we understand the "keep" value, we don't
handle it differently than the case of any other unknown value.
So let's retain UNTRACKED_CACHE_KEEP and remove the
UNTRACKED_CACHE_UNSET setting (which was always implicitly
UNTRACKED_CACHE_KEEP before). We don't need to inform any code
after prepare_repo_settings() that the setting was "unset", as far
as anyone else is concerned it's core.untrackedCache=keep. if
"core.untrackedcache" isn't present in the config.
* FETCH_NEGOTIATION_UNSET & FETCH_NEGOTIATION_NONE:
Since these two two enum fields added in [5] don't rely on the
memzero() setting them to "-1" anymore we don't have to provide
them with explicit values.
1. c6cc4c5afd (repo-settings: create feature.manyFiles setting,
2019-08-13)
2. 31b1de6a09 (commit-graph: turn on commit-graph by default,
2019-08-13)
3. 31b1de6a09 (commit-graph: turn on commit-graph by default,
2019-08-13)
4. ad0fb65999 (repo-settings: parse core.untrackedCache,
2019-08-13)
5. aaf633c2ad (repo-settings: create feature.experimental setting,
2019-08-13)
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Change tweak_untracked_cache() in "read-cache.c" to use a switch() to
have the compiler assert that we checked all possible values in the
"enum untracked_cache_setting" type, and likewise remove the "default"
case in fetch_negotiator_init() in favor of checking for
"FETCH_NEGOTIATION_UNSET" and "FETCH_NEGOTIATION_NONE".
As will be discussed in a subsequent we'll only ever have either of
these set to FETCH_NEGOTIATION_NONE, FETCH_NEGOTIATION_UNSET and
UNTRACKED_CACHE_UNSET within the prepare_repo_settings() function
itself. In preparation for fixing that code let's add a BUG() here to
mark this as unreachable code.
See ad0fb65999 (repo-settings: parse core.untrackedCache, 2019-08-13)
for when the "unset" and "keep" handling for core.untrackedCache was
consolidated, and aaf633c2ad (repo-settings: create
feature.experimental setting, 2019-08-13) for the addition of the
"default" pattern in "fetch-negotiator.c".
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Running tests with GIT_TEST_SPLIT_INDEX=1 is supposed to turn on the
split index feature and trigger index splitting (mostly) randomly.
Alas, this has been broken since 6e37c8ed3c (read-cache.c: fix writing
"link" index ext with null base oid, 2019-02-13), and
GIT_TEST_SPLIT_INDEX=1 hasn't triggered any index splitting since
then.
This patch makes GIT_TEST_SPLIT_INDEX work again, though it doesn't
restore the pre-6e37c8ed3c behavior. To understand the bug, the fix,
and the behavior change we first have to look at how
GIT_TEST_SPLIT_INDEX used to work before 6e37c8ed3c:
There are two places where we check the value of
GIT_TEST_SPLIT_INDEX, and before 6e37c8ed3c they worked like this:
1) In the lower-level do_write_index(), where, if
GIT_TEST_SPLIT_INDEX is enabled, we call init_split_index().
This call merely allocates and zero-initializes
'istate->split_index', but does nothing else (i.e. doesn't fill
the base/shared index with cache entries, doesn't actually
write a shared index file, etc.). Pertinent to this issue, the
hash of the base index remains all zeroed out.
2) In the higher-level write_locked_index(), but only when
'istate->split_index' has already been initialized. Then, if
GIT_TEST_SPLIT_INDEX is enabled, it randomly sets the flag that
triggers index splitting later in this function. This
randomness comes from the first byte of the hash of the base
index via an 'if ((first_byte & 15) < 6)' condition.
However, if 'istate->split_index' hasn't been initialized (i.e.
it is still NULL), then write_locked_index() just calls
do_write_locked_index(), which internally calls the above
mentioned do_write_index().
This means that while GIT_TEST_SPLIT_INDEX=1 usually triggered index
splitting randomly, the first two index writes were always
deterministic (though I suspect this was unintentional):
- The initial index write never splits the index.
During the first index write write_locked_index() is called with
'istate->split_index' still uninitialized, so the check in 2) is
not executed. It still calls do_write_index(), though, which
then executes the check in 1). The resulting all zero base
index hash then leads to the 'link' extension being written to
'.git/index', though a shared index file is not written:
$ rm .git/index
$ GIT_TEST_SPLIT_INDEX=1 git update-index --add file
$ test-tool dump-split-index .git/index
own c6ef71168597caec8553c83d9d0048f1ef416170
base 0000000000000000000000000000000000000000
100644 d00491fd7e5bb6fa28c517a0bb32b8b506539d4d 0 file
replacements:
deletions:
$ ls -l .git/sharedindex.*
ls: cannot access '.git/sharedindex.*': No such file or directory
- The second index write always splits the index.
When the index written in the previous point is read,
'istate->split_index' is initialized because of the presence of
the 'link' extension. So during the second write
write_locked_index() does run the check in 2), and the first
byte of the all zero base index hash always fulfills the
randomness condition, which in turn always triggers the index
splitting.
- Subsequent index writes will find the 'link' extension with a
real non-zero base index hash, so from then on the check in 2)
is executed and the first byte of the base index hash is as
random as it gets (coming from the SHA-1 of index data including
timestamps and inodes...).
All this worked until 6e37c8ed3c came along, and stopped writing the
'link' extension if the hash of the base index was all zero:
$ rm .git/index
$ GIT_TEST_SPLIT_INDEX=1 git update-index --add file
$ test-tool dump-split-index .git/index
own abbd6f6458d5dee73ae8e210ca15a68a390c6fd7
not a split index
$ ls -l .git/sharedindex.*
ls: cannot access '.git/sharedindex.*': No such file or directory
So, since the first index write with GIT_TEST_SPLIT_INDEX=1 doesn't
write a 'link' extension, in the second index write
'istate->split_index' remains uninitialized, and the check in 2) is
not executed, and ultimately the index is never split.
Fix this by modifying write_locked_index() to make sure to check
GIT_TEST_SPLIT_INDEX even if 'istate->split_index' is still
uninitialized, and initialize it if necessary. The check for
GIT_TEST_SPLIT_INDEX and separate init_split_index() call in
do_write_index() thus becomes unnecessary, so remove it. Furthermore,
add a test to 't1700-split-index.sh' to make sure that
GIT_TEST_SPLIT_INDEX=1 will keep working (though only check the
index splitting on the first index write, because after that it will
be random).
Note that this change does not restore the pre-6e37c8ed3c behaviour,
as it will deterministically split the index already on the first
index write. Since GIT_TEST_SPLIT_INDEX is purely a developer aid,
there is no backwards compatibility issue here. The new behaviour did
trigger test failures in 't0003-attributes.sh' and 't1600-index.sh',
though, which have been fixed in preparatory patches in this series.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Acked-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When reading a split index git always looks for its referenced shared
base index in the gitdir of the current repository, even when reading
an alternate index specified via GIT_INDEX_FILE, and even when that
alternate index file is the "main" '.git/index' file of an other
repository. However, if that split index and its referenced shared
index files were written by a git command running entirely in that
other repository, then, naturally, the shared index file is written to
that other repository's gitdir. Consequently, a git command
attempting to read that shared index file while running in a different
repository won't be able find it and will error out.
I'm not sure in what use case it is necessary to read the index of one
repository by a git command running in a different repository, but it
is certainly possible to do so, and in fact the test 'bare repository:
check that --cached honors index' in 't0003-attributes.sh' does
exactly that. If GIT_TEST_SPLIT_INDEX=1 were to split the index in
just the right moment [1], then this test would indeed fail, because
the referenced shared index file could not be found.
Let's look for the referenced shared index file not only in the gitdir
of the current directory, but, if the shared index is not there, right
next to the split index as well.
[1] We haven't seen this issue trigger a failure in t0003 yet,
because:
- While GIT_TEST_SPLIT_INDEX=1 is supposed to trigger index
splitting randomly, the first index write has always been
deterministic and it has never split the index.
- That alternate index file in the other repository is written
only once in the entire test script, so it's never split.
However, the next patch will fix GIT_TEST_SPLIT_INDEX, and while
doing so it will slightly change its behavior to always split the
index already on the first index write, and t0003 would always
fail without this patch.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Acked-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The convert_to_sparse() method checks for the GIT_TEST_SPARSE_INDEX
environment variable or the "index.sparse" config setting before
converting the index to a sparse one. This is for ease of use since all
current consumers are preparing to compress the index before writing it
to disk. If these settings are not enabled, then convert_to_sparse()
silently returns without doing anything.
We will add a consumer in the next change that wants to use the sparse
index as an in-memory data structure, regardless of whether the on-disk
format should be sparse.
To that end, create the SPARSE_INDEX_MEMORY_ONLY flag that will skip
these config checks when enabled. All current consumers are modified to
pass '0' in the new 'flags' parameter.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Reviewed-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Leak plugging.
* ah/plugleaks:
reset: clear_unpack_trees_porcelain to plug leak
builtin/rebase: fix options.strategy memory lifecycle
builtin/merge: free found_ref when done
builtin/mv: free or UNLEAK multiple pointers at end of cmd_mv
convert: release strbuf to avoid leak
read-cache: call diff_setup_done to avoid leak
ref-filter: also free head for ATOM_HEAD to avoid leak
diffcore-rename: move old_dir/new_dir definition to plug leak
builtin/for-each-repo: remove unnecessary argv copy to plug leak
builtin/submodule--helper: release unused strbuf to avoid leak
environment: move strbuf into block to plug leak
fmt-merge-msg: free newly allocated temporary strings when done
"git read-tree" had a codepath where blobs are fetched one-by-one
from the promisor remote, which has been corrected to fetch in bulk.
* jt/bulk-prefetch:
cache-tree: prefetch in partial clone read-tree
unpack-trees: refactor prefetching code
"git status" codepath learned to work with sparsely populated index
without hydrating it fully.
* ds/status-with-sparse-index:
t1092: document bad sparse-checkout behavior
fsmonitor: integrate with sparse index
wt-status: expand added sparse directory entries
status: use sparse-index throughout
status: skip sparse-checkout percentage with sparse-index
diff-lib: handle index diffs with sparse dirs
dir.c: accept a directory as part of cone-mode patterns
unpack-trees: unpack sparse directory entries
unpack-trees: rename unpack_nondirectories()
unpack-trees: compare sparse directories correctly
unpack-trees: preserve cache_bottom
t1092: add tests for status/add and sparse files
t1092: expand repository data shape
t1092: replace incorrect 'echo' with 'cat'
sparse-index: include EXTENDED flag when expanding
sparse-index: skip indexes with unmerged entries
repo_diff_setup() calls through to diff.c's static prep_parse_options(),
which in turn allocates a new array into diff_opts.parseopts.
diff_setup_done() is responsible for freeing that array, and has the
benefit of verifying diff_opts too - hence we add a call to
diff_setup_done() to avoid leaking parseopts.
Output from the leak as found while running t0090 with LSAN:
Direct leak of 7120 byte(s) in 1 object(s) allocated from:
#0 0x49a82d in malloc ../projects/compiler-rt/lib/asan/asan_malloc_linux.cpp:145:3
#1 0xa8bf89 in do_xmalloc wrapper.c:41:8
#2 0x7a7bae in prep_parse_options diff.c:5636:2
#3 0x7a7bae in repo_diff_setup diff.c:4611:2
#4 0x93716c in repo_index_has_changes read-cache.c:2518:3
#5 0x872233 in unclean merge-ort-wrappers.c:12:14
#6 0x872233 in merge_ort_recursive merge-ort-wrappers.c:53:6
#7 0x5d5b11 in try_merge_strategy builtin/merge.c:752:12
#8 0x5d0b6b in cmd_merge builtin/merge.c:1666:9
#9 0x4ce83e in run_builtin git.c:475:11
#10 0x4ccafe in handle_builtin git.c:729:3
#11 0x4cb01c in run_argv git.c:818:4
#12 0x4cb01c in cmd_main git.c:949:19
#13 0x6bdc2d in main common-main.c:52:11
#14 0x7f551eb51349 in __libc_start_main (/lib64/libc.so.6+0x24349)
SUMMARY: AddressSanitizer: 7120 byte(s) leaked in 1 allocation(s)
Signed-off-by: Andrzej Hunt <andrzej@ahunt.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Refactor the prefetching code in unpack-trees.c into its own function,
because it will be used elsewhere in a subsequent commit.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
By testing 'git -c core.fsmonitor= status -uno', we can check for the
simplest index operations that can be made sparse-aware. The necessary
implementation details are already integrated with sparse-checkout, so
modify command_requires_full_index to be zero for cmd_status().
In refresh_index(), we loop through the index entries to refresh their
stat() information. However, sparse directories have no stat()
information to populate. Ignore these entries.
This allows 'git status' to no longer expand a sparse index to a full
one. This is further tested by dropping the "-uno" option and adding an
untracked file into the worktree.
The performance test p2000-sparse-checkout-operations.sh demonstrates
these improvements:
Test HEAD~1 HEAD
-----------------------------------------------------------------------------
2000.2: git status (full-index-v3) 0.31(0.30+0.05) 0.31(0.29+0.06) +0.0%
2000.3: git status (full-index-v4) 0.31(0.29+0.07) 0.34(0.30+0.08) +9.7%
2000.4: git status (sparse-index-v3) 2.35(2.28+0.10) 0.04(0.04+0.05) -98.3%
2000.5: git status (sparse-index-v4) 2.35(2.24+0.15) 0.05(0.04+0.06) -97.9%
Note that since HEAD~1 was expanding the sparse index by parsing trees,
it was artificially slower than the full index case. Thus, the 98%
improvement is misleading, and instead we should celebrate the 0.34s to
0.05s improvement of 85%. This is more indicative of the peformance
gains we are expecting by using a sparse index.
Note: we are dropping the assignment of core.fsmonitor here. This is not
necessary for the test script as we are not altering the config any
other way. Correct integration with FS Monitor will be validated in
later changes.
Reviewed-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Linux users may benefit from additional information on how to
avoid ENOMEM from mmap despite the system having enough RAM to
accomodate them. We can't reliably unmap pack windows to work
around the issue since malloc and other library routines may
mmap without our knowledge.
Signed-off-by: Eric Wong <e@80x24.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Don't guard the calls to the progress.c API with "if (progress)". The
API itself will check this. This doesn't change any behavior, but
makes this code consistent with the rest of the codebase.
See ae9af12287 (status: show progress bar if refreshing the index
takes too long, 2018-09-15) for the commit that added the pattern
we're changing here.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
These methods were marked as MAYBE_UNUSED in the previous change to
avoid a complicated diff. Delete them entirely, since we now use the
hashfile API instead of this custom hashing code.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The do_write_index() method in read-cache.c has its own hashing logic
and buffering mechanism. Specifically, the ce_write() method was
introduced by 4990aadc (Speed up index file writing by chunking it
nicely, 2005-04-20) and similar mechanisms were introduced a few months
later in c38138cd (git-pack-objects: write the pack files with a SHA1
csum, 2005-06-26). Based on the timing, in the early days of the Git
codebase, I figured that these roughly equivalent code paths were never
unified only because it got lost in the shuffle. The hashfile API has
since been used extensively in other file formats, such as pack-indexes,
multi-pack-indexes, and commit-graphs. Therefore, it seems prudent to
unify the index writing code to use the same mechanism.
I discovered this disparity while trying to create a new index format
that uses the chunk-format API. That API uses a hashfile as its base, so
it is incompatible with the custom code in read-cache.c.
This rewrite is rather straightforward. It replaces all writes to the
temporary file with writes to the hashfile struct. This takes care of
many of the direct interactions with the_hash_algo.
There are still some git_hash_ctx uses remaining: the extension headers
are hashed for use in the End of Index Entries (EOIE) extension. This
use of the git_hash_ctx is left as-is. There are multiple reasons to not
use a hashfile here, including the fact that the data is not actually
writing to a file, just a hash computation. These hashes do not block
our adoption of the chunk-format API in a future change to the index, so
leave it as-is.
The internals of the algorithms are mostly identical. Previously, the
hashfile API used a smaller 8KB buffer instead of the 128KB buffer from
read-cache.c. The previous change already unified these sizes.
There is one subtle point: we do not pass the CSUM_FSYNC to the
finalize_hashfile() method, which differs from most consumers of the
hashfile API. The extra fsync() call indicated by this flag causes a
significant peformance degradation that is noticeable for quick
commands that write the index, such as "git add". Other consumers can
absorb this cost with their more complicated data structure
organization, and further writing structures such as pack-files and
commit-graphs is rarely in the critical path for common user
interactions.
Some static methods become orphaned in this diff, so I marked them as
MAYBE_UNUSED. The diff is much harder to read if they are deleted during
this change. Instead, they will be deleted in the following change.
In addition to the test suite passing, I computed indexes using the
previous binaries and the binaries compiled after this change, and found
the index data to be exactly equal. Finally, I did extensive performance
testing of "git update-index --force-write" on repos of various sizes,
including one with over 2 million paths at HEAD. These tests
demonstrated less than 1% difference in behavior. As expected, the
performance should be considered unchanged. The previous changes to
increase the hashfile buffer size from 8K to 128K ensured this change
would not create a peformance regression.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The final part of "parallel checkout".
* mt/parallel-checkout-part-3:
ci: run test round with parallel-checkout enabled
parallel-checkout: add tests related to .gitattributes
t0028: extract encoding helpers to lib-encoding.sh
parallel-checkout: add tests related to path collisions
parallel-checkout: add tests for basic operations
checkout-index: add parallel checkout support
builtin/checkout.c: complete parallel checkout support
make_transient_cache_entry(): optionally alloc from mem_pool
SHA-256 transition.
* bc/hash-transition-interop-part-1:
hex: print objects using the hash algorithm member
hex: default to the_hash_algo on zero algorithm value
builtin/pack-objects: avoid using struct object_id for pack hash
commit-graph: don't store file hashes as struct object_id
builtin/show-index: set the algorithm for object IDs
hash: provide per-algorithm null OIDs
hash: set, copy, and use algo field in struct object_id
builtin/pack-redundant: avoid casting buffers to struct object_id
Use the final_oid_fn to finalize hashing of object IDs
hash: add a function to finalize object IDs
http-push: set algorithm when reading object ID
Always use oidread to read into struct object_id
hash: add an algo member to struct object_id
"git add" and "git rm" learned not to touch those paths that are
outside of sparse checkout.
* mt/add-rm-in-sparse-checkout:
rm: honor sparse checkout patterns
add: warn when asked to update SKIP_WORKTREE entries
refresh_index(): add flag to ignore SKIP_WORKTREE entries
pathspec: allow to ignore SKIP_WORKTREE entries on index matching
add: make --chmod and --renormalize honor sparse checkouts
t3705: add tests for `git add` in sparse checkouts
add: include magic part of pathspec on --refresh error
Allow make_transient_cache_entry() to optionally receive a mem_pool
struct in which it should allocate the entry. This will be used in the
following patch, to store some transient entries which should persist
until parallel checkout finishes.
Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Builds on top of the sparse-index infrastructure to mark operations
that are not ready to mark with the sparse index, causing them to
fall back on fully-populated index that they always have worked with.
* ds/sparse-index-protections: (47 commits)
name-hash: use expand_to_path()
sparse-index: expand_to_path()
name-hash: don't add directories to name_hash
revision: ensure full index
resolve-undo: ensure full index
read-cache: ensure full index
pathspec: ensure full index
merge-recursive: ensure full index
entry: ensure full index
dir: ensure full index
update-index: ensure full index
stash: ensure full index
rm: ensure full index
merge-index: ensure full index
ls-files: ensure full index
grep: ensure full index
fsck: ensure full index
difftool: ensure full index
commit: ensure full index
checkout: ensure full index
...
The backslash character is not a valid part of a file name on Windows.
If, in Windows, Git attempts to write a file that has a backslash
character in the filename, it will be incorrectly interpreted as a
directory separator.
This caused CVE-2019-1354 in MinGW, as this behaviour can be manipulated
to cause the checkout to write to files it ought not write to, such as
adding code to the .git/hooks directory. This was fixed by e1d911dd4c
(mingw: disallow backslash characters in tree objects' file names,
2019-09-12). However, the vulnerability also exists in Cygwin: while
Cygwin mostly provides a POSIX-like path system, it will still interpret
a backslash as a directory separator.
To avoid this vulnerability, CVE-2021-29468, extend the previous fix to
also apply to Cygwin.
Similarly, extend the test case added by the previous version of the
commit. The test suite doesn't have an easy way to say "run this test
if in MinGW or Cygwin", so add a new test prerequisite that covers both.
As well as checking behaviour in the presence of paths containing
backslashes, the existing test also checks behaviour in the presence of
paths that differ only by the presence of a trailing ".". MinGW follows
normal Windows application behaviour and treats them as the same path,
but Cygwin more closely emulates *nix systems (at the expense of
compatibility with native Windows applications) and will create and
distinguish between such paths. Gate the relevant bit of that test
accordingly.
Reported-by: RyotaK <security@ryotak.me>
Helped-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Adam Dinwoodie <adam@dinwoodie.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In the future, we'll want oidread to automatically set the hash
algorithm member for an object ID we read into it, so ensure we use
oidread instead of hashcpy everywhere we're copying a hash value into a
struct object_id.
Signed-off-by: brian m. carlson <sandals@crustytoothpaste.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Before iterating over all cache entries, ensure that a sparse index is
expanded to a full index to avoid unexpected behavior.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Reviewed-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Callers to index_name_pos() or index_name_stage_pos() have a specific
path in mind. If that happens to be a path with an ancestor being a
sparse-directory entry, it can lead to unexpected results.
In the case that we did not find the requested path, check to see if the
position _before_ the inserted position is a sparse directory entry that
matches the initial segment of the input path (including the directory
separator at the end of the directory name). If so, then expand the
index to be a full index and search again. This expansion will only
happen once per index read.
Future enhancements could be more careful to expand only the necessary
sparse directory entry, but then we would have a special "not fully
sparse, but also not fully expanded" mode that could affect writing the
index to file. Since this only occurs if a specific file is requested
outside of the sparse checkout definition, this is unlikely to be a
common situation.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Reviewed-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Several methods specify that they take a 'struct index_state' pointer
with the 'const' qualifier because they intend to only query the data,
not change it. However, we will be introducing a step very low in the
method stack that might modify a sparse-index to become a full index in
the case that our queries venture inside a sparse-directory entry.
This change only removes the 'const' qualifiers that are necessary for
the following change which will actually modify the implementation of
index_name_stage_pos().
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Reviewed-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
refresh_index() doesn't update SKIP_WORKTREE entries, but it still
matches them against the given pathspecs, marks the matches on the
seen[] array, check if unmerged, etc. In the following patch, one caller
will need refresh_index() to ignore SKIP_WORKTREE entries entirely, so
add a flag that implements this behavior.
While we are here, also realign the REFRESH_* flags and convert the hex
values to the more natural bit shift format, which makes it easier to
spot holes.
Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
If we have a full index, then we can convert it to a sparse index by
replacing directories outside of the sparse cone with sparse directory
entries. The convert_to_sparse() method does this, when the situation is
appropriate.
For now, we avoid converting the index to a sparse index if:
1. the index is split.
2. the index is already sparse.
3. sparse-checkout is disabled.
4. sparse-checkout does not use cone mode.
Finally, we currently limit the conversion to when the
GIT_TEST_SPARSE_INDEX environment variable is enabled. A mode using Git
config will be added in a later change.
The trickiest thing about this conversion is that we might not be able
to mark a directory as a sparse directory just because it is outside the
sparse cone. There might be unmerged files within that directory, so we
need to look for those. Also, if there is some strange reason why a file
is not marked with CE_SKIP_WORKTREE, then we should give up on
converting that directory. There is still hope that some of its
subdirectories might be able to convert to sparse, so we keep looking
deeper.
The conversion process is assisted by the cache-tree extension. This is
calculated from the full index if it does not already exist. We then
abandon the cache-tree as it no longer applies to the newly-sparse
index. Thus, this cache-tree will be recalculated in every
sparse-full-sparse round-trip until we integrate the cache-tree
extension with the sparse index.
Some Git commands use the index after writing it. For example, 'git add'
will update the index, then write it to disk, then read its entries to
report information. To keep the in-memory index in a full state after
writing, we re-expand it to a full one after the write. This is wasteful
for commands that only write the index and do not read from it again,
but that is only the case until we make those commands "sparse aware."
We can compare the behavior of the sparse-index in
t1092-sparse-checkout-compability.sh by using GIT_TEST_SPARSE_INDEX=1
when operating on the 'sparse-index' repo. We can also compare the two
sparse repos directly, such as comparing their indexes (when expanded to
full in the case of the 'sparse-index' repo). We also verify that the
index is actually populated with sparse directory entries.
The 'checkout and reset (mixed)' test is marked for failure when
comparing a sparse repo to a full repo, but we can compare the two
sparse-checkout cases directly to ensure that we are not changing the
behavior when using a sparse index.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The index format does not currently allow for sparse directory entries.
This violates some expectations that older versions of Git or
third-party tools might not understand. We need an indicator inside the
index file to warn these tools to not interact with a sparse index
unless they are aware of sparse directory entries.
Add a new _required_ index extension, 'sdir', that indicates that the
index may contain sparse directory entries. This allows us to continue
to use the differences in index formats 2, 3, and 4 before we create a
new index version 5 in a later change.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We will mark an in-memory index_state as having sparse directory entries
with the sparse_index bit. These currently cannot exist, but we will add
a mechanism for collapsing a full index to a sparse one in a later
change. That will happen at write time, so we must first allow parsing
the format before writing it.
Commands or methods that require a full index in order to operate can
call ensure_full_index() to expand that index in-memory. This requires
parsing trees using that index's repository.
Sparse directory entries have a specific 'ce_mode' value. The macro
S_ISSPARSEDIR(ce->ce_mode) can check if a cache_entry 'ce' has this type.
This ce_mode is not possible with the existing index formats, so we don't
also verify all properties of a sparse-directory entry, which are:
1. ce->ce_mode == 0040000
2. ce->flags & CE_SKIP_WORKTREE is true
3. ce->name[ce->namelen - 1] == '/' (ends in dir separator)
4. ce->oid references a tree object.
These are all semi-enforced in ensure_full_index() to some extent. Any
deviation will cause a warning at minimum or a failure in the worst
case.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
CALLOC_ARRAY() macro replaces many uses of xcalloc().
* rs/calloc-array:
cocci: allow xcalloc(1, size)
use CALLOC_ARRAY
git-compat-util.h: drop trailing semicolon from macro definition
The data structure used by fsmonitor interface was not properly
duplicated during an in-core merge, leading to use-after-free etc.
* js/fsmonitor-unpack-fix:
fsmonitor: do not forget to release the token in `discard_index()`
fsmonitor: fix memory corruption in some corner cases
In 56c6910028 (fsmonitor: change last update timestamp on the
index_state to opaque token, 2020-01-07), we forgot to adjust
`discard_index()` to release the "last-update" token: it is no longer a
64-bit number, but a free-form string that has been allocated.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add and apply a semantic patch for converting code that open-codes
CALLOC_ARRAY to use it instead. It shortens the code and infers the
element size automatically.
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Raise the buffer size used when writing the index file out from
(obviously too small) 8kB to (clearly sufficiently large) 128kB.
* ns/raise-write-index-buffer-size:
read-cache: make the index write buffer size 128K
Preliminary changes to fsmonitor integration.
* jh/fsmonitor-prework:
fsmonitor: refactor initialization of fsmonitor_last_update token
fsmonitor: allow all entries for a folder to be invalidated
fsmonitor: log FSMN token when reading and writing the index
fsmonitor: log invocation of FSMonitor hook to trace2
read-cache: log the number of scanned files to trace2
read-cache: log the number of lstat calls to trace2
preload-index: log the number of lstat calls to trace2
p7519: add trace logging during perf test
p7519: move watchman cleanup earlier in the test
p7519: fix watchman watch-list test on Windows
p7519: do not rely on "xargs -d" in test
Writing an index 8K at a time invokes the OS filesystem and caching code
very frequently, introducing noticeable overhead while writing large
indexes. When experimenting with different write buffer sizes on Windows
writing the Windows OS repo index (260MB), most of the benefit came by
bumping the index write buffer size to 64K. I picked 128K to ensure that
we're past the knee of the curve.
With this change, the time under do_write_index for an index with 3M
files goes from ~1.02s to ~0.72s.
Signed-off-by: Neeraj Singh <neerajsi@ntdev.microsoft.com>
Acked-by: Jeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Report the number of files in the working directory that were read and
their hashes verified in `refresh_index()`.
FSMonitor improves the performance of commands like `git status` by
avoiding scanning the disk for changed files. Let's measure this.
Signed-off-by: Jeff Hostetler <jeffhost@microsoft.com>
Reviewed-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Report the total number of calls made to lstat() inside of refresh_index().
FSMonitor improves the performance of commands like `git status` by
avoiding scanning the disk for changed files. This can be seen in
`refresh_index()`. Let's measure this.
Signed-off-by: Jeff Hostetler <jeffhost@microsoft.com>
Reviewed-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Similar to the previous commits, try to avoid peeking into the `struct
lock_file`. We also have some `struct tempfile`s -- let's avoid looking
into those as well.
Note that `do_write_index()` takes a tempfile and that when we call it,
we either have a tempfile which we can easily hand down, or we have a
lock file, from which we need to somehow obtain the internal tempfile.
So we need to leave that one instance of peeking-into. Nevertheless,
this commit leaves us not relying on exactly how the path of the
tempfile / lock file is stored internally.
Signed-off-by: Martin Ågren <martin.agren@gmail.com>
Reviewed-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
44c7e1a7e0 (mem-pool: use more standard initialization and finalization,
2020-08-15) moved the allocation of the mem-pool structure to callers.
It also added an allocation to load_cache_entries_threaded(), but for an
unrelated mem-pool. Fix that by allocating the correct one instead --
the one that is initialized two lines later.
Reported-by: Sandor Bodo-Merle <sbodomerle@gmail.com>
Signed-off-by: René Scharfe <l.s.r@web.de>
Reviewed-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A typical memory type, such as strbuf, hashmap, or string_list can be
stored on the stack or embedded within another structure. mem_pool
cannot be, because of how mem_pool_init() and mem_pool_discard() are
written. mem_pool_init() does essentially the following (simplified
for purposes of explanation here):
void mem_pool_init(struct mem_pool **pool...)
{
*pool = xcalloc(1, sizeof(*pool));
It seems weird to require that mem_pools can only be accessed through a
pointer. It also seems slightly dangerous: unlike strbuf_release() or
strbuf_reset() or string_list_clear(), all of which put the data
structure into a state where it can be re-used after the call,
mem_pool_discard(pool) will leave pool pointing at free'd memory.
read-cache (and split-index) are the only current users of mem_pools,
and they haven't fallen into a use-after-free mistake here, but it seems
likely to be problematic for future users especially since several of
the current callers of mem_pool_init() will only call it when the
mem_pool* is not already allocated (i.e. is NULL).
This type of mechanism also prevents finding synchronization
points where one can free existing memory and then resume more
operations. It would be natural at such points to run something like
mem_pool_discard(pool...);
and, if necessary,
mem_pool_init(&pool...);
and then carry on continuing to use the pool. However, this fails badly
if several objects had a copy of the value of pool from before these
commands; in such a case, those objects won't get the updated value of
pool that mem_pool_init() overwrites pool with and they'll all instead
be reading and writing from free'd memory.
Modify mem_pool_init()/mem_pool_discard() to behave more like
strbuf_init()/strbuf_release()
or
string_list_init()/string_list_clear()
In particular: (1) make mem_pool_init() just take a mem_pool* and have
it only worry about allocating struct mp_blocks, not the struct mem_pool
itself, (2) make mem_pool_discard() free the memory that the pool was
responsible for, but leave it in a state where it can be used to
allocate more memory afterward (without the need to call mem_pool_init()
again).
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
has_dir_name() has some optimizations for the case where entries are
added to an index in the correct order. They kick in if the new entry
sorts after the last one. One of them exits early if the last entry has
a longer name than the directory of the new entry. Here's its comment:
/*
* The directory prefix lines up with part of
* a longer file or directory name, but sorts
* after it, so this sub-directory cannot
* collide with a file.
*
* last: xxx/yy-file (because '-' sorts before '/')
* this: xxx/yy/abc
*/
However, a file named xxx/yy would be sorted before xxx/yy-file because
'-' sorts after NUL, so the length check against the last entry is not
sufficient to rule out a collision. Remove it.
Reported-by: SZEDER Gábor <szeder.dev@gmail.com>
Suggested-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Further tweak to a "no backslash in indexed paths" for Windows port
we applied earlier.
* js/mingw-loosen-overstrict-tree-entry-checks:
mingw: safeguard better against backslashes in file names
In 224c7d70fa (mingw: only test index entries for backslashes, not tree
entries, 2019-12-31), we relaxed the check for backslashes in tree
entries to check only index entries.
However, the code change was incorrect: it was added to
`add_index_entry_with_check()`, not to `add_index_entry()`, so under
certain circumstances it was possible to side-step the protection.
Besides, the description of that commit purported that all index entries
would be checked when in fact they were only checked when being added to
the index (there are code paths that do not do that, constructing
"transient" index entries).
In any case, it was pointed out in one insightful review at
https://github.com/git-for-windows/git/pull/2437#issuecomment-566771835
that it would be a much better idea to teach `verify_path()` to perform
the check for a backslash. This is safer, even if it comes with two
notable drawbacks:
- `verify_path()` cannot say _what_ is wrong with the path, therefore
the user will no longer be told that there was a backslash in the
path, only that the path was invalid.
- The `git apply` command also calls the `verify_path()` function, and
might have been able to handle Windows-style paths (i.e. with
backslashes instead of forward slashes). This will no longer be
possible unless the user (temporarily) sets `core.protectNTFS=false`.
Note that `git add <windows-path>` will _still_ work because
`normalize_path_copy_len()` will convert the backslashes to forward
slashes before hitting the code path that creates an index entry.
The clear advantage is that `verify_path()`'s purpose is to check the
validity of the file name, therefore we naturally tap into all the code
paths that need safeguarding, also implicitly into future code paths.
The benefits of that approach outweigh the downsides, so let's move the
check from `add_index_entry_with_check()` to `verify_path()`.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
An earlier update to Git for Windows declared that a tree object is
invalid if it has a path component with backslash in it, which was
overly strict, which has been corrected. The only protection the
Windows users need is to prevent such path (or any path that their
filesystem cannot check out) from entering the index.
* js/mingw-loosen-overstrict-tree-entry-checks:
mingw: only test index entries for backslashes, not tree entries
During a clone of a repository that contained a file with a backslash in
its name in the past, as of v2.24.1(2), Git for Windows prints errors
like this:
error: filename in tree entry contains backslash: '\'
The idea is to prevent Git from even trying to write files with
backslashes in their file names: while these characters are valid in
file names on other platforms, on Windows it is interpreted as directory
separator (which would obviously lead to ambiguities, e.g. when there is
a file `a\b` and there is also a file `a/b`).
Arguably, this is the wrong layer for that error: As long as the user
never checks out the files whose names contain backslashes, there should
not be any problem in the first place.
So let's loosen the requirements: we now leave tree entries with
backslashes in their file names alone, but we do require any entries
that are added to the Git index to contain no backslashes on Windows.
Note: just as before, the check is guarded by `core.protectNTFS` (to
allow overriding the check by toggling that config setting), and it
is _only_ performed on Windows, as the backslash is not a directory
separator elsewhere, even when writing to NTFS-formatted volumes.
An alternative approach would be to try to prevent creating files with
backslashes in their file names. However, that comes with its own set of
problems. For example, `git config -f C:\ProgramData\Git\config ...` is
a very valid way to specify a custom config location, and we obviously
do _not_ want to prevent that. Therefore, the approach chosen in this
patch would appear to be better.
This addresses https://github.com/git-for-windows/git/issues/2435
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* maint-2.23: (44 commits)
Git 2.23.1
Git 2.22.2
Git 2.21.1
mingw: sh arguments need quoting in more circumstances
mingw: fix quoting of empty arguments for `sh`
mingw: use MSYS2 quoting even when spawning shell scripts
mingw: detect when MSYS2's sh is to be spawned more robustly
t7415: drop v2.20.x-specific work-around
Git 2.20.2
t7415: adjust test for dubiously-nested submodule gitdirs for v2.20.x
Git 2.19.3
Git 2.18.2
Git 2.17.3
Git 2.16.6
test-drop-caches: use `has_dos_drive_prefix()`
Git 2.15.4
Git 2.14.6
mingw: handle `subst`-ed "DOS drives"
mingw: refuse to access paths with trailing spaces or periods
mingw: refuse to access paths with illegal characters
...
