Measure the time taken to apply the FSMonitor query result
to the index and the untracked-cache.
Set the `FSMONITOR_CHANGED` bit on `istate->cache_changed` when
FSMonitor returns a very large repsonse to ensure that the index is
written to disk.
Normally, when the FSMonitor response includes a tracked file, the
index is always updated. Similarly, the index might be updated when
the response alters the untracked-cache (when enabled). However, in
cases where neither of those cause the index to be considered changed,
the FSMonitor response is wasted. Subsequent Git commands will make
requests with the same token and receive the same response.
If that response is very large, performance may suffer. It would be
more efficient to force update the index now (and the token in the
index extension) in order to reduce the size of the response received
by future commands.
This was observed on Windows after a large checkout. On Windows, the
kernel emits events for the files that are changed as they are
changed. However, it might delay events for the containing
directories until the system is more idle (or someone scans the
directory (so it seems)). The first status following a checkout would
get the list of files. The subsequent status commands would get the
list of directories as the events trickled out. But they would never
catch up because the token was not advanced because the index wasn't
updated.
This list of directories caused `wt_status_collect_untracked()` to
unnecessarily spend time actually scanning them during each command.
Signed-off-by: Jeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Use simple IPC to directly communicate with the new builtin file
system monitor daemon when `core.fsmonitor` is set to true.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Jeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Move fsmonitor config settings to a new and opaque
`struct fsmonitor_settings` structure. Add a lazily-loaded pointer
to this into `struct repo_settings`
Create an `enum fsmonitor_mode` type in `struct fsmonitor_settings` to
represent the state of fsmonitor. This lets us represent which, if
any, fsmonitor provider (hook or IPC) is enabled.
Create `fsm_settings__get_*()` getters to lazily look up fsmonitor-
related config settings.
Get rid of the `core_fsmonitor` global variable. Move the code to
lookup the existing `core.fsmonitor` config value into the fsmonitor
settings.
Create a hook pathname variable in `struct fsmonitor-settings` and
only set it when in hook mode.
Extend the definition of `core.fsmonitor` to be either a boolean
or a hook pathname. When true, the builtin FSMonitor is used.
When false or unset, no FSMonitor (neither builtin nor hook) is
used.
The existing `core_fsmonitor` global variable was used to store the
pathname to the fsmonitor hook *and* it was used as a boolean to see
if fsmonitor was enabled. This dual usage and global visibility leads
to confusion when we add the IPC-based provider. So lets hide the
details in fsmonitor-settings.c and let it decide which provider to
use in the case of multiple settings. This avoids cluttering up
repo-settings.c with these private details.
A future commit in builtin-fsmonitor series will add the ability to
disqualify worktrees for various reasons, such as being mounted from a
remote volume, where fsmonitor should not be started. Having the
config settings hidden in fsmonitor-settings.c allows such worktree
restrictions to override the config values used.
Signed-off-by: Jeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The fsmonitor interface read from its input without making sure
there is something to read from. This bug is new in 2.31
timeframe.
* jh/fsmonitor-prework:
fsmonitor: avoid global-buffer-overflow READ when checking trivial response
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
Isolate and document initialization of `istate->fsmonitor_last_update`.
This field should contain a fsmonitor-specific opaque token, but we
need to initialize it before we can actually talk to a fsmonitor process,
so we create a generic default value.
Signed-off-by: Jeff Hostetler <jeffhost@microsoft.com>
Reviewed-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Allow fsmonitor to report directory changes by reporting paths with a
trailing slash.
Signed-off-by: Jeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: Kevin Willford <Kevin.Willford@microsoft.com>
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Reviewed-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Let's measure the time taken to request and receive FSMonitor data
via the hook API and the size of the response.
Signed-off-by: Jeff Hostetler <jeffhost@microsoft.com>
Reviewed-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The index has an fsmonitor_dirty bitmap that records which index entries
are "dirty" based on the response from the FSMonitor. If this bitmap
ever grows larger than the index, then there was an error in how it was
constructed, and it was probably a developer's bug.
There are several BUG() statements that are very similar, so replace
these uses with a simpler assert_index_minimum(). Since there is one
caller that uses a custom 'pos' value instead of the bit_size member, we
cannot simplify it too much. However, the error string is identical in
each, so this simplifies things.
Be sure to add one when checking if a position if valid, since the
minimum is a bound on the expected size.
The end result is that the code is simpler to read while also preserving
these assertions for developers in the FSMonitor space.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The argv_array API is useful for not just managing argv but any
"vector" (NULL-terminated array) of strings, and has seen adoption
to a certain degree. It has been renamed to "strvec" to reduce the
barrier to adoption.
* jk/strvec:
strvec: rename struct fields
strvec: drop argv_array compatibility layer
strvec: update documention to avoid argv_array
strvec: fix indentation in renamed calls
strvec: convert remaining callers away from argv_array name
strvec: convert more callers away from argv_array name
strvec: convert builtin/ callers away from argv_array name
quote: rename sq_dequote_to_argv_array to mention strvec
strvec: rename files from argv-array to strvec
argv-array: rename to strvec
argv-array: use size_t for count and alloc
We eventually want to drop the argv_array name and just use strvec
consistently. There's no particular reason we have to do it all at once,
or care about interactions between converted and unconverted bits.
Because of our preprocessor compat layer, the names are interchangeable
to the compiler (so even a definition and declaration using different
names is OK).
This patch converts remaining files from the first half of the alphabet,
to keep the diff to a manageable size.
The conversion was done purely mechanically with:
git ls-files '*.c' '*.h' |
xargs perl -i -pe '
s/ARGV_ARRAY/STRVEC/g;
s/argv_array/strvec/g;
'
and then selectively staging files with "git add '[abcdefghjkl]*'".
We'll deal with any indentation/style fallouts separately.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Some file monitors like watchman will use something other than a timestamp
to keep better track of what changes happen in between calls to query
the fsmonitor. The clockid in watchman is a string. Now that the index
is storing an opaque token for the last update the code needs to be
updated to pass that opaque token to a verion 2 of the fsmonitor hook.
Because there are repos that already have version 1 of the hook and we
want them to continue to work when git is updated, we need to handle
both version 1 and version 2 of the hook. In order to do that a
config value is being added core.fsmonitorHookVersion to force what
version of the hook should be used. When this is not set it will default
to -1 and then the code will attempt to call version 2 of the hook first.
If that fails it will fallback to trying version 1.
Signed-off-by: Kevin Willford <Kevin.Willford@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Some file system monitors might not use or take a timestamp for processing
and in the case of watchman could have race conditions with using a
timestamp. Watchman uses something called a clockid that is used for race
free queries to it. The clockid for watchman is simply a string.
Change the fsmonitor_last_update from being a uint64_t to a char pointer
so that any arbitrary data can be stored in it and passed back to the
fsmonitor.
Signed-off-by: Kevin Willford <Kevin.Willford@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The index might be aware that a file hasn't modified via fsmonitor, but
unpack-trees did not pay attention to it and checked via ie_match_stat
which can be inefficient on certain filesystems. This significantly slows
down commands that run oneway_merge, like checkout and reset --hard.
This patch makes oneway_merge check whether a file is considered
unchanged through fsmonitor and skips ie_match_stat on it. unpack-trees
also now correctly copies over fsmonitor validity state from the source
index. Finally, for correctness, we force a refresh of fsmonitor state in
tweak_fsmonitor.
After this change, commands like stash (that use reset --hard
internally) go from 8s or more to ~2s on a 250k file repository on a
mac.
Helped-by: Junio C Hamano <gitster@pobox.com>
Helped-by: Kevin Willford <Kevin.Willford@microsoft.com>
Signed-off-by: Utsav Shah <utsav@dropbox.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
3444ec2e ("fsmonitor: don't fill bitmap with entries to be removed",
2019-10-11) added a handful of sanity checks that make sure that a
bit position in fsmonitor bitmap does not go beyond the end of the
index. As each bit in the bitmap corresponds to a path in the
index, this is the right check most of the time.
Except for the case when we are in the split-index mode and looking
at a delta index that is to be overlayed on the base index but
before the base index has actually been merged in, namely in read_
and write_fsmonitor_extension(). In these codepaths, the entries in
the split/delta index is typically a small subset of the entire set
of paths (otherwise why would we be using split-index?), so the
bitmap used by the fsmonitor is almost always larger than the number
of entries in the partial index, and the incorrect comparison would
trigger the BUG().
Reported-by: Utsav Shah <ukshah2@illinois.edu>
Helped-by: Kevin Willford <Kevin.Willford@microsoft.com>
Helped-by: William Baker <William.Baker@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
While doing some testing with fsmonitor enabled I found
that git commands would segfault after staging and
unstaging an untracked file. Looking at the crash it
appeared that fsmonitor_ewah_callback was attempting to
adjust bits beyond the bounds of the index cache.
Digging into how this could happen it became clear that
the fsmonitor extension must have been written with
more bits than there were entries in the index. The
root cause ended up being that fill_fsmonitor_bitmap was
populating fsmonitor_dirty with bits for all entries in
the index, even those that had been marked for removal.
To solve this problem fill_fsmonitor_bitmap has been
updated to skip entries with the the CE_REMOVE flag set.
With this change the bits written for the fsmonitor
extension will be consistent with the index entries
written by do_write_index. Additionally, BUG checks
have been added to detect if the number of bits in
fsmonitor_dirty should ever exceed the number of
entries in the index again.
Another option that was considered was moving the call
to fill_fsmonitor_bitmap closer to where the index is
written (and where the fsmonitor extension itself is
written). However, that did not work as the
fsmonitor_dirty bitmap must be filled before the index
is split during writing.
Signed-off-by: William Baker <William.Baker@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
883e248b8a ("fsmonitor: teach git to optionally utilize a file system
monitor to speed up detecting new or changed files.", 2017-09-22) uses
an int in a loop that would wrap if index_state->cache_nr (unsigned)
is bigger than INT_MAX
Signed-off-by: Carlo Marcelo Arenas Belón <carenas@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
With this change, the `index_state` struct becomes the new home for the
flag that says whether the fsmonitor hook has been run, i.e. it is now
per-index.
It also gets re-set when the index is discarded, fixing the bug
demonstrated by the "test_expect_failure" test added in the preceding
commit. In that case fsmonitor-enabled Git would miss updates under
certain circumstances, see that preceding commit for details.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Avoid unchecked snprintf() to make future code auditing easier.
* jk/snprintf-truncation:
fmt_with_err: add a comment that truncation is OK
shorten_unambiguous_ref: use xsnprintf
fsmonitor: use internal argv_array of struct child_process
log_write_email_headers: use strbufs
http: use strbufs instead of fixed buffers
Avoid magic array sizes and indexes by constructing the fsmonitor
command line using the embedded argv_array of the child_process. The
resulting code is shorter and easier to extend.
Getting rid of the snprintf() calls is a bonus -- even though the
buffers were big enough here to avoid truncation -- as it makes auditing
the remaining callers easier.
Inspired-by: Jeff King <peff@peff.net>
Signed-off-by: Rene Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The index file is updated to record the fsmonitor section after a
full scan was made, to avoid wasting the effort that has already
spent.
* bp/fsmonitor-prime-index:
fsmonitor: force index write after full scan
Fix an unexploitable (because the oversized contents are not under
attacker's control) buffer overflow.
* bp/fsmonitor-bufsize-fix:
fsmonitor: fix incorrect buffer size when printing version number
fsmonitor currently only flags the index as dirty if the extension is being
added or removed. This is a performance optimization that recognizes you can
stat() a lot of files in less time than it takes to write out an updated index.
This patch makes a small enhancement and flags the index dirty if we end up
having to stat() all files and scan the entire working directory. The assumption
being that must be expensive or you would not have turned on the feature.
Signed-off-by: Ben Peart <benpeart@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This is a trivial bug fix for passing the incorrect size to snprintf() when
outputting the version. It should be passing the size of the destination buffer
rather than the size of the value being printed.
Signed-off-by: Ben Peart <benpeart@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
read_directory() code ignores all paths named ".git" even if it's not
a valid git repository. See treat_path() for details. Since ".git" is
basically invisible to read_directory(), when we are asked to
invalidate a path that contains ".git", we can safely ignore it
because the slow path would not consider it anyway.
This helps when fsmonitor is used and we have a real ".git" repo at
worktree top. Occasionally .git/index will be updated and if the
fsmonitor hook does not filter it, untracked cache is asked to
invalidate the path ".git/index".
Without this patch, we invalidate the root directory unncessarily,
which:
- makes read_directory() fall back to slow path for root directory
(slower)
- makes the index dirty (because UNTR extension is updated). Depending
on the index size, writing it down could also be slow.
A note about the new "safe_path" knob. Since this new check could be
relatively expensive, avoid it when we know it's not needed. If the
path comes from the index, it can't contain ".git". If it does
contain, we may be screwed up at many more levels, not just this one.
Noticed-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
ba1b9cac ("fsmonitor: delay updating state until after split index
is merged", 2017-10-27) resolved the problem of the fsmonitor data
being applied to the non-base index when reading; however, a similar
problem exists when writing the index. Specifically, writing of the
fsmonitor extension happens only after the work to split the index
has been applied -- as such, the information in the index is only
for the non-"base" index, and thus the extension information
contains only partial data.
When saving, compute the ewah bitmap before the index is split, and
store it in the fsmonitor_dirty field, mirroring the behavior that
occurred during reading. fsmonitor_dirty is kept from being leaked by
being freed when the extension data is written -- which always happens
precisely once, no matter the split index configuration.
Signed-off-by: Alex Vandiver <alexmv@dropbox.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
If the fsmonitor extension is used in conjunction with the split index
extension, the set of entries in the index when it is first loaded is
only a subset of the real index. This leads to only the non-"base"
index being marked as CE_FSMONITOR_VALID.
Delay the expansion of the ewah bitmap until after tweak_split_index
has been called to merge in the base index as well.
The new fsmonitor_dirty is kept from being leaked by dint of being
cleaned up in post_read_index_from, which is guaranteed to be called
after do_read_index in read_index_from.
Signed-off-by: Alex Vandiver <alexmv@dropbox.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The fsmonitor command inherits the PWD of its caller, which may be
anywhere in the working copy. This makes is difficult for the
fsmonitor command to operate on the whole repository. Specifically,
for the watchman integration, this causes each subdirectory to get its
own watch entry.
Set the CWD to the top of the working directory, for consistency.
Signed-off-by: Alex Vandiver <alexmv@dropbox.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When the index is read from disk, the fsmonitor index extension is used
to flag the last known potentially dirty index entries. The registered
core.fsmonitor command is called with the time the index was last
updated and returns the list of files changed since that time. This list
is used to flag any additional dirty cache entries and untracked cache
directories.
We can then use this valid state to speed up preload_index(),
ie_match_stat(), and refresh_cache_ent() as they do not need to lstat()
files to detect potential changes for those entries marked
CE_FSMONITOR_VALID.
In addition, if the untracked cache is turned on valid_cached_dir() can
skip checking directories for new or changed files as fsmonitor will
invalidate the cache only for those directories that have been
identified as having potential changes.
To keep the CE_FSMONITOR_VALID state accurate during git operations;
when git updates a cache entry to match the current state on disk,
it will now set the CE_FSMONITOR_VALID bit.
Inversely, anytime git changes a cache entry, the CE_FSMONITOR_VALID bit
is cleared and the corresponding untracked cache directory is marked
invalid.
Signed-off-by: Ben Peart <benpeart@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Jeff Hostetler
Junio C Hamano
Andrzej Hunt
Kevin Willford
Derrick Stolee
Jeff King
Elijah Newren
Utsav Shah
William Baker
Carlo Marcelo Arenas Belón
Johannes Schindelin
René Scharfe
Ben Peart
Ben Peart
Nguyễn Thái Ngọc Duy
Alex Vandiver