The update-index functionality is used internally by 'git stash push' to
setup the internal stashed commit.
This change enables bulk-checkin for update-index infrastructure to
speed up adding new objects to the object database by leveraging the
pack functionality and the new bulk-fsync functionality.
There is some risk with this change, since under batch fsync, the object
files will not be available until the update-index is entirely complete.
This usage is unlikely, since any tool invoking update-index and
expecting to see objects would have to synchronize with the update-index
process after passing it a file path.
Signed-off-by: Neeraj Singh <neerajsi@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This commit adds a win32 implementation for fsync_no_flush that is
called git_fsync. The 'NtFlushBuffersFileEx' function being called is
available since Windows 8. If the function is not available, we
return -1 and Git falls back to doing a full fsync.
The operating system is told to flush data only without a hardware
flush primitive. A later full fsync will cause the metadata log
to be flushed and then the disk cache to be flushed on NTFS and
ReFS. Other filesystems will treat this as a full flush operation.
I added a new file here for this system call so as not to conflict with
downstream changes in the git-for-windows repository related to fscache.
Signed-off-by: Neeraj Singh <neerajsi@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When adding many objects to a repo with core.fsyncObjectFiles set to
true, the cost of fsync'ing each object file can become prohibitive.
One major source of the cost of fsync is the implied flush of the
hardware writeback cache within the disk drive. Fortunately, Windows,
and macOS offer mechanisms to write data from the filesystem page cache
without initiating a hardware flush. Linux has the sync_file_range API,
which issues a pagecache writeback request reliably after version 5.2.
This patch introduces a new 'core.fsyncObjectFiles = batch' option that
batches up hardware flushes. It hooks into the bulk-checkin plugging and
unplugging functionality and takes advantage of tmp-objdir.
When the new mode is enabled do the following for each new object:
1. Create the object in a tmp-objdir.
2. Issue a pagecache writeback request and wait for it to complete.
At the end of the entire transaction when unplugging bulk checkin:
1. Issue an fsync against a dummy file to flush the hardware writeback
cache, which should by now have processed the tmp-objdir writes.
2. Rename all of the tmp-objdir files to their final names.
3. When updating the index and/or refs, we assume that Git will issue
another fsync internal to that operation. This is not the case today,
but may be a good extension to those components.
On a filesystem with a singular journal that is updated during name
operations (e.g. create, link, rename, etc), such as NTFS, HFS+, or XFS
we would expect the fsync to trigger a journal writeout so that this
sequence is enough to ensure that the user's data is durable by the time
the git command returns.
This change also updates the macOS code to trigger a real hardware flush
via fnctl(fd, F_FULLFSYNC) when fsync_or_die is called. Previously, on
macOS there was no guarantee of durability since a simple fsync(2) call
does not flush any hardware caches.
_Performance numbers_:
Linux - Hyper-V VM running Kernel 5.11 (Ubuntu 20.04) on a fast SSD.
Mac - macOS 11.5.1 running on a Mac mini on a 1TB Apple SSD.
Windows - Same host as Linux, a preview version of Windows 11.
This number is from a patch later in the series.
Adding 500 files to the repo with 'git add' Times reported in seconds.
core.fsyncObjectFiles | Linux | Mac | Windows
----------------------|-------|-------|--------
false | 0.06 | 0.35 | 0.61
true | 1.88 | 11.18 | 2.47
batch | 0.15 | 0.41 | 1.53
Signed-off-by: Neeraj Singh <neerajsi@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Preparation for adding bulk-fsync to the bulk-checkin.c infrastructure.
* Rename 'state' variable to 'bulk_checkin_state', since we will later
be adding 'bulk_fsync_state'. This also makes the variable easier to
find in the debugger, since the name is more unique.
* Move the 'plugged' data member of 'bulk_checkin_state' into a separate
static variable. Doing this avoids resetting the variable in
finish_bulk_checkin when zeroing the 'bulk_checkin_state'. As-is, we
seem to unintentionally disable the plugging functionality the first
time a new packfile must be created due to packfile size limits. While
disabling the plugging state only results in suboptimal behavior for
the current code, it would be fatal for the bulk-fsync functionality
later in this patch series.
Signed-off-by: Neeraj Singh <neerajsi@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When creating a subprocess with a temporary ODB, we set the
GIT_QUARANTINE_ENVIRONMENT env var to tell child Git processes not
to update refs, since the tmp-objdir may go away.
Introduce a similar mechanism for in-process temporary ODBs when
we call tmp_objdir_replace_primary_odb. Now both mechanisms set
the disable_ref_updates flag on the odb, which is queried by
the ref_transaction_prepare function.
Note: This change adds an assumption that the state of
the_repository is relevant for any ref transaction that might
be initiated. Unwinding this assumption should be straightforward
by saving the relevant repository to query in the transaction or
the ref_store.
Peff's test case was invoking ref updates via the cachetextconv
setting. That particular code silently does nothing when a ref
update is forbidden. See the call to notes_cache_put in
fill_textconv where errors are ignored.
Reported-by: Jeff King <peff@peff.net>
Signed-off-by: Neeraj Singh <neerajsi@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The tmp_objdir API provides the ability to create temporary object
directories, but was designed with the goal of having subprocesses
access these object stores, followed by the main process migrating
objects from it to the main object store or just deleting it. The
subprocesses would view it as their primary datastore and write to it.
Here we add the tmp_objdir_replace_primary_odb function that replaces
the current process's writable "main" object directory with the
specified one. The previous main object directory is restored in either
tmp_objdir_migrate or tmp_objdir_destroy.
For the --remerge-diff usecase, add a new `will_destroy` flag in `struct
object_database` to mark ephemeral object databases that do not require
fsync durability.
Add 'git prune' support for removing temporary object databases, and
make sure that they have a name starting with tmp_ and containing an
operation-specific name.
Based-on-patch-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Neeraj Singh <neerajsi@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When checking typos in file "po/ko.po", "git-po-helper" reports lots of
false positives because there are no spaces between ASCII and Korean
characters. After applied commit adee197 "(dict: add smudge table for
Korean language, 2021-11-11)" of "git-l10n/git-po-helper" to suppress
these false positives, some easy-to-fix typos are found and fixed.
Signed-off-by: Jiang Xin <worldhello.net@gmail.com>
When we added a new event type to trace2 event stream, we forgot to
raise the format version number, which has been corrected.
* js/trace2-raise-format-version:
trace2: increment event format version
Regression fix.
* ab/fsck-unexpected-type:
object-file: free(*contents) only in read_loose_object() caller
object-file: fix SEGV on free() regression in v2.34.0-rc2
In 64bc752 (trace2: add trace2_child_ready() to report on background
children, 2021-09-20), we added a new "child_ready" event. In
Documentation/technical/api-trace2.txt, we promise that adding a new
event type will result in incrementing the trace2 event format version
number, but this was not done. Correct this in code & docs.
Signed-off-by: Josh Steadmon <steadmon@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In the preceding commit a free() of uninitialized memory regression in
96e41f58fe (fsck: report invalid object type-path combinations,
2021-10-01) was fixed, but we'd still have an issue with leaking
memory from fsck_loose(). Let's fix that issue too.
That issue was introduced in my 31deb28f5e (fsck: don't hard die on
invalid object types, 2021-10-01). It can be reproduced under
SANITIZE=leak with the test I added in 093fffdfbe (fsck tests: add
test for fsck-ing an unknown type, 2021-10-01):
./t1450-fsck.sh --run=84 -vixd
In some sense it's not a problem, we lost the same amount of memory in
terms of things malloc'd and not free'd. It just moved from the "still
reachable" to "definitely lost" column in valgrind(1) nomenclature[1],
since we'd have die()'d before.
But now that we don't hard die() anymore in the library let's properly
free() it. Doing so makes this code much easier to follow, since we'll
now have one function owning the freeing of the "contents" variable,
not two.
For context on that memory management pattern the read_loose_object()
function was added in f6371f9210 (sha1_file: add read_loose_object()
function, 2017-01-13) and subsequently used in c68b489e56 (fsck:
parse loose object paths directly, 2017-01-13). The pattern of it
being the task of both sides to free() the memory has been there in
this form since its inception.
1. https://valgrind.org/docs/manual/mc-manual.html#mc-manual.leaks
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This reverts commit f45022dc2f,
as this is like breakage in the traversal more likely. In a
history with 10 single strand of pearls,
1-->2-->3--...->7-->8-->9-->10
asking "rev-list --unsorted-input 1 10 --not 9 8 7 6 5 4" fails to
paint the bottom 1 uninteresting as the traversal stops, without
completing the propagation of uninteresting bit starting at 4 down
through 3 and 2 to 1.
Fix a regression introduced in my 96e41f58fe (fsck: report invalid
object type-path combinations, 2021-10-01). When fsck-ing blobs larger
than core.bigFileThreshold, we'd free() a pointer to uninitialized
memory.
This issue would have been caught by SANITIZE=address, but since it
involves core.bigFileThreshold, none of the existing tests in our test
suite covered it.
Running them with the "big_file_threshold" in "environment.c" changed
to say "6" would have shown this failure, but let's add a dedicated
test for this scenario based on Han Xin's report[1].
The bug was introduced between v9 and v10[2] of the fsck series merged
in 061a21d36d (Merge branch 'ab/fsck-unexpected-type', 2021-10-25).
1. https://lore.kernel.org/git/20211111030302.75694-1-hanxin.hx@alibaba-inc.com/
2. https://lore.kernel.org/git/cover-v10-00.17-00000000000-20211001T091051Z-avarab@gmail.com/
Reported-by: Han Xin <chiyutianyi@gmail.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The way Cygwin emulates a unix-domain socket, on top of which the
simple-ipc mechanism is implemented, can race with the program on
the other side that wants to use the socket, and briefly make it
appear as a regular file before lstat(2) starts reporting it as a
socket. We now have a workaround on the side that connects to a
unix domain socket.
* js/simple-ipc-cygwin-socket-fix:
simple-ipc: work around issues with Cygwin's Unix socket emulation
"git maintenance run" learned to use system supplied scheduler
backend, but cron on macOS turns out to be unusable for this
purpose.
* ds/no-usable-cron-on-macos:
maintenance: disable cron on macOS
"git pull --ff-only" and "git pull --rebase --ff-only" should make
it a no-op to attempt pulling from a remote that is behind us, but
instead the command errored out by saying it was impossible to
fast-forward, which may technically be true, but not a useful thing
to diagnose as an error. This has been corrected.
* jc/fix-pull-ff-only-when-already-up-to-date:
pull: --ff-only should make it a noop when already-up-to-date
The "-Y find-principals" option of ssh-keygen seems to be broken in
Debian's openssh-client 1:8.7p1-1, whereas it works fine in 1:8.4p1-5.
This causes several failures for GPGSSH tests. We fulfill the
prerequisite because generating the keys works fine, but actually
verifying a signature causes results ranging from bogus results to
ssh-keygen segfaulting.
We can find the broken version during the prereq check by feeding it
empty input. This should result in it complaining to stderr, but in the
broken version it triggers the segfault, causing the GPGSSH tests to be
skipped.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In eba1ba9 (maintenance: `git maintenance run` learned
`--scheduler=<scheduler>`, 2021-09-04), we introduced the ability to
specify a scheduler explicitly. This led to some extra checks around
whether an alternative scheduler was available. This added the
functionality of removing background maintenance from schedulers other
than the one selected.
On macOS, cron is technically available, but running 'crontab' triggers
a UI prompt asking for special permissions. This is the major reason why
launchctl is used as the default scheduler. The is_crontab_available()
method triggers this UI prompt, causing user disruption.
Remove this disruption by using an #ifdef to prevent running crontab
this way on macOS. This has the unfortunate downside that if a user
manually selects cron via the '--scheduler' option, then adjusting the
scheduler later will not remove the schedule from cron. The
'--scheduler' option ignores the is_available checks, which is how we
can get into this situation.
Extract the new check_crontab_process() method to avoid making the
'child' variable unused on macOS. The method is marked MAYBE_UNUSED
because it has no callers on macOS.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Cygwin emulates Unix sockets by writing files with custom contents and
then marking them as system files.
The tricky problem is that while the file is written and its `system`
bit is set, it is still identified as a file. This caused test failures
when Git is too fast looking for the Unix sockets and then complains
that there is a plain file in the way.
Let's work around this by adding a delayed retry loop, specifically for
Cygwin.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Tested-by: Ramsay Jones <ramsay@ramsayjones.plus.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* 'master' of github.com:git/git:
Git 2.34-rc2
parse-options.[ch]: revert use of "enum" for parse_options()
t/lib-git.sh: fix ACL-related permissions failure
A few fixes before -rc2
async_die_is_recursing: work around GCC v11.x issue on Fedora
Document positive variant of commit and merge option "--no-verify"
pull: honor --no-verify and do not call the commit-msg hook
http-backend: remove a duplicated code branch
The filter system allows for alterations to file contents when they're
moved between the database and the worktree. We already made sure that
it is possible for smudge filters to produce contents that are larger
than `unsigned long` can represent (which matters on systems where
`unsigned long` is narrower than `size_t`, most notably 64-bit Windows).
Now we make sure that clean filters can _consume_ contents that are
larger than that.
Note that this commit only allows clean filters' _input_ to be larger
than can be represented by `unsigned long`.
This change makes only a very minute dent into the much larger project
to teach Git to use `size_t` instead of `unsigned long` wherever
appropriate.
Helped-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Matt Cooper <vtbassmatt@gmail.com>
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
