The clean/smudge conversion code path has been prepared to better
work on platforms where ulong is narrower than size_t.
* mc/clean-smudge-with-llp64:
clean/smudge: allow clean filters to process extremely large files
odb: guard against data loss checking out a huge file
git-compat-util: introduce more size_t helpers
odb: teach read_blob_entry to use size_t
t1051: introduce a smudge filter test for extremely large files
test-lib: add prerequisite for 64-bit platforms
test-tool genzeros: generate large amounts of data more efficiently
test-genzeros: allow more than 2G zeros in Windows
There is mixed use of size_t and unsigned long to deal with sizes in the
codebase. Recall that Windows defines unsigned long as 32 bits even on
64-bit platforms, meaning that converting size_t to unsigned long narrows
the range. This mostly doesn't cause a problem since Git rarely deals
with files larger than 2^32 bytes.
But adjunct systems such as Git LFS, which use smudge/clean filters to
keep huge files out of the repository, may have huge file contents passed
through some of the functions in entry.c and convert.c. On Windows, this
results in a truncated file being written to the workdir. I traced this to
one specific use of unsigned long in write_entry (and a similar instance
in write_pc_item_to_fd for parallel checkout). That appeared to be for
the call to read_blob_entry, which expects a pointer to unsigned long.
By altering the signature of read_blob_entry to expect a size_t,
write_entry can be switched to use size_t internally (which all of its
callers and most of its callees already used). To avoid touching dozens of
additional files, read_blob_entry uses a local unsigned long to call a
chain of functions which aren't prepared to accept size_t.
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>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This function was added in 4981fe750b (pkt-line: share
buffer/descriptor reading implementation, 2013-02-23), but in
01f9ec64c8 (Use packet_reader instead of packet_read_line,
2018-12-29) the code that was using it was removed.
Since it's being removed we can in turn remove the "src" and "src_len"
arguments to packet_read(), all the remaining users just passed a
NULL/NULL pair to it.
That function is only a thin wrapper for packet_read_with_status()
which still needs those arguments, but for the thin packet_read()
convenience wrapper we can do away with it for now.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
An object_id storing a SHA-1 name has some unused bytes at the end of
the hash array. Since these bytes are not used, they are usually not
initialized to any value either. However, at
parallel_checkout.c:send_one_item() the object_id of a cache entry is
copied into a buffer which is later sent to a checkout worker through a
pipe write(). This makes Valgrind complain about passing uninitialized
bytes to a syscall. The worker won't use these uninitialized bytes
either, but the warning could confuse someone trying to debug this code;
So instead of using oidcpy(), send_one_item() uses hashcpy() to only
copy the used/initialized bytes of the object_id, and leave the
remaining part with zeros.
However, since cf0983213c ("hash: add an algo member to struct
object_id", 2021-04-26), using hashcpy() is no longer sufficient here as
it won't copy the new algo field from the object_id. Let's add and use a
new function which meets both our requirements of copying all the
important object_id data while still avoiding the uninitialized bytes,
by padding the end of the hash array in the destination object_id. With
this change, we also no longer need the destination buffer from
send_one_item() to be initialized with zeros, so let's switch from
xcalloc() to xmalloc() to make this clear.
Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We already have tests for the basic parallel-checkout operations. But
this code can also run be executed by other commands, such as
git-read-tree and git-sparse-checkout, which are currently not tested
with multiple workers. To promote a wider test coverage without
duplicating tests:
1. Add the GIT_TEST_CHECKOUT_WORKERS environment variable, to optionally
force parallel-checkout execution during the whole test suite.
2. Set this variable (with a value of 2) in the second test round of our
linux-gcc CI job. This round runs `make test` again with some
optional GIT_TEST_* variables enabled, so there is no additional
overhead in exercising the parallel-checkout code here.
Note that tests checking out less than two parallel-eligible entries
will fall back to the sequential mode. Nevertheless, it's still a good
exercise for the parallel-checkout framework as the fallback codepath
also writes the queued entries using the parallel-checkout functions
(only without spawning any worker).
Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add tests to confirm that path collisions are properly detected by
checkout workers, both to avoid race conditions and to report colliding
entries on clone.
Co-authored-by: Jeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Make parallel checkout configurable by introducing two new settings:
checkout.workers and checkout.thresholdForParallelism. The first defines
the number of workers (where one means sequential checkout), and the
second defines the minimum number of entries to attempt parallel
checkout.
To decide the default value for checkout.workers, the parallel version
was benchmarked during three operations in the linux repo, with cold
cache: cloning v5.8, checking out v5.8 from v2.6.15 (checkout I) and
checking out v5.8 from v5.7 (checkout II). The four tables below show
the mean run times and standard deviations for 5 runs in: a local file
system on SSD, a local file system on HDD, a Linux NFS server, and
Amazon EFS (all on Linux). Each parallel checkout test was executed with
the number of workers that brings the best overall results in that
environment.
Local SSD:
Sequential 10 workers Speedup
Clone 8.805 s ± 0.043 s 3.564 s ± 0.041 s 2.47 ± 0.03
Checkout I 9.678 s ± 0.057 s 4.486 s ± 0.050 s 2.16 ± 0.03
Checkout II 5.034 s ± 0.072 s 3.021 s ± 0.038 s 1.67 ± 0.03
Local HDD:
Sequential 10 workers Speedup
Clone 32.288 s ± 0.580 s 30.724 s ± 0.522 s 1.05 ± 0.03
Checkout I 54.172 s ± 7.119 s 54.429 s ± 6.738 s 1.00 ± 0.18
Checkout II 40.465 s ± 2.402 s 38.682 s ± 1.365 s 1.05 ± 0.07
Linux NFS server (v4.1, on EBS, single availability zone):
Sequential 32 workers Speedup
Clone 240.368 s ± 6.347 s 57.349 s ± 0.870 s 4.19 ± 0.13
Checkout I 242.862 s ± 2.215 s 58.700 s ± 0.904 s 4.14 ± 0.07
Checkout II 65.751 s ± 1.577 s 23.820 s ± 0.407 s 2.76 ± 0.08
EFS (v4.1, replicated over multiple availability zones):
Sequential 32 workers Speedup
Clone 922.321 s ± 2.274 s 210.453 s ± 3.412 s 4.38 ± 0.07
Checkout I 1011.300 s ± 7.346 s 297.828 s ± 0.964 s 3.40 ± 0.03
Checkout II 294.104 s ± 1.836 s 126.017 s ± 1.190 s 2.33 ± 0.03
The above benchmarks show that parallel checkout is most effective on
repositories located on an SSD or over a distributed file system. For
local file systems on spinning disks, and/or older machines, the
parallelism does not always bring a good performance. For this reason,
the default value for checkout.workers is one, a.k.a. sequential
checkout.
To decide the default value for checkout.thresholdForParallelism,
another benchmark was executed in the "Local SSD" setup, where parallel
checkout showed to be beneficial. This time, we compared the runtime of
a `git checkout -f`, with and without parallelism, after randomly
removing an increasing number of files from the Linux working tree. The
"sequential fallback" column below corresponds to the executions where
checkout.workers was 10 but checkout.thresholdForParallelism was equal
to the number of to-be-updated files plus one (so that we end up writing
sequentially). Each test case was sampled 15 times, and each sample had
a randomly different set of files removed. Here are the results:
sequential fallback 10 workers speedup
10 files 772.3 ms ± 12.6 ms 769.0 ms ± 13.6 ms 1.00 ± 0.02
20 files 780.5 ms ± 15.8 ms 775.2 ms ± 9.2 ms 1.01 ± 0.02
50 files 806.2 ms ± 13.8 ms 767.4 ms ± 8.5 ms 1.05 ± 0.02
100 files 833.7 ms ± 21.4 ms 750.5 ms ± 16.8 ms 1.11 ± 0.04
200 files 897.6 ms ± 30.9 ms 730.5 ms ± 14.7 ms 1.23 ± 0.05
500 files 1035.4 ms ± 48.0 ms 677.1 ms ± 22.3 ms 1.53 ± 0.09
1000 files 1244.6 ms ± 35.6 ms 654.0 ms ± 38.3 ms 1.90 ± 0.12
2000 files 1488.8 ms ± 53.4 ms 658.8 ms ± 23.8 ms 2.26 ± 0.12
From the above numbers, 100 files seems to be a reasonable default value
for the threshold setting.
Note: Up to 1000 files, we observe a drop in the execution time of the
parallel code with an increase in the number of files. This is a rather
odd behavior, but it was observed in multiple repetitions. Above 1000
files, the execution time increases according to the number of files, as
one would expect.
About the test environments: Local SSD tests were executed on an
i7-7700HQ (4 cores with hyper-threading) running Manjaro Linux. Local
HDD tests were executed on an Intel(R) Xeon(R) E3-1230 (also 4 cores
with hyper-threading), HDD Seagate Barracuda 7200.14 SATA 3.1, running
Debian. NFS and EFS tests were executed on an Amazon EC2 c5n.xlarge
instance, with 4 vCPUs. The Linux NFS server was running on a m6g.large
instance with 2 vCPUSs and a 1 TB EBS GP2 volume. Before each timing,
the linux repository was removed (or checked out back to its previous
state), and `sync && sysctl vm.drop_caches=3` was executed.
Co-authored-by: Jeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Use multiple worker processes to distribute the queued entries and call
write_pc_item() in parallel for them. The items are distributed
uniformly in contiguous chunks. This minimizes the chances of two
workers writing to the same directory simultaneously, which could affect
performance due to lock contention in the kernel. Work stealing (or any
other format of re-distribution) is not implemented yet.
The protocol between the main process and the workers is quite simple.
They exchange binary messages packed in pkt-line format, and use
PKT-FLUSH to mark the end of input (from both sides). The main process
starts the communication by sending N pkt-lines, each corresponding to
an item that needs to be written. These packets contain all the
necessary information to load, smudge, and write the blob associated
with each item. Then it waits for the worker to send back N pkt-lines
containing the results for each item. The resulting packet must contain:
the identification number of the item that it refers to, the status of
the operation, and the lstat() data gathered after writing the file (iff
the operation was successful).
For now, checkout always uses a hardcoded value of 2 workers, only to
demonstrate that the parallel checkout framework correctly divides and
writes the queued entries. The next patch will add user configurations
and define a more reasonable default, based on tests with the said
settings.
Co-authored-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Co-authored-by: Jeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This new interface allows us to enqueue some of the entries being
checked out to later uncompress them, apply in-process filters, and
write out the files in parallel. For now, the parallel checkout
machinery is enabled by default and there is no user configuration, but
run_parallel_checkout() just writes the queued entries in sequence
(without spawning additional workers). The next patch will actually
implement the parallelism and, later, we will make it configurable.
Note that, to avoid potential data races, not all entries are eligible
for parallel checkout. Also, paths that collide on disk (e.g.
case-sensitive paths in case-insensitive file systems), are detected by
the parallel checkout code and skipped, so that they can be safely
sequentially handled later. The collision detection works like the
following:
- If the collision was at basename (e.g. 'a/b' and 'a/B'), the framework
detects it by looking for EEXIST and EISDIR errors after an
open(O_CREAT | O_EXCL) failure.
- If the collision was at dirname (e.g. 'a/b' and 'A'), it is detected
at the has_dirs_only_path() check, which is done for the leading path
of each item in the parallel checkout queue.
Both verifications rely on the fact that, before enqueueing an entry for
parallel checkout, checkout_entry() makes sure that there is no file at
the entry's path and that its leading components are all real
directories. So, any later change in these conditions indicates that
there was a collision (either between two parallel-eligible entries or
between an eligible and an ineligible one).
After all parallel-eligible entries have been processed, the collided
(and thus, skipped) entries are sequentially fed to checkout_entry()
again. This is similar to the way the current code deals with
collisions, overwriting the previously checked out entries with the
subsequent ones. The only difference is that, since we no longer create
the files in the same order that they appear on index, we are not able
to determine which of the colliding entries will survive on disk (for
the classic code, it is always the last entry).
Co-authored-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Co-authored-by: Jeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Junio C Hamano
Matt Cooper
Ævar Arnfjörð Bjarmason
Matheus Tavares