mirror/ git
1
0
Fork 0
mirror of https://github.com/git/git.git synced 2024-06-01 07:46:19 +02:00
Code

fetch: use "quick" has_sha1_file for tag following 

When we auto-follow tags in a fetch, we look at all of the
tags advertised by the remote and fetch ones where we don't
already have the tag, but we do have the object it peels to.
This involves a lot of calls to has_sha1_file(), some of
which we can reasonably expect to fail. Since 45e8a74
(has_sha1_file: re-check pack directory before giving up,
2013-08-30), this may cause many calls to
reprepare_packed_git(), which is potentially expensive.

This has gone unnoticed for several years because it
requires a fairly unique setup to matter:

  1. You need to have a lot of packs on the client side to
     make reprepare_packed_git() expensive (the most
     expensive part is finding duplicates in an unsorted
     list, which is currently quadratic).

  2. You need a large number of tag refs on the server side
     that are candidates for auto-following (i.e., that the
     client doesn't have). Each one triggers a re-read of
     the pack directory.

  3. Under normal circumstances, the client would
     auto-follow those tags and after one large fetch, (2)
     would no longer be true. But if those tags point to
     history which is disconnected from what the client
     otherwise fetches, then it will never auto-follow, and
     those candidates will impact it on every fetch.

So when all three are true, each fetch pays an extra
O(nr_tags * nr_packs^2) cost, mostly in string comparisons
on the pack names. This was exacerbated by 47bf4b0
(prepare_packed_git_one: refactor duplicate-pack check,
2014-06-30) which uses a slightly more expensive string
check, under the assumption that the duplicate check doesn't
happen very often (and it shouldn't; the real problem here
is how often we are calling reprepare_packed_git()).

This patch teaches fetch to use HAS_SHA1_QUICK to sacrifice
accuracy for speed, in cases where we might be racy with a
simultaneous repack. This is similar to the fix in 0eeb077
(index-pack: avoid excessive re-reading of pack directory,
2015-06-09). As with that case, it's OK for has_sha1_file()
occasionally say "no I don't have it" when we do, because
the worst case is not a corruption, but simply that we may
fail to auto-follow a tag that points to it.

Here are results from the included perf script, which sets
up a situation similar to the one described above:

Test            HEAD^               HEAD
----------------------------------------------------------
5550.4: fetch   11.21(10.42+0.78)   0.08(0.04+0.02) -99.3%

Reported-by: Vegard Nossum <vegard.nossum@oracle.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This commit is contained in:
Jeff King 2016-10-13 12:53:44 -04:00 committed by Junio C Hamano
parent 0202c411ed
commit 5827a03545
4 changed files with 112 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
builtin/fetch.c
View File

@ -232,9 +232,10 @@ static void find_non_local_tags(struct transport *transport,
* as one to ignore by setting util to NULL.
*/
if (ends_with(ref->name, "^{}")) {
if (item && !has_object_file(&ref->old_oid) &&
if (item &&
!has_object_file_with_flags(&ref->old_oid, HAS_SHA1_QUICK) &&
!will_fetch(head, ref->old_oid.hash) &&
!has_sha1_file(item->util) &&
!has_sha1_file_with_flags(item->util, HAS_SHA1_QUICK) &&
!will_fetch(head, item->util))
item->util = NULL;
item = NULL;
@ -247,7 +248,8 @@ static void find_non_local_tags(struct transport *transport,
* to check if it is a lightweight tag that we want to
* fetch.
*/
if (item && !has_sha1_file(item->util) &&
if (item &&
!has_sha1_file_with_flags(item->util, HAS_SHA1_QUICK) &&
!will_fetch(head, item->util))
item->util = NULL;
@ -267,7 +269,8 @@ static void find_non_local_tags(struct transport *transport,
* We may have a final lightweight tag that needs to be
* checked to see if it needs fetching.
*/
if (item && !has_sha1_file(item->util) &&
if (item &&
!has_sha1_file_with_flags(item->util, HAS_SHA1_QUICK) &&
!will_fetch(head, item->util))
item->util = NULL;

1
cache.h
View File

@ -1116,6 +1116,7 @@ static inline int has_sha1_file(const unsigned char *sha1)
/* Same as the above, except for struct object_id. */
extern int has_object_file(const struct object_id *oid);
extern int has_object_file_with_flags(const struct object_id *oid, int flags);
/*
* Return true iff an alternate object database has a loose object

5
sha1_file.c
View File

@ -3239,6 +3239,11 @@ int has_object_file(const struct object_id *oid)
return has_sha1_file(oid->hash);
}
int has_object_file_with_flags(const struct object_id *oid, int flags)
{
return has_sha1_file_with_flags(oid->hash, flags);
}
static void check_tree(const void *buf, size_t size)
{
struct tree_desc desc;

99
t/perf/p5550-fetch-tags.sh Executable file
View File

@ -0,0 +1,99 @@
#!/bin/sh
test_description='performance of tag-following with many tags
This tests a fairly pathological case, so rather than rely on a real-world
case, we will construct our own repository. The situation is roughly as
follows.
The parent repository has a large number of tags which are disconnected from
the rest of history. That makes them candidates for tag-following, but we never
actually grab them (and thus they will impact each subsequent fetch).
The child repository is a clone of parent, without the tags, and is at least
one commit behind the parent (meaning that we will fetch one object and then
examine the tags to see if they need followed). Furthermore, it has a large
number of packs.
The exact values of "large" here are somewhat arbitrary; I picked values that
start to show a noticeable performance problem on my machine, but without
taking too long to set up and run the tests.
'
. ./perf-lib.sh
# make a long nonsense history on branch $1, consisting of $2 commits, each
# with a unique file pointing to the blob at $2.
create_history () {
perl -le '
my ($branch, $n, $blob) = @ARGV;
for (1..$n) {
print "commit refs/heads/$branch";
print "committer nobody <nobody@example.com> now";
print "data 4";
print "foo";
print "M 100644 $blob $_";
}
' "$@" |
git fast-import --date-format=now
}
# make a series of tags, one per commit in the revision range given by $@
create_tags () {
git rev-list "$@" |
perl -lne 'print "create refs/tags/$. $_"' |
git update-ref --stdin
}
# create $1 nonsense packs, each with a single blob
create_packs () {
perl -le '
my ($n) = @ARGV;
for (1..$n) {
print "blob";
print "data <<EOF";
print "$_";
print "EOF";
}
' "$@" |
git fast-import &&
git cat-file --batch-all-objects --batch-check='%(objectname)' |
while read sha1
do
echo $sha1 | git pack-objects .git/objects/pack/pack
done
}
test_expect_success 'create parent and child' '
git init parent &&
git -C parent commit --allow-empty -m base &&
git clone parent child &&
git -C parent commit --allow-empty -m trigger-fetch
'
test_expect_success 'populate parent tags' '
(
cd parent &&
blob=$(echo content | git hash-object -w --stdin) &&
create_history cruft 3000 $blob &&
create_tags cruft &&
git branch -D cruft
)
'
test_expect_success 'create child packs' '
(
cd child &&
git config gc.auto 0 &&
git config gc.autopacklimit 0 &&
create_packs 500
)
'
test_perf 'fetch' '
# make sure there is something to fetch on each iteration
git -C child update-ref -d refs/remotes/origin/master &&
git -C child fetch
'
test_done