git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Stefan Beller <sbeller@google.com>
To: git@vger.kernel.org, gitster@pobox.com, mhagger@alum.mit.edu,
	peff@peff.net, loic@dachary.org
Cc: Stefan Beller <sbeller@google.com>
Subject: [PATCH 6/6] refs.c: enable large transactions
Date: Wed, 21 Jan 2015 15:23:45 -0800	[thread overview]
Message-ID: <1421882625-916-7-git-send-email-sbeller@google.com> (raw)
In-Reply-To: <1421882625-916-1-git-send-email-sbeller@google.com>

By closing the file descriptors after creating the lock file we are not
limiting the size of the transaction by the number of available file
descriptors.

Signed-off-by: Stefan Beller <sbeller@google.com>
---
 refs.c                | 14 +++++++++++++-
 t/t1400-update-ref.sh |  4 ++--
 2 files changed, 15 insertions(+), 3 deletions(-)

diff --git a/refs.c b/refs.c
index 0161667..014ef59 100644
--- a/refs.c
+++ b/refs.c
@@ -3068,7 +3068,17 @@ static ssize_t write_in_full_to_lockfile(struct lock_file *lock,
 					 const void *buf,
 					 size_t count)
 {
-	return write_in_full(lock->fd, buf, count);
+	if (lock->fd == -1) {
+		int ret;
+		if (reopen_lock_file(lock) < 0)
+			return -1;
+		ret = write_in_full(lock->fd, buf, count);
+		if (close_lock_file(lock) < 0)
+			return -1;
+		return ret;
+	} else {
+		return write_in_full(lock->fd, buf, count);
+	}
 }
 
 /*
@@ -3797,6 +3807,8 @@ int ref_transaction_commit(struct ref_transaction *transaction,
 				    update->refname);
 			goto cleanup;
 		}
+		/* Do not keep all lock files open at the same time. */
+		close_lock_file(update->lock->lk);
 	}
 
 	/* Perform updates first so live commits remain referenced */
diff --git a/t/t1400-update-ref.sh b/t/t1400-update-ref.sh
index ea98b9b..751b6dc 100755
--- a/t/t1400-update-ref.sh
+++ b/t/t1400-update-ref.sh
@@ -1071,7 +1071,7 @@ run_with_limited_open_files () {
 
 test_lazy_prereq ULIMIT 'run_with_limited_open_files true'
 
-test_expect_failure ULIMIT 'large transaction creating branches does not burst open file limit' '
+test_expect_success ULIMIT 'large transaction creating branches does not burst open file limit' '
 (
 	for i in $(seq 33)
 	do
@@ -1082,7 +1082,7 @@ test_expect_failure ULIMIT 'large transaction creating branches does not burst o
 )
 '
 
-test_expect_failure ULIMIT 'large transaction deleting branches does not burst open file limit' '
+test_expect_success ULIMIT 'large transaction deleting branches does not burst open file limit' '
 (
 	for i in $(seq 33)
 	do
-- 
2.2.1.62.g3f15098

  parent reply	other threads:[~2015-01-21 23:24 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-01-21 23:23 [PATCHv1 0/6] Fix bug in large transactions Stefan Beller
2015-01-21 23:23 ` [PATCH 1/6] update-ref: Test handling large transactions properly Stefan Beller
2015-01-21 23:34   ` Jeff King
2015-01-21 23:23 ` [PATCH 2/6] refs.c: remove lock_fd from struct ref_lock Stefan Beller
2015-01-21 23:23 ` [PATCH 3/6] refs.c: replace write_str_in_full by write_in_full Stefan Beller
2015-01-21 23:38   ` Jeff King
2015-01-21 23:44     ` Stefan Beller
2015-01-21 23:52       ` Jeff King
2015-01-21 23:23 ` [PATCH 4/6] refs.c: Have a write_in_full_to_lock_file wrapping write_in_full Stefan Beller
2015-01-21 23:40   ` Jeff King
2015-01-21 23:23 ` [PATCH 5/6] refs.c: write to a lock file only once Stefan Beller
2015-01-21 23:44   ` Jeff King
2015-01-21 23:23 ` Stefan Beller [this message]
2015-01-21 23:47 ` [PATCHv1 0/6] Fix bug in large transactions Jeff King
2015-01-22  8:00   ` Junio C Hamano
2015-01-22 17:52     ` Stefan Beller
2015-01-22 17:58       ` Junio C Hamano
2015-01-22 18:29         ` Stefan Beller

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1421882625-916-7-git-send-email-sbeller@google.com \
    --to=sbeller@google.com \
    --cc=git@vger.kernel.org \
    --cc=gitster@pobox.com \
    --cc=loic@dachary.org \
    --cc=mhagger@alum.mit.edu \
    --cc=peff@peff.net \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).