From: Taylor Blau <me@ttaylorr.com>
To: git@vger.kernel.org
Cc: Johannes.Schindelin@gmx.de, dstolee@microsoft.com,
jeffhost@microsoft.com, peff@peff.net
Subject: [PATCH 0/4] midx: avoid potential rename-while-mapped on Windows
Date: Fri, 8 Oct 2021 17:46:25 -0400 [thread overview]
Message-ID: <cover.1633729502.git.me@ttaylorr.com> (raw)
(This topic is based on tb/repack-write-midx).
While looking through some failures in t5319 when Git is compiled with
SANITIZE=leak, I noticed a curious pattern that basically looks like:
struct multi_pack_index *m = load_multi_pack_index(object_dir, 0);
/* ... */
write_midx_internal();
This can cause problems on platforms where a memory mapped file cannot be
overridden via a rename, because load_multi_pack_index() builds a new MIDX from
scratch each time it is called without adding that MIDX to the object store. But
write_midx_internal() loads a MIDX from the object store, causing the same
region to be mapped twice.
The details are fully spelled out in the second patch, but the gist is that
while write_midx_internal() unmaps *its* copy of the MIDX, the caller's copy is
still mapped in memory.
Strangely, I could not get this to produce a failure on Windows, even though I
would have expected one. I asked around off-list, and it sounds like it's
possible this behavior has changed, or Windows' mmap-equivalent may not lock
files that fit within a single page.
In any case, we let's side step the issue entirely by only interacting with
MIDXs that can be found from the object store.
Taylor Blau (4):
midx.c: extract MIDX lookup by object_dir
midx.c: lookup MIDX by object directory during expire
midx.c: lookup MIDX by object directory during repack
midx.c: guard against commit_lock_file() failures
midx.c | 42 +++++++++++++++++++++++-------------------
1 file changed, 23 insertions(+), 19 deletions(-)
--
2.33.0.96.g73915697e6
next reply other threads:[~2021-10-08 21:46 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-10-08 21:46 Taylor Blau [this message]
2021-10-08 21:46 ` [PATCH 1/4] midx.c: extract MIDX lookup by object_dir Taylor Blau
2021-10-08 21:46 ` [PATCH 2/4] midx.c: lookup MIDX by object directory during expire Taylor Blau
2021-10-13 13:28 ` Derrick Stolee
2021-10-08 21:46 ` [PATCH 3/4] midx.c: lookup MIDX by object directory during repack Taylor Blau
2021-10-08 21:46 ` [PATCH 4/4] midx.c: guard against commit_lock_file() failures Taylor Blau
2021-10-09 2:07 ` Ævar Arnfjörð Bjarmason
2021-10-13 13:29 ` [PATCH 0/4] midx: avoid potential rename-while-mapped on Windows Derrick Stolee
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
List information: http://vger.kernel.org/majordomo-info.html
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=cover.1633729502.git.me@ttaylorr.com \
--to=me@ttaylorr.com \
--cc=Johannes.Schindelin@gmx.de \
--cc=dstolee@microsoft.com \
--cc=git@vger.kernel.org \
--cc=jeffhost@microsoft.com \
--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.
Code repositories for project(s) associated with this public inbox
https://80x24.org/mirrors/git.git
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 read-only IMAP folder(s) and NNTP newsgroup(s).