From: Rafael Silva <rafaeloliveira.cs@gmail.com>
To: git@vger.kernel.org
Cc: "Derrick Stolee" <dstolee@microsoft.com>,
"Nguyễn Thái Ngọc Duy" <pclouds@gmail.com>,
"Jeff King" <peff@peff.net>,
"Rafael Silva" <rafaeloliveira.cs@gmail.com>
Subject: [PATCH 0/1] maintenance: Fix a SEGFAULT when no repository when running git maintenance run/start
Date: Tue, 24 Nov 2020 16:44:04 +0000 [thread overview]
Message-ID: <20201124164405.29327-1-rafaeloliveira.cs@gmail.com> (raw)
In d7514f6ed5 (maintenance: take a lock on the objects directory, 2020-09-17) [1],
and 2fec604f8d (maintenance: add start/stop subcommands, 2020-09-11) [2] The
"git maintenance run" and "git maintenance start" was taught to hold a file-based
lock at the .git/maintenance.lock and .git/schedule.lock respectively because these
operations involves writing data into the .git/repository.
The lock file path string is built using the the_repository->objects->odb->path,
in case the_repository->objects->odb is NULL when there is not repository available,
resulting in a SEGFAULT.
[1] https://lore.kernel.org/git/1a0a3eebb825ac3eabfdd86f82ed7ef6abb454c5.1600366313.git.gitgitgadget@gmail.com/
[2] https://lore.kernel.org/git/5194f6b1facbd14cc17eea0337c0cc397a2a51fc.1602782524.git.gitgitgadget@gmail.com/
In order to reproduce the error, one can execute maintenance "run" and/or
"start" subcommand with a non valid repository:
$ git -C /tmp maintenance start
Segmentation fault
$ git -C /tmp maintenance run
Segmentation fault
The above test was executed from a git built from commit: faefdd61ec (Sixth batch, 2020-11-18):
For reference here's the output from GDB when debugging the "start" command
Program received signal SIGSEGV, Segmentation fault.
0x00005555555b9b4c in maintenance_run_tasks (opts=0x7fffffffded4) at builtin/gc.c:1268
1268 char *lock_path = xstrfmt("%s/maintenance", r->objects->odb->path);
This patch serie adds a check on the maintenance_run_tasks() and update_background_schedule()
to check if the_repository->git_dir is set to validate if a git repository is available and
return error otherwise. The logic is borrowed from maintenance_unregister() that performs
the same validation with different error message.
A new test case is added into t/t7900-maintenance.sh to cover these cases. The test is based on
the assumption that the `/tmp` directory is available, readable and is not a git repository
which I hope is fine for the running the test is all platforms.
This patch is based on faefdd61ec (Sixth batch, 2020-11-18) (master branch) as the both commits
that introduced the file-based lock are graduated to master already. Hope this also plays nice
with the integration branches maintenance-part-v[1,4].
Rafael Silva (1):
maintenance: fix a SEGFAULT when no repository
builtin/gc.c | 14 ++++++++++++--
t/t7900-maintenance.sh | 5 +++++
2 files changed, 17 insertions(+), 2 deletions(-)
--
2.29.2.505.g04529851e5
next reply other threads:[~2020-11-24 16:45 UTC|newest]
Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-11-24 16:44 Rafael Silva [this message]
2020-11-24 16:44 ` [PATCH 1/1] maintenance: fix a SEGFAULT when no repository Rafael Silva
2020-11-24 17:22 ` Derrick Stolee
2020-11-24 21:48 ` Junio C Hamano
2020-11-24 21:59 ` Derrick Stolee
2020-11-26 8:22 ` Rafael Silva
2020-11-26 11:21 ` Derrick Stolee
2020-11-24 17:24 ` Eric Sunshine
2020-11-24 19:14 ` SZEDER Gábor
2020-11-24 19:34 ` Eric Sunshine
2020-11-26 7:13 ` Rafael Silva
2020-11-24 19:03 ` Martin Ågren
2020-11-26 7:07 ` Rafael Silva
2020-11-26 20:41 ` [PATCH v2 0/1] maintenance: Fix SEGFAULT when running outside of a repository Rafael Silva
2020-11-26 20:41 ` [PATCH v2 1/1] maintenance: fix SEGFAULT when no repository Rafael Silva
2020-11-27 20:43 ` Derrick Stolee
2020-12-08 20:12 ` Josh Steadmon
2020-12-08 21:58 ` Junio C Hamano
2020-12-08 22:17 ` Junio C Hamano
2020-12-24 8:12 ` Christian Couder
2020-12-24 14:14 ` Junio C Hamano
2020-12-09 9:29 ` Rafael Silva
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=20201124164405.29327-1-rafaeloliveira.cs@gmail.com \
--to=rafaeloliveira.cs@gmail.com \
--cc=dstolee@microsoft.com \
--cc=git@vger.kernel.org \
--cc=pclouds@gmail.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).