git@vger.kernel.org mailing list mirror (one of many)
 help / color / mirror / code / Atom feed
From: Johannes Schindelin <johannes.schindelin@gmx.de>
To: git@vger.kernel.org
Cc: Junio C Hamano <gitster@pobox.com>,
	Lars Schneider <larsxschneider@gmail.com>,
	Eric Sunshine <sunshine@sunshineco.com>
Subject: [PATCH v2 0/2] Do not lock temporary files via child processes on Windows
Date: Thu, 18 Aug 2016 16:51:01 +0200 (CEST)	[thread overview]
Message-ID: <cover.1471531799.git.johannes.schindelin@gmx.de> (raw)
In-Reply-To: <cover.1471437637.git.johannes.schindelin@gmx.de>

This issue was originally reported and fixed in
https://github.com/git-for-windows/git/pull/755

The problem is that file handles to temporary files (such as
index.lock) were inherited by spawned processes. If those spawned
processes do not exit before the parent process wants to delete or
rename them, we are in big trouble.

The original use case triggering the bug is a merge driver that does
not quit, but listen to subsequent merge requests.

However, the same issue turned up in Lars Schneider's work on making
clean/smudge filters batchable (i.e. more efficient by avoiding
possibly thousands of child processes, one per file).

Changes since v1:

- the two commit messages have been corrected, as per Junio's and Eric's
  suggestion,
- lockfile.h and tempfile.h now sport explicit documentation that the
  current process needs to write to the files, no spawned processes.


Ben Wijen (2):
  t6026-merge-attr: child processes must not inherit index.lock handles
  mingw: ensure temporary file handles are not inherited by child
    processes

 compat/mingw.h        |  4 ++++
 lockfile.h            |  4 ++++
 t/t6026-merge-attr.sh | 13 +++++++++++++
 tempfile.c            |  2 +-
 tempfile.h            |  4 ++++
 5 files changed, 26 insertions(+), 1 deletion(-)

Published-As: https://github.com/dscho/git/releases/tag/mingw-index-lock-v2
Fetch-It-Via: git fetch https://github.com/dscho/git mingw-index-lock-v2

Interdiff vs v1:

 diff --git a/lockfile.h b/lockfile.h
 index 3d30193..d26ad27 100644
 --- a/lockfile.h
 +++ b/lockfile.h
 @@ -55,6 +55,10 @@
   *   * calling `fdopen_lock_file()` to get a `FILE` pointer for the
   *     open file and writing to the file using stdio.
   *
 + *   Note that the file descriptor returned by hold_lock_file_for_update()
 + *   is marked O_CLOEXEC, so the new contents must be written by the
 + *   current process, not a spawned one.
 + *
   * When finished writing, the caller can:
   *
   * * Close the file descriptor and rename the lockfile to its final
 diff --git a/tempfile.h b/tempfile.h
 index 4219fe4..2f0038d 100644
 --- a/tempfile.h
 +++ b/tempfile.h
 @@ -33,6 +33,10 @@
   *   * calling `fdopen_tempfile()` to get a `FILE` pointer for the
   *     open file and writing to the file using stdio.
   *
 + *   Note that the file descriptor returned by create_tempfile()
 + *   is marked O_CLOEXEC, so the new contents must be written by
 + *   the current process, not any spawned one.
 + *
   * When finished writing, the caller can:
   *
   * * Close the file descriptor and remove the temporary file by

-- 
2.9.2.691.g78954f3

base-commit: d63263a4dee8fc7da9b97bbdedf9c0d1f33024d4

  parent reply	other threads:[~2016-08-19  2:30 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-08-17 12:40 [PATCH 0/2] Do not lock temporary files via child processes on Windows Johannes Schindelin
2016-08-17 12:40 ` [PATCH 1/2] t6026-merge-attr: child processes must not inherit index.lock handles Johannes Schindelin
2016-08-17 17:55   ` Junio C Hamano
2016-08-17 12:41 ` [PATCH 2/2] mingw: ensure temporary file handles are not inherited by child processes Johannes Schindelin
2016-08-17 12:48   ` Eric Sunshine
2016-08-18 14:42     ` Johannes Schindelin
2016-08-17 13:14   ` Lars Schneider
2016-08-17 18:22   ` Junio C Hamano
2016-08-18 14:50     ` Johannes Schindelin
2016-08-18 17:35   ` Eric Wong
2016-08-18 21:53     ` Junio C Hamano
2016-08-18 22:48       ` Eric Wong
2016-08-19 15:57         ` Junio C Hamano
2016-08-22 12:47           ` Johannes Schindelin
2016-08-18 14:51 ` Johannes Schindelin [this message]
2016-08-18 14:51   ` [PATCH v2 1/2] t6026-merge-attr: child processes must not inherit index.lock handles Johannes Schindelin
2016-08-18 14:51   ` [PATCH v2 2/2] mingw: ensure temporary file handles are not inherited by child processes Johannes Schindelin
2016-08-22 12:47   ` [PATCH v3 0/2] Do not lock temporary files via child processes on Windows Johannes Schindelin
2016-08-22 12:47     ` [PATCH v3 1/2] t6026-merge-attr: child processes must not inherit index.lock handles Johannes Schindelin
2016-08-22 12:47     ` [PATCH v3 2/2] mingw: ensure temporary file handles are not inherited by child processes Johannes Schindelin
2016-08-22 17:58       ` Eric Wong

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.1471531799.git.johannes.schindelin@gmx.de \
    --to=johannes.schindelin@gmx.de \
    --cc=git@vger.kernel.org \
    --cc=gitster@pobox.com \
    --cc=larsxschneider@gmail.com \
    --cc=sunshine@sunshineco.com \
    /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).