git@vger.kernel.org mailing list mirror (one of many)
 help / color / mirror / code / Atom feed
From: Junio C Hamano <gitster@pobox.com>
To: Jeff King <peff@peff.net>
Cc: Jonathan Nieder <jrnieder@gmail.com>, git@vger.kernel.org
Subject: Re: [PATCH v2 7/8] verify_path(): disallow symlinks in .gitattributes and .gitignore
Date: Tue, 27 Oct 2020 15:00:36 -0700	[thread overview]
Message-ID: <xmqqv9ev9vnf.fsf@gitster.c.googlers.com> (raw)
In-Reply-To: <20201027075853.GH3005508@coredump.intra.peff.net> (Jeff King's message of "Tue, 27 Oct 2020 03:58:53 -0400")

Jeff King <peff@peff.net> writes:

> diff --git a/environment.c b/environment.c
> index bb518c61cd..7c233e0e0e 100644
> --- a/environment.c
> +++ b/environment.c
> @@ -73,6 +73,7 @@ int merge_log_config = -1;
>  int precomposed_unicode = -1; /* see probe_utf8_pathname_composition() */
>  unsigned long pack_size_limit_cfg;
>  enum log_refs_config log_all_ref_updates = LOG_REFS_UNSET;
> +int allow_external_symlinks = 1;

OK, so by default it is not blocked...

> +static int symlink_leaves_repo(const char *target, const char *linkpath)
> +{
> +	/*
> +	 * Absolute paths are always considered to leave the repository (even
> +	 * if they happen to point to the working tree path).
> +	 */
> +	if (is_absolute_path(target))
> +		return 1;

Very sensible.

> +	/*
> +	 * Allow relative paths that start with a sequence of "../",
> +	 * as long as they do not break out of the symlink's root.
> +	 * This loop will detect break-out cases and return; otherwise, at the
> +	 * end of the loop "target" will point to the first non-".." component.
> +	 *
> +	 * We count the depth of linkpath by eating up directory components left
> +	 * to right. Technically the symlink would resolve right-to-left, but
> +	 * we don't care about the actual values, only the number.
> +	 */
> +	while (target[0] == '.') {
> +		if (!target[1]) {
> +			/* trailing "." -- ignore */
> +			target++;
> +		} else if (is_dir_sep(target[1])) {
> +			/* "./" -- ignore */
> +			target += 2;
> +		} else if (target[1] == '.' &&
> +			   (!target[2] || is_dir_sep(target[2]))) {
> +			/* ".." or "../" -- drop one from linkpath depth */
> +			while (!is_dir_sep(*linkpath)) {
> +				/* end-of-string; target exceeded our depth */
> +				if (!*linkpath)
> +					return 1;
> +				linkpath++;
> +			}
> +			/* skip final "/" */
> +			linkpath++;
> +
> +			/* skip past ".." */
> +			target += 2;
> +			/* and "/" if present */
> +			if (is_dir_sep(*target))
> +				target++;
> +		}
> +	}
> +
> +	/*
> +	 * Now we have a path in "target" that only go down into the tree.
> +	 * Disallow any interior "../", like "foo/../bar". These might be
> +	 * OK, but we cannot know unless we know whether "foo" is itself a
> +	 * symlink. So err on the side of caution.
> +	 */
> +	while (*target) {
> +		const char *v;
> +		if (skip_prefix(target, "..", &v) && (!*v || is_dir_sep(*v)))
> +			return 1;
> +		target++;
> +	}
> +
> +	return 0;
> +}
> +
> +int safe_symlink(const char *target, const char *linkpath)
> +{
> +	if (!allow_external_symlinks &&
> +	    symlink_leaves_repo(target, linkpath)) {
> +		errno = EPERM;
> +		return -1;
> +	}
> +
> +	return symlink(target, linkpath);
> +}

OK.  This is only about blocking creation of new symbolic links that
goes outside the working tree.  It obviously is a good thing to do.

We have some "symlink safety" in various parts of the system [*1*],
and I wonder if we can somehow consolidate the support to a more
central place.

Thanks.


[Footnote]

*1* For example, apply tries to be careful not to take the "path"
    recorded in the incoming patch blindly, and instead checks if
    any path component in it is a symbolic link before touching.
    Similarly, callers of has_symlink_leading_path() all try to be
    careful when the "path" they want to use to access a filesystem
    entity has a symbolic link in the middle on the filesystem.

  reply	other threads:[~2020-10-27 22:00 UTC|newest]

Thread overview: 55+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-10-05  7:17 [PATCH 0/7] forbidding symlinked .gitattributes and .gitignore Jeff King
2020-10-05  7:19 ` [PATCH 1/7] fsck_tree(): fix shadowed variable Jeff King
2020-10-05  7:44   ` Jonathan Nieder
2020-10-05  8:20     ` Jeff King
2020-10-05  8:29       ` Jonathan Nieder
2020-10-05  7:19 ` [PATCH 2/7] fsck_tree(): wrap some long lines Jeff King
2020-10-05  7:46   ` Jonathan Nieder
2020-10-05  7:19 ` [PATCH 3/7] t7415: rename to expand scope Jeff King
2020-10-05  7:50   ` Jonathan Nieder
2020-10-05  8:24     ` Jeff King
2020-10-05  8:34       ` Jonathan Nieder
2020-10-05  8:49         ` Jeff King
2020-10-05  7:20 ` [PATCH 4/7] t7450: test verify_path() handling of gitmodules Jeff King
2020-10-05  7:53   ` Jonathan Nieder
2020-10-05  8:30     ` Jeff King
2020-10-05  8:38       ` Jonathan Nieder
2020-10-05  7:21 ` [PATCH 5/7] t0060: test obscured .gitattributes and .gitignore matching Jeff King
2020-10-05  8:03   ` Jonathan Nieder
2020-10-05  8:40     ` Jeff King
2020-10-05 21:20       ` Johannes Schindelin
2020-10-06 14:01         ` Jeff King
2020-10-05  7:24 ` [PATCH 6/7] verify_path(): disallow symlinks in .gitattributes and .gitignore Jeff King
2020-10-05  8:09   ` Jonathan Nieder
2020-10-05 12:07     ` Jeff King
2020-10-05  7:25 ` [PATCH 7/7] fsck: complain when .gitattributes or .gitignore is a symlink Jeff King
2020-10-05  8:12   ` Jonathan Nieder
2020-10-05  8:53     ` Jeff King
2020-10-05  7:32 ` [PATCH 0/7] forbidding symlinked .gitattributes and .gitignore Jonathan Nieder
2020-10-05  8:58   ` Jeff King
2020-10-05 12:16 ` [PATCH v2 0/8] " Jeff King
2020-10-05 12:16   ` [PATCH v2 1/8] fsck_tree(): fix shadowed variable Jeff King
2020-10-05 12:16   ` [PATCH v2 2/8] fsck_tree(): wrap some long lines Jeff King
2020-10-05 12:16   ` [PATCH v2 3/8] t7415: rename to expand scope Jeff King
2020-10-05 12:16   ` [PATCH v2 4/8] t7450: test verify_path() handling of gitmodules Jeff King
2020-10-05 12:16   ` [PATCH v2 5/8] t7450: test .gitmodules symlink matching against obscured names Jeff King
2020-10-05 12:16   ` [PATCH v2 6/8] t0060: test obscured .gitattributes and .gitignore matching Jeff King
2020-10-05 12:16   ` [PATCH v2 7/8] verify_path(): disallow symlinks in .gitattributes and .gitignore Jeff King
2020-10-27  3:35     ` Jonathan Nieder
2020-10-27  7:58       ` Jeff King
2020-10-27 22:00         ` Junio C Hamano [this message]
2020-10-28  9:41           ` Jeff King
2020-10-27 23:43         ` Jonathan Nieder
2020-10-28 19:18           ` Junio C Hamano
2020-10-05 12:16   ` [PATCH v2 8/8] fsck: complain when .gitattributes or .gitignore is a symlink Jeff King
2020-10-06 20:41   ` [PATCH v2 0/8] forbidding symlinked .gitattributes and .gitignore Junio C Hamano
2020-10-20 23:19   ` Philip Oakley
2020-10-23  8:17     ` [PATCH] documentation symlink restrictions for .git* files Jeff King
2020-10-23  8:27       ` Jeff King
2020-10-26 22:18       ` Philip Oakley
2020-10-26 22:53         ` Jeff King
2020-10-26 23:32           ` Junio C Hamano
2020-10-27  7:26             ` Jeff King
2020-10-27 18:45               ` Junio C Hamano
2020-10-27 21:00                 ` Philip Oakley
2020-10-28 19:14                   ` Junio C Hamano

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=xmqqv9ev9vnf.fsf@gitster.c.googlers.com \
    --to=gitster@pobox.com \
    --cc=git@vger.kernel.org \
    --cc=jrnieder@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).