From: Andreas Schwab <schwab@suse.de>
To: Florian Weimer <fweimer@redhat.com>
Cc: libc-alpha@sourceware.org
Subject: Re: [PATCH] Avoid concurrency problem in ldconfig (bug 23973)
Date: Thu, 18 Apr 2019 14:59:46 +0200 [thread overview]
Message-ID: <mvmv9zbpiyl.fsf@suse.de> (raw)
In-Reply-To: <87bm13y11o.fsf@oldenburg2.str.redhat.com> (Florian Weimer's message of "Thu, 18 Apr 2019 14:01:55 +0200")
On Apr 18 2019, Florian Weimer <fweimer@redhat.com> wrote:
> * Andreas Schwab:
>
>> Use a unique name for the temporary file when updating the ld.so cache, so
>> that two concurrent runs of ldconfig don't write to the same file.
>>
>> * elf/cache.c (save_cache): Use unique temporary name.
>> (save_aux_cache): Likewise.
>
> The downside of this change is that if ldconfig is interrupted, the
> temporary file never goes away.
>
> Ideally, we would use O_TMPFILE if supported by the (file) system, but
> that can get quite involved.
That shortens the window, but won't close it, since we need a name to
pass to rename in any case.
>> diff --git a/elf/cache.c b/elf/cache.c
>> index b8e9e6ccc3..ec7d94b0bc 100644
>> --- a/elf/cache.c
>> +++ b/elf/cache.c
>> @@ -427,12 +427,12 @@ save_cache (const char *cache_name)
>> /* Write out the cache. */
>>
>> /* Write cache first to a temporary file and rename it later. */
>> - char *temp_name = xmalloc (strlen (cache_name) + 2);
>> - sprintf (temp_name, "%s~", cache_name);
>> + char *temp_name;
>> + if (asprintf (&temp_name, "%s.XXXXXX", cache_name) < 0)
>> + error (EXIT_FAILURE, errno, _("Can't allocate temporary name for cache file"));
>>
>> /* Create file. */
>> - int fd = open (temp_name, O_CREAT|O_WRONLY|O_TRUNC|O_NOFOLLOW,
>> - S_IRUSR|S_IWUSR);
>> + int fd = mkostemp (temp_name, O_NOFOLLOW);
>
> I think you can use mkstemp because O_NOFOLLOW is implied by its use of
> O_EXCL.
Yes, O_NOFOLLOW is useless anyway, since mk[o]stemp is about creating a
new file.
Andreas.
--
Andreas Schwab, SUSE Labs, schwab@suse.de
GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE 1748 E4D4 88E3 0EEA B9D7
"And now for something completely different."
next prev parent reply other threads:[~2019-04-18 12:59 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-02-14 16:51 [PATCH] Avoid concurrency problem in ldconfig (bug 23973) Andreas Schwab
2019-04-18 12:01 ` Florian Weimer
2019-04-18 12:09 ` Christian Brauner
2019-04-18 12:11 ` Florian Weimer
2019-04-18 12:59 ` Andreas Schwab [this message]
2019-04-18 14:47 ` Florian Weimer
2019-04-18 15:16 ` Andreas Schwab
2019-04-18 15:52 ` Florian Weimer
2019-04-18 16:04 ` Andreas Schwab
-- strict thread matches above, loose matches on Subject: below --
2019-01-21 15:30 Andreas Schwab
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: https://www.gnu.org/software/libc/involved.html
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=mvmv9zbpiyl.fsf@suse.de \
--to=schwab@suse.de \
--cc=fweimer@redhat.com \
--cc=libc-alpha@sourceware.org \
/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 read-only IMAP folder(s) and NNTP newsgroup(s).