On Fri, 21 Nov 2003, Jim Carter wrote:
> Here at UCLA-Mathnet we use the automounter extensively, and we have had
> a couple of associated problems:
Thanks for your efforts but ...
>
> A. Our NFS mounted filesystems invariably require submounts, and they
> are not auto-dismounted. Ever. This proved to be due to a duplicated
> slash between the main and sub mount points. Once this was fixed,
> auto-dismount started working. There is a rumor that a related fix was
> posted somewhere, but I couldn't find it.
I believe I fixed this in the 4.0.0 release. There are a couple of other
fixes in it also.
>
> B. If the daemon is hit with SIGUSR1, it goes into an infinite loop
> trying unsuccessfully to dismount eligible filesystems, spitting out
> typically 1000 syslog messages over 2 seconds until item C (below)
> supervenes. I put in both a rate throttle (20/second) and a dynamic
> limit on the number of dismounts.
This sounds like a problem that needs to be identified and fixed.
Rate throttling seems more of a workaround that a solution.
Can you give more information please.
Come to think about it this might have already been fixed in ????
Maybe it's not in the 4.0.0 release. I'll check.
>
> C. Upon auto-dismount or SIGUSR1 looping, st_prepare_shutdown is called
> when ap.state != ST_READY and an assertion fails, killing the thread.
> I changed it to die on ST_SHUTDOWN_PENDING, i.e. a recursive call. I'm
> not 100% sure that this is the correct contingency, but automount does
> dismount the unused filesystems and does exit.
Have seen this. I'm not sure if I fixed this in the 4.0.0 release either.
Will check into it.
>
> D. It would appear that a maliciously constructed directory name could
> overflow a buffer in several places, at least causing denial of service
> and possibly allowing the execution of code. Perhaps other O.S. limits
> on the length of a filename (PATH_MAX) invariably protect the daemon
> from this exposure, but defense in depth in this area seems both
> warranted and not burdensome. I changed sprintf to snprintf wherever
> occurring, and the subroutine which joins the dirname and basename
> checks the buffer size.
Definately will check into this for 4.0 and 4.1.
>
> E. There were a few cases where size_t and int were mixed together,
> causing compiler warnings. As there have been exploits against mixed
> signed-unsigned variables, I took the opportunity to fix this issue.
Ditto.
>
> On one machine we have been running the patched automount for a week,
> and all Linux boxes have had it in production for 48 hours. No peculiar
> log messages or crashes have been seen. In one test, various random NFS
> filesystems were automounted and allowed to time out, with varying
> numbers of filesystems simultaneously mounted. Almost 500
> mount-autodismount pairs were done; processes always exited when they
> should, and no error messages appeared. There's a pretty good chance
> that this patch is working.
Assume you mean the attached patch.
>
> The patches follow. They are against autofs-4.0.0pre10, which is the
> version distributed with SuSE 8.2, the distro we are using.
The SuSE maintainer contacted me a while ago, sent me a copy of his
autofs which was much appreciated. I merged some of the SuSE patches into
the current 4.1.0 beta.
I hope to encourage him to adopt 4.1.0 when a final version is released.
--
,-._|\ Ian Kent
/ \ Perth, Western Australia
*_.--._/ E-mail: [EMAIL PROTECTED]
v Web: http://themaw.net/
_______________________________________________
autofs mailing list
[EMAIL PROTECTED]
http://linux.kernel.org/mailman/listinfo/autofs
