Hello Junjiro,
I remember you wrote that email but I thought Jan took necessary steps to apply
your suggestion to posixovl. He didn't. So I did :)
I've replaced fstatvfs() call in posixovl by fstatfs() as you suggest in aufs
man page.
And it works now, thank you very much!
I have to reread aufs man page. Because I see there is a lot of new features
since the time I last checked it.
Tomas M
[EMAIL PROTECTED] wrote:
> Tomas M:
>> These processes are blocked:
>> 1 (init)
>> 2190 (posixovl)
>> 3236 (udevd)
>> 5291 (klogd)
>> 5352 (dbus-daemon)
>
> Your log shows two processes, dbus-daemon and posixovl are in deadlock.
> And I am afraid there may be other locked process.
>
> It seems a known problem which I have reported to you.
> See the attached my old mail. That says a problem is in fuse branch and
> aufs create=mfs mode. But the same problem may happen regardless the mfs
> mode actually.
> Please try the patch first.
>
>
> Junjiro R. Okajima
>
> ----------------------------------------------------------------------
>
> From: [EMAIL PROTECTED]
> To: [EMAIL PROTECTED], [EMAIL PROTECTED]
> Subject: posixovl and aufs most-free-space mode
> Date: Mon, 05 May 2008 17:48:29 +0900
>
> ------- =_aaaaaaaaaa0
> Content-Type: text/plain; charset="us-ascii"
> Content-ID: <[EMAIL PROTECTED]>
>
> Hello Tomas and Jan,
> I hope you are doing well.
>
> In last month, I received a question from one of aufs user, saying he
> wanted to use vfat as an aufs branch. I answered him to try posixovl,
> while my recommendation is loopback-mount of ext2 or something. :-)
> He tried it and found a trouble in aufs most-free-space mode (create=mfs
> option).
>
> I dived into posixovl and found it is very hard to solve this problem in
> aufs. Finally, I made a patch against posixovl. I know SLAX doesn't use
> mfs mode, but I think it is better to let you know this issue.
> Here is a note in the aufs manual I will add in next Monday release, and
> a patch against the latest posixovl.c which I got today by,
> $ git clone git://dev.computergmbh.de/posixovl posixovl.git
>
>
> Junjiro Okajima
>
> ----------------------------------------------------------------------
>
> diff -x CVS -x RCS -x .pdiff -rup ../../sf.anon/aufs/util/aufs.in.5
> ./util/aufs.in.5
> --- ../../sf.anon/aufs/util/aufs.in.5 2008-04-28 12:30:53.000000000 +0900
> +++ ./util/aufs.in.5 2008-05-05 17:16:51.000000000 +0900
> @@ -1064,9 +1064,61 @@ the performance, you can specify the dur
> aufs hold the index of last selected writable branch until the
> specified seconds expires. The first time you create something in aufs
> after the specified seconds expired, aufs checks the amount of free
> -space of all writable branches and the held branch index will be
> -updated.
> +space of all writable branches by internal statfs call
> +and the held branch index will be updated.
> The default value is \*[AUFS_MFS_SECOND_DEF] seconds.
> +
> +In this mode, a FUSE branch needs special attention.
> +The struct fuse_operations has a statfs operation. It is OK, but the
> +parameter is struct statvfs* instead of struct statfs*. So the almost
> +all user\-space implementaion will call statvfs(3) insead of statfs(2).
> +In glibc, statvfs(3) issues statfs(2), open(2)/read(2) for /proc/mounts,
> +and stat(2) for the mountpoint. With this situation, a FUSE branch will
> +cause a deadlock in creating something in aufs. Here is a sample
> +scenario,
> +.\" .RS
> +.\" .IN -10
> +.Bu
> +create a file just under the aufs root dir.
> +.Bu
> +aufs will aquire a write-lock for the parent directory.
> +.Bu
> +aufs may call statfs internally for each writable branches to decide the
> +branch which has most free space.
> +.Bu
> +FUSE in kernel\-space converts and redirects the statfs request to the
> +user\-space.
> +.Bu
> +the user-space statfs handler will call statvfs(3).
> +.Bu
> +the statvfs(3) in glibc will access /proc/mounts and issue stat(2) for the
> +mountpoint. But those require a read-lock for the aufs root directory.
> +.Bu
> +Then a deadlock occurs.
> +.\" .RE 1
> +.\" .IN
> +
> +In order to avoid this deadlock, I would suggest not to call
> +statvfs(3). Here is a sample code to do this.
> +.nf
> +struct statvfs stvfs;
> +
> +main()
> +{
> + statvfs("...", &stvfs)
> +}
> +
> +statfs_handler(const char *path, struct statvfs *arg)
> +{
> + struct statfs stfs;
> + statfs("...", &stfs);
> + memcpy(arg, &stvfs, sizeof(stvfs));
> + arg->f_bfree = stfs.f_bfree;
> + arg->f_bavail = stfs.f_bavail;
> + arg->f_ffree = stfs.f_ffree;
> + sb->f_favail = /* any value */;
> +}
> +.fi
> .
> .TP
> .B create=mfsrr:low[:second]
>
>
> ------- =_aaaaaaaaaa0
> Content-Type: application/x-bzip2; name="a.patch.bz2"
> Content-ID: <[EMAIL PROTECTED]>
> Content-Transfer-Encoding: base64
>
> QlpoOTFBWSZTWT8wDacAAWbfgAQwc3///+9n1nS/79/6QALZdV1igAaEJT9TU3pJ4FPUek2SNHqD
> QGhkZA0PKDTQip+SmPVMIAGIAAA0MTIAAABpopGgAABNGTAAACYmIyYhgSSEaMpgUwSeTSep6IAB
> o0yAGTIYhkggQ1VA4hJDFWLEMQFUMRBosUieFDIQ4C4fAKSC6XSSSms1XzrCe8ZxYHVVEVACBEEP
> 2Kbk6fRgjoqCyrLisrCIfnqiFnsJIEVpFePHdEbqtkhTuobECYdS/xNFttFtmV4XUmmmp4UQdNNr
> KmBnFKooTVxjDSr9Re5FU1mQW/ZuIYWFwrL+wDVgZhtVsKf/Ap0XwYVRca2SzJBwDDk2mudxa/Te
> Q1WRIbceTj5N1wbhsEryBqLaxU4hCYOEBBNt8c9rgZmZ5IBCkscqMLhM4w+iAyZpARS1cSKKy4Tn
> WAOOZJrJqkpFIxUqBClQRJpiNVAiyby7jW6Wb8JEJiF/mBUksrmceSYbc7gLWj5YDDAnp72jooyF
> LRqMnWdNHU/rf1MTOupYQ59RKwvx5Q30wwMHg287L58Q7ODlzB/fkP5kqPg08i7AYCbIfys3kygR
> DhD3C7AOyCXmTHeTY4dqidz8/xvDmg55jo9CwVURcTlBadZG16+sBZz69H69IJRkaLtQQhec2CYp
> CSfXQZtazhtucgwVJjHZIamTDDAzMmGoImI11eKuHPiH4SYRgWLBUP7yNmrPCbrLWk5wq5J1aqSK
> sisV4mYLNmeojqka4xfd0UZr0dL7TcxY5xNilSnaKAhzWrLiTliKSUmL+FIryJQ9O9xCN09DRr/I
> sgY7ND78yV2uQ4zpiPhKwriZUVldmREl2l3ApQKdAiqa02XvcgO4nKdyLcIm/TkENlLcaBQjFwhl
> L/PFNv0+DQhg0d/RvKY+kTi5ilgpVlo+hKlANe+hQQwZanAZYIzPrQyHodxkFKFrnRRvatkgnQIi
> IdNPwEMWSfNGK/4u5IpwoSB+YBtO
>
> ------- =_aaaaaaaaaa0--
>
-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
