Re: [Exim] check_spool_inodes / reiserfs / Linux 2.6 bug

Page principale
Supprimer ce message
Répondre à ce message
Auteur: Andreas Metzler
Date:  
À: exim-users
Sujet: Re: [Exim] check_spool_inodes / reiserfs / Linux 2.6 bug
On 2004-02-24 Nigel Metheringham <Nigel.Metheringham@???> wrote:
> On Tue, 2004-02-24 at 16:01, Andreas Metzler wrote:
> > I can confirm the kernel-side. With Linux 2.4 both f_ffree and f_files
> > are reported as -1 for reiserfs, while they are reported as 0 on Linux
> > 2.6. (I have checked with filesystems in reiserfs-3.5 and
> > reiserfs-3.6.) It might be interesting to know what other tree-based
> > filesystems (XFS, JFS) return.


> Thats obviously not the case on all 2.4 kernels:-
>         statfs("/", {f_type="REISERFS_SUPER_MAGIC", f_bsize=4096,
>         f_blocks=17492465, f_bfree=16922849, f_bavail=16922849,
>         f_files=4294967295, f_ffree=4294967295, f_fsid={0, 0},
>         f_namelen=255, f_frsize=0}) = 0

[...]

D'oh. 4294967295 is what you get (instead of -1) if you interpret the
value correctly as unsigned long, which is obviously correct (types.h,
exim, and df -i agree). Thanks for the correction.

BTW if my memory does not trick me completely this is the second time
reiserfs has switched, it started with returning zero (yielding an
incompatibility with inn of this time) before it switched to returning
the big number.
               cu andreas


PS: printf '%lu\n' -1
--
"See, I told you they'd listen to Reason," [SPOILER] Svfurlr fnlf,
fuhggvat qbja gur juveyvat tha.
Neal Stephenson in "Snow Crash"