Hi everybody,
I have a question about a kernel message that appears
in my /var/log/messages:
----------
Oct 25 11:01:00 judicator kernel: blk: queue c03746a4, I/O limit 4095Mb
(mask 0xffffffff)
----------
I'm getting this since I installed Eric's RH7.1 XFS 2.4.9-6PR3 on my roswell
system,
recompiled (kgcc) from sources with lvm enabled.
It occurs several times per hour when I'm writing
to my xfs-lvm-stripeset, two maxtor 80gb drives,
but I did not notice any problems -
Is this message xfs related or lvm? or something completely different?
I'm on a Athlon system, VIA kt133 board, the harddrives are both attached to
a Promise PDC20268.
thx
-simon
----- Original Message -----
From: "Eric Sandeen" <sandeen@xxxxxxx>
To: "Simon Pabst" <simon.pabst@xxxxxxxxxxxxxxxxxxxxxxxx>
Sent: Thursday, October 25, 2001 4:16 PM
Subject: Re: For Testing - updated RH7.1 XFS 2.4.9-6 kernels
> Not sure about gcc... 3.0 maybe.
>
> I also don't know about the error message... perhaps this is one for the
> list?
>
> -Eric
>
> Simon Pabst wrote:
> >
> > Hi Eric,
> > thx again, ntfs compiles fine now. I don't need ntfs on my
> > server box, but compiled it to test if it works at all,
> > cause I'll need it on my desktop when the installer for 7.2
> > is released ;-)
> > What do you think is a _new_ version of gcc?
> > Doesn't roswell's gcc 2.96-96 do the trick?
> > BTW my lvm-xfs volume seems to work fine
> > I'm using two 80gb maxtor drives striped to a
> > raid0 - but I'm getting kernel messages in /var/log/messages
> > like this one:
> > ----
> > Oct 25 11:01:00 judicator kernel: blk: queue c03746a4, I/O limit 4095Mb
> > (mask 0xffffffff)
> > ----
> >
> > is this related to lvm/xfs? I searched through the list and the web and
> > found nothing appropriate.
> >
> > thx
> > -simon
> >
> > ----- Original Message -----
> > From: "Eric Sandeen" <sandeen@xxxxxxx>
> > To: "Simon Pabst" <simon.pabst@xxxxxxxxxxxxxxxxxxxxxxxx>
> > Sent: Thursday, October 25, 2001 4:23 AM
> > Subject: Re: For Testing - updated RH7.1 XFS 2.4.9-6 kernels
> >
> > > Hi Simon -
> > >
> > > Try changing the "..." on line 25 of fs/ntfs/support.h to "ARGS..." -
> > > i.e.
> > >
> > > #define ntfs_debug(mask, fmt, ARGS...) do {} while (0)
> > >
> > > Apparently _newer_ versions of gcc/cpp can handle it the way it is in
> > > the original code...
> > >
> > > FWIW, we never _touched_ this file. :)
> > >
> > > -Eric
> > >
> > > Simon Pabst wrote:
> > > >
> > > > Hi Eric,
> > > > just compiled a PR3-kernel, checked my old (reiser) lvm drives and
> > changed
> > > > them to xfs without even a hint of a problem.
> > > > Great work, thanks!
> > > > I'll let you now if anything goes wrong.
> > > > But I have another small problem - the NTFS module.
> > > > Once again one of those RH doesn't compile by default.
> > > > I get :
> > > > -----------------------------
> > > > ....
> > > >
> >
kgcc -D__KERNEL__ -I/usr/src/linux-2.4.9-6SGI_XFS_PR3/include -Wall -Wstri
> > > >
> >
ct-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common -Wno-unuse
> > > >
> >
d -fomit-frame-pointer -pipe -march=i686 -malign-functions=4 -DMODULE -DNT
> > > > FS_VERSION=\"1.1.18\" -c -o fs.o fs.c
> > > > In file included from fs.c:21:
> > > > support.h:25: badly punctuated parameter list in `#define'
> > > > fs.c: In function `ntfs_read':
> > > > fs.c:75: warning: implicit declaration of function `ntfs_debug'
> > > > make[2]: *** [fs.o] Error 1
> > > > make[2]: Leaving directory
`/usr/src/linux-2.4.9-6SGI_XFS_PR3/fs/ntfs'
> > > > make[1]: *** [_modsubdir_ntfs] Error 2
> > > > make[1]: Leaving directory `/usr/src/linux-2.4.9-6SGI_XFS_PR3/fs'
> > > > make: *** [_mod_fs] Error 2
> > > > ----------------------------------
> > > > (both kgcc and gcc)
> > > >
> > > > -simon
> > > >
>
|