XFS corruption on move from xscale to i686

XFS corruption on move from xscale to i686

[Yura Pakhuchiy]

Hi,

I'm creadted XFS volume on 2.6.10 linux xscale/iq31244 box, then I
copyied files on it and moved this hard drive to i686 machine. When I
mounted it on i686, I found no files on it. I runned xfs_check, here is
output:

pc299:/home/yura# xfs_check /dev/sda5
dir 128 size is 31, should be 13
dir 128 offsets too high
root directory 128 has .. 0
dir 131 entry  bad inode number 4030121216
dir 131 entry                                      SRPMS.base\x02 bad offset 0
dir 131 bad size in entry at 68
link count mismatch for inode 0 (name ?), nlink 0, counted 3
link count mismatch for inode 6303282 (name ?), nlink 0, counted 1
link count mismatch for inode 128 (name ?), nlink 3, counted 1
link count mismatch for inode 131 (name ?), nlink 5, counted 4
link count mismatch for inode 33554560 (name ?), nlink 2, counted 1
link count mismatch for inode 67108992 (name ?), nlink 2, counted 1
link count mismatch for inode 100663424 (name ?), nlink 2, counted 1

-- 
Best regards,
        Yura

Re: XFS corruption on move from xscale to i686

[Nathan Scott]

On Thu, Jul 07, 2005 at 08:15:52PM +0300, Yura Pakhuchiy wrote:
> Hi,
> 
> I'm creadted XFS volume on 2.6.10 linux xscale/iq31244 box, then I
> copyied files on it and moved this hard drive to i686 machine. When I
> mounted it on i686, I found no files on it. I runned xfs_check, here is
> output:

Someone else was doing this awhile back, and also had issues.
Their trouble seemed to be related to xscale gcc miscompiling
parts of XFS - search the linux-xfs archives for details.

cheers.

-- 
Nathan

Re: XFS corruption on move from xscale to i686

[Yura Pakhuchiy]

2005/7/8, Nathan Scott <nathans@sgi.com>:
> On Thu, Jul 07, 2005 at 08:15:52PM +0300, Yura Pakhuchiy wrote:
> > Hi,
> >
> > I'm creadted XFS volume on 2.6.10 linux xscale/iq31244 box, then I
> > copyied files on it and moved this hard drive to i686 machine. When I
> > mounted it on i686, I found no files on it. I runned xfs_check, here is
> > output:
> 
> Someone else was doing this awhile back, and also had issues.
> Their trouble seemed to be related to xscale gcc miscompiling
> parts of XFS - search the linux-xfs archives for details.

I found patch by Greg Ungreger to fix this problem, but why it's still
not in mainline? Or it's a gcc problem and should be fixed by gcc folks?
BTW, my kernel on xscale is compiled using gcc 3.4.3.

Thanks,
        Yura

Re: XFS corruption on move from xscale to i686

[Nathan Scott]

On Wed, Jul 13, 2005 at 06:22:28PM +0300, Yura Pakhuchiy wrote:
> I found patch by Greg Ungreger to fix this problem, but why it's still
> not in mainline? Or it's a gcc problem and should be fixed by gcc folks?

Yes, IIRC the patch was incorrect for other platforms, and it sure
looked like an arm-specific gcc problem (this was ages back, so
perhaps its fixed by now).

cheers.

-- 
Nathan

Re: XFS corruption on move from xscale to i686

[Yura Pakhuchiy]

2005/7/14, Nathan Scott <nathans@sgi.com>:
> On Wed, Jul 13, 2005 at 06:22:28PM +0300, Yura Pakhuchiy wrote:
> > I found patch by Greg Ungreger to fix this problem, but why it's still
> > not in mainline? Or it's a gcc problem and should be fixed by gcc folks?
> 
> Yes, IIRC the patch was incorrect for other platforms, and it sure
> looked like an arm-specific gcc problem (this was ages back, so
> perhaps its fixed by now).

AFAIR gcc-3.4.3 was released after this conversation take place at linux-xfs,
maybe add something like this:

#ifdef XSCALE
    /* We need this because some gcc versions for xscale are broken. */
    [patched version here]
#else
    [original version here]
#endif

Best regards,
        Yura

Re: XFS corruption on move from xscale to i686

[Christoph Hellwig]

On Thu, Jul 14, 2005 at 04:50:01PM +0300, Yura Pakhuchiy wrote:
> 2005/7/14, Nathan Scott <nathans@sgi.com>:
> > On Wed, Jul 13, 2005 at 06:22:28PM +0300, Yura Pakhuchiy wrote:
> > > I found patch by Greg Ungreger to fix this problem, but why it's still
> > > not in mainline? Or it's a gcc problem and should be fixed by gcc folks?
> > 
> > Yes, IIRC the patch was incorrect for other platforms, and it sure
> > looked like an arm-specific gcc problem (this was ages back, so
> > perhaps its fixed by now).
> 
> AFAIR gcc-3.4.3 was released after this conversation take place at linux-xfs,
> maybe add something like this:
> 
> #ifdef XSCALE
>     /* We need this because some gcc versions for xscale are broken. */
>     [patched version here]
> #else
>     [original version here]
> #endif

no, just fix your compiler or let the gcc folks do it.  Did anyone of
the arm folks ever open a PR at the gcc bugzilla with a reproduced
testcase?  You're never get your compiler fixed with that attitude.

Re: XFS corruption on move from xscale to i686

[Yura Pakhuchiy]

2005/7/14, Christoph Hellwig <hch@infradead.org>:
> On Thu, Jul 14, 2005 at 04:50:01PM +0300, Yura Pakhuchiy wrote:
> > 2005/7/14, Nathan Scott <nathans@sgi.com>:
> > > On Wed, Jul 13, 2005 at 06:22:28PM +0300, Yura Pakhuchiy wrote:
> > > > I found patch by Greg Ungreger to fix this problem, but why it's still
> > > > not in mainline? Or it's a gcc problem and should be fixed by gcc folks?
> > >
> > > Yes, IIRC the patch was incorrect for other platforms, and it sure
> > > looked like an arm-specific gcc problem (this was ages back, so
> > > perhaps its fixed by now).
> >
> > AFAIR gcc-3.4.3 was released after this conversation take place at linux-xfs,
> > maybe add something like this:
> >
> > #ifdef XSCALE
> >     /* We need this because some gcc versions for xscale are broken. */
> >     [patched version here]
> > #else
> >     [original version here]
> > #endif
> 
> no, just fix your compiler or let the gcc folks do it.  Did anyone of
> the arm folks ever open a PR at the gcc bugzilla with a reproduced
> testcase?  You're never get your compiler fixed with that attitude.

Yes, but a lof of people use older versions of compilers and suffer
from this bug.
I personally was very unhappy when lost my data.

Best regards,
        Yura

Re: XFS corruption on move from xscale to i686

[Christoph Hellwig]

On Thu, Jul 14, 2005 at 05:45:15PM +0300, Yura Pakhuchiy wrote:
> Yes, but a lof of people use older versions of compilers and suffer
> from this bug.
> I personally was very unhappy when lost my data.

then host the patch somewhere and make sure to apply it.