xfs
[Top] [All Lists]

Re: Failure of test case : 250

To: Amit Sahrawat <amit.sahrawat83@xxxxxxxxx>
Subject: Re: Failure of test case : 250
From: Dave Chinner <david@xxxxxxxxxxxxx>
Date: Sat, 25 Jun 2011 10:09:24 +1000
Cc: xfs@xxxxxxxxxxx
In-reply-to: <BANLkTinCLAb3X59J=SnXgEkeRCLgB6FAGg@xxxxxxxxxxxxxx>
References: <BANLkTinCLAb3X59J=SnXgEkeRCLgB6FAGg@xxxxxxxxxxxxxx>
User-agent: Mutt/1.5.20 (2009-06-14)
On Fri, Jun 24, 2011 at 11:49:45AM +0530, Amit Sahrawat wrote:
> Dear All,
> 
> Test Case : 250 results in failure on x86(2.6.31.y) and ARM(2.6.35.13..)

The XFS_WANT_CORRUPTED_GOTO bug wasn't fixed until 3.0-rc1 (bf59170
xfs: obey minleft values during extent allocation correctly), so of
course you'll get that failure on .31 and .35 kernels.

Amit, I'd suggest that you try to reproduce the problems you have on
a top of tree kernel first before reporting them. Most of us XFS
developers don't really have time to triage or solve failures on old
kernels like 2.6.31 or 2.6.35 and so are ignoring your bug reports.
Reporting bugs on old kernels that have since been fixed is really
just a waste of our (limited) time. If you find a bug that isn't
fixed on TOT, then we'll triage it and get it fixed in TOT.

However, if you can't reproduce them on top of tree, then you need
to go an find the commit (kernel, xfsprogs or xfstests) yourself
that makes the test pass and back port it to your tree. Same for if
we fix a bug you report in TOT - we're not going to backport the
fixes to your trees for you. We'll review backported patches if you
post them to the list, but we don't have time (or motivation) to do
the backports and QA them for you.

And if you are backporting to a long term stable tree, feel free to
submit the tested backports to stable@xxxxxxxxxx (cc'ing the xfs
list so they can be reviewed, of course) so that everyone else can
also benefit from your work....

Cheers,

Dave.
-- 
Dave Chinner
david@xxxxxxxxxxxxx

<Prev in Thread] Current Thread [Next in Thread>