xfs
[Top] [All Lists]

Re: 2.6.30-rc6: BUG at fs/xfs/support/debug.c:109!

To: Alexander Beregalov <a.beregalov@xxxxxxxxx>
Subject: Re: 2.6.30-rc6: BUG at fs/xfs/support/debug.c:109!
From: Felix Blyakher <felixb@xxxxxxx>
Date: Mon, 1 Jun 2009 12:58:56 -0500
Cc: Kernel Testers List <kernel-testers@xxxxxxxxxxxxxxx>, xfs@xxxxxxxxxxx
In-reply-to: <a4423d670906011038qd809ebay22aebca5aaf271a4@xxxxxxxxxxxxxx>
References: <a4423d670906010822l6e8e112dl3d6f7de9e2179850@xxxxxxxxxxxxxx> <B5660F1A-7940-4E4A-92E8-8C0F93C274BD@xxxxxxx> <a4423d670906011038qd809ebay22aebca5aaf271a4@xxxxxxxxxxxxxx>

On Jun 1, 2009, at 12:38 PM, Alexander Beregalov wrote:

It was Gentoo's `emerge --metadata`. It reads many small files (ebuilds).
I am sure it cannot be easily reproduced.

Right. Otherwise many more people would've reported it.

It runs fine everyday. I do not have a testcase.
I can try to run it forever in loop if you need.

Yes, that will be good.
The problem is that the traces are confusing. From one
side readdir stack makes sense based on your description
of the load, but OTOH the panic is from xfs_bmapi(), which
doesn't fit in that backtrace at all. It seems like
backtrace is from the concurrently running different thread.
Don't have good idea for this bug atm.

Felix

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