Re: [PATCH] drm/i915: Record error batch buffers using iomem

From: Chris Wilson <chris@chris-wilson.co.uk>
To: Jaswinder Singh Rajput <jaswinderlinux@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>,
	dri-devel@lists.freedesktop.org, Dave Airlie <airlied@linux.ie>,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH] drm/i915: Record error batch buffers using iomem
Date: Wed, 12 May 2010 14:50:50 +0100	[thread overview]
Message-ID: <89k83a$87e1b2@azsmga001.ch.intel.com> (raw)
In-Reply-To: <AANLkTilbqi5crUD8KXmZ5SIv92KioEAN7Ra6nk2ALU6F@mail.gmail.com>

On Wed, 12 May 2010 18:45:55 +0530, Jaswinder Singh Rajput <jaswinderlinux@gmail.com> wrote:
> Hello Chris,
> 
> With this patch after XWindows freezes, I get :
[snip]
> freeze dmesg : http://userweb.kernel.org/~jaswinder/acer_netbook/dmesg_2634_chris_hang.txt
> 
> freeze Xorg.log : http://userweb.kernel.org/~jaswinder/acer_netbook/Xorg_log.txt

Jaswinder can you also upload the /sys/kernel/debug/dri/0/i915_error_state
following a freeze as well, please. If your /sys/kernel/debug is empty,
you will need to "mount -tdebugfs debug /sys/kernel/debug".

> So it means this patches shifted the BUG and warning messages after
> some time. So I can only work on XWindows for few minutes with this
> patch.
> Andrew patch is in linus git tree. Can you please update your patch
> above Andrew patch. So that I can do further testing.

What is in the tree is adequate for the time being. It will capture the
batch buffer into the error state. My follow-on patch only increases the
level of paranoia. Thanks for testing.

-- 
Chris Wilson, Intel Open Source Technology Centre