Re: [ 04/16] drm/i915: correctly order the ring init sequence

From: Daniel Vetter <daniel.vetter@ffwll.ch>
To: Herton Ronaldo Krzesinski <herton.krzesinski@canonical.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	linux-kernel@vger.kernel.org, stable@vger.kernel.org,
	torvalds@linux-foundation.org, akpm@linux-foundation.org,
	alan@lxorguk.ukuu.org.uk, Jani Nikula <jani.nikula@intel.com>,
	Yang Guang <guang.a.yang@intel.com>
Subject: Re: [ 04/16] drm/i915: correctly order the ring init sequence
Date: Tue, 21 Aug 2012 08:42:35 +0200	[thread overview]
Message-ID: <CAKMK7uFpj+qctqKOk6PJqRn_7XAiDbE7Bv2r2VseGyUHUTD2mA@mail.gmail.com> (raw)
In-Reply-To: <20120821051318.GB3625@herton-Z68MA-D2H-B3>

On Tue, Aug 21, 2012 at 7:13 AM, Herton Ronaldo Krzesinski
<herton.krzesinski@canonical.com> wrote:
> On Sun, Aug 19, 2012 at 08:56:03PM -0700, Greg Kroah-Hartman wrote:
>> From: Greg KH <gregkh@linuxfoundation.org>
>>
>> 3.0-stable review patch.  If anyone has any objections, please let me know.
>>
>> ------------------
>>
>> From: Daniel Vetter <daniel.vetter@ffwll.ch>
>>
>> commit 0d8957c8a90bbb5d34fab9a304459448a5131e06 upstream.
>>
>> We may only start to set up the new register values after having
>> confirmed that the ring is truely off. Otherwise the hw might lose the
>> newly written register values. This is caught later on in the init
>> sequence, when we check whether the register writes have stuck.
>>
>> Reviewed-by: Jani Nikula <jani.nikula@intel.com>
>> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=50522
>> Tested-by: Yang Guang <guang.a.yang@intel.com>
>> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
>> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> [...]
>
> I had the same problem as on 3.2 with this change, i915 stopped working
> unable to initialize render ring, eg. on one of the boots here:
> [drm:init_ring_common] *ERROR* render ring initialization failed ctl 0001f003 head 00001020 tail 00000000 start 00001000
>
> But unlike I was expecting as with 3.2 case, picking commit
> f01db988ef6f6c70a6cc36ee71e4a98a68901229 ("drm/i915: Add wait_for in
> init_ring_common") here isn't enough, it continues to fail even if I
> try to increase the delay in the wait_for, I'm not sure why yet... may
> be something else is going on, or 3.0 has something else missing.
>
> Also the same proposed patch for 3.4.10 gives the same problem, but
> picking f01db988ef6f6c70a6cc36ee71e4a98a68901229 there made things work
> again like happend on first 3.2.28 proposed update. Only 3.0
> is misteriously failing either way here.

I guess we're missing something then still in the stable backports for
3.0. Herton, what machine do you have exaclty (lspci -nn)?

Greg, I think for now it's better if you hold off on merging this
patch to 3.0 until this is sorted out.

Thanks, Daniel
-- 
Daniel Vetter
daniel.vetter@ffwll.ch - +41 (0) 79 365 57 48 - http://blog.ffwll.ch