Re: [PATCH] drm/i915: Add OACONTROL to the command parser register whitelist.

From: Kenneth Graunke <kenneth@whitecape.org>
To: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: "intel-gfx@lists.freedesktop.org" <intel-gfx@lists.freedesktop.org>
Subject: Re: [PATCH] drm/i915: Add OACONTROL to the command parser register whitelist.
Date: Thu, 27 Mar 2014 14:34:41 -0700	[thread overview]
Message-ID: <53349971.3000806@whitecape.org> (raw)
In-Reply-To: <CAKMK7uGiHGx=PDAoiLFidVr34=FZ2-p1NTh5y-99TXR+dVp42w@mail.gmail.com>

[-- Attachment #1.1: Type: text/plain, Size: 1623 bytes --]

On 03/27/2014 01:16 PM, Daniel Vetter wrote:
> On Thu, Mar 27, 2014 at 4:57 PM, Volkin, Bradley D
> <bradley.d.volkin@intel.com> wrote:
>> On Thu, Mar 27, 2014 at 12:57:21AM -0700, Daniel Vetter wrote:
>>> Another one that blows is igt/gen7_forcewake_mt. Not sure yet whether it's
>>> an issue with the test or the checker:
>>>
>>> https://bugs.freedesktop.org/show_bug.cgi?id=76670
>>
>> For this one, the parser rejects an MI_STORE_REGISTER_MEM with the GGTT bit
>> set. We don't currently allow that, even from master. It sounds like there
>> might be released versions of the ddx that do this as well. If that's the
>> case, or if there are other situations where tests, etc rely on being able
>> to do whatever they want when setting the I915_DISPATCH_SECURE flag, then I
>> think we might as well stop parsing secure batches and let them go through
>> as before.
> 
> Well for the testcase I think we can just add the missing flag. If
> there's indeed shipping userspace out there which is getting these
> flags wrong then I think we need to silently upgrade them when copying
> the cmds over to the 2nd batch. But I guess until that need is really
> established we can hope we don't need this.
> -Daniel

Why are we parsing batches with I915_EXEC_SECURE at all?  Secure batches
are only issued from trusted code which is guaranteed to be running as
root.  I don't see any benefit to scanning those batches, and there's
definitely overhead.

I mean, sure, it may be reasonable in the short term as a way to test
the command parser, but I certainly hope we don't *ship* that.

--Ken

[-- Attachment #1.2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

[-- Attachment #2: Type: text/plain, Size: 159 bytes --]

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx