Re: [PATCH 1/3] drm: Widen vblank count to 64 bits. Change vblank time precision to ns

From: Keith Packard <keithp@keithp.com>
To: "Michel Dänzer" <michel@daenzer.net>,
	"Dave Airlie" <airlied@redhat.com>,
	"Daniel Vetter" <daniel@ffwll.ch>
Cc: linux-kernel@vger.kernel.org, dri-devel@lists.freedesktop.org
Subject: Re: [PATCH 1/3] drm: Widen vblank count to 64 bits. Change vblank time precision to ns
Date: Thu, 06 Jul 2017 08:04:47 -0700	[thread overview]
Message-ID: <86o9sxd2sw.fsf@keithp.com> (raw)
In-Reply-To: <83a2dbe9-dfcb-e31e-fc7b-14c4364daea5@daenzer.net>

[-- Attachment #1: Type: text/plain, Size: 1357 bytes --]

Michel Dänzer <michel@daenzer.net> writes:

> Subtle breakage here: vblwait->request.sequence must still get updated
> for _DRM_VBLANK_RELATIVE, in case we're interrupted by a signal.

Thanks for finding this.

I think it might be better to just not modify the request.type field
instead, so that on re-entry it gets recomputed? That would mean that a
signal might cause the value to be different if the application takes a
long time processing the signal, but I'm not sure that's wrong?

>> @@ -317,6 +317,9 @@ int via_driver_irq_postinstall(struct drm_device *dev)
>>  	if (!dev_priv)
>>  		return -EINVAL;
>>  
>> +	if (dev->driver->get_vblank_counter)
>> +		dev->max_vblank_count = 0xffffffff;
>
> What's the purpose of this? All drivers providing get_vblank_counter
> should already initialize max_vblank_count correctly.

Yeah, I couldn't prove that this driver did that, and as Daniel says, we
haven't ever audited the drivers to make sure they do.

We have a check to see that they don't set max_vblank_count if they
don't provide a get function, but I can't find the matching check for
drivers that do provide a function and aren't setting max_vblank_count.

Do you have any thoughts on the wisdom of changing this API before we
have a driver that needs it?

And, of course, thanks for your review!

-- 
-keith

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 832 bytes --]