Re: [PATCH] drm/i915: Remove support for legacy debugfs crc interface

From: "Ville Syrjälä" <ville.syrjala@linux.intel.com>
To: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Jani Nikula <jani.nikula@intel.com>,
	Tomi Sarvela <tomi.p.sarvela@intel.com>,
	intel-gfx@lists.freedesktop.org
Subject: Re: [PATCH] drm/i915: Remove support for legacy debugfs crc interface
Date: Tue, 21 Nov 2017 16:07:43 +0200	[thread overview]
Message-ID: <20171121140743.GA10981@intel.com> (raw)
In-Reply-To: <f9e136d3-006b-ef34-4d9b-bb39944712c8@linux.intel.com>

On Tue, Nov 21, 2017 at 11:39:01AM +0100, Maarten Lankhorst wrote:
> Op 08-11-17 om 12:25 schreef Ville Syrjälä:
> > On Wed, Nov 08, 2017 at 11:40:19AM +0100, Maarten Lankhorst wrote:
> >> Op 02-11-17 om 17:11 schreef Ville Syrjälä:
> >>> On Thu, Nov 02, 2017 at 05:19:07PM +0200, Jani Nikula wrote:
> >>>> On Thu, 02 Nov 2017, Ville Syrjälä <ville.syrjala@linux.intel.com> wrote:
> >>>>> On Thu, Nov 02, 2017 at 02:56:51PM +0100, Maarten Lankhorst wrote:
> >>>>>> This interface is deprecated, and has been replaced by the upstream
> >>>>>> drm crc interface.
> >>>>> Before we nuke this I would like to see an option in the new interface
> >>>>> to not filter out the "bad" CRCs. When analyzing how the hardware
> >>>>> behaves seeing every CRC can be valuable. And I'm not at all convinced
> >>>>> we should be dropping as many CRCs as we are currently.
> >>>> I'm not against it, but do you have a concrete proposal on how that
> >>>> option would look like?
> >>> Some kind of of filter_bad_crcs file with a bool value perhaps?
> >> You can set sources, might as well add a nofilter option.. But I don't see what it has to do
> >> with this patch? This problem existed since before the api was introduced.. Only difference
> >> is kernel eats possibly corrupt CRCs now instead of IGT.
> > I don't use igt for this.
> >
> If it's not in IGT then I'm not sure we should hold upthis patch for it tbh. You can
> always change skipped = 0 to skipped = 2 for the new debugfs interface to find bugs
> with garbage CRC values, or add a flag to pipe source parsing for not skipping
> garbage CRC's.

Why do you want to intentionally generate more pointless work for me?

If you don't want to fix the new interface to have feature parity
with the old one then just don't remove the old one. It's not
hurting anyone AFAICS.

-- 
Ville Syrjälä
Intel OTC
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx