[Intel-gfx] [PATCH] drm/i915: Remove pre-production pooled-EU w/a for Broxton

[Chris Wilson]

Quoting Jani Nikula (2017-11-17 07:44:08)
> On Thu, 16 Nov 2017, Rodrigo Vivi <rodrigo.vivi at intel.com> wrote:
> > On Thu, Nov 16, 2017 at 09:22:23AM +0000, Jani Nikula wrote:
> >> On Wed, 15 Nov 2017, Chris Wilson <chris at chris-wilson.co.uk> wrote:
> >> > Quoting David Weinehall (2017-11-15 18:01:41)
> >> >> On Tue, Nov 14, 2017 at 01:51:16PM +0000, Chris Wilson wrote:
> >> >> > WaEnablePooledEuFor2x6 only applies to preproduction models, unsupported
> >> >> > since commit 0102ba1fd8af ("drm/i915: Add early BXT sdv to the list of
> >> >> > preproduction machines").
> >> >> > 
> >> >> > Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> >> >> > Cc: Jani Nikula <jani.nikula at intel.com>
> >> >> 
> >> >> Reviewed-by: David Weinehall <david.weinehall at linux.intel.com>
> >> >
> >> > Ta, hope your sdp is ok in the bin ;)
> >> >
> >> > Any volunteers for IS_GLK_REVID, IS_KBL_REVID ?
> >> >
> >> > I see one IS_SKL_REVID that can easily be removed.
> >> 
> >> Need to know what the first shipped rev was. For example with bxt we
> >> apparently can't remove some of the rev checks.
> >> 
> >> Seems like we should add more checks to intel_detect_preproduction_hw()
> >> too.
> >
> > With the IS_*_REVID in place I'd advocate to leave the workarounds around.
> >
> > I understand that codes get cleaner if we remove them from here,
> > but usually previous pre-prod workarounds are useful reference for
> > new platforms power-on...
> 
> Perhaps for the gen-1, but not older.

That seems like a reasonably policy.

diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index 57dfaf04d819..0be79cf993fa 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -833,6 +833,11 @@ static void i915_workqueues_cleanup(struct drm_i915_private *dev_priv)
  * We don't keep the workarounds for pre-production hardware, so we expect our
  * driver to fail on these machines in one way or another. A little warning on
  * dmesg may help both the user and the bug triagers.
+ *
+ * Our policy for removing workarounds is to keep the current gen workarounds
+ * as a guide to the bring-up of the next gen (workarounds have a habit of
+ * persisting). Anything older than that should be removed along with the
+ * complications they introduce.
  */
 static void intel_detect_preproduction_hw(struct drm_i915_private *dev_priv)
 {