public inbox for intel-gfx@lists.freedesktop.org
 help / color / mirror / Atom feed
From: Imre Deak <imre.deak@intel.com>
To: intel-gfx@lists.freedesktop.org
Cc: Eero Tamminen <eero.t.tamminen@intel.com>,
	Valtteri Rantala <valtteri.rantala@intel.com>,
	Michael T Frederick <michael.t.frederick@intel.com>
Subject: [PATCH v2 2/2] drm/i915/bxt: Fix inadvertent CPU snooping due to incorrect MOCS config
Date: Tue, 26 Apr 2016 15:44:22 +0300	[thread overview]
Message-ID: <1461674662-29353-2-git-send-email-imre.deak@intel.com> (raw)
In-Reply-To: <1461674662-29353-1-git-send-email-imre.deak@intel.com>

Setting a write-back cache policy in the MOCS entry definition also
implies snooping, which has a considerable overhead. This is
unexpected for a few reasons:
- From user-space's point of view since it didn't want a coherent
  surface (it didn't set the buffer as such via the set caching IOCTL).
- There is a separate MOCS entry field for snooping (which we never
  set).
- This MOCS table is about caching in (e)LLC and there is no (e)LLC on
  BXT. There is a separate table for L3 cache control.

Considering the above the current behavior of snooping looks like an
unintentional side-effect of the WB setting. Changing it to be PTE
based cacheability gets rid of the snooping without any ill-effects.
For a coherent surface the application would use a separate MOCS entry
(at index 1) and call the set caching IOCTL to setup the PTE entries
for the corresponding buffer to be snooped.

This resulted in 70% improvement in synthetic texturing benchmarks.

Kudos to Valtteri Rantala, Eero Tamminen and Michael T Frederick and
Ville who helped to narrow the source of problem to the kernel and to
the snooping behaviour in particular.

With a follow-up change to adjust the 3rd entry value
igt/gem_mocs_settings is passing after this change.

v2:
- Rebase on v2 of patch 1/2.

CC: Valtteri Rantala <valtteri.rantala@intel.com>
CC: Eero Tamminen <eero.t.tamminen@intel.com>
CC: Michael T Frederick <michael.t.frederick@intel.com>
CC: Ville Syrjälä <ville.syrjala@linux.intel.com>
CC: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Imre Deak <imre.deak@intel.com>
---
 drivers/gpu/drm/i915/intel_mocs.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/i915/intel_mocs.c b/drivers/gpu/drm/i915/intel_mocs.c
index f391ad6..d38f65e 100644
--- a/drivers/gpu/drm/i915/intel_mocs.c
+++ b/drivers/gpu/drm/i915/intel_mocs.c
@@ -149,8 +149,8 @@ static const struct drm_i915_mocs_entry broxton_mocs_table[] = {
 	  .l3cc_value =    L3_ESC(0) | L3_SCC(0) | L3_CACHEABILITY(L3_WB),
 	},
 	{
-	  /* 0x0000003b */
-	  .control_value = LE_CACHEABILITY(LE_WB) |
+	  /* 0x00000038 */
+	  .control_value = LE_CACHEABILITY(LE_PAGETABLE) |
 			   LE_TGT_CACHE(LE_TC_LLC_ELLC) |
 			   LE_LRUM(3) | LE_AOM(0) | LE_RSC(0) | LE_SCC(0) |
 			   LE_PFM(0) | LE_SCF(0),
-- 
2.5.0

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

  reply	other threads:[~2016-04-26 12:44 UTC|newest]

Thread overview: 32+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-04-26 12:44 [PATCH v2 1/2] drm/i915/gen9: Clean up MOCS table definitions Imre Deak
2016-04-26 12:44 ` Imre Deak [this message]
2016-04-26 12:57   ` [PATCH v2 2/2] drm/i915/bxt: Fix inadvertent CPU snooping due to incorrect MOCS config Chris Wilson
2016-04-26 13:17     ` Imre Deak
2016-04-26 13:23       ` Chris Wilson
2016-04-26 13:43         ` Imre Deak
2016-04-26 13:58           ` Chris Wilson
2016-04-26 14:26         ` Eero Tamminen
2016-04-26 14:30           ` Daniel Vetter
2016-04-26 17:18             ` Eero Tamminen
2016-04-26 17:25               ` Frederick, Michael T
2016-04-27 13:25                 ` Eero Tamminen
2016-04-27 14:53                   ` Chris Wilson
2016-04-27 18:42                     ` Dave Gordon
2016-04-29  8:01                     ` Eero Tamminen
2016-04-26 17:57             ` Ville Syrjälä
2016-04-28  8:13               ` Daniel Vetter
2016-04-28 10:48                 ` Ville Syrjälä
2016-04-28 14:44                   ` Daniel Vetter
2016-04-28 17:21                     ` Ville Syrjälä
2016-04-26 14:42           ` Chris Wilson
2016-04-26 16:01             ` Imre Deak
2016-04-28  8:17               ` Daniel Vetter
2016-04-28  8:38                 ` Imre Deak
2016-04-28 14:48                   ` Daniel Vetter
2016-04-28 17:15                     ` Imre Deak
2016-05-02  8:28                       ` Daniel Vetter
2016-05-02 11:18                         ` Ville Syrjälä
2016-05-02 13:50                         ` Imre Deak
2016-04-28 17:25                     ` Ville Syrjälä
2016-04-26 13:12   ` Chris Wilson
2016-04-26 16:55 ` ✗ Fi.CI.BAT: failure for series starting with [v2,1/2] drm/i915/gen9: Clean up MOCS table definitions Patchwork

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1461674662-29353-2-git-send-email-imre.deak@intel.com \
    --to=imre.deak@intel.com \
    --cc=eero.t.tamminen@intel.com \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=michael.t.frederick@intel.com \
    --cc=valtteri.rantala@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox