linux-stable/drivers/gpu/drm
Francisco Jerez 4fc020d864 drm/i915: Remove WaDisableLSQCROPERFforOCL KBL workaround.
The WaDisableLSQCROPERFforOCL workaround has the side effect of
disabling an L3SQ optimization that has huge performance implications
and is unlikely to be necessary for the correct functioning of usual
graphic workloads.  Userspace is free to re-enable the workaround on
demand, and is generally in a better position to determine whether the
workaround is necessary than the DRM is (e.g. only during the
execution of compute kernels that rely on both L3 fences and HDC R/W
requests).

The same workaround seems to apply to BDW (at least to production
stepping G1) and SKL as well (the internal workaround database claims
that it does for all steppings, while the BSpec workaround table only
mentions pre-production steppings), but the DRM doesn't do anything
beyond whitelisting the L3SQCREG4 register so userspace can enable it
when it sees fit.  Do the same on KBL platforms.

Improves performance of the GFXBench4 gl_manhattan31 benchmark by 60%,
and gl_4 (AKA car chase) by 14% on a KBL GT2 running Mesa master --
This is followed by a regression of 35% and 10% respectively for the
same benchmarks and platform caused by my recent patch series
switching userspace to use the dataport constant cache instead of the
sampler to implement uniform pull constant loads, which caused us to
hit more heavily the L3 cache (and on platforms other than KBL had the
opposite effect of improving performance of the same two benchmarks).
The overall effect on KBL of this change combined with the recent
userspace change is respectively 4.6% and 2.6%.  SynMark2 OglShMapPcf
was affected by the constant cache changes (though it improved as it
did on other platforms rather than regressing), but is not
significantly affected by this patch (with statistical significance of
5% and sample size 20).

v2: Drop some more code to avoid unused variable warning.

Fixes: 738fa1b312 ("drm/i915/kbl: Add WaDisableLSQCROPERFforOCL")
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=99256
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Cc: Matthew Auld <matthew.william.auld@gmail.com>
Cc: Eero Tamminen <eero.t.tamminen@intel.com>
Cc: Jani Nikula <jani.nikula@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Cc: beignet@lists.freedesktop.org
Cc: <stable@vger.kernel.org> # v4.7+
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
[Removed double Fixes tag]
Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1484217894-20505-1-git-send-email-mika.kuoppala@intel.com
(cherry picked from commit 8726f2faa3)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
2017-01-18 16:31:11 +02:00
..
amd ktime: Cleanup ktime_set() usage 2016-12-25 17:21:22 +01:00
arc Merge remote-tracking branch 'airlied/drm-next' into drm-misc-next 2016-12-06 10:26:48 +01:00
arm Linux 4.9-rc8 2016-12-05 17:11:48 +10:00
armada mm: use vmf->address instead of of vmf->virtual_address 2016-12-14 16:04:09 -08:00
ast drivers/gpu/drm/ast: Fix infinite loop if read fails 2016-12-18 14:39:50 +01:00
atmel-hlcdc drm: define drm_compat_ioctl NULL on CONFIG_COMPAT=n and reduce #ifdefs 2016-11-02 11:33:47 -04:00
bochs drm: Make the connector .detect() callback optional 2016-12-01 10:05:53 -05:00
bridge drm/bridge: analogix: Don't return -EINVAL when panel doesn't support PSR 2016-12-07 08:27:36 +05:30
cirrus drm: Make the connector .detect() callback optional 2016-12-01 10:05:53 -05:00
etnaviv mm: use vmf->address instead of of vmf->virtual_address 2016-12-14 16:04:09 -08:00
exynos mm: use vmf->address instead of of vmf->virtual_address 2016-12-14 16:04:09 -08:00
fsl-dcu Merge tag 'drm-fsl-dcu-for-next' of http://git.agner.ch/git/linux-drm-fsl-dcu into drm-next 2016-12-09 12:43:17 +10:00
gma500 mm: use vmf->address instead of of vmf->virtual_address 2016-12-14 16:04:09 -08:00
hisilicon drm: Make the connector .detect() callback optional 2016-12-01 10:05:53 -05:00
i2c drm/i2c: tda998x: fix spelling mistake 2016-11-18 00:00:40 +00:00
i810 drm: define drm_compat_ioctl NULL on CONFIG_COMPAT=n and reduce #ifdefs 2016-11-02 11:33:47 -04:00
i915 drm/i915: Remove WaDisableLSQCROPERFforOCL KBL workaround. 2017-01-18 16:31:11 +02:00
imx Merge remote-tracking branch 'airlied/drm-next' into drm-misc-next 2016-12-06 10:26:48 +01:00
mediatek Merge remote-tracking branch 'airlied/drm-next' into drm-misc-next 2016-12-06 10:26:48 +01:00
meson drm: Add support for Amlogic Meson Graphic Controller 2016-12-01 10:07:20 +01:00
mga
mgag200 drm: Make the connector .detect() callback optional 2016-12-01 10:05:53 -05:00
msm mm: use vmf->address instead of of vmf->virtual_address 2016-12-14 16:04:09 -08:00
mxsfb drm: mxsfb: Select DRM_PANEL 2016-12-08 10:25:37 +10:00
nouveau ktime: Cleanup ktime_set() usage 2016-12-25 17:21:22 +01:00
omapdrm mm: use vmf->address instead of of vmf->virtual_address 2016-12-14 16:04:09 -08:00
panel drm/panel: simple: Add support for AUO G185HAN01 2016-12-06 17:06:32 +01:00
qxl Merge remote-tracking branch 'airlied/drm-next' into drm-misc-next 2016-12-06 10:26:48 +01:00
r128
radeon Merge branches 'acpica' and 'acpi-scan' 2016-12-22 14:34:24 +01:00
rcar-du drm: Make the connector .detect() callback optional 2016-12-01 10:05:53 -05:00
rockchip drm: Make the connector .detect() callback optional 2016-12-01 10:05:53 -05:00
savage drm: define drm_compat_ioctl NULL on CONFIG_COMPAT=n and reduce #ifdefs 2016-11-02 11:33:47 -04:00
shmobile drm: Make the connector .detect() callback optional 2016-12-01 10:05:53 -05:00
sis drm: define drm_compat_ioctl NULL on CONFIG_COMPAT=n and reduce #ifdefs 2016-11-02 11:33:47 -04:00
sti drm: Make the connector .detect() callback optional 2016-12-01 10:05:53 -05:00
sun4i Merge remote-tracking branch 'airlied/drm-next' into drm-misc-next 2016-12-06 10:26:48 +01:00
tdfx drm: define drm_compat_ioctl NULL on CONFIG_COMPAT=n and reduce #ifdefs 2016-11-02 11:33:47 -04:00
tegra mm: use vmf->address instead of of vmf->virtual_address 2016-12-14 16:04:09 -08:00
tilcdc ktime: Cleanup ktime_set() usage 2016-12-25 17:21:22 +01:00
ttm mm: use vmf->address instead of of vmf->virtual_address 2016-12-14 16:04:09 -08:00
udl mm: use vmf->address instead of of vmf->virtual_address 2016-12-14 16:04:09 -08:00
vc4 This pull request brings in VEC (TV-out) support for vc4, along with a 2016-12-13 12:05:12 +10:00
vgem mm: use vmf->address instead of of vmf->virtual_address 2016-12-14 16:04:09 -08:00
via drm: define drm_compat_ioctl NULL on CONFIG_COMPAT=n and reduce #ifdefs 2016-11-02 11:33:47 -04:00
virtio drm/virtio: annotate virtio_gpu_queue_ctrl_buffer_locked 2016-12-15 06:59:17 +02:00
vmwgfx drm/vmwgfx: Switch to mode_cmd2 2016-12-02 17:15:20 +01:00
zte drm: zte: checking for NULL instead of IS_ERR() 2016-11-15 11:00:42 +01:00
ati_pcigart.c
drm_agpsupport.c
drm_atomic.c This pull request brings in VEC (TV-out) support for vc4, along with a 2016-12-13 12:05:12 +10:00
drm_atomic_helper.c drm/atomic: doc: remove old comment about nonblocking commits 2016-12-06 16:28:30 -02:00
drm_auth.c
drm_blend.c
drm_bridge.c
drm_bufs.c
drm_cache.c
drm_color_mgmt.c drm/color: document NULL values and default settings better 2016-11-15 22:39:48 +01:00
drm_connector.c drm: document standard connector properties 2016-11-18 09:45:07 +01:00
drm_context.c
drm_crtc.c drm/fence: add drm_crtc_create_fence() 2016-12-07 11:16:55 -02:00
drm_crtc_helper.c
drm_crtc_helper_internal.h
drm_crtc_internal.h drm/fence: add drm_crtc_create_fence() 2016-12-07 11:16:55 -02:00
drm_debugfs.c drm/atomic: cleanup debugfs entries on un-registering the driver. 2016-11-21 13:22:08 -05:00
drm_debugfs_crc.c
drm_dma.c
drm_dp_aux_dev.c
drm_dp_dual_mode_helper.c
drm_dp_helper.c
drm_dp_mst_topology.c
drm_drv.c drm: Add fake controlD* symlinks for backwards compat 2016-12-10 22:46:19 +01:00
drm_dumb_buffers.c drm: Consolidate dumb buffer docs 2016-11-15 12:51:49 +01:00
drm_edid.c Merge tag 'drm-misc-next-2016-11-29' of git://anongit.freedesktop.org/git/drm-misc into drm-next 2016-11-30 14:28:20 +10:00
drm_edid_load.c
drm_encoder.c drm: Fix locking cargo-cult in encoder/plane init/cleanup 2016-11-29 23:34:36 +01:00
drm_encoder_slave.c
drm_fb_cma_helper.c drm/fb_cma_helper: Add drm_fb_cma_prepare_fb() helper 2016-11-14 12:43:58 +01:00
drm_fb_helper.c drm: Protect fb_helper list manipulation with a mutex 2016-11-29 15:52:32 -05:00
drm_flip_work.c
drm_fops.c drm: define drm_compat_ioctl NULL on CONFIG_COMPAT=n and reduce #ifdefs 2016-11-02 11:33:47 -04:00
drm_fourcc.c drm: move allocation out of drm_get_format_name() 2016-11-12 14:19:38 +01:00
drm_framebuffer.c drm: Nuke modifier[1-3] 2016-11-17 11:58:09 +01:00
drm_gem.c
drm_gem_cma_helper.c
drm_global.c
drm_hashtab.c
drm_info.c
drm_internal.h drm: drm_irq.h header cleanup 2016-11-15 23:33:48 +01:00
drm_ioc32.c
drm_ioctl.c drm: Allow CAP_PRIME on !MODESET 2016-12-08 10:00:15 +01:00
drm_irq.c drm/irq: Unexport drm_vblank_on/off 2016-11-15 23:33:48 +01:00
drm_kms_helper_common.c
drm_legacy.h
drm_lock.c drm: Avoid NULL dereference for DRM_LEGACY debug message 2016-11-28 08:39:41 +01:00
drm_memory.c
drm_mipi_dsi.c
drm_mm.c drm: Initialise drm_mm.head_node.allocated 2016-12-01 10:35:45 +01:00
drm_mode_config.c drm/fence: add out-fences support 2016-11-16 14:36:27 +01:00
drm_mode_object.c
drm_modes.c Revert "drm: Add aspect ratio parsing in DRM layer" 2016-11-15 15:01:42 +01:00
drm_modeset_helper.c drm: also move DSI panels to the front of the connector list 2016-11-17 11:59:26 +01:00
drm_modeset_lock.c drm/doc: Fix indenting in drm_modeset_lock.c comment 2016-11-29 23:34:36 +01:00
drm_of.c
drm_panel.c
drm_pci.c
drm_plane.c Merge tag 'drm-misc-next-2016-12-08' of git://anongit.freedesktop.org/git/drm-misc into drm-next 2016-12-09 12:29:10 +10:00
drm_plane_helper.c drm: add helpers to go from plane state to drm_rect 2016-11-08 16:38:03 -05:00
drm_platform.c
drm_prime.c drm: Take ownership of the dmabuf->obj when exporting 2016-12-08 10:29:22 +01:00
drm_print.c drm/print: Move kerneldoc next to definition 2016-11-15 12:55:24 +01:00
drm_probe_helper.c drm: Don't block the kworker waiting for mode_config.mutex in output_poll() 2016-12-06 13:41:15 +01:00
drm_property.c drm: Use u64_to_user_ptr() helper for blob ioctls 2016-11-28 08:12:18 +01:00
drm_rect.c drm: helper macros to print composite types 2016-11-08 16:38:03 -05:00
drm_scatter.c
drm_simple_kms_helper.c
drm_sysfs.c
drm_trace.h
drm_trace_points.c
drm_vm.c mm: use vmf->address instead of of vmf->virtual_address 2016-12-14 16:04:09 -08:00
drm_vma_manager.c
Kconfig Add support for the Amlogic Meson Video Processing Unit 2016-12-08 10:35:52 +10:00
Makefile Add support for the Amlogic Meson Video Processing Unit 2016-12-08 10:35:52 +10:00