]> git.proxmox.com Git - mirror_ubuntu-bionic-kernel.git/commit - drivers/gpu/drm/i915/i915_gem_gtt.c
drm/i915: preallocate pdps for 32 bit vgpu
authorZhiyuan Lv <zhiyuan.lv@intel.com>
Fri, 28 Aug 2015 07:41:14 +0000 (15:41 +0800)
committerDaniel Vetter <daniel.vetter@ffwll.ch>
Wed, 2 Sep 2015 09:42:53 +0000 (11:42 +0200)
commit331f38e77d4519d2ca60cea232c7f2bfad5e4d28
treee1dc9418ac39b73315d4679f96f9b353122c4a72
parent42a8ca4cb4a48ddbf40e8edb291425e76bcdc230
drm/i915: preallocate pdps for 32 bit vgpu

This is based on Mika Kuoppala's patch below:
http://article.gmane.org/gmane.comp.freedesktop.xorg.drivers.intel/61104/match=workaround+hw+preload

The patch will preallocate the page directories for 32-bit PPGTT when
i915 runs inside a virtual machine with Intel GVT-g. With this change,
the root pointers in EXECLIST context will always keep the same.

The change is needed for vGPU because Intel GVT-g will do page table
shadowing, and needs to track all the page table changes from guest
i915 driver. However, if guest PPGTT is modified through GPU commands
like LRI, it is not possible to trap the operations in the right time,
so it will be hard to make shadow PPGTT to work correctly.

Shadow PPGTT could be much simpler with this change. Meanwhile
hypervisor could simply prohibit any attempt of PPGTT modification
through GPU command for security.

The function gen8_preallocate_top_level_pdps() in the patch is from
Mika, with only one change to set "used_pdpes" to avoid duplicated
allocation later.

Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Cc: Dave Gordon <david.s.gordon@intel.com>
Cc: Michel Thierry <michel.thierry@intel.com>
Signed-off-by: Zhiyuan Lv <zhiyuan.lv@intel.com>
Signed-off-by: Zhi Wang <zhi.a.wang@intel.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
drivers/gpu/drm/i915/i915_gem_gtt.c
drivers/gpu/drm/i915/intel_lrc.c