mbox series

[v2,0/7] drm/i915/pxp: Prepare intel_pxp entry points for MTL

Message ID 20221006043834.606220-1-alan.previn.teres.alexis@intel.com (mailing list archive)
Headers show
Series drm/i915/pxp: Prepare intel_pxp entry points for MTL | expand

Message

Alan Previn Oct. 6, 2022, 4:38 a.m. UTC
MTL has two tiles that is represented by the intel_gt structure in the i915
code. The PXP feature has a control-structure that contains the PXP context
and this hangs of the intel_gt structure. In MTL, the standalone media tile
(i.e. not the root tile) contains the VDBOX and KCR engine which is what
PXP relies on for establishing and tearing down the PXP session. However
PXP is a global feature as other engines on other tiles can reference the
PXP session in object info within batch buffer instructions.That coherrency
is handled implicitly by the HW. However current intel_pxp functions such
as intel_pxp_enabled, intel_pxp_start and others take in the intel_gt
structure pointer as its input thus creation the perception that PXP is
a GT-tile specific domain that is independant from other GT tiles.

This series updates all of the intel_pxp_foo functions that are accessed
from outside the PXP subsystem so that the callers only need to pass in the
i915 structure as the input param (being a global handle). Internally,
these functions will loop through all available GT structures on the GPU
and find the one GT structure that contains the one PXP+TEE control
structure before proceeding with the rest of its operation.

Changes from prior revs:
   v1: Add one more patch to the series for the intel_pxp suspend/resume
       for similiar refactoring

Alan Previn (7):
  drm/i915/pxp: Make gt and pxp init/fini aware of PXP-owning-GT
  drm/i915/pxp: Make intel_pxp_is_enabled implicitly sort PXP-owning-GT
  drm/i915/pxp: Make intel_pxp_is_active implicitly sort PXP-owning-GT
  drm/i915/pxp: Make PXP tee component bind/unbind aware of
    PXP-owning-GT
  drm/i915/pxp: Make intel_pxp_start implicitly sort PXP-owning-GT
  drm/i915/pxp: Make intel_pxp_key_check implicitly sort PXP-owning-GT
  drm/i915/pxp: Make intel_pxp power management implicitly sort
    PXP-owning-GT

 .../drm/i915/display/skl_universal_plane.c    |  2 +-
 drivers/gpu/drm/i915/gem/i915_gem_context.c   |  6 +-
 drivers/gpu/drm/i915/gem/i915_gem_create.c    |  2 +-
 .../gpu/drm/i915/gem/i915_gem_execbuffer.c    |  2 +-
 drivers/gpu/drm/i915/gt/intel_gt.c            |  4 +
 drivers/gpu/drm/i915/gt/intel_gt_pm.c         | 10 +--
 drivers/gpu/drm/i915/gt/intel_gt_types.h      |  5 ++
 drivers/gpu/drm/i915/gt/intel_sa_media.c      |  4 +
 drivers/gpu/drm/i915/i915_drv.h               |  6 +-
 drivers/gpu/drm/i915/i915_pci.c               |  1 +
 drivers/gpu/drm/i915/intel_device_info.h      |  1 +
 drivers/gpu/drm/i915/pxp/intel_pxp.c          | 79 ++++++++++++++++---
 drivers/gpu/drm/i915/pxp/intel_pxp.h          | 10 ++-
 drivers/gpu/drm/i915/pxp/intel_pxp_cmd.c      |  2 +-
 drivers/gpu/drm/i915/pxp/intel_pxp_debugfs.c  |  8 +-
 drivers/gpu/drm/i915/pxp/intel_pxp_irq.c      |  7 +-
 drivers/gpu/drm/i915/pxp/intel_pxp_pm.c       | 44 ++++++++---
 drivers/gpu/drm/i915/pxp/intel_pxp_pm.h       | 22 +++---
 drivers/gpu/drm/i915/pxp/intel_pxp_tee.c      | 21 ++++-
 19 files changed, 177 insertions(+), 59 deletions(-)

Comments

Alan Previn Oct. 10, 2022, 7:05 p.m. UTC | #1
I dont believe either of these failures are related to my changes as ICL and SKL doesn't support PXP and after re-
looking at the change to "intel_pxp_is_enabled", I am confident it remains consistent with prior code in that it would
return FALSE for any HW without PXP support after it checks "pxp->ce" like before.
...alan


On Thu, 2022-10-06 at 18:19 +0000, Patchwork wrote:
> Patch Details
> Series:drm/i915/pxp: Prepare intel_pxp entry points for MTL
> (rev2)URL:https://patchwork.freedesktop.org/series/109429/State:failure
> Details:https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109429v2/index.html
> CI Bug Log - changes from CI_DRM_12221_full -> Patchwork_109429v2_fullSummaryPossible new issues
> Here are the unknown changes that may have been introduced in Patchwork_109429v2_full:
> CI changesPossible regressions * boot:shard-iclb: (PASS, PASS, PASS, PASS, PASS, PASS, PASS, PASS, PASS, PASS, PASS, PASS, PASS, PASS, PASS, PASS,
>    PASS, PASS, PASS, PASS, PASS, PASS, PASS, PASS, PASS) -> (PASS, PASS, PASS, PASS, PASS, PASS, PASS, PASS, PASS, PASS,
>    PASS, PASS, PASS, PASS, PASS, PASS, PASS, PASS, PASS, PASS, FAIL, PASS, PASS, PASS, PASS)
> IGT changesPossible regressions * igt@kms_sequence@queue-idle@edp-1-pipe-b:shard-skl: PASS -> FAIL +1 similar issue
> Suppressed