[RFC 0/5] drm/xe: Support capture and dump of devcoredump for general debug

[John.C.Harrison@Intel.com]

From: John Harrison <John.C.Harrison@Intel.com>

It is useful to be able to dump driver/hardware state when various
unexpected errors occur. E.g. on an internal error in the GuC
communication layer, there is a dump of the GuC state. Currently the
CT code rolls its own capture and print. However, the devcoredump
mechanism is basically doing exactly the same thing. So tweak that to
allow it to be called from arbitrary places and use it instead.

Signed-off-by: John Harrison <John.C.Harrison@Intel.com>


John Harrison (5):
  drm/xe/devcoredump: Support coredumps without jobs
  drm/xe: Trigger a devcoredump capture on a GT reset
  drm/xe: Disconnect coredump structure from xe_device structure
  drm/xe: Make coredump printing to in-memory cache optional
  drm/xe: Support devcoredump capture from dead CT handler

 drivers/gpu/drm/xe/xe_devcoredump.c       | 266 ++++++++++++++++------
 drivers/gpu/drm/xe/xe_devcoredump.h       |   9 +-
 drivers/gpu/drm/xe/xe_devcoredump_types.h |   2 +
 drivers/gpu/drm/xe/xe_gt.c                |   4 +
 drivers/gpu/drm/xe/xe_guc_ct.c            |  57 +++--
 drivers/gpu/drm/xe/xe_guc_ct_types.h      |   6 +-
 drivers/gpu/drm/xe/xe_guc_submit.c        |   2 +-
 7 files changed, 241 insertions(+), 105 deletions(-)

-- 
2.47.0