Intel-XE Archive on lore.kernel.org
 help / color / mirror / Atom feed
* [FOR CI 0/5] drm/xe/display: Test series with PREEMPT_RT and preemption disable.
@ 2025-10-23 15:34 Maarten Lankhorst
  2025-10-23 15:34 ` [FOR CI 1/5] drm/xe: Bump xe_device_l2_flush even higher Maarten Lankhorst
                   ` (9 more replies)
  0 siblings, 10 replies; 17+ messages in thread
From: Maarten Lankhorst @ 2025-10-23 15:34 UTC (permalink / raw)
  To: intel-xe

Fixing the issues with waiting from the previous version.

Maarten Lankhorst (3):
  drm/xe: Bump xe_device_l2_flush even higher
  drm/xe/display: Disable preemption in the most critical section
  PREEMPT_RT injection

Mike Galbraith (2):
  drm/i915: Use preempt_disable/enable_rt() where recommended
  drm/i915: Don't disable interrupts on PREEMPT_RT during atomic updates

 drivers/gpu/drm/i915/Kconfig.debug          | 15 ----
 drivers/gpu/drm/i915/display/intel_crtc.c   | 23 +++++-
 drivers/gpu/drm/i915/display/intel_cursor.c | 27 ++++++-
 drivers/gpu/drm/i915/display/intel_vblank.c | 88 ++++++++++++++-------
 drivers/gpu/drm/xe/Kconfig.debug            |  5 ++
 drivers/gpu/drm/xe/xe_device.c              |  4 +-
 kernel/Kconfig.preempt                      |  4 +-
 7 files changed, 112 insertions(+), 54 deletions(-)

-- 
2.51.0


^ permalink raw reply	[flat|nested] 17+ messages in thread
* [FOR CI 0/5] Testing PREEMPT_RT with disabling preemption in the most critical section.
@ 2025-10-29  9:59 Maarten Lankhorst
  2025-10-29  9:59 ` [FOR CI 3/5] drm/i915: Don't disable interrupts on PREEMPT_RT during atomic updates Maarten Lankhorst
  0 siblings, 1 reply; 17+ messages in thread
From: Maarten Lankhorst @ 2025-10-29  9:59 UTC (permalink / raw)
  To: intel-xe

Fix get_vblank_counter() and try to disable preemption in the most critical part of the code.

Maarten Lankhorst (3):
  drm/i915/display: Make get_vblank_counter use intel_de_read_fw()
  drm/xe/display: Disable preemption in the most critical section
  PREEMPT_RT injection

Mike Galbraith (2):
  drm/i915: Use preempt_disable/enable_rt() where recommended
  drm/i915: Don't disable interrupts on PREEMPT_RT during atomic updates

 drivers/gpu/drm/i915/Kconfig.debug          | 15 ----
 drivers/gpu/drm/i915/display/intel_crtc.c   | 23 +++++-
 drivers/gpu/drm/i915/display/intel_cursor.c | 27 ++++++-
 drivers/gpu/drm/i915/display/intel_vblank.c | 90 ++++++++++++++-------
 drivers/gpu/drm/xe/Kconfig.debug            |  5 ++
 kernel/Kconfig.preempt                      |  4 +-
 6 files changed, 111 insertions(+), 53 deletions(-)

-- 
2.51.0


^ permalink raw reply	[flat|nested] 17+ messages in thread
* [FOR CI 0/5] Testing PREEMPT_RT with disabling interrupts in the most critical section.
@ 2025-10-23  8:55 Maarten Lankhorst
  2025-10-23  8:55 ` [FOR CI 3/5] drm/i915: Don't disable interrupts on PREEMPT_RT during atomic updates Maarten Lankhorst
  0 siblings, 1 reply; 17+ messages in thread
From: Maarten Lankhorst @ 2025-10-23  8:55 UTC (permalink / raw)
  To: intel-xe

Only disable preemption in this series.

Maarten Lankhorst (3):
  drm/xe: Bump xe_device_l2_flush even higher
  drm/xe/display: Disable preemption in the most critical section
  PREEMPT_RT injection

Mike Galbraith (2):
  drm/i915: Use preempt_disable/enable_rt() where recommended
  drm/i915: Don't disable interrupts on PREEMPT_RT during atomic updates

 drivers/gpu/drm/i915/Kconfig.debug          | 15 ----
 drivers/gpu/drm/i915/display/intel_crtc.c   | 23 +++++-
 drivers/gpu/drm/i915/display/intel_cursor.c | 27 ++++++-
 drivers/gpu/drm/i915/display/intel_vblank.c | 81 ++++++++++++++-------
 drivers/gpu/drm/xe/Kconfig.debug            |  5 ++
 drivers/gpu/drm/xe/xe_device.c              |  4 +-
 kernel/Kconfig.preempt                      |  4 +-
 7 files changed, 108 insertions(+), 51 deletions(-)

-- 
2.51.0


^ permalink raw reply	[flat|nested] 17+ messages in thread
* [FOR CI 0/5] drm/xe/display: Test series with PREEMPT_RT and preemption disable.
@ 2025-10-22 12:13 Maarten Lankhorst
  2025-10-22 12:13 ` [FOR CI 3/5] drm/i915: Don't disable interrupts on PREEMPT_RT during atomic updates Maarten Lankhorst
  0 siblings, 1 reply; 17+ messages in thread
From: Maarten Lankhorst @ 2025-10-22 12:13 UTC (permalink / raw)
  To: intel-xe; +Cc: Maarten Lankhorst

Keep interrupts enabled, see what happens...

Maarten Lankhorst (3):
  drm/xe: Bump xe_device_l2_flush even higher
  drm/xe/display: Disable preemption in the most critical section
  PREEMPT_RT injection

Mike Galbraith (2):
  drm/i915: Use preempt_disable/enable_rt() where recommended
  drm/i915: Don't disable interrupts on PREEMPT_RT during atomic updates

 drivers/gpu/drm/i915/Kconfig.debug          | 15 ------
 drivers/gpu/drm/i915/display/intel_crtc.c   | 23 ++++++--
 drivers/gpu/drm/i915/display/intel_cursor.c | 27 ++++++++--
 drivers/gpu/drm/i915/display/intel_vblank.c | 58 ++++++++++++++++-----
 drivers/gpu/drm/xe/Kconfig.debug            |  5 ++
 drivers/gpu/drm/xe/xe_device.c              |  4 +-
 kernel/Kconfig.preempt                      |  4 +-
 7 files changed, 97 insertions(+), 39 deletions(-)

-- 
2.51.0


^ permalink raw reply	[flat|nested] 17+ messages in thread
* [FOR CI 0/5] drm/xe/display: Disable migration for the critical
@ 2025-10-21 12:43 Maarten Lankhorst
  2025-10-21 12:43 ` [FOR CI 3/5] drm/i915: Don't disable interrupts on PREEMPT_RT during atomic updates Maarten Lankhorst
  0 siblings, 1 reply; 17+ messages in thread
From: Maarten Lankhorst @ 2025-10-21 12:43 UTC (permalink / raw)
  To: intel-xe

Using the principle of minimal impact to latency possible, only prevent migration
from occurring first.

Maarten Lankhorst (3):
  drm/xe: Bump xe_device_l2_flush even higher
  drm/xe/display: Disable migration in the most critical section
  [FOR-CI] PREEMPT_RT injection

Mike Galbraith (2):
  drm/i915: Use preempt_disable/enable_rt() where recommended
  drm/i915: Don't disable interrupts on PREEMPT_RT during atomic updates

 drivers/gpu/drm/i915/Kconfig.debug          | 15 ------
 drivers/gpu/drm/i915/display/intel_crtc.c   | 23 ++++++--
 drivers/gpu/drm/i915/display/intel_cursor.c | 27 ++++++++--
 drivers/gpu/drm/i915/display/intel_dmc.c    |  4 +-
 drivers/gpu/drm/i915/display/intel_vblank.c | 58 ++++++++++++++++-----
 drivers/gpu/drm/xe/Kconfig.debug            |  5 ++
 drivers/gpu/drm/xe/xe_device.c              |  4 +-
 kernel/Kconfig.preempt                      |  3 +-
 8 files changed, 98 insertions(+), 41 deletions(-)

-- 
2.51.0


^ permalink raw reply	[flat|nested] 17+ messages in thread
* [FOR CI 0/5] Testing PREEMPT_RT with disabling interrupts in the most critical section.
@ 2025-10-17 15:37 Maarten Lankhorst
  2025-10-17 15:37 ` [FOR CI 3/5] drm/i915: Don't disable interrupts on PREEMPT_RT during atomic updates Maarten Lankhorst
  0 siblings, 1 reply; 17+ messages in thread
From: Maarten Lankhorst @ 2025-10-17 15:37 UTC (permalink / raw)
  To: intel-xe

*** BLURB HERE ***

Maarten Lankhorst (3):
  drm/xe: Bump xe_device_l2_flush even higher
  drm/i915/display: Disable interrupts on xe for realtime preemption
  [FOR-CI] PREEMPT_RT injection

Mike Galbraith (2):
  drm/i915: Use preempt_disable/enable_rt() where recommended
  drm/i915: Don't disable interrupts on PREEMPT_RT during atomic updates

 drivers/gpu/drm/i915/Kconfig.debug          | 15 -------
 drivers/gpu/drm/i915/display/intel_crtc.c   | 13 ++++--
 drivers/gpu/drm/i915/display/intel_cursor.c |  9 ++--
 drivers/gpu/drm/i915/display/intel_vblank.c | 49 ++++++++++++++++-----
 drivers/gpu/drm/i915/display/intel_vblank.h |  9 ++++
 drivers/gpu/drm/xe/Kconfig.debug            |  5 +++
 drivers/gpu/drm/xe/xe_device.c              |  4 +-
 kernel/Kconfig.preempt                      |  3 +-
 8 files changed, 69 insertions(+), 38 deletions(-)

-- 
2.51.0


^ permalink raw reply	[flat|nested] 17+ messages in thread
* [FOR CI 0/5] Test PREEMPT_RT with irqs disabled.
@ 2025-10-17 11:57 Maarten Lankhorst
  2025-10-17 11:57 ` [FOR CI 3/5] drm/i915: Don't disable interrupts on PREEMPT_RT during atomic updates Maarten Lankhorst
  0 siblings, 1 reply; 17+ messages in thread
From: Maarten Lankhorst @ 2025-10-17 11:57 UTC (permalink / raw)
  To: intel-xe

Just a quick check to see if PREEMPT_RT runs into any issues without
IRQs.

Maarten Lankhorst (3):
  drm/xe: Bump xe_device_l2_flush even higher
  drm/i915/display: Disable interrupts on xe for realtime preemption
  [FOR-CI] PREEMPT_RT injection

Mike Galbraith (2):
  drm/i915: Use preempt_disable/enable_rt() where recommended
  drm/i915: Don't disable interrupts on PREEMPT_RT during atomic updates

 drivers/gpu/drm/i915/Kconfig.debug          | 15 -------
 drivers/gpu/drm/i915/display/intel_crtc.c   | 13 ++++--
 drivers/gpu/drm/i915/display/intel_cursor.c |  9 ++--
 drivers/gpu/drm/i915/display/intel_vblank.c | 49 ++++++++++++++++-----
 drivers/gpu/drm/i915/display/intel_vblank.h |  9 ++++
 drivers/gpu/drm/xe/Kconfig.debug            |  5 +++
 drivers/gpu/drm/xe/xe_device.c              |  4 +-
 kernel/Kconfig.preempt                      |  3 +-
 8 files changed, 69 insertions(+), 38 deletions(-)

-- 
2.51.0


^ permalink raw reply	[flat|nested] 17+ messages in thread

end of thread, other threads:[~2025-10-29  9:59 UTC | newest]

Thread overview: 17+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-10-23 15:34 [FOR CI 0/5] drm/xe/display: Test series with PREEMPT_RT and preemption disable Maarten Lankhorst
2025-10-23 15:34 ` [FOR CI 1/5] drm/xe: Bump xe_device_l2_flush even higher Maarten Lankhorst
2025-10-23 15:34 ` [FOR CI 2/5] drm/i915: Use preempt_disable/enable_rt() where recommended Maarten Lankhorst
2025-10-23 15:34 ` [FOR CI 3/5] drm/i915: Don't disable interrupts on PREEMPT_RT during atomic updates Maarten Lankhorst
2025-10-23 15:34 ` [FOR CI 4/5] drm/xe/display: Disable preemption in the most critical section Maarten Lankhorst
2025-10-23 15:34 ` [FOR CI 5/5] PREEMPT_RT injection Maarten Lankhorst
2025-10-23 18:23 ` ✗ CI.checkpatch: warning for drm/xe/display: Test series with PREEMPT_RT and preemption disable. (rev2) Patchwork
2025-10-23 18:24 ` ✓ CI.KUnit: success " Patchwork
2025-10-23 18:38 ` ✗ CI.checksparse: warning " Patchwork
2025-10-23 19:16 ` ✗ Xe.CI.BAT: failure " Patchwork
2025-10-24  7:42 ` ✗ Xe.CI.Full: " Patchwork
  -- strict thread matches above, loose matches on Subject: below --
2025-10-29  9:59 [FOR CI 0/5] Testing PREEMPT_RT with disabling preemption in the most critical section Maarten Lankhorst
2025-10-29  9:59 ` [FOR CI 3/5] drm/i915: Don't disable interrupts on PREEMPT_RT during atomic updates Maarten Lankhorst
2025-10-23  8:55 [FOR CI 0/5] Testing PREEMPT_RT with disabling interrupts in the most critical section Maarten Lankhorst
2025-10-23  8:55 ` [FOR CI 3/5] drm/i915: Don't disable interrupts on PREEMPT_RT during atomic updates Maarten Lankhorst
2025-10-22 12:13 [FOR CI 0/5] drm/xe/display: Test series with PREEMPT_RT and preemption disable Maarten Lankhorst
2025-10-22 12:13 ` [FOR CI 3/5] drm/i915: Don't disable interrupts on PREEMPT_RT during atomic updates Maarten Lankhorst
2025-10-21 12:43 [FOR CI 0/5] drm/xe/display: Disable migration for the critical Maarten Lankhorst
2025-10-21 12:43 ` [FOR CI 3/5] drm/i915: Don't disable interrupts on PREEMPT_RT during atomic updates Maarten Lankhorst
2025-10-17 15:37 [FOR CI 0/5] Testing PREEMPT_RT with disabling interrupts in the most critical section Maarten Lankhorst
2025-10-17 15:37 ` [FOR CI 3/5] drm/i915: Don't disable interrupts on PREEMPT_RT during atomic updates Maarten Lankhorst
2025-10-17 11:57 [FOR CI 0/5] Test PREEMPT_RT with irqs disabled Maarten Lankhorst
2025-10-17 11:57 ` [FOR CI 3/5] drm/i915: Don't disable interrupts on PREEMPT_RT during atomic updates Maarten Lankhorst

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox