Skip to content

L
linux
Switch branch/tag
  1. 31 May, 2023 3 commits
  3. 26 May, 2023 5 commits
  5. 24 May, 2023 7 commits
  7. 23 May, 2023 1 commit
  9. 22 May, 2023 2 commits
  11. 19 May, 2023 10 commits
  13. 18 May, 2023 1 commit
  15. 17 May, 2023 3 commits
  17. 16 May, 2023 8 commits
    • Imre Deak's avatar
      drm/i915/tc: Reset TypeC PHYs left enabled in DP-alt mode after the sink disconnects · c598c335
      Imre Deak authored 
      If the output on a DP-alt link with its sink disconnected is kept
enabled for too long (about 20 sec), then some IOM/TCSS firmware timeout
will cause havoc on the PCI bus, at least for other GFX devices on it
which will stop powering up. Since user space is not guaranteed to do a
disabling modeset in time, switch such disconnected but active links to
TBT mode - which is without such shortcomings - with a 2 second delay.

If the above condition is detected already during the driver load/system
resume sanitization step disable the output instead, as at that point no
user space or kernel client depends on a consistent output state yet and
because subsequent atomic modeset on such connectors - without the
actual sink capabilities available - can fail.

An active/disconnected port as above will also block the HPD status of
other active/disconnected ports to get updated (stuck in the connected
state), until the former port is disabled, its PHY is disconnected and
a ~10 ms delay has elapsed. This means the link state for all TypeC
ports/CRTCs must be rechecked after a CRTC is disabled due to the above
reason. For this disconnect the PHY synchronously after the CRTC/port is
disabled and recheck all CRTCs for the above condition whenever such a
port is disabled.

To account for a race condition during driver loading where the sink is
disconnected after the above sanitization step and before the HPD
interrupts get enabled, do an explicit check/link reset if needed from
the encoder's late_register hook, which is called after the HPD
interrupts are enabled already.

v2:
- Handle an active/disconnected port blocking the HPD state update of
  another active/disconnected port.
- Cancel the delayed work resetting the link also from the encoder
  enable/suspend/shutdown hooks.
- Rebase on the earlier intel_modeset_lock_ctx_retry() addition,
  fixing here the missed atomic state reset in case of a retry.
- Fix handling of an error return from intel_atomic_get_crtc_state().
- Recheck if the port needs to be reset after all the atomic state
  is locked and async commits are waited on.

v3:
- Add intel_crtc_needs_link_reset(), instead of open-coding it,
  keep intel_crtc_has_encoders(). (Ville)
- Fix state dumping and use a bitmask to track disabled CRTCs in
  intel_sanitize_all_crtcs(). (Ville)
- Set internal in intel_atomic_state right after allocating it.
  (Ville)
- Recheck all CRTCs (not yet force-disabled) after a CRTC is
  force-disabled for any reason (not only due to a link state)
  in intel_sanitize_all_crtcs().
- Reduce delay after CRTC disabling to 20ms, and use the simpler
  msleep().
- Clarify code comment about HPD behaviour in
  intel_sanitize_all_crtcs().
- Move all the TC link reset logic to intel_tc.c .
- Cancel the link reset work synchronously during system suspend,
  driver unload and shutdown.

v4:
- Rebased on previous patch, which allows calling the TC port
  suspend/cleanup handlers without modeset locks held; remove the
  display driver suspended assert from the link reset work
  accordingly.

v5: (Ville)
- Remove reset work canceling from intel_ddi_pre_pll_enable().
- Track a crtc vs. pipe mask in intel_sanitize_all_crtcs().
- Add reset_link_commit() to clarify the
  intel_modeset_lock_ctx_retry loop.

Cc: Kai-Heng Feng <kai.heng.feng@canonical.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Tested-by: default avatarKai-Heng Feng <kai.heng.feng@canonical.com>
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/5860Reviewed-by: default avatarVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: default avatarImre Deak <imre.deak@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230512195513.2699-2-imre.deak@intel.com
      c598c335
    • Imre Deak's avatar
      drm/i915/tc: Call TypeC port flush_work/cleanup without modeset locks held · b61fad5f
      Imre Deak authored 
      Call the TypeC port flush_work and cleanup handlers without the modeset
locks held. These don't require the locks, as the work takes - as it
should be able to at any point in time - any locks it needs and by the
time cleanup is called and after cleanup returns the encoder is not in
use.

This is required by the next patch canceling a TypeC port work
synchronously during encoder suspend and shutdown, where the work can
take modeset locks as well, hence the canceling must be done without
holding the locks.

I also considered moving the modeset locking down to each encoder
suspend()/shutdown() hook instead, however locking the full modeset
state for each encoder separately would be odd, and the bigger change -
affecting all encoders - is beyond the scope of this patchset.

v2:
- Add a TODO: comment to remove modeset locks if no encoder depends
  on this. (Ville)

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: default avatarVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: default avatarImre Deak <imre.deak@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230512195513.2699-1-imre.deak@intel.com
      b61fad5f
    • Imre Deak's avatar
      drm/i915: Factor out a helper for handling atomic modeset locks/state · 60ded7cc
      Imre Deak authored 
      This patch simplifying the handling of modeset locks and atomic state
for an atomic commit is based on

https://lore.kernel.org/all/20210715184954.7794-2-ville.syrjala@linux.intel.com/

adding the helper to i915. I find this approach preferrable than
open-coding the corresponding steps (fixed for me an atomic
state reset during a DEADLK retry, which I missed in the open-coded
version) and also better than the existing
DRM_MODESET_LOCK_ALL_BEGIN/END macros for the reasons described in the
above original patchset.

This change takes the helper into use only for atomic commits during DDI
hotplug handling, as a preparation for a follow-up patch adding a
similar commit started from the same spot. Other places doing a
driver-internal atomic commit is to be converted by a follow-up
patchset.

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: default avatarVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: default avatarImre Deak <imre.deak@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230510103131.1618266-13-imre.deak@intel.com
      60ded7cc
    • Imre Deak's avatar
      drm/i915/dp: Factor out intel_dp_get_active_pipes() · 7e4460c3
      Imre Deak authored 
      Factor out a helper used by a follow up patch to reset an active DP
link.

No functional changes.

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: default avatarVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: default avatarImre Deak <imre.deak@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230510103131.1618266-12-imre.deak@intel.com
      7e4460c3
    • Imre Deak's avatar
      drm/i915/dp: Prevent link training fallback on disconnected port · f45156ff
      Imre Deak authored 
      Prevent downgrading the link training maximum lane count/rate if the
sink is disconnected - and so the link training failure is expected. In
such cases modeset failures due to the reduced max link params would be
just confusing for user space (instead of which the correct thing it
should act on is the sink disconnect signaled by a hotplug event,
requiring a disabling modeset).

v2:
- Check the actual HPD state to handle the forced connector state case.
  (Vinod)

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Vinod Govindapillai <vinod.govindapillai@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> (v1)
Reviewed-by: default avatarVinod Govindapillai <vinod.govindapillai@intel.com>
Signed-off-by: default avatarImre Deak <imre.deak@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230510103131.1618266-11-imre.deak@intel.com
      f45156ff
    • Imre Deak's avatar
      drm/i915/dp: Convert link training error to debug message on disconnected sink · 9031c6d4
      Imre Deak authored 
      If a sink is disconnected it's expected that link training actions will
fail on it, so downgrade the error messages about such actions to be a
debug message. Such - expected - link training failures are more
frequent after a follow up patch, after which an active TypeC link is
reset after the sink is disconnected which also involves a link
training.

v2:
- Check the actual HPD state to handle the forced connector state case.
  (Vinod)

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Vinod Govindapillai <vinod.govindapillai@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> (v1)
Reviewed-by: default avatarVinod Govindapillai <vinod.govindapillai@intel.com>
Signed-off-by: default avatarImre Deak <imre.deak@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230510103131.1618266-10-imre.deak@intel.com
      9031c6d4
    • Imre Deak's avatar
      drm/i915/dp: Add link training debug and error printing helpers · f48eab29
      Imre Deak authored 
      Add functions for printing link training debug and error messages, both
to prepare for the next patch, which downgrades an error to a debug
message if the sink is disconnected and to remove some code duplication.

v2: (Ville)
- Always print the connector prefix.
- Preserve the drm_dbg_kms() debug category.
v3:
- Keep printing the name of functions calling the helpers. (Jani)

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> (v2)
Signed-off-by: default avatarImre Deak <imre.deak@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230510103131.1618266-9-imre.deak@intel.com
      f48eab29
    • Imre Deak's avatar
      drm/i915: Add support for disabling any CRTCs during HW readout/sanitization · e826839e
      Imre Deak authored 
      During HW readout/sanitization CRTCs can be disabled only if they don't
have an attached encoder (and so the encoder disable hooks don't need to
be called). An upcoming patch will need to disable CRTCs also with an
attached encoder, so add support for this.

For bigjoiner configs the encoder disabling hooks require the slave CRTC
states, so add these too to the atomic state. Since the connector atomic
state is already up-to-date when the CRTC is disabled the connector
state needs to be updated (reset) after the CRTC is disabled, make this
so. Follow the proper order of disabling first all bigjoiner slaves,
then any port synced CRTC slaves followed by the CRTC originally
requested to be disabled.

v2:
- Fix calculating the bigjoiner_masters mask in a port sync config,
  (Ville)
- Keep _noatomic suffix in intel_crtc_disable_noatomic(). (Ville)
- Rebase on full CRTC state reset in this patchset, not requiring
  resetting the bigjoiner state separately and (instead) resetting
  the full atomic CRTC and related global state after all linked
  pipes got disabled.
- Disable portsync slaves before a portsync master.
- Disable a portsync master if a linked portsync slave is disabled.

v3: (Ville)
- Use s/u32/u8 for transcoder and pipe masks.
- Use is_power_of_2() instead of hweight()==1.

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: default avatarVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: default avatarImre Deak <imre.deak@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230510103131.1618266-8-imre.deak@intel.com
      e826839e
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7