]> www.infradead.org Git - nvme.git/commit
drm/i915: Init DRM connector polled field early
authorImre Deak <imre.deak@intel.com>
Thu, 4 Jan 2024 08:29:57 +0000 (10:29 +0200)
committerImre Deak <imre.deak@intel.com>
Mon, 8 Jan 2024 16:51:31 +0000 (18:51 +0200)
commit289d4180bda98bfd47e0dac402a1caf2a8f50cf7
tree56defee96049b35a1e745110a0b50f382103d6ab
parent0fa647659c492c0a4342f7da70f5f946a40df250
drm/i915: Init DRM connector polled field early

After an HPD IRQ storm on a connector intel_hpd_irq_storm_detect() will
set the connector's HPD pin state to HPD_MARK_DISABLED and the IRQ gets
disabled. Subsequently intel_hpd_irq_storm_switch_to_polling() will
enable polling for these connectors, setting the pin state to
HPD_DISABLED, but only if the connector's base.polled field is set to
DRM_CONNECTOR_POLL_HPD. intel_hpd_irq_storm_reenable_work() will
reenable the IRQ - after 2 minutes -  if the pin state is HPD_DISABLED.

The connectors will be created with their base.polled field set to 0,
which gets initialized only later in i915_hpd_poll_init_work() (using
intel_connector::polled). If a storm is detected on a connector after
it's created and IRQs are enabled on it - by intel_hpd_init() - and
before its bease.polled field is initialized in the above work, the
connector's HPD pin will stay in the HPD_MARK_DISABLED state - leaving
the IRQ disabled indefinitely - and polling will not get enabled on it as
intended.

I can't see a reason for initializing base.polled in a delayed manner,
so do this already when creating the connector, to prevent the above
race condition.

Link: https://patchwork.freedesktop.org/patch/msgid/20240104083008.2715733-2-imre.deak@intel.com
Reviewed-by: Jouni Högander <jouni.hogander@intel.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
drivers/gpu/drm/i915/display/intel_crt.c
drivers/gpu/drm/i915/display/intel_dp.c
drivers/gpu/drm/i915/display/intel_dvo.c
drivers/gpu/drm/i915/display/intel_hdmi.c
drivers/gpu/drm/i915/display/intel_sdvo.c
drivers/gpu/drm/i915/display/intel_tv.c