Messages in this thread | | | Date | Wed, 11 May 2022 16:21:54 -0700 | Subject | Re: [PATCH v4] drm/msm/dp: Always clear mask bits to disable interrupts at dp_ctrl_reset_irq_ctrl() | From | Kuogee Hsieh <> |
| |
On 5/11/2022 2:47 PM, Dmitry Baryshkov wrote: > On 12/05/2022 00:40, Kuogee Hsieh wrote: >> dp_catalog_ctrl_reset() will software reset DP controller. But it will >> not reset programmable registers to default value. DP driver still have >> to clear mask bits to interrupt status registers to disable interrupts >> after software reset of controller. This patch removes the enable flag >> condition checking to always clear mask bits of interrupt status >> registers to disable interrupts if enable flag is false. >> >> This patch also will fix the potential problem happen at system >> suspend where >> dp_ctrl_reset_irq_ctrl() was called to try to disable HPD related >> irqs but >> the irq is still unmasked unexpectedly > > Why is it unmasked? Especially unexpectedly.
due to dp_catalog_ctrl_reset() does not clear hpd interrupt mask bits.
This is the problem this patch try to fix it.
I will revise commit text to explain more.
> >> and can come in while system are >> suspending. This leads to bus hangs if the irq is handled after we >> power down >> the DP hardware because we run the irq handler and access a device >> register >> assuming that no irq could ever come in if we powered down the >> device. We >> don't know when the irq will be handled though, so it's possible the >> irq is >> pending from before we disable the irq in the hardware. > > Please split into two patches.
this is just the same problem mention above.
> >> >> Changes in v2: >> -- add more details commit text >> >> Changes in v3: >> -- add synchrons_irq() >> -- add atomic_t suspended >> >> Changes in v4: >> -- correct Fixes's commit ID >> >> Fixes: 989ebe7bc446 ("drm/msm/dp: do not initialize phy until plugin >> interrupt received") >> Signed-off-by: Kuogee Hsieh <quic_khsieh@quicinc.com> >> --- >> drivers/gpu/drm/msm/dp/dp_ctrl.c | 9 +++++++-- >> drivers/gpu/drm/msm/dp/dp_display.c | 18 ++++++++++++++++++ >> 2 files changed, 25 insertions(+), 2 deletions(-) >> >> diff --git a/drivers/gpu/drm/msm/dp/dp_ctrl.c >> b/drivers/gpu/drm/msm/dp/dp_ctrl.c >> index af7a80c..f3e333e 100644 >> --- a/drivers/gpu/drm/msm/dp/dp_ctrl.c >> +++ b/drivers/gpu/drm/msm/dp/dp_ctrl.c >> @@ -1389,8 +1389,13 @@ void dp_ctrl_reset_irq_ctrl(struct dp_ctrl >> *dp_ctrl, bool enable) >> dp_catalog_ctrl_reset(ctrl->catalog); >> - if (enable) >> - dp_catalog_ctrl_enable_irq(ctrl->catalog, enable); >> + /* >> + * all dp controller programmable registers will not >> + * be reset to default value after DP_SW_RESET >> + * therefore interrupt mask bits have to be updated >> + * to enable/disable interrupts >> + */ >> + dp_catalog_ctrl_enable_irq(ctrl->catalog, enable); >> } >> void dp_ctrl_phy_init(struct dp_ctrl *dp_ctrl) >> diff --git a/drivers/gpu/drm/msm/dp/dp_display.c >> b/drivers/gpu/drm/msm/dp/dp_display.c >> index c388323..c34dbfc 100644 >> --- a/drivers/gpu/drm/msm/dp/dp_display.c >> +++ b/drivers/gpu/drm/msm/dp/dp_display.c >> @@ -98,6 +98,8 @@ struct dp_display_private { >> struct dp_ctrl *ctrl; >> struct dp_debug *debug; >> + atomic_t suspended; > > Usage of atomic_t signifies that something is wrong here. Bool should > be enough here.
suspended flag are shared by both event_thread and display_irq_handler, i think it need either atomic
or protected by spinlock.
> >> + >> struct dp_usbpd_cb usbpd_cb; >> struct dp_display_mode dp_mode; >> struct msm_dp dp_display; >> @@ -187,6 +189,11 @@ static int dp_add_event(struct >> dp_display_private *dp_priv, u32 event, >> int pndx; >> spin_lock_irqsave(&dp_priv->event_lock, flag); >> + if (atomic_read(&dp_priv->suspended)) { >> + spin_unlock_irqrestore(&dp_priv->event_lock, flag); >> + return -EPERM; > > Why EPERM? > >> + } >> + >> pndx = dp_priv->event_pndx + 1; >> pndx %= DP_EVENT_Q_MAX; >> if (pndx == dp_priv->event_gndx) { >> @@ -454,6 +461,13 @@ static void dp_display_host_deinit(struct >> dp_display_private *dp) >> dp->dp_display.connector_type, dp->core_initialized, >> dp->phy_initialized); >> + if (!dp->core_initialized) { > > Can this happen?
not likely,
but It should not in this patch. will double check.
> >> + DRM_DEBUG_DP("DP core not initialized\n"); >> + return; >> + } >> + >> + synchronize_irq(dp->irq); > > Why? If you need to run with IRQs disabled, you can use > suspend_late/early_resume. But generally it should be enough to check > for the !suspended in the IRQ. > >> + >> dp_ctrl_reset_irq_ctrl(dp->ctrl, false); >> dp_aux_deinit(dp->aux); >> dp_power_deinit(dp->power); >> @@ -1362,6 +1376,8 @@ static int dp_pm_resume(struct device *dev) >> dp->dp_display.connector_type, dp->core_initialized, >> dp->phy_initialized, dp_display->power_on); >> + atomic_set(&dp->suspended, 0); >> + >> /* start from disconnected state */ >> dp->hpd_state = ST_DISCONNECTED; >> @@ -1431,6 +1447,8 @@ static int dp_pm_suspend(struct device *dev) >> dp->dp_display.connector_type, dp->core_initialized, >> dp->phy_initialized, dp_display->power_on); >> + atomic_inc(&dp->suspended); >> + >> /* mainlink enabled */ >> if (dp_power_clk_status(dp->power, DP_CTRL_PM)) >> dp_ctrl_off_link_stream(dp->ctrl); > >
| |