Re: [PATCH v3] drm/msm/dp: check core_initialized flag at both host_init() and host_deinit()

[Dmitry Baryshkov]

On 04/03/2023 00:45, Kuogee Hsieh wrote:
> On 3/2/2023 11:04 AM, Dmitry Baryshkov wrote:
> > On Thu, 2 Mar 2023 at 20:41, Kuogee Hsieh <quic_khsieh@xxxxxxxxxxx> 
> > wrote:
> > > On 3/1/2023 1:15 PM, Dmitry Baryshkov wrote:
> > > > On 01/03/2023 18:57, Kuogee Hsieh wrote:
> > > > > On 2/28/2023 6:16 PM, Dmitry Baryshkov wrote:
> > > > > > On Wed, 1 Mar 2023 at 02:17, Kuogee Hsieh <quic_khsieh@xxxxxxxxxxx>
> > > > > > wrote:
> > > > > > > There is a reboot/suspend test case where system suspend is forced
> > > > > > > during system booting up. Since dp_display_host_init() of external
> > > > > > > DP is executed at hpd thread context, this test case may created a
> > > > > > > scenario that dp_display_host_deinit() from pm_suspend() run before
> > > > > > > dp_display_host_init() if hpd thread has no chance to run during
> > > > > > > booting up while suspend request command was issued. At this 
> > > > > > > scenario
> > > > > > > system will crash at aux register access at dp_display_host_deinit()
> > > > > > > since aux clock had not yet been enabled by dp_display_host_init().
> > > > > > > Therefore we have to ensure aux clock enabled by checking
> > > > > > > core_initialized flag before access aux registers at pm_suspend.
> > > > > > Can a call to dp_display_host_init() be moved from
> > > > > > dp_display_config_hpd() to dp_display_bind()?
> > > > > yes,  Sankeerth's  "drm/msm/dp: enable pm_runtime support for dp
> > > > > driver" patch is doing that which is under review.
> > > > >
> > > > > https://patchwork.freedesktop.org/patch/523879/?series=114297&rev=1
> > > > No, he is doing another thing. He is moving these calls to pm_runtime
> > > > callbacks, not to the dp_display_bind().
> > > >
> > > > > > Related question: what is the primary reason for having
> > > > > > EV_HPD_INIT_SETUP and calling dp_display_config_hpd() via the event
> > > > > > thread? Does DP driver really depend on DPU irqs being installed? As
> > > > > > far as I understand, DP device uses MDSS interrupts and those IRQs 
> > > > > > are
> > > > > > available and working at the time of dp_display_probe() /
> > > > > > dp_display_bind().
> > > > > HDP gpio pin has to run through DP aux module 100ms denouncing logic
> > > > > and have its mask bits.
> > > > >
> > > > > Therefore DP irq has to be enabled to receive DP isr with mask bits 
> > > > > set.
> > > > So... DP irq is enabled by the MDSS, not by the DPU. Again, why does
> > > > DP driver depend on DPU irqs being installed?
> > > sorry, previously i mis understand your question -- why does DP driver
> > > depend on DPU irqs being installed?
> > >
> > > now, I think you are asking why  dpu_irq_postinstall() ==>
> > > msm_dp_irq_postinstall() ==> event_thread ==> dp_display_config_hdp()
> > > ==> enable_irq(dp->irq)
> > >
> > > With the below test i had run, i think the reason is to make sure
> > > dp->irq be requested before enable it.
> > >
> > > I just run the execution timing order test and collect execution order
> > > as descending order at below,
> > >
> > > 1) dp_display_probe() -- start
> > >
> > > 2) dp_display_bind()
> > >
> > > 3) msm_dp_modeset_init()  ==> dp_display_request_irq() ==>
> > > dp_display_get_next_bridge()
> > >
> > > 4) dpu_irq_postinstall() ==> msm_dp_irq_postinstall() ==>
> > > enable_irq(dp->irq)
> > >
> > > 5) dp_display_probe() -- end
> > >
> > > dp->irq is request at msm_dp_modeset_init() and enabled after.
> > Should be moved to probe.
> >
> > > That bring up the issue to move DP's dp_display_host_init() executed at
> > > dp_display_bind().
> > >
> > > Since eDP have dp_dispaly_host_init() executed at
> > > dp_display_get_next_bridge() which executed after dp_display_bind().
> > >
> > > If moved DP's dp_display_host_init() to dp_dispaly_bind() which means DP
> > > will be ready to receive HPD irq before eDP ready.
> > And the AUX bus population should also be moved to probe(), which
> > means we should call dp_display_host_init() from probe() too.
> > Having aux_bus_populate in probe would allow moving component_add() to
> > the done_probing() callback, making probe/defer case more robust
> >
> > > This may create some uncertainties at execution flow and complicate
> > > things up.
> > Hopefully the changes suggested above will make it simpler.
>
> ok, I will create another patch to

patchset

> 1) move dp_display_host_init() to probe()
>
> 2) move component_add() to done_probing() for eDP
>
> 3) keep DP as simple platform device (component_add() still executed in 
> probe())

4) move devm_request_irq() to probe, add IRQF_NO_AUTOEN instead of 
calling disable_irq() right after request_irq()

5) drop DP_HPD_INIT_SETUP and related code

> Meanwhile, can you approve this patch so that it will not block our 
> internal daily testing?

Quoting your commit message: "Since dp_display_host_init() of external
DP is executed at hpd thread context...". After these changes the 
mentioned function will no longer be executed from the hpd thread. So, 
let's rework the probe/init sequence first, then we can see if this 
patch is still necessary and how should it look.

--
With best wishes
Dmitry