Re: [PATCH v2 0/7] drm/i915/display: Handle vblank evasion with CONFIG_PREEMPT_RT

[Sebastian Andrzej Siewior <bigeasy@linutronix.de>]

On 2025-11-10 19:17:46 [+0100], Maarten Lankhorst wrote:
> Hey,
Hi Maarten,

> I would like to recommend dropping the patch below. The hardware doesn't like being
> programmed during vblank time, and may lock up or show glitches on the screen,
> especially at older machines.

Okay. I did include them for testing as requested. I would keep them in
the series until we settle on something upstream.

> That's why the whole complicated preparations exist, to be able to complete
> programming the hardware before the vblank.
> 
> I created my series to be able to run that the timing sensitive parts safely without
> any jitter from locking in between.
> 
> I tried running the following series through CI in response:
> https://patchwork.freedesktop.org/series/157258/ 
> 
> After looking at the results and investigating more closely, I think the FBC warning
> you mentioned is a false positive.
> 
> The code either runs in intel_pre_plane_update() when programming the hardware directly,
> or it gets added to the list of mmio's programmed by the hardware, without the
> vblank evasion mode active.
> 
> Still I would like to clean it up, but it's not as urgent as it was.

Okay. Good.
If there is anything you want me to test or look at, just le me know.

> Kind regards,
> ~Maarten Lankhorst

Sebastian