Re: [PATCH v8 06/14] drm/bridge: it6505: bail out of the IRQ handler when status reads fail

From: Luca Ceresoli

Date: Thu Nov 19 2026 - 12:10:01 EST


On Mon Nov 16, 2026 at 9:55 AM CET, Daniel Golle wrote:
> Hi Luca,
>
> On Tue, Oct 20, 2026 at 06:51:39PM +0200, Luca Ceresoli wrote:
>> One first question it whether a read failure is realistic or not, and
>> whether you have experienced it on real hardware.
>
> I have not hit it on hardware. It takes a failing i2c transfer, e.g.
> the chip NACKing while held in reset, which can be provoked through
> the force_power_on_off debugfs knob: it6505_poweroff() only calls
> disable_irq_nosync(), so a running handler keeps reading from a chip
> that is being put into reset.
>
> And hostesly, my interest here is basically getting audio going on
> that Lenovo Ideapad, and for that
>
> https://lore.kernel.org/all/cover.1784561622.git.daniel@xxxxxxxxxxxxxx/
>
> is sufficient. The 12 other patches are what Sahsiko has been chasing
> me into, and all the complains it had looked valid and legitimate, so
> I took on writing the fixes for that as well.

Ah, well, I understand.

I had missed the initial history of this series, thanks for the recap!

>> I would consider replacing all the 3 it6505_read() calls with a single
>> regmap_bulk_read(), making error management cleaner.
>>
>> What do you think about this option?
>>
>> Perhaps the three it6505_write() below could also become a bulk write.
>
> Makes sense, INT_STATUS_01..03 are contiguous and the driver already
> uses multi-byte transfers elsewhere. I'll switch to a bulk read and
> write in v9.

Good, thanks!

Luca

--
Luca Ceresoli, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com