Re: [PATCH] media: ipu6: Don't stop or close a stream that failed to start

From: Jake Steinman

Date: Tue Nov 10 2026 - 16:03:57 EST


Hi Sakari,

On Tue, Nov 10, 2026 at 12:23:18PM +0200, Sakari Ailus wrote:
> Please be careful with git send-email arguments in the future.

Sorry, that was a quoting mistake in my Cc list. It's fixed for future
sends.

> I wonder if any firmware communication failure, a timeout or other error
> should trigger the IPU reset (once the users are gone). The functions
> communicating with the firmware should also probably check that condition.
> I just sent a PR on IPU reset changes, they probably don't address this
> problem though.

Agreed. This patch also applies cleanly on top of your ipu6 branch. With
"media: ipu6: Return errors from stream stop and close" there, a failed
start already marks need_reset in ipu6_isys_start_stream_firmware(), and
this patch only skips the stop and close calls that can't succeed.

One thing I've seen on Panther Lake: the PSYS is a child of the ISYS, and
as far as I can tell the ISYS isn't runtime suspended while the PSYS is in
use. So need_reset can stay set until the last PSYS user goes away, which
in practice means restarting PipeWire.

"media: ipu6: Don't mark need_reset based on returned buffers" covers the
IPU7 case from my need_reset RFC, so please drop that one. I'll test your
ipu6 branch on Panther Lake and report back.

Jake