Re: [PATCH v21 06/14] dmaengine: qcom: bam_dma: add support for BAM locking

[Bartosz Golaszewski]

On Tue, 14 Jul 2026 11:49:14 +0200, Stephan Gerhold
<stephan.gerhold@xxxxxxxxxx> said:
>
>> +
>>  /**
>>   * bam_prep_slave_sg - Prep slave sg transaction
>>   *
>> [...]
>> @@ -802,6 +851,7 @@ static int bam_dma_terminate_all(struct dma_chan *chan)
>>  		}
>>
>>  		vchan_get_all_descriptors(&bchan->vc, &head);
>> +		bchan->bam_locked = false;
>
> I wonder about the implications of this. If the LOCK descriptor was
> already processed, will we cause a deadlock if we never submit the
> UNLOCK descriptor? Or I guess bam_reset_channel() might reset the lock
> as well?
>

There's nothing in the HPG that would imply it but I instrumented a test where
I excercised this situation and QCE was not locked up in any way. If there
were more execution environments doing the same, I'm not sure but this series
was tested by the NAND team and they didn't notice anything.

Bart