Re: [PATCH v3 0/6] media: iris: prepare support for video codecs on Qcom vpu4 platform

From: Vikash Garodia

Date: Fri Nov 21 2025 - 06:03:31 EST



On 11/20/2025 3:19 PM, Bryan O'Donoghue wrote:
On 12/11/2025 05:09, Vikash Garodia wrote:
One additional bonus is that fluster runs several process in parallel
by default, catching issues caused by several decode threads running
in parallel.

multi process issue is due to below [1] (tried it on lemans). Due to
higher concurrency, we can see that the DMA buffer is mapped into
un-addressable range (0-0x25800000) i.e 0x24b00000, and leading to
global fault. This was the reason i was keeping 2 memory-region in
kaanapali binding, to restrict certain ranges of IOVA.

If that is currently true then we should restrict concurrent sessions entirely unless/until it can be safely supported.


I would prefer to fix it than disable the feature. I would raising a fix for it.

Also, to keep it clear, this is nothing to do with this patch series. The issue about non usable IOVA range has been there for qcom vpu, it not newly introduced in iris driver. It was not exposed in venus and it is seen now in iris with concurrency when IOVA falls into that non addressable range.

Regards,
Vikash