Thanks for pointing this out. Returning ABORTED was not generally a problem, but it limited user space ability to distinguish between different failure modes. Changing this improves debugability and allows applications to take actions based on separate return codes accordingly.
I have improved clarification of the change in commit message in PATCHv2. Karol On 10/7/2025 4:54 PM, Jeff Hugo wrote: > On 10/7/2025 2:35 AM, Karol Wachowski wrote: >> From: Andrzej Kacprowski <[email protected]> >> >> The driver was returning ABORTED for all error that trigger engine >> reset. > > "is returning" right? The driver currently does this (without this > patch)? > > Is this a bad thing? Should the driver do something different? I feel > like there should be more explanation here. > >> Refactor ivpu_job_signal_and_destroy() by extracting engine error >> handling logic into a new function ivpu_job_handle_engine_error(). >> This simplifies engine error handling logic by removing necessity of >> calling ivpu_job_singal_and_destroy() multiple times by a single job >> changing it's behavior based on job status. >> >> Signed-off-by: Andrzej Kacprowski <[email protected]> >> Signed-off-by: Karol Wachowski <[email protected]>
