Re: [PATCH] firmware: stratix10-rsu: Add synchronous fallback for async SVC operations

[NG, TZE YEE]

On 9/7/2026 11:19 pm, Dinh Nguyen wrote:
> 
> 
> On 7/9/26 04:13, NG, TZE YEE wrote:
>> On 26/5/2026 5:41 pm, NG, TZE YEE wrote:
>>> From: Mahesh Rao <mahesh.rao@xxxxxxxxxx>
>>>
>>> The RSU driver was migrated to the Stratix10 asynchronous service
>>> framework, which assumes async client registration always succeeds. On
>>> platforms where the async controller is unavailable or RSU async support
>>> is not present, probe fails or sysfs operations cannot reach firmware.
>>>
>>> Detect async capability at probe via stratix10_svc_add_async_client() 
>>> and
>>> fall back to the legacy synchronous rsu_send_msg() path when 
>>> registration
>>> fails. Track capability in priv->async and branch accordingly for:
>>>
>>>    - Initial RSU status and retry counter during probe (retry is 
>>> bundled in
>>>      async status; issue COMMAND_RSU_RETRY separately in sync mode)
>>>    - notify sysfs store (notify, status, and retry)
>>>    - SPT table retrieval (COMMAND_RSU_GET_SPT_TABLE async vs mbox
>>>      COMMAND_MBOX_SEND_CMD synchronous)
>>>
>>> Restore synchronous callbacks (rsu_status_callback, rsu_retry_callback,
>>> rsu_get_spt_callback) and wire COMMAND_MBOX_SEND_CMD payload handling in
>>> rsu_send_msg().
>>>
>>> Improve probe and remove cleanup: return on intermediate probe failures,
>>> remove the async client when registered, and free the SPT response 
>>> buffer
>>> on sync-path allocation errors.
>>>
> 
> 
>>>    }
>>
>> Hi,
>>
>> Gentle ping on this patch — any chance for a review when you have time?
>> Happy to rebase or revise based on your feedback.
>>
> 
> Sorry for missing this. Can you please rebase this to 
> socfpga_firmware_for_v7.3 on my tree and I will review it.
> 
> Thanks,
> Dinh
> 

Hi Dinh,

Sure. I have rebased to socfpga_firmware_for_v7.3 in v2: 
https://lore.kernel.org/all/e991a29b0c7bd567da6b9e9a3a59486774a34ab2.1783666551.git.tze.yee.ng@xxxxxxxxxx/

Thanks for checking.
Tze Yee