Re: [PATCH v2 03/20] scsi: ibmvfc: complete NVMe FCP requests on H_CLOSED send failure

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] scsi: ibmvfc: misleading commit message contradicts safe implementation
--

commit a9d59d0a40253b9e7dff6fac485fc0f17bf38172
Author: Tyrel Datwyler <tyreld@linux.ibm.com>

scsi: ibmvfc: complete NVMe FCP requests on H_CLOSED send failure

This commit addresses a hang caused by silently dropped NVMe-FC requests 
when ibmvfc_send_event() returns H_CLOSED. It changes the code to correctly
signal a retryable transport-level failure to the NVMe-FC transport.

>     Fix this by calling fcp_req->done() with a -EBUSY status before freeing
>     the event in the H_CLOSED path, consistent with how other FC-NVMe LLDDs
>     (e.g. lpfc) signal a retryable transport-level failure to the NVMe-FC
>     transport.

[ ... ]

> diff --git a/drivers/scsi/ibmvscsi/ibmvfc-nvme.c b/drivers/scsi/ibmvscsi/ibmvfc-nvme.c
> index 9ed6522c2e9e3..08dd897b8e64e 100644
> --- a/drivers/scsi/ibmvscsi/ibmvfc-nvme.c
> +++ b/drivers/scsi/ibmvscsi/ibmvfc-nvme.c
> @@ -367,8 +367,12 @@ static int ibmvfc_nvme_fcp_io(struct nvme_fc_local_port *lport,
>  
>  	vfc_cmd->correlation = cpu_to_be64((u64)evt);
>  
> -	if (likely(!(rc = ibmvfc_nvme_map_sg_data(fcp_req, evt, vfc_cmd))))
> -		return ibmvfc_send_event(evt, vhost, 0);
> +	if (likely(!(rc = ibmvfc_nvme_map_sg_data(fcp_req, evt, vfc_cmd)))) {
> +		rc = ibmvfc_send_event(evt, vhost, 0);
> +		if (rc)
> +			return -EBUSY;

[Severity: Medium]
Does the commit message accurately describe the implementation here?

The commit message explicitly claims the fix is implemented by "calling
fcp_req->done() with a -EBUSY status before freeing the event". However,
the code safely returns -EBUSY directly to the NVMe-FC transport without
invoking the done() callback.

If the implementation actually matched the commit message, it would have
caused a double-completion bug by calling done() while also failing the
submission API. The code appears correct by returning -EBUSY, but the
commit message misrepresents the implemented transport interaction.

Could the commit message be updated to match the code?

> +		return 0;
> +	}
>  
>  	ibmvfc_free_event(evt);

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260919013306.2948028-1-tyreld@linux.ibm.com?part=3