On Fri, 2021-09-10 at 14:04 -0500, wenxiong@xxxxxxxxxxxxxxxxxx wrote:
> From: Wen Xiong <root@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx>
>
> We saw two errors with Slider drawer:
> - Failed to get diagnostic page 0x1 during booting up
> - /sys/class/enclosure is empty with ipr adapter + Slider drawer
>
> From scsi logging level with error=3, looks ses_recv_diag not try on
> a UA. The patch addes retrying on a UA in ses_recv_diag();
Do we know why the device is returning a UA? Presumably it's a check
condition UA meaning the device is trying to tell us something and
wants us to request sense?
> Signed-Off-by: Wen Xiong<wenxiong@xxxxxxxxxxxxxxxxxx>
> Reviewed-by: Brian King <brking@xxxxxxxxxxxxxxxxxx>
> ---
> drivers/scsi/ses.c | 11 +++++++++--
> 1 file changed, 9 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/scsi/ses.c b/drivers/scsi/ses.c
> index c2afba2a5414..93f6a8ce1bea 100644
> --- a/drivers/scsi/ses.c
> +++ b/drivers/scsi/ses.c
> @@ -87,9 +87,16 @@ static int ses_recv_diag(struct scsi_device *sdev,
> int page_code,
> 0
> };
> unsigned char recv_page_code;
> + struct scsi_sense_hdr sshdr;
> + int retries = SES_RETRIES;
> +
> + do {
> + ret = scsi_execute_req(sdev, cmd, DMA_FROM_DEVICE,
> buf,
> + bufflen, &sshdr, NULL, SES_TIMEOUT,
This grew an additional argument: you want to replace the NULL with the
sshdr, I think ... and compile test it next time.
> SES_RETRIES, NULL);
I think you want a 1 here instead of SES_RETRIES because you're
retrying for SES_RETRIES in an outer loop now, so if you maxed out both
sets of retries, you'd retry for SES_RETRIES^2.
If this is a CC/UA, you can simplify all this by setting
cmd->expecting_cc_ua = 1
and avoiding the loop.
James
