Re: [PATCH v4] scsi: trace: show rtn in string for scsi_dispatch_cmd_error

[James Bottomley <James.Bottomley@HansenPartnership.com>]

On Tue, 2025-05-20 at 09:59 +0800, Kassey Li wrote:
> By default it showed rtn in decimal:
> 
> kworker/3:1H-183 [003] ....  51.035474: scsi_dispatch_cmd_error:
> host_no=0 channel=0 id=0 lun=4 data_sgl=1  prot_sgl=0
> prot_op=SCSI_PROT
> _NORMAL cmnd=(READ_10 lba=3907214  txlen=1 protect=0 raw=28 00 00 3b
> 9e 8e 00 00 01 00) rtn=4181
> 
> In source code we define these possible values as hexadecimal:
> 
> include/scsi/scsi.h
> 
> SCSI_MLQUEUE_HOST_BUSY   0x1055
> SCSI_MLQUEUE_DEVICE_BUSY 0x1056
> SCSI_MLQUEUE_EH_RETRY    0x1057
> SCSI_MLQUEUE_TARGET_BUSY 0x1058
> 
> This change shows the rtn in strings:
> 
> dd-1059    [007] .....    31.689529: scsi_dispatch_cmd_error:
> host_no=0 channel=0 id=0 lun=4 data_sgl=65 prot_sgl=0
> prot_op=SCSI_PROT_NORMAL driver_tag=23 scheduler_tag=117
> cmnd=(READ_10 lba=0 txlen=128 protect=0 raw=28 00 00 00 00 00 00 00
> 80 00) rtn=SCSI_MLQUEUE_DEVICE_BUSY
> 
> Signed-off-by: Kassey Li <quic_yingangl@quicinc.com>
> ---
>  include/trace/events/scsi.h | 13 +++++++++++--
>  1 file changed, 11 insertions(+), 2 deletions(-)
> 
> diff --git a/include/trace/events/scsi.h
> b/include/trace/events/scsi.h
> index 8e2d9b1b0e77..38743635f2c1 100644
> --- a/include/trace/events/scsi.h
> +++ b/include/trace/events/scsi.h
> @@ -199,6 +199,14 @@ TRACE_EVENT(scsi_dispatch_cmd_start,
>  		  __print_hex(__get_dynamic_array(cmnd), __entry-
> >cmd_len))
>  );
>  
> +#define scsi_rtn_name(result)	{ result, #result }
> +#define show_rtn_name(val)					\
> +	__print_symbolic(val,					\
> +		scsi_rtn_name(SCSI_MLQUEUE_HOST_BUSY),		\
> +		scsi_rtn_name(SCSI_MLQUEUE_DEVICE_BUSY),	\
> +		scsi_rtn_name(SCSI_MLQUEUE_EH_RETRY),	\
> +		scsi_rtn_name(SCSI_MLQUEUE_TARGET_BUSY))
> +

This looks much more consistent with the style of the file, thanks;
just make sure the last backslash aligns.

Also, could you assure us you have actually compiled and tested v4?  It
was pretty obvious from the kbuild errors for v2 and v3 that they
hadn't even been compiled.

Regards,

James