On 10/28/2014 06:33 PM, Christoph Hellwig wrote: > On Fri, Oct 24, 2014 at 09:00:15PM +0000, Elliott, Robert (Server Storage) > wrote: >> With MLQUEUE level 0 and MLCOMPLETE level 0, v8 runs those prints, >> and they work as expected: >> * the value in the ratelimit message matches the block layer (82) >> * the number of print_result/print_sense/print_command matches >> the number of block layer prints (10) >> * the SCSI prints end up interleaved, but that's for the next >> patch series to fix >> >> Assuming the MLQUEUE -> MLCOMPLETE fix is put back in, >> you may add: >> Tested-by: Robert Elliott <elli...@hp.com> >> Reviewed-by: Robert Elliott <elli...@hp.com> > > Hannes, can you send me a quick incremental fix for this issue? > Of course.
Here it is. Cheers, Hannes -- Dr. Hannes Reinecke zSeries & Storage h...@suse.de +49 911 74053 688 SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)
>From 7da8db4d39ae71b065e0167339a6e4107f1243b8 Mon Sep 17 00:00:00 2001 From: Hannes Reinecke <h...@suse.de> Date: Tue, 28 Oct 2014 18:42:15 +0100 Subject: [PATCH] scsi: fixup SCSI_LOG_MLCOMPLETE instead of SCSI_LOG_MLQUEUE Robert Elliott pointed out that we need to use SCSI_LOG_MLCOMPLETE in scsi_io_completion() instead of SCSI_LOG_MLQUEUE. Signed-off-by: Hannes Reinecke <h...@suse.de> --- drivers/scsi/scsi_lib.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/scsi/scsi_lib.c b/drivers/scsi/scsi_lib.c index 4572991..7711dbd 100644 --- a/drivers/scsi/scsi_lib.c +++ b/drivers/scsi/scsi_lib.c @@ -1045,8 +1045,8 @@ void scsi_io_completion(struct scsi_cmnd *cmd, unsigned int good_bytes) DEFAULT_RATELIMIT_BURST); if (unlikely(scsi_logging_level)) - level = SCSI_LOG_LEVEL(SCSI_LOG_MLQUEUE_SHIFT, - SCSI_LOG_MLQUEUE_BITS); + level = SCSI_LOG_LEVEL(SCSI_LOG_MLCOMPLETE_SHIFT, + SCSI_LOG_MLCOMPLETE_BITS); /* * if logging is enabled the failure will be printed -- 1.8.4.5
