[PATCH] [SCSI] sd: Lower printk level for drive cache information

[PATCH] [SCSI] sd: Lower printk level for drive cache information

[Josh Boyer]

Numerous devices don't present cache data, particularly USB sticks.  The
devices operate correctly despite this, so lower the printk severity to
KERN_NOTICE.

Signed-off-by: Josh Boyer <jwboyer@redhat.com>

---

diff --git a/drivers/scsi/sd.c b/drivers/scsi/sd.c
index fa3a591..7cf9ec5 100644
--- a/drivers/scsi/sd.c
+++ b/drivers/scsi/sd.c
@@ -2119,7 +2119,7 @@ sd_read_cache_type(struct scsi_disk *sdkp, unsigned char *buffer)
 		}
 
 		if (modepage == 0x3F) {
-			sd_printk(KERN_ERR, sdkp, "No Caching mode page "
+			sd_printk(KERN_NOTICE, sdkp, "No Caching mode page "
 				  "present\n");
 			goto defaults;
 		} else if ((buffer[offset] & 0x3f) != modepage) {
@@ -2164,7 +2164,7 @@ bad_sense:
 		sd_printk(KERN_ERR, sdkp, "Asking for cache data failed\n");
 
 defaults:
-	sd_printk(KERN_ERR, sdkp, "Assuming drive cache: write through\n");
+	sd_printk(KERN_NOTICE, sdkp, "Assuming drive cache: write through\n");
 	sdkp->WCE = 0;
 	sdkp->RCD = 0;
 	sdkp->DPOFUA = 0;

Re: [PATCH] [SCSI] sd: Lower printk level for drive cache information

[Christoph Hellwig]

On Mon, Nov 28, 2011 at 11:31:55AM -0500, Josh Boyer wrote:
> Numerous devices don't present cache data, particularly USB sticks.  The
> devices operate correctly despite this, so lower the printk severity to
> KERN_NOTICE.

Do they?  The last reports on the list were for devices that didn't
present it but actually needed cache flushes despite not having the
cache mode page.

Re: [PATCH] [SCSI] sd: Lower printk level for drive cache information

[Josh Boyer]

On Mon, Nov 28, 2011 at 11:44:42PM -0500, Christoph Hellwig wrote:
> On Mon, Nov 28, 2011 at 11:31:55AM -0500, Josh Boyer wrote:
> > Numerous devices don't present cache data, particularly USB sticks.  The
> > devices operate correctly despite this, so lower the printk severity to
> > KERN_NOTICE.
> 
> Do they?  The last reports on the list were for devices that didn't
> present it but actually needed cache flushes despite not having the
> cache mode page.

We've seen a number of reports for liveusb images that spit out those
printks, but seem to otherwise work fine.

Even if the devices need cache flushes, what is a typical user supposed
to do about that?  I'm not particularly tied to this patch, but there is
nothing a normal user can do with the information spit out by these
printks, so making it KERN_ERR seems excessive.

josh
>