From: Martin Peschke <mpeschke@linux.vnet.ibm.com>
Date: Fri, 30 Jan 2009 14:46:23 +0000 (+0100)
Subject: [SCSI] sg: fix device number in blktrace data
X-Git-Url: https://git.karo-electronics.de/?a=commitdiff_plain;h=76e3a19d0691bbfcc559ce77ab3004818fab8f22;p=linux-beck.git

[SCSI] sg: fix device number in blktrace data

Hi,

we have run into an issue with blktrace being started for sg devices.
Please apply.

Thanks,
Martin

From: Martin Peschke <mpeschke@linux.vnet.ibm.com>

The device number denoting a generic SCSI devices (sg) in a blktrace
trace is broken; major and minor are always 0. It looks like
sdp->device->sdev_gendev.devt is not initialized properly.
The fix below uses other data to make up a valid device number,
similar to the way an sg device number is generated for sysfs output.

Reported-by: Stefan Raspl <raspl@linux.vnet.ibm.com>
Signed-off-by: Martin Peschke <mpeschke@linux.vnet.ibm.com>
Acked-by: Douglas Gilbert <dgilbert@interlog.com>
Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
---

diff --git a/drivers/scsi/sg.c b/drivers/scsi/sg.c
index 8f0bd3f7a59f..516925d8b570 100644
--- a/drivers/scsi/sg.c
+++ b/drivers/scsi/sg.c
@@ -1078,7 +1078,7 @@ sg_ioctl(struct inode *inode, struct file *filp,
 	case BLKTRACESETUP:
 		return blk_trace_setup(sdp->device->request_queue,
 				       sdp->disk->disk_name,
-				       sdp->device->sdev_gendev.devt,
+				       MKDEV(SCSI_GENERIC_MAJOR, sdp->index),
 				       (char *)arg);
 	case BLKTRACESTART:
 		return blk_trace_startstop(sdp->device->request_queue, 1);