On Fri, May 1, 2020 at 9:40 AM Luis Chamberlain [off-list ref] wrote:

On Fri, May 01, 2020 at 08:34:23AM -0700, Christoph Hellwig wrote:

quoted

On Fri, May 01, 2020 at 03:06:26PM +0000, Luis Chamberlain wrote:

quoted

quoted

You have access to a block device here, please use dev_warn() instead
here for that, that makes it obvious as to what device a "concurrent
blktrace" was attempted for.

The block device may be empty, one example is for scsi-generic, but I'll
use buts->name.

Is blktrace on /dev/sg something we intentionally support, or just by
some accident of history?  Given all the pains it causes I'd be tempted
to just remove the support and see if anyone screams.

From what I can tell I think it was a historic and brutal mistake. I am
more than happy to remove it.

I take that back:

commit 6da127ad0918f93ea93678dad62ce15ffed18797
Author: Christof Schmitt [off-list ref]
Date:   Fri Jan 11 10:09:43 2008 +0100

    blktrace: Add blktrace ioctls to SCSI generic devices

    Since the SCSI layer uses the request queues from the block layer,
blktrace can
    also be used to trace the requests to all SCSI devices (like SCSI
tape drives),
    not only disks. The only missing part is the ioctl interface to
start and stop
    tracing.

    This patch adds the SETUP, START, STOP and TEARDOWN ioctls from
blktrace to the
    sg device files. With this change, blktrace can be used for SCSI
devices like
    for disks, e.g.: blktrace -d /dev/sg1 -o - | blkparse -i -

    Signed-off-by: Christof Schmitt [off-list ref]
    Signed-off-by: Jens Axboe [off-list ref]

Christof, any thoughts on removing this support?

 Luis