On 04/10/15 23:21, Nicholas A. Bellinger wrote:
On Fri, 2015-04-10 at 14:59 +0200, Bart Van Assche wrote:
__transport_register_session() must be called with se_tpg->session_lock
held. Fix the call sites where this lock is not held.
Signed-off-by: Bart Van Assche <bart.vanassche@xxxxxxxxxxx>
Cc: Christoph Hellwig <hch@xxxxxx>
Cc: <qla2xxx-upstream@xxxxxxxxxx>
Cc: Felipe Balbi <balbi@xxxxxx>
Cc: Michael S. Tsirkin <mst@xxxxxxxxxx>
Cc: Juergen Gross <jgross@xxxxxxxx>
---
drivers/scsi/qla2xxx/tcm_qla2xxx.c | 6 ++----
drivers/target/loopback/tcm_loop.c | 7 ++-----
drivers/usb/gadget/legacy/tcm_usb_gadget.c | 5 ++---
drivers/vhost/scsi.c | 5 ++---
drivers/xen/xen-scsiback.c | 7 ++-----
5 files changed, 10 insertions(+), 20 deletions(-)
This series was broken out and merged in v4.0-rc5 code.
The qla2xxx one has already been pushed to stable, and you've been CC'ed
on a dozen emails for it.
The other fabrics only ever use a single session per endpoint, so they
can't run into problems here.
Hello Nic,
The only reason why the qla2xxx changes were present in my patch is
because I had started from your for-next branch and because the qla2xxx
changes are not yet present in your for-next branch. But you are right,
it would have been better if I would have left out these changes from my
patch.
Even though the other target drivers only use a single session per
endpoint, what these other target drivers do conflicts with the comment
above __transport_register_session():
/*
* Called with spin_lock_irqsave(&struct se_portal_group->session_lock
* called.
*/
I'm afraid that kernel developers who have not followed this discussion,
who notice the __transport_register_session() calls and who see the
above comment will get confused. Is it OK for you if I send a second
version of my patch that clears up this confusion ?
Thanks,
Bart.
--
To unsubscribe from this list: send the line "unsubscribe target-devel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
