On Fri, 2011-12-02 at 13:43 -0800, Nicholas A. Bellinger wrote:
> On Fri, 2011-12-02 at 22:13 +0100, Sebastian Andrzej Siewior wrote:
> > Commit f281233d aka ("SCSI host lock push-down") was converting all
> > drivers in tree. It looks to me that those two were not omitted on
> > purpose. tcm_loop joined the tree after the conversion and virtio_scsi
> > lives only in Nicholas' lio tree. So here is an equivalent change for
> > those two.
> >
> > Signed-off-by: Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx>
> > ---
> > Having a differnt callback name for converted and not converted users
> > would probably make it easier to tell them apart.
> >
> > drivers/scsi/virtio_scsi.c | 4 +++-
> > drivers/target/loopback/tcm_loop.c | 4 +++-
> > 2 files changed, 6 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/scsi/virtio_scsi.c b/drivers/scsi/virtio_scsi.c
> > index 7946f40..4c577fa 100644
> > --- a/drivers/scsi/virtio_scsi.c
> > +++ b/drivers/scsi/virtio_scsi.c
> > @@ -151,7 +151,7 @@ static void virtscsi_map_cmd(struct virtio_scsi *vscsi, struct scsi_cmnd *sc,
> > sg_set_buf(&sg[idx++], &cmd->footer, sizeof(cmd->footer));
> > }
> >
> > -static int virtscsi_queuecommand(struct Scsi_Host *sh, struct scsi_cmnd *sc)
> > +static int virtscsi_queuecommand_lck(struct Scsi_Host *sh, struct scsi_cmnd *sc)
> > {
> > struct virtio_scsi *vscsi = shost_priv(sh);
> > struct virtio_scsi_cmd *cmd;
> > @@ -190,6 +190,8 @@ static int virtscsi_queuecommand(struct Scsi_Host *sh, struct scsi_cmnd *sc)
> > return ret;
> > }
> >
> > +static DEF_SCSI_QCMD(virtscsi_queuecommand);
> > +
> > static struct scsi_host_template virtscsi_host_template = {
> > .module = THIS_MODULE,
> > .name = "Virtio SCSI HBA",
> > diff --git a/drivers/target/loopback/tcm_loop.c b/drivers/target/loopback/tcm_loop.c
> > index 5f69594..92ef58a 100644
> > --- a/drivers/target/loopback/tcm_loop.c
> > +++ b/drivers/target/loopback/tcm_loop.c
> > @@ -266,7 +266,7 @@ static int tcm_loop_change_queue_depth(
> > * Main entry point from struct scsi_host_template for incoming SCSI CDB+Data
> > * from Linux/SCSI subsystem for SCSI low level device drivers (LLDs)
> > */
> > -static int tcm_loop_queuecommand(
> > +static int tcm_loop_queuecommand_lck(
> > struct Scsi_Host *sh,
> > struct scsi_cmnd *sc)
> > {
> > @@ -310,6 +310,8 @@ static int tcm_loop_queuecommand(
> > return 0;
> > }
> >
> > +static DEF_SCSI_QCMD(tcm_loop_queuecommand)
> > +
> > /*
> > * Called from SCSI EH process context to issue a LUN_RESET TMR
> > * to struct scsi_device
>
> Both of these are already running in 'hock_lock-less' mode as they
> should be, and there is no reason why it should be using DEF_SCSI_QCMD()
> in order to revert to holding hock_lock.
Or host_lock even..
--nab
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
