Hi James, After further review of this patch please do not include it for 3.12-rc. Thanks, ~Saurav >From: Chad Dupuis <chad.dupuis@xxxxxxxxxx> > >If the call to abort a request fails we need to clear it's entry in the >in-flight >commands array as it won't be cleared in interrupt context. Not doing so >could >potentially lead to a double completion. > >Signed-off-by: Chad Dupuis <chad.dupuis@xxxxxxxxxx> >Signed-off-by: Saurav Kashyap <saurav.kashyap@xxxxxxxxxx> >--- > drivers/scsi/qla2xxx/qla_os.c | 8 ++++++++ > 1 files changed, 8 insertions(+), 0 deletions(-) > >diff --git a/drivers/scsi/qla2xxx/qla_os.c b/drivers/scsi/qla2xxx/qla_os.c >index 9f01bbb..66ed1c9 100644 >--- a/drivers/scsi/qla2xxx/qla_os.c >+++ b/drivers/scsi/qla2xxx/qla_os.c >@@ -951,6 +951,7 @@ qla2xxx_eh_abort(struct scsi_cmnd *cmd) > unsigned long flags; > int wait = 0; > struct qla_hw_data *ha = vha->hw; >+ uint32_t handle = 0; > > if (!CMD_SP(cmd)) > return SUCCESS; >@@ -989,7 +990,14 @@ qla2xxx_eh_abort(struct scsi_cmnd *cmd) > } > > spin_lock_irqsave(&ha->hardware_lock, flags); >+ handle = sp->handle; > sp->done(ha, sp, 0); >+ /* >+ * If the mailbox command failed, clear the entry in the in-flight >+ * commands array as the entry won't be cleared in interrupt context. >+ */ >+ if (ret == FAILED) >+ vha->req->outstanding_cmds[handle] = NULL; > spin_unlock_irqrestore(&ha->hardware_lock, flags); > > /* Did the command return during mailbox execution? */ >-- >1.7.7 >
<<attachment: winmail.dat>>
