| Message ID | 20201102072746.27410-1-sreekanth.reddy@broadcom.com (mailing list archive) |
|---|---|
| State | Accepted |
| Headers | show |
| Series | mpt3sas: Fix timeouts observed while reenabling IRQ | expand |
On 11/2/20 8:27 AM, Sreekanth Reddy wrote: > While reenabling the IRQ after irq poll there may be small time > window where HBA firmware has posted some replies and raise the > interrupts but driver has not received the interrupts. So we may > observe IO timeouts as the driver has not processed the replies > as interrupts got missed while reenabling the IRQ. > > So, to fix this issue, the driver has to go for one more > round of processing the reply descriptors from reply descriptor > post queue after enabling the IRQ. > > Signed-off-by: Sreekanth Reddy <sreekanth.reddy@broadcom.com> > Reported-by: Tomas Henzl <thenzl@redhat.com> > --- > drivers/scsi/mpt3sas/mpt3sas_base.c | 7 +++++++ > 1 file changed, 7 insertions(+) > > diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.c b/drivers/scsi/mpt3sas/mpt3sas_base.c > index b096917..a0ab44d 100644 > --- a/drivers/scsi/mpt3sas/mpt3sas_base.c > +++ b/drivers/scsi/mpt3sas/mpt3sas_base.c > @@ -1740,6 +1740,13 @@ _base_irqpoll(struct irq_poll *irqpoll, int budget) > reply_q->irq_poll_scheduled = false; > reply_q->irq_line_enable = true; > enable_irq(reply_q->os_irq); > + /* > + * Go for one more round of processing the > + * reply descriptor post queue incase if HBA > + * Firmware has posted some reply descriptors > + * while reenabling the IRQ. > + */ > + _base_process_reply_queue(reply_q); > } > > return num_entries; > I've tested the fix and the system now works. Reviewed-by: Tomas Henzl <thenzl@redhat.com>
On Mon, 2 Nov 2020 12:57:46 +0530, Sreekanth Reddy wrote: > While reenabling the IRQ after irq poll there may be small time > window where HBA firmware has posted some replies and raise the > interrupts but driver has not received the interrupts. So we may > observe IO timeouts as the driver has not processed the replies > as interrupts got missed while reenabling the IRQ. > > So, to fix this issue, the driver has to go for one more > round of processing the reply descriptors from reply descriptor > post queue after enabling the IRQ. Applied to 5.10/scsi-fixes, thanks! [1/1] scsi: mpt3sas: Fix timeouts observed while reenabling IRQ https://git.kernel.org/mkp/scsi/c/5feed64f9199
diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.c b/drivers/scsi/mpt3sas/mpt3sas_base.c index b096917..a0ab44d 100644 --- a/drivers/scsi/mpt3sas/mpt3sas_base.c +++ b/drivers/scsi/mpt3sas/mpt3sas_base.c @@ -1740,6 +1740,13 @@ _base_irqpoll(struct irq_poll *irqpoll, int budget) reply_q->irq_poll_scheduled = false; reply_q->irq_line_enable = true; enable_irq(reply_q->os_irq); + /* + * Go for one more round of processing the + * reply descriptor post queue incase if HBA + * Firmware has posted some reply descriptors + * while reenabling the IRQ. + */ + _base_process_reply_queue(reply_q); } return num_entries;
While reenabling the IRQ after irq poll there may be small time window where HBA firmware has posted some replies and raise the interrupts but driver has not received the interrupts. So we may observe IO timeouts as the driver has not processed the replies as interrupts got missed while reenabling the IRQ. So, to fix this issue, the driver has to go for one more round of processing the reply descriptors from reply descriptor post queue after enabling the IRQ. Signed-off-by: Sreekanth Reddy <sreekanth.reddy@broadcom.com> Reported-by: Tomas Henzl <thenzl@redhat.com> --- drivers/scsi/mpt3sas/mpt3sas_base.c | 7 +++++++ 1 file changed, 7 insertions(+)