ata_port_wait_eh(): Change irqsave into unconditional closing of interrupts

PATCH In review Kirill Tkhai 2 participants 3 messages Sep 26, 2013 libata

lore

Share & tools

lore

Kirill Tkhai

author

2013-09-26 13:27 UTC #0 lore raw

PATCH 1/1 ata_port_wait_eh(): Change irqsave into unconditional closing of interrupts

Hi,

ata_port_wait_eh() uses spin_lock_irqsave() and this can confuse in fact that it is suitable to use in irqs_disabled() context. But we can't.

(schedule() returns with interrupts enabled so it's possible ata_port_wait_eh() enters with disabled interrupts but returns with enabled)

So, replace irqsave to unconditional closing of interrupts.

I propose to consider to add patch like this. (If you don't have a magic with flags which is not obvious for me :)

Signed-off-by Kirill Tkhai <[email protected]>

---
 drivers/ata/libata-eh.c |    9 ++++-----
 1 files changed, 4 insertions(+), 5 deletions(-)

drivers/ata/libata-eh.c +4 -5

index c69fcce..502c1f9 100644
@@ -841,21 +841,20 @@ EXPORT_SYMBOL_GPL(ata_scsi_port_error_handler);
  */
 void ata_port_wait_eh(struct ata_port *ap)
 {
-	unsigned long flags;
 	DEFINE_WAIT(wait);
 
  retry:
-	spin_lock_irqsave(ap->lock, flags);
+	spin_lock_irq(ap->lock);
 
 	while (ap->pflags & (ATA_PFLAG_EH_PENDING | ATA_PFLAG_EH_IN_PROGRESS)) {
 		prepare_to_wait(&ap->eh_wait_q, &wait, TASK_UNINTERRUPTIBLE);
-		spin_unlock_irqrestore(ap->lock, flags);
+		spin_unlock_irq(ap->lock);
 		schedule();
-		spin_lock_irqsave(ap->lock, flags);
+		spin_lock_irq(ap->lock);
 	}
 	finish_wait(&ap->eh_wait_q, &wait);
 
-	spin_unlock_irqrestore(ap->lock, flags);
+	spin_unlock_irq(ap->lock);
 
 	/* make sure SCSI EH is complete */
 	if (scsi_host_in_recovery(ap->scsi_host)) {

Tejun Heo

2013-09-26 13:32 UTC #1 lore raw

Hello, Kirill.

Quoted 5 lines · Kirill Tkhai

Hi,
ata_port_wait_eh() uses spin_lock_irqsave() and this can confuse in fact that it is suitable to use in irqs_disabled() context. But we can't.

Quoted 7 lines

(schedule() returns with interrupts enabled so it's possible ata_port_wait_eh() enters with disabled interrupts but returns with enabled)
So, replace irqsave to unconditional closing of interrupts.
I propose to consider to add patch like this. (If you don't have a magic with flags which is not obvious for me :)

Hmmm... yeah, this was Jeff's preference, at least way back, so libata has a lot of spin_lock_irqsave()'s where spin_lock_irq() should do. At this point, I don't really mind either way but if you wanna change it can you please do a full sweep through libata? While I don't mind either state too much, I do want them to be mostly consistent one way or the other.

Thanks.

Signature · 1 line

tejun

Kirill Tkhai

author

2013-09-26 13:43 UTC #2 lore raw

26.09.2013, 17:32, "Tejun Heo" <[email protected]>:

Quoted 17 lines

Hello, Kirill.
Quoted 5 lines · Kirill Tkhai
Hi,
ata_port_wait_eh() uses spin_lock_irqsave() and this can confuse in fact that it is suitable to use in irqs_disabled() context. But we can't.
Quoted 7 lines
(schedule() returns with interrupts enabled so it's possible ata_port_wait_eh() enters with disabled interrupts but returns with enabled)
So, replace irqsave to unconditional closing of interrupts.
I propose to consider to add patch like this. (If you don't have a magic with flags which is not obvious for me :)

Quoted 6 lines

Hmmm... yeah, this was Jeff's preference, at least way back, so libata has a lot of spin_lock_irqsave()'s where spin_lock_irq() should do. At this point, I don't really mind either way but if you wanna change it can you please do a full sweep through libata? While I don't mind either state too much, I do want them to be mostly consistent one way or the other.

I don't exclude this, maybe, I do this in my spare time ;)

Kirill