* [PATCH 2/2] libata: disable preemption during PIO copy
@ 2010-03-21 21:53 Sebastian Andrzej Siewior
2010-03-23 9:49 ` FUJITA Tomonori
0 siblings, 1 reply; 5+ messages in thread
From: Sebastian Andrzej Siewior @ 2010-03-21 21:53 UTC (permalink / raw)
To: Jeff Garzik; +Cc: linux-ide
If the request has been made in user context it could be moved to a
different CPU on a SMP machine between the copy and cache flush. Thus we
could flush the dcache on the wrong CPU.
Signed-off-by: Sebastian Andrzej Siewior <sebastian@breakpoint.cc>
---
drivers/ata/libata-sff.c | 6 ++++++
1 files changed, 6 insertions(+), 0 deletions(-)
diff --git a/drivers/ata/libata-sff.c b/drivers/ata/libata-sff.c
index 83ecf48..dca9f90 100644
--- a/drivers/ata/libata-sff.c
+++ b/drivers/ata/libata-sff.c
@@ -874,6 +874,9 @@ static void ata_pio_sector(struct ata_queued_cmd *qc)
DPRINTK("data %s\n", qc->tf.flags & ATA_TFLAG_WRITE ? "write" : "read");
+#ifdef CONFIG_SMP
+ preempt_disable();
+#endif
if (PageHighMem(page)) {
unsigned long flags;
@@ -896,6 +899,9 @@ static void ata_pio_sector(struct ata_queued_cmd *qc)
if (!do_write && !PageSlab(page))
flush_dcache_page(page);
+#ifdef CONFIG_SMP
+ preempt_enable();
+#endif
qc->curbytes += qc->sect_size;
qc->cursg_ofs += qc->sect_size;
--
1.6.6
^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: [PATCH 2/2] libata: disable preemption during PIO copy
2010-03-21 21:53 [PATCH 2/2] libata: disable preemption during PIO copy Sebastian Andrzej Siewior
@ 2010-03-23 9:49 ` FUJITA Tomonori
2010-03-23 11:01 ` Sebastian Andrzej Siewior
0 siblings, 1 reply; 5+ messages in thread
From: FUJITA Tomonori @ 2010-03-23 9:49 UTC (permalink / raw)
To: sebastian; +Cc: jgarzik, linux-ide
On Sun, 21 Mar 2010 22:53:08 +0100
Sebastian Andrzej Siewior <sebastian@breakpoint.cc> wrote:
> If the request has been made in user context it could be moved to a
> different CPU on a SMP machine between the copy and cache flush. Thus we
> could flush the dcache on the wrong CPU.
This issue should be addressed in flush_dcache_page() inside?
> Signed-off-by: Sebastian Andrzej Siewior <sebastian@breakpoint.cc>
> ---
> drivers/ata/libata-sff.c | 6 ++++++
> 1 files changed, 6 insertions(+), 0 deletions(-)
>
> diff --git a/drivers/ata/libata-sff.c b/drivers/ata/libata-sff.c
> index 83ecf48..dca9f90 100644
> --- a/drivers/ata/libata-sff.c
> +++ b/drivers/ata/libata-sff.c
> @@ -874,6 +874,9 @@ static void ata_pio_sector(struct ata_queued_cmd *qc)
>
> DPRINTK("data %s\n", qc->tf.flags & ATA_TFLAG_WRITE ? "write" : "read");
>
> +#ifdef CONFIG_SMP
> + preempt_disable();
> +#endif
> if (PageHighMem(page)) {
> unsigned long flags;
>
> @@ -896,6 +899,9 @@ static void ata_pio_sector(struct ata_queued_cmd *qc)
> if (!do_write && !PageSlab(page))
> flush_dcache_page(page);
>
> +#ifdef CONFIG_SMP
> + preempt_enable();
> +#endif
> qc->curbytes += qc->sect_size;
> qc->cursg_ofs += qc->sect_size;
>
> --
> 1.6.6
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-ide" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH 2/2] libata: disable preemption during PIO copy
2010-03-23 9:49 ` FUJITA Tomonori
@ 2010-03-23 11:01 ` Sebastian Andrzej Siewior
2010-03-23 11:51 ` FUJITA Tomonori
0 siblings, 1 reply; 5+ messages in thread
From: Sebastian Andrzej Siewior @ 2010-03-23 11:01 UTC (permalink / raw)
To: FUJITA Tomonori; +Cc: jgarzik, linux-ide
* FUJITA Tomonori | 2010-03-23 18:49:04 [+0900]:
>On Sun, 21 Mar 2010 22:53:08 +0100
>Sebastian Andrzej Siewior <sebastian@breakpoint.cc> wrote:
>
>> If the request has been made in user context it could be moved to a
>> different CPU on a SMP machine between the copy and cache flush. Thus we
>> could flush the dcache on the wrong CPU.
>
>This issue should be addressed in flush_dcache_page() inside?
How so? You have to remain on the same CPU on which you started the copy
process. Once you get to flush_dcache_page() you may have allready
switched the CPU.
Sebastian
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH 2/2] libata: disable preemption during PIO copy
2010-03-23 11:01 ` Sebastian Andrzej Siewior
@ 2010-03-23 11:51 ` FUJITA Tomonori
2010-03-24 21:52 ` Sebastian Andrzej Siewior
0 siblings, 1 reply; 5+ messages in thread
From: FUJITA Tomonori @ 2010-03-23 11:51 UTC (permalink / raw)
To: sebastian; +Cc: fujita.tomonori, jgarzik, linux-ide
On Tue, 23 Mar 2010 12:01:24 +0100
Sebastian Andrzej Siewior <sebastian@breakpoint.cc> wrote:
> * FUJITA Tomonori | 2010-03-23 18:49:04 [+0900]:
>
> >On Sun, 21 Mar 2010 22:53:08 +0100
> >Sebastian Andrzej Siewior <sebastian@breakpoint.cc> wrote:
> >
> >> If the request has been made in user context it could be moved to a
> >> different CPU on a SMP machine between the copy and cache flush. Thus we
> >> could flush the dcache on the wrong CPU.
> >
> >This issue should be addressed in flush_dcache_page() inside?
>
> How so? You have to remain on the same CPU on which you started the copy
> process. Once you get to flush_dcache_page() you may have allready
> switched the CPU.
Oops, I thought that you trying to address on the deferred
flush_dcache_page() issue. I don't know enough about how this pio code
works, but if your description is correct, then the patch is fine, I
guess.
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH 2/2] libata: disable preemption during PIO copy
2010-03-23 11:51 ` FUJITA Tomonori
@ 2010-03-24 21:52 ` Sebastian Andrzej Siewior
0 siblings, 0 replies; 5+ messages in thread
From: Sebastian Andrzej Siewior @ 2010-03-24 21:52 UTC (permalink / raw)
To: FUJITA Tomonori; +Cc: sebastian, jgarzik, linux-ide
* FUJITA Tomonori | 2010-03-23 20:51:34 [+0900]:
>> How so? You have to remain on the same CPU on which you started the copy
>> process. Once you get to flush_dcache_page() you may have allready
>> switched the CPU.
>
>Oops, I thought that you trying to address on the deferred
>flush_dcache_page() issue. I don't know enough about how this pio code
>works, but if your description is correct, then the patch is fine, I
>guess.
No, it is not. The dcache flush is a global one so a CPU switch does not
matter. So I NAK that one myself.
Sebastian
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2010-03-24 21:52 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2010-03-21 21:53 [PATCH 2/2] libata: disable preemption during PIO copy Sebastian Andrzej Siewior
2010-03-23 9:49 ` FUJITA Tomonori
2010-03-23 11:01 ` Sebastian Andrzej Siewior
2010-03-23 11:51 ` FUJITA Tomonori
2010-03-24 21:52 ` Sebastian Andrzej Siewior
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.