2012-02-23 20:58:19

by Paul E. McKenney

[permalink] [raw]
Subject: [PATCH RFC] acpi: Fix unprotected smp_processor_id() in acpi_processor_cst_has_changed()

The acpi_processor_cst_has_changed() function is invoked from a
CPU_ONLINE or CPU_DEAD function, which might well execute on CPU 0
even though the CPU being hotplugged is some other CPU. In addition,
acpi_processor_cst_has_changed() invokes smp_processor_id() without
protection, resulting in splats when onlining CPUs.

This commit therefore changes the smp_processor_id() to pr->id, as is
used elsewhere in the code, for example, in acpi_processor_add().

This works for me, but I don't claim to understand ACPI.

Signed-off-by: Paul E. McKenney <[email protected]>

diff --git a/drivers/acpi/processor_idle.c b/drivers/acpi/processor_idle.c
index 0e8e2de..9e57b06 100644
--- a/drivers/acpi/processor_idle.c
+++ b/drivers/acpi/processor_idle.c
@@ -1159,8 +1159,7 @@ int acpi_processor_cst_has_changed(struct acpi_processor *pr)
* to make the code that updates C-States be called once.
*/

- if (smp_processor_id() == 0 &&
- cpuidle_get_driver() == &acpi_idle_driver) {
+ if (pr->id == 0 && cpuidle_get_driver() == &acpi_idle_driver) {

cpuidle_pause_and_lock();
/* Protect against cpu-hotplug */


2012-02-28 06:13:46

by Srivatsa S. Bhat

[permalink] [raw]
Subject: Re: [PATCH RFC] acpi: Fix unprotected smp_processor_id() in acpi_processor_cst_has_changed()

On 02/24/2012 02:28 AM, Paul E. McKenney wrote:

> The acpi_processor_cst_has_changed() function is invoked from a
> CPU_ONLINE or CPU_DEAD function, which might well execute on CPU 0
> even though the CPU being hotplugged is some other CPU. In addition,
> acpi_processor_cst_has_changed() invokes smp_processor_id() without
> protection, resulting in splats when onlining CPUs.
>
> This commit therefore changes the smp_processor_id() to pr->id, as is
> used elsewhere in the code, for example, in acpi_processor_add().
>
> This works for me, but I don't claim to understand ACPI.
>
> Signed-off-by: Paul E. McKenney <[email protected]>
>


This looks like the right fix to me, observing the disparity between what
the comment says (and hence what was intended) vs what the code originally
did.

Reviewed-by: Srivatsa S. Bhat <[email protected]>


> diff --git a/drivers/acpi/processor_idle.c b/drivers/acpi/processor_idle.c
> index 0e8e2de..9e57b06 100644
> --- a/drivers/acpi/processor_idle.c
> +++ b/drivers/acpi/processor_idle.c
> @@ -1159,8 +1159,7 @@ int acpi_processor_cst_has_changed(struct acpi_processor *pr)
> * to make the code that updates C-States be called once.
> */
>
> - if (smp_processor_id() == 0 &&
> - cpuidle_get_driver() == &acpi_idle_driver) {
> + if (pr->id == 0 && cpuidle_get_driver() == &acpi_idle_driver) {
>
> cpuidle_pause_and_lock();
> /* Protect against cpu-hotplug */
>


Regards,
Srivatsa S. Bhat

2012-02-28 07:19:56

by Yong Zhang

[permalink] [raw]
Subject: Re: [PATCH RFC] acpi: Fix unprotected smp_processor_id() in acpi_processor_cst_has_changed()

On Thu, Feb 23, 2012 at 12:58:05PM -0800, Paul E. McKenney wrote:
> The acpi_processor_cst_has_changed() function is invoked from a
> CPU_ONLINE or CPU_DEAD function, which might well execute on CPU 0
> even though the CPU being hotplugged is some other CPU. In addition,
> acpi_processor_cst_has_changed() invokes smp_processor_id() without
> protection, resulting in splats when onlining CPUs.
>
> This commit therefore changes the smp_processor_id() to pr->id, as is
> used elsewhere in the code, for example, in acpi_processor_add().
>
> This works for me, but I don't claim to understand ACPI.
>
> Signed-off-by: Paul E. McKenney <[email protected]>

Got the same warning.
And the patch works on my side.

Tested-by: Yong Zhang <[email protected]>

>
> diff --git a/drivers/acpi/processor_idle.c b/drivers/acpi/processor_idle.c
> index 0e8e2de..9e57b06 100644
> --- a/drivers/acpi/processor_idle.c
> +++ b/drivers/acpi/processor_idle.c
> @@ -1159,8 +1159,7 @@ int acpi_processor_cst_has_changed(struct acpi_processor *pr)
> * to make the code that updates C-States be called once.
> */
>
> - if (smp_processor_id() == 0 &&
> - cpuidle_get_driver() == &acpi_idle_driver) {
> + if (pr->id == 0 && cpuidle_get_driver() == &acpi_idle_driver) {
>
> cpuidle_pause_and_lock();
> /* Protect against cpu-hotplug */
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/

--
Only stand for myself

2012-02-28 13:30:42

by Paul E. McKenney

[permalink] [raw]
Subject: Re: [PATCH RFC] acpi: Fix unprotected smp_processor_id() in acpi_processor_cst_has_changed()

On Tue, Feb 28, 2012 at 03:19:43PM +0800, Yong Zhang wrote:
> On Thu, Feb 23, 2012 at 12:58:05PM -0800, Paul E. McKenney wrote:
> > The acpi_processor_cst_has_changed() function is invoked from a
> > CPU_ONLINE or CPU_DEAD function, which might well execute on CPU 0
> > even though the CPU being hotplugged is some other CPU. In addition,
> > acpi_processor_cst_has_changed() invokes smp_processor_id() without
> > protection, resulting in splats when onlining CPUs.
> >
> > This commit therefore changes the smp_processor_id() to pr->id, as is
> > used elsewhere in the code, for example, in acpi_processor_add().
> >
> > This works for me, but I don't claim to understand ACPI.
> >
> > Signed-off-by: Paul E. McKenney <[email protected]>
>
> Got the same warning.
> And the patch works on my side.
>
> Tested-by: Yong Zhang <[email protected]>

Thank you, I have added your Tested-by (and Srivatsa's Reviewed-by,
for that matter).

Any takers for this, or should I just push it up the -tip tree?

Thanx, Paul

> > diff --git a/drivers/acpi/processor_idle.c b/drivers/acpi/processor_idle.c
> > index 0e8e2de..9e57b06 100644
> > --- a/drivers/acpi/processor_idle.c
> > +++ b/drivers/acpi/processor_idle.c
> > @@ -1159,8 +1159,7 @@ int acpi_processor_cst_has_changed(struct acpi_processor *pr)
> > * to make the code that updates C-States be called once.
> > */
> >
> > - if (smp_processor_id() == 0 &&
> > - cpuidle_get_driver() == &acpi_idle_driver) {
> > + if (pr->id == 0 && cpuidle_get_driver() == &acpi_idle_driver) {
> >
> > cpuidle_pause_and_lock();
> > /* Protect against cpu-hotplug */
> >
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> > the body of a message to [email protected]
> > More majordomo info at http://vger.kernel.org/majordomo-info.html
> > Please read the FAQ at http://www.tux.org/lkml/
>
> --
> Only stand for myself
>

2012-02-28 20:53:53

by Rafael J. Wysocki

[permalink] [raw]
Subject: Re: [PATCH RFC] acpi: Fix unprotected smp_processor_id() in acpi_processor_cst_has_changed()

On Tuesday, February 28, 2012, Paul E. McKenney wrote:
> On Tue, Feb 28, 2012 at 03:19:43PM +0800, Yong Zhang wrote:
> > On Thu, Feb 23, 2012 at 12:58:05PM -0800, Paul E. McKenney wrote:
> > > The acpi_processor_cst_has_changed() function is invoked from a
> > > CPU_ONLINE or CPU_DEAD function, which might well execute on CPU 0
> > > even though the CPU being hotplugged is some other CPU. In addition,
> > > acpi_processor_cst_has_changed() invokes smp_processor_id() without
> > > protection, resulting in splats when onlining CPUs.
> > >
> > > This commit therefore changes the smp_processor_id() to pr->id, as is
> > > used elsewhere in the code, for example, in acpi_processor_add().
> > >
> > > This works for me, but I don't claim to understand ACPI.
> > >
> > > Signed-off-by: Paul E. McKenney <[email protected]>
> >
> > Got the same warning.
> > And the patch works on my side.
> >
> > Tested-by: Yong Zhang <[email protected]>
>
> Thank you, I have added your Tested-by (and Srivatsa's Reviewed-by,
> for that matter).
>
> Any takers for this, or should I just push it up the -tip tree?

I think this patch is for Len, either through ACPI, or through the idle tree.
Len?

Anyway:

Acked-by: Rafael J. Wysocki <[email protected]>

Thanks,
Rafael


> > > diff --git a/drivers/acpi/processor_idle.c b/drivers/acpi/processor_idle.c
> > > index 0e8e2de..9e57b06 100644
> > > --- a/drivers/acpi/processor_idle.c
> > > +++ b/drivers/acpi/processor_idle.c
> > > @@ -1159,8 +1159,7 @@ int acpi_processor_cst_has_changed(struct acpi_processor *pr)
> > > * to make the code that updates C-States be called once.
> > > */
> > >
> > > - if (smp_processor_id() == 0 &&
> > > - cpuidle_get_driver() == &acpi_idle_driver) {
> > > + if (pr->id == 0 && cpuidle_get_driver() == &acpi_idle_driver) {
> > >
> > > cpuidle_pause_and_lock();
> > > /* Protect against cpu-hotplug */
> > >
> > > --
> > > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> > > the body of a message to [email protected]
> > > More majordomo info at http://vger.kernel.org/majordomo-info.html
> > > Please read the FAQ at http://www.tux.org/lkml/
> >
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
>
>

2012-02-28 21:28:15

by Paul E. McKenney

[permalink] [raw]
Subject: Re: [PATCH RFC] acpi: Fix unprotected smp_processor_id() in acpi_processor_cst_has_changed()

On Tue, Feb 28, 2012 at 09:57:53PM +0100, Rafael J. Wysocki wrote:
> On Tuesday, February 28, 2012, Paul E. McKenney wrote:
> > On Tue, Feb 28, 2012 at 03:19:43PM +0800, Yong Zhang wrote:
> > > On Thu, Feb 23, 2012 at 12:58:05PM -0800, Paul E. McKenney wrote:
> > > > The acpi_processor_cst_has_changed() function is invoked from a
> > > > CPU_ONLINE or CPU_DEAD function, which might well execute on CPU 0
> > > > even though the CPU being hotplugged is some other CPU. In addition,
> > > > acpi_processor_cst_has_changed() invokes smp_processor_id() without
> > > > protection, resulting in splats when onlining CPUs.
> > > >
> > > > This commit therefore changes the smp_processor_id() to pr->id, as is
> > > > used elsewhere in the code, for example, in acpi_processor_add().
> > > >
> > > > This works for me, but I don't claim to understand ACPI.
> > > >
> > > > Signed-off-by: Paul E. McKenney <[email protected]>
> > >
> > > Got the same warning.
> > > And the patch works on my side.
> > >
> > > Tested-by: Yong Zhang <[email protected]>
> >
> > Thank you, I have added your Tested-by (and Srivatsa's Reviewed-by,
> > for that matter).
> >
> > Any takers for this, or should I just push it up the -tip tree?
>
> I think this patch is for Len, either through ACPI, or through the idle tree.
> Len?

Whichever way it is supposed to go, it reminds me of the need for it every
time I forget to merge it in before testing. ;-)

> Anyway:
>
> Acked-by: Rafael J. Wysocki <[email protected]>

Thank you, applied! The patch appears below with the recent review,
testing, and ack.

Thanx, Paul

------------------------------------------------------------------------

acpi: Fix unprotected smp_processor_id() in acpi_processor_cst_has_changed()

The acpi_processor_cst_has_changed() function is invoked from a
CPU_ONLINE or CPU_DEAD function, which might well execute on CPU 0
even though the CPU being hotplugged is some other CPU. In addition,
acpi_processor_cst_has_changed() invokes smp_processor_id() without
protection, resulting in splats when onlining CPUs.

This commit therefore changes the smp_processor_id() to pr->id, as is
used elsewhere in the code, for example, in acpi_processor_add().

Signed-off-by: Paul E. McKenney <[email protected]>
Signed-off-by: Paul E. McKenney <[email protected]>
Reviewed-by: Srivatsa S. Bhat <[email protected]>
Tested-by: Yong Zhang <[email protected]>
Acked-by: Rafael J. Wysocki <[email protected]>

diff --git a/drivers/acpi/processor_idle.c b/drivers/acpi/processor_idle.c
index 0e8e2de..9e57b06 100644
--- a/drivers/acpi/processor_idle.c
+++ b/drivers/acpi/processor_idle.c
@@ -1159,8 +1159,7 @@ int acpi_processor_cst_has_changed(struct acpi_processor *pr)
* to make the code that updates C-States be called once.
*/

- if (smp_processor_id() == 0 &&
- cpuidle_get_driver() == &acpi_idle_driver) {
+ if (pr->id == 0 && cpuidle_get_driver() == &acpi_idle_driver) {

cpuidle_pause_and_lock();
/* Protect against cpu-hotplug */