* Handling ARM_ERRATA_364296 when CONFIG_SMP=y
@ 2013-07-08 18:16 Fabio Estevam
2013-07-08 18:22 ` Willy Tarreau
2013-07-12 15:14 ` Catalin Marinas
0 siblings, 2 replies; 6+ messages in thread
From: Fabio Estevam @ 2013-07-08 18:16 UTC (permalink / raw)
To: linux-arm-kernel
Hi Catalin,
On commit 145e10e1 (ARM: 7015/1: ARM errata: Possible cache data
corruption with hit-under-miss enabled) you added support for
ARM_ERRATA_364296 option:
+config ARM_ERRATA_364296
+ bool "ARM errata: Possible cache data corruption with
hit-under-miss enabled"
+ depends on CPU_V6 && !SMP
+ help
As this erratum affects mx31/mx35, I would like to enable it, but I am
not allowed to do it because mx31/mx35 share the same config file as
mx6 (imx_v6_v7_defconfig), which has SMP turned on.
What would be the correct approach for applying this erratum?
Is it safe to do "depends on CPU_V6" only or should this erratum be
handled in the bootloader?
Regards,
Fabio Estevam
^ permalink raw reply [flat|nested] 6+ messages in thread
* Handling ARM_ERRATA_364296 when CONFIG_SMP=y
2013-07-08 18:16 Handling ARM_ERRATA_364296 when CONFIG_SMP=y Fabio Estevam
@ 2013-07-08 18:22 ` Willy Tarreau
2013-07-08 20:29 ` Fabio Estevam
2013-07-12 15:14 ` Catalin Marinas
1 sibling, 1 reply; 6+ messages in thread
From: Willy Tarreau @ 2013-07-08 18:22 UTC (permalink / raw)
To: linux-arm-kernel
On Mon, Jul 08, 2013 at 03:16:34PM -0300, Fabio Estevam wrote:
> Hi Catalin,
>
> On commit 145e10e1 (ARM: 7015/1: ARM errata: Possible cache data
> corruption with hit-under-miss enabled) you added support for
> ARM_ERRATA_364296 option:
>
> +config ARM_ERRATA_364296
> + bool "ARM errata: Possible cache data corruption with
> hit-under-miss enabled"
> + depends on CPU_V6 && !SMP
> + help
>
> As this erratum affects mx31/mx35, I would like to enable it, but I am
> not allowed to do it because mx31/mx35 share the same config file as
> mx6 (imx_v6_v7_defconfig), which has SMP turned on.
>
> What would be the correct approach for applying this erratum?
>
> Is it safe to do "depends on CPU_V6" only or should this erratum be
> handled in the bootloader?
Or maybe fix the config so that it does not force SMP on a CPU which
only exists as a single core ?
Regards,
Willy
^ permalink raw reply [flat|nested] 6+ messages in thread
* Handling ARM_ERRATA_364296 when CONFIG_SMP=y
2013-07-08 18:22 ` Willy Tarreau
@ 2013-07-08 20:29 ` Fabio Estevam
2013-07-08 21:10 ` Willy Tarreau
0 siblings, 1 reply; 6+ messages in thread
From: Fabio Estevam @ 2013-07-08 20:29 UTC (permalink / raw)
To: linux-arm-kernel
Hi Willy,
On Mon, Jul 8, 2013 at 3:22 PM, Willy Tarreau <w@1wt.eu> wrote:
> Or maybe fix the config so that it does not force SMP on a CPU which
> only exists as a single core ?
The original config is imx_v6_v7_defconfig and it selects by default
both mx6 (quad core) and other single core CPUs.
That's the reason why CONFIG_SMP=y for this config by default.
Regards,
Fabio Estevam
^ permalink raw reply [flat|nested] 6+ messages in thread
* Handling ARM_ERRATA_364296 when CONFIG_SMP=y
2013-07-08 20:29 ` Fabio Estevam
@ 2013-07-08 21:10 ` Willy Tarreau
0 siblings, 0 replies; 6+ messages in thread
From: Willy Tarreau @ 2013-07-08 21:10 UTC (permalink / raw)
To: linux-arm-kernel
Hi Fabio,
On Mon, Jul 08, 2013 at 05:29:34PM -0300, Fabio Estevam wrote:
> Hi Willy,
>
> On Mon, Jul 8, 2013 at 3:22 PM, Willy Tarreau <w@1wt.eu> wrote:
>
> > Or maybe fix the config so that it does not force SMP on a CPU which
> > only exists as a single core ?
>
> The original config is imx_v6_v7_defconfig and it selects by default
> both mx6 (quad core) and other single core CPUs.
>
> That's the reason why CONFIG_SMP=y for this config by default.
OK, so if it's just a default, it's not really a problem, I understand
better now. So probably your solution to make the workaround not depend
on !SMP makes sense.
Regards,
Willy
^ permalink raw reply [flat|nested] 6+ messages in thread
* Handling ARM_ERRATA_364296 when CONFIG_SMP=y
2013-07-08 18:16 Handling ARM_ERRATA_364296 when CONFIG_SMP=y Fabio Estevam
2013-07-08 18:22 ` Willy Tarreau
@ 2013-07-12 15:14 ` Catalin Marinas
2013-07-12 15:19 ` Russell King - ARM Linux
1 sibling, 1 reply; 6+ messages in thread
From: Catalin Marinas @ 2013-07-12 15:14 UTC (permalink / raw)
To: linux-arm-kernel
On Mon, Jul 08, 2013 at 07:16:34PM +0100, Fabio Estevam wrote:
> On commit 145e10e1 (ARM: 7015/1: ARM errata: Possible cache data
> corruption with hit-under-miss enabled) you added support for
> ARM_ERRATA_364296 option:
>
> +config ARM_ERRATA_364296
> + bool "ARM errata: Possible cache data corruption with
> hit-under-miss enabled"
> + depends on CPU_V6 && !SMP
> + help
>
> As this erratum affects mx31/mx35, I would like to enable it, but I am
> not allowed to do it because mx31/mx35 share the same config file as
> mx6 (imx_v6_v7_defconfig), which has SMP turned on.
>
> What would be the correct approach for applying this erratum?
>
> Is it safe to do "depends on CPU_V6" only or should this erratum be
> handled in the bootloader?
I think you can just remove the !SMP case. The workaround (setting an
undocumented bit) already checks for the right CPU Id. The only problem
issue is that oldconfig will ask whether to enable it or not.
Bootloader would be even better but it's pretty late to enforce this on
ARMv6.
--
Catalin
^ permalink raw reply [flat|nested] 6+ messages in thread
* Handling ARM_ERRATA_364296 when CONFIG_SMP=y
2013-07-12 15:14 ` Catalin Marinas
@ 2013-07-12 15:19 ` Russell King - ARM Linux
0 siblings, 0 replies; 6+ messages in thread
From: Russell King - ARM Linux @ 2013-07-12 15:19 UTC (permalink / raw)
To: linux-arm-kernel
On Fri, Jul 12, 2013 at 04:14:35PM +0100, Catalin Marinas wrote:
> On Mon, Jul 08, 2013 at 07:16:34PM +0100, Fabio Estevam wrote:
> > On commit 145e10e1 (ARM: 7015/1: ARM errata: Possible cache data
> > corruption with hit-under-miss enabled) you added support for
> > ARM_ERRATA_364296 option:
> >
> > +config ARM_ERRATA_364296
> > + bool "ARM errata: Possible cache data corruption with
> > hit-under-miss enabled"
> > + depends on CPU_V6 && !SMP
> > + help
> >
> > As this erratum affects mx31/mx35, I would like to enable it, but I am
> > not allowed to do it because mx31/mx35 share the same config file as
> > mx6 (imx_v6_v7_defconfig), which has SMP turned on.
> >
> > What would be the correct approach for applying this erratum?
> >
> > Is it safe to do "depends on CPU_V6" only or should this erratum be
> > handled in the bootloader?
>
> I think you can just remove the !SMP case. The workaround (setting an
> undocumented bit) already checks for the right CPU Id. The only problem
> issue is that oldconfig will ask whether to enable it or not.
Good thing you don't disagree, because it's already on its way to Linus
as of this morning.
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2013-07-12 15:19 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2013-07-08 18:16 Handling ARM_ERRATA_364296 when CONFIG_SMP=y Fabio Estevam
2013-07-08 18:22 ` Willy Tarreau
2013-07-08 20:29 ` Fabio Estevam
2013-07-08 21:10 ` Willy Tarreau
2013-07-12 15:14 ` Catalin Marinas
2013-07-12 15:19 ` Russell King - ARM Linux
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.