| Message ID | 0f0a77ddee48027efa35f36d6c809d2d3ff91ce1.1431064527.git.jan.kiszka@siemens.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Fri, May 08, 2015 at 07:55:26AM +0200, Jan Kiszka wrote: > Ensure that we can use psci_smp_available without checking for > CONFIG_SMP first. > > Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com> > --- > arch/arm/include/asm/psci.h | 6 +++++- > 1 file changed, 5 insertions(+), 1 deletion(-) > > diff --git a/arch/arm/include/asm/psci.h b/arch/arm/include/asm/psci.h > index c25ef3e..eef6a0a 100644 > --- a/arch/arm/include/asm/psci.h > +++ b/arch/arm/include/asm/psci.h > @@ -39,9 +39,13 @@ extern struct smp_operations psci_smp_ops; > > #ifdef CONFIG_ARM_PSCI > int psci_init(void); > -bool psci_smp_available(void); > #else > static inline int psci_init(void) { return 0; } > +#endif > + > +#if defined(CONFIG_ARM_PSCI) && defined(CONFIG_SMP) > +bool psci_smp_available(void); > +#else > static inline bool psci_smp_available(void) { return false; } > #endif > Looks good to me. Russell, would you be willing to ack this patch so that I can take it through the Tegra tree along with patch 2/2 that depends on this? Thierry
On Fri, May 08, 2015 at 01:26:17PM +0200, Thierry Reding wrote: > On Fri, May 08, 2015 at 07:55:26AM +0200, Jan Kiszka wrote: > > Ensure that we can use psci_smp_available without checking for > > CONFIG_SMP first. > > > > Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com> > > --- > > arch/arm/include/asm/psci.h | 6 +++++- > > 1 file changed, 5 insertions(+), 1 deletion(-) > > > > diff --git a/arch/arm/include/asm/psci.h b/arch/arm/include/asm/psci.h > > index c25ef3e..eef6a0a 100644 > > --- a/arch/arm/include/asm/psci.h > > +++ b/arch/arm/include/asm/psci.h > > @@ -39,9 +39,13 @@ extern struct smp_operations psci_smp_ops; > > > > #ifdef CONFIG_ARM_PSCI > > int psci_init(void); > > -bool psci_smp_available(void); > > #else > > static inline int psci_init(void) { return 0; } > > +#endif > > + > > +#if defined(CONFIG_ARM_PSCI) && defined(CONFIG_SMP) > > +bool psci_smp_available(void); > > +#else > > static inline bool psci_smp_available(void) { return false; } > > #endif > > > > Looks good to me. > > Russell, would you be willing to ack this patch so that I can take it > through the Tegra tree along with patch 2/2 that depends on this? I'd prefer to see an ack from Will on this, as it's code which Will contributed and presumably understands better than me.
On Fri, May 08, 2015 at 03:56:20PM +0100, Russell King - ARM Linux wrote: > On Fri, May 08, 2015 at 01:26:17PM +0200, Thierry Reding wrote: > > On Fri, May 08, 2015 at 07:55:26AM +0200, Jan Kiszka wrote: > > > Ensure that we can use psci_smp_available without checking for > > > CONFIG_SMP first. > > > > > > Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com> > > > --- > > > arch/arm/include/asm/psci.h | 6 +++++- > > > 1 file changed, 5 insertions(+), 1 deletion(-) > > > > > > diff --git a/arch/arm/include/asm/psci.h b/arch/arm/include/asm/psci.h > > > index c25ef3e..eef6a0a 100644 > > > --- a/arch/arm/include/asm/psci.h > > > +++ b/arch/arm/include/asm/psci.h > > > @@ -39,9 +39,13 @@ extern struct smp_operations psci_smp_ops; > > > > > > #ifdef CONFIG_ARM_PSCI > > > int psci_init(void); > > > -bool psci_smp_available(void); > > > #else > > > static inline int psci_init(void) { return 0; } > > > +#endif > > > + > > > +#if defined(CONFIG_ARM_PSCI) && defined(CONFIG_SMP) > > > +bool psci_smp_available(void); > > > +#else > > > static inline bool psci_smp_available(void) { return false; } > > > #endif > > > > > > > Looks good to me. > > > > Russell, would you be willing to ack this patch so that I can take it > > through the Tegra tree along with patch 2/2 that depends on this? > > I'd prefer to see an ack from Will on this, as it's code which Will > contributed and presumably understands better than me. Sorry, I'm late to the thread so I'm a tad confused. If I understand correctly, these two patches are trying to avoid using the tegra cpuidle driver if PSCI is available instead? That raises the question as to whether or not you're actually trying to discover whether you're in the secure or non-secure world. I assume PSCI only works if the kernel is running in non-secure mode, but does the Tegra cpuidle driver only work in secure mode? Will
On 2015-05-08 18:14, Will Deacon wrote: > On Fri, May 08, 2015 at 03:56:20PM +0100, Russell King - ARM Linux wrote: >> On Fri, May 08, 2015 at 01:26:17PM +0200, Thierry Reding wrote: >>> On Fri, May 08, 2015 at 07:55:26AM +0200, Jan Kiszka wrote: >>>> Ensure that we can use psci_smp_available without checking for >>>> CONFIG_SMP first. >>>> >>>> Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com> >>>> --- >>>> arch/arm/include/asm/psci.h | 6 +++++- >>>> 1 file changed, 5 insertions(+), 1 deletion(-) >>>> >>>> diff --git a/arch/arm/include/asm/psci.h b/arch/arm/include/asm/psci.h >>>> index c25ef3e..eef6a0a 100644 >>>> --- a/arch/arm/include/asm/psci.h >>>> +++ b/arch/arm/include/asm/psci.h >>>> @@ -39,9 +39,13 @@ extern struct smp_operations psci_smp_ops; >>>> >>>> #ifdef CONFIG_ARM_PSCI >>>> int psci_init(void); >>>> -bool psci_smp_available(void); >>>> #else >>>> static inline int psci_init(void) { return 0; } >>>> +#endif >>>> + >>>> +#if defined(CONFIG_ARM_PSCI) && defined(CONFIG_SMP) >>>> +bool psci_smp_available(void); >>>> +#else >>>> static inline bool psci_smp_available(void) { return false; } >>>> #endif >>>> >>> >>> Looks good to me. >>> >>> Russell, would you be willing to ack this patch so that I can take it >>> through the Tegra tree along with patch 2/2 that depends on this? >> >> I'd prefer to see an ack from Will on this, as it's code which Will >> contributed and presumably understands better than me. > > Sorry, I'm late to the thread so I'm a tad confused. > > If I understand correctly, these two patches are trying to avoid using > the tegra cpuidle driver if PSCI is available instead? That raises the > question as to whether or not you're actually trying to discover whether > you're in the secure or non-secure world. > > I assume PSCI only works if the kernel is running in non-secure mode, > but does the Tegra cpuidle driver only work in secure mode? To my understanding of the hardware, the flow controller can be accessed from both worlds. So, testing the mode would neither be sufficient nor more logical - it's PSCI which occupies this resource, not automatically all secure monitors. Jan
On Sun, May 10, 2015 at 11:48:20AM +0200, Jan Kiszka wrote: > On 2015-05-08 18:14, Will Deacon wrote: > > On Fri, May 08, 2015 at 03:56:20PM +0100, Russell King - ARM Linux wrote: > >> On Fri, May 08, 2015 at 01:26:17PM +0200, Thierry Reding wrote: > >>> On Fri, May 08, 2015 at 07:55:26AM +0200, Jan Kiszka wrote: > >>>> Ensure that we can use psci_smp_available without checking for > >>>> CONFIG_SMP first. > >>>> > >>>> Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com> > >>>> --- > >>>> arch/arm/include/asm/psci.h | 6 +++++- > >>>> 1 file changed, 5 insertions(+), 1 deletion(-) > >>>> > >>>> diff --git a/arch/arm/include/asm/psci.h b/arch/arm/include/asm/psci.h > >>>> index c25ef3e..eef6a0a 100644 > >>>> --- a/arch/arm/include/asm/psci.h > >>>> +++ b/arch/arm/include/asm/psci.h > >>>> @@ -39,9 +39,13 @@ extern struct smp_operations psci_smp_ops; > >>>> > >>>> #ifdef CONFIG_ARM_PSCI > >>>> int psci_init(void); > >>>> -bool psci_smp_available(void); > >>>> #else > >>>> static inline int psci_init(void) { return 0; } > >>>> +#endif > >>>> + > >>>> +#if defined(CONFIG_ARM_PSCI) && defined(CONFIG_SMP) > >>>> +bool psci_smp_available(void); > >>>> +#else > >>>> static inline bool psci_smp_available(void) { return false; } > >>>> #endif > >>>> > >>> > >>> Looks good to me. > >>> > >>> Russell, would you be willing to ack this patch so that I can take it > >>> through the Tegra tree along with patch 2/2 that depends on this? > >> > >> I'd prefer to see an ack from Will on this, as it's code which Will > >> contributed and presumably understands better than me. > > > > Sorry, I'm late to the thread so I'm a tad confused. > > > > If I understand correctly, these two patches are trying to avoid using > > the tegra cpuidle driver if PSCI is available instead? That raises the > > question as to whether or not you're actually trying to discover whether > > you're in the secure or non-secure world. > > > > I assume PSCI only works if the kernel is running in non-secure mode, > > but does the Tegra cpuidle driver only work in secure mode? > > To my understanding of the hardware, the flow controller can be accessed > from both worlds. So, testing the mode would neither be sufficient nor > more logical - it's PSCI which occupies this resource, not automatically > all secure monitors. This is correct. The flow controller doesn't check the NS bit. Cheers, Peter.
diff --git a/arch/arm/include/asm/psci.h b/arch/arm/include/asm/psci.h index c25ef3e..eef6a0a 100644 --- a/arch/arm/include/asm/psci.h +++ b/arch/arm/include/asm/psci.h @@ -39,9 +39,13 @@ extern struct smp_operations psci_smp_ops; #ifdef CONFIG_ARM_PSCI int psci_init(void); -bool psci_smp_available(void); #else static inline int psci_init(void) { return 0; } +#endif + +#if defined(CONFIG_ARM_PSCI) && defined(CONFIG_SMP) +bool psci_smp_available(void); +#else static inline bool psci_smp_available(void) { return false; } #endif
Ensure that we can use psci_smp_available without checking for CONFIG_SMP first. Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com> --- arch/arm/include/asm/psci.h | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-)