Message ID | 1323862781-3465-4-git-send-email-dave.martin@linaro.org |
---|---|
State | Superseded |
Headers | show |
* Dave Martin <dave.martin@linaro.org> [111214 03:08]: > If running in the Normal World on a TrustZone-enabled SoC, Linux > does not have complete control over the L2 cache controller > configuration. The kernel cannot work reliably on such platforms > without the l2x0 cache support code built in. There are HS and GP omaps (High Security and General Purpose). GP omaps do have full control of the L2. Also HS omaps most likely provide control over enabling and disabling L2 depending how the secure code is implemented. BTW, the real problem is that because the secure code is implemented in various ways, we don't really have any handling for it in Linux. The SMI instruction numbers don't seem to be standardized at all, and can mean different things on different boards, even different board versions :( Sounds like devicetree is the only safe way to deal with the L2 control options. Regards, Tony > This patch unconditionally enables l2x0 support for the OMAP4 SoCs. > > Thanks to Rob Herring for this suggestion. [1] > > Signed-off-by: Dave Martin <dave.martin@linaro.org> > > [1] http://lists.infradead.org/pipermail/linux-arm-kernel/2011-November/074495.html > --- > arch/arm/mach-omap2/Kconfig | 2 +- > 1 files changed, 1 insertions(+), 1 deletions(-) > > diff --git a/arch/arm/mach-omap2/Kconfig b/arch/arm/mach-omap2/Kconfig > index bb1b670..94e568a 100644 > --- a/arch/arm/mach-omap2/Kconfig > +++ b/arch/arm/mach-omap2/Kconfig > @@ -41,11 +41,11 @@ config ARCH_OMAP4 > bool "TI OMAP4" > default y > depends on ARCH_OMAP2PLUS > + select CACHE_L2X0 > select CPU_V7 > select ARM_GIC > select HAVE_SMP > select LOCAL_TIMERS if SMP > - select MIGHT_HAVE_CACHE_L2X0 > select PL310_ERRATA_588369 > select PL310_ERRATA_727915 > select ARM_ERRATA_720789 > -- > 1.7.4.1 >
On Wed, Dec 14, 2011 at 10:14:25AM -0800, Tony Lindgren wrote: > * Dave Martin <dave.martin@linaro.org> [111214 03:08]: > > If running in the Normal World on a TrustZone-enabled SoC, Linux > > does not have complete control over the L2 cache controller > > configuration. The kernel cannot work reliably on such platforms > > without the l2x0 cache support code built in. > > There are HS and GP omaps (High Security and General Purpose). > GP omaps do have full control of the L2. Also HS omaps most likely > provide control over enabling and disabling L2 depending how the > secure code is implemented. > > BTW, the real problem is that because the secure code is implemented > in various ways, we don't really have any handling for it in Linux. > > The SMI instruction numbers don't seem to be standardized at all, > and can mean different things on different boards, even different > board versions :( > > Sounds like devicetree is the only safe way to deal with the L2 > control options. > > Regards, > > Tony > > > > This patch unconditionally enables l2x0 support for the OMAP4 SoCs. > > > > Thanks to Rob Herring for this suggestion. [1] > > > > Signed-off-by: Dave Martin <dave.martin@linaro.org> > > > > [1] http://lists.infradead.org/pipermail/linux-arm-kernel/2011-November/074495.html > > --- > > arch/arm/mach-omap2/Kconfig | 2 +- > > 1 files changed, 1 insertions(+), 1 deletions(-) > > > > diff --git a/arch/arm/mach-omap2/Kconfig b/arch/arm/mach-omap2/Kconfig > > index bb1b670..94e568a 100644 > > --- a/arch/arm/mach-omap2/Kconfig > > +++ b/arch/arm/mach-omap2/Kconfig > > @@ -41,11 +41,11 @@ config ARCH_OMAP4 > > bool "TI OMAP4" > > default y > > depends on ARCH_OMAP2PLUS > > + select CACHE_L2X0 > > select CPU_V7 > > select ARM_GIC > > select HAVE_SMP > > select LOCAL_TIMERS if SMP > > - select MIGHT_HAVE_CACHE_L2X0 > > select PL310_ERRATA_588369 > > select PL310_ERRATA_727915 > > select ARM_ERRATA_720789 > > -- > > 1.7.4.1 > > To clarify, are you suggesting we retain this patch, or not? (If we only know what l2x0 support will be needed once the dts is parsed at runtime, there could be an argument for keeping the select CACHE_L2X0 here -- unless we have specific kconfigs for the different security variants of omap.) Cheers ---Dave
* Dave Martin <dave.martin@linaro.org> [111214 09:58]: > On Wed, Dec 14, 2011 at 10:14:25AM -0800, Tony Lindgren wrote: > > * Dave Martin <dave.martin@linaro.org> [111214 03:08]: > > > If running in the Normal World on a TrustZone-enabled SoC, Linux > > > does not have complete control over the L2 cache controller > > > configuration. The kernel cannot work reliably on such platforms > > > without the l2x0 cache support code built in. > > > > There are HS and GP omaps (High Security and General Purpose). > > GP omaps do have full control of the L2. Also HS omaps most likely > > provide control over enabling and disabling L2 depending how the > > secure code is implemented. > > > > BTW, the real problem is that because the secure code is implemented > > in various ways, we don't really have any handling for it in Linux. > > > > The SMI instruction numbers don't seem to be standardized at all, > > and can mean different things on different boards, even different > > board versions :( > > > > Sounds like devicetree is the only safe way to deal with the L2 > > control options. > > > > Regards, > > > > Tony > > > > > > > This patch unconditionally enables l2x0 support for the OMAP4 SoCs. > > > > > > Thanks to Rob Herring for this suggestion. [1] > > > > > > Signed-off-by: Dave Martin <dave.martin@linaro.org> > > > > > > [1] http://lists.infradead.org/pipermail/linux-arm-kernel/2011-November/074495.html > > > --- > > > arch/arm/mach-omap2/Kconfig | 2 +- > > > 1 files changed, 1 insertions(+), 1 deletions(-) > > > > > > diff --git a/arch/arm/mach-omap2/Kconfig b/arch/arm/mach-omap2/Kconfig > > > index bb1b670..94e568a 100644 > > > --- a/arch/arm/mach-omap2/Kconfig > > > +++ b/arch/arm/mach-omap2/Kconfig > > > @@ -41,11 +41,11 @@ config ARCH_OMAP4 > > > bool "TI OMAP4" > > > default y > > > depends on ARCH_OMAP2PLUS > > > + select CACHE_L2X0 > > > select CPU_V7 > > > select ARM_GIC > > > select HAVE_SMP > > > select LOCAL_TIMERS if SMP > > > - select MIGHT_HAVE_CACHE_L2X0 > > > select PL310_ERRATA_588369 > > > select PL310_ERRATA_727915 > > > select ARM_ERRATA_720789 > > > -- > > > 1.7.4.1 > > > > > To clarify, are you suggesting we retain this patch, or not? I think we should keep L2 configurable for omaps until we have some way of getting the configuration dynamically or from device tree. > (If we only know what l2x0 support will be needed once the dts is > parsed at runtime, there could be an argument for keeping the > select CACHE_L2X0 here -- unless we have specific kconfigs for > the different security variants of omap.) Well we can detect if it's an HS omap, but we may not know what SMI it uses for enabling and disabling L2.. That can depend on the board version. Is there some problem keeping MIGHT_HAVE_CACHE_L2X0? This is pretty important from debugging cache issues point of view. Regards, Tony
On 12/14/2011 12:39 PM, Tony Lindgren wrote: > * Dave Martin <dave.martin@linaro.org> [111214 09:58]: >> On Wed, Dec 14, 2011 at 10:14:25AM -0800, Tony Lindgren wrote: >>> * Dave Martin <dave.martin@linaro.org> [111214 03:08]: >>>> If running in the Normal World on a TrustZone-enabled SoC, Linux >>>> does not have complete control over the L2 cache controller >>>> configuration. The kernel cannot work reliably on such platforms >>>> without the l2x0 cache support code built in. >>> >>> There are HS and GP omaps (High Security and General Purpose). >>> GP omaps do have full control of the L2. Also HS omaps most likely >>> provide control over enabling and disabling L2 depending how the >>> secure code is implemented. >>> >>> BTW, the real problem is that because the secure code is implemented >>> in various ways, we don't really have any handling for it in Linux. >>> >>> The SMI instruction numbers don't seem to be standardized at all, >>> and can mean different things on different boards, even different >>> board versions :( >>> >>> Sounds like devicetree is the only safe way to deal with the L2 >>> control options. >>> >>> Regards, >>> >>> Tony >>> >>> >>>> This patch unconditionally enables l2x0 support for the OMAP4 SoCs. >>>> >>>> Thanks to Rob Herring for this suggestion. [1] >>>> >>>> Signed-off-by: Dave Martin <dave.martin@linaro.org> >>>> >>>> [1] http://lists.infradead.org/pipermail/linux-arm-kernel/2011-November/074495.html >>>> --- >>>> arch/arm/mach-omap2/Kconfig | 2 +- >>>> 1 files changed, 1 insertions(+), 1 deletions(-) >>>> >>>> diff --git a/arch/arm/mach-omap2/Kconfig b/arch/arm/mach-omap2/Kconfig >>>> index bb1b670..94e568a 100644 >>>> --- a/arch/arm/mach-omap2/Kconfig >>>> +++ b/arch/arm/mach-omap2/Kconfig >>>> @@ -41,11 +41,11 @@ config ARCH_OMAP4 >>>> bool "TI OMAP4" >>>> default y >>>> depends on ARCH_OMAP2PLUS >>>> + select CACHE_L2X0 >>>> select CPU_V7 >>>> select ARM_GIC >>>> select HAVE_SMP >>>> select LOCAL_TIMERS if SMP >>>> - select MIGHT_HAVE_CACHE_L2X0 >>>> select PL310_ERRATA_588369 >>>> select PL310_ERRATA_727915 >>>> select ARM_ERRATA_720789 >>>> -- >>>> 1.7.4.1 >>>> >> >> To clarify, are you suggesting we retain this patch, or not? > > I think we should keep L2 configurable for omaps until we have some > way of getting the configuration dynamically or from device tree. > This already exists with l2x0_of_init. OMAP just needs to start using it. It will initialize the l2 if present in the DT and skip it if not present. Rob >> (If we only know what l2x0 support will be needed once the dts is >> parsed at runtime, there could be an argument for keeping the >> select CACHE_L2X0 here -- unless we have specific kconfigs for >> the different security variants of omap.) > > Well we can detect if it's an HS omap, but we may not know what > SMI it uses for enabling and disabling L2.. That can depend on > the board version. > > Is there some problem keeping MIGHT_HAVE_CACHE_L2X0? This is > pretty important from debugging cache issues point of view. > > Regards, > > Tony
* Rob Herring <robherring2@gmail.com> [111214 12:30]: > > On 12/14/2011 12:39 PM, Tony Lindgren wrote: > > > > I think we should keep L2 configurable for omaps until we have some > > way of getting the configuration dynamically or from device tree. > > > > This already exists with l2x0_of_init. OMAP just needs to start using > it. It will initialize the l2 if present in the DT and skip it if not > present. That's great, that will allow doing the right thing with setting up how to enable and disable it :) Considering that, this patch should be OK to apply: Acked-by: Tony Lindgren <tony@atomide.com>
diff --git a/arch/arm/mach-omap2/Kconfig b/arch/arm/mach-omap2/Kconfig index bb1b670..94e568a 100644 --- a/arch/arm/mach-omap2/Kconfig +++ b/arch/arm/mach-omap2/Kconfig @@ -41,11 +41,11 @@ config ARCH_OMAP4 bool "TI OMAP4" default y depends on ARCH_OMAP2PLUS + select CACHE_L2X0 select CPU_V7 select ARM_GIC select HAVE_SMP select LOCAL_TIMERS if SMP - select MIGHT_HAVE_CACHE_L2X0 select PL310_ERRATA_588369 select PL310_ERRATA_727915 select ARM_ERRATA_720789
If running in the Normal World on a TrustZone-enabled SoC, Linux does not have complete control over the L2 cache controller configuration. The kernel cannot work reliably on such platforms without the l2x0 cache support code built in. This patch unconditionally enables l2x0 support for the OMAP4 SoCs. Thanks to Rob Herring for this suggestion. [1] Signed-off-by: Dave Martin <dave.martin@linaro.org> [1] http://lists.infradead.org/pipermail/linux-arm-kernel/2011-November/074495.html --- arch/arm/mach-omap2/Kconfig | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)