[PATCH 0/5] arm64: Allow booting with GICv3 in GICv2 mode

[Marc Zyngier]

On 08/10/15 16:56, Catalin Marinas wrote:
> On Fri, Oct 02, 2015 at 05:37:49PM +0100, Marc Zyngier wrote:
>> Recent evolutions of the ARM Trusted Firmware have outlined issues
>> when the system is equipped with a GICv3 interrupt controller, but the
>> firmware has decided to restrict it to GICv2 compatibility mode.
>>
>> In this mode, system registers cannot be enabled, and the firmware is
>> expected to pass a GICv2 description (DT or ACPI tables).
>>
>> This series makes sure that system register access is checked at EL2
>> setup time and when the feature detection is performed. Additionally,
>> the GICv2 driver checks that system registers are disabled, and warns
>> if they are enabled.
>>
>> The booting requirements are also updated to make the above explicit.
>>
>> Marc Zyngier (5):
>>   arm64: el2_setup: Make sure ICC_SRE_EL2.SRE sticks before using GICv3
>>     sysregs
>>   irqchip/gic-v3: Make gic_enable_sre an inline function
>>   arm64: cpufeatures: Check ICC_EL1_SRE.SRE before enabling
>>     ARM64_HAS_SYSREG_GIC_CPUIF
>>   irqchip/gic: Warn if GICv3 system registers are enabled
>>   arm64: Update booting requirements for GICv3 in GICv2 mode
>>
>>  Documentation/arm64/booting.txt    | 11 ++++++++++-
>>  arch/arm64/kernel/cpufeature.c     | 19 ++++++++++++++++++-
>>  arch/arm64/kernel/head.S           |  2 ++
>>  drivers/irqchip/irq-gic-v3.c       | 32 +++++++++-----------------------
>>  drivers/irqchip/irq-gic.c          | 15 +++++++++++++++
>>  include/linux/irqchip/arm-gic-v3.h | 16 ++++++++++++++++
>>  6 files changed, 70 insertions(+), 25 deletions(-)
> 
> Reviewed-by: Catalin Marinas <catalin.marinas at arm.com>
> 
> How do you plan to merge these patches? I'm fine for them to go via the
> irqchip maintainers since they are all GIC related.

I guess we can direct this via tip if Thomas is OK with it.

Thanks,

	M.
-- 
Jazz is not dead. It just smells funny...