Re: [BOOTWRAPPER][PATCH] Make GICv3.1 extended ranges available to non-secure

From: Marc Zyngier <marc.zyngier@arm.com>
To: Mark Rutland <mark.rutland@arm.com>
Cc: linux-arm-kernel@lists.infradead.org
Subject: Re: [BOOTWRAPPER][PATCH] Make GICv3.1 extended ranges available to non-secure
Date: Tue, 23 Jul 2019 15:28:11 +0100	[thread overview]
Message-ID: <ad153393-29d9-1c14-b747-9af4adb1859f@arm.com> (raw)
In-Reply-To: <20190723141017.GB5004@lakrids.cambridge.arm.com>

On 23/07/2019 15:10, Mark Rutland wrote:
> Hi Marc,
> 
> On Tue, Jul 23, 2019 at 12:55:45PM +0100, Marc Zyngier wrote:
>> If we have a GICv3.1-capable system, configure the EPPI/ESPI ranges
>> to be accessible from the non-secure world.
>>
>> Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
>> ---
>>  gic-v3.c | 14 ++++++++++++--
>>  1 file changed, 12 insertions(+), 2 deletions(-)
>>
>> diff --git a/gic-v3.c b/gic-v3.c
>> index 476f703..4b733ba 100644
>> --- a/gic-v3.c
>> +++ b/gic-v3.c
>> @@ -17,6 +17,8 @@
>>  #define GICD_TYPER			0x4
>>  #define GICD_IGROUP0			0x80
>>  #define GICD_IGRPMOD0			0xd00
>> +#define GICD_IGROUPR0E			0x1000
>> +#define GICD_IGRPMODR0E			0x3400
>>  
>>  #define GICD_CTLR_EnableGrp0		(1 << 0)
>>  #define GICD_CTLR_EnableGrp1ns		(1 << 1)
>> @@ -57,6 +59,7 @@ void gic_secure_init_primary(void)
>>  		 * ChildrenAsleep to be 0.
>>  		 */
>>  		uint32_t waker = raw_readl(gicr_ptr + GICR_WAKER);
>> +		uint32_t regs;
>>  		waker &= ~GICR_WAKER_ProcessorSleep;
>>  		raw_writel(waker, gicr_ptr + GICR_WAKER);
>>  		dsb(st);
>> @@ -72,8 +75,11 @@ void gic_secure_init_primary(void)
>>  		typer = raw_readl(gicr_ptr + GICR_TYPER);
>>  
>>  		gicr_ptr += 0x10000; /* Go to SGI_Base */
>> -		raw_writel(~0x0, gicr_ptr + GICR_IGROUP0);
>> -		raw_writel(0x0, gicr_ptr + GICR_IGRPMOD0);
>> +		regs = 1 + ((typer >> 27) & 0x1f);
> 
> IIUC that's GICR_TYPER.PPInum, right?

Indeed.

> 
>> +		for (i = 0; i < regs; i++) {
>> +			raw_writel(~0x0, gicr_ptr + GICR_IGROUP0 + i * 4);
>> +			raw_writel(0x0, gicr_ptr + GICR_IGRPMOD0 + i * 4);
>> +		}
>>  
>>  		/* Next redist */
>>  		gicr_ptr += 0x10000;
>> @@ -87,6 +93,10 @@ void gic_secure_init_primary(void)
>>  		raw_writel(~0x0, gicd_base + GICD_IGROUP0 + i * 4);
>>  		raw_writel(0x0, gicd_base + GICD_IGRPMOD0 + i * 4);
>>  	}
>> +	for (i = 0; i < ((typer >> 27) & GICD_TYPER_ITLineNumber); i++) {
> 
> IIUC this is ESPI_range_field, so using the GICD_TYPER_ITLineNumber
> mnemonic is a bit misleading.
> 
> Given the verbose regfield names, _SHIFT and _MASK definitions will make
> this illegible, so could we please add extactors:
> 
> #define GICR_TYPER_PPInum(r)		(((r) >> 27) & 0x1f)
> #define GICD_TYPER_ESPI_range(r)	(((r) >> 27) & 0x1f)
> 
> ... and use those in the for loops?
> 
> With that, this looks good to me. If you respin, I'm happy to pick this
> up. :)

Sure, I'll do that shortly.

Thanks,

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

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel