| Message ID | 20160526075843.25236-4-Wei.Chen@linaro.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Hi Wei, On 26/05/16 08:58, Wei Chen wrote: > Currently, MPIDR_HWID_MASK is using the bit definition of AArch32 MPIDR. > From ARMv8 ARM we can see there are 4 levels of affinity on AArch64 > whilst AArch32 has only 3. So, this value is not correct when Xen is > running on AArch64. > > Now, we use the value 0xff00ffffff for this macro on AArch64. But neither > of this value and its bitwise invert value can be used in mov instruction > with the encoding of {imm16:shift} or {imms:immr}. So we have to use ldr > to load the bitwise invert value to register. > > The details of mov immediate encoding are listed in ARMv8 ARM C4.2.5. The section numbering may change between two versions of the spec. I would mention the version (e.g DDI 0487A.i, I guess?). Regards,
On 27 May 2016 at 18:50, Julien Grall <julien.grall@arm.com> wrote: > Hi Wei, > > On 26/05/16 08:58, Wei Chen wrote: >> >> Currently, MPIDR_HWID_MASK is using the bit definition of AArch32 MPIDR. >> From ARMv8 ARM we can see there are 4 levels of affinity on AArch64 >> whilst AArch32 has only 3. So, this value is not correct when Xen is >> running on AArch64. >> >> Now, we use the value 0xff00ffffff for this macro on AArch64. But neither >> of this value and its bitwise invert value can be used in mov instruction >> with the encoding of {imm16:shift} or {imms:immr}. So we have to use ldr >> to load the bitwise invert value to register. >> >> The details of mov immediate encoding are listed in ARMv8 ARM C4.2.5. > > > The section numbering may change between two versions of the spec. I would > mention the version (e.g DDI 0487A.i, I guess?). > Sure, I would do it in v3. > Regards, > > -- > Julien Grall
diff --git a/xen/arch/arm/arm64/head.S b/xen/arch/arm/arm64/head.S index d5831f2..3090beb 100644 --- a/xen/arch/arm/arm64/head.S +++ b/xen/arch/arm/arm64/head.S @@ -270,7 +270,7 @@ common_start: tbz x0, _MPIDR_SMP, 1f /* Multiprocessor extension not supported? */ tbnz x0, _MPIDR_UP, 1f /* Uniprocessor system? */ - mov x13, #(~MPIDR_HWID_MASK) + ldr x13, =(~MPIDR_HWID_MASK) bic x24, x0, x13 /* Mask out flags to get CPU ID */ 1: diff --git a/xen/include/asm-arm/processor.h b/xen/include/asm-arm/processor.h index b4cce7e..284ad6a 100644 --- a/xen/include/asm-arm/processor.h +++ b/xen/include/asm-arm/processor.h @@ -18,7 +18,11 @@ #define MPIDR_SMP (_AC(1,U) << _MPIDR_SMP) #define MPIDR_AFF0_SHIFT (0) #define MPIDR_AFF0_MASK (_AC(0xff,U) << MPIDR_AFF0_SHIFT) +#ifdef CONFIG_ARM_64 +#define MPIDR_HWID_MASK _AC(0xff00ffffff,UL) +#else #define MPIDR_HWID_MASK _AC(0xffffff,U) +#endif #define MPIDR_INVALID (~MPIDR_HWID_MASK) #define MPIDR_LEVEL_BITS (8)
Currently, MPIDR_HWID_MASK is using the bit definition of AArch32 MPIDR. From ARMv8 ARM we can see there are 4 levels of affinity on AArch64 whilst AArch32 has only 3. So, this value is not correct when Xen is running on AArch64. Now, we use the value 0xff00ffffff for this macro on AArch64. But neither of this value and its bitwise invert value can be used in mov instruction with the encoding of {imm16:shift} or {imms:immr}. So we have to use ldr to load the bitwise invert value to register. The details of mov immediate encoding are listed in ARMv8 ARM C4.2.5. Signed-off-by: Wei Chen <Wei.Chen@linaro.org> --- v2: Address Julien's comments 1. Fix typos in commit messages. 2. Explain valid MPIDR_HWID_MASK value in AArch64. 3. Simply explain mov immediate encoding. --- xen/arch/arm/arm64/head.S | 2 +- xen/include/asm-arm/processor.h | 4 ++++ 2 files changed, 5 insertions(+), 1 deletion(-)