[PATCH v2 2/2] arm64: smccc: Support SMCCC v1.3 SVE register saving hint
Ard Biesheuvel
ardb at kernel.org
Wed May 19 00:05:41 PDT 2021
Hi Mark,
Some more comments on this patch below. Apologies for not mentioning
these the first time around, I only commented on Marc's reply and did
not look carefully at the rest.
On Tue, 18 May 2021 at 20:35, Mark Brown <broonie at kernel.org> wrote:
>
> SMCCC v1.2 requires that all SVE state be preserved over SMC calls which
> introduces substantial overhead in the common case where there is no SVE
> state in the registers. To avoid this SMCCC v1.3 introduces a flag which
> allows the caller to say that there is no state that needs to be preserved
> in the registers. Make use of this flag, setting it if the SMCCC version
> indicates support for it and the TIF_ flags indicate that there is no live
> SVE state in the registers, this avoids placing any constraints on when
> SMCCC calls can be done or triggering extra saving and reloading of SVE
> register state in the kernel.
>
> This would be straightforward enough except for the rather entertaining
> inline assembly we use to do SMCCC v1.1 calls to allow us to take advantage
> of the limited number of registers it clobbers. Deal with this by having a
> slightly non-standard function which we call immediately before issuing the
> SMCCC call to make our checks. This causes an extra function call on any
> system built with SVE support, and extra checks when SVE is detected at
> runtime, but these costs are expected to be reasonable in the context of
> doing a SMCCC call in the first place.
>
> Since we can't check the TIF_ flags for current and would need to map
> the helper function separately when in the nVHE hypervisor we don't
> attempt to do the optimisation for it, this could be done as a followup.
>
> Signed-off-by: Mark Brown <broonie at kernel.org>
> ---
> arch/arm64/kernel/smccc-call.S | 38 ++++++++++++++++++++++++++++++++++
> drivers/firmware/smccc/smccc.c | 4 ++++
> include/linux/arm-smccc.h | 23 ++++++++++++++++++--
> 3 files changed, 63 insertions(+), 2 deletions(-)
>
> diff --git a/arch/arm64/kernel/smccc-call.S b/arch/arm64/kernel/smccc-call.S
> index d62447964ed9..2713ea37b536 100644
> --- a/arch/arm64/kernel/smccc-call.S
> +++ b/arch/arm64/kernel/smccc-call.S
> @@ -7,8 +7,46 @@
>
> #include <asm/asm-offsets.h>
> #include <asm/assembler.h>
> +#include <asm/thread_info.h>
> +
> +/*
> + * If we have SMCCC v1.3 and (as is likely) no SVE state in
> + * the registers then set the SMCCC hint bit to say there's no
> + * need to preserve it. Do this by directly adjusting the SMCCC
> + * function value which is already stored in x0 ready to be called.
> + *
> + * Since we need scratch registers but wish to avoid having to handle
> + * the stack we expect the caller to preserve x15 and x16 if needed,
You should use x16 and x17 here. The calling convention dictates that
the linker may use x16 and x17 in a veneer (or PLT lookup sequence) if
the branch to __smccc_sve_check() goes out of range, and even if our
PLT code only uses x16 at this time, it is better to follow the AAPCS
here.
> + * the only callers are expected to be the call below and the inline
> + * asm in linux/arm-smccc.h for SMCCC 1.1 and later calls.
> + */
> +SYM_CODE_START(__smccc_sve_check)
> + BTI_C
> +
Any reason not to use SYM_FUNC_START() here?
> +alternative_if ARM64_SVE
> +
> + ldr_l x15, smccc_has_sve_hint
> + cbz x15, 2f
> +
> + get_current_task x15
> + ldr x15, [x15, #TSK_TI_FLAGS]
> + and x16, x15, #_TIF_FOREIGN_FPSTATE // Any live FP state?
> + cbnz x16, 1f
You could use tbnz here, with the actual shift in TIF_FOREIGN_FPSTATE
> + mov x16, #_TIF_SVE // Does that state include SVE?
> + and x16, x15, x16
> + cbnz x16, 2f
> +
Same but with TIF_SVE. That way, you don't need to corrupt x15 at all
(but we should still mark x16 and x17 as clobbered)
> +1: orr x0, x0, ARM_SMCCC_1_3_SVE_HINT
> +alternative_else_nop_endif
> +
> +2: ret
> +SYM_CODE_END(__smccc_sve_check)
> +EXPORT_SYMBOL(__smccc_sve_check)
>
> .macro SMCCC instr
> +alternative_if ARM64_SVE
> + bl __smccc_sve_check
> +alternative_else_nop_endif
Do we need the redundant alternatives here? Could we use an
ALTERNATIVE() below instead, and drop the one in __smccc_sve_check?
> \instr #0
> ldr x4, [sp]
> stp x0, x1, [x4, #ARM_SMCCC_RES_X0_OFFS]
> diff --git a/drivers/firmware/smccc/smccc.c b/drivers/firmware/smccc/smccc.c
> index 028f81d702cc..9f937b125ab0 100644
> --- a/drivers/firmware/smccc/smccc.c
> +++ b/drivers/firmware/smccc/smccc.c
> @@ -15,6 +15,7 @@ static u32 smccc_version = ARM_SMCCC_VERSION_1_0;
> static enum arm_smccc_conduit smccc_conduit = SMCCC_CONDUIT_NONE;
>
> bool __ro_after_init smccc_trng_available = false;
> +u64 __ro_after_init smccc_has_sve_hint = false;
>
> void __init arm_smccc_version_init(u32 version, enum arm_smccc_conduit conduit)
> {
> @@ -22,6 +23,9 @@ void __init arm_smccc_version_init(u32 version, enum arm_smccc_conduit conduit)
> smccc_conduit = conduit;
>
> smccc_trng_available = smccc_probe_trng();
> + if (IS_ENABLED(CONFIG_ARM64_SVE) &&
> + smccc_version >= ARM_SMCCC_VERSION_1_3)
> + smccc_has_sve_hint = true;
> }
>
> enum arm_smccc_conduit arm_smccc_1_1_get_conduit(void)
> diff --git a/include/linux/arm-smccc.h b/include/linux/arm-smccc.h
> index 6861489a1890..9e974a9f8993 100644
> --- a/include/linux/arm-smccc.h
> +++ b/include/linux/arm-smccc.h
> @@ -63,6 +63,9 @@
> #define ARM_SMCCC_VERSION_1_0 0x10000
> #define ARM_SMCCC_VERSION_1_1 0x10001
> #define ARM_SMCCC_VERSION_1_2 0x10002
> +#define ARM_SMCCC_VERSION_1_3 0x10003
> +
> +#define ARM_SMCCC_1_3_SVE_HINT 0x10000
>
> #define ARM_SMCCC_VERSION_FUNC_ID \
> ARM_SMCCC_CALL_VAL(ARM_SMCCC_FAST_CALL, \
> @@ -216,6 +219,8 @@ u32 arm_smccc_get_version(void);
>
> void __init arm_smccc_version_init(u32 version, enum arm_smccc_conduit conduit);
>
> +extern u64 smccc_has_sve_hint;
> +
> /**
> * struct arm_smccc_res - Result from SMC/HVC call
> * @a0-a3 result values from registers 0 to 3
> @@ -297,6 +302,19 @@ asmlinkage void __arm_smccc_hvc(unsigned long a0, unsigned long a1,
>
> #endif
>
> +/* nVHE hypervisor doesn't have a current thread so needs separate checks */
> +#if defined(CONFIG_ARM64_SVE) && !defined(__KVM_NVHE_HYPERVISOR__)
> +
> +#define SMCCC_SVE_CHECK "bl __smccc_sve_check \n"
I.e., make this
#define SMCCC_SVE_CHECK ALTERNATIVE("bl __smccc_sve_check \n", "nop", ARM64_SVE)
(or whichever way around they should go in this particular case)
> +#define smccc_sve_clobbers "x15", "x16", "lr",
> +
> +#else
> +
> +#define SMCCC_SVE_CHECK
> +#define smccc_sve_clobbers
> +
> +#endif
> +
> #define ___count_args(_0, _1, _2, _3, _4, _5, _6, _7, _8, x, ...) x
>
> #define __count_args(...) \
> @@ -364,7 +382,7 @@ asmlinkage void __arm_smccc_hvc(unsigned long a0, unsigned long a1,
>
> #define ___constraints(count) \
> : __constraint_read_ ## count \
> - : "memory"
> + : smccc_sve_clobbers "memory"
> #define __constraints(count) ___constraints(count)
>
> /*
> @@ -379,7 +397,8 @@ asmlinkage void __arm_smccc_hvc(unsigned long a0, unsigned long a1,
> register unsigned long r2 asm("r2"); \
> register unsigned long r3 asm("r3"); \
> __declare_args(__count_args(__VA_ARGS__), __VA_ARGS__); \
> - asm volatile(inst "\n" : \
> + asm volatile(SMCCC_SVE_CHECK \
> + inst "\n" : \
> "=r" (r0), "=r" (r1), "=r" (r2), "=r" (r3) \
> __constraints(__count_args(__VA_ARGS__))); \
> if (___res) \
> --
> 2.20.1
>
More information about the linux-arm-kernel
mailing list