[PATCH 11/18] arm64: fpsimd: Split FPSR/FPCR from SVE save/restore
Vladimir Murzin
vladimir.murzin at arm.com
Wed May 27 06:44:29 PDT 2026
On 5/21/26 14:25, Mark Rutland wrote:
> Regardless of whether the vector registers are saved in FPSIMD or SVE
> format, we store FPSR and FPCR in user_fpsimd_state::{fpsr,fpcr}.
>
> For historical reasons, the functions which save/restore SVE context
> take a pointer to user_fpsimd_state::fpsr, and use this to access both
> user_fpsimd_state::fpsr and user_fpsimd_state::fpcr. This is
> unnecessarily fragile.
>
> Move the save/restore of FPSR and FPCR into separate helper functions
> which take a pointer to user_fpsimd_state. I've used read_sysreg_s() and
> write_sysreg_s() as contemporary versions of LLVM will refuse to
> directly assemble accesses to FPCR or FPSR unless the "fp" arch
> extension is enabled.
>
> Note that the SVE assembly sequence for restoring FPCR uses an
> unconditional write to FPCR. The plain FPSIMD assembly sequence has used
> a conditional write to FPCR since 2014 in commit:
>
> 5959e25729a5 ("arm64: fpsimd: avoid restoring fpcr if the contents haven't change")
>
> ... but this was not followed for the SVE restore assembly implemented
> in 2017 in commit:
>
> 1fc5dce78ad1 ("arm64/sve: Low-level SVE architectural state manipulation functions")
>
> ... so I've assumed that this doesn't actually matter in practice, and
> implemented the C version matching the existing SVE assembly.
>
> For the moment, fpsimd_save_state() and fpsimd_load_state() are left
> as-is with their own logic to save/restore FPSR and FPCR. This will be
> unified in subsequent patches.
>
> Signed-off-by: Mark Rutland <mark.rutland at arm.com>
> Cc: Catalin Marinas <catalin.marinas at arm.com>
> Cc: Fuad Tabba <tabba at google.com>
> Cc: James Morse <james.morse at arm.com>
> Cc: Marc Zyngier <maz at kernel.org>
> Cc: Mark Brown <broonie at kernel.org>
> Cc: Oliver Upton <oupton at kernel.org>
> Cc: Will Deacon <will at kernel.org>
> ---
> arch/arm64/include/asm/fpsimd.h | 17 ++++++++++++++---
> arch/arm64/include/asm/fpsimdmacros.h | 13 ++-----------
> arch/arm64/include/asm/kvm_hyp.h | 4 ++--
> arch/arm64/kernel/entry-fpsimd.S | 10 ++++------
> arch/arm64/kernel/fpsimd.c | 5 +++--
> arch/arm64/kvm/hyp/fpsimd.S | 4 ++--
> arch/arm64/kvm/hyp/include/hyp/switch.h | 4 ++--
> arch/arm64/kvm/hyp/nvhe/hyp-main.c | 5 +++--
> 8 files changed, 32 insertions(+), 30 deletions(-)
>
> diff --git a/arch/arm64/include/asm/fpsimd.h b/arch/arm64/include/asm/fpsimd.h
> index 36cf528e64971..6fd5cdf5e5f17 100644
> --- a/arch/arm64/include/asm/fpsimd.h
> +++ b/arch/arm64/include/asm/fpsimd.h
> @@ -74,6 +74,18 @@ static inline void cpacr_restore(unsigned long cpacr)
>
> struct task_struct;
>
> +static inline void fpsimd_save_common(struct user_fpsimd_state *state)
> +{
> + state->fpsr = read_sysreg_s(SYS_FPSR);
> + state->fpcr = read_sysreg_s(SYS_FPCR);
> +}
> +
> +static inline void fpsimd_load_common(const struct user_fpsimd_state *state)
> +{
> + write_sysreg_s(state->fpsr, SYS_FPSR);
> + write_sysreg_s(state->fpcr, SYS_FPCR);
> +}
> +
> extern void fpsimd_save_state(struct user_fpsimd_state *state);
> extern void fpsimd_load_state(struct user_fpsimd_state *state);
>
> @@ -157,9 +169,8 @@ static inline unsigned int sve_get_vl(void)
> return vl;
> }
>
> -extern void sve_save_state(void *state, u32 *pfpsr, int save_ffr);
> -extern void sve_load_state(void const *state, u32 const *pfpsr,
> - int restore_ffr);
> +extern void sve_save_state(void *state, int save_ffr);
> +extern void sve_load_state(void const *state, int restore_ffr);
> extern void sve_flush_live(bool flush_ffr, unsigned long vq_minus_1);
> extern void sme_save_state(void *state, int zt);
> extern void sme_load_state(void const *state, int zt);
> diff --git a/arch/arm64/include/asm/fpsimdmacros.h b/arch/arm64/include/asm/fpsimdmacros.h
> index d75c9d4c9989b..c79ae7ec1ff05 100644
> --- a/arch/arm64/include/asm/fpsimdmacros.h
> +++ b/arch/arm64/include/asm/fpsimdmacros.h
> @@ -235,7 +235,7 @@
> _sve_wrffr 0
> .endm
>
> -.macro sve_save nxbase, xpfpsr, save_ffr, nxtmp
> +.macro sve_save nxbase, save_ffr
> _for n, 0, 31, _sve_str_v \n, \nxbase, \n - 34
> _for n, 0, 15, _sve_str_p \n, \nxbase, \n - 16
> cbz \save_ffr, 921f
> @@ -246,24 +246,15 @@
> 922:
> _sve_str_p 0, \nxbase
> _sve_ldr_p 0, \nxbase, -16
> - mrs x\nxtmp, fpsr
> - str w\nxtmp, [\xpfpsr]
> - mrs x\nxtmp, fpcr
> - str w\nxtmp, [\xpfpsr, #4]
> .endm
>
> -.macro sve_load nxbase, xpfpsr, restore_ffr, nxtmp
> +.macro sve_load nxbase, restore_ffr
> _for n, 0, 31, _sve_ldr_v \n, \nxbase, \n - 34
> cbz \restore_ffr, 921f
> _sve_ldr_p 0, \nxbase
> _sve_wrffr 0
> 921:
> _for n, 0, 15, _sve_ldr_p \n, \nxbase, \n - 16
> -
> - ldr w\nxtmp, [\xpfpsr]
> - msr fpsr, x\nxtmp
> - ldr w\nxtmp, [\xpfpsr, #4]
> - msr fpcr, x\nxtmp
> .endm
>
> .macro sme_save_za nxbase, xvl, nw
> diff --git a/arch/arm64/include/asm/kvm_hyp.h b/arch/arm64/include/asm/kvm_hyp.h
> index 8d06b62e7188c..0030cc1b52197 100644
> --- a/arch/arm64/include/asm/kvm_hyp.h
> +++ b/arch/arm64/include/asm/kvm_hyp.h
> @@ -123,8 +123,8 @@ void __debug_restore_host_buffers_nvhe(struct kvm_vcpu *vcpu);
>
> void __fpsimd_save_state(struct user_fpsimd_state *fp_regs);
> void __fpsimd_restore_state(struct user_fpsimd_state *fp_regs);
> -void __sve_save_state(void *sve_pffr, u32 *fpsr, int save_ffr);
> -void __sve_restore_state(void *sve_pffr, u32 *fpsr, int restore_ffr);
> +void __sve_save_state(void *sve, int save_ffr);
> +void __sve_restore_state(void *sve, int restore_ffr);
>
> u64 __guest_enter(struct kvm_vcpu *vcpu);
>
> diff --git a/arch/arm64/kernel/entry-fpsimd.S b/arch/arm64/kernel/entry-fpsimd.S
> index 7f2d31dff8c17..83fe9c32bbd1c 100644
> --- a/arch/arm64/kernel/entry-fpsimd.S
> +++ b/arch/arm64/kernel/entry-fpsimd.S
> @@ -37,11 +37,10 @@ SYM_FUNC_END(fpsimd_load_state)
> * Save the SVE state
> *
> * x0 - pointer to buffer for state
> - * x1 - pointer to storage for FPSR
> - * x2 - Save FFR if non-zero
> + * x1 - Save FFR if non-zero
> */
> SYM_FUNC_START(sve_save_state)
> - sve_save 0, x1, x2, 3
> + sve_save 0, x1
> ret
> SYM_FUNC_END(sve_save_state)
>
> @@ -49,11 +48,10 @@ SYM_FUNC_END(sve_save_state)
> * Load the SVE state
> *
> * x0 - pointer to buffer for state
> - * x1 - pointer to storage for FPSR
> - * x2 - Restore FFR if non-zero
> + * x1 - Restore FFR if non-zero
> */
> SYM_FUNC_START(sve_load_state)
> - sve_load 0, x1, x2, 4
> + sve_load 0, x1
> ret
> SYM_FUNC_END(sve_load_state)
>
> diff --git a/arch/arm64/kernel/fpsimd.c b/arch/arm64/kernel/fpsimd.c
> index 2578c2372c89e..9806fea8fea7c 100644
> --- a/arch/arm64/kernel/fpsimd.c
> +++ b/arch/arm64/kernel/fpsimd.c
> @@ -426,8 +426,8 @@ static void task_fpsimd_load(void)
> if (restore_sve_regs) {
> WARN_ON_ONCE(current->thread.fp_type != FP_STATE_SVE);
> sve_load_state(sve_pffr(¤t->thread),
> - ¤t->thread.uw.fpsimd_state.fpsr,
> restore_ffr);
> + fpsimd_load_common(¤t->thread.uw.fpsimd_state);
> } else {
> WARN_ON_ONCE(current->thread.fp_type != FP_STATE_FPSIMD);
> fpsimd_load_state(¤t->thread.uw.fpsimd_state);
> @@ -509,7 +509,8 @@ static void fpsimd_save_user_state(void)
>
> sve_save_state((char *)last->sve_state +
> sve_ffr_offset(vl),
> - &last->st->fpsr, save_ffr);
> + save_ffr);
> + fpsimd_save_common(last->st);
> *last->fp_type = FP_STATE_SVE;
> } else {
> fpsimd_save_state(last->st);
> diff --git a/arch/arm64/kvm/hyp/fpsimd.S b/arch/arm64/kvm/hyp/fpsimd.S
> index 6e16cbfc5df27..8575e32977d19 100644
> --- a/arch/arm64/kvm/hyp/fpsimd.S
> +++ b/arch/arm64/kvm/hyp/fpsimd.S
> @@ -21,11 +21,11 @@ SYM_FUNC_START(__fpsimd_restore_state)
> SYM_FUNC_END(__fpsimd_restore_state)
>
> SYM_FUNC_START(__sve_restore_state)
> - sve_load 0, x1, x2, 3
> + sve_load 0, x1
> ret
> SYM_FUNC_END(__sve_restore_state)
>
> SYM_FUNC_START(__sve_save_state)
> - sve_save 0, x1, x2, 3
> + sve_save 0, x1
> ret
> SYM_FUNC_END(__sve_save_state)
> diff --git a/arch/arm64/kvm/hyp/include/hyp/switch.h b/arch/arm64/kvm/hyp/include/hyp/switch.h
> index 6512dd3f75ae4..eb76a863ebb84 100644
> --- a/arch/arm64/kvm/hyp/include/hyp/switch.h
> +++ b/arch/arm64/kvm/hyp/include/hyp/switch.h
> @@ -468,8 +468,8 @@ static inline void __hyp_sve_restore_guest(struct kvm_vcpu *vcpu)
> */
> sve_cond_update_zcr_vq(vcpu_sve_max_vq(vcpu) - 1, SYS_ZCR_EL2);
> __sve_restore_state(vcpu_sve_pffr(vcpu),
> - &vcpu->arch.ctxt.fp_regs.fpsr,
> true);
> + fpsimd_load_common(&vcpu->arch.ctxt.fp_regs);
>
> /*
> * The effective VL for a VM could differ from the max VL when running a
> @@ -490,8 +490,8 @@ static inline void __hyp_sve_save_host(void)
> ctxt_sys_reg(hctxt, ZCR_EL1) = read_sysreg_el1(SYS_ZCR);
> write_sysreg_s(sve_vq_from_vl(kvm_host_sve_max_vl) - 1, SYS_ZCR_EL2);
> __sve_save_state(sve_regs + sve_ffr_offset(kvm_host_sve_max_vl),
> - &hctxt->fp_regs.fpsr,
> true);
> + fpsimd_save_common(&hctxt->fp_regs);
> }
>
> static inline void fpsimd_lazy_switch_to_guest(struct kvm_vcpu *vcpu)
> diff --git a/arch/arm64/kvm/hyp/nvhe/hyp-main.c b/arch/arm64/kvm/hyp/nvhe/hyp-main.c
> index 04a6d2e0ea73f..0be4577a67e7b 100644
> --- a/arch/arm64/kvm/hyp/nvhe/hyp-main.c
> +++ b/arch/arm64/kvm/hyp/nvhe/hyp-main.c
> @@ -35,7 +35,8 @@ static void __hyp_sve_save_guest(struct kvm_vcpu *vcpu)
> * on the VL, so use a consistent (i.e., the maximum) guest VL.
> */
> sve_cond_update_zcr_vq(vcpu_sve_max_vq(vcpu) - 1, SYS_ZCR_EL2);
> - __sve_save_state(vcpu_sve_pffr(vcpu), &vcpu->arch.ctxt.fp_regs.fpsr, true);
> + __sve_save_state(vcpu_sve_pffr(vcpu), true);
> + fpsimd_save_common(&vcpu->arch.ctxt.fp_regs);
> write_sysreg_s(sve_vq_from_vl(kvm_host_sve_max_vl) - 1, SYS_ZCR_EL2);
> }
>
> @@ -55,8 +56,8 @@ static void __hyp_sve_restore_host(void)
> */
> write_sysreg_s(sve_vq_from_vl(kvm_host_sve_max_vl) - 1, SYS_ZCR_EL2);
> __sve_restore_state(sve_regs + sve_ffr_offset(kvm_host_sve_max_vl),
> - &hctxt->fp_regs.fpsr,
> true);
> + fpsimd_load_common(&hctxt->fp_regs);
> write_sysreg_el1(ctxt_sys_reg(hctxt, ZCR_EL1), SYS_ZCR);
> }
>
> -- 2.30.2
>
FWIW,
Reviewed-by: Vladimir Murzin <vladimir.murzin at arm.com>
More information about the linux-arm-kernel
mailing list