@@ -767,6 +767,11 @@ static __always_inline bool system_supports_tpidr2(void)
return system_supports_sme();
}
+static __always_inline bool system_supports_fpmr(void)
+{
+ return alternative_has_cap_unlikely(ARM64_HAS_FPMR);
+}
+
static __always_inline bool system_supports_cnp(void)
{
return alternative_has_cap_unlikely(ARM64_HAS_CNP);
@@ -89,6 +89,7 @@ struct cpu_fp_state {
void *sve_state;
void *sme_state;
u64 *svcr;
+ u64 *fpmr;
unsigned int sve_vl;
unsigned int sme_vl;
enum fp_type *fp_type;
@@ -154,6 +155,7 @@ extern void cpu_enable_sve(const struct arm64_cpu_capabilities *__unused);
extern void cpu_enable_sme(const struct arm64_cpu_capabilities *__unused);
extern void cpu_enable_sme2(const struct arm64_cpu_capabilities *__unused);
extern void cpu_enable_fa64(const struct arm64_cpu_capabilities *__unused);
+extern void cpu_enable_fpmr(const struct arm64_cpu_capabilities *__unused);
extern u64 read_smcr_features(void);
@@ -517,6 +517,7 @@ struct kvm_vcpu_arch {
enum fp_type fp_type;
unsigned int sve_max_vl;
u64 svcr;
+ u64 fpmr;
/* Stage 2 paging state used by the hardware on next switch */
struct kvm_s2_mmu *hw_mmu;
@@ -158,6 +158,8 @@ struct thread_struct {
struct user_fpsimd_state fpsimd_state;
} uw;
+ u64 fpmr; /* Adjacent to fpsimd_state for KVM */
+
enum fp_type fp_type; /* registers FPSIMD or SVE? */
unsigned int fpsimd_cpu;
void *sve_state; /* SVE registers, if any */
@@ -272,6 +272,7 @@ static const struct arm64_ftr_bits ftr_id_aa64pfr1[] = {
};
static const struct arm64_ftr_bits ftr_id_aa64pfr2[] = {
+ ARM64_FTR_BITS(FTR_VISIBLE, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64PFR2_EL1_FPMR_SHIFT, 4, 0),
ARM64_FTR_END,
};
@@ -2759,6 +2760,14 @@ static const struct arm64_cpu_capabilities arm64_features[] = {
.matches = has_cpuid_feature,
ARM64_CPUID_FIELDS(ID_AA64MMFR2_EL1, EVT, IMP)
},
+ {
+ .desc = "FPMR",
+ .type = ARM64_CPUCAP_SYSTEM_FEATURE,
+ .capability = ARM64_HAS_FPMR,
+ .matches = has_cpuid_feature,
+ .cpu_enable = cpu_enable_fpmr,
+ ARM64_CPUID_FIELDS(ID_AA64PFR2_EL1, FPMR, IMP)
+ },
{},
};
@@ -385,6 +385,9 @@ static void task_fpsimd_load(void)
WARN_ON(!system_supports_fpsimd());
WARN_ON(!have_cpu_fpsimd_context());
+ if (system_supports_fpmr())
+ write_sysreg_s(current->thread.fpmr, SYS_FPMR);
+
if (system_supports_sve() || system_supports_sme()) {
switch (current->thread.fp_type) {
case FP_STATE_FPSIMD:
@@ -472,6 +475,9 @@ static void fpsimd_save(void)
if (test_thread_flag(TIF_FOREIGN_FPSTATE))
return;
+ if (system_supports_fpmr())
+ *(last->fpmr) = read_sysreg_s(SYS_FPMR);
+
/*
* If a task is in a syscall the ABI allows us to only
* preserve the state shared with FPSIMD so don't bother
@@ -714,6 +720,12 @@ static void sve_to_fpsimd(struct task_struct *task)
}
}
+void cpu_enable_fpmr(const struct arm64_cpu_capabilities *__always_unused p)
+{
+ write_sysreg_s(read_sysreg_s(SYS_SCTLR_EL1) | SCTLR_EL1_EnFPM_MASK,
+ SYS_SCTLR_EL1);
+}
+
#ifdef CONFIG_ARM64_SVE
/*
* Call __sve_free() directly only if you know task can't be scheduled
@@ -1671,6 +1683,7 @@ static void fpsimd_bind_task_to_cpu(void)
last->sve_vl = task_get_sve_vl(current);
last->sme_vl = task_get_sme_vl(current);
last->svcr = ¤t->thread.svcr;
+ last->fpmr = ¤t->thread.fpmr;
last->fp_type = ¤t->thread.fp_type;
last->to_save = FP_STATE_CURRENT;
current->thread.fpsimd_cpu = smp_processor_id();
@@ -153,6 +153,7 @@ void kvm_arch_vcpu_ctxsync_fp(struct kvm_vcpu *vcpu)
fp_state.sve_vl = vcpu->arch.sve_max_vl;
fp_state.sme_state = NULL;
fp_state.svcr = &vcpu->arch.svcr;
+ fp_state.fpmr = &vcpu->arch.fpmr;
fp_state.fp_type = &vcpu->arch.fp_type;
if (vcpu_has_sve(vcpu))
@@ -26,6 +26,7 @@ HAS_ECV
HAS_ECV_CNTPOFF
HAS_EPAN
HAS_EVT
+HAS_FPMR
HAS_FGT
HAS_FPSIMD
HAS_GENERIC_AUTH
FEAT_FPMR defines a new EL0 accessible register FPMR use to configure the FP8 related features added to the architecture at the same time. Detect support for this register and context switch it for EL0 when present. Due to the sharing of responsibility for saving floating point state between the host kernel and KVM FP8 support is not yet implemented in KVM and a stub similar to that used for SVCR is provided for FPMR in order to avoid bisection issues. To make it easier to share host state with the hypervisor we store FPMR immediately after the base floating point state, existing usage means that it is not practical to extend that directly. Signed-off-by: Mark Brown <broonie@kernel.org> --- arch/arm64/include/asm/cpufeature.h | 5 +++++ arch/arm64/include/asm/fpsimd.h | 2 ++ arch/arm64/include/asm/kvm_host.h | 1 + arch/arm64/include/asm/processor.h | 2 ++ arch/arm64/kernel/cpufeature.c | 9 +++++++++ arch/arm64/kernel/fpsimd.c | 13 +++++++++++++ arch/arm64/kvm/fpsimd.c | 1 + arch/arm64/tools/cpucaps | 1 + 8 files changed, 34 insertions(+)