| Message ID | 20231123044219.896776-19-richard.henderson@linaro.org |
|---|---|
| State | New |
| Headers | show |
| Series | target/arm: kvm cleanups | expand |
On 23/11/23 05:42, Richard Henderson wrote: > There is no need to do this in kvm_arch_init_vcpu per vcpu. > Inline kvm_arm_init_serror_injection rather than keep separate. > > Signed-off-by: Richard Henderson <richard.henderson@linaro.org> > --- > target/arm/kvm_arm.h | 8 -------- > target/arm/kvm.c | 13 ++++--------- > 2 files changed, 4 insertions(+), 17 deletions(-) Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
On 23/11/23 05:42, Richard Henderson wrote: > There is no need to do this in kvm_arch_init_vcpu per vcpu. > Inline kvm_arm_init_serror_injection rather than keep separate. > > Signed-off-by: Richard Henderson <richard.henderson@linaro.org> > --- > target/arm/kvm_arm.h | 8 -------- > target/arm/kvm.c | 13 ++++--------- > 2 files changed, 4 insertions(+), 17 deletions(-) > @@ -562,6 +556,10 @@ int kvm_arch_init(MachineState *ms, KVMState *s) > > cap_has_mp_state = kvm_check_extension(s, KVM_CAP_MP_STATE); > > + /* Check whether user space can specify guest syndrome value */ > + cap_has_inject_serror_esr = > + kvm_check_extension(s, KVM_CAP_ARM_INJECT_SERROR_ESR); > + > if (ms->smp.cpus > 256 && > !kvm_check_extension(s, KVM_CAP_ARM_IRQ_LINE_LAYOUT_2)) { > error_report("Using more than 256 vcpus requires a host kernel " > @@ -1948,9 +1946,6 @@ int kvm_arch_init_vcpu(CPUState *cs) > } > cpu->mp_affinity = mpidr & ARM64_AFFINITY_MASK; > > - /* Check whether user space can specify guest syndrome value */ > - kvm_arm_init_serror_injection(cs); > - > return kvm_arm_init_cpreg_list(cpu); > } Just checking, in a heterogeneous setup we still want to keep these 2 calls per-vCPU, right? if (kvm_check_extension(cs->kvm_state, KVM_CAP_ARM_PSCI_0_2)) { cpu->psci_version = QEMU_PSCI_VERSION_0_2; cpu->kvm_init_features[0] |= 1 << KVM_ARM_VCPU_PSCI_0_2; } if (!kvm_check_extension(cs->kvm_state, KVM_CAP_ARM_PMU_V3)) { cpu->has_pmu = false; }
On 11/23/23 15:42, Richard Henderson wrote: > There is no need to do this in kvm_arch_init_vcpu per vcpu. > Inline kvm_arm_init_serror_injection rather than keep separate. > > Signed-off-by: Richard Henderson <richard.henderson@linaro.org> > --- > target/arm/kvm_arm.h | 8 -------- > target/arm/kvm.c | 13 ++++--------- > 2 files changed, 4 insertions(+), 17 deletions(-) > Reviewed-by: Gavin Shan <gshan@redhat.com>
On 11/24/23 03:54, Philippe Mathieu-Daudé wrote: > On 23/11/23 05:42, Richard Henderson wrote: >> There is no need to do this in kvm_arch_init_vcpu per vcpu. >> Inline kvm_arm_init_serror_injection rather than keep separate. >> >> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> >> --- >> target/arm/kvm_arm.h | 8 -------- >> target/arm/kvm.c | 13 ++++--------- >> 2 files changed, 4 insertions(+), 17 deletions(-) > > >> @@ -562,6 +556,10 @@ int kvm_arch_init(MachineState *ms, KVMState *s) >> cap_has_mp_state = kvm_check_extension(s, KVM_CAP_MP_STATE); >> + /* Check whether user space can specify guest syndrome value */ >> + cap_has_inject_serror_esr = >> + kvm_check_extension(s, KVM_CAP_ARM_INJECT_SERROR_ESR); >> + >> if (ms->smp.cpus > 256 && >> !kvm_check_extension(s, KVM_CAP_ARM_IRQ_LINE_LAYOUT_2)) { >> error_report("Using more than 256 vcpus requires a host kernel " >> @@ -1948,9 +1946,6 @@ int kvm_arch_init_vcpu(CPUState *cs) >> } >> cpu->mp_affinity = mpidr & ARM64_AFFINITY_MASK; >> - /* Check whether user space can specify guest syndrome value */ >> - kvm_arm_init_serror_injection(cs); >> - >> return kvm_arm_init_cpreg_list(cpu); >> } > > > Just checking, in a heterogeneous setup we still want to keep > these 2 calls per-vCPU, right? There is no hetrogeneous kvm -- every vcpu must match the host cpu. r~ > if (kvm_check_extension(cs->kvm_state, KVM_CAP_ARM_PSCI_0_2)) { > cpu->psci_version = QEMU_PSCI_VERSION_0_2; > cpu->kvm_init_features[0] |= 1 << KVM_ARM_VCPU_PSCI_0_2; > } > > if (!kvm_check_extension(cs->kvm_state, KVM_CAP_ARM_PMU_V3)) { > cpu->has_pmu = false; > }
On 11/12/23 18:09, Richard Henderson wrote: > On 11/24/23 03:54, Philippe Mathieu-Daudé wrote: >> On 23/11/23 05:42, Richard Henderson wrote: >>> There is no need to do this in kvm_arch_init_vcpu per vcpu. >>> Inline kvm_arm_init_serror_injection rather than keep separate. >>> >>> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> >>> --- >>> target/arm/kvm_arm.h | 8 -------- >>> target/arm/kvm.c | 13 ++++--------- >>> 2 files changed, 4 insertions(+), 17 deletions(-) >> >> >>> @@ -562,6 +556,10 @@ int kvm_arch_init(MachineState *ms, KVMState *s) >>> cap_has_mp_state = kvm_check_extension(s, KVM_CAP_MP_STATE); >>> + /* Check whether user space can specify guest syndrome value */ >>> + cap_has_inject_serror_esr = >>> + kvm_check_extension(s, KVM_CAP_ARM_INJECT_SERROR_ESR); >>> + >>> if (ms->smp.cpus > 256 && >>> !kvm_check_extension(s, KVM_CAP_ARM_IRQ_LINE_LAYOUT_2)) { >>> error_report("Using more than 256 vcpus requires a host >>> kernel " >>> @@ -1948,9 +1946,6 @@ int kvm_arch_init_vcpu(CPUState *cs) >>> } >>> cpu->mp_affinity = mpidr & ARM64_AFFINITY_MASK; >>> - /* Check whether user space can specify guest syndrome value */ >>> - kvm_arm_init_serror_injection(cs); >>> - >>> return kvm_arm_init_cpreg_list(cpu); >>> } >> >> >> Just checking, in a heterogeneous setup we still want to keep >> these 2 calls per-vCPU, right? > > There is no hetrogeneous kvm -- every vcpu must match the host cpu. So big.LITTLE will never be a KVM thing?
On 12/11/23 10:43, Philippe Mathieu-Daudé wrote: > On 11/12/23 18:09, Richard Henderson wrote: >> On 11/24/23 03:54, Philippe Mathieu-Daudé wrote: >>> On 23/11/23 05:42, Richard Henderson wrote: >>>> There is no need to do this in kvm_arch_init_vcpu per vcpu. >>>> Inline kvm_arm_init_serror_injection rather than keep separate. >>>> >>>> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> >>>> --- >>>> target/arm/kvm_arm.h | 8 -------- >>>> target/arm/kvm.c | 13 ++++--------- >>>> 2 files changed, 4 insertions(+), 17 deletions(-) >>> >>> >>>> @@ -562,6 +556,10 @@ int kvm_arch_init(MachineState *ms, KVMState *s) >>>> cap_has_mp_state = kvm_check_extension(s, KVM_CAP_MP_STATE); >>>> + /* Check whether user space can specify guest syndrome value */ >>>> + cap_has_inject_serror_esr = >>>> + kvm_check_extension(s, KVM_CAP_ARM_INJECT_SERROR_ESR); >>>> + >>>> if (ms->smp.cpus > 256 && >>>> !kvm_check_extension(s, KVM_CAP_ARM_IRQ_LINE_LAYOUT_2)) { >>>> error_report("Using more than 256 vcpus requires a host kernel " >>>> @@ -1948,9 +1946,6 @@ int kvm_arch_init_vcpu(CPUState *cs) >>>> } >>>> cpu->mp_affinity = mpidr & ARM64_AFFINITY_MASK; >>>> - /* Check whether user space can specify guest syndrome value */ >>>> - kvm_arm_init_serror_injection(cs); >>>> - >>>> return kvm_arm_init_cpreg_list(cpu); >>>> } >>> >>> >>> Just checking, in a heterogeneous setup we still want to keep >>> these 2 calls per-vCPU, right? >> >> There is no hetrogeneous kvm -- every vcpu must match the host cpu. > > So big.LITTLE will never be a KVM thing? Not as far as I'm aware. There are even issues *running* on big.LITTLE hosts -- you must use taskset to limit qemu to one cpu type, either big or little. r~
diff --git a/target/arm/kvm_arm.h b/target/arm/kvm_arm.h index 350ba6cb96..1ec2476de7 100644 --- a/target/arm/kvm_arm.h +++ b/target/arm/kvm_arm.h @@ -98,14 +98,6 @@ void kvm_arm_cpu_post_load(ARMCPU *cpu); */ void kvm_arm_reset_vcpu(ARMCPU *cpu); -/** - * kvm_arm_init_serror_injection: - * @cs: CPUState - * - * Check whether KVM can set guest SError syndrome. - */ -void kvm_arm_init_serror_injection(CPUState *cs); - /** * kvm_get_vcpu_events: * @cpu: ARMCPU diff --git a/target/arm/kvm.c b/target/arm/kvm.c index bc4ba7628b..3250919273 100644 --- a/target/arm/kvm.c +++ b/target/arm/kvm.c @@ -96,12 +96,6 @@ static int kvm_arm_vcpu_finalize(CPUState *cs, int feature) return kvm_vcpu_ioctl(cs, KVM_ARM_VCPU_FINALIZE, &feature); } -void kvm_arm_init_serror_injection(CPUState *cs) -{ - cap_has_inject_serror_esr = kvm_check_extension(cs->kvm_state, - KVM_CAP_ARM_INJECT_SERROR_ESR); -} - bool kvm_arm_create_scratch_host_vcpu(const uint32_t *cpus_to_try, int *fdarray, struct kvm_vcpu_init *init) @@ -562,6 +556,10 @@ int kvm_arch_init(MachineState *ms, KVMState *s) cap_has_mp_state = kvm_check_extension(s, KVM_CAP_MP_STATE); + /* Check whether user space can specify guest syndrome value */ + cap_has_inject_serror_esr = + kvm_check_extension(s, KVM_CAP_ARM_INJECT_SERROR_ESR); + if (ms->smp.cpus > 256 && !kvm_check_extension(s, KVM_CAP_ARM_IRQ_LINE_LAYOUT_2)) { error_report("Using more than 256 vcpus requires a host kernel " @@ -1948,9 +1946,6 @@ int kvm_arch_init_vcpu(CPUState *cs) } cpu->mp_affinity = mpidr & ARM64_AFFINITY_MASK; - /* Check whether user space can specify guest syndrome value */ - kvm_arm_init_serror_injection(cs); - return kvm_arm_init_cpreg_list(cpu); }
There is no need to do this in kvm_arch_init_vcpu per vcpu. Inline kvm_arm_init_serror_injection rather than keep separate. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/arm/kvm_arm.h | 8 -------- target/arm/kvm.c | 13 ++++--------- 2 files changed, 4 insertions(+), 17 deletions(-)