* [Qemu-devel] [PATCH V3] forward cpuid leaves when using -cpu host @ 2013-09-02 15:06 Benoît Canet 2013-09-02 15:06 ` [Qemu-devel] [PATCH V3] target-i386: forward CPUID cache leaves when -cpu host is used Benoît Canet 0 siblings, 1 reply; 7+ messages in thread From: Benoît Canet @ 2013-09-02 15:06 UTC (permalink / raw) To: qemu-devel; +Cc: imammedo, Benoît Canet, afaerber, kvm, ehabkost This patch uses directly cpuid_host to forward the informations instead of storing a variable number of leaves in the cpu states. v3: s/i386/target-i386/ [Andrea] move forward field to X86CPU structure and document it [Andrea] rename forward field [Eduardo] Rebase on top of eduardo cache flags cleanup [Andrea] v2: use index as argument to cpuid_host Benoît Canet (1): target-i386: forward CPUID cache leaves when -cpu host is used target-i386/cpu-qom.h | 3 +++ target-i386/cpu.c | 19 +++++++++++++++++++ 2 files changed, 22 insertions(+) -- 1.7.10.4 ^ permalink raw reply [flat|nested] 7+ messages in thread
* [Qemu-devel] [PATCH V3] target-i386: forward CPUID cache leaves when -cpu host is used 2013-09-02 15:06 [Qemu-devel] [PATCH V3] forward cpuid leaves when using -cpu host Benoît Canet @ 2013-09-02 15:06 ` Benoît Canet 2013-09-02 16:19 ` Eduardo Habkost 2013-09-19 17:33 ` Paolo Bonzini 0 siblings, 2 replies; 7+ messages in thread From: Benoît Canet @ 2013-09-02 15:06 UTC (permalink / raw) To: qemu-devel; +Cc: imammedo, Benoît Canet, afaerber, kvm, ehabkost Some users running cpu intensive tasks checking the cache CPUID leaves at startup and making decisions based on the result reported that the guest was not reflecting the host CPUID leaves when -cpu host is used. This patch fix this. Signed-off-by: Benoit Canet <benoit@irqsave.net> --- target-i386/cpu-qom.h | 3 +++ target-i386/cpu.c | 19 +++++++++++++++++++ 2 files changed, 22 insertions(+) diff --git a/target-i386/cpu-qom.h b/target-i386/cpu-qom.h index c4447c2..b1d1bd8 100644 --- a/target-i386/cpu-qom.h +++ b/target-i386/cpu-qom.h @@ -70,6 +70,9 @@ typedef struct X86CPU { bool hyperv_relaxed_timing; int hyperv_spinlock_attempts; + /* if true the CPUID code directly forward host cache leaves to the guest */ + bool fwd_host_cache_info; + /* Features that were filtered out because of missing host capabilities */ uint32_t filtered_features[FEATURE_WORDS]; diff --git a/target-i386/cpu.c b/target-i386/cpu.c index c36345e..f0df4db 100644 --- a/target-i386/cpu.c +++ b/target-i386/cpu.c @@ -486,6 +486,7 @@ typedef struct x86_def_t { int stepping; FeatureWordArray features; char model_id[48]; + bool fwd_host_cache_info; } x86_def_t; #define I486_FEATURES (CPUID_FP87 | CPUID_VME | CPUID_PSE) @@ -1139,6 +1140,7 @@ static void kvm_cpu_fill_host(x86_def_t *x86_cpu_def) assert(kvm_enabled()); x86_cpu_def->name = "host"; + x86_cpu_def->fwd_host_cache_info = true; host_cpuid(0x0, 0, &eax, &ebx, &ecx, &edx); x86_cpu_vendor_words2str(x86_cpu_def->vendor, ebx, edx, ecx); @@ -1888,6 +1890,7 @@ static void cpu_x86_register(X86CPU *cpu, const char *name, Error **errp) env->features[FEAT_C000_0001_EDX] = def->features[FEAT_C000_0001_EDX]; env->features[FEAT_7_0_EBX] = def->features[FEAT_7_0_EBX]; env->cpuid_xlevel2 = def->xlevel2; + cpu->fwd_host_cache_info = def->fwd_host_cache_info; object_property_set_str(OBJECT(cpu), def->model_id, "model-id", errp); } @@ -2062,6 +2065,10 @@ void cpu_x86_cpuid(CPUX86State *env, uint32_t index, uint32_t count, break; case 2: /* cache info: needed for Pentium Pro compatibility */ + if (cpu->fwd_host_cache_info) { + host_cpuid(index, 0, eax, ebx, ecx, edx); + break; + } *eax = 1; /* Number of CPUID[EAX=2] calls required */ *ebx = 0; *ecx = 0; @@ -2071,6 +2078,10 @@ void cpu_x86_cpuid(CPUX86State *env, uint32_t index, uint32_t count, break; case 4: /* cache info: needed for Core compatibility */ + if (cpu->fwd_host_cache_info) { + host_cpuid(index, count, eax, ebx, ecx, edx); + break; + } if (cs->nr_cores > 1) { *eax = (cs->nr_cores - 1) << 26; } else { @@ -2228,6 +2239,10 @@ void cpu_x86_cpuid(CPUX86State *env, uint32_t index, uint32_t count, break; case 0x80000005: /* cache info (L1 cache) */ + if (cpu->fwd_host_cache_info) { + host_cpuid(index, 0, eax, ebx, ecx, edx); + break; + } *eax = (L1_DTLB_2M_ASSOC << 24) | (L1_DTLB_2M_ENTRIES << 16) | \ (L1_ITLB_2M_ASSOC << 8) | (L1_ITLB_2M_ENTRIES); *ebx = (L1_DTLB_4K_ASSOC << 24) | (L1_DTLB_4K_ENTRIES << 16) | \ @@ -2239,6 +2254,10 @@ void cpu_x86_cpuid(CPUX86State *env, uint32_t index, uint32_t count, break; case 0x80000006: /* cache info (L2 cache) */ + if (cpu->fwd_host_cache_info) { + host_cpuid(index, 0, eax, ebx, ecx, edx); + break; + } *eax = (AMD_ENC_ASSOC(L2_DTLB_2M_ASSOC) << 28) | \ (L2_DTLB_2M_ENTRIES << 16) | \ (AMD_ENC_ASSOC(L2_ITLB_2M_ASSOC) << 12) | \ -- 1.7.10.4 ^ permalink raw reply related [flat|nested] 7+ messages in thread
* Re: [Qemu-devel] [PATCH V3] target-i386: forward CPUID cache leaves when -cpu host is used 2013-09-02 15:06 ` [Qemu-devel] [PATCH V3] target-i386: forward CPUID cache leaves when -cpu host is used Benoît Canet @ 2013-09-02 16:19 ` Eduardo Habkost 2013-09-02 17:09 ` Benoît Canet 2013-09-19 13:56 ` Benoît Canet 2013-09-19 17:33 ` Paolo Bonzini 1 sibling, 2 replies; 7+ messages in thread From: Eduardo Habkost @ 2013-09-02 16:19 UTC (permalink / raw) To: Benoît Canet; +Cc: imammedo, qemu-devel, kvm, afaerber On Mon, Sep 02, 2013 at 05:06:37PM +0200, Benoît Canet wrote: > Some users running cpu intensive tasks checking the cache CPUID leaves at > startup and making decisions based on the result reported that the guest was > not reflecting the host CPUID leaves when -cpu host is used. > > This patch fix this. > > Signed-off-by: Benoit Canet <benoit@irqsave.net> Reviewed-by: Eduardo Habkost <ehabkost@redhat.com> -- Eduardo ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [Qemu-devel] [PATCH V3] target-i386: forward CPUID cache leaves when -cpu host is used 2013-09-02 16:19 ` Eduardo Habkost @ 2013-09-02 17:09 ` Benoît Canet 2013-09-04 17:35 ` Eduardo Habkost 2013-09-19 13:56 ` Benoît Canet 1 sibling, 1 reply; 7+ messages in thread From: Benoît Canet @ 2013-09-02 17:09 UTC (permalink / raw) To: Eduardo Habkost; +Cc: imammedo, qemu-devel, kvm, afaerber > > > > Signed-off-by: Benoit Canet <benoit@irqsave.net> > > Reviewed-by: Eduardo Habkost <ehabkost@redhat.com> Thanks. Do you have an idea on how QEMU could reflect the real host clock frequency to the guest when the host cpu scaling governor kicks in ? Giving a false value to cloud customers is mildly annoying. Best regards Benoît ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [Qemu-devel] [PATCH V3] target-i386: forward CPUID cache leaves when -cpu host is used 2013-09-02 17:09 ` Benoît Canet @ 2013-09-04 17:35 ` Eduardo Habkost 0 siblings, 0 replies; 7+ messages in thread From: Eduardo Habkost @ 2013-09-04 17:35 UTC (permalink / raw) To: Benoît Canet; +Cc: imammedo, qemu-devel, kvm, afaerber On Mon, Sep 02, 2013 at 07:09:47PM +0200, Benoît Canet wrote: > > > > > > Signed-off-by: Benoit Canet <benoit@irqsave.net> > > > > Reviewed-by: Eduardo Habkost <ehabkost@redhat.com> > > Thanks. > > Do you have an idea on how QEMU could reflect the real host clock frequency > to the guest when the host cpu scaling governor kicks in ? > Giving a false value to cloud customers is mildly annoying. Probably you will need changes on KVM, SeaBIOS and QEMU to implement the interfaces to let the system notify the OS about CPU frequency changes. I don't know much a lot about ACPI and power management, to know how much of that is already implemented and how much is missing. -- Eduardo ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [Qemu-devel] [PATCH V3] target-i386: forward CPUID cache leaves when -cpu host is used 2013-09-02 16:19 ` Eduardo Habkost 2013-09-02 17:09 ` Benoît Canet @ 2013-09-19 13:56 ` Benoît Canet 1 sibling, 0 replies; 7+ messages in thread From: Benoît Canet @ 2013-09-19 13:56 UTC (permalink / raw) To: Eduardo Habkost; +Cc: imammedo, pbonzini, qemu-devel, kvm, afaerber Le Monday 02 Sep 2013 à 13:19:16 (-0300), Eduardo Habkost a écrit : > On Mon, Sep 02, 2013 at 05:06:37PM +0200, Benoît Canet wrote: > > Some users running cpu intensive tasks checking the cache CPUID leaves at > > startup and making decisions based on the result reported that the guest was > > not reflecting the host CPUID leaves when -cpu host is used. > > > > This patch fix this. > > > > Signed-off-by: Benoit Canet <benoit@irqsave.net> > > Reviewed-by: Eduardo Habkost <ehabkost@redhat.com> ping Paolo: is this patch ok for qemu/uq ? > > -- > Eduardo > ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [Qemu-devel] [PATCH V3] target-i386: forward CPUID cache leaves when -cpu host is used 2013-09-02 15:06 ` [Qemu-devel] [PATCH V3] target-i386: forward CPUID cache leaves when -cpu host is used Benoît Canet 2013-09-02 16:19 ` Eduardo Habkost @ 2013-09-19 17:33 ` Paolo Bonzini 1 sibling, 0 replies; 7+ messages in thread From: Paolo Bonzini @ 2013-09-19 17:33 UTC (permalink / raw) To: Benoît Canet; +Cc: imammedo, ehabkost, qemu-devel, kvm, afaerber Il 02/09/2013 17:06, Benoît Canet ha scritto: > Some users running cpu intensive tasks checking the cache CPUID leaves at > startup and making decisions based on the result reported that the guest was > not reflecting the host CPUID leaves when -cpu host is used. > > This patch fix this. > > Signed-off-by: Benoit Canet <benoit@irqsave.net> > --- > target-i386/cpu-qom.h | 3 +++ > target-i386/cpu.c | 19 +++++++++++++++++++ > 2 files changed, 22 insertions(+) > > diff --git a/target-i386/cpu-qom.h b/target-i386/cpu-qom.h > index c4447c2..b1d1bd8 100644 > --- a/target-i386/cpu-qom.h > +++ b/target-i386/cpu-qom.h > @@ -70,6 +70,9 @@ typedef struct X86CPU { > bool hyperv_relaxed_timing; > int hyperv_spinlock_attempts; > > + /* if true the CPUID code directly forward host cache leaves to the guest */ > + bool fwd_host_cache_info; > + > /* Features that were filtered out because of missing host capabilities */ > uint32_t filtered_features[FEATURE_WORDS]; > > diff --git a/target-i386/cpu.c b/target-i386/cpu.c > index c36345e..f0df4db 100644 > --- a/target-i386/cpu.c > +++ b/target-i386/cpu.c > @@ -486,6 +486,7 @@ typedef struct x86_def_t { > int stepping; > FeatureWordArray features; > char model_id[48]; > + bool fwd_host_cache_info; > } x86_def_t; > > #define I486_FEATURES (CPUID_FP87 | CPUID_VME | CPUID_PSE) > @@ -1139,6 +1140,7 @@ static void kvm_cpu_fill_host(x86_def_t *x86_cpu_def) > assert(kvm_enabled()); > > x86_cpu_def->name = "host"; > + x86_cpu_def->fwd_host_cache_info = true; > host_cpuid(0x0, 0, &eax, &ebx, &ecx, &edx); > x86_cpu_vendor_words2str(x86_cpu_def->vendor, ebx, edx, ecx); > > @@ -1888,6 +1890,7 @@ static void cpu_x86_register(X86CPU *cpu, const char *name, Error **errp) > env->features[FEAT_C000_0001_EDX] = def->features[FEAT_C000_0001_EDX]; > env->features[FEAT_7_0_EBX] = def->features[FEAT_7_0_EBX]; > env->cpuid_xlevel2 = def->xlevel2; > + cpu->fwd_host_cache_info = def->fwd_host_cache_info; > > object_property_set_str(OBJECT(cpu), def->model_id, "model-id", errp); > } > @@ -2062,6 +2065,10 @@ void cpu_x86_cpuid(CPUX86State *env, uint32_t index, uint32_t count, > break; > case 2: > /* cache info: needed for Pentium Pro compatibility */ > + if (cpu->fwd_host_cache_info) { > + host_cpuid(index, 0, eax, ebx, ecx, edx); > + break; > + } > *eax = 1; /* Number of CPUID[EAX=2] calls required */ > *ebx = 0; > *ecx = 0; > @@ -2071,6 +2078,10 @@ void cpu_x86_cpuid(CPUX86State *env, uint32_t index, uint32_t count, > break; > case 4: > /* cache info: needed for Core compatibility */ > + if (cpu->fwd_host_cache_info) { > + host_cpuid(index, count, eax, ebx, ecx, edx); > + break; > + } > if (cs->nr_cores > 1) { > *eax = (cs->nr_cores - 1) << 26; > } else { > @@ -2228,6 +2239,10 @@ void cpu_x86_cpuid(CPUX86State *env, uint32_t index, uint32_t count, > break; > case 0x80000005: > /* cache info (L1 cache) */ > + if (cpu->fwd_host_cache_info) { > + host_cpuid(index, 0, eax, ebx, ecx, edx); > + break; > + } > *eax = (L1_DTLB_2M_ASSOC << 24) | (L1_DTLB_2M_ENTRIES << 16) | \ > (L1_ITLB_2M_ASSOC << 8) | (L1_ITLB_2M_ENTRIES); > *ebx = (L1_DTLB_4K_ASSOC << 24) | (L1_DTLB_4K_ENTRIES << 16) | \ > @@ -2239,6 +2254,10 @@ void cpu_x86_cpuid(CPUX86State *env, uint32_t index, uint32_t count, > break; > case 0x80000006: > /* cache info (L2 cache) */ > + if (cpu->fwd_host_cache_info) { > + host_cpuid(index, 0, eax, ebx, ecx, edx); > + break; > + } > *eax = (AMD_ENC_ASSOC(L2_DTLB_2M_ASSOC) << 28) | \ > (L2_DTLB_2M_ENTRIES << 16) | \ > (AMD_ENC_ASSOC(L2_ITLB_2M_ASSOC) << 12) | \ > I renamed the new field to cache_info_passthrough (Eduardo had a "pmu_passthrough" patch a few weeks ago) and will push it tomorrow to uq/master. Thanks, Paolo ^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2013-09-19 17:33 UTC | newest] Thread overview: 7+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2013-09-02 15:06 [Qemu-devel] [PATCH V3] forward cpuid leaves when using -cpu host Benoît Canet 2013-09-02 15:06 ` [Qemu-devel] [PATCH V3] target-i386: forward CPUID cache leaves when -cpu host is used Benoît Canet 2013-09-02 16:19 ` Eduardo Habkost 2013-09-02 17:09 ` Benoît Canet 2013-09-04 17:35 ` Eduardo Habkost 2013-09-19 13:56 ` Benoît Canet 2013-09-19 17:33 ` Paolo Bonzini
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).