* [Qemu-devel] [PATCH 0/2] Disable kvm_mmu_op by default on pc-1.4 @ 2013-01-04 14:52 Eduardo Habkost 2013-01-04 14:52 ` [Qemu-devel] [PATCH 1/2] target-i386: Don't set any KVM flag by default if KVM is disabled Eduardo Habkost 2013-01-04 14:52 ` [Qemu-devel] [PATCH 2/2] target-i386: Disable kvm_mmu_op by default on pc-1.4 Eduardo Habkost 0 siblings, 2 replies; 8+ messages in thread From: Eduardo Habkost @ 2013-01-04 14:52 UTC (permalink / raw) To: qemu-devel Cc: kvm, Gleb Natapov, Michael S. Tsirkin, Marcelo Tosatti, Igor Mammedov, Andreas Färber The kvm_mmu_op feature was removed from the kernel since v3.3 (released in March 2012), it was marked for removal since January 2011 and it's slower than shadow or hardware assisted paging (see kernel commit fb92045843). It doesn't make sense to keep it enabled by default. Also, keeping it enabled by default would cause unnecessary hassle when libvirt start using the "enforce" option. Eduardo Habkost (2): target-i386: don't set any KVM flag by default if KVM is disabled target-i386: disable kvm_mmu_op by default on pc-1.4 hw/pc_piix.c | 11 ++++++++++- target-i386/cpu.c | 15 ++++++++++++--- target-i386/cpu.h | 1 + 3 files changed, 23 insertions(+), 4 deletions(-) -- 1.7.11.7 ^ permalink raw reply [flat|nested] 8+ messages in thread
* [Qemu-devel] [PATCH 1/2] target-i386: Don't set any KVM flag by default if KVM is disabled 2013-01-04 14:52 [Qemu-devel] [PATCH 0/2] Disable kvm_mmu_op by default on pc-1.4 Eduardo Habkost @ 2013-01-04 14:52 ` Eduardo Habkost 2013-01-04 20:47 ` Blue Swirl 2013-01-04 14:52 ` [Qemu-devel] [PATCH 2/2] target-i386: Disable kvm_mmu_op by default on pc-1.4 Eduardo Habkost 1 sibling, 1 reply; 8+ messages in thread From: Eduardo Habkost @ 2013-01-04 14:52 UTC (permalink / raw) To: qemu-devel Cc: kvm, Gleb Natapov, Michael S. Tsirkin, Marcelo Tosatti, Igor Mammedov, Andreas Färber This is a cleanup that tries to solve two small issues: - We don't need a separate kvm_pv_eoi_features variable just to keep a constant calculated at compile-time, and this style would require adding a separate variable (that's declared twice because of the CONFIG_KVM ifdef) for each feature that's going to be enabled/disable by machine-type compat code. - The pc-1.3 code is setting the kvm_pv_eoi flag on cpuid_kvm_features even when KVM is disabled at runtime. This small incosistency in the cpuid_kvm_features field isn't a problem today because cpuid_kvm_features is ignored by the TCG code, but it may cause unexpected problems later when refactoring the CPUID handling code. This patch eliminates the kvm_pv_eoi_features variable and simply uses CONFIG_KVM and kvm_enabled() inside the enable_kvm_pv_eoi() compat function, so it enables kvm_pv_eoi only if KVM is enabled. I believe this makes the behavior of enable_kvm_pv_eoi() clearer and easier to understand. Signed-off-by: Eduardo Habkost <ehabkost@redhat.com> --- Cc: kvm@vger.kernel.org Cc: Michael S. Tsirkin <mst@redhat.com> Cc: Gleb Natapov <gleb@redhat.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> --- target-i386/cpu.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/target-i386/cpu.c b/target-i386/cpu.c index 82685dc..808001a 100644 --- a/target-i386/cpu.c +++ b/target-i386/cpu.c @@ -145,15 +145,16 @@ static uint32_t kvm_default_features = (1 << KVM_FEATURE_CLOCKSOURCE) | (1 << KVM_FEATURE_ASYNC_PF) | (1 << KVM_FEATURE_STEAL_TIME) | (1 << KVM_FEATURE_CLOCKSOURCE_STABLE_BIT); -static const uint32_t kvm_pv_eoi_features = (0x1 << KVM_FEATURE_PV_EOI); #else static uint32_t kvm_default_features = 0; -static const uint32_t kvm_pv_eoi_features = 0; #endif void enable_kvm_pv_eoi(void) { - kvm_default_features |= kvm_pv_eoi_features; +#ifdef CONFIG_KVM + if (kvm_enabled()) + kvm_default_features |= (1UL << KVM_FEATURE_PV_EOI); +#endif } void host_cpuid(uint32_t function, uint32_t count, -- 1.7.11.7 ^ permalink raw reply related [flat|nested] 8+ messages in thread
* Re: [Qemu-devel] [PATCH 1/2] target-i386: Don't set any KVM flag by default if KVM is disabled 2013-01-04 14:52 ` [Qemu-devel] [PATCH 1/2] target-i386: Don't set any KVM flag by default if KVM is disabled Eduardo Habkost @ 2013-01-04 20:47 ` Blue Swirl 2013-01-04 21:02 ` Eduardo Habkost 0 siblings, 1 reply; 8+ messages in thread From: Blue Swirl @ 2013-01-04 20:47 UTC (permalink / raw) To: Eduardo Habkost Cc: kvm, Gleb Natapov, Michael S. Tsirkin, Marcelo Tosatti, qemu-devel, Igor Mammedov, Andreas Färber On Fri, Jan 4, 2013 at 2:52 PM, Eduardo Habkost <ehabkost@redhat.com> wrote: > This is a cleanup that tries to solve two small issues: > > - We don't need a separate kvm_pv_eoi_features variable just to keep a > constant calculated at compile-time, and this style would require > adding a separate variable (that's declared twice because of the > CONFIG_KVM ifdef) for each feature that's going to be enabled/disable > by machine-type compat code. > - The pc-1.3 code is setting the kvm_pv_eoi flag on cpuid_kvm_features > even when KVM is disabled at runtime. This small incosistency in > the cpuid_kvm_features field isn't a problem today because > cpuid_kvm_features is ignored by the TCG code, but it may cause > unexpected problems later when refactoring the CPUID handling code. > > This patch eliminates the kvm_pv_eoi_features variable and simply uses > CONFIG_KVM and kvm_enabled() inside the enable_kvm_pv_eoi() compat > function, so it enables kvm_pv_eoi only if KVM is enabled. I believe > this makes the behavior of enable_kvm_pv_eoi() clearer and easier to > understand. > > Signed-off-by: Eduardo Habkost <ehabkost@redhat.com> > --- > Cc: kvm@vger.kernel.org > Cc: Michael S. Tsirkin <mst@redhat.com> > Cc: Gleb Natapov <gleb@redhat.com> > Cc: Marcelo Tosatti <mtosatti@redhat.com> > --- > target-i386/cpu.c | 7 ++++--- > 1 file changed, 4 insertions(+), 3 deletions(-) > > diff --git a/target-i386/cpu.c b/target-i386/cpu.c > index 82685dc..808001a 100644 > --- a/target-i386/cpu.c > +++ b/target-i386/cpu.c > @@ -145,15 +145,16 @@ static uint32_t kvm_default_features = (1 << KVM_FEATURE_CLOCKSOURCE) | > (1 << KVM_FEATURE_ASYNC_PF) | > (1 << KVM_FEATURE_STEAL_TIME) | > (1 << KVM_FEATURE_CLOCKSOURCE_STABLE_BIT); > -static const uint32_t kvm_pv_eoi_features = (0x1 << KVM_FEATURE_PV_EOI); > #else > static uint32_t kvm_default_features = 0; > -static const uint32_t kvm_pv_eoi_features = 0; > #endif > > void enable_kvm_pv_eoi(void) > { > - kvm_default_features |= kvm_pv_eoi_features; > +#ifdef CONFIG_KVM > + if (kvm_enabled()) Missing braces, please read CODING_STYLE and use checkpatch.pl to find problems in patches. > + kvm_default_features |= (1UL << KVM_FEATURE_PV_EOI); > +#endif > } > > void host_cpuid(uint32_t function, uint32_t count, > -- > 1.7.11.7 > > ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [Qemu-devel] [PATCH 1/2] target-i386: Don't set any KVM flag by default if KVM is disabled 2013-01-04 20:47 ` Blue Swirl @ 2013-01-04 21:02 ` Eduardo Habkost 2013-01-04 21:25 ` Eduardo Habkost 0 siblings, 1 reply; 8+ messages in thread From: Eduardo Habkost @ 2013-01-04 21:02 UTC (permalink / raw) To: Blue Swirl Cc: kvm, Gleb Natapov, Michael S. Tsirkin, Marcelo Tosatti, qemu-devel, Igor Mammedov, Andreas Färber On Fri, Jan 04, 2013 at 08:47:07PM +0000, Blue Swirl wrote: > > { > > - kvm_default_features |= kvm_pv_eoi_features; > > +#ifdef CONFIG_KVM > > + if (kvm_enabled()) > > Missing braces, please read CODING_STYLE and use checkpatch.pl to find > problems in patches. Sorry (again). I will soon send a new version fixing the coding style problems. -- Eduardo ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [Qemu-devel] [PATCH 1/2] target-i386: Don't set any KVM flag by default if KVM is disabled 2013-01-04 21:02 ` Eduardo Habkost @ 2013-01-04 21:25 ` Eduardo Habkost 0 siblings, 0 replies; 8+ messages in thread From: Eduardo Habkost @ 2013-01-04 21:25 UTC (permalink / raw) To: Blue Swirl Cc: kvm, Gleb Natapov, Michael S. Tsirkin, Marcelo Tosatti, qemu-devel, Igor Mammedov, Andreas Färber On Fri, Jan 04, 2013 at 07:02:17PM -0200, Eduardo Habkost wrote: > On Fri, Jan 04, 2013 at 08:47:07PM +0000, Blue Swirl wrote: > > > { > > > - kvm_default_features |= kvm_pv_eoi_features; > > > +#ifdef CONFIG_KVM > > > + if (kvm_enabled()) > > > > Missing braces, please read CODING_STYLE and use checkpatch.pl to find > > problems in patches. > > Sorry (again). I will soon send a new version fixing the coding style > problems. checkpatch.pl is behaving weirdly, here: $ git show 193e2f6 | grep 'if (kvm' + if (kvm_enabled()) $ git show 193e2f6 | ./scripts/checkpatch.pl - total: 0 errors, 0 warnings, 19 lines checked Your patch has no obvious style problems and is ready for submission. $ -- Eduardo ^ permalink raw reply [flat|nested] 8+ messages in thread
* [Qemu-devel] [PATCH 2/2] target-i386: Disable kvm_mmu_op by default on pc-1.4 2013-01-04 14:52 [Qemu-devel] [PATCH 0/2] Disable kvm_mmu_op by default on pc-1.4 Eduardo Habkost 2013-01-04 14:52 ` [Qemu-devel] [PATCH 1/2] target-i386: Don't set any KVM flag by default if KVM is disabled Eduardo Habkost @ 2013-01-04 14:52 ` Eduardo Habkost 2013-01-04 20:48 ` Blue Swirl 1 sibling, 1 reply; 8+ messages in thread From: Eduardo Habkost @ 2013-01-04 14:52 UTC (permalink / raw) To: qemu-devel Cc: kvm, Gleb Natapov, Michael S. Tsirkin, libvir-list, Marcelo Tosatti, Igor Mammedov, Jiri Denemark, Andreas Färber The kvm_mmu_op feature was removed from the kernel since v3.3 (released in March 2012), it was marked for removal since January 2011 and it's slower than shadow or hardware assisted paging (see kernel commit fb92045843). It doesn't make sense to keep it enabled by default. Also, keeping it enabled by default would cause unnecessary hassle when libvirt start using the "enforce" option. Signed-off-by: Eduardo Habkost <ehabkost@redhat.com> --- Cc: kvm@vger.kernel.org Cc: Michael S. Tsirkin <mst@redhat.com> Cc: Gleb Natapov <gleb@redhat.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: libvir-list@redhat.com Cc: Jiri Denemark <jdenemar@redhat.com> I was planning to reverse the logic of the compat init functions and make pc_init_pci_1_3() enable kvm_mmu_op and then call pc_init_pci_1_4(). But that would require changing pc_init_pci_no_kvmclock() and pc_init_isa() as well. So to keep the changes simple, I am keeping the pattern used when pc_init_pci_1_3() was introduced, making pc_init_pci_1_4() disable kvm_mmu_op and then call pc_init_pci_1_3(). --- hw/pc_piix.c | 11 ++++++++++- target-i386/cpu.c | 8 ++++++++ target-i386/cpu.h | 1 + 3 files changed, 19 insertions(+), 1 deletion(-) diff --git a/hw/pc_piix.c b/hw/pc_piix.c index 99747a7..a6bf645 100644 --- a/hw/pc_piix.c +++ b/hw/pc_piix.c @@ -217,6 +217,7 @@ static void pc_init1(MemoryRegion *system_memory, } } +/* machine init function for pc-0.14 - pc-1.2 */ static void pc_init_pci(QEMUMachineInitArgs *args) { ram_addr_t ram_size = args->ram_size; @@ -232,12 +233,20 @@ static void pc_init_pci(QEMUMachineInitArgs *args) initrd_filename, cpu_model, 1, 1); } +/* machine init function for pc-1.3 */ static void pc_init_pci_1_3(QEMUMachineInitArgs *args) { enable_kvm_pv_eoi(); pc_init_pci(args); } +/* machine init function for pc-1.4 */ +static void pc_init_pci_1_4(QEMUMachineInitArgs *args) +{ + disable_kvm_mmu_op(); + pc_init_pci_1_3(args); +} + static void pc_init_pci_no_kvmclock(QEMUMachineInitArgs *args) { ram_addr_t ram_size = args->ram_size; @@ -285,7 +294,7 @@ static QEMUMachine pc_machine_v1_4 = { .name = "pc-1.4", .alias = "pc", .desc = "Standard PC", - .init = pc_init_pci_1_3, + .init = pc_init_pci_1_4, .max_cpus = 255, .is_default = 1, }; diff --git a/target-i386/cpu.c b/target-i386/cpu.c index 808001a..ec877c7 100644 --- a/target-i386/cpu.c +++ b/target-i386/cpu.c @@ -157,6 +157,14 @@ void enable_kvm_pv_eoi(void) #endif } +void disable_kvm_mmu_op(void) +{ +#ifdef CONFIG_KVM + if (kvm_enabled()) + kvm_default_features &= ~(1UL << KVM_FEATURE_MMU_OP); +#endif +} + void host_cpuid(uint32_t function, uint32_t count, uint32_t *eax, uint32_t *ebx, uint32_t *ecx, uint32_t *edx) { diff --git a/target-i386/cpu.h b/target-i386/cpu.h index 1283537..27c8d0c 100644 --- a/target-i386/cpu.h +++ b/target-i386/cpu.h @@ -1219,5 +1219,6 @@ void do_smm_enter(CPUX86State *env1); void cpu_report_tpr_access(CPUX86State *env, TPRAccess access); void enable_kvm_pv_eoi(void); +void disable_kvm_mmu_op(void); #endif /* CPU_I386_H */ -- 1.7.11.7 ^ permalink raw reply related [flat|nested] 8+ messages in thread
* Re: [Qemu-devel] [PATCH 2/2] target-i386: Disable kvm_mmu_op by default on pc-1.4 2013-01-04 14:52 ` [Qemu-devel] [PATCH 2/2] target-i386: Disable kvm_mmu_op by default on pc-1.4 Eduardo Habkost @ 2013-01-04 20:48 ` Blue Swirl 2013-01-04 21:48 ` Eduardo Habkost 0 siblings, 1 reply; 8+ messages in thread From: Blue Swirl @ 2013-01-04 20:48 UTC (permalink / raw) To: Eduardo Habkost Cc: kvm, Gleb Natapov, Michael S. Tsirkin, libvir-list, Marcelo Tosatti, qemu-devel, Igor Mammedov, Jiri Denemark, Andreas Färber On Fri, Jan 4, 2013 at 2:52 PM, Eduardo Habkost <ehabkost@redhat.com> wrote: > The kvm_mmu_op feature was removed from the kernel since v3.3 (released > in March 2012), it was marked for removal since January 2011 and it's > slower than shadow or hardware assisted paging (see kernel commit > fb92045843). It doesn't make sense to keep it enabled by default. > > Also, keeping it enabled by default would cause unnecessary hassle when > libvirt start using the "enforce" option. > > Signed-off-by: Eduardo Habkost <ehabkost@redhat.com> > --- > Cc: kvm@vger.kernel.org > Cc: Michael S. Tsirkin <mst@redhat.com> > Cc: Gleb Natapov <gleb@redhat.com> > Cc: Marcelo Tosatti <mtosatti@redhat.com> > Cc: libvir-list@redhat.com > Cc: Jiri Denemark <jdenemar@redhat.com> > > I was planning to reverse the logic of the compat init functions and make > pc_init_pci_1_3() enable kvm_mmu_op and then call pc_init_pci_1_4(). But that > would require changing pc_init_pci_no_kvmclock() and pc_init_isa() as well. So > to keep the changes simple, I am keeping the pattern used when pc_init_pci_1_3() > was introduced, making pc_init_pci_1_4() disable kvm_mmu_op and then call > pc_init_pci_1_3(). > > --- > hw/pc_piix.c | 11 ++++++++++- > target-i386/cpu.c | 8 ++++++++ > target-i386/cpu.h | 1 + > 3 files changed, 19 insertions(+), 1 deletion(-) > > diff --git a/hw/pc_piix.c b/hw/pc_piix.c > index 99747a7..a6bf645 100644 > --- a/hw/pc_piix.c > +++ b/hw/pc_piix.c > @@ -217,6 +217,7 @@ static void pc_init1(MemoryRegion *system_memory, > } > } > > +/* machine init function for pc-0.14 - pc-1.2 */ > static void pc_init_pci(QEMUMachineInitArgs *args) > { > ram_addr_t ram_size = args->ram_size; > @@ -232,12 +233,20 @@ static void pc_init_pci(QEMUMachineInitArgs *args) > initrd_filename, cpu_model, 1, 1); > } > > +/* machine init function for pc-1.3 */ The comment does give much information compared to the function name. > static void pc_init_pci_1_3(QEMUMachineInitArgs *args) > { > enable_kvm_pv_eoi(); > pc_init_pci(args); > } > > +/* machine init function for pc-1.4 */ Ditto. > +static void pc_init_pci_1_4(QEMUMachineInitArgs *args) > +{ > + disable_kvm_mmu_op(); > + pc_init_pci_1_3(args); > +} > + > static void pc_init_pci_no_kvmclock(QEMUMachineInitArgs *args) > { > ram_addr_t ram_size = args->ram_size; > @@ -285,7 +294,7 @@ static QEMUMachine pc_machine_v1_4 = { > .name = "pc-1.4", > .alias = "pc", > .desc = "Standard PC", > - .init = pc_init_pci_1_3, > + .init = pc_init_pci_1_4, > .max_cpus = 255, > .is_default = 1, > }; > diff --git a/target-i386/cpu.c b/target-i386/cpu.c > index 808001a..ec877c7 100644 > --- a/target-i386/cpu.c > +++ b/target-i386/cpu.c > @@ -157,6 +157,14 @@ void enable_kvm_pv_eoi(void) > #endif > } > > +void disable_kvm_mmu_op(void) > +{ > +#ifdef CONFIG_KVM > + if (kvm_enabled()) Braces. > + kvm_default_features &= ~(1UL << KVM_FEATURE_MMU_OP); > +#endif > +} > + > void host_cpuid(uint32_t function, uint32_t count, > uint32_t *eax, uint32_t *ebx, uint32_t *ecx, uint32_t *edx) > { > diff --git a/target-i386/cpu.h b/target-i386/cpu.h > index 1283537..27c8d0c 100644 > --- a/target-i386/cpu.h > +++ b/target-i386/cpu.h > @@ -1219,5 +1219,6 @@ void do_smm_enter(CPUX86State *env1); > void cpu_report_tpr_access(CPUX86State *env, TPRAccess access); > > void enable_kvm_pv_eoi(void); > +void disable_kvm_mmu_op(void); > > #endif /* CPU_I386_H */ > -- > 1.7.11.7 > > ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [Qemu-devel] [PATCH 2/2] target-i386: Disable kvm_mmu_op by default on pc-1.4 2013-01-04 20:48 ` Blue Swirl @ 2013-01-04 21:48 ` Eduardo Habkost 0 siblings, 0 replies; 8+ messages in thread From: Eduardo Habkost @ 2013-01-04 21:48 UTC (permalink / raw) To: Blue Swirl Cc: kvm, Gleb Natapov, Michael S. Tsirkin, libvir-list, Marcelo Tosatti, qemu-devel, Igor Mammedov, Jiri Denemark, Andreas Färber On Fri, Jan 04, 2013 at 08:48:42PM +0000, Blue Swirl wrote: [...] > > +/* machine init function for pc-0.14 - pc-1.2 */ > > static void pc_init_pci(QEMUMachineInitArgs *args) > > { > > ram_addr_t ram_size = args->ram_size; > > @@ -232,12 +233,20 @@ static void pc_init_pci(QEMUMachineInitArgs *args) > > initrd_filename, cpu_model, 1, 1); > > } > > > > +/* machine init function for pc-1.3 */ > > The comment does give much information compared to the function name. You can't be sure from the function name if it's just for 1.3 or for 1.3 and higher/lower. e.g. before this patch, the pc_init_pci_1_3() function was used by both pc-1.3 and pc-1.4. But I will change this, anyway: I will document only the non-trivial cases (that don't apply to only one machine-type version), it should be better than redundant comments. [...] > > +void disable_kvm_mmu_op(void) > > +{ > > +#ifdef CONFIG_KVM > > + if (kvm_enabled()) > > Braces. Will be fixed in the next version. -- Eduardo ^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2013-01-04 21:46 UTC | newest] Thread overview: 8+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2013-01-04 14:52 [Qemu-devel] [PATCH 0/2] Disable kvm_mmu_op by default on pc-1.4 Eduardo Habkost 2013-01-04 14:52 ` [Qemu-devel] [PATCH 1/2] target-i386: Don't set any KVM flag by default if KVM is disabled Eduardo Habkost 2013-01-04 20:47 ` Blue Swirl 2013-01-04 21:02 ` Eduardo Habkost 2013-01-04 21:25 ` Eduardo Habkost 2013-01-04 14:52 ` [Qemu-devel] [PATCH 2/2] target-i386: Disable kvm_mmu_op by default on pc-1.4 Eduardo Habkost 2013-01-04 20:48 ` Blue Swirl 2013-01-04 21:48 ` Eduardo Habkost
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).