If a HVM/PVH guest writes to MSR_IA32_TSC{_ADJUST} and thus changes
the value of the time stamp counter the vcpu time info must also be
updated, or the time calculated by the guest using the Xen PV clock
interface will be skewed.
Update the vcpu time info when the guest writes to either MSR_IA32_TSC
or MSR_IA32_TSC_ADJUST. This fixes lockups seen when running the
pv-shim on AMD hardware, since the shim will aggressively try to keep
TSCs in sync by periodically writing to MSR_IA32_TSC if the TSC is not
reliable.
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
---
Cc: Jürgen Groß <jgross@suse.com>
---
xen/arch/x86/hvm/hvm.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/xen/arch/x86/hvm/hvm.c b/xen/arch/x86/hvm/hvm.c
index 9acd359c99..e4c0425330 100644
--- a/xen/arch/x86/hvm/hvm.c
+++ b/xen/arch/x86/hvm/hvm.c
@@ -442,6 +442,8 @@ static void hvm_set_guest_tsc_msr(struct vcpu *v, u64 guest_tsc)
hvm_set_guest_tsc(v, guest_tsc);
v->arch.hvm.msr_tsc_adjust += v->arch.hvm.cache_tsc_offset - tsc_offset;
+ if ( v == current )
+ update_vcpu_system_time(v);
}
static void hvm_set_guest_tsc_adjust(struct vcpu *v, u64 tsc_adjust)
@@ -449,6 +451,8 @@ static void hvm_set_guest_tsc_adjust(struct vcpu *v, u64 tsc_adjust)
v->arch.hvm.cache_tsc_offset += tsc_adjust - v->arch.hvm.msr_tsc_adjust;
hvm_set_tsc_offset(v, v->arch.hvm.cache_tsc_offset, 0);
v->arch.hvm.msr_tsc_adjust = tsc_adjust;
+ if ( v == current )
+ update_vcpu_system_time(v);
}
u64 hvm_get_guest_tsc_fixed(struct vcpu *v, uint64_t at_tsc)
--
2.23.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
On 23.10.19 10:57, Roger Pau Monne wrote: > If a HVM/PVH guest writes to MSR_IA32_TSC{_ADJUST} and thus changes > the value of the time stamp counter the vcpu time info must also be > updated, or the time calculated by the guest using the Xen PV clock > interface will be skewed. > > Update the vcpu time info when the guest writes to either MSR_IA32_TSC > or MSR_IA32_TSC_ADJUST. This fixes lockups seen when running the > pv-shim on AMD hardware, since the shim will aggressively try to keep > TSCs in sync by periodically writing to MSR_IA32_TSC if the TSC is not > reliable. > > Signed-off-by: Roger Pau Monné <roger.pau@citrix.com> Release-acked-by: Juergen Gross <jgross@suse.com> Juergen _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel
On Wed, Oct 23, 2019 at 10:57:39AM +0200, Roger Pau Monne wrote: > If a HVM/PVH guest writes to MSR_IA32_TSC{_ADJUST} and thus changes > the value of the time stamp counter the vcpu time info must also be > updated, or the time calculated by the guest using the Xen PV clock > interface will be skewed. > > Update the vcpu time info when the guest writes to either MSR_IA32_TSC > or MSR_IA32_TSC_ADJUST. This fixes lockups seen when running the > pv-shim on AMD hardware, since the shim will aggressively try to keep > TSCs in sync by periodically writing to MSR_IA32_TSC if the TSC is not > reliable. > > Signed-off-by: Roger Pau Monné <roger.pau@citrix.com> > --- > Cc: Jürgen Groß <jgross@suse.com> > --- > xen/arch/x86/hvm/hvm.c | 4 ++++ > 1 file changed, 4 insertions(+) > > diff --git a/xen/arch/x86/hvm/hvm.c b/xen/arch/x86/hvm/hvm.c > index 9acd359c99..e4c0425330 100644 > --- a/xen/arch/x86/hvm/hvm.c > +++ b/xen/arch/x86/hvm/hvm.c > @@ -442,6 +442,8 @@ static void hvm_set_guest_tsc_msr(struct vcpu *v, u64 guest_tsc) > > hvm_set_guest_tsc(v, guest_tsc); > v->arch.hvm.msr_tsc_adjust += v->arch.hvm.cache_tsc_offset - tsc_offset; > + if ( v == current ) > + update_vcpu_system_time(v); Why not get rid of the test and always call update_vcpu_system_time(v)? Wei. > } > > static void hvm_set_guest_tsc_adjust(struct vcpu *v, u64 tsc_adjust) > @@ -449,6 +451,8 @@ static void hvm_set_guest_tsc_adjust(struct vcpu *v, u64 tsc_adjust) > v->arch.hvm.cache_tsc_offset += tsc_adjust - v->arch.hvm.msr_tsc_adjust; > hvm_set_tsc_offset(v, v->arch.hvm.cache_tsc_offset, 0); > v->arch.hvm.msr_tsc_adjust = tsc_adjust; > + if ( v == current ) > + update_vcpu_system_time(v); > } > > u64 hvm_get_guest_tsc_fixed(struct vcpu *v, uint64_t at_tsc) > -- > 2.23.0 > _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel
On Wed, Oct 23, 2019 at 11:55:13AM +0100, Wei Liu wrote: > On Wed, Oct 23, 2019 at 10:57:39AM +0200, Roger Pau Monne wrote: > > If a HVM/PVH guest writes to MSR_IA32_TSC{_ADJUST} and thus changes > > the value of the time stamp counter the vcpu time info must also be > > updated, or the time calculated by the guest using the Xen PV clock > > interface will be skewed. > > > > Update the vcpu time info when the guest writes to either MSR_IA32_TSC > > or MSR_IA32_TSC_ADJUST. This fixes lockups seen when running the > > pv-shim on AMD hardware, since the shim will aggressively try to keep > > TSCs in sync by periodically writing to MSR_IA32_TSC if the TSC is not > > reliable. > > > > Signed-off-by: Roger Pau Monné <roger.pau@citrix.com> > > --- > > Cc: Jürgen Groß <jgross@suse.com> > > --- > > xen/arch/x86/hvm/hvm.c | 4 ++++ > > 1 file changed, 4 insertions(+) > > > > diff --git a/xen/arch/x86/hvm/hvm.c b/xen/arch/x86/hvm/hvm.c > > index 9acd359c99..e4c0425330 100644 > > --- a/xen/arch/x86/hvm/hvm.c > > +++ b/xen/arch/x86/hvm/hvm.c > > @@ -442,6 +442,8 @@ static void hvm_set_guest_tsc_msr(struct vcpu *v, u64 guest_tsc) > > > > hvm_set_guest_tsc(v, guest_tsc); > > v->arch.hvm.msr_tsc_adjust += v->arch.hvm.cache_tsc_offset - tsc_offset; > > + if ( v == current ) > > + update_vcpu_system_time(v); > > Why not get rid of the test and always call update_vcpu_system_time(v)? If the vcpu is not running the update will be done by the scheduling code in schedule_tail. Calling update_vcpu_system_time unconditionally would work, I was just trying to avoid the overhead if it's not really needed. Thanks, Roger. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel
On Wed, Oct 23, 2019 at 01:05:14PM +0200, Roger Pau Monné wrote: > On Wed, Oct 23, 2019 at 11:55:13AM +0100, Wei Liu wrote: > > On Wed, Oct 23, 2019 at 10:57:39AM +0200, Roger Pau Monne wrote: > > > If a HVM/PVH guest writes to MSR_IA32_TSC{_ADJUST} and thus changes > > > the value of the time stamp counter the vcpu time info must also be > > > updated, or the time calculated by the guest using the Xen PV clock > > > interface will be skewed. > > > > > > Update the vcpu time info when the guest writes to either MSR_IA32_TSC > > > or MSR_IA32_TSC_ADJUST. This fixes lockups seen when running the > > > pv-shim on AMD hardware, since the shim will aggressively try to keep > > > TSCs in sync by periodically writing to MSR_IA32_TSC if the TSC is not > > > reliable. > > > > > > Signed-off-by: Roger Pau Monné <roger.pau@citrix.com> > > > --- > > > Cc: Jürgen Groß <jgross@suse.com> > > > --- > > > xen/arch/x86/hvm/hvm.c | 4 ++++ > > > 1 file changed, 4 insertions(+) > > > > > > diff --git a/xen/arch/x86/hvm/hvm.c b/xen/arch/x86/hvm/hvm.c > > > index 9acd359c99..e4c0425330 100644 > > > --- a/xen/arch/x86/hvm/hvm.c > > > +++ b/xen/arch/x86/hvm/hvm.c > > > @@ -442,6 +442,8 @@ static void hvm_set_guest_tsc_msr(struct vcpu *v, u64 guest_tsc) > > > > > > hvm_set_guest_tsc(v, guest_tsc); > > > v->arch.hvm.msr_tsc_adjust += v->arch.hvm.cache_tsc_offset - tsc_offset; > > > + if ( v == current ) > > > + update_vcpu_system_time(v); > > > > Why not get rid of the test and always call update_vcpu_system_time(v)? > > If the vcpu is not running the update will be done by the scheduling > code in schedule_tail. Calling update_vcpu_system_time unconditionally > would work, I was just trying to avoid the overhead if it's not really > needed. > Fair enough. Reviewed-by: Wei Liu <wl@xen.org> _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel
On 23.10.2019 10:57, Roger Pau Monne wrote: > If a HVM/PVH guest writes to MSR_IA32_TSC{_ADJUST} and thus changes > the value of the time stamp counter the vcpu time info must also be > updated, or the time calculated by the guest using the Xen PV clock > interface will be skewed. > > Update the vcpu time info when the guest writes to either MSR_IA32_TSC > or MSR_IA32_TSC_ADJUST. This fixes lockups seen when running the > pv-shim on AMD hardware, since the shim will aggressively try to keep > TSCs in sync by periodically writing to MSR_IA32_TSC if the TSC is not > reliable. > > Signed-off-by: Roger Pau Monné <roger.pau@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com> _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel
© 2016 - 2024 Red Hat, Inc.