Skip to contentSkip to conversation

x86/nvmx: fixes for interrupt injection

PATCH Superseded Roger Pau Monne 2 participants 12 messages Mar 25, 2020 – Mar 26, 2020

v2 of 4 Superseded Latest is v4 →

lore
Share & tools
lore
#0 lore raw

v2 cover x86/nvmx: fixes for interrupt injection

Hello,

osstest identified a regression caused by my earlier attempt to fix interrupt injection when using nested VMX. This series aims to fix the regression, and should unblock several osstest branches.

The following report is from osstest with this series applied:

http://logs.test-lab.xenproject.org/osstest/logs/148979/

Note the last patch (4/4) is the one that actually fixes the issue. Xen will always use the Ack on exit feature so patches 2/4 and 3/4 don't change the functionality when running a nested Xen, as it always requires SVI to be updated.

Thanks, Roger.

Roger Pau Monne (4):
Revert "x86/vvmx: fix virtual interrupt injection when Ack on exit
control is used"
x86/nvmx: only update SVI when using Ack on exit x86/nvmx: split updating RVI from SVI in nvmx_update_apicv x86/nvmx: update exit bitmap when using virtual interrupt delivery

 xen/arch/x86/hvm/vmx/intr.c       | 21 ++++++++++++--------
 xen/arch/x86/hvm/vmx/vvmx.c       | 33 ++++++++++++++++++++-----------
 xen/include/asm-x86/hvm/vmx/vmx.h |  2 ++
 3 files changed, 36 insertions(+), 20 deletions(-)
Signature · 1 line
2.26.0
#1 lore raw

PATCH v2 1/4 Revert "x86/vvmx: fix virtual interrupt injection when Ack on exit control is used"

This reverts commit f96e1469ad06b61796c60193daaeb9f8a96d7458.

The commit is wrong, as the whole point of nvmx_update_apicv is to update the guest interrupt status field when the Ack on exit VMEXIT control feature is enabled.

---
 xen/arch/x86/hvm/vmx/vvmx.c | 7 +------
 1 file changed, 1 insertion(+), 6 deletions(-)
xen/arch/x86/hvm/vmx/vvmx.c +1 -6
index f049920196..1b8461ba30 100644
@@ -1456,12 +1456,7 @@ static void virtual_vmexit(struct cpu_user_regs *regs)
     /* updating host cr0 to sync TS bit */
     __vmwrite(HOST_CR0, v->arch.hvm.vmx.host_cr0);
 
-    if ( cpu_has_vmx_virtual_intr_delivery &&
-         /*
-          * Only inject the vector if the Ack on exit bit is not set, else the
-          * interrupt will be signaled in the vmcs VM_EXIT_INTR_INFO field.
-          */
-         !(get_vvmcs(v, VM_EXIT_CONTROLS) & VM_EXIT_ACK_INTR_ON_EXIT) )
+    if ( cpu_has_vmx_virtual_intr_delivery )
         nvmx_update_apicv(v);
 
     nvcpu->nv_vmswitch_in_progress = 0;
Signature · 1 line
2.26.0
#2 lore raw

PATCH v2 2/4 x86/nvmx: only update SVI when using Ack on exit

In reply to Roger Pau Monne

Check whether there's a valid interrupt in VM_EXIT_INTR_INFO in order to decide whether to update SVI in nvmx_update_apicv. If Ack on exit is not being used VM_EXIT_INTR_INFO won't have a valid interrupt and hence SVI shouldn't be updated to signal the interrupt is currently in service because it won't be Acked.

---
 xen/arch/x86/hvm/vmx/vvmx.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
xen/arch/x86/hvm/vmx/vvmx.c +1 -1
index 1b8461ba30..1753005c91 100644
@@ -1383,7 +1383,7 @@ static void nvmx_update_apicv(struct vcpu *v)
 {
     struct nestedvmx *nvmx = &vcpu_2_nvmx(v);
     unsigned long reason = get_vvmcs(v, VM_EXIT_REASON);
-    uint32_t intr_info = nvmx->intr.intr_info;
+    unsigned long intr_info = get_vvmcs(v, VM_EXIT_INTR_INFO);
 
     if ( reason == EXIT_REASON_EXTERNAL_INTERRUPT &&
          nvmx->intr.source == hvm_intsrc_lapic &&
Signature · 1 line
2.26.0
Reviewed-by on v2 2/4
#3 lore raw

Re: [Xen-devel] [PATCH v2 2/4] x86/nvmx: only update SVI when using Ack on exit

Quoted 10 lines

From: Roger Pau Monne <[email protected]>
Sent: Wednesday, March 25, 2020 6:19 PM

Check whether there's a valid interrupt in VM_EXIT_INTR_INFO in order to decide whether to update SVI in nvmx_update_apicv. If Ack on exit is not being used VM_EXIT_INTR_INFO won't have a valid interrupt and hence SVI shouldn't be updated to signal the interrupt is currently in service because it won't be Acked.

#4 lore raw

PATCH v2 3/4 x86/nvmx: split updating RVI from SVI in nvmx_update_apicv

In reply to Roger Pau Monne

Updating SVI is required when an interrupt has been injected using the Ack on exit VMEXIT feature, so that the in service interrupt in the GUEST_INTR_STATUS matches the vector that is signaled in VM_EXIT_INTR_INFO.

Updating RVI however is not tied to the Ack on exit feature, as it signals the next vector to be injected, and hence should always be updated to the next pending vector, regardless of whether Ack on exit is enabled.

---
 xen/arch/x86/hvm/vmx/vvmx.c | 22 +++++++++++++++++-----
 1 file changed, 17 insertions(+), 5 deletions(-)
xen/arch/x86/hvm/vmx/vvmx.c +17 -5
index 1753005c91..8431c912a1 100644
@@ -1384,26 +1384,38 @@ static void nvmx_update_apicv(struct vcpu *v)
     struct nestedvmx *nvmx = &vcpu_2_nvmx(v);
     unsigned long reason = get_vvmcs(v, VM_EXIT_REASON);
     unsigned long intr_info = get_vvmcs(v, VM_EXIT_INTR_INFO);
+    int rvi;
 
     if ( reason == EXIT_REASON_EXTERNAL_INTERRUPT &&
          nvmx->intr.source == hvm_intsrc_lapic &&
          (intr_info & INTR_INFO_VALID_MASK) )
     {
         uint16_t status;
-        uint32_t rvi, ppr;
-        uint32_t vector = intr_info & 0xff;
+        uint32_t ppr;
+        unsigned int vector = intr_info & INTR_INFO_VECTOR_MASK;
         struct vlapic *vlapic = vcpu_vlapic(v);
 
+        /*
+         * Update SVI to record the currently in service interrupt that's
+         * signaled in EXIT_INTR_INFO.
+         */
         vlapic_ack_pending_irq(v, vector, 1);
 
         ppr = vlapic_set_ppr(vlapic);
         WARN_ON((ppr & 0xf0) != (vector & 0xf0));
 
         status = vector << VMX_GUEST_INTR_STATUS_SVI_OFFSET;
-        rvi = vlapic_has_pending_irq(v);
-        if ( rvi != -1 )
-            status |= rvi & VMX_GUEST_INTR_STATUS_SUBFIELD_BITMASK;
+        __vmwrite(GUEST_INTR_STATUS, status);
+    }
+
+    rvi = vlapic_has_pending_irq(v);
+    if ( rvi != -1 )
+    {
+        unsigned long status;
 
+        __vmread(GUEST_INTR_STATUS, &status);
+        status &= ~VMX_GUEST_INTR_STATUS_SUBFIELD_BITMASK;
+        status |= rvi & VMX_GUEST_INTR_STATUS_SUBFIELD_BITMASK;
         __vmwrite(GUEST_INTR_STATUS, status);
     }
 }
Signature · 1 line
2.26.0
#5 lore raw

Re: [Xen-devel] [PATCH v2 3/4] x86/nvmx: split updating RVI from SVI in nvmx_update_apicv

Quoted 55 lines

From: Roger Pau Monne <[email protected]>
Sent: Wednesday, March 25, 2020 6:19 PM

Updating SVI is required when an interrupt has been injected using the Ack on exit VMEXIT feature, so that the in service interrupt in the GUEST_INTR_STATUS matches the vector that is signaled in VM_EXIT_INTR_INFO.

Updating RVI however is not tied to the Ack on exit feature, as it signals the next vector to be injected, and hence should always be updated to the next pending vector, regardless of whether Ack on exit is enabled.

---
 xen/arch/x86/hvm/vmx/vvmx.c | 22 +++++++++++++++++-----
 1 file changed, 17 insertions(+), 5 deletions(-)
xen/arch/x86/hvm/vmx/vvmx.c +10 -5
index 1753005c91..8431c912a1 100644
@@ -1384,26 +1384,38 @@ static void nvmx_update_apicv(struct vcpu *v)
     struct nestedvmx *nvmx = &vcpu_2_nvmx(v);
     unsigned long reason = get_vvmcs(v, VM_EXIT_REASON);
     unsigned long intr_info = get_vvmcs(v, VM_EXIT_INTR_INFO);
+    int rvi;
 
     if ( reason == EXIT_REASON_EXTERNAL_INTERRUPT &&
          nvmx->intr.source == hvm_intsrc_lapic &&
          (intr_info & INTR_INFO_VALID_MASK) )
     {
         uint16_t status;
-        uint32_t rvi, ppr;
-        uint32_t vector = intr_info & 0xff;
+        uint32_t ppr;
+        unsigned int vector = intr_info & INTR_INFO_VECTOR_MASK;
         struct vlapic *vlapic = vcpu_vlapic(v);
 
+        /*
+         * Update SVI to record the currently in service interrupt that's
+         * signaled in EXIT_INTR_INFO.
+         */
         vlapic_ack_pending_irq(v, vector, 1);
 
         ppr = vlapic_set_ppr(vlapic);
         WARN_ON((ppr & 0xf0) != (vector & 0xf0));
 
         status = vector << VMX_GUEST_INTR_STATUS_SVI_OFFSET;
-        rvi = vlapic_has_pending_irq(v);
-        if ( rvi != -1 )
-            status |= rvi & VMX_GUEST_INTR_STATUS_SUBFIELD_BITMASK;
+        __vmwrite(GUEST_INTR_STATUS, status);
+    }
+

xen/arch/x86/hvm/vmx/vvmx.c static void nvmx_update_apicv(struct vcpu *v) · line 1420

Quoted 10 lines
xen/arch/x86/hvm/vmx/vvmx.c +7 -0
+    rvi = vlapic_has_pending_irq(v);
+    if ( rvi != -1 )
+    {
+        unsigned long status;
 
+        __vmread(GUEST_INTR_STATUS, &status);
+        status &= ~VMX_GUEST_INTR_STATUS_SUBFIELD_BITMASK;
+        status |= rvi & VMX_GUEST_INTR_STATUS_SUBFIELD_BITMASK;
         __vmwrite(GUEST_INTR_STATUS, status);
     }

I have two minor comments. First, original code requires only one __vmwrite but now needs two writes and one read for Ack-on-exit is set. Can we optimize to sustain the original behavior? Second, although I didn't come up a case where always updating RVI (w/ pending interrupt) may bring some bad effect, it is anyway safer to put the whole logic within the check of vmexit reason and intr source.

Thanks
Kevin

#6 lore raw

Re: [Xen-devel] [PATCH v2 3/4] x86/nvmx: split updating RVI from SVI in nvmx_update_apicv

Quoted 60 lines · Kevin
Quoted 60 lines

From: Roger Pau Monne <[email protected]>
Sent: Wednesday, March 25, 2020 6:19 PM

Updating SVI is required when an interrupt has been injected using the Ack on exit VMEXIT feature, so that the in service interrupt in the GUEST_INTR_STATUS matches the vector that is signaled in VM_EXIT_INTR_INFO.

Updating RVI however is not tied to the Ack on exit feature, as it signals the next vector to be injected, and hence should always be updated to the next pending vector, regardless of whether Ack on exit is enabled.

---
 xen/arch/x86/hvm/vmx/vvmx.c | 22 +++++++++++++++++-----
 1 file changed, 17 insertions(+), 5 deletions(-)
xen/arch/x86/hvm/vmx/vvmx.c +14 -5
index 1753005c91..8431c912a1 100644
@@ -1384,26 +1384,38 @@ static void nvmx_update_apicv(struct vcpu *v)
     struct nestedvmx *nvmx = &vcpu_2_nvmx(v);
     unsigned long reason = get_vvmcs(v, VM_EXIT_REASON);
     unsigned long intr_info = get_vvmcs(v, VM_EXIT_INTR_INFO);
+    int rvi;
 
     if ( reason == EXIT_REASON_EXTERNAL_INTERRUPT &&
          nvmx->intr.source == hvm_intsrc_lapic &&
          (intr_info & INTR_INFO_VALID_MASK) )
     {
         uint16_t status;
-        uint32_t rvi, ppr;
-        uint32_t vector = intr_info & 0xff;
+        uint32_t ppr;
+        unsigned int vector = intr_info & INTR_INFO_VECTOR_MASK;
         struct vlapic *vlapic = vcpu_vlapic(v);
 
+        /*
+         * Update SVI to record the currently in service interrupt that's
+         * signaled in EXIT_INTR_INFO.
+         */
         vlapic_ack_pending_irq(v, vector, 1);
 
         ppr = vlapic_set_ppr(vlapic);
         WARN_ON((ppr & 0xf0) != (vector & 0xf0));
 
         status = vector << VMX_GUEST_INTR_STATUS_SVI_OFFSET;
-        rvi = vlapic_has_pending_irq(v);
-        if ( rvi != -1 )
-            status |= rvi & VMX_GUEST_INTR_STATUS_SUBFIELD_BITMASK;
+        __vmwrite(GUEST_INTR_STATUS, status);
+    }
+
+    rvi = vlapic_has_pending_irq(v);
+    if ( rvi != -1 )
+    {
+        unsigned long status;
 
Quoted 9 lines

xen/arch/x86/hvm/vmx/vvmx.c static void nvmx_update_apicv(struct vcpu *v) · line 1420

Quoted 5 lines
xen/arch/x86/hvm/vmx/vvmx.c +3 -0
+        __vmread(GUEST_INTR_STATUS, &status);
+        status &= ~VMX_GUEST_INTR_STATUS_SUBFIELD_BITMASK;
+        status |= rvi & VMX_GUEST_INTR_STATUS_SUBFIELD_BITMASK;
         __vmwrite(GUEST_INTR_STATUS, status);
     }

I have two minor comments. First, original code requires only one __vmwrite but now needs two writes and one read for Ack-on-exit is set. Can we optimize to sustain the original behavior?

I think I can manage to do a single write by using a global status variable initialized to 0 and only do the vmwrite if the value is != 0 by the end of the function.

Quoted 5 lines

Second,
although I didn't come up a case where always updating RVI (w/ pending interrupt) may bring some bad effect, it is anyway safer to put the whole logic within the check of vmexit reason and intr source.

I also considered this and decided there was no harm in always updating RVI if there's a pending interrupt, it could even prevent a future VMEXIT to update RVI AFAICT?

If you prefer I can return early from the function if reason != EXIT_REASON_EXTERNAL_INTERRUPT.

Thanks, Roger.

#7 lore raw

Re: [Xen-devel] [PATCH v2 3/4] x86/nvmx: split updating RVI from SVI in nvmx_update_apicv

Quoted 86 lines

From: Roger Pau Monné <[email protected]>
Sent: Thursday, March 26, 2020 5:20 PM

Quoted 13 lines · Kevin
Quoted 13 lines

From: Roger Pau Monne <[email protected]>
Sent: Wednesday, March 25, 2020 6:19 PM

Updating SVI is required when an interrupt has been injected using the Ack on exit VMEXIT feature, so that the in service interrupt in the GUEST_INTR_STATUS matches the vector that is signaled in VM_EXIT_INTR_INFO.

Updating RVI however is not tied to the Ack on exit feature, as it signals the next vector to be injected, and hence should always be updated to the next pending vector, regardless of whether Ack on exit is enabled.

Quoted 6 lines

Review

Quoted 6 lines
---
 xen/arch/x86/hvm/vmx/vvmx.c | 22 +++++++++++++++++-----
 1 file changed, 17 insertions(+), 5 deletions(-)
(quoted diff) +0 -0

b/xen/arch/x86/hvm/vmx/vvmx.c

Quoted 4 lines

xen/arch/x86/hvm/vmx/vvmx.c static void nvmx_update_apicv(struct vcpu

Quoted 4 lines

index 1753005c91..8431c912a1 100644

xen/arch/x86/hvm/vmx/vvmx.c +0 -0
@@ -1384,26 +1384,38 @@ static void nvmx_update_apicv(struct vcpu

*v)

Quoted 43 lines
Quoted 32 lines
xen/arch/x86/hvm/vmx/vvmx.c +10 -5
     struct nestedvmx *nvmx = &vcpu_2_nvmx(v);
     unsigned long reason = get_vvmcs(v, VM_EXIT_REASON);
     unsigned long intr_info = get_vvmcs(v, VM_EXIT_INTR_INFO);
+    int rvi;
 
     if ( reason == EXIT_REASON_EXTERNAL_INTERRUPT &&
          nvmx->intr.source == hvm_intsrc_lapic &&
          (intr_info & INTR_INFO_VALID_MASK) )
     {
         uint16_t status;
-        uint32_t rvi, ppr;
-        uint32_t vector = intr_info & 0xff;
+        uint32_t ppr;
+        unsigned int vector = intr_info & INTR_INFO_VECTOR_MASK;
         struct vlapic *vlapic = vcpu_vlapic(v);
 
+        /*
+         * Update SVI to record the currently in service interrupt that's
+         * signaled in EXIT_INTR_INFO.
+         */
         vlapic_ack_pending_irq(v, vector, 1);
 
         ppr = vlapic_set_ppr(vlapic);
         WARN_ON((ppr & 0xf0) != (vector & 0xf0));
 
         status = vector << VMX_GUEST_INTR_STATUS_SVI_OFFSET;
-        rvi = vlapic_has_pending_irq(v);
-        if ( rvi != -1 )
-            status |= rvi & VMX_GUEST_INTR_STATUS_SUBFIELD_BITMASK;
+        __vmwrite(GUEST_INTR_STATUS, status);
+    }
+

xen/arch/x86/hvm/vmx/vvmx.c static void nvmx_update_apicv(struct vcpu *v) · line 1420

Quoted 10 lines
xen/arch/x86/hvm/vmx/vvmx.c +7 -0
+    rvi = vlapic_has_pending_irq(v);
+    if ( rvi != -1 )
+    {
+        unsigned long status;
 
+        __vmread(GUEST_INTR_STATUS, &status);
+        status &= ~VMX_GUEST_INTR_STATUS_SUBFIELD_BITMASK;
+        status |= rvi & VMX_GUEST_INTR_STATUS_SUBFIELD_BITMASK;
         __vmwrite(GUEST_INTR_STATUS, status);
     }
Quoted 3 lines

I have two minor comments. First, original code requires only one __vmwrite but now needs two writes and one read for Ack-on-exit is set. Can we optimize to sustain the original behavior?

I think I can manage to do a single write by using a global status variable initialized to 0 and only do the vmwrite if the value is != 0 by the end of the function.

Quoted 5 lines

Second,
although I didn't come up a case where always updating RVI (w/ pending interrupt) may bring some bad effect, it is anyway safer to put the whole logic within the check of vmexit reason and intr source.

Quoted 7 lines

I also considered this and decided there was no harm in always updating RVI if there's a pending interrupt, it could even prevent a future VMEXIT to update RVI AFAICT?

If you prefer I can return early from the function if reason != EXIT_REASON_EXTERNAL_INTERRUPT.

I agree with your but given the trickiness of nested intr handling I prefer to a more conservative but safer approach. So yes, return early sounds good here.

Thanks
Kevin

#8 lore raw

PATCH v2 4/4 x86/nvmx: update exit bitmap when using virtual interrupt delivery

In reply to Roger Pau Monne

Force an update of the EOI exit bitmap in nvmx_update_apicv, because the one performed in vmx_intr_assist might not be reached if the interrupt is intercepted by nvmx_intr_intercept returning true.

Extract the code to update the exit bitmap from vmx_intr_assist into a helper and use it in nvmx_update_apicv.

---
Changes since v1:
- Reword commit message.

---
 xen/arch/x86/hvm/vmx/intr.c       | 21 +++++++++++++--------
 xen/arch/x86/hvm/vmx/vvmx.c       |  2 ++
 xen/include/asm-x86/hvm/vmx/vmx.h |  2 ++
 3 files changed, 17 insertions(+), 8 deletions(-)
xen/arch/x86/hvm/vmx/intr.c +13 -8
index 49a1295f09..000e14af49 100644
@@ -224,6 +224,18 @@ static int nvmx_intr_intercept(struct vcpu *v, struct hvm_intack intack)
     return 0;
 }
 
+void vmx_sync_exit_bitmap(struct vcpu *v)
+{
+    const unsigned int n = ARRAY_SIZE(v->arch.hvm.vmx.eoi_exit_bitmap);
+    unsigned int i;
+
+    while ( (i = find_first_bit(&v->arch.hvm.vmx.eoi_exitmap_changed, n)) < n )
+    {
+        clear_bit(i, &v->arch.hvm.vmx.eoi_exitmap_changed);
+        __vmwrite(EOI_EXIT_BITMAP(i), v->arch.hvm.vmx.eoi_exit_bitmap[i]);
+    }
+}
+
 void vmx_intr_assist(void)
 {
     struct hvm_intack intack;
@@ -318,7 +330,6 @@ void vmx_intr_assist(void)
               intack.source != hvm_intsrc_vector )
     {
         unsigned long status;
-        unsigned int i, n;
 
        /*
         * intack.vector is the highest priority vector. So we set eoi_exit_bitmap
@@ -379,13 +390,7 @@ void vmx_intr_assist(void)
                     intack.vector;
         __vmwrite(GUEST_INTR_STATUS, status);
 
-        n = ARRAY_SIZE(v->arch.hvm.vmx.eoi_exit_bitmap);
-        while ( (i = find_first_bit(&v->arch.hvm.vmx.eoi_exitmap_changed,
-                                    n)) < n )
-        {
-            clear_bit(i, &v->arch.hvm.vmx.eoi_exitmap_changed);
-            __vmwrite(EOI_EXIT_BITMAP(i), v->arch.hvm.vmx.eoi_exit_bitmap[i]);
-        }
+        vmx_sync_exit_bitmap(v);
 
         pt_intr_post(v, intack);
     }
xen/arch/x86/hvm/vmx/vvmx.c +2 -0
index 8431c912a1..845dd87f75 100644
@@ -1418,6 +1418,8 @@ static void nvmx_update_apicv(struct vcpu *v)
         status |= rvi & VMX_GUEST_INTR_STATUS_SUBFIELD_BITMASK;
         __vmwrite(GUEST_INTR_STATUS, status);
     }
+
+    vmx_sync_exit_bitmap(v);
 }
 
 static void virtual_vmexit(struct cpu_user_regs *regs)
xen/include/asm-x86/hvm/vmx/vmx.h +2 -0
index b334e1ec94..111ccd7e61 100644
@@ -610,6 +610,8 @@ void update_guest_eip(void);
 void vmx_pi_per_cpu_init(unsigned int cpu);
 void vmx_pi_desc_fixup(unsigned int cpu);
 
+void vmx_sync_exit_bitmap(struct vcpu *v);
+
 #ifdef CONFIG_HVM
 void vmx_pi_hooks_assign(struct domain *d);
 void vmx_pi_hooks_deassign(struct domain *d);
Signature · 1 line
2.26.0
#9 lore raw

Re: [Xen-devel] [PATCH v2 4/4] x86/nvmx: update exit bitmap when using virtual interrupt delivery

Quoted 73 lines

From: Roger Pau Monne <[email protected]>
Sent: Wednesday, March 25, 2020 6:19 PM

Force an update of the EOI exit bitmap in nvmx_update_apicv, because the one performed in vmx_intr_assist might not be reached if the interrupt is intercepted by nvmx_intr_intercept returning true.

Extract the code to update the exit bitmap from vmx_intr_assist into a helper and use it in nvmx_update_apicv.

---
Changes since v1:
- Reword commit message.

---
 xen/arch/x86/hvm/vmx/intr.c       | 21 +++++++++++++--------
 xen/arch/x86/hvm/vmx/vvmx.c       |  2 ++
 xen/include/asm-x86/hvm/vmx/vmx.h |  2 ++
 3 files changed, 17 insertions(+), 8 deletions(-)
xen/arch/x86/hvm/vmx/intr.c +12 -0
index 49a1295f09..000e14af49 100644
@@ -224,6 +224,18 @@ static int nvmx_intr_intercept(struct vcpu *v, struct
 hvm_intack intack)
     return 0;
 }
 
+void vmx_sync_exit_bitmap(struct vcpu *v)
+{
+    const unsigned int n = ARRAY_SIZE(v->arch.hvm.vmx.eoi_exit_bitmap);
+    unsigned int i;
+
+    while ( (i = find_first_bit(&v->arch.hvm.vmx.eoi_exitmap_changed, n)) <
 n )
+    {
+        clear_bit(i, &v->arch.hvm.vmx.eoi_exitmap_changed);
+        __vmwrite(EOI_EXIT_BITMAP(i), v->arch.hvm.vmx.eoi_exit_bitmap[i]);
+    }
+}
+
 void vmx_intr_assist(void)
 {
     struct hvm_intack intack;
xen/arch/x86/hvm/vmx/intr.c +1 -8
@@ -318,7 +330,6 @@ void vmx_intr_assist(void)
               intack.source != hvm_intsrc_vector )
     {
         unsigned long status;
-        unsigned int i, n;
 
        /*
         * intack.vector is the highest priority vector. So we set eoi_exit_bitmap
@@ -379,13 +390,7 @@ void vmx_intr_assist(void)
                     intack.vector;
         __vmwrite(GUEST_INTR_STATUS, status);
 
-        n = ARRAY_SIZE(v->arch.hvm.vmx.eoi_exit_bitmap);
-        while ( (i = find_first_bit(&v->arch.hvm.vmx.eoi_exitmap_changed,
-                                    n)) < n )
-        {
-            clear_bit(i, &v->arch.hvm.vmx.eoi_exitmap_changed);
-            __vmwrite(EOI_EXIT_BITMAP(i), v->arch.hvm.vmx.eoi_exit_bitmap[i]);
-        }
+        vmx_sync_exit_bitmap(v);
 
         pt_intr_post(v, intack);
     }
xen/arch/x86/hvm/vmx/vvmx.c +0 -0
index 8431c912a1..845dd87f75 100644

xen/arch/x86/hvm/vmx/vvmx.c static void nvmx_update_apicv(struct vcpu *v) · line 1422

Quoted 5 lines
xen/arch/x86/hvm/vmx/vvmx.c +2 -0
@@ -1418,6 +1418,8 @@ static void nvmx_update_apicv(struct vcpu *v)
         status |= rvi & VMX_GUEST_INTR_STATUS_SUBFIELD_BITMASK;
         __vmwrite(GUEST_INTR_STATUS, status);
     }
+
+    vmx_sync_exit_bitmap(v);

Similarly, I'd like to do the sync within the conditional block, when intr status is actually changed. Otherwise, it becomes checking bitmap change in every vmentry when apicv is enabled.

Quoted 19 lines
xen/arch/x86/hvm/vmx/vvmx.c +0 -0
 }
 
 static void virtual_vmexit(struct cpu_user_regs *regs)
xen/include/asm- renamed from xen/include/asm-x86/hvm/vmx/vmx.h +0 -0

x86/hvm/vmx/vmx.h
index b334e1ec94..111ccd7e61 100644

xen/include/asm-x86/hvm/vmx/vmx.h +2 -1
@@ -610,6 +610,8 @@ void update_guest_eip(void);
 void vmx_pi_per_cpu_init(unsigned int cpu);
 void vmx_pi_desc_fixup(unsigned int cpu);
 
+void vmx_sync_exit_bitmap(struct vcpu *v);
+
 #ifdef CONFIG_HVM
 void vmx_pi_hooks_assign(struct domain *d);
 void vmx_pi_hooks_deassign(struct domain *d);
--

2.26.0

#10 lore raw

Re: [Xen-devel] [PATCH v2 4/4] x86/nvmx: update exit bitmap when using virtual interrupt delivery

Quoted 66 lines · Kevin
Quoted 66 lines

From: Roger Pau Monne <[email protected]>
Sent: Wednesday, March 25, 2020 6:19 PM

Force an update of the EOI exit bitmap in nvmx_update_apicv, because the one performed in vmx_intr_assist might not be reached if the interrupt is intercepted by nvmx_intr_intercept returning true.

Extract the code to update the exit bitmap from vmx_intr_assist into a helper and use it in nvmx_update_apicv.

---
Changes since v1:
- Reword commit message.

---
 xen/arch/x86/hvm/vmx/intr.c       | 21 +++++++++++++--------
 xen/arch/x86/hvm/vmx/vvmx.c       |  2 ++
 xen/include/asm-x86/hvm/vmx/vmx.h |  2 ++
 3 files changed, 17 insertions(+), 8 deletions(-)
xen/arch/x86/hvm/vmx/intr.c +12 -0
index 49a1295f09..000e14af49 100644
@@ -224,6 +224,18 @@ static int nvmx_intr_intercept(struct vcpu *v, struct
 hvm_intack intack)
     return 0;
 }
 
+void vmx_sync_exit_bitmap(struct vcpu *v)
+{
+    const unsigned int n = ARRAY_SIZE(v->arch.hvm.vmx.eoi_exit_bitmap);
+    unsigned int i;
+
+    while ( (i = find_first_bit(&v->arch.hvm.vmx.eoi_exitmap_changed, n)) <
 n )
+    {
+        clear_bit(i, &v->arch.hvm.vmx.eoi_exitmap_changed);
+        __vmwrite(EOI_EXIT_BITMAP(i), v->arch.hvm.vmx.eoi_exit_bitmap[i]);
+    }
+}
+
 void vmx_intr_assist(void)
 {
     struct hvm_intack intack;
xen/arch/x86/hvm/vmx/intr.c +1 -8
@@ -318,7 +330,6 @@ void vmx_intr_assist(void)
               intack.source != hvm_intsrc_vector )
     {
         unsigned long status;
-        unsigned int i, n;
 
        /*
         * intack.vector is the highest priority vector. So we set eoi_exit_bitmap
@@ -379,13 +390,7 @@ void vmx_intr_assist(void)
                     intack.vector;
         __vmwrite(GUEST_INTR_STATUS, status);
 
-        n = ARRAY_SIZE(v->arch.hvm.vmx.eoi_exit_bitmap);
-        while ( (i = find_first_bit(&v->arch.hvm.vmx.eoi_exitmap_changed,
-                                    n)) < n )
-        {
-            clear_bit(i, &v->arch.hvm.vmx.eoi_exitmap_changed);
-            __vmwrite(EOI_EXIT_BITMAP(i), v->arch.hvm.vmx.eoi_exit_bitmap[i]);
-        }
+        vmx_sync_exit_bitmap(v);
 
Quoted 16 lines

xen/arch/x86/hvm/vmx/vvmx.c static void nvmx_update_apicv(struct vcpu *v) · line 1422

Quoted 12 lines
xen/arch/x86/hvm/vmx/intr.c +0 -0
         pt_intr_post(v, intack);
     }
xen/arch/x86/hvm/vmx/vvmx.c +2 -0
index 8431c912a1..845dd87f75 100644
@@ -1418,6 +1418,8 @@ static void nvmx_update_apicv(struct vcpu *v)
         status |= rvi & VMX_GUEST_INTR_STATUS_SUBFIELD_BITMASK;
         __vmwrite(GUEST_INTR_STATUS, status);
     }
+
+    vmx_sync_exit_bitmap(v);

Similarly, I'd like to do the sync within the conditional block, when intr status is actually changed. Otherwise, it becomes checking bitmap change in every vmentry when apicv is enabled.

No - it will only check the bitmap when there's a virtual vmexit (which is where nvmx_update_apicv gets called), not on every vmentry. I can try to do this conditionally on whether GUEST_INTR_STATUS is actually changed.

Thanks, Roger.

#11 lore raw

Re: [Xen-devel] [PATCH v2 4/4] x86/nvmx: update exit bitmap when using virtual interrupt delivery

Quoted 93 lines

From: Roger Pau Monné <[email protected]>
Sent: Thursday, March 26, 2020 5:22 PM

Quoted 20 lines · Kevin
Quoted 20 lines

From: Roger Pau Monne <[email protected]>
Sent: Wednesday, March 25, 2020 6:19 PM

Force an update of the EOI exit bitmap in nvmx_update_apicv, because the one performed in vmx_intr_assist might not be reached if the interrupt is intercepted by nvmx_intr_intercept returning true.

Extract the code to update the exit bitmap from vmx_intr_assist into a helper and use it in nvmx_update_apicv.

---
Changes since v1:
- Reword commit message.

---
 xen/arch/x86/hvm/vmx/intr.c       | 21 +++++++++++++--------
 xen/arch/x86/hvm/vmx/vvmx.c       |  2 ++
 xen/include/asm-x86/hvm/vmx/vmx.h |  2 ++
 3 files changed, 17 insertions(+), 8 deletions(-)
Quoted 5 lines

xen/arch/x86/hvm/vmx/intr.c static int nvmx_intr_intercept(struct vcpu *v,

Quoted 5 lines
xen/arch/x86/hvm/vmx/intr.c +0 -0
index 49a1295f09..000e14af49 100644
@@ -224,6 +224,18 @@ static int nvmx_intr_intercept(struct vcpu *v,

struct

Quoted 10 lines

xen/arch/x86/hvm/vmx/intr.c static int nvmx_intr_intercept(struct vcpu *v, struct hvm_intack intack)

Quoted 10 lines
xen/arch/x86/hvm/vmx/intr.c +6 -0
 hvm_intack intack)
     return 0;
 }
 
+void vmx_sync_exit_bitmap(struct vcpu *v)
+{
+    const unsigned int n = ARRAY_SIZE(v->arch.hvm.vmx.eoi_exit_bitmap);
+    unsigned int i;
+
+    while ( (i = find_first_bit(&v->arch.hvm.vmx.eoi_exitmap_changed, n))

<

Quoted 17 lines

xen/arch/x86/hvm/vmx/intr.c static int nvmx_intr_intercept(struct vcpu *v, struct hvm_intack intack)

Quoted 4 lines
xen/arch/x86/hvm/vmx/intr.c +3 -0
 n )
+    {
+        clear_bit(i, &v->arch.hvm.vmx.eoi_exitmap_changed);
+        __vmwrite(EOI_EXIT_BITMAP(i), v-

arch.hvm.vmx.eoi_exit_bitmap[i]);

Quoted 12 lines
xen/arch/x86/hvm/vmx/intr.c +3 -0
+    }
+}
+
 void vmx_intr_assist(void)
 {
     struct hvm_intack intack;
xen/arch/x86/hvm/vmx/intr.c +0 -1
@@ -318,7 +330,6 @@ void vmx_intr_assist(void)
               intack.source != hvm_intsrc_vector )
     {
         unsigned long status;
-        unsigned int i, n;
 
Quoted 2 lines

xen/arch/x86/hvm/vmx/intr.c void vmx_intr_assist(void) · line 335

Quoted 2 lines
xen/arch/x86/hvm/vmx/intr.c +0 -0
        /*
         * intack.vector is the highest priority vector. So we set

eoi_exit_bitmap

Quoted 11 lines

xen/arch/x86/hvm/vmx/intr.c void vmx_intr_assist(void) · line 387

Quoted 10 lines
xen/arch/x86/hvm/vmx/intr.c +0 -6
@@ -379,13 +390,7 @@ void vmx_intr_assist(void)
                     intack.vector;
         __vmwrite(GUEST_INTR_STATUS, status);
 
-        n = ARRAY_SIZE(v->arch.hvm.vmx.eoi_exit_bitmap);
-        while ( (i = find_first_bit(&v->arch.hvm.vmx.eoi_exitmap_changed,
-                                    n)) < n )
-        {
-            clear_bit(i, &v->arch.hvm.vmx.eoi_exitmap_changed);
-            __vmwrite(EOI_EXIT_BITMAP(i), v-

arch.hvm.vmx.eoi_exit_bitmap[i]);

Quoted 6 lines

Review

Quoted 6 lines
xen/arch/x86/hvm/vmx/intr.c +1 -1
-        }
+        vmx_sync_exit_bitmap(v);
 
         pt_intr_post(v, intack);
     }
(quoted diff) +0 -0

b/xen/arch/x86/hvm/vmx/vvmx.c

Quoted 5 lines
Quoted 5 lines

index 8431c912a1..845dd87f75 100644

xen/arch/x86/hvm/vmx/vvmx.c +0 -0
@@ -1418,6 +1418,8 @@ static void nvmx_update_apicv(struct vcpu *v)
         status |= rvi & VMX_GUEST_INTR_STATUS_SUBFIELD_BITMASK;
Quoted 8 lines

xen/arch/x86/hvm/vmx/vvmx.c static void nvmx_update_apicv(struct vcpu *v) · line 1422

Quoted 4 lines
xen/arch/x86/hvm/vmx/vvmx.c +2 -0
         __vmwrite(GUEST_INTR_STATUS, status);
     }
+
+    vmx_sync_exit_bitmap(v);

Similarly, I'd like to do the sync within the conditional block, when intr status is actually changed. Otherwise, it becomes checking bitmap change in every vmentry when apicv is enabled.

Quoted 2 lines

No - it will only check the bitmap when there's a virtual vmexit (which is where nvmx_update_apicv gets called), not on every vmentry.

you are right. I overlooked it. but still it's unnecessary to do it for every virtual vmexit. 😊

Quoted 3 lines

I can try to do this conditionally on whether GUEST_INTR_STATUS is actually changed.

yes.

Thanks
Kevin