X86: Handle Hyper-V vmbus interrupts as special hypervisor interrupts
Subsystems (3)
PATCH RESEND 1/1 X86: Handle Hyper-V vmbus interrupts as special hypervisor interrupts
Starting with win8, vmbus interrupts can be delivered on any VCPU in the guest and furthermore can be concurrently active on multiple VCPUs. Support this interrupt delivery model by setting up a separate IDT entry for Hyper-V vmbus. interrupts. I would like to thank Jan Beulich <[email protected]> and Thomas Gleixner <[email protected]>, for their help.
In this version of the patch, based on the feedback, I have merged the IDT vector for Xen and Hyper-V and made the necessary adjustments.
--- arch/x86/include/asm/irq_vectors.h | 4 +- arch/x86/include/asm/mshyperv.h | 4 +++ arch/x86/kernel/cpu/mshyperv.c | 39 ++++++++++++++++++++++++++++++++++++ arch/x86/kernel/entry_32.S | 9 +++++++- arch/x86/kernel/entry_64.S | 7 +++++- drivers/xen/events.c | 6 ++-- 6 files changed, 62 insertions(+), 7 deletions(-)
arch/x86/include/asm/irq_vectors.h +2 -2
@@ -109,8 +109,8 @@#define UV_BAU_MESSAGE 0xf5 -/* Xen vector callback to receive events in a HVM domain */ -#define XEN_HVM_EVTCHN_CALLBACK 0xf3 +/* Vector on which hypervisor callbacks will be delivered */ +#define HYPERVISOR_CALLBACK_VECTOR 0xf3/* * Local APIC timer IRQ vector is on a different priority level,
arch/x86/include/asm/mshyperv.h +4 -0
@@ -11,4 +11,8 @@ struct ms_hyperv_info {extern struct ms_hyperv_info ms_hyperv; +void hyperv_callback_vector (void ); +void hyperv_vector_handler (struct pt_regs *regs); +void hv_register_vmbus_handler (int irq,irq_handler_t handler); +#endif
arch/x86/kernel/cpu/mshyperv.c +39 -0
@@ -14,10 +14,15 @@#include <linux/time.h> #include <linux/clocksource.h> #include <linux/module.h> +#include <linux/hardirq.h> +#include <linux/interrupt.h> #include <asm/processor.h> #include <asm/hypervisor.h> #include <asm/hyperv.h> #include <asm/mshyperv.h> +#include <asm/desc.h> +#include <asm/idle.h> +#include <asm/irq_regs.h> struct ms_hyperv_info ms_hyperv;EXPORT_SYMBOL_GPL (ms_hyperv);
@@ -69,6 +74,11 @@ static void __init ms_hyperv_init_platform(void) ms_hyperv.features, ms_hyperv.hints);clocksource_register_hz (&hyperv_cs,NSEC_PER_SEC /100 ); + +/* +* Setup the IDT for hypervisor callback. +*/ +alloc_intr_gate (HYPERVISOR_CALLBACK_VECTOR , hyperv_callback_vector); }const __refconst struct hypervisor_x86 x86_hyper_ms_hyperv = {
@@ -77,3 +87,32 @@ const __refconst struct hypervisor_x86 x86_hyper_ms_hyperv = { .init_platform = ms_hyperv_init_platform, };EXPORT_SYMBOL (x86_hyper_ms_hyperv); + +static int vmbus_irq =-1 ; +static irq_handler_t vmbus_isr; + +void hv_register_vmbus_handler (int irq,irq_handler_t handler) +{ + vmbus_irq = irq; + vmbus_isr = handler; +} +EXPORT_SYMBOL_GPL (hv_register_vmbus_handler); + +void hyperv_vector_handler (struct pt_regs *regs) +{ +struct pt_regs *old_regs =set_irq_regs (regs); +struct irq_desc *desc; + +irq_enter (); +#ifdef CONFIG_X86 +exit_idle (); +#endif + + desc =irq_to_desc (vmbus_irq); + +if (desc) +generic_handle_irq_desc (vmbus_irq, desc); + +irq_exit (); +set_irq_regs (old_regs); +}
arch/x86/kernel/entry_32.S +8 -1
@@ -1047,11 +1047,18 @@ ENTRY(xen_failsafe_callback) _ASM_EXTABLE(4b,9b) ENDPROC(xen_failsafe_callback) -BUILD_INTERRUPT3(xen_hvm_callback_vector, XEN_HVM_EVTCHN_CALLBACK, +BUILD_INTERRUPT3(xen_hvm_callback_vector, HYPERVISOR_CALLBACK_VECTOR, xen_evtchn_do_upcall) #endif /* CONFIG_XEN */ +#if IS_ENABLED(CONFIG_HYPERV) + +BUILD_INTERRUPT3(hyperv_callback_vector, HYPERVISOR_CALLBACK_VECTOR, + hyperv_vector_handler) + +#endif /* CONFIG_HYPERV */ + #ifdef CONFIG_FUNCTION_TRACER #ifdef CONFIG_DYNAMIC_FTRACE
arch/x86/kernel/entry_64.S +6 -1
@@ -1446,11 +1446,16 @@ ENTRY(xen_failsafe_callback) CFI_ENDPROC END(xen_failsafe_callback) -apicinterrupt XEN_HVM_EVTCHN_CALLBACK \ +apicinterrupt HYPERVISOR_CALLBACK_VECTOR \ xen_hvm_callback_vector xen_evtchn_do_upcall #endif /* CONFIG_XEN */ +#if IS_ENABLED(CONFIG_HYPERV) +apicinterrupt HYPERVISOR_CALLBACK_VECTOR \ + hyperv_callback_vector hyperv_vector_handler +#endif /* CONFIG_HYPERV */ + /* * Some functions should be protected against kprobes */
drivers/xen/events.c +3 -3
@@ -1787,7 +1787,7 @@ void xen_callback_vector(void)int rc;uint64_t callback_via;if (xen_have_vector_callback) { - callback_via =HVM_CALLBACK_VECTOR (XEN_HVM_EVTCHN_CALLBACK ); + callback_via =HVM_CALLBACK_VECTOR (HYPERVISOR_CALLBACK_VECTOR ); rc =xen_set_callback_via (callback_via);if (rc) { printk(KERN_ERR "Request for Xen HVM callback vector"
@@ -1798,8 +1798,8 @@ void xen_callback_vector(void)printk (KERN_INFO "Xen HVM callback vector for event delivery is " "enabled\n" );/* in the restore case the vector has already been allocated */ -if (!test_bit (XEN_HVM_EVTCHN_CALLBACK , used_vectors)) - alloc_intr_gate(XEN_HVM_EVTCHN_CALLBACK, xen_hvm_callback_vector); +if (!test_bit (HYPERVISOR_CALLBACK_VECTOR , used_vectors)) + alloc_intr_gate(HYPERVISOR_CALLBACK_VECTOR, xen_hvm_callback_vector); } }#else
Signature · 1 line
1.7.4.1
arch/x86/kernel/cpu/mshyperv.c static void __init ms_hyperv_init_platform(void) · line 81
Quoted 10 lines
Quoted 1 line
Quoted 1 line · K. Y. Srinivasan
arch/x86/kernel/cpu/mshyperv.c +5
-0@@ -69,6 +74,11 @@ static void __init ms_hyperv_init_platform(void) ms_hyperv.features, ms_hyperv.hints);clocksource_register_hz (&hyperv_cs,NSEC_PER_SEC /100 ); + +/* +* Setup the IDT for hypervisor callback. +*/ +alloc_intr_gate (HYPERVISOR_CALLBACK_VECTOR , hyperv_callback_vector);
Isn't doing this unconditionally here as problematic as the call to clocksource_register_hz() turned out to be when Xen's Hyper-V shim reacts to the CPUID inquiry above?
Quoted 14 lines
arch/x86/kernel/cpu/mshyperv.c +10
-0@@ -77,3 +87,32 @@ const __refconst struct hypervisor_x86 x86_hyper_ms_hyperv = { .init_platform = ms_hyperv_init_platform, };EXPORT_SYMBOL (x86_hyper_ms_hyperv); + +static int vmbus_irq =-1 ; +static irq_handler_t vmbus_isr; + +void hv_register_vmbus_handler (int irq,irq_handler_t handler) +{ + vmbus_irq = irq; + vmbus_isr = handler; +} +EXPORT_SYMBOL_GPL (hv_register_vmbus_handler);
arch/x86/kernel/cpu/mshyperv.c const __refconst struct hypervisor_x86 x86_hyper_ms_hyperv = { · line 109
Quoted 10 lines
arch/x86/kernel/cpu/mshyperv.c +10
-0+ +void hyperv_vector_handler (struct pt_regs *regs) +{ +struct pt_regs *old_regs =set_irq_regs (regs); +struct irq_desc *desc; + +irq_enter (); +#ifdef CONFIG_X86 +exit_idle (); +#endif
This being in a file underneath arch/x86 - why the conditional?
Jan
Quoted 24 lines
-----Original Message-----
From: Jan Beulich [mailto:[email protected]]
Sent: Thursday, January 24, 2013 3:48 AM
To: KY Srinivasan
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]
Subject: Re: [PATCH RESEND 1/1] X86: Handle Hyper-V vmbus interrupts as special hypervisor interruptsarch/x86/kernel/cpu/mshyperv.c static void __init ms_hyperv_init_platform(void) · line 81
Quoted 10 lines
Quoted 1 line
Quoted 1 line · K. Y. Srinivasan
arch/x86/kernel/cpu/mshyperv.c +5
-0@@ -69,6 +74,11 @@ static void __init ms_hyperv_init_platform(void) ms_hyperv.features, ms_hyperv.hints);clocksource_register_hz (&hyperv_cs,NSEC_PER_SEC /100 ); + +/* +* Setup the IDT for hypervisor callback. +*/ + alloc_intr_gate(HYPERVISOR_CALLBACK_VECTOR,
hyperv_callback_vector);
Quoted 3 lines
Isn't doing this unconditionally here as problematic as the call to
clocksource_register_hz()turned out to be when Xen's Hyper-V shim reacts to the CPUID inquiry above?
I was not sure what to make this conditional on at run-time. To the extent that Xen emulation of Hyper-V is complete, this will be a problem. Does Xen return all the "feature bits" of Hyper-V. I could discriminate on a feature that Xen does not plan to emulate.
Quoted 20 lines
Review const __refconst struct hypervisor_x86
Quoted 2 lines
(quoted diff) +0
-0@@ -77,3 +87,32 @@ const __refconst struct hypervisor_x86
x86_hyper_ms_hyperv= {Quoted 17 lines
arch/x86/kernel/cpu/mshyperv.c +14
-0.init_platform = ms_hyperv_init_platform, };EXPORT_SYMBOL (x86_hyper_ms_hyperv); + +static int vmbus_irq =-1 ; +static irq_handler_t vmbus_isr; + +void hv_register_vmbus_handler (int irq,irq_handler_t handler) +{ + vmbus_irq = irq; + vmbus_isr = handler; +} +EXPORT_SYMBOL_GPL (hv_register_vmbus_handler); + +void hyperv_vector_handler (struct pt_regs *regs) +{ +struct pt_regs *old_regs =set_irq_regs (regs);
Quoted 8 lines
arch/x86/kernel/cpu/mshyperv.c const __refconst struct hypervisor_x86 x86_hyper_ms_hyperv = { · line 109
Quoted 6 lines
arch/x86/kernel/cpu/mshyperv.c +6
-0+struct irq_desc *desc; + +irq_enter (); +#ifdef CONFIG_X86 +exit_idle (); +#endif This being in a file underneath
arch/x86- why the conditional?
Good point, I will fix that.
Quoted 4 lines
Jan
Quoted 30 lines
Quoted 9 lines
Quoted 1 line · KY Srinivasan
-----Original Message-----
From: Jan Beulich [mailto:[email protected]]
Sent: Thursday, January 24, 2013 3:48 AM
To: KY Srinivasan
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected];Quoted 16 lines
[email protected]; [email protected]; [email protected]
Subject: Re: [PATCH RESEND 1/1] X86: Handle Hyper-V vmbus interrupts as special hypervisor interruptsarch/x86/kernel/cpu/mshyperv.c static void __init ms_hyperv_init_platform(void) · line 81
Quoted 10 lines
Quoted 1 line
Quoted 1 line · K. Y. Srinivasan
arch/x86/kernel/cpu/mshyperv.c +5
-0@@ -69,6 +74,11 @@ static void __init ms_hyperv_init_platform(void) ms_hyperv.features, ms_hyperv.hints);clocksource_register_hz (&hyperv_cs,NSEC_PER_SEC /100 ); + +/* +* Setup the IDT for hypervisor callback. +*/ + alloc_intr_gate(HYPERVISOR_CALLBACK_VECTOR,
hyperv_callback_vector);Quoted 3 lines
Isn't doing this unconditionally here as problematic as the call to
clocksource_register_hz()turned out to be when Xen's Hyper-V shim reacts to the CPUID inquiry above?
Quoted 5 lines
I was not sure what to make this conditional on at run-time. To the extent that Xen emulation of Hyper-V is complete, this will be a problem. Does Xen return all the "feature bits" of Hyper-V. I could discriminate on a feature that
Xen does not plan to emulate.
I've no idea what plans there might be, but that's the code there is currently:
int cpuid_viridian_leaves(unsigned int leaf, unsigned int *eax,
unsigned int *ebx, unsigned int *ecx, unsigned int *edx)
{
struct domain *d = current->domain;
if ( !is_viridian_domain(d) )
return 0;
leaf -= 0x40000000;
if ( leaf > 6 )
return 0;
*eax = *ebx = *ecx = *edx = 0;
switch ( leaf )
{
case 0:
*eax = 0x40000006; /* Maximum leaf */
*ebx = 0x7263694d; /* Magic numbers */
*ecx = 0x666F736F;
*edx = 0x76482074;
break;
case 1:
*eax = 0x31237648; /* Version number */
break;
case 2:
/* Hypervisor information, but only if the guest has set its
own version number. */
if ( d->arch.hvm_domain.viridian.guest_os_id.raw == 0 )
break;
*eax = 1; /* Build number */
*ebx = (xen_major_version() << 16) | xen_minor_version();
*ecx = 0; /* SP */
*edx = 0; /* Service branch and number */
break;
case 3:
/* Which hypervisor MSRs are available to the guest */
*eax = (CPUID3A_MSR_APIC_ACCESS |
CPUID3A_MSR_HYPERCALL |
CPUID3A_MSR_VP_INDEX);
break;
case 4:
/* Recommended hypercall usage. */
if ( (d->arch.hvm_domain.viridian.guest_os_id.raw == 0) ||
(d->arch.hvm_domain.viridian.guest_os_id.fields.os < 4) )
break;
*eax = (CPUID4A_MSR_BASED_APIC |
CPUID4A_RELAX_TIMER_INT);
*ebx = 2047; /* long spin count */
break;
}
return 1;
}Question is - considering you stated that this is supported starting in Win8, doesn't Hyper-V itself announce that capability in some explicit way?
Jan
Quoted 106 lines
-----Original Message-----
From: Jan Beulich [mailto:[email protected]]
Sent: Thursday, January 24, 2013 11:16 AM
To: KY Srinivasan
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]
Subject: RE: [PATCH RESEND 1/1] X86: Handle Hyper-V vmbus interrupts as special hypervisor interruptsQuoted 30 lines
Quoted 9 lines
Quoted 1 line · KY Srinivasan
-----Original Message-----
From: Jan Beulich [mailto:[email protected]]
Sent: Thursday, January 24, 2013 3:48 AM
To: KY Srinivasan
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected];Quoted 16 lines
[email protected]; [email protected]; [email protected]
Subject: Re: [PATCH RESEND 1/1] X86: Handle Hyper-V vmbus interrupts as special hypervisor interruptsarch/x86/kernel/cpu/mshyperv.c static void __init ms_hyperv_init_platform(void) · line 81
Quoted 10 lines
Quoted 1 line
Quoted 1 line · K. Y. Srinivasan
arch/x86/kernel/cpu/mshyperv.c +5
-0@@ -69,6 +74,11 @@ static void __init ms_hyperv_init_platform(void) ms_hyperv.features, ms_hyperv.hints);clocksource_register_hz (&hyperv_cs,NSEC_PER_SEC /100 ); + +/* +* Setup the IDT for hypervisor callback. +*/ + alloc_intr_gate(HYPERVISOR_CALLBACK_VECTOR,
hyperv_callback_vector);Quoted 3 lines
Isn't doing this unconditionally here as problematic as the call to
clocksource_register_hz()turned out to be when Xen's Hyper-V shim reacts to the CPUID inquiry above?Quoted 5 lines
I was not sure what to make this conditional on at run-time. To the extent that Xen emulation of Hyper-V is complete, this will be a problem. Does Xen return all the "feature bits" of Hyper-V. I could discriminate on a feature that
Xen does not plan to emulate.I've no idea what plans there might be, but that's the code there is currently:
int cpuid_viridian_leaves(unsigned int leaf, unsigned int *eax,
unsigned int *ebx, unsigned int *ecx, unsigned int *edx){ struct domain *d = current->domain; if ( !is_viridian_domain(d) ) return 0; leaf -= 0x40000000; if ( leaf > 6 ) return 0; *eax = *ebx = *ecx = *edx = 0; switch ( leaf ) { case 0: *eax = 0x40000006; /* Maximum leaf */ *ebx = 0x7263694d; /* Magic numbers */ *ecx = 0x666F736F; *edx = 0x76482074; break; case 1: *eax = 0x31237648; /* Version number */ break; case 2: /* Hypervisor information, but only if the guest has set its own version number. */ if ( d->arch.hvm_domain.viridian.guest_os_id.raw == 0 ) break; *eax = 1; /* Build number */ *ebx = (xen_major_version() << 16) | xen_minor_version(); *ecx = 0; /* SP */ *edx = 0; /* Service branch and number */ break; case 3: /* Which hypervisor MSRs are available to the guest */ *eax = (CPUID3A_MSR_APIC_ACCESS | CPUID3A_MSR_HYPERCALL | CPUID3A_MSR_VP_INDEX); break; case 4: /* Recommended hypercall usage. */ if ( (d->arch.hvm_domain.viridian.guest_os_id.raw == 0) || (d->arch.hvm_domain.viridian.guest_os_id.fields.os < 4) ) break; *eax = (CPUID4A_MSR_BASED_APIC | CPUID4A_RELAX_TIMER_INT); *ebx = 2047; /* long spin count */ break; } return 1; }
Quoted 3 lines
Question is - considering you stated that this is supported starting in Win8, doesn't Hyper-V itself announce that capability in some explicit way?
Thanks Jan. Unfortunately I don't think tis interrupt delivery model is specified via a "feature" bit (I will check with the Windows guys). Perhaps, we can have a Hyper-V specific compilation switch to address the Xen emulation issue.
Regards,
K. Y
Quoted 4 lines
Jan
Quoted 8 lines · KY Srinivasan
Quoted 3 lines
Question is - considering you stated that this is supported starting in Win8, doesn't Hyper-V itself announce that capability in some explicit way?
Thanks Jan. Unfortunately I don't think tis interrupt delivery model is specified via a "feature" bit (I will check with the Windows guys). Perhaps, we can have a Hyper-V specific compilation switch to address the Xen emulation issue.
Would that really help if both pvops XEN_PVHVM and HYPERV are enabled in the config? I assume at this point the access to the DMI data is not yet possible.
Olaf
Quoted 19 lines
-----Original Message-----
From: Olaf Hering [mailto:[email protected]]
Sent: Thursday, January 24, 2013 2:00 PM
To: KY Srinivasan
Cc: Jan Beulich; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]
Subject: Re: [PATCH RESEND 1/1] X86: Handle Hyper-V vmbus interrupts as special hypervisor interruptsQuoted 4 lines · KY Srinivasan
Quoted 3 lines
Question is - considering you stated that this is supported starting in Win8, doesn't Hyper-V itself announce that capability in some explicit way?
Quoted 8 lines
Quoted 4 lines
Thanks Jan. Unfortunately I don't think tis interrupt delivery model is specified via a "feature" bit (I will check with the Windows guys). Perhaps, we can have a Hyper-V specific compilation switch to address the Xen emulation issue.
Would that really help if both pvops
XEN_PVHVMand HYPERV are enabled in the config? I assume at this point the access to the DMI data is not yet possible.
I was under the impression that only Dom0 Xen would emulate Hyper-V.
K. Y
Quoted 3 lines
Olaf
���{.n�+�������+%�����ݶ��w��{.n�+����{��G�����{ay� ʇڙ�,j��f���h��������z_�(�階�ݢj"���m������G������?����&���~��iO��z��v�^�m���� ������?�I�
Quoted 2 lines
Quoted 1 line
Quoted 1 line · KY Srinivasan
I was under the impression that only Dom0 Xen would emulate Hyper-V.
How that? It's the hypervisor that does the emulation, and obviously not for Dom0, but for (HVM) DomU-s.
Jan
H. Peter Anvin
reviewerCan't you discover Xen emulation by looking for Xen first? Or does Xen go "stealth"?
Quoted 20 lines · Olaf Hering
Quoted 6 lines · KY Srinivasan
Quoted 3 lines
Question is - considering you stated that this is supported starting in Win8, doesn't Hyper-V itself announce that capability in some explicit way?
Thanks Jan. Unfortunately I don't think tis interrupt delivery model is specified via a "feature" bit (I will check with the Windows
guys).
Quoted 2 lines
Perhaps, we can have a Hyper-V specific compilation switch to address the Xen emulation issue.
Would that really help if both pvops
XEN_PVHVMand HYPERV are enabled in
the config? I assume at this point the access to the DMI data is not yet
possible.Olaf
Signature · 1 line
Sent from my mobile phone. Please excuse brevity and lack of formatting.
Quoted 1 line · H. Peter Anvin
Can't you discover Xen emulation by looking for Xen first? Or does Xen go "stealth"?
The current code looks at cpuid to decide what the host type is. One thing to detect a Xen host is to look for the DIM strings, which still have "Xen HVM DomU" or similar.
Olaf
H. Peter Anvin
reviewerAnd there is no evidence of Xen in CPUID?
Quoted 10 lines · Olaf Hering
Quoted 1 line · H. Peter Anvin
Can't you discover Xen emulation by looking for Xen first? Or does
Xen go "stealth"?
The current code looks at cpuid to decide what the host type is. One thing to detect a Xen host is to look for the DIM strings, which still have "Xen HVM DomU" or similar.
Olaf
Signature · 1 line
Sent from my mobile phone. Please excuse brevity and lack of formatting.
Quoted 3 lines
Quoted 1 line
Quoted 1 line · H. Peter Anvin
Can't you discover Xen emulation by looking for Xen first? Or does Xen go "stealth"?
That's already being done - the ordering currently is Xen, VMWare, Hyper-V, KVM. I.e. there's only a problem when !CONFIG_XEN_PVHVM.
Jan
Quoted 8 lines
Quoted 1 line
Quoted 1 line · Olaf Hering
Quoted 4 lines · KY Srinivasan
Quoted 3 lines
Question is - considering you stated that this is supported starting in Win8, doesn't Hyper-V itself announce that capability in some explicit way?
Quoted 8 lines
Quoted 4 lines
Thanks Jan. Unfortunately I don't think tis interrupt delivery model is specified via a "feature" bit (I will check with the Windows guys). Perhaps, we can have a Hyper-V specific compilation switch to address the Xen emulation issue.
Would that really help if both pvops
XEN_PVHVMand HYPERV are enabled in the config? I assume at this point the access to the DMI data is not yet possible.
As just said in another reply - with XEN_PVHVM, there's no problem, since Xen gets checked for first. There's only a problem when !XEN_PVHVM, because in that case Hyper-V will be probed for. And the problem can only get bigger when on top of that the out-of-tree PV drivers are intended to be used.
Jan
Quoted 17 lines · Jan Beulich
Quoted 8 lines
Quoted 1 line
Quoted 1 line · Olaf Hering
Quoted 4 lines · KY Srinivasan
Quoted 3 lines
Question is - considering you stated that this is supported starting in Win8, doesn't Hyper-V itself announce that capability in some explicit way?
Quoted 8 lines
Quoted 4 lines
Thanks Jan. Unfortunately I don't think tis interrupt delivery model is specified via a "feature" bit (I will check with the Windows guys). Perhaps, we can have a Hyper-V specific compilation switch to address the Xen emulation issue.
Would that really help if both pvops
XEN_PVHVMand HYPERV are enabled in the config? I assume at this point the access to the DMI data is not yet possible.
Quoted 5 lines
As just said in another reply - with
XEN_PVHVM, there's no problem, since Xen gets checked for first. There's only a problem when !XEN_PVHVM, because in that case Hyper-V will be probed for. And the problem can only get bigger when on top of that the out-of-tree PV drivers are intended to be used.
Yes, xen_cpuid_base() recognizes Xen as hypervisor even with 'viridian=1' in the .cfg file.
I think if there is no feature bit for this, the DMI can be used
because its appearently available before the host is checked:
...
[ 0.000000] NX (Execute Disable) protection: active [ 0.000000] DMI 2.4 present. [ 0.000000] DMI: Xen HVM domU, BIOS 4.1.3_06-0.7.1 12/05/2012 [ 0.000000] Hypervisor detected: Microsoft HyperV [ 0.000000] HyperV: features 0x70, hints 0x0 [ 0.000000] e820 update range: 0000000000000000 - 0000000000010000 (usable) ==> (reserved)
...
Olaf
H. Peter Anvin
reviewerI also have to point out the obvious problem with Xen masquerading as HyperV without implementing the same feature set...
Quoted 134 lines · KY Srinivasan
Quoted 8 lines
-----Original Message-----
From: Jan Beulich [mailto:[email protected]]
Sent: Thursday, January 24, 2013 11:16 AM
To: KY Srinivasan
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected];Quoted 2 lines
[email protected]; [email protected]; [email protected]
Subject: RE: [PATCH RESEND 1/1] X86: Handle Hyper-V vmbus interruptsas
Quoted 10 lines
special hypervisor interrupts
Quoted 8 lines
Quoted 8 lines
Quoted 1 line · KY Srinivasan
-----Original Message-----
From: Jan Beulich [mailto:[email protected]]
Sent: Thursday, January 24, 2013 3:48 AM
To: KY Srinivasan
Cc: [email protected]; [email protected]; [email protected];Quoted 4 lines
Quoted 4 lines
Quoted 1 line
Quoted 2 lines
[email protected]; [email protected]; [email protected]
Subject: Re: [PATCH RESEND 1/1] X86: Handle Hyper-V vmbusinterrupts as
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
special hypervisor interrupts
Quoted 1 line
Quoted 1 line
Quoted 1 line
On 24.01.13 at 02:17, "K. Y. Srinivasan" <[email protected]>
Quoted 1 line · wrote:
Quoted 1 line
Quoted 1 line
Review static void __init
Quoted 1 line
(quoted diff) +0
-0@@ -69,6 +74,11 @@ static void __init
ms_hyperv_init_platform(void)Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
arch/x86/kernel/cpu/mshyperv.c static void __init ms_hyperv_init_platform(void)
Quoted 8 lines
arch/x86/kernel/cpu/mshyperv.c +5
-0ms_hyperv.features, ms_hyperv.hints);clocksource_register_hz (&hyperv_cs,NSEC_PER_SEC /100 ); + +/* +* Setup the IDT for hypervisor callback. +*/ + alloc_intr_gate(HYPERVISOR_CALLBACK_VECTOR,
hyperv_callback_vector);Isn't doing this unconditionally here as problematic as the call
to
Quoted 4 lines
Quoted 4 lines
Quoted 2 lines
clocksource_register_hz()turned out to be when Xen's Hyper-V shim reacts to the CPUID inquiry above?I was not sure what to make this conditional on at run-time. To the
extent
Quoted 1 line
Quoted 1 line
that Xen emulation of Hyper-V is complete, this will be a problem.
Does Xen
Quoted 1 line
Quoted 1 line
return all the "feature bits" of Hyper-V. I could discriminate on a
feature
Quoted 43 lines
Quoted 2 lines
that
Xen does not plan to emulate.I've no idea what plans there might be, but that's the code there is currently:
int cpuid_viridian_leaves(unsigned int leaf, unsigned int *eax,
unsigned int *ebx, unsigned int *ecx, unsigned int *edx){ struct domain *d = current->domain; if ( !is_viridian_domain(d) ) return 0; leaf -= 0x40000000; if ( leaf > 6 ) return 0; *eax = *ebx = *ecx = *edx = 0; switch ( leaf ) { case 0: *eax = 0x40000006; /* Maximum leaf */ *ebx = 0x7263694d; /* Magic numbers */ *ecx = 0x666F736F; *edx = 0x76482074; break; case 1: *eax = 0x31237648; /* Version number */ break; case 2: /* Hypervisor information, but only if the guest has set its own version number. */ if ( d->arch.hvm_domain.viridian.guest_os_id.raw == 0 ) break; *eax = 1; /* Build number */ *ebx = (xen_major_version() << 16) | xen_minor_version(); *ecx = 0; /* SP */ *edx = 0; /* Service branch and number */ break; case 3: /* Which hypervisor MSRs are available to the guest */Quoted 8 lines
*eax = (CPUID3A_MSR_APIC_ACCESS | CPUID3A_MSR_HYPERCALL | CPUID3A_MSR_VP_INDEX); break; case 4: /* Recommended hypercall usage. */ if ( (d->arch.hvm_domain.viridian.guest_os_id.raw == 0) || (d->arch.hvm_domain.viridian.guest_os_id.fields.os < 4))
Quoted 10 lines
break; *eax = (CPUID4A_MSR_BASED_APIC | CPUID4A_RELAX_TIMER_INT); *ebx = 2047; /* long spin count */ break; } return 1; }Quoted 3 lines
Question is - considering you stated that this is supported starting in Win8, doesn't Hyper-V itself announce that capability in some explicit way?
Thanks Jan. Unfortunately I don't think tis interrupt delivery model is specified via a "feature" bit (I will check with the Windows guys). Perhaps, we
can have a Hyper-V specific compilation switch to address the Xen emulation issue.Regards,
K. Y
Quoted 4 lines
Jan
Signature · 1 line
Sent from my mobile phone. Please excuse brevity and lack of formatting.
Quoted 3 lines
Quoted 1 line
Quoted 1 line · H. Peter Anvin
I also have to point out the obvious problem with Xen masquerading as HyperV without implementing the same feature set...
I can't resist to state that this is what feature bits have been introduced for in CPUID leaves. As long as Xen's shim doesn't advertise a feature bit it doesn't implement, all should be fine.
Jan