LinuxLists.cc - [PATCH v2 0/5] Add support running nested Microsoft Hypervisor

2022-11-02 17:05:31

Subject: [PATCH v2 0/5] Add support running nested Microsoft Hypervisor

2022-11-02 17:09:49

Subject: [PATCH v2 3/5] hv: Add an interface to do nested hypercalls

According to TLFS, in order to communicate to L0 hypervisor there needs
to be an additional bit set in the control register. This communication
is required to perform priviledged instructions which can only be
performed by L0 hypervisor. An example of that could be setting up the
VMBus infrastructure.

Signed-off-by: Jinank Jain <[email protected]>
---
arch/x86/include/asm/hyperv-tlfs.h | 3 ++-
arch/x86/include/asm/mshyperv.h | 42 +++++++++++++++++++++++++++---
include/asm-generic/hyperv-tlfs.h | 1 +
include/asm-generic/mshyperv.h | 1 +
4 files changed, 42 insertions(+), 5 deletions(-)

diff --git a/arch/x86/include/asm/hyperv-tlfs.h b/arch/x86/include/asm/hyperv-tlfs.h
index 0319091e2019..fd066226f12b 100644
--- a/arch/x86/include/asm/hyperv-tlfs.h
+++ b/arch/x86/include/asm/hyperv-tlfs.h
@@ -380,7 +380,8 @@ struct hv_nested_enlightenments_control {
__u32 reserved:31;
} features;
struct {
- __u32 reserved;
+ __u32 inter_partition_comm:1;
+ __u32 reserved:31;
} hypercallControls;
} __packed;

diff --git a/arch/x86/include/asm/mshyperv.h b/arch/x86/include/asm/mshyperv.h
index 415289757428..451d8c3ab63b 100644
--- a/arch/x86/include/asm/mshyperv.h
+++ b/arch/x86/include/asm/mshyperv.h
@@ -74,10 +74,16 @@ static inline u64 hv_do_hypercall(u64 control, void *input, void *output)
return hv_status;
}

+/* Hypercall to the L0 hypervisor */
+static inline u64 hv_do_nested_hypercall(u64 control, void *input, void *output)
+{
+ return hv_do_hypercall(control | HV_HYPERCALL_NESTED, input, output);
+}
+
/* Fast hypercall with 8 bytes of input and no output */
-static inline u64 hv_do_fast_hypercall8(u16 code, u64 input1)
+static inline u64 _hv_do_fast_hypercall8(u64 control, u16 code, u64 input1)
{
- u64 hv_status, control = (u64)code | HV_HYPERCALL_FAST_BIT;
+ u64 hv_status;

#ifdef CONFIG_X86_64
{
@@ -105,10 +111,24 @@ static inline u64 hv_do_fast_hypercall8(u16 code, u64 input1)
return hv_status;
}

+static inline u64 hv_do_fast_hypercall8(u16 code, u64 input1)
+{
+ u64 control = (u64)code | HV_HYPERCALL_FAST_BIT;
+
+ return _hv_do_fast_hypercall8(control, code, input1);
+}
+
+static inline u64 hv_do_fast_nested_hypercall8(u16 code, u64 input1)
+{
+ u64 control = (u64)code | HV_HYPERCALL_FAST_BIT | HV_HYPERCALL_NESTED;
+
+ return _hv_do_fast_hypercall8(control, code, input1);
+}
+
/* Fast hypercall with 16 bytes of input */
-static inline u64 hv_do_fast_hypercall16(u16 code, u64 input1, u64 input2)
+static inline u64 _hv_do_fast_hypercall16(u64 control, u16 code, u64 input1, u64 input2)
{
- u64 hv_status, control = (u64)code | HV_HYPERCALL_FAST_BIT;
+ u64 hv_status;

#ifdef CONFIG_X86_64
{
@@ -139,6 +159,20 @@ static inline u64 hv_do_fast_hypercall16(u16 code, u64 input1, u64 input2)
return hv_status;
}

+static inline u64 hv_do_fast_hypercall16(u16 code, u64 input1, u64 input2)
+{
+ u64 control = (u64)code | HV_HYPERCALL_FAST_BIT;
+
+ return _hv_do_fast_hypercall16(control, code, input1, input2);
+}
+
+static inline u64 hv_do_fast_nested_hypercall16(u16 code, u64 input1, u64 input2)
+{
+ u64 control = (u64)code | HV_HYPERCALL_FAST_BIT | HV_HYPERCALL_NESTED;
+
+ return _hv_do_fast_hypercall16(control, code, input1, input2);
+}
+
extern struct hv_vp_assist_page **hv_vp_assist_page;

static inline struct hv_vp_assist_page *hv_get_vp_assist_page(unsigned int cpu)
diff --git a/include/asm-generic/hyperv-tlfs.h b/include/asm-generic/hyperv-tlfs.h
index fdce7a4cfc6f..c67836dd8468 100644
--- a/include/asm-generic/hyperv-tlfs.h
+++ b/include/asm-generic/hyperv-tlfs.h
@@ -185,6 +185,7 @@ enum HV_GENERIC_SET_FORMAT {
#define HV_HYPERCALL_VARHEAD_OFFSET 17
#define HV_HYPERCALL_VARHEAD_MASK GENMASK_ULL(26, 17)
#define HV_HYPERCALL_RSVD0_MASK GENMASK_ULL(31, 27)
+#define HV_HYPERCALL_NESTED BIT(31)
#define HV_HYPERCALL_REP_COMP_OFFSET 32
#define HV_HYPERCALL_REP_COMP_1 BIT_ULL(32)
#define HV_HYPERCALL_REP_COMP_MASK GENMASK_ULL(43, 32)
diff --git a/include/asm-generic/mshyperv.h b/include/asm-generic/mshyperv.h
index bfb9eb9d7215..a2524d96ce2d 100644
--- a/include/asm-generic/mshyperv.h
+++ b/include/asm-generic/mshyperv.h
@@ -53,6 +53,7 @@ extern void * __percpu *hyperv_pcpu_input_arg;
extern void * __percpu *hyperv_pcpu_output_arg;

extern u64 hv_do_hypercall(u64 control, void *inputaddr, void *outputaddr);
+extern u64 hv_do_nested_hypercall(u64 control, void *inputaddr, void *outputaddr);
extern u64 hv_do_fast_hypercall8(u16 control, u64 input8);
extern bool hv_isolation_type_snp(void);

--
2.25.1

2022-11-02 17:34:42

by Jinank Jain

[permalink] [raw]

Detect if Linux is running as a nested hypervisor in the root
partition for Microsoft Hypervisor, using flags provided by MSHV.
Expose a new variable hv_nested that is used later for decisions
specific to the nested use case.

Signed-off-by: Jinank Jain <[email protected]>
---
arch/x86/include/asm/hyperv-tlfs.h | 3 +++
arch/x86/kernel/cpu/mshyperv.c | 7 +++++++
drivers/hv/hv_common.c | 9 ++++++---
include/asm-generic/mshyperv.h | 1 +
4 files changed, 17 insertions(+), 3 deletions(-)

diff --git a/arch/x86/include/asm/hyperv-tlfs.h b/arch/x86/include/asm/hyperv-tlfs.h
index 6d9368ea3701..58c03d18c235 100644
--- a/arch/x86/include/asm/hyperv-tlfs.h
+++ b/arch/x86/include/asm/hyperv-tlfs.h
@@ -114,6 +114,9 @@
/* Recommend using the newer ExProcessorMasks interface */
#define HV_X64_EX_PROCESSOR_MASKS_RECOMMENDED BIT(11)

+/* Indicates that the hypervisor is nested within a Hyper-V partition. */
+#define HV_X64_HYPERV_NESTED BIT(12)
+
/* Recommend using enlightened VMCS */
#define HV_X64_ENLIGHTENED_VMCS_RECOMMENDED BIT(14)

diff --git a/arch/x86/kernel/cpu/mshyperv.c b/arch/x86/kernel/cpu/mshyperv.c
index 46668e255421..f9b78d4829e3 100644
--- a/arch/x86/kernel/cpu/mshyperv.c
+++ b/arch/x86/kernel/cpu/mshyperv.c
@@ -37,6 +37,8 @@

/* Is Linux running as the root partition? */
bool hv_root_partition;
+/* Is Linux running on nested Microsoft Hypervisor */
+bool hv_nested;
struct ms_hyperv_info ms_hyperv;

#if IS_ENABLED(CONFIG_HYPERV)
@@ -301,6 +303,11 @@ static void __init ms_hyperv_init_platform(void)
pr_info("Hyper-V: running as root partition\n");
}

+ if (ms_hyperv.hints & HV_X64_HYPERV_NESTED) {
+ hv_nested = true;
+ pr_info("Hyper-V: running on a nested hypervisor\n");
+ }
+
/*
* Extract host information.
*/
diff --git a/drivers/hv/hv_common.c b/drivers/hv/hv_common.c
index ae68298c0dca..52a6f89ccdbd 100644
--- a/drivers/hv/hv_common.c
+++ b/drivers/hv/hv_common.c
@@ -25,17 +25,20 @@
#include <asm/mshyperv.h>

/*
- * hv_root_partition and ms_hyperv are defined here with other Hyper-V
- * specific globals so they are shared across all architectures and are
+ * hv_root_partition, ms_hyperv and hv_nested are defined here with other
+ * Hyper-V specific globals so they are shared across all architectures and are
* built only when CONFIG_HYPERV is defined. But on x86,
* ms_hyperv_init_platform() is built even when CONFIG_HYPERV is not
- * defined, and it uses these two variables. So mark them as __weak
+ * defined, and it uses these three variables. So mark them as __weak
* here, allowing for an overriding definition in the module containing
* ms_hyperv_init_platform().
*/
bool __weak hv_root_partition;
EXPORT_SYMBOL_GPL(hv_root_partition);

+bool __weak hv_nested;
+EXPORT_SYMBOL_GPL(hv_nested);
+
struct ms_hyperv_info __weak ms_hyperv;
EXPORT_SYMBOL_GPL(ms_hyperv);

diff --git a/include/asm-generic/mshyperv.h b/include/asm-generic/mshyperv.h
index bfb9eb9d7215..f131027830c3 100644
--- a/include/asm-generic/mshyperv.h
+++ b/include/asm-generic/mshyperv.h
@@ -48,6 +48,7 @@ struct ms_hyperv_info {
u64 shared_gpa_boundary;
};
extern struct ms_hyperv_info ms_hyperv;
+extern bool hv_nested;

extern void * __percpu *hyperv_pcpu_input_arg;
extern void * __percpu *hyperv_pcpu_output_arg;
--
2.25.1

2022-12-09 05:46:12

by Jinank Jain

[permalink] [raw]

On Mon, Jan 02, 2023 at 07:12:55AM +0000, Jinank Jain wrote:
> Traditionally we have been using the HYPERVISOR_CALLBACK_VECTOR to relay

Who's "we"?

Please use passive voice in your commit message: no "we" or "I", etc,
and describe your changes in imperative mood.

Also, pls read section "2) Describe your changes" in
Documentation/process/submitting-patches.rst for more details.

Also, see section "Changelog" in
Documentation/process/maintainer-tip.rst

Bottom line is: personal pronouns are ambiguous in text, especially with
so many parties/companies/etc developing the kernel so let's avoid them
please.

> the VMBus interrupt. But this does not work in case of nested
> hypervisor. Microsoft Hypervisor reserves 0x31 to 0x34 as the interrupt
> vector range for VMBus and thus we have to use one of the vectors from
> that range and setup the IDT accordingly.
>
> Signed-off-by: Jinank Jain <[email protected]>
> ---
> arch/x86/include/asm/idtentry.h | 2 ++
> arch/x86/include/asm/irq_vectors.h | 6 ++++++
> arch/x86/kernel/cpu/mshyperv.c | 15 +++++++++++++++
> arch/x86/kernel/idt.c | 10 ++++++++++
> drivers/hv/vmbus_drv.c | 3 ++-
> 5 files changed, 35 insertions(+), 1 deletion(-)
>
> diff --git a/arch/x86/include/asm/idtentry.h b/arch/x86/include/asm/idtentry.h
> index 72184b0b2219..c0648e3e4d4a 100644
> --- a/arch/x86/include/asm/idtentry.h
> +++ b/arch/x86/include/asm/idtentry.h
> @@ -686,6 +686,8 @@ DECLARE_IDTENTRY_SYSVEC(POSTED_INTR_NESTED_VECTOR, sysvec_kvm_posted_intr_nested
> DECLARE_IDTENTRY_SYSVEC(HYPERVISOR_CALLBACK_VECTOR, sysvec_hyperv_callback);
> DECLARE_IDTENTRY_SYSVEC(HYPERV_REENLIGHTENMENT_VECTOR, sysvec_hyperv_reenlightenment);
> DECLARE_IDTENTRY_SYSVEC(HYPERV_STIMER0_VECTOR, sysvec_hyperv_stimer0);
> +DECLARE_IDTENTRY_SYSVEC(HYPERV_INTR_NESTED_VMBUS_VECTOR,
> + sysvec_hyperv_nested_vmbus_intr);
> #endif
>
> #if IS_ENABLED(CONFIG_ACRN_GUEST)
> diff --git a/arch/x86/include/asm/irq_vectors.h b/arch/x86/include/asm/irq_vectors.h
> index 43dcb9284208..729d19eab7f5 100644
> --- a/arch/x86/include/asm/irq_vectors.h
> +++ b/arch/x86/include/asm/irq_vectors.h
> @@ -102,6 +102,12 @@
> #if IS_ENABLED(CONFIG_HYPERV)
> #define HYPERV_REENLIGHTENMENT_VECTOR 0xee
> #define HYPERV_STIMER0_VECTOR 0xed
> +/*
> + * FIXME: Change this, once Microsoft Hypervisor changes its assumption
^^^^^^

This patch looks like it is not ready to go anywhere yet...

> + * around VMBus interrupt vector allocation for nested root partition.

When is that going to happen? If at all...

> + * Or provides a better interface to detect this instead of hardcoding.
> + */
> +#define HYPERV_INTR_NESTED_VMBUS_VECTOR 0x31
> #endif
>
> #define LOCAL_TIMER_VECTOR 0xec
> diff --git a/arch/x86/kernel/cpu/mshyperv.c b/arch/x86/kernel/cpu/mshyperv.c
> index 938fc82edf05..4dfe0f9d7be3 100644
> --- a/arch/x86/kernel/cpu/mshyperv.c
> +++ b/arch/x86/kernel/cpu/mshyperv.c
> @@ -126,6 +126,21 @@ DEFINE_IDTENTRY_SYSVEC(sysvec_hyperv_callback)
> set_irq_regs(old_regs);
> }
>
> +DEFINE_IDTENTRY_SYSVEC(sysvec_hyperv_nested_vmbus_intr)
> +{
> + struct pt_regs *old_regs = set_irq_regs(regs);
> +
> + inc_irq_stat(irq_hv_callback_count);
> +
> + if (vmbus_handler)
> + vmbus_handler();
> +
> + if (ms_hyperv.hints & HV_DEPRECATING_AEOI_RECOMMENDED)
> + ack_APIC_irq();
> +
> + set_irq_regs(old_regs);
> +}
> +
> void hv_setup_vmbus_handler(void (*handler)(void))
> {
> vmbus_handler = handler;
> diff --git a/arch/x86/kernel/idt.c b/arch/x86/kernel/idt.c
> index a58c6bc1cd68..3536935cea39 100644
> --- a/arch/x86/kernel/idt.c
> +++ b/arch/x86/kernel/idt.c
> @@ -160,6 +160,16 @@ static const __initconst struct idt_data apic_idts[] = {
> # endif
> INTG(SPURIOUS_APIC_VECTOR, asm_sysvec_spurious_apic_interrupt),
> INTG(ERROR_APIC_VECTOR, asm_sysvec_error_interrupt),
> +#ifdef CONFIG_HYPERV
> + /*
> + * This is a hack because we cannot install this interrupt handler
> + * via alloc_intr_gate as it does not allow interrupt vector less
> + * than FIRST_SYSTEM_VECTORS. And hyperv does not want anything other
> + * than 0x31-0x34 as the interrupt vector for vmbus interrupt in case

Well:

/*
* IDT vectors usable for external interrupt sources start at 0x20.
* (0x80 is the syscall vector, 0x30-0x3f are for ISA)
^^^^^^^^^^^^^^^^^^^^^^

*/
#define FIRST_EXTERNAL_VECTOR 0x20

I guess HyperV decided to reuse those...?

--
Regards/Gruss,
Boris.

https://people.kernel.org/tglx/notes-about-netiquette

2023-01-12 20:52:44

by Thomas Gleixner

[permalink] [raw]

Subject: Re: [PATCH v10 5/5] x86/hyperv: Change interrupt vector for nested root partition

Jinank!

On Mon, Jan 02 2023 at 07:12, Jinank Jain wrote:
> --- a/arch/x86/include/asm/irq_vectors.h
> +++ b/arch/x86/include/asm/irq_vectors.h
> +/*
> + * FIXME: Change this, once Microsoft Hypervisor changes its assumption
> + * around VMBus interrupt vector allocation for nested root partition.
> + * Or provides a better interface to detect this instead of hardcoding.
> + */
> +#define HYPERV_INTR_NESTED_VMBUS_VECTOR 0x31

arch/x86/include/asm/irq_vectors.h line 47:

/*
* Vectors 0x30-0x3f are used for ISA interrupts.
* round up to the next 16-vector boundary
*/
#define ISA_IRQ_VECTOR(irq) (((FIRST_EXTERNAL_VECTOR + 16) & ~15) + irq)

So this overlaps with the legacy interrupt vector space.

> +#ifdef CONFIG_HYPERV
> + /*
> + * This is a hack because we cannot install this interrupt handler
> + * via alloc_intr_gate as it does not allow interrupt vector less
> + * than FIRST_SYSTEM_VECTORS. And hyperv does not want anything other
> + * than 0x31-0x34 as the interrupt vector for vmbus interrupt in case
> + * of nested setup.
> + */
> + INTG(HYPERV_INTR_NESTED_VMBUS_VECTOR, asm_sysvec_hyperv_nested_vmbus_intr),
> +#endif

I agree, that this is a hack, but that puts it mildly: It's a completely
broken hack.

> +DECLARE_IDTENTRY_SYSVEC(HYPERV_INTR_NESTED_VMBUS_VECTOR, sysvec_hyperv_nested_vmbus_intr);

This generates the low level entry stub for vector 0x31 at compile time,
which competes with the interrupt stub for external interrupts generated
by:

SYM_CODE_START(irq_entries_start)

Now the above INTG() hard-codes the IDT entry for vector 0x31 into the
apic_idts table. That marks it as system vector which in turn prevents
idt_setup_apic_and_irq_gates() to install the IDT entry for the external
vector on _ALL_ systems unconditionally.

IOW, you broke world except for systems which do not use the legacy
interrupt space. Congrats!

That legacy space is hardcoded and that's clearly documented so.

0x31 becomes IRQ1 - usually the i8042 - which makes it pretty much
guaranteed that this collides and fails. The worst case consequence is a
fully uncontrolled interrupt storm which is not even detectable.

So this patch is /dev/null material and either the hypervisor side makes
it possible to use a different vector space or this needs some very
careful modifications to the legacy ISA vector assignment.

Thanks,

tglx