diff mbox series

[Part2,RFC,v4,35/40] KVM: Add arch hooks to track the host write to guest memory

Message ID 20210707183616.5620-36-brijesh.singh@amd.com
State New
Headers show
Series [Part2,RFC,v4,01/40] KVM: SVM: Add support to handle AP reset MSR protocol | expand

Commit Message

Brijesh Singh July 7, 2021, 6:36 p.m. UTC 
The kvm_write_guest{_page} and kvm_vcpu_write_guest{_page} are used by
the hypevisor to write to the guest memory. The kvm_vcpu_map() and
kvm_map_gfn() are used by the hypervisor to map the guest memory and
and access it later.

When SEV-SNP is enabled in the guest VM, the guest memory pages can
either be a private or shared. A write from the hypervisor goes through
the RMP checks. If hardware sees that hypervisor is attempting to write
to a guest private page, then it triggers an RMP violation (i.e, #PF with
RMP bit set).

Enhance the KVM guest write helpers to invoke an architecture specific
hooks (kvm_arch_write_gfn_{begin,end}) to track the write access from the
hypervisor.

When SEV-SNP is enabled, the guest uses the PAGE_STATE vmgexit to ask the
hypervisor to change the page state from shared to private or vice versa.
While changing the page state to private, use the
kvm_host_write_track_is_active() to check whether the page is being
tracked for the host write access (i.e either mapped or kvm_write_guest
is in progress). If its tracked, then do not change the page state.

Signed-off-by: Brijesh Singh <brijesh.singh@amd.com>
---
 arch/x86/include/asm/kvm_host.h |  6 +++
 arch/x86/kvm/svm/sev.c          | 51 +++++++++++++++++++++
 arch/x86/kvm/svm/svm.c          |  2 +
 arch/x86/kvm/svm/svm.h          |  1 +
 arch/x86/kvm/x86.c              | 78 +++++++++++++++++++++++++++++++++
 include/linux/kvm_host.h        |  3 ++
 virt/kvm/kvm_main.c             | 21 +++++++--
 7 files changed, 159 insertions(+), 3 deletions(-)

Comments

Sean Christopherson July 19, 2021, 11:30 p.m. UTC | #1 
On Wed, Jul 07, 2021, Brijesh Singh wrote:
> The kvm_write_guest{_page} and kvm_vcpu_write_guest{_page} are used by

> the hypevisor to write to the guest memory. The kvm_vcpu_map() and

> kvm_map_gfn() are used by the hypervisor to map the guest memory and

> and access it later.

> 

> When SEV-SNP is enabled in the guest VM, the guest memory pages can

> either be a private or shared. A write from the hypervisor goes through

> the RMP checks. If hardware sees that hypervisor is attempting to write

> to a guest private page, then it triggers an RMP violation (i.e, #PF with

> RMP bit set).

> 

> Enhance the KVM guest write helpers to invoke an architecture specific

> hooks (kvm_arch_write_gfn_{begin,end}) to track the write access from the

> hypervisor.

> 

> When SEV-SNP is enabled, the guest uses the PAGE_STATE vmgexit to ask the

> hypervisor to change the page state from shared to private or vice versa.

> While changing the page state to private, use the

> kvm_host_write_track_is_active() to check whether the page is being

> tracked for the host write access (i.e either mapped or kvm_write_guest

> is in progress). If its tracked, then do not change the page state.

> 

> Signed-off-by: Brijesh Singh <brijesh.singh@amd.com>

> ---


...

> @@ -3468,3 +3489,33 @@ int sev_get_tdp_max_page_level(struct kvm_vcpu *vcpu, gpa_t gpa, int max_level)

>  

>  	return min_t(uint32_t, level, max_level);

>  }

> +

> +void sev_snp_write_page_begin(struct kvm *kvm, struct kvm_memory_slot *slot, gfn_t gfn)

> +{

> +	struct rmpentry *e;

> +	int level, rc;

> +	kvm_pfn_t pfn;

> +

> +	if (!sev_snp_guest(kvm))

> +		return;

> +

> +	pfn = gfn_to_pfn(kvm, gfn);

> +	if (is_error_noslot_pfn(pfn))

> +		return;

> +

> +	e = snp_lookup_page_in_rmptable(pfn_to_page(pfn), &level);

> +	if (unlikely(!e))

> +		return;

> +

> +	/*

> +	 * A hypervisor should never write to the guest private page. A write to the

> +	 * guest private will cause an RMP violation. If the guest page is private,

> +	 * then make it shared.


NAK on converting RMP entries in response to guest accesses.  Corrupting guest
data (due to dropping the "validated" flag) on a rogue/incorrect guest emulation
request or misconfigured PV feature is double ungood.  The potential kernel panic
below isn't much better.

And I also don't think we need this heavyweight flow for user access, e.g.
__copy_to_user(), just eat the RMP violation #PF like all other #PFs and exit
to userspace with -EFAULT.

kvm_vcpu_map() and friends might need the manual lookup, at least initially, but
in an ideal world that would be naturally handled by gup(), e.g. by unmapping
guest private memory or whatever approach TDX ends up employing to avoid #MCs.

> +	 */

> +	if (rmpentry_assigned(e)) {

> +		pr_err("SEV-SNP: write to guest private gfn %llx\n", gfn);

> +		rc = snp_make_page_shared(kvm_get_vcpu(kvm, 0),

> +				gfn << PAGE_SHIFT, pfn, PG_LEVEL_4K);

> +		BUG_ON(rc != 0);

> +	}

> +}


...

> +void kvm_arch_write_gfn_begin(struct kvm *kvm, struct kvm_memory_slot *slot, gfn_t gfn)

> +{

> +	update_gfn_track(slot, gfn, KVM_PAGE_TRACK_WRITE, 1);


Tracking only writes isn't correct, as KVM reads to guest private memory will
return garbage.  Pulling the rug out from under KVM reads won't fail as
spectacularly as writes (at least not right away), but they'll still fail.  I'm
actually ok reading garbage if the guest screws up, but KVM needs consistent
semantics.

Good news is that per-gfn tracking is probably overkill anyways.  As mentioned
above, user access don't need extra magic, they either fail or they don't.

For kvm_vcpu_map(), one thought would be to add a "short-term" map variant that
is not allowed to be retained across VM-Entry, and then use e.g. SRCU to block
PSC requests until there are no consumers.

> +	if (kvm_x86_ops.write_page_begin)

> +		kvm_x86_ops.write_page_begin(kvm, slot, gfn);

> +}
Brijesh Singh July 20, 2021, 3:15 p.m. UTC | #2 
On 7/19/21 6:30 PM, Sean Christopherson wrote:
...>
> NAK on converting RMP entries in response to guest accesses.  Corrupting guest

> data (due to dropping the "validated" flag) on a rogue/incorrect guest emulation

> request or misconfigured PV feature is double ungood.  The potential kernel panic

> below isn't much better.

> 


I also debated myself whether its okay to transition the page state to 
shared to complete the write operation. I am good with removing the 
converting RMP entries from the patch, and that will also remove the 
kernel panic code.


> And I also don't think we need this heavyweight flow for user access, e.g.

> __copy_to_user(), just eat the RMP violation #PF like all other #PFs and exit

> to userspace with -EFAULT.

>


Yes, we could improve the __copy_to_user() to eat the RMP violation. I 
was tempted to go on that path but struggled to find a strong reason for 
it and was not sure if that accepted. I can add that support in next rev.



> kvm_vcpu_map() and friends might need the manual lookup, at least initially, 


Yes, the enhancement to the __copy_to_user() does not solve this problem 
and for it we need to do the manually lookup.


but
> in an ideal world that would be naturally handled by gup(), e.g. by unmapping

> guest private memory or whatever approach TDX ends up employing to avoid #MCs.


> 

>> +	 */

>> +	if (rmpentry_assigned(e)) {

>> +		pr_err("SEV-SNP: write to guest private gfn %llx\n", gfn);

>> +		rc = snp_make_page_shared(kvm_get_vcpu(kvm, 0),

>> +				gfn << PAGE_SHIFT, pfn, PG_LEVEL_4K);

>> +		BUG_ON(rc != 0);

>> +	}

>> +}

> 

> ...

> 

>> +void kvm_arch_write_gfn_begin(struct kvm *kvm, struct kvm_memory_slot *slot, gfn_t gfn)

>> +{

>> +	update_gfn_track(slot, gfn, KVM_PAGE_TRACK_WRITE, 1);

> 

> Tracking only writes isn't correct, as KVM reads to guest private memory will

> return garbage.  Pulling the rug out from under KVM reads won't fail as

> spectacularly as writes (at least not right away), but they'll still fail.  I'm

> actually ok reading garbage if the guest screws up, but KVM needs consistent

> semantics.

> 

> Good news is that per-gfn tracking is probably overkill anyways.  As mentioned

> above, user access don't need extra magic, they either fail or they don't.

> 

> For kvm_vcpu_map(), one thought would be to add a "short-term" map variant that

> is not allowed to be retained across VM-Entry, and then use e.g. SRCU to block

> PSC requests until there are no consumers.

> 


Sounds good to me, i will add the support in the next rev.

thanks
diff mbox series

Patch

diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index 59185b6bc82a..678992e9966a 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -865,10 +865,13 @@  struct kvm_lpage_info {
 	int disallow_lpage;
 };
 
+bool kvm_host_write_track_is_active(struct kvm *kvm, gfn_t gfn);
+
 struct kvm_arch_memory_slot {
 	struct kvm_rmap_head *rmap[KVM_NR_PAGE_SIZES];
 	struct kvm_lpage_info *lpage_info[KVM_NR_PAGE_SIZES - 1];
 	unsigned short *gfn_track[KVM_PAGE_TRACK_MAX];
+	unsigned short *host_write_track[KVM_PAGE_TRACK_MAX];
 };
 
 /*
@@ -1393,6 +1396,9 @@  struct kvm_x86_ops {
 	void (*vcpu_deliver_sipi_vector)(struct kvm_vcpu *vcpu, u8 vector);
 	void *(*alloc_apic_backing_page)(struct kvm_vcpu *vcpu);
 	int (*get_tdp_max_page_level)(struct kvm_vcpu *vcpu, gpa_t gpa, int max_level);
+
+	void (*write_page_begin)(struct kvm *kvm, struct kvm_memory_slot *slot, gfn_t gfn);
+	void (*write_page_end)(struct kvm *kvm, struct kvm_memory_slot *slot, gfn_t gfn);
 };
 
 struct kvm_x86_nested_ops {
diff --git a/arch/x86/kvm/svm/sev.c b/arch/x86/kvm/svm/sev.c
index 0155d9b3127d..839cf321c6dd 100644
--- a/arch/x86/kvm/svm/sev.c
+++ b/arch/x86/kvm/svm/sev.c
@@ -2884,6 +2884,19 @@  static int snp_make_page_shared(struct kvm_vcpu *vcpu, gpa_t gpa, kvm_pfn_t pfn,
 	return rmpupdate(pfn_to_page(pfn), &val);
 }
 
+static inline bool kvm_host_write_track_gpa_range_is_active(struct kvm *kvm,
+							    gpa_t start, gpa_t end)
+{
+	while (start < end) {
+		if (kvm_host_write_track_is_active(kvm, gpa_to_gfn(start)))
+			return 1;
+
+		start += PAGE_SIZE;
+	}
+
+	return false;
+}
+
 static int snp_make_page_private(struct kvm_vcpu *vcpu, gpa_t gpa, kvm_pfn_t pfn, int level)
 {
 	struct kvm_sev_info *sev = &to_kvm_svm(vcpu->kvm)->sev_info;
@@ -2895,6 +2908,14 @@  static int snp_make_page_private(struct kvm_vcpu *vcpu, gpa_t gpa, kvm_pfn_t pfn
 	if (!e)
 		return -EINVAL;
 
+	/*
+	 * If the GPA is tracked for the write access then do not change the
+	 * page state from shared to private.
+	 */
+	if (kvm_host_write_track_gpa_range_is_active(vcpu->kvm,
+		gpa, gpa + page_level_size(level)))
+		return -EBUSY;
+
 	/* Log if the entry is validated */
 	if (rmpentry_validated(e))
 		pr_warn_ratelimited("Asked to make a pre-validated gpa %llx private\n", gpa);
@@ -3468,3 +3489,33 @@  int sev_get_tdp_max_page_level(struct kvm_vcpu *vcpu, gpa_t gpa, int max_level)
 
 	return min_t(uint32_t, level, max_level);
 }
+
+void sev_snp_write_page_begin(struct kvm *kvm, struct kvm_memory_slot *slot, gfn_t gfn)
+{
+	struct rmpentry *e;
+	int level, rc;
+	kvm_pfn_t pfn;
+
+	if (!sev_snp_guest(kvm))
+		return;
+
+	pfn = gfn_to_pfn(kvm, gfn);
+	if (is_error_noslot_pfn(pfn))
+		return;
+
+	e = snp_lookup_page_in_rmptable(pfn_to_page(pfn), &level);
+	if (unlikely(!e))
+		return;
+
+	/*
+	 * A hypervisor should never write to the guest private page. A write to the
+	 * guest private will cause an RMP violation. If the guest page is private,
+	 * then make it shared.
+	 */
+	if (rmpentry_assigned(e)) {
+		pr_err("SEV-SNP: write to guest private gfn %llx\n", gfn);
+		rc = snp_make_page_shared(kvm_get_vcpu(kvm, 0),
+				gfn << PAGE_SHIFT, pfn, PG_LEVEL_4K);
+		BUG_ON(rc != 0);
+	}
+}
diff --git a/arch/x86/kvm/svm/svm.c b/arch/x86/kvm/svm/svm.c
index 2632eae52aa3..4ff6fc86dd18 100644
--- a/arch/x86/kvm/svm/svm.c
+++ b/arch/x86/kvm/svm/svm.c
@@ -4577,6 +4577,8 @@  static struct kvm_x86_ops svm_x86_ops __initdata = {
 
 	.alloc_apic_backing_page = svm_alloc_apic_backing_page,
 	.get_tdp_max_page_level = sev_get_tdp_max_page_level,
+
+	.write_page_begin = sev_snp_write_page_begin,
 };
 
 static struct kvm_x86_init_ops svm_init_ops __initdata = {
diff --git a/arch/x86/kvm/svm/svm.h b/arch/x86/kvm/svm/svm.h
index af4cce39b30f..e0276ad8a1ae 100644
--- a/arch/x86/kvm/svm/svm.h
+++ b/arch/x86/kvm/svm/svm.h
@@ -576,6 +576,7 @@  void sev_es_prepare_guest_switch(struct vcpu_svm *svm, unsigned int cpu);
 void sev_es_unmap_ghcb(struct vcpu_svm *svm);
 struct page *snp_safe_alloc_page(struct kvm_vcpu *vcpu);
 int sev_get_tdp_max_page_level(struct kvm_vcpu *vcpu, gpa_t gpa, int max_level);
+void sev_snp_write_page_begin(struct kvm *kvm, struct kvm_memory_slot *slot, gfn_t gfn);
 
 /* vmenter.S */
 
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index bbc4e04e67ad..1398b8021982 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -9076,6 +9076,48 @@  void kvm_arch_mmu_notifier_invalidate_range(struct kvm *kvm,
 		kvm_make_all_cpus_request(kvm, KVM_REQ_APIC_PAGE_RELOAD);
 }
 
+static void update_gfn_track(struct kvm_memory_slot *slot, gfn_t gfn,
+			     enum kvm_page_track_mode mode, short count)
+{
+	int index, val;
+
+	index = gfn_to_index(gfn, slot->base_gfn, PG_LEVEL_4K);
+
+	val = slot->arch.host_write_track[mode][index];
+
+	if (WARN_ON(val + count < 0 || val + count > USHRT_MAX))
+		return;
+
+	slot->arch.host_write_track[mode][index] += count;
+}
+
+bool kvm_host_write_track_is_active(struct kvm *kvm, gfn_t gfn)
+{
+	struct kvm_memory_slot *slot;
+	int index;
+
+	slot = gfn_to_memslot(kvm, gfn);
+	if (!slot)
+		return false;
+
+	index = gfn_to_index(gfn, slot->base_gfn, PG_LEVEL_4K);
+	return !!READ_ONCE(slot->arch.host_write_track[KVM_PAGE_TRACK_WRITE][index]);
+}
+EXPORT_SYMBOL_GPL(kvm_host_write_track_is_active);
+
+void kvm_arch_write_gfn_begin(struct kvm *kvm, struct kvm_memory_slot *slot, gfn_t gfn)
+{
+	update_gfn_track(slot, gfn, KVM_PAGE_TRACK_WRITE, 1);
+
+	if (kvm_x86_ops.write_page_begin)
+		kvm_x86_ops.write_page_begin(kvm, slot, gfn);
+}
+
+void kvm_arch_write_gfn_end(struct kvm *kvm, struct kvm_memory_slot *slot, gfn_t gfn)
+{
+	update_gfn_track(slot, gfn, KVM_PAGE_TRACK_WRITE, -1);
+}
+
 void kvm_vcpu_reload_apic_access_page(struct kvm_vcpu *vcpu)
 {
 	if (!lapic_in_kernel(vcpu))
@@ -10896,6 +10938,36 @@  void kvm_arch_destroy_vm(struct kvm *kvm)
 	kvm_hv_destroy_vm(kvm);
 }
 
+static void kvm_write_page_track_free_memslot(struct kvm_memory_slot *slot)
+{
+	int i;
+
+	for (i = 0; i < KVM_PAGE_TRACK_MAX; i++) {
+		kvfree(slot->arch.host_write_track[i]);
+		slot->arch.host_write_track[i] = NULL;
+	}
+}
+
+static int kvm_write_page_track_create_memslot(struct kvm_memory_slot *slot,
+					       unsigned long npages)
+{
+	int  i;
+
+	for (i = 0; i < KVM_PAGE_TRACK_MAX; i++) {
+		slot->arch.host_write_track[i] =
+			kvcalloc(npages, sizeof(*slot->arch.host_write_track[i]),
+				 GFP_KERNEL_ACCOUNT);
+		if (!slot->arch.host_write_track[i])
+			goto track_free;
+	}
+
+	return 0;
+
+track_free:
+	kvm_write_page_track_free_memslot(slot);
+	return -ENOMEM;
+}
+
 void kvm_arch_free_memslot(struct kvm *kvm, struct kvm_memory_slot *slot)
 {
 	int i;
@@ -10969,8 +11041,14 @@  static int kvm_alloc_memslot_metadata(struct kvm_memory_slot *slot,
 	if (kvm_page_track_create_memslot(slot, npages))
 		goto out_free;
 
+	if (kvm_write_page_track_create_memslot(slot, npages))
+		goto e_free_page_track;
+
 	return 0;
 
+e_free_page_track:
+	kvm_page_track_free_memslot(slot);
+
 out_free:
 	for (i = 0; i < KVM_NR_PAGE_SIZES; ++i) {
 		kvfree(slot->arch.rmap[i]);
diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
index 2f34487e21f2..f22e22cd2179 100644
--- a/include/linux/kvm_host.h
+++ b/include/linux/kvm_host.h
@@ -1550,6 +1550,9 @@  static inline long kvm_arch_vcpu_async_ioctl(struct file *filp,
 void kvm_arch_mmu_notifier_invalidate_range(struct kvm *kvm,
 					    unsigned long start, unsigned long end);
 
+void kvm_arch_write_gfn_begin(struct kvm *kvm, struct kvm_memory_slot *slot, gfn_t gfn);
+void kvm_arch_write_gfn_end(struct kvm *kvm, struct kvm_memory_slot *slot, gfn_t gfn);
+
 #ifdef CONFIG_HAVE_KVM_VCPU_RUN_PID_CHANGE
 int kvm_arch_vcpu_run_pid_change(struct kvm_vcpu *vcpu);
 #else
diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index 6b4feb92dc79..bc805c15d0de 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -160,6 +160,14 @@  __weak void kvm_arch_mmu_notifier_invalidate_range(struct kvm *kvm,
 {
 }
 
+__weak void kvm_arch_write_gfn_begin(struct kvm *kvm, struct kvm_memory_slot *slot, gfn_t gfn)
+{
+}
+
+__weak void kvm_arch_write_gfn_end(struct kvm *kvm, struct kvm_memory_slot *slot, gfn_t gfn)
+{
+}
+
 bool kvm_is_zone_device_pfn(kvm_pfn_t pfn)
 {
 	/*
@@ -2309,7 +2317,8 @@  static void kvm_cache_gfn_to_pfn(struct kvm_memory_slot *slot, gfn_t gfn,
 	cache->generation = gen;
 }
 
-static int __kvm_map_gfn(struct kvm_memslots *slots, gfn_t gfn,
+static int __kvm_map_gfn(struct kvm *kvm,
+			 struct kvm_memslots *slots, gfn_t gfn,
 			 struct kvm_host_map *map,
 			 struct gfn_to_pfn_cache *cache,
 			 bool atomic)
@@ -2361,20 +2370,22 @@  static int __kvm_map_gfn(struct kvm_memslots *slots, gfn_t gfn,
 	map->pfn = pfn;
 	map->gfn = gfn;
 
+	kvm_arch_write_gfn_begin(kvm, slot, map->gfn);
+
 	return 0;
 }
 
 int kvm_map_gfn(struct kvm_vcpu *vcpu, gfn_t gfn, struct kvm_host_map *map,
 		struct gfn_to_pfn_cache *cache, bool atomic)
 {
-	return __kvm_map_gfn(kvm_memslots(vcpu->kvm), gfn, map,
+	return __kvm_map_gfn(vcpu->kvm, kvm_memslots(vcpu->kvm), gfn, map,
 			cache, atomic);
 }
 EXPORT_SYMBOL_GPL(kvm_map_gfn);
 
 int kvm_vcpu_map(struct kvm_vcpu *vcpu, gfn_t gfn, struct kvm_host_map *map)
 {
-	return __kvm_map_gfn(kvm_vcpu_memslots(vcpu), gfn, map,
+	return __kvm_map_gfn(vcpu->kvm, kvm_vcpu_memslots(vcpu), gfn, map,
 		NULL, false);
 }
 EXPORT_SYMBOL_GPL(kvm_vcpu_map);
@@ -2412,6 +2423,8 @@  static void __kvm_unmap_gfn(struct kvm *kvm,
 	else
 		kvm_release_pfn(map->pfn, dirty, NULL);
 
+	kvm_arch_write_gfn_end(kvm, memslot, map->gfn);
+
 	map->hva = NULL;
 	map->page = NULL;
 }
@@ -2612,7 +2625,9 @@  static int __kvm_write_guest_page(struct kvm *kvm,
 	addr = gfn_to_hva_memslot(memslot, gfn);
 	if (kvm_is_error_hva(addr))
 		return -EFAULT;
+	kvm_arch_write_gfn_begin(kvm, memslot, gfn);
 	r = __copy_to_user((void __user *)addr + offset, data, len);
+	kvm_arch_write_gfn_end(kvm, memslot, gfn);
 	if (r)
 		return -EFAULT;
 	mark_page_dirty_in_slot(kvm, memslot, gfn);