FAILED: patch "[PATCH] MIPS: KVM: Fix modular KVM under QEMU" failed to apply to 3.14-stable tree

PATCH Applied [email protected] 4 participants 4 messages Jul 11, 2016 – Oct 8, 2017

Subsystems (4)

Stable branch KVM MIPS KVM MIPS

lore

Share & tools

lore

[email protected]

author maintainer

2016-07-11 23:55 UTC #0 lore raw

The patch below does not apply to the 3.14-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

Quoted 1 line

From 797179bc4fe06c89e47a9f36f886f68640b423f8 Mon Sep 17 00:00:00 2001

From: James Hogan <[email protected]>
Date: Thu, 9 Jun 2016 10:50:43 +0100
Subject: [PATCH] MIPS: KVM: Fix modular KVM under QEMU
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Copy __kvm_mips_vcpu_run() into unmapped memory, so that we can never get a TLB refill exception in it when KVM is built as a module.

This was observed to happen with the host MIPS kernel running under QEMU, due to a not entirely transparent optimisation in the QEMU TLB handling where TLB entries replaced with TLBWR are copied to a separate part of the TLB array. Code in those pages continue to be executable, but those mappings persist only until the next ASID switch, even if they are marked global.

An ASID switch happens in __kvm_mips_vcpu_run() at exception level after switching to the guest exception base. Subsequent TLB mapped kernel instructions just prior to switching to the guest trigger a TLB refill exception, which enters the guest exception handlers without updating EPC. This appears as a guest triggered TLB refill on a host kernel mapped (host KSeg2) address, which is not handled correctly as user (guest) mode accesses to kernel (host) segments always generate address error exceptions.

Signed-off-by James Hogan <[email protected]>

Cc Paolo Bonzini <[email protected]>

Cc Radim Krčmář <[email protected]>

Cc Ralf Baechle <[email protected]>

Cc [email protected]

Cc <[email protected]> # 3.10.x-

Signed-off-by Paolo Bonzini <[email protected]>

arch/mips/include/asm/kvm_host.h +1 -0

index 6733ac575da4..2d5bb133d11a 100644
@@ -338,6 +338,7 @@ struct kvm_mips_tlb {
 #define KVM_MIPS_GUEST_TLB_SIZE	64
 struct kvm_vcpu_arch {
 	void *host_ebase, *guest_ebase;
+	int (*vcpu_run)(struct kvm_run *run, struct kvm_vcpu *vcpu);
 	unsigned long host_stack;
 	unsigned long host_gp;

arch/mips/kvm/interrupt.h +1 -0

index 4ab4bdfad703..2143884709e4 100644
@@ -28,6 +28,7 @@
 #define MIPS_EXC_MAX                12
 /* XXXSL More to follow */
 
+extern char __kvm_mips_vcpu_run_end[];
 extern char mips32_exception[], mips32_exceptionEnd[];
 extern char mips32_GuestException[], mips32_GuestExceptionEnd[];

arch/mips/kvm/locore.S +1 -0

index 3ef03009de5f..828fcfc1cd7f 100644
@@ -202,6 +202,7 @@ FEXPORT(__kvm_mips_load_k0k1)
 
 	/* Jump to guest */
 	eret
+EXPORT(__kvm_mips_vcpu_run_end)
 
 VECTOR(MIPSX(exception), unknown)
 /* Find out what mode we came from and jump to the proper handler. */

arch/mips/kvm/mips.c +10 -1

index dc052fb5c7a2..44da5259f390 100644
@@ -315,6 +315,15 @@ struct kvm_vcpu *kvm_arch_vcpu_create(struct kvm *kvm, unsigned int id)
 	memcpy(gebase + offset, mips32_GuestException,
 	       mips32_GuestExceptionEnd - mips32_GuestException);
 
+#ifdef MODULE
+	offset += mips32_GuestExceptionEnd - mips32_GuestException;
+	memcpy(gebase + offset, (char *)__kvm_mips_vcpu_run,
+	       __kvm_mips_vcpu_run_end - (char *)__kvm_mips_vcpu_run);
+	vcpu->arch.vcpu_run = gebase + offset;
+#else
+	vcpu->arch.vcpu_run = __kvm_mips_vcpu_run;
+#endif
+
 	/* Invalidate the icache for these ranges */
 	local_flush_icache_range((unsigned long)gebase,
 				(unsigned long)gebase + ALIGN(size, PAGE_SIZE));

@@ -404,7 +413,7 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu, struct kvm_run *run)
 	/* Disable hardware page table walking while in guest */
 	htw_stop();
 
-	r = __kvm_mips_vcpu_run(run, vcpu);
+	r = vcpu->arch.vcpu_run(run, vcpu);
 
 	/* Re-enable HTW before enabling interrupts */
 	htw_start();

James Hogan

2016-07-13 17:14 UTC #1 lore raw

[BACKPORT 3.14.y] MIPS: KVM: Fix modular KVM under QEMU

commit 797179bc4fe06c89e47a9f36f886f68640b423f8 upstream.

Copy __kvm_mips_vcpu_run() into unmapped memory, so that we can never get a TLB refill exception in it when KVM is built as a module.

Signed-off-by James Hogan <[email protected]>

Cc Paolo Bonzini <[email protected]>

Cc Radim Krčmář <[email protected]>

Cc Ralf Baechle <[email protected]>

Cc [email protected]

Signed-off-by Paolo Bonzini <[email protected]>

[[email protected]: backported for stable 3.14]

Signed-off-by James Hogan <[email protected]>

---
 arch/mips/include/asm/kvm_host.h |  1 +
 arch/mips/kvm/kvm_locore.S       |  1 +
 arch/mips/kvm/kvm_mips.c         | 11 ++++++++++-
 arch/mips/kvm/kvm_mips_int.h     |  2 ++
 4 files changed, 14 insertions(+), 1 deletion(-)

arch/mips/include/asm/kvm_host.h +1 -0

index a995fce87791..3ff5b4921b76 100644
@@ -342,6 +342,7 @@ struct kvm_mips_tlb {
 #define KVM_MIPS_GUEST_TLB_SIZE     64
 struct kvm_vcpu_arch {
 	void *host_ebase, *guest_ebase;
+	int (*vcpu_run)(struct kvm_run *run, struct kvm_vcpu *vcpu);
 	unsigned long host_stack;
 	unsigned long host_gp;

arch/mips/kvm/kvm_locore.S +1 -0

index ba5ce99c021d..d1fa2a57218b 100644
@@ -229,6 +229,7 @@ FEXPORT(__kvm_mips_load_k0k1)
 
 	/* Jump to guest */
 	eret
+EXPORT(__kvm_mips_vcpu_run_end)
 
 VECTOR(MIPSX(exception), unknown)
 /*

arch/mips/kvm/kvm_mips.c +10 -1

index 12d850b68763..2b2dd4ec03fb 100644
@@ -348,6 +348,15 @@ struct kvm_vcpu *kvm_arch_vcpu_create(struct kvm *kvm, unsigned int id)
 	memcpy(gebase + offset, mips32_GuestException,
 	       mips32_GuestExceptionEnd - mips32_GuestException);
 
+#ifdef MODULE
+	offset += mips32_GuestExceptionEnd - mips32_GuestException;
+	memcpy(gebase + offset, (char *)__kvm_mips_vcpu_run,
+	       __kvm_mips_vcpu_run_end - (char *)__kvm_mips_vcpu_run);
+	vcpu->arch.vcpu_run = gebase + offset;
+#else
+	vcpu->arch.vcpu_run = __kvm_mips_vcpu_run;
+#endif
+
 	/* Invalidate the icache for these ranges */
 	mips32_SyncICache((unsigned long) gebase, ALIGN(size, PAGE_SIZE));

@@ -431,7 +440,7 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu, struct kvm_run *run)
 
 	kvm_guest_enter();
 
-	r = __kvm_mips_vcpu_run(run, vcpu);
+	r = vcpu->arch.vcpu_run(run, vcpu);
 
 	kvm_guest_exit();
 	local_irq_enable();

arch/mips/kvm/kvm_mips_int.h +2 -0

index 20da7d29eede..bf41ea36210e 100644
@@ -27,6 +27,8 @@
 #define MIPS_EXC_MAX                12
 /* XXXSL More to follow */
 
+extern char __kvm_mips_vcpu_run_end[];
+
 #define C_TI        (_ULCAST_(1) << 30)
 
 #define KVM_MIPS_IRQ_DELIVER_ALL_AT_ONCE (0)

Signature · 1 line

2.4.10

Jiri Slaby

Applied

2016-07-14 07:58 UTC #2 lore raw

Re: [BACKPORT 3.14.y] MIPS: KVM: Fix modular KVM under QEMU

Quoted 1 line · James Hogan

commit 797179bc4fe06c89e47a9f36f886f68640b423f8 upstream.

Applied to 3.12 too. Thanks!

Signature · 2 lines

js
suse labs

Ben Hutchings

2017-10-08 21:14 UTC #3 lore raw

Re: [BACKPORT 3.14.y] MIPS: KVM: Fix modular KVM under QEMU

In reply to James Hogan

Quoted 21 lines · James Hogan

commit 797179bc4fe06c89e47a9f36f886f68640b423f8 upstream.
Copy __kvm_mips_vcpu_run() into unmapped memory, so that we can never get a TLB refill exception in it when KVM is built as a module.
This was observed to happen with the host MIPS kernel running under QEMU, due to a not entirely transparent optimisation in the QEMU TLB handling where TLB entries replaced with TLBWR are copied to a separate part of the TLB array. Code in those pages continue to be executable, but those mappings persist only until the next ASID switch, even if they are marked global.
An ASID switch happens in __kvm_mips_vcpu_run() at exception level after switching to the guest exception base. Subsequent TLB mapped kernel instructions just prior to switching to the guest trigger a TLB refill exception, which enters the guest exception handlers without updating EPC. This appears as a guest triggered TLB refill on a host kernel mapped (host KSeg2) address, which is not handled correctly as user (guest) mode accesses to kernel (host) segments always generate address error exceptions.

Quoted 9 lines

Signed-off-by James Hogan <[email protected]>
Cc Paolo Bonzini <[email protected]>
Cc Radim Krčmář <[email protected]>
Cc Ralf Baechle <[email protected]>
Cc [email protected]
Cc [email protected]
Signed-off-by Paolo Bonzini <[email protected]>
[[email protected]: backported for stable 3.14]
Signed-off-by James Hogan <[email protected]>

[...]

Belatedly queued this up for 3.16.

Ben.

Signature · 2 lines

Ben Hutchings
compatible: Gracefully accepts erroneous data from any source