public inbox for kvm@vger.kernel.org
 help / color / mirror / Atom feed
* [kvm-unit-tests PATCH] x86/emulator64: Extend non-canonical memory access tests with CR2 coverage
@ 2025-06-12 14:16 Mathias Krause
  2025-06-26 14:49 ` Sean Christopherson
  0 siblings, 1 reply; 4+ messages in thread
From: Mathias Krause @ 2025-06-12 14:16 UTC (permalink / raw)
  To: Sean Christopherson, Paolo Bonzini; +Cc: kvm, Mathias Krause

Extend the non-canonical memory access tests to verify CR2 stays
unchanged.

There's currently a bug in QEMU/TCG that breaks that assumption.

Link: https://gitlab.com/qemu-project/qemu/-/issues/928
Signed-off-by: Mathias Krause <minipli@grsecurity.net>
---
 x86/emulator64.c | 26 ++++++++++++++++++++++++++
 1 file changed, 26 insertions(+)

diff --git a/x86/emulator64.c b/x86/emulator64.c
index 5d1bb0f06d4f..abef2bda29f1 100644
--- a/x86/emulator64.c
+++ b/x86/emulator64.c
@@ -325,16 +325,39 @@ static void test_mmx_movq_mf(uint64_t *mem)
 	report(exception_vector() == MF_VECTOR, "movq mmx generates #MF");
 }
 
+#define CR2_REF_VALUE	0xdecafbadUL
+
+static void setup_cr2(void)
+{
+	write_cr2(CR2_REF_VALUE);
+}
+
+static void check_cr2(void)
+{
+	unsigned long cr2 = read_cr2();
+
+	if (cr2 == CR2_REF_VALUE) {
+		report(true, "CR2 unchanged");
+	} else {
+		report(false, "CR2 changed from %#lx to %#lx", CR2_REF_VALUE, cr2);
+		setup_cr2();
+	}
+}
+
 static void test_jmp_noncanonical(uint64_t *mem)
 {
+	setup_cr2();
 	*mem = NONCANONICAL;
 	asm volatile (ASM_TRY("1f") "jmp *%0; 1:" : : "m"(*mem));
 	report(exception_vector() == GP_VECTOR,
 	       "jump to non-canonical address");
+	check_cr2();
 }
 
 static void test_reg_noncanonical(void)
 {
+	setup_cr2();
+
 	/* RAX based, should #GP(0) */
 	asm volatile(ASM_TRY("1f") "orq $0, (%[noncanonical]); 1:"
 		     : : [noncanonical]"a"(NONCANONICAL));
@@ -342,6 +365,7 @@ static void test_reg_noncanonical(void)
 	       "non-canonical memory access, should %s(0), got %s(%u)",
 	       exception_mnemonic(GP_VECTOR),
 	       exception_mnemonic(exception_vector()), exception_error_code());
+	check_cr2();
 
 	/* RSP based, should #SS(0) */
 	asm volatile(ASM_TRY("1f") "orq $0, (%%rsp,%[noncanonical],1); 1:"
@@ -350,6 +374,7 @@ static void test_reg_noncanonical(void)
 	       "non-canonical rsp-based access, should %s(0), got %s(%u)",
 	       exception_mnemonic(SS_VECTOR),
 	       exception_mnemonic(exception_vector()), exception_error_code());
+	check_cr2();
 
 	/* RBP based, should #SS(0) */
 	asm volatile(ASM_TRY("1f") "orq $0, (%%rbp,%[noncanonical],1); 1:"
@@ -358,6 +383,7 @@ static void test_reg_noncanonical(void)
 	       "non-canonical rbp-based access, should %s(0), got %s(%u)",
 	       exception_mnemonic(SS_VECTOR),
 	       exception_mnemonic(exception_vector()), exception_error_code());
+	check_cr2();
 }
 
 static void test_movabs(uint64_t *mem)
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 4+ messages in thread

* Re: [kvm-unit-tests PATCH] x86/emulator64: Extend non-canonical memory access tests with CR2 coverage
  2025-06-12 14:16 [kvm-unit-tests PATCH] x86/emulator64: Extend non-canonical memory access tests with CR2 coverage Mathias Krause
@ 2025-06-26 14:49 ` Sean Christopherson
  2025-06-27 14:33   ` Mathias Krause
  2025-11-18 22:39   ` Sean Christopherson
  0 siblings, 2 replies; 4+ messages in thread
From: Sean Christopherson @ 2025-06-26 14:49 UTC (permalink / raw)
  To: Mathias Krause; +Cc: Paolo Bonzini, kvm

[-- Attachment #1: Type: text/plain, Size: 3659 bytes --]

On Thu, Jun 12, 2025, Mathias Krause wrote:
> Extend the non-canonical memory access tests to verify CR2 stays
> unchanged.
> 
> There's currently a bug in QEMU/TCG that breaks that assumption.
> 
> Link: https://gitlab.com/qemu-project/qemu/-/issues/928
> Signed-off-by: Mathias Krause <minipli@grsecurity.net>
> ---
>  x86/emulator64.c | 26 ++++++++++++++++++++++++++
>  1 file changed, 26 insertions(+)
> 
> diff --git a/x86/emulator64.c b/x86/emulator64.c
> index 5d1bb0f06d4f..abef2bda29f1 100644
> --- a/x86/emulator64.c
> +++ b/x86/emulator64.c
> @@ -325,16 +325,39 @@ static void test_mmx_movq_mf(uint64_t *mem)
>  	report(exception_vector() == MF_VECTOR, "movq mmx generates #MF");
>  }
>  
> +#define CR2_REF_VALUE	0xdecafbadUL
> +
> +static void setup_cr2(void)
> +{
> +	write_cr2(CR2_REF_VALUE);
> +}
> +
> +static void check_cr2(void)
> +{
> +	unsigned long cr2 = read_cr2();
> +
> +	if (cr2 == CR2_REF_VALUE) {
> +		report(true, "CR2 unchanged");
> +	} else {
> +		report(false, "CR2 changed from %#lx to %#lx", CR2_REF_VALUE, cr2);
> +		setup_cr2();

Writing CR2 isn't expensive in the grand scheme, so rather than conditionally
re-write CR2, I think it makes sense to write CR2 at the start of every testcase,
and then just do "report(cr2 == CR2_REF_VALUE".

> +	}
> +}
> +
>  static void test_jmp_noncanonical(uint64_t *mem)
>  {
> +	setup_cr2();
>  	*mem = NONCANONICAL;
>  	asm volatile (ASM_TRY("1f") "jmp *%0; 1:" : : "m"(*mem));
>  	report(exception_vector() == GP_VECTOR,
>  	       "jump to non-canonical address");
> +	check_cr2();
>  }
>  
>  static void test_reg_noncanonical(void)
>  {
> +	setup_cr2();
> +
>  	/* RAX based, should #GP(0) */
>  	asm volatile(ASM_TRY("1f") "orq $0, (%[noncanonical]); 1:"
>  		     : : [noncanonical]"a"(NONCANONICAL));
> @@ -342,6 +365,7 @@ static void test_reg_noncanonical(void)
>  	       "non-canonical memory access, should %s(0), got %s(%u)",
>  	       exception_mnemonic(GP_VECTOR),
>  	       exception_mnemonic(exception_vector()), exception_error_code());
> +	check_cr2();

And then rather than add more copy+paste, what if we add a macro to handle the
checks?  Then the CR2 validation can slot in nicely (and maybe someday the macro
could be used outside of the x86/emulator64.c).

Attached patches yield:	

#define CR2_REF_VALUE	0xdecafbadUL

#define ASM_TRY_NONCANONICAL(insn, inputs, access, ex_vector)			\
do {										\
	unsigned int vector, ec;						\
										\
	write_cr2(CR2_REF_VALUE);						\
										\
	asm volatile(ASM_TRY("1f") insn "; 1:" :: inputs);			\
										\
	vector = exception_vector();						\
	ec = exception_error_code();						\
										\
	report(vector == ex_vector && !ec,					\
	      "non-canonical " access ", should %s(0), got %s(%u)",		\
	      exception_mnemonic(ex_vector), exception_mnemonic(vector), ec);	\
										\
	if (vector != PF_VECTOR) {						\
		unsigned long cr2  = read_cr2();				\
										\
		report(cr2 == CR2_REF_VALUE,					\
		       "Wanted CR2 '0x%lx', got '0x%lx", CR2_REF_VALUE, cr2);	\
	}									\
} while (0)

static void test_jmp_noncanonical(uint64_t *mem)
{
	*mem = NONCANONICAL;

	ASM_TRY_NONCANONICAL("jmp *%0", "m"(*mem), "jmp", GP_VECTOR);
}

static void test_reg_noncanonical(void)
{
	/* RAX based, should #GP(0) */
	ASM_TRY_NONCANONICAL("orq $0, (%[nc])", [nc]"a"(NONCANONICAL),
			     "memory access", GP_VECTOR);

	/* RSP based, should #SS(0) */
	ASM_TRY_NONCANONICAL("orq $0, (%%rsp,%[nc],1)", [nc]"r"(NONCANONICAL),
			     "rsp-based access", SS_VECTOR);

	/* RBP based, should #SS(0) */
	ASM_TRY_NONCANONICAL("orq $0, (%%rbp,%[nc],1)", [nc]"r"(NONCANONICAL),
			     "rbp-based access", SS_VECTOR);
}

[-- Attachment #2: 0001-x86-emulator64-Add-macro-to-test-emulation-of-non-ca.patch --]
[-- Type: text/x-diff, Size: 3492 bytes --]

From 0a7ee3543ef899ea36614ddff56c306dd63c341c Mon Sep 17 00:00:00 2001
From: Sean Christopherson <seanjc@google.com>
Date: Thu, 26 Jun 2025 07:39:43 -0700
Subject: [PATCH 1/2] x86/emulator64: Add macro to test emulation of
 non-canonical accesses

Add a macro to "try" and check a non-canonical access.  In addition to
de-duplicating the checking logic, this will allow extending the logic to
verify that CR2 isn't incorrectly modified, e.g. on #GP/#SS.

No functional change intended (ignoring the newly added check on the error
code for the JMP case).

Signed-off-by: Sean Christopherson <seanjc@google.com>
---
 x86/emulator64.c | 43 ++++++++++++++++++++++---------------------
 1 file changed, 22 insertions(+), 21 deletions(-)

diff --git a/x86/emulator64.c b/x86/emulator64.c
index 138903af..21df3b0a 100644
--- a/x86/emulator64.c
+++ b/x86/emulator64.c
@@ -325,39 +325,40 @@ static void test_mmx_movq_mf(uint64_t *mem)
 	report(exception_vector() == MF_VECTOR, "movq mmx generates #MF");
 }
 
+#define ASM_TRY_NONCANONICAL(insn, inputs, access, ex_vector)			\
+do {										\
+	unsigned int vector, ec;						\
+										\
+	asm volatile(ASM_TRY("1f") insn "; 1:" :: inputs);			\
+										\
+	vector = exception_vector();						\
+	ec = exception_error_code();						\
+										\
+	report(vector == ex_vector && !ec,					\
+	      "non-canonical " access ", should %s(0), got %s(%u)",		\
+	      exception_mnemonic(ex_vector), exception_mnemonic(vector), ec);	\
+} while (0)
+
 static void test_jmp_noncanonical(uint64_t *mem)
 {
 	*mem = NONCANONICAL;
-	asm volatile (ASM_TRY("1f") "jmp *%0; 1:" : : "m"(*mem));
-	report(exception_vector() == GP_VECTOR,
-	       "jump to non-canonical address");
+
+	ASM_TRY_NONCANONICAL("jmp *%0", "m"(*mem), "jmp", GP_VECTOR);
 }
 
 static void test_reg_noncanonical(void)
 {
 	/* RAX based, should #GP(0) */
-	asm volatile(ASM_TRY("1f") "orq $0, (%[noncanonical]); 1:"
-		     : : [noncanonical]"a"(NONCANONICAL));
-	report(exception_vector() == GP_VECTOR && exception_error_code() == 0,
-	       "non-canonical memory access, should %s(0), got %s(%u)",
-	       exception_mnemonic(GP_VECTOR),
-	       exception_mnemonic(exception_vector()), exception_error_code());
+	ASM_TRY_NONCANONICAL("orq $0, (%[nc])", [nc]"a"(NONCANONICAL),
+			     "memory access", GP_VECTOR);
 
 	/* RSP based, should #SS(0) */
-	asm volatile(ASM_TRY("1f") "orq $0, (%%rsp,%[noncanonical],1); 1:"
-		     : : [noncanonical]"r"(NONCANONICAL));
-	report(exception_vector() == SS_VECTOR && exception_error_code() == 0,
-	       "non-canonical rsp-based access, should %s(0), got %s(%u)",
-	       exception_mnemonic(SS_VECTOR),
-	       exception_mnemonic(exception_vector()), exception_error_code());
+	ASM_TRY_NONCANONICAL("orq $0, (%%rsp,%[nc],1)", [nc]"r"(NONCANONICAL),
+			     "rsp-based access", SS_VECTOR);
 
 	/* RBP based, should #SS(0) */
-	asm volatile(ASM_TRY("1f") "orq $0, (%%rbp,%[noncanonical],1); 1:"
-		     : : [noncanonical]"r"(NONCANONICAL));
-	report(exception_vector() == SS_VECTOR && exception_error_code() == 0,
-	       "non-canonical rbp-based access, should %s(0), got %s(%u)",
-	       exception_mnemonic(SS_VECTOR),
-	       exception_mnemonic(exception_vector()), exception_error_code());
+	ASM_TRY_NONCANONICAL("orq $0, (%%rbp,%[nc],1)", [nc]"r"(NONCANONICAL),
+			     "rbp-based access", SS_VECTOR);
 }
 
 static void test_movabs(uint64_t *mem)

base-commit: 525bdb5d65d51a367341f471eb1bcd505d73c51f
-- 
2.50.0.727.gbf7dc18ff4-goog


[-- Attachment #3: 0002-x86-emulator64-Extend-non-canonical-memory-access-te.patch --]
[-- Type: text/x-diff, Size: 1767 bytes --]

From efb11a007a0e041ef029753b0b98abe071008334 Mon Sep 17 00:00:00 2001
From: Mathias Krause <minipli@grsecurity.net>
Date: Thu, 26 Jun 2025 07:42:51 -0700
Subject: [PATCH 2/2] x86/emulator64: Extend non-canonical memory access tests
 with CR2 coverage

Extend the non-canonical memory access tests to verify CR2 stays
unchanged.

There's currently a bug in QEMU/TCG that breaks that assumption.

Link: https://gitlab.com/qemu-project/qemu/-/issues/928
Signed-off-by: Mathias Krause <minipli@grsecurity.net>
Signed-off-by: Sean Christopherson <seanjc@google.com>
---
 x86/emulator64.c | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/x86/emulator64.c b/x86/emulator64.c
index 21df3b0a..6a85122f 100644
--- a/x86/emulator64.c
+++ b/x86/emulator64.c
@@ -325,10 +325,14 @@ static void test_mmx_movq_mf(uint64_t *mem)
 	report(exception_vector() == MF_VECTOR, "movq mmx generates #MF");
 }
 
+#define CR2_REF_VALUE	0xdecafbadUL
+
 #define ASM_TRY_NONCANONICAL(insn, inputs, access, ex_vector)			\
 do {										\
 	unsigned int vector, ec;						\
 										\
+	write_cr2(CR2_REF_VALUE);						\
+										\
 	asm volatile(ASM_TRY("1f") insn "; 1:" :: inputs);			\
 										\
 	vector = exception_vector();						\
@@ -337,6 +341,13 @@ do {										\
 	report(vector == ex_vector && !ec,					\
 	      "non-canonical " access ", should %s(0), got %s(%u)",		\
 	      exception_mnemonic(ex_vector), exception_mnemonic(vector), ec);	\
+										\
+	if (vector != PF_VECTOR) {						\
+		unsigned long cr2  = read_cr2();				\
+										\
+		report(cr2 == CR2_REF_VALUE,					\
+		       "Wanted CR2 '0x%lx', got '0x%lx", CR2_REF_VALUE, cr2);	\
+	}									\
 } while (0)
 
 static void test_jmp_noncanonical(uint64_t *mem)
-- 
2.50.0.727.gbf7dc18ff4-goog


^ permalink raw reply related	[flat|nested] 4+ messages in thread

* Re: [kvm-unit-tests PATCH] x86/emulator64: Extend non-canonical memory access tests with CR2 coverage
  2025-06-26 14:49 ` Sean Christopherson
@ 2025-06-27 14:33   ` Mathias Krause
  2025-11-18 22:39   ` Sean Christopherson
  1 sibling, 0 replies; 4+ messages in thread
From: Mathias Krause @ 2025-06-27 14:33 UTC (permalink / raw)
  To: Sean Christopherson; +Cc: Paolo Bonzini, kvm

On 26.06.25 16:49, Sean Christopherson wrote:
> On Thu, Jun 12, 2025, Mathias Krause wrote:
>> Extend the non-canonical memory access tests to verify CR2 stays
>> unchanged.
>>
>> There's currently a bug in QEMU/TCG that breaks that assumption.
>>
>> Link: https://gitlab.com/qemu-project/qemu/-/issues/928
>> Signed-off-by: Mathias Krause <minipli@grsecurity.net>
>> ---
>>  x86/emulator64.c | 26 ++++++++++++++++++++++++++
>>  1 file changed, 26 insertions(+)
>>
>> diff --git a/x86/emulator64.c b/x86/emulator64.c
>> index 5d1bb0f06d4f..abef2bda29f1 100644
>> --- a/x86/emulator64.c
>> +++ b/x86/emulator64.c
>> @@ -325,16 +325,39 @@ static void test_mmx_movq_mf(uint64_t *mem)
>>  	report(exception_vector() == MF_VECTOR, "movq mmx generates #MF");
>>  }
>>  
>> +#define CR2_REF_VALUE	0xdecafbadUL
>> +
>> +static void setup_cr2(void)
>> +{
>> +	write_cr2(CR2_REF_VALUE);
>> +}
>> +
>> +static void check_cr2(void)
>> +{
>> +	unsigned long cr2 = read_cr2();
>> +
>> +	if (cr2 == CR2_REF_VALUE) {
>> +		report(true, "CR2 unchanged");
>> +	} else {
>> +		report(false, "CR2 changed from %#lx to %#lx", CR2_REF_VALUE, cr2);
>> +		setup_cr2();
> 
> Writing CR2 isn't expensive in the grand scheme, so rather than conditionally
> re-write CR2, I think it makes sense to write CR2 at the start of every testcase,
> and then just do "report(cr2 == CR2_REF_VALUE".

Probably makes sense.

> 
>> +	}
>> +}
>> +
>>  static void test_jmp_noncanonical(uint64_t *mem)
>>  {
>> +	setup_cr2();
>>  	*mem = NONCANONICAL;
>>  	asm volatile (ASM_TRY("1f") "jmp *%0; 1:" : : "m"(*mem));
>>  	report(exception_vector() == GP_VECTOR,
>>  	       "jump to non-canonical address");
>> +	check_cr2();
>>  }
>>  
>>  static void test_reg_noncanonical(void)
>>  {
>> +	setup_cr2();
>> +
>>  	/* RAX based, should #GP(0) */
>>  	asm volatile(ASM_TRY("1f") "orq $0, (%[noncanonical]); 1:"
>>  		     : : [noncanonical]"a"(NONCANONICAL));
>> @@ -342,6 +365,7 @@ static void test_reg_noncanonical(void)
>>  	       "non-canonical memory access, should %s(0), got %s(%u)",
>>  	       exception_mnemonic(GP_VECTOR),
>>  	       exception_mnemonic(exception_vector()), exception_error_code());
>> +	check_cr2();
> 
> And then rather than add more copy+paste, what if we add a macro to handle the
> checks?  Then the CR2 validation can slot in nicely (and maybe someday the macro
> could be used outside of the x86/emulator64.c).
> 
> Attached patches yield:	
> 
> #define CR2_REF_VALUE	0xdecafbadUL
> 
> #define ASM_TRY_NONCANONICAL(insn, inputs, access, ex_vector)			\
> do {										\
> 	unsigned int vector, ec;						\
> 										\
> 	write_cr2(CR2_REF_VALUE);						\
> 										\
> 	asm volatile(ASM_TRY("1f") insn "; 1:" :: inputs);			\
> 										\
> 	vector = exception_vector();						\
> 	ec = exception_error_code();						\
> 										\
> 	report(vector == ex_vector && !ec,					\
> 	      "non-canonical " access ", should %s(0), got %s(%u)",		\
> 	      exception_mnemonic(ex_vector), exception_mnemonic(vector), ec);	\
> 										\
> 	if (vector != PF_VECTOR) {						\
> 		unsigned long cr2  = read_cr2();				\
> 										\
> 		report(cr2 == CR2_REF_VALUE,					\
> 		       "Wanted CR2 '0x%lx', got '0x%lx", CR2_REF_VALUE, cr2);	\
> 	}									\
> } while (0)
> 
> static void test_jmp_noncanonical(uint64_t *mem)
> {
> 	*mem = NONCANONICAL;
> 
> 	ASM_TRY_NONCANONICAL("jmp *%0", "m"(*mem), "jmp", GP_VECTOR);
> }
> 
> static void test_reg_noncanonical(void)
> {
> 	/* RAX based, should #GP(0) */
> 	ASM_TRY_NONCANONICAL("orq $0, (%[nc])", [nc]"a"(NONCANONICAL),
> 			     "memory access", GP_VECTOR);
> 
> 	/* RSP based, should #SS(0) */
> 	ASM_TRY_NONCANONICAL("orq $0, (%%rsp,%[nc],1)", [nc]"r"(NONCANONICAL),
> 			     "rsp-based access", SS_VECTOR);
> 
> 	/* RBP based, should #SS(0) */
> 	ASM_TRY_NONCANONICAL("orq $0, (%%rbp,%[nc],1)", [nc]"r"(NONCANONICAL),
> 			     "rbp-based access", SS_VECTOR);
> }

Yeah, looks much nicer indeed!

Both patches:

Reviewed-by: Mathias Krause <minipli@grsecurity.net>
Tested-by: Mathias Krause <minipli@grsecurity.net>

Thanks a lot,
Mathias

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [kvm-unit-tests PATCH] x86/emulator64: Extend non-canonical memory access tests with CR2 coverage
  2025-06-26 14:49 ` Sean Christopherson
  2025-06-27 14:33   ` Mathias Krause
@ 2025-11-18 22:39   ` Sean Christopherson
  1 sibling, 0 replies; 4+ messages in thread
From: Sean Christopherson @ 2025-11-18 22:39 UTC (permalink / raw)
  To: Mathias Krause; +Cc: Paolo Bonzini, kvm

On Thu, Jun 26, 2025, Sean Christopherson wrote:
> On Thu, Jun 12, 2025, Mathias Krause wrote:
> And then rather than add more copy+paste, what if we add a macro to handle the
> checks?  Then the CR2 validation can slot in nicely (and maybe someday the macro
> could be used outside of the x86/emulator64.c).
> 
> Attached patches yield:	

Applied said patches to kvm-x86 next.

[1/2] x86/emulator64: Add macro to test emulation of non-canonical accesses
      https://github.com/kvm-x86/kvm-unit-tests/commit/9f5f06b1
[2/2] x86/emulator64: Extend non-canonical memory access tests with CR2 coverage
      https://github.com/kvm-x86/kvm-unit-tests/commit/a55d5602

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2025-11-18 22:39 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-06-12 14:16 [kvm-unit-tests PATCH] x86/emulator64: Extend non-canonical memory access tests with CR2 coverage Mathias Krause
2025-06-26 14:49 ` Sean Christopherson
2025-06-27 14:33   ` Mathias Krause
2025-11-18 22:39   ` Sean Christopherson

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox