arm64/crash_core: Export KERNELPACMASK in vmcoreinfo
Subsystems (5)
PATCH v2 1/2 arm64/crash_core: Export KERNELPACMASK in vmcoreinfo
Recently arm64 linux kernel added support for Armv8.3-A Pointer Authentication feature. If this feature is enabled in the kernel and the hardware supports address authentication then the return addresses are signed and stored in the stack to prevent ROP kind of attack. Kdump tool will now dump the kernel with signed lr values in the stack.
Any user analysis tool for this kernel dump may need the kernel pac mask information in vmcoreinfo to generate the correct return address for stacktrace purpose as well as to resolve the symbol name.
This patch is similar to commit ec6e822d1a22d0eef ("arm64: expose user PAC bit positions via ptrace") which exposes pac mask information via ptrace interfaces.
---
Changes since v1:
* Rebased to kernel 5.7-rc3.
* commit log change.
An implementation of this new KERNELPACMASK vmcoreinfo field used by crash tool can be found here[1]. This change is accepted by crash utility maintainer [2].
[1]: https://www.redhat.com/archives/crash-utility/2020-April/msg00095.html
[2]: https://www.redhat.com/archives/crash-utility/2020-April/msg00099.html
arch/arm64/include/asm/compiler.h | 3 +++ arch/arm64/kernel/crash_core.c | 4 ++++ 2 files changed, 7 insertions(+)
arch/arm64/include/asm/compiler.h +3 -0
@@ -19,6 +19,9 @@#define __builtin_return_address (val) \ (void *)(ptrauth_clear_pac((unsigned long)__builtin_return_address(val))) +#else /* !CONFIG_ARM64_PTR_AUTH */ +#define ptrauth_user_pac_mask () 0ULL +#define ptrauth_kernel_pac_mask () 0ULL#endif /* CONFIG_ARM64_PTR_AUTH */ #endif /* __ASM_COMPILER_H */
arch/arm64/kernel/crash_core.c +4 -0
@@ -6,6 +6,7 @@#include <linux/crash_core.h> #include <asm/memory.h> +#include <asm/pointer_auth.h> void arch_crash_save_vmcoreinfo (void ) {
@@ -16,4 +17,7 @@ void arch_crash_save_vmcoreinfo(void)vmcoreinfo_append_str ("NUMBER(PHYS_OFFSET)=0x%llx\n" ,PHYS_OFFSET );vmcoreinfo_append_str ("KERNELOFFSET=%lx\n" ,kaslr_offset ()); +vmcoreinfo_append_str ("NUMBER(KERNELPACMASK)=0x%llx\n" , +system_supports_address_auth () ? +ptrauth_kernel_pac_mask () :0 ); }
Signature · 7 lines
2.7.4 _______________________________________________ kexec mailing list [email protected] http://lists.infradead.org/mailman/listinfo/kexec
PATCH v2 2/2 Documentation/vmcoreinfo: Add documentation for 'KERNELPACMASK'
Add documentation for KERNELPACMASK variable being added to the vmcoreinfo.
It indicates the PAC bits mask information of signed kernel pointers if Armv8.3-A Pointer Authentication feature is present.
--- Documentation/admin-guide/kdump/vmcoreinfo.rst | 6 ++++++ 1 file changed, 6 insertions(+)
Documentation/admin-guide/kdump/vmcoreinfo.rst +6 -0
@@ -393,6 +393,12 @@ KERNELOFFSET The kernel randomization offset. Used to compute the page offset. If KASLR is disabled, this value is zero. +KERNELPACMASK +------------- + +Indicates the PAC bits mask information if Pointer Authentication is +enabled and address authentication feature is present. + arm ===
Signature · 7 lines
2.7.4 _______________________________________________ kexec mailing list [email protected] http://lists.infradead.org/mailman/listinfo/kexec
Will Deacon
maintainerRe: [PATCH v2 2/2] Documentation/vmcoreinfo: Add documentation for 'KERNELPACMASK'
Quoted 20 lines · Amit Daniel Kachhap
Add documentation for KERNELPACMASK variable being added to the vmcoreinfo.
It indicates the PAC bits mask information of signed kernel pointers if Armv8.3-A Pointer Authentication feature is present.
Cc Catalin Marinas <[email protected]>Cc Will Deacon <[email protected]>Cc Mark Rutland <[email protected]>Cc Dave Young <[email protected]>Cc Baoquan He <[email protected]>Signed-off-by Amit Daniel Kachhap <[email protected]>--- Documentation/admin-guide/kdump/vmcoreinfo.rst | 6 ++++++ 1 file changed, 6 insertions(+)Documentation/admin-guide/kdump/vmcoreinfo.rst +0
-0
Documentation/admin-guide/kdump/vmcoreinfo.rst KERNELOFFSET · line 400
Quoted 8 lines
Documentation/admin-guide/kdump/vmcoreinfo.rst +5
-0@@ -393,6 +393,12 @@ KERNELOFFSET The kernel randomization offset. Used to compute the page offset. If KASLR is disabled, this value is zero. +KERNELPACMASK +------------- + +Indicates the PAC bits mask information if Pointer Authentication is +enabled and address authentication feature is present.
This is a bit cryptic. How about:
The mask to extract the Pointer Authentication Code from a kernel virtual address.
Will
_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec
Re: [PATCH v2 2/2] Documentation/vmcoreinfo: Add documentation for 'KERNELPACMASK'
Hi,
Quoted 32 lines · Will Deacon
Quoted 20 lines · Amit Daniel Kachhap
Add documentation for KERNELPACMASK variable being added to the vmcoreinfo.
It indicates the PAC bits mask information of signed kernel pointers if Armv8.3-A Pointer Authentication feature is present.
Cc Catalin Marinas <[email protected]>Cc Will Deacon <[email protected]>Cc Mark Rutland <[email protected]>Cc Dave Young <[email protected]>Cc Baoquan He <[email protected]>Signed-off-by Amit Daniel Kachhap <[email protected]>--- Documentation/admin-guide/kdump/vmcoreinfo.rst | 6 ++++++ 1 file changed, 6 insertions(+)Documentation/admin-guide/kdump/vmcoreinfo.rst +0
-0Documentation/admin-guide/kdump/vmcoreinfo.rst KERNELOFFSET · line 400
Quoted 8 lines
Documentation/admin-guide/kdump/vmcoreinfo.rst +5
-0@@ -393,6 +393,12 @@ KERNELOFFSET The kernel randomization offset. Used to compute the page offset. If KASLR is disabled, this value is zero. +KERNELPACMASK +------------- + +Indicates the PAC bits mask information if Pointer Authentication is +enabled and address authentication feature is present.This is a bit cryptic. How about:
Quoted 2 lines
The mask to extract the Pointer Authentication Code from a kernel virtual address.
Ok sure. I will update like this in the v3 version.
Cheers,
Amit Daniel
Quoted 3 lines
Will
_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec
Hi Will/Catalin,
Quoted 10 lines · Amit Daniel Kachhap
Recently arm64 linux kernel added support for Armv8.3-A Pointer Authentication feature. If this feature is enabled in the kernel and the hardware supports address authentication then the return addresses are signed and stored in the stack to prevent ROP kind of attack. Kdump tool will now dump the kernel with signed lr values in the stack.
Any user analysis tool for this kernel dump may need the kernel pac mask information in vmcoreinfo to generate the correct return address for stacktrace purpose as well as to resolve the symbol name.
Quoted 8 lines
This patch is similar to commit
ec6e822d1a22d0eef("arm64: expose user PAC bit positions via ptrace") which exposes pac mask information via ptrace interfaces.Cc Catalin Marinas <[email protected]>Cc Will Deacon <[email protected]>Cc Mark Rutland <[email protected]>Signed-off-by Amit Daniel Kachhap <[email protected]>
This patch user side changes are accepted by crash-utility maintainer [1] so I think this is in a good shape to go in.
Thanks,
Amit Daniel
[1]: https://www.redhat.com/archives/crash-utility/2020-April/msg00099.html
Quoted 44 lines
---
Changes since v1:
* Rebased to kernel 5.7-rc3.
* commit log change.An implementation of this new KERNELPACMASK vmcoreinfo field used by crash tool can be found here[1]. This change is accepted by crash utility maintainer [2].
[1]: https://www.redhat.com/archives/crash-utility/2020-April/msg00095.html
[2]: https://www.redhat.com/archives/crash-utility/2020-April/msg00099.htmlarch/arm64/include/asm/compiler.h | 3 +++ arch/arm64/kernel/crash_core.c | 4 ++++ 2 files changed, 7 insertions(+)arch/arm64/include/asm/compiler.h +3
-0@@ -19,6 +19,9 @@#define __builtin_return_address (val) \ (void *)(ptrauth_clear_pac((unsigned long)__builtin_return_address(val))) +#else /* !CONFIG_ARM64_PTR_AUTH */ +#define ptrauth_user_pac_mask () 0ULL +#define ptrauth_kernel_pac_mask () 0ULL#endif /* CONFIG_ARM64_PTR_AUTH */ #endif /* __ASM_COMPILER_H */ arch/arm64/kernel/crash_core.c +1
-0@@ -6,6 +6,7 @@#include <linux/crash_core.h> #include <asm/memory.h> +#include <asm/pointer_auth.h> void arch_crash_save_vmcoreinfo (void ) {
arch/arm64/kernel/crash_core.c void arch_crash_save_vmcoreinfo(void) · line 23
Quoted 7 lines
arch/arm64/kernel/crash_core.c +3
-0@@ -16,4 +17,7 @@ void arch_crash_save_vmcoreinfo(void)vmcoreinfo_append_str ("NUMBER(PHYS_OFFSET)=0x%llx\n" ,PHYS_OFFSET );vmcoreinfo_append_str ("KERNELOFFSET=%lx\n" ,kaslr_offset ()); +vmcoreinfo_append_str ("NUMBER(KERNELPACMASK)=0x%llx\n" , +system_supports_address_auth () ? +ptrauth_kernel_pac_mask () :0 ); }
_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec
Hi Will/Catalin,
Sorry: Resending with correct To list.
Quoted 6 lines · Amit Daniel Kachhap
Recently arm64 linux kernel added support for Armv8.3-A Pointer Authentication feature. If this feature is enabled in the kernel and the hardware supports address authentication then the return addresses are signed and stored in the stack to prevent ROP kind of attack. Kdump tool will now dump the kernel with signed lr values in the stack.
Quoted 7 lines
Any user analysis tool for this kernel dump may need the kernel pac mask information in vmcoreinfo to generate the correct return address for stacktrace purpose as well as to resolve the symbol name.
This patch is similar to commit
ec6e822d1a22d0eef("arm64: expose user PAC bit positions via ptrace") which exposes pac mask information via ptrace interfaces.
This patch user side changes are accepted by crash-utility maintainer [1] so I think this is in a good shape to go in.
Thanks,
Amit Daniel
[1]: https://www.redhat.com/archives/crash-utility/2020-April/msg00095.html
Quoted 49 lines
Cc Catalin Marinas <[email protected]>Cc Will Deacon <[email protected]>Cc Mark Rutland <[email protected]>Signed-off-by Amit Daniel Kachhap <[email protected]>---
Changes since v1:
* Rebased to kernel 5.7-rc3.
* commit log change.An implementation of this new KERNELPACMASK vmcoreinfo field used by crash tool can be found here[1]. This change is accepted by crash utility maintainer [2].
[1]: https://www.redhat.com/archives/crash-utility/2020-April/msg00095.html
[2]: https://www.redhat.com/archives/crash-utility/2020-April/msg00099.htmlarch/arm64/include/asm/compiler.h | 3 +++ arch/arm64/kernel/crash_core.c | 4 ++++ 2 files changed, 7 insertions(+)arch/arm64/include/asm/compiler.h +3
-0@@ -19,6 +19,9 @@#define __builtin_return_address (val) \ (void *)(ptrauth_clear_pac((unsigned long)__builtin_return_address(val))) +#else /* !CONFIG_ARM64_PTR_AUTH */ +#define ptrauth_user_pac_mask () 0ULL +#define ptrauth_kernel_pac_mask () 0ULL#endif /* CONFIG_ARM64_PTR_AUTH */ #endif /* __ASM_COMPILER_H */ arch/arm64/kernel/crash_core.c +1
-0@@ -6,6 +6,7 @@#include <linux/crash_core.h> #include <asm/memory.h> +#include <asm/pointer_auth.h> void arch_crash_save_vmcoreinfo (void ) {
arch/arm64/kernel/crash_core.c void arch_crash_save_vmcoreinfo(void) · line 23
Quoted 7 lines
arch/arm64/kernel/crash_core.c +3
-0@@ -16,4 +17,7 @@ void arch_crash_save_vmcoreinfo(void)vmcoreinfo_append_str ("NUMBER(PHYS_OFFSET)=0x%llx\n" ,PHYS_OFFSET );vmcoreinfo_append_str ("KERNELOFFSET=%lx\n" ,kaslr_offset ()); +vmcoreinfo_append_str ("NUMBER(KERNELPACMASK)=0x%llx\n" , +system_supports_address_auth () ? +ptrauth_kernel_pac_mask () :0 ); }
_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec
Will Deacon
maintainerQuoted 39 lines · Amit Daniel Kachhap
Recently arm64 linux kernel added support for Armv8.3-A Pointer Authentication feature. If this feature is enabled in the kernel and the hardware supports address authentication then the return addresses are signed and stored in the stack to prevent ROP kind of attack. Kdump tool will now dump the kernel with signed lr values in the stack.
Any user analysis tool for this kernel dump may need the kernel pac mask information in vmcoreinfo to generate the correct return address for stacktrace purpose as well as to resolve the symbol name.
This patch is similar to commit
ec6e822d1a22d0eef("arm64: expose user PAC bit positions via ptrace") which exposes pac mask information via ptrace interfaces.Cc Catalin Marinas <[email protected]>Cc Will Deacon <[email protected]>Cc Mark Rutland <[email protected]>Signed-off-by Amit Daniel Kachhap <[email protected]>---
Changes since v1:
* Rebased to kernel 5.7-rc3.
* commit log change.An implementation of this new KERNELPACMASK vmcoreinfo field used by crash tool can be found here[1]. This change is accepted by crash utility maintainer [2].
[1]: https://www.redhat.com/archives/crash-utility/2020-April/msg00095.html
[2]: https://www.redhat.com/archives/crash-utility/2020-April/msg00099.htmlarch/arm64/include/asm/compiler.h | 3 +++ arch/arm64/kernel/crash_core.c | 4 ++++ 2 files changed, 7 insertions(+)arch/arm64/include/asm/compiler.h +0
-0
arch/arm64/include/asm/compiler.h line 24
Quoted 6 lines
arch/arm64/include/asm/compiler.h +3
-0@@ -19,6 +19,9 @@#define __builtin_return_address (val) \ (void *)(ptrauth_clear_pac((unsigned long)__builtin_return_address(val))) +#else /* !CONFIG_ARM64_PTR_AUTH */ +#define ptrauth_user_pac_mask () 0ULL +#define ptrauth_kernel_pac_mask () 0ULL
This doesn't look quite right to me, since you still have to take into account the case where CONFIG_ARM64_PTR_AUTH=y but the feature is not available at runtime:
arch/arm64/include/asm/compiler.h void arch_crash_save_vmcoreinfo(void) · line 22
Quoted 7 lines
arch/arm64/include/asm/compiler.h +3
-0@@ -16,4 +17,7 @@ void arch_crash_save_vmcoreinfo(void)vmcoreinfo_append_str ("NUMBER(PHYS_OFFSET)=0x%llx\n" ,PHYS_OFFSET );vmcoreinfo_append_str ("KERNELOFFSET=%lx\n" ,kaslr_offset ()); +vmcoreinfo_append_str ("NUMBER(KERNELPACMASK)=0x%llx\n" , +system_supports_address_auth () ? +ptrauth_kernel_pac_mask () :0 );
In which case, would it make more sense to define ptrauth_{kernel,user}_pac_mask() unconditionally? In fact, I'd probably just remove the guards completely from asm/compiler.h because I think they're misleading.
Will
--->8
arch/arm64/include/asm/compiler.h +0 -4
@@ -2,8 +2,6 @@#ifndef __ASM_COMPILER_H#define __ASM_COMPILER_H -#if defined(CONFIG_ARM64_PTR_AUTH ) -/* * The EL0/EL1 pointer bits used by a pointer authentication code. * This is dependent on TBI0/TBI1 being enabled, or bits 63:56 would also apply.
@@ -19,6 +17,4 @@#define __builtin_return_address (val) \ (void *)(ptrauth_clear_pac((unsigned long)__builtin_return_address(val))) -#endif /* CONFIG_ARM64_PTR_AUTH */ -#endif /* __ASM_COMPILER_H */
_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec
Hi Will,
Quoted 43 lines · Will Deacon
Quoted 42 lines · Amit Daniel Kachhap
Recently arm64 linux kernel added support for Armv8.3-A Pointer Authentication feature. If this feature is enabled in the kernel and the hardware supports address authentication then the return addresses are signed and stored in the stack to prevent ROP kind of attack. Kdump tool will now dump the kernel with signed lr values in the stack.
Any user analysis tool for this kernel dump may need the kernel pac mask information in vmcoreinfo to generate the correct return address for stacktrace purpose as well as to resolve the symbol name.
This patch is similar to commit
ec6e822d1a22d0eef("arm64: expose user PAC bit positions via ptrace") which exposes pac mask information via ptrace interfaces.Cc Catalin Marinas <[email protected]>Cc Will Deacon <[email protected]>Cc Mark Rutland <[email protected]>Signed-off-by Amit Daniel Kachhap <[email protected]>---
Changes since v1:
* Rebased to kernel 5.7-rc3.
* commit log change.An implementation of this new KERNELPACMASK vmcoreinfo field used by crash tool can be found here[1]. This change is accepted by crash utility maintainer [2].
[1]: https://www.redhat.com/archives/crash-utility/2020-April/msg00095.html
[2]: https://www.redhat.com/archives/crash-utility/2020-April/msg00099.htmlarch/arm64/include/asm/compiler.h | 3 +++ arch/arm64/kernel/crash_core.c | 4 ++++ 2 files changed, 7 insertions(+)arch/arm64/include/asm/compiler.h +0
-0@@ -19,6 +19,9 @@#define __builtin_return_address (val) \ (void *)(ptrauth_clear_pac((unsigned long)__builtin_return_address(val)))
Quoted 7 lines
arch/arm64/include/asm/compiler.h line 24
Quoted 3 lines
arch/arm64/include/asm/compiler.h +3
-0+#else /* !CONFIG_ARM64_PTR_AUTH */ +#define ptrauth_user_pac_mask () 0ULL +#define ptrauth_kernel_pac_mask () 0ULLThis doesn't look quite right to me, since you still have to take into account the case where
CONFIG_ARM64_PTR_AUTH=ybut the feature is not available at runtime:
Yes agree with you here. However the config gaurd is saving some extra computation for __builtin_return_address. There are some compiler support being added in __builtin_extract_return_address to mask the PAC. Hopefully that will improve this code. In the meantime let it be like this.
I can remove this else case and as other users of ptrauth_{kernel,user}_pac_mask(ptrace.c) protect it with a config gaurd there.
Quoted 13 lines
arch/arm64/include/asm/compiler.h void arch_crash_save_vmcoreinfo(void) · line 22
Quoted 8 lines
arch/arm64/include/asm/compiler.h +3
-0@@ -16,4 +17,7 @@ void arch_crash_save_vmcoreinfo(void)vmcoreinfo_append_str ("NUMBER(PHYS_OFFSET)=0x%llx\n" ,PHYS_OFFSET );vmcoreinfo_append_str ("KERNELOFFSET=%lx\n" ,kaslr_offset ()); +vmcoreinfo_append_str ("NUMBER(KERNELPACMASK)=0x%llx\n" , +system_supports_address_auth () ? +ptrauth_kernel_pac_mask () :0 );In which case, would it make more sense to define ptrauth_{kernel,user}_pac_mask() unconditionally? In fact, I'd probably just remove the guards completely from
asm/compiler.hbecause I think they're misleading.
As answered above. Let me know your opinion. Although I don't have strong reservation in keeping the config gaurd.
Thanks,
Amit Daniel
Quoted 20 lines
Will
--->8
arch/arm64/include/asm/compiler.h +0
-2@@ -2,8 +2,6 @@#ifndef __ASM_COMPILER_H#define __ASM_COMPILER_H -#if defined(CONFIG_ARM64_PTR_AUTH ) -/* * The EL0/EL1 pointer bits used by a pointer authentication code. * This is dependent on TBI0/TBI1 being enabled, or bits 63:56 would also apply.
arch/arm64/include/asm/compiler.h line 20
Quoted 6 lines
arch/arm64/include/asm/compiler.h +0
-2@@ -19,6 +17,4 @@#define __builtin_return_address (val) \ (void *)(ptrauth_clear_pac((unsigned long)__builtin_return_address(val))) -#endif /* CONFIG_ARM64_PTR_AUTH */ -#endif /* __ASM_COMPILER_H */
_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec
Will Deacon
maintainerQuoted 17 lines · Amit Kachhap
Quoted 15 lines · Will Deacon
arch/arm64/include/asm/compiler.h line 23
Quoted 10 lines · Amit Daniel Kachhap
arch/arm64/include/asm/compiler.h +3
-0@@ -19,6 +19,9 @@#define __builtin_return_address (val) \ (void *)(ptrauth_clear_pac((unsigned long)__builtin_return_address(val))) +#else /* !CONFIG_ARM64_PTR_AUTH */ +#define ptrauth_user_pac_mask () 0ULL +#define ptrauth_kernel_pac_mask () 0ULLThis doesn't look quite right to me, since you still have to take into account the case where
CONFIG_ARM64_PTR_AUTH=ybut the feature is not available at runtime:
Quoted 4 lines
Yes agree with you here. However the config gaurd is saving some extra computation for
__builtin_return_address. There are some compiler support being added in__builtin_extract_return_addressto mask the PAC. Hopefully that will improve this code. In the meantime let it be like this.
Does the extra computation matter? Isn't it just a couple of instructions?
Will
_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec
Hi,
Quoted 18 lines · Will Deacon
Quoted 17 lines · Amit Kachhap
Quoted 15 lines · Will Deacon
arch/arm64/include/asm/compiler.h line 23
Quoted 10 lines · Amit Daniel Kachhap
arch/arm64/include/asm/compiler.h +3
-0@@ -19,6 +19,9 @@#define __builtin_return_address (val) \ (void *)(ptrauth_clear_pac((unsigned long)__builtin_return_address(val))) +#else /* !CONFIG_ARM64_PTR_AUTH */ +#define ptrauth_user_pac_mask () 0ULL +#define ptrauth_kernel_pac_mask () 0ULLThis doesn't look quite right to me, since you still have to take into account the case where
CONFIG_ARM64_PTR_AUTH=ybut the feature is not available at runtime:
Quoted 6 lines
Quoted 4 lines
Yes agree with you here. However the config gaurd is saving some extra computation for
__builtin_return_address. There are some compiler support being added in__builtin_extract_return_addressto mask the PAC. Hopefully that will improve this code. In the meantime let it be like this.Does the extra computation matter? Isn't it just a couple of instructions?
ok sure. I will push v3 as you suggested.
Thanks,
Amit
Quoted 3 lines
Will
_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec