Re: [PATCH V2] powerpc/kasan: require memintrinsic prefix support for KASAN

["Christophe Leroy (CS GROUP)" <chleroy@kernel.org>]

Hi Mukesh,

Le 11/09/2026 à 20:05, Mukesh Kumar Chaurasiya (IBM) a écrit :
> powerpc unconditionally selects GENERIC_ENTRY. The GENERIC_ENTRY
> infrastructure relies on the compiler emitting __asan_mem*() calls at
> instrumented mem*() sites rather than plain memset/memcpy/memmove, so
> that entry/exit paths calling those functions are not instrumented.
> 
> This assumption is encoded in two places:
> 
>    mm/kasan/shadow.c:
>      #if !defined(CONFIG_CC_HAS_KASAN_MEMINTRINSIC_PREFIX) &&
>          !defined(CONFIG_GENERIC_ENTRY)
> 
>    include/linux/fortify-string.h:
>      #if !defined(CONFIG_CC_HAS_KASAN_MEMINTRINSIC_PREFIX) &&
>          !defined(CONFIG_GENERIC_ENTRY)
> 
> When GENERIC_ENTRY is set, both guards suppress the C wrappers for
> memset/memcpy/memmove and the __underlying_mem*() redirections. This
> is only safe when the compiler supports the prefixed __asan_mem*()
> intrinsics. On older toolchains (e.g. GCC 9) that lack this support,
> plain mem*() calls from instrumented code fall through to the raw
> assembly implementations in mem_64.S / copy_32.S, completely bypassing
> the KASAN shadow check.
> 
> Other arches with GENERIC_ENTRY (x86, s390, loongarch, riscv) do not
> hit this because their CI toolchains are always new enough to support
> the prefix flag.
> 
> Background: the !GENERIC_ENTRY guard was introduced by commit 69d4c0d32186
> ("entry, kasan, x86: Disallow overriding mem*() functions", Peter Zijlstra,
> Jan 2023). The root problem is that the KASAN C wrappers override the
> linker symbol memset/memcpy/memmove globally, so any call from noinstr or
> __no_sanitize_address code (e.g. irqentry_enter/irqentry_exit) would still
> reach the KASAN shadow-check wrapper -- at a point where KASAN invariants
> may not hold. The compiler prefix approach (Marco Elver, Feb 2023,
> commit 51287dcb00cc) solves this by having the compiler emit __asan_memset
> at instrumented call sites and bare memset inside __no_sanitize_address
> functions, splitting the decision at code-generation time rather than at
> link time.
> 
> A manual C-level override cannot replicate this split: a single linker
> symbol cannot be made to resolve differently depending on the caller.
> 
> x86 also placed its raw memset/memcpy/memmove implementations in
> .noinstr.text (same commit, 69d4c0d32186), which is the other half of
> the fix: noinstr callers hit the raw assembly directly, safely bypassing
> KASAN. PowerPC has not done this. Placing mem_64.S / memcpy_64.S /
> copy_32.S implementations in .noinstr.text would be the complementary
> long-term fix that could re-enable KASAN on older toolchains, but it
> requires care around linker stub overflow on large PPC64 kernels (the
> same reason powerpc uses NOKPROBE_SYMBOL rather than noinstr for its
> interrupt handlers -- see the comment in asm/interrupt.h). That work
> is left as a follow-up.
> 
> For now, introduce PPC_CC_HAS_KASAN_MEMINTRINSIC_PREFIX, an arch-local
> compiler probe that mirrors the same check as CC_HAS_KASAN_MEMINTRINSIC_PREFIX
> in lib/Kconfig.kasan but lives outside the 'if KASAN' block to avoid a
> recursive dependency (CC_HAS_KASAN_MEMINTRINSIC_PREFIX depends on KASAN
> which depends on HAVE_ARCH_KASAN). Gate the three HAVE_ARCH_KASAN selects
> on this new symbol so that KASAN is not offered as a config option on
> toolchains that cannot support it correctly with GENERIC_ENTRY.
> 
> Since KASAN on powerpc now unconditionally implies
> CC_HAS_KASAN_MEMINTRINSIC_PREFIX, the old !CC_HAS_KASAN_MEMINTRINSIC_PREFIX
> code paths in asm/kasan.h and asm/string.h are dead. Clean them up:
> 
> - asm/kasan.h: remove the dual-entry-point variant of _GLOBAL_KASAN /
>    _GLOBAL_TOC_KASAN / EXPORT_SYMBOL_KASAN that emitted both memset and
>    __memset as entry points to the same assembly. These aliases were only
>    needed so the C KASAN wrappers in shadow.c could call __memset() to
>    reach raw memory ops; with the compiler prefix approach those wrappers
>    are not used for mem* on powerpc.
> 
> - asm/string.h: remove the separate __memset/__memcpy/__memmove symbol
>    declarations and the memset/memcpy/memmove macro redirections for
>    uninstrumented files that were needed on old toolchains. Simplify the
>    CONFIG_KASAN block to just the three #define aliases (still used by
>    shadow.c as raw backends to bypass KASAN checking).
> 
> - cputable.c, prom_init.c: the original comments said "use memcpy() so
>    GCC emits __memcpy() under KASAN". The real reason is that these run
>    pre-relocation: the destination pointer is PTRRELOC-adjusted to its
>    current physical address, and the kernel is loaded at a different
>    address than it was linked at. A struct assignment (*t = *s) may cause
>    the compiler to emit an implicit memcpy() call that resolves through
>    the unrelocated virtual symbol address -- before the MMU mapping is set
>    up -- jumping to garbage. An explicit memcpy(t, ...) uses the already-
>    corrected pointer and is safe. Update the comments accordingly.

Nope, see my other email.

> 
> Reported-by: Venkat Rao Bagalkote <venkat88@linux.ibm.com>
> Closes: https://lore.kernel.org/all/96dae110-f79b-4e54-8a9b-514369019b5d@linux.ibm.com
> Tested-by: Venkat Rao Bagalkote <venkat88@linux.ibm.com>
> Signed-off-by: Mukesh Kumar Chaurasiya (IBM) <mkchauras@gmail.com>
> ---
> Changelog:
> 
> V1 -> V2:
> 	- Comments reworded
> 	- Commit message reworded
> V1: https://lore.kernel.org/all/20260908064948.999530-1-mkchauras@gmail.com
> 
>   arch/powerpc/Kconfig              | 10 +++++++---
>   arch/powerpc/include/asm/kasan.h  | 11 -----------
>   arch/powerpc/include/asm/string.h | 21 +--------------------
>   arch/powerpc/kernel/cputable.c    | 14 +++++++++++---
>   arch/powerpc/kernel/prom_init.c   | 10 ++++++----
>   5 files changed, 25 insertions(+), 41 deletions(-)
> 
> diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig
> index 2580e27e4328..b27ed9739eea 100644
> --- a/arch/powerpc/Kconfig
> +++ b/arch/powerpc/Kconfig
> @@ -7,6 +7,10 @@ config CC_HAS_ELFV2
>   config CC_HAS_PREFIXED
>   	def_bool PPC64 && $(cc-option, -mcpu=power10 -mprefixed)
>   
> +config PPC_CC_HAS_KASAN_MEMINTRINSIC_PREFIX
> +	def_bool (CC_IS_CLANG && $(cc-option,-fsanitize=kernel-address -mllvm -asan-kernel-mem-intrinsic-prefix=1)) || \
> +		 (CC_IS_GCC && $(cc-option,-fsanitize=kernel-address --param asan-kernel-mem-intrinsic-prefix=1))
> +
>   config CC_HAS_PCREL
>   	# Clang has a bug (https://github.com/llvm/llvm-project/issues/62372)
>   	# where pcrel code is not generated if -msoft-float, -mno-altivec, or
> @@ -220,9 +224,9 @@ config PPC
>   	select HAVE_ARCH_HUGE_VMAP		if PPC_RADIX_MMU || PPC_8xx
>   	select HAVE_ARCH_JUMP_LABEL
>   	select HAVE_ARCH_JUMP_LABEL_RELATIVE
> -	select HAVE_ARCH_KASAN			if PPC32 && PAGE_SHIFT <= 14
> -	select HAVE_ARCH_KASAN			if PPC_RADIX_MMU
> -	select HAVE_ARCH_KASAN			if PPC_BOOK3E_64
> +	select HAVE_ARCH_KASAN			if PPC32 && PAGE_SHIFT <= 14 && PPC_CC_HAS_KASAN_MEMINTRINSIC_PREFIX
> +	select HAVE_ARCH_KASAN			if PPC_RADIX_MMU && PPC_CC_HAS_KASAN_MEMINTRINSIC_PREFIX
> +	select HAVE_ARCH_KASAN			if PPC_BOOK3E_64 && PPC_CC_HAS_KASAN_MEMINTRINSIC_PREFIX
>   	select HAVE_ARCH_KASAN_VMALLOC		if HAVE_ARCH_KASAN
>   	select HAVE_ARCH_KCSAN
>   	select HAVE_ARCH_KFENCE			if ARCH_SUPPORTS_DEBUG_PAGEALLOC
> diff --git a/arch/powerpc/include/asm/kasan.h b/arch/powerpc/include/asm/kasan.h
> index a690e7da53c2..d62756b87ba4 100644
> --- a/arch/powerpc/include/asm/kasan.h
> +++ b/arch/powerpc/include/asm/kasan.h
> @@ -2,20 +2,9 @@
>   #ifndef __ASM_KASAN_H
>   #define __ASM_KASAN_H
>   
> -#if defined(CONFIG_KASAN) && !defined(CONFIG_CC_HAS_KASAN_MEMINTRINSIC_PREFIX)
> -#define _GLOBAL_KASAN(fn)			\
> -	_GLOBAL(fn);				\
> -	_GLOBAL(__##fn)
> -#define _GLOBAL_TOC_KASAN(fn)			\
> -	_GLOBAL_TOC(fn);			\
> -	_GLOBAL_TOC(__##fn)
> -#define EXPORT_SYMBOL_KASAN(fn)			\
> -	EXPORT_SYMBOL(__##fn)
> -#else /* CONFIG_KASAN && !CONFIG_CC_HAS_KASAN_MEMINTRINSIC_PREFIX */
>   #define _GLOBAL_KASAN(fn)	_GLOBAL(fn)
>   #define _GLOBAL_TOC_KASAN(fn)	_GLOBAL_TOC(fn)
>   #define EXPORT_SYMBOL_KASAN(fn)
> -#endif /* CONFIG_KASAN && !CONFIG_CC_HAS_KASAN_MEMINTRINSIC_PREFIX */
>   
>   #ifndef __ASSEMBLER__
>   
> diff --git a/arch/powerpc/include/asm/string.h b/arch/powerpc/include/asm/string.h
> index 1981bd4036b5..72b5c93a2b84 100644
> --- a/arch/powerpc/include/asm/string.h
> +++ b/arch/powerpc/include/asm/string.h
> @@ -29,29 +29,10 @@ extern void * memchr(const void *,int,__kernel_size_t);
>   void memcpy_flushcache(void *dest, const void *src, size_t size);
>   
>   #ifdef CONFIG_KASAN
> -/* __mem variants are used by KASAN to implement instrumented meminstrinsics. */
> -#ifdef CONFIG_CC_HAS_KASAN_MEMINTRINSIC_PREFIX
> +/* Used by mm/kasan/shadow.c as raw backends to bypass KASAN checking. */
>   #define __memset memset
>   #define __memcpy memcpy
>   #define __memmove memmove
> -#else /* CONFIG_CC_HAS_KASAN_MEMINTRINSIC_PREFIX */
> -void *__memset(void *s, int c, __kernel_size_t count);
> -void *__memcpy(void *to, const void *from, __kernel_size_t n);
> -void *__memmove(void *to, const void *from, __kernel_size_t n);
> -#ifndef __SANITIZE_ADDRESS__
> -/*
> - * For files that are not instrumented (e.g. mm/slub.c) we
> - * should use not instrumented version of mem* functions.
> - */
> -#define memcpy(dst, src, len) __memcpy(dst, src, len)
> -#define memmove(dst, src, len) __memmove(dst, src, len)
> -#define memset(s, c, n) __memset(s, c, n)
> -
> -#ifndef __NO_FORTIFY
> -#define __NO_FORTIFY /* FORTIFY_SOURCE uses __builtin_memcpy, etc. */
> -#endif
> -#endif /* !__SANITIZE_ADDRESS__ */
> -#endif /* CONFIG_CC_HAS_KASAN_MEMINTRINSIC_PREFIX */
>   #endif /* CONFIG_KASAN */
>   
>   #ifdef CONFIG_PPC64
> diff --git a/arch/powerpc/kernel/cputable.c b/arch/powerpc/kernel/cputable.c
> index 6f6801da9dc1..233b5c650d1e 100644
> --- a/arch/powerpc/kernel/cputable.c
> +++ b/arch/powerpc/kernel/cputable.c
> @@ -36,8 +36,12 @@ void __init set_cur_cpu_spec(struct cpu_spec *s)
>   
>   	t = PTRRELOC(t);
>   	/*
> -	 * use memcpy() instead of *t = *s so that GCC replaces it
> -	 * by __memcpy() when KASAN is active
> +	 * Use memcpy() instead of *t = *s because t is a PTRRELOC-adjusted
> +	 * pointer and this code runs before the MMU mapping is established.
> +	 * A struct assignment is a compiler-generated aggregate copy whose
> +	 * implementation is not under our control in relocation-sensitive code;
> +	 * memcpy() ensures the adjusted pointer is explicitly passed to the
> +	 * copy routine.

This comment is wrong, see my response to your explanation. I think the 
comment should be removed completely, now it is possible to use *t = *s 
without any issue again.


>   	 */
>   	memcpy(t, s, sizeof(*t));
>   
> @@ -55,7 +59,11 @@ static struct cpu_spec * __init setup_cpu_spec(unsigned long offset,
>   
>   	/*
>   	 * Copy everything, then do fixups. Use memcpy() instead of *t = *s
> -	 * so that GCC replaces it by __memcpy() when KASAN is active
> +	 * because t is a PTRRELOC-adjusted pointer and this code runs before
> +	 * the MMU mapping is established. A struct assignment is a
> +	 * compiler-generated aggregate copy whose implementation is not under
> +	 * our control in relocation-sensitive code; memcpy() ensures the
> +	 * adjusted pointer is explicitly passed to the copy routine.

Same

>   	 */
>   	memcpy(t, s, sizeof(*t));
>   
> diff --git a/arch/powerpc/kernel/prom_init.c b/arch/powerpc/kernel/prom_init.c
> index eb9f556b0937..d6d7f1ede319 100644
> --- a/arch/powerpc/kernel/prom_init.c
> +++ b/arch/powerpc/kernel/prom_init.c
> @@ -1363,10 +1363,12 @@ static void __init prom_check_platform_support(void)
>   				       "ibm,arch-vec-5-platform-support");
>   
>   	/*
> -	 * First copy the architecture vec template
> -	 *
> -	 * use memcpy() instead of *vec = *vec_template so that GCC replaces it
> -	 * by __memcpy() when KASAN is active
> +	 * First copy the architecture vec template. Use memcpy() instead of
> +	 * a struct assignment because this code runs before the MMU mapping
> +	 * is established. A struct assignment is a compiler-generated
> +	 * aggregate copy whose implementation is not under our control in
> +	 * relocation-sensitive code; memcpy() ensures the adjusted pointer
> +	 * is explicitly passed to the copy routine.

Same.

>   	 */
>   	memcpy(&ibm_architecture_vec, &ibm_architecture_vec_template,
>   	       sizeof(ibm_architecture_vec));