Linux Hardening
 help / color / mirror / Atom feed
From: "Christophe Leroy (CS GROUP)" <chleroy@kernel.org>
To: Xie Yuanbin <xieyuanbin1@huawei.com>,
	andriy.shevchenko@intel.com, maddy@linux.ibm.com,
	mpe@ellerman.id.au, npiggin@gmail.com, kees@kernel.org,
	andy@kernel.org
Cc: linuxppc-dev@lists.ozlabs.org, linux-kernel@vger.kernel.org,
	linux-hardening@vger.kernel.org, lilinjie8@huawei.com,
	liaohua4@huawei.com
Subject: Re: [PATCH 2/2] powerpc/text-patching: Fix possible stringop-overread compilation error
Date: Thu, 30 Apr 2026 11:41:49 +0200	[thread overview]
Message-ID: <c755a2d3-6447-4531-aea7-f279eea8f2f8@kernel.org> (raw)
In-Reply-To: <20260430072913.62348-1-xieyuanbin1@huawei.com>



Le 30/04/2026 à 09:28, Xie Yuanbin a écrit :
> Hi, Christophe Leroy!
> 
> On Mon, 9 Feb 2026 14:41:51 +0100, Christophe Leroy (CS GROUP) wrote:
>> On 09/02/2026 14:25, Xie Yuanbin wrote:
>>> When CONFIG_PPC64_ELF_ABI_V1=y, it seems that the try of lookupinp
>>> the original non-dot symbol is missing.
>>>
>>> What about this (Only the compilation test is performed):
>>> ```c
>>> static inline unsigned long __ppc_kallsyms_lookup_name(const char *name)
>>> {
>>> 	unsigned long addr = kallsyms_lookup_name(name);
>>>
>>> 	if (IS_ENABLED(CONFIG_PPC64_ELF_ABI_V2) && addr)
>>> 		addr = ppc_function_entry((void *)addr);
>>>
>>> 	return addr;
>>> }
>>>
>>> #define ppc_kallsyms_lookup_name(x) ({					\
>>> 		unsigned long addr = 0;					\
>>> 		if (IS_ENABLED(CONFIG_PPC64_ELF_ABI_V1))		\
>>> 			addr = __ppc_kallsyms_lookup_name("." x);	\
>>> 		if (!addr)						\
>>> 			addr = __ppc_kallsyms_lookup_name(x);		\
>>> 		addr;							\
>>> 	})
>>> ```
>>
>> Good point.
>>
>> To avoid duplicating the string I'd suggest:
>>
>> static inline unsigned long __ppc_kallsyms_lookup_name(const char *name)
>> {
>> 	unsigned long addr = kallsyms_lookup_name(name);
>>
>> 	if (IS_ENABLED(CONFIG_PPC64_ELF_ABI_V1) && !addr)
>> 		addr = kallsyms_lookup_name(name + 1);
>> 	if (IS_ENABLED(CONFIG_PPC64_ELF_ABI_V2) && addr)
>> 		addr = ppc_function_entry((void *)addr);
>>
>> 	return addr;
>> }
>>
>> #ifdef CONFIG_PPC64_ELF_ABI_V1
>> #define ppc_kallsyms_lookup_name(x)	__ppc_kallsyms_lookup_name("." x);
>> #else
>> #define ppc_kallsyms_lookup_name(x)	__ppc_kallsyms_lookup_name(x)
>> #endif
> 
> I would like to submit the modification as a new patch, and include you
> as Co-developer. Do you agree?

Add me as Suggested-by: instead.

Thanks
Christophe

> 
> Full of the patch:
> ```patch
>  From fbf07e5c1a97da7c8572435537f2b92213ede39d Mon Sep 17 00:00:00 2001
> From: Xie Yuanbin <xieyuanbin1@huawei.com>
> Date: Thu, 30 Apr 2026 14:15:26 +0800
> Subject: [PATCH V2] powerpc/text-patching: simplify the implementation of ppc_kallsyms_lookup_name()
> 
> ppc_kallsyms_lookup_name() is called only twice in the kernel code, and
> the parameters are all constant strings. strnlen(name, KSYM_NAME_LEN) is
> called inside ppc_kallsyms_lookup_name(), when the compiler detects that
> KSYM_NAME_LEN is larger then the constant strings,
> the following error will be triggered:
> ```log
>    CC      arch/powerpc/kernel/optprobes.o
> In file included from ./arch/powerpc/include/asm/kprobes.h:24,
>                   from ./include/linux/kprobes.h:31,
>                   from arch/powerpc/kernel/optprobes.c:8:
> In function ‘ppc_kallsyms_lookup_name’,
>      inlined from ‘arch_prepare_optimized_kprobe’ at arch/powerpc/kernel/optprobes.c:209:21:
> ./arch/powerpc/include/asm/text-patching.h:232:13: error: ‘strnlen’ specified bound 512 exceeds source size 19 [-Werror=stringop-overread]
>    232 |         if (strnlen(name, KSYM_NAME_LEN) >= KSYM_NAME_LEN)
>        |             ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
> In function ‘ppc_kallsyms_lookup_name’,
>      inlined from ‘arch_prepare_optimized_kprobe’ at arch/powerpc/kernel/optprobes.c:210:22:
> ./arch/powerpc/include/asm/text-patching.h:232:13: error: ‘strnlen’ specified bound 512 exceeds source size 13 [-Werror=stringop-overread]
>    232 |         if (strnlen(name, KSYM_NAME_LEN) >= KSYM_NAME_LEN)
>        |             ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
> cc1: all warnings being treated as errors
> ```
> 
> The error can be reproduced in the following ways:
> Use lastest linux-next source, change ppc_kallsyms_lookup_name() to
> __always_inline, use default ppc64_defconfig, set CONFIG_EXPERT=y,
> CONFIG_PPC64_BIG_ENDIAN_ELF_ABI_V2=n, CONFIG_CC_OPTIMIZE_FOR_SIZE=y,
> and use gcc-14 or a later version for compilation.
> 
> Since ppc_kallsyms_lookup_name() is called only twice in the kernel,
> and the parameters are all constant strins, simplify the implementation
> of ppc_kallsyms_lookup_name() and avoid calling strnlen().
> 
> Cc: Andy Shevchenko <andriy.shevchenko@intel.com>
> Cc: Kees Cook <kees@kernel.org>
> Co-developed-by: Christophe Leroy (CS GROUP) <chleroy@kernel.org>
> Signed-off-by: Christophe Leroy (CS GROUP) <chleroy@kernel.org>
> Signed-off-by: Xie Yuanbin <xieyuanbin1@huawei.com>
> ---
> v1->v2: https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Flore.kernel.org%2F20260205100517.292858-2-xieyuanbin1%40huawei.com&data=05%7C02%7Cchristophe.leroy%40csgroup.eu%7C525ecaba66ea4e3b15f208dea68a3fa1%7C8b87af7d86474dc78df45f69a2011bb5%7C0%7C0%7C639131309913493421%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=aaK9PO2mGdxM3vuJBVP0lcl6VKfqs3VKih5VPcG0IxI%3D&reserved=0
>    - Not use strlen()
> 
>   arch/powerpc/include/asm/text-patching.h | 42 ++++++++----------------
>   1 file changed, 17 insertions(+), 33 deletions(-)
> 
> diff --git a/arch/powerpc/include/asm/text-patching.h b/arch/powerpc/include/asm/text-patching.h
> index e7f14720f630..2d3f698cb4f1 100644
> --- a/arch/powerpc/include/asm/text-patching.h
> +++ b/arch/powerpc/include/asm/text-patching.h
> @@ -221,39 +221,23 @@ static inline unsigned long ppc_global_function_entry(void *func)
>    * - For ABIv1, we lookup the dot variant.
>    * - For ABIv2, we return the local entry point.
>    */
> -static inline unsigned long ppc_kallsyms_lookup_name(const char *name)
> -{
> -	unsigned long addr;
> -#ifdef CONFIG_PPC64_ELF_ABI_V1
> -	/* check for dot variant */
> -	char dot_name[1 + KSYM_NAME_LEN];
> -	bool dot_appended = false;
> -
> -	if (strnlen(name, KSYM_NAME_LEN) >= KSYM_NAME_LEN)
> -		return 0;
> -
> -	if (name[0] != '.') {
> -		dot_name[0] = '.';
> -		dot_name[1] = '\0';
> -		strlcat(dot_name, name, sizeof(dot_name));
> -		dot_appended = true;
> -	} else {
> -		dot_name[0] = '\0';
> -		strlcat(dot_name, name, sizeof(dot_name));
> -	}
> -	addr = kallsyms_lookup_name(dot_name);
> -	if (!addr && dot_appended)
> -		/* Let's try the original non-dot symbol lookup	*/
> -		addr = kallsyms_lookup_name(name);
> -#elif defined(CONFIG_PPC64_ELF_ABI_V2)
> -	addr = kallsyms_lookup_name(name);
> -	if (addr)
> -		addr = ppc_function_entry((void *)addr);
> -#else
> -	addr = kallsyms_lookup_name(name);
> -#endif
> -	return addr;
> -}
> +static inline unsigned long __ppc_kallsyms_lookup_name(const char *name)
> +{
> +	unsigned long addr = kallsyms_lookup_name(name);
> +
> +	if (IS_ENABLED(CONFIG_PPC64_ELF_ABI_V1) && !addr)
> +		addr = kallsyms_lookup_name(name + 1);
> +	if (IS_ENABLED(CONFIG_PPC64_ELF_ABI_V2) && addr)
> +		addr = ppc_function_entry((void *)addr);
> +
> +	return addr;
> +}
> +
> +#ifdef CONFIG_PPC64_ELF_ABI_V1
> +#define ppc_kallsyms_lookup_name(x)	__ppc_kallsyms_lookup_name("." x)
> +#else
> +#define ppc_kallsyms_lookup_name(x)	__ppc_kallsyms_lookup_name(x)
> +#endif
>   
>   /*
>    * Some instruction encodings commonly used in dynamic ftracing


      reply	other threads:[~2026-04-30  9:41 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-02-05 10:05 [PATCH 1/2] string: move __compiletime_strlen() to string.h Xie Yuanbin
2026-02-05 10:05 ` [PATCH 2/2] powerpc/text-patching: Fix possible stringop-overread compilation error Xie Yuanbin
2026-02-05 16:40   ` Andy Shevchenko
2026-02-06 11:14     ` Xie Yuanbin
2026-02-06 18:26   ` Kees Cook
2026-02-06 19:53     ` Christophe Leroy (CS GROUP)
2026-02-09 13:25       ` Xie Yuanbin
2026-02-09 13:41         ` Christophe Leroy (CS GROUP)
2026-02-09 14:11           ` Xie Yuanbin
2026-04-30  7:28           ` Xie Yuanbin
2026-04-30  9:41             ` Christophe Leroy (CS GROUP) [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=c755a2d3-6447-4531-aea7-f279eea8f2f8@kernel.org \
    --to=chleroy@kernel.org \
    --cc=andriy.shevchenko@intel.com \
    --cc=andy@kernel.org \
    --cc=kees@kernel.org \
    --cc=liaohua4@huawei.com \
    --cc=lilinjie8@huawei.com \
    --cc=linux-hardening@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linuxppc-dev@lists.ozlabs.org \
    --cc=maddy@linux.ibm.com \
    --cc=mpe@ellerman.id.au \
    --cc=npiggin@gmail.com \
    --cc=xieyuanbin1@huawei.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox