netdev - Re: [Patch v2] vmalloc: check CONFIG_EXECMEM in is_vmalloc_or_module

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date: Tue, 28 May 2024 19:22:30 +0300
From: Mike Rapoport <rppt@...nel.org>
To: Cong Wang <xiyou.wangcong@...il.com>
Cc: netdev@...r.kernel.org, linux-mm@...ck.org, akpm@...ux-foundation.org,
	bpf@...r.kernel.org, Cong Wang <cong.wang@...edance.com>,
	Luis Chamberlain <mcgrof@...nel.org>
Subject: Re: [Patch v2] vmalloc: check CONFIG_EXECMEM in
 is_vmalloc_or_module_addr()

On Tue, May 28, 2024 at 09:08:38AM -0700, Cong Wang wrote:
> From: Cong Wang <cong.wang@...edance.com>
> 
> After commit 2c9e5d4a0082 ("bpf: remove CONFIG_BPF_JIT dependency on CONFIG_MODULES of")
> CONFIG_BPF_JIT does not depend on CONFIG_MODULES any more and bpf jit
> also uses the [MODULES_VADDR, MODULES_END] memory region. But
> is_vmalloc_or_module_addr() still checks CONFIG_MODULES, which then
> returns false for a bpf jit memory region when CONFIG_MODULES is not
> defined. It leads to the following kernel BUG:

...

> Fix this by checking CONFIG_EXECMEM instead.
> 
> Fixes: 2c9e5d4a0082 ("bpf: remove CONFIG_BPF_JIT dependency on CONFIG_MODULES of")
> Cc: Luis Chamberlain <mcgrof@...nel.org>
> Cc: Mike Rapoport (IBM) <rppt@...nel.org>
> Signed-off-by: Cong Wang <cong.wang@...edance.com>

Acked-by: Mike Rapoport (IBM) <rppt@...nel.org>

> ---
>  mm/vmalloc.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/mm/vmalloc.c b/mm/vmalloc.c
> index 6641be0ca80b..94e1d2dbdec0 100644
> --- a/mm/vmalloc.c
> +++ b/mm/vmalloc.c
> @@ -722,7 +722,7 @@ int is_vmalloc_or_module_addr(const void *x)
>  	 * and fall back on vmalloc() if that fails. Others
>  	 * just put it in the vmalloc space.
>  	 */
> -#if defined(CONFIG_MODULES) && defined(MODULES_VADDR)
> +#if defined(CONFIG_EXECMEM) && defined(MODULES_VADDR)
>  	unsigned long addr = (unsigned long)kasan_reset_tag(x);
>  	if (addr >= MODULES_VADDR && addr < MODULES_END)
>  		return 1;
> -- 
> 2.34.1
> 
> 

-- 
Sincerely yours,
Mike.