linux-kernel - Re: [PATCH] kfence: fix implicit function declaration

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

Date:   Fri, 4 Dec 2020 13:38:25 +0100
From:   Marco Elver <elver@...gle.com>
To:     Anders Roxell <anders.roxell@...aro.org>
Cc:     akpm@...ux-foundation.org, glider@...gle.com, dvyukov@...gle.com,
        catalin.marinas@....com, will@...nel.org,
        kasan-dev@...glegroups.com, linux-arm-kernel@...ts.infradead.org,
        linux-kernel@...r.kernel.org, rppt@...nel.org, david@...hat.com
Subject: Re: [PATCH] kfence: fix implicit function declaration

On Fri, Dec 04, 2020 at 01:18PM +0100, Anders Roxell wrote:
> When building kfence the following error shows up:
> 
> In file included from mm/kfence/report.c:13:
> arch/arm64/include/asm/kfence.h: In function ‘kfence_protect_page’:
> arch/arm64/include/asm/kfence.h:12:2: error: implicit declaration of function ‘set_memory_valid’ [-Werror=implicit-function-declaration]
>    12 |  set_memory_valid(addr, 1, !protect);
>       |  ^~~~~~~~~~~~~~~~
> 
> Use the correct include both
> f2b7c491916d ("set_memory: allow querying whether set_direct_map_*() is actually enabled")
> and 4c4c75881536 ("arm64, kfence: enable KFENCE for ARM64") went in the

Note that -mm does not have stable commit hashes.

> same day via different trees.
> 
> Signed-off-by: Anders Roxell <anders.roxell@...aro.org>
> ---

Ack, we need this patch somewhere but we should probably fix the patch
that does the move, otherwise we'll have a build-broken kernel still.

> I got this build error in todays next-20201204.
> Andrew, since both patches are in your -mm tree, I think this can be
> folded into 4c4c75881536 ("arm64, kfence: enable KFENCE for ARM64")

I don't think that's the right way around. This would result in a
build-broken commit point as well.

Looking at current -next, I see that "set_memory: allow querying whether
set_direct_map_*() is actually enabled" is after "arm64, kfence: enable
KFENCE for ARM64".

I think the patch that introduces set_memory.h for arm64 simply needs to
squash in this patch (assuming the order is retained as-is in -mm).

Thanks,
-- Marco

>  arch/arm64/include/asm/kfence.h | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/arch/arm64/include/asm/kfence.h b/arch/arm64/include/asm/kfence.h
> index 6c0afeeab635..c44bb368a810 100644
> --- a/arch/arm64/include/asm/kfence.h
> +++ b/arch/arm64/include/asm/kfence.h
> @@ -3,7 +3,7 @@
>  #ifndef __ASM_KFENCE_H
>  #define __ASM_KFENCE_H
>  
> -#include <asm/cacheflush.h>
> +#include <asm/set_memory.h>
>  
>  static inline bool arch_kfence_init_pool(void) { return true; }
>