linux-kernel - Re: [PATCH 2/3] mm/vmalloc: do not call kmemleak

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]

Message-ID: <b761b165-9892-0761-cd33-14300e39e36f@virtuozzo.com>
Date:   Fri, 11 Jan 2019 22:26:39 +0300
From:   Andrey Ryabinin <aryabinin@...tuozzo.com>
To:     Roman Penyaev <rpenyaev@...e.de>
Cc:     Andrew Morton <akpm@...ux-foundation.org>,
        Michal Hocko <mhocko@...e.com>, Joe Perches <joe@...ches.com>,
        "Luis R. Rodriguez" <mcgrof@...nel.org>, linux-mm@...ck.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH 2/3] mm/vmalloc: do not call kmemleak_free() on not yet
 accounted memory



On 1/3/19 5:59 PM, Roman Penyaev wrote:
> __vmalloc_area_node() calls vfree() on error path, which in turn calls
> kmemleak_free(), but area is not yet accounted by kmemleak_vmalloc().
> 
> Signed-off-by: Roman Penyaev <rpenyaev@...e.de>
> Cc: Andrew Morton <akpm@...ux-foundation.org>
> Cc: Michal Hocko <mhocko@...e.com>
> Cc: Andrey Ryabinin <aryabinin@...tuozzo.com>
> Cc: Joe Perches <joe@...ches.com>
> Cc: "Luis R. Rodriguez" <mcgrof@...nel.org>
> Cc: linux-mm@...ck.org
> Cc: linux-kernel@...r.kernel.org
> ---
>  mm/vmalloc.c | 16 +++++++++++-----
>  1 file changed, 11 insertions(+), 5 deletions(-)
> 
> diff --git a/mm/vmalloc.c b/mm/vmalloc.c
> index 2cd24186ba84..dc6a62bca503 100644
> --- a/mm/vmalloc.c
> +++ b/mm/vmalloc.c
> @@ -1565,6 +1565,14 @@ void vfree_atomic(const void *addr)
>  	__vfree_deferred(addr);
>  }
>  
> +static void __vfree(const void *addr)
> +{
> +	if (unlikely(in_interrupt()))
> +		__vfree_deferred(addr);
> +	else
> +		__vunmap(addr, 1);
> +}
> +
>  /**
>   *	vfree  -  release memory allocated by vmalloc()
>   *	@addr:		memory base address
> @@ -1591,10 +1599,8 @@ void vfree(const void *addr)
>  
>  	if (!addr)
>  		return;
> -	if (unlikely(in_interrupt()))
> -		__vfree_deferred(addr);
> -	else
> -		__vunmap(addr, 1);
> +
> +	__vfree(addr);
>  }
>  EXPORT_SYMBOL(vfree);
>  
> @@ -1709,7 +1715,7 @@ static void *__vmalloc_area_node(struct vm_struct *area, gfp_t gfp_mask,
>  	warn_alloc(gfp_mask, NULL,
>  			  "vmalloc: allocation failure, allocated %ld of %ld bytes",
>  			  (area->nr_pages*PAGE_SIZE), area->size);
> -	vfree(area->addr);
> +	__vfree(area->addr);

This can't be an interrupt context for a several reasons. One of them is BUG_ON(in_interrupt()) in __get_vm_area_node()
which is called right before __vmalloc_are_node().

So you can just do __vunmap(area->addr, 1); instead of __vfree().


>  	return NULL;
>  }
>  
>