linux-kernel - Re: [PATCH] ata: sata_mv: Fix incorrect string length computation in mv_dump

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: <2b36ae3d-de9b-4125-a0e3-e1083a67cf4a@wanadoo.fr>
Date:   Fri, 8 Sep 2023 07:18:58 +0200
From:   Christophe JAILLET <christophe.jaillet@...adoo.fr>
To:     Damien Le Moal <dlemoal@...nel.org>
Cc:     linux-kernel@...r.kernel.org, kernel-janitors@...r.kernel.org,
        linux-ide@...r.kernel.org
Subject: Re: [PATCH] ata: sata_mv: Fix incorrect string length computation in
 mv_dump_mem()

Le 06/09/2023 à 03:11, Damien Le Moal a écrit :
> On 9/6/23 00:28, Christophe JAILLET wrote:
>>
>>
>> Le 05/09/2023 à 07:04, Damien Le Moal a écrit :
>>> On 9/5/23 04:54, Christophe JAILLET wrote:
>>>> snprintf() returns the "number of characters which *would* be generated for
>>>> the given input", not the size *really* generated.
>>>>
>>>> In order to avoid too large values for 'o' (and potential negative values
>>>> for "sizeof(linebuf) o") use scnprintf() instead of snprintf().
>>>>
>>>> Note that given the "w < 4" in the for loop, the buffer can NOT
>>>> overflow, but using the *right* function is always better.
>>>>
>>>> Signed-off-by: Christophe JAILLET <christophe.jaillet@...adoo.fr>
>>>
>>> Doesn't this need Fixes and CC stable tags ?
>>
>> I don't think so.
>> As said in the commit message :
>>      Note that given the "w < 4" in the for loop, the buffer can NOT
>>      overflow, but using the *right* function is always better.
>>
>> linebuf is 38 chars.
>> In each iteration, we write 9 bytes + NULL.
>> We write only 4 elements per line (because of w < 4), so 9 * 4 + 1 = 37
>> bytes are needed.
>> 9 is for %08x<space>
>>
>> It can't overflow.
>> Moreover, it is really unlikely that the size of linebuf or the number
>> of elements on each line change in a stable kernel.
>>
>> So, from my POV, this patch is more a clean-up than anything else.
>>
>> I would even agree that it is maybe not even needed. But should someone
>> cut'n'paste it one day, then using the correct function could maybe help
>> him.
> 
> OK. Fine. But then maybe the patch title should be something like "Improve
> string length computation in mv_dump_mem()" as the "Fix" word you used seems to
> be somewhat misleading. With the patch title as is, the stable bot will likely
> pick up that patch for stable. Fine with me, unless you see an issue with that.

Hi,

up to you to tweak it if desired.

My POV is that it *fixes* the length calculation, but having it "wrong" 
is harmless.
Should it trigger a backport, it wouldn't be a real issue either. And we 
can still ask to remove it from the backport list when notified.

And as said, leaving things as-is looks also fine to me.

I let you choose the best option.

CJ

> 
>>
>> CJ
>>
>>>
>>>> ---
>>>>    drivers/ata/sata_mv.c | 4 ++--
>>>>    1 file changed, 2 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/drivers/ata/sata_mv.c b/drivers/ata/sata_mv.c
>>>> index d105db5c7d81..45e48d653c60 100644
>>>> --- a/drivers/ata/sata_mv.c
>>>> +++ b/drivers/ata/sata_mv.c
>>>> @@ -1255,8 +1255,8 @@ static void mv_dump_mem(struct device *dev, void __iomem *start, unsigned bytes)
>>>>    
>>>>    	for (b = 0; b < bytes; ) {
>>>>    		for (w = 0, o = 0; b < bytes && w < 4; w++) {
>>>> -			o += snprintf(linebuf + o, sizeof(linebuf) - o,
>>>> -				      "%08x ", readl(start + b));
>>>> +			o += scnprintf(linebuf + o, sizeof(linebuf) - o,
>>>> +				       "%08x ", readl(start + b));
>>>>    			b += sizeof(u32);
>>>>    		}
>>>>    		dev_dbg(dev, "%s: %p: %s\n",
>>>
>