Re: [PATCH] crypto: inside-secure: safexcel - fix memory allocation

[Kees Cook]

On Wed, Oct 17, 2018 at 7:41 AM, Gustavo A. R. Silva
<gustavo@xxxxxxxxxxxxxx> wrote:
>
>
> On 10/17/18 9:20 AM, Antoine Tenart wrote:
>> Hi,
>>
>> On Wed, Oct 17, 2018 at 02:17:41PM +0800, Herbert Xu wrote:
>>> On Tue, Oct 16, 2018 at 09:44:02PM +0200, Gustavo A. R. Silva wrote:
>>>> On 10/9/18 12:20 AM, Kees Cook wrote:
>>>>> On Mon, Oct 8, 2018 at 12:17 PM, Gustavo A. R. Silva
>>>>> <gustavo@xxxxxxxxxxxxxx> wrote:
>>>>>> The original intention is to allocate space for EIP197_DEFAULT_RING_SIZE
>>>>>> *pointers* to struct, so sizeof(priv->ring[i].rdr_req) should be
>>>>>> sizeof(*priv->ring[i].rdr_req).
>>>>>>
>>>>>> Addresses-Coverity-ID: 1473962 ("Sizeof not portable")
>>>>>> Fixes: 9744fec95f06 ("crypto: inside-secure - remove request list to improve performance")
>>>>>> Signed-off-by: Gustavo A. R. Silva <gustavo@xxxxxxxxxxxxxx>
>>>>>
>>>>> Reviewed-by: Kees Cook <keescook@xxxxxxxxxxxx>
>>>>>
>>>>
>>>>
>>>> Friendly ping. Who can take this?
>>>
>>> Well I tried to take it but it doesn't apply against cryptodev.
>>> So I presume this can go into the tree that carried the change
>>> which it depended on?
>>
>> I would say this should go in cryptodev. The issue is probably because
>> of other changes that got applied in the meantime. Gustavo can probably
>> rebase his patch on top of cryptodev, and re-send it.
>>
>
> cryptodev is missing the previous commit 329e09893909d409039f6a79757d9b80b67efe39
> to which this patch applies.
>
> Kees, did you apply the commit above to your tree?
>
> If so, could you take this patch?

Since this has no functional exposure (the sizes are the same), let's
just wait until after the merge window to get this into crypto-next.

-Kees

-- 
Kees Cook
Pixel Security