On Thu, Jan 18, 2018 at 1:31 PM, Laura Abbott <labbott@xxxxxxxxxx> wrote:
> On 01/10/2018 06:02 PM, Kees Cook wrote:
>>
>> From: David Windsor <dave@xxxxxxxxxxxx>
>>
>> The autoclose field can be copied with put_user(), so there is no need to
>> use copy_to_user(). In both cases, hardened usercopy is being bypassed
>> since the size is constant, and not open to runtime manipulation.
>>
>> This patch is verbatim from Brad Spengler/PaX Team's PAX_USERCOPY
>> whitelisting code in the last public patch of grsecurity/PaX based on my
>> understanding of the code. Changes or omissions from the original code are
>> mine and don't reflect the original grsecurity/PaX code.
>>
>
> Just tried a quick rebase and it looks like this conflicts with
> c76f97c99ae6 ("sctp: make use of pre-calculated len")
> I don't think we can use put_user if we're copying via the full
> len?
It should be fine, since:
len = sizeof(int);
c76f97c99ae6 just does a swap of sizeof(int) with len, put_user() will
work in either case, since autoclose will always be int sized.
-Kees
>
> Thanks,
> Laura
>
>
>> Signed-off-by: David Windsor <dave@xxxxxxxxxxxx>
>> [kees: adjust commit log]
>> Cc: Vlad Yasevich <vyasevich@xxxxxxxxx>
>> Cc: Neil Horman <nhorman@xxxxxxxxxxxxx>
>> Cc: "David S. Miller" <davem@xxxxxxxxxxxxx>
>> Cc: linux-sctp@xxxxxxxxxxxxxxx
>> Cc: netdev@xxxxxxxxxxxxxxx
>> Signed-off-by: Kees Cook <keescook@xxxxxxxxxxxx>
>> ---
>> net/sctp/socket.c | 2 +-
>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/net/sctp/socket.c b/net/sctp/socket.c
>> index efbc8f52c531..15491491ec88 100644
>> --- a/net/sctp/socket.c
>> +++ b/net/sctp/socket.c
>> @@ -5011,7 +5011,7 @@ static int sctp_getsockopt_autoclose(struct sock
>> *sk, int len, char __user *optv
>> len = sizeof(int);
>> if (put_user(len, optlen))
>> return -EFAULT;
>> - if (copy_to_user(optval, &sctp_sk(sk)->autoclose, sizeof(int)))
>> + if (put_user(sctp_sk(sk)->autoclose, (int __user *)optval))
>> return -EFAULT;
>> return 0;
>> }
>>
>
--
Kees Cook
Pixel Security
