Re: [PATCH] cfq: fix NULL pointer dereference in cfq_choose_cfqg.

[Namjae Jeon]

Hi. Richard.

You're right. need to add exception code more. I will look for more.

Thanks.

2012/3/4 richard -rw- weinberger <richard.weinberger@xxxxxxxxx>:
> On Sun, Mar 4, 2012 at 1:10 PM, Namjae Jeon <linkinjeon@xxxxxxxxx> wrote:
>> I found null pointer dereferencing possibility while reviewing.
>> If cfq_get_next_cfqg return null by RB_EMPTY_ROOT, null pointer dereferencing error can occur.
>>
>> Signed-off-by: Namjae Jeon <linkinjeon@xxxxxxxxx>
>> ---
>> Âblock/cfq-iosched.c | Â Â2 +-
>> Â1 files changed, 1 insertions(+), 1 deletions(-)
>>
>> diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c
>> index d0ba505..d924272 100644
>> --- a/block/cfq-iosched.c
>> +++ b/block/cfq-iosched.c
>> @@ -2306,7 +2306,7 @@ static void cfq_choose_cfqg(struct cfq_data *cfqd)
>> Â Â Â Âcfqd->serving_group = cfqg;
>>
>> Â Â Â Â/* Restore the workload type data */
>> - Â Â Â if (cfqg->saved_workload_slice) {
>> + Â Â Â if (!cfqg && cfqg->saved_workload_slice) {
>> Â Â Â Â Â Â Â Âcfqd->workload_expires = jiffies + cfqg->saved_workload_slice;
>> Â Â Â Â Â Â Â Âcfqd->serving_type = cfqg->saved_workload;
>> Â Â Â Â Â Â Â Âcfqd->serving_prio = cfqg->saved_serving_prio;
>
> I don't think that this is a correct fix.
>
> 1. Can it really happen that cfq_get_next_cfqg() returns NULL at this point?
>
> 2. Adding a if(!cfgg ...) does not solve the problem, it ignores it.
> After the if block we call choose_service_tree().
> I'm sure choose_service_tree() it not happy when we pass a NULL pointer to it...
>
> --
> Thanks,
> //richard
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/