Re: [PATCH] RDMA/hfi: add a judgment on the availability of cpumask

[xiaolinkui <xiaolinkui@126.com>]

Thanks for your reply.

When CONFIG_CPUMASK_OFFSTACK=y, "ret" will be false if diff==NULL.

However, when CONFIG_CPUMASK_OFFSTACK=n, these two are not necessarily 
equivalent.

Thanks

On 4/4/23 14:05, Leon Romanovsky wrote:
> On Tue, Apr 04, 2023 at 11:05:25AM +0800, xiaolinkui wrote:
>> From: Linkui Xiao <xiaolinkui@kylinos.cn>
>>
>> When CONFIG_CPUMASK_OFFSTACK is n, cpumask may fail to allocate, cpumask may
>> be NULL, and performing a bitmap operation on cpumask may cause problems at
>> this time.
>>
>> Of course, this is a unlikely event.
>>
>> Signed-off-by: Linkui Xiao <xiaolinkui@kylinos.cn>
>> ---
>>   drivers/infiniband/hw/hfi1/affinity.c | 8 ++++----
>>   1 file changed, 4 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/infiniband/hw/hfi1/affinity.c b/drivers/infiniband/hw/hfi1/affinity.c
>> index 77ee77d4000f..3caa861f4d1d 100644
>> --- a/drivers/infiniband/hw/hfi1/affinity.c
>> +++ b/drivers/infiniband/hw/hfi1/affinity.c
>> @@ -1047,16 +1047,16 @@ int hfi1_get_proc_affinity(int node)
>>   	 */
>>   
>>   	ret = zalloc_cpumask_var(&diff, GFP_KERNEL);
>> -	if (!ret)
>> +	if (!ret || unlikely(!diff))
> Why do you think that check of "ret" is not enough?
> "ret" will be false if diff == NULL.
>
> Thanks