Re: [PATCH 1/3] mm/slub: fix the race between validate_slab and slab_free

[Vlastimil Babka]

On 7/19/22 16:15, Rongwei Wang wrote:
> 
...
>> +
>> +    slab_unlock(slab, &flags2);
>> +    spin_unlock_irqrestore(&n->list_lock, flags);
>> +    if (!ret)
>> +        slab_fix(s, "Object at 0x%p not freed", object);
>> +    if (slab_to_discard) {
>> +        stat(s, FREE_SLAB);
>> +        discard_slab(s, slab);
>> +    }
>> +
>> +    return ret;
>> +}
> I had test this patch, and it indeed deal with the bug that I described.

Thanks.

> Though I am also has prepared this part of code, your code is ok to me.

Aha, feel free to post your version, maybe it's simpler? We can compare.