On 2022/12/26 16:08, lijiang wrote:
> Isn't this better? If no NR_SLAB_RECLAIMABLE, there is no
need to
> search for NR_SLAB_UNRECLAIMABLE.
>
>
Originally, I had the same idea as you. But later, I noticed that there
was too much duplication of code. So, eventually I used the current fix.
But anyway, if you would prefer the following change, It's also good to me.
ok, it doesn't look too redundant to me :) and the following
corresponds to the kernel change logically and can avoid an
unnecessary call on 5.9 and later.
will change and merge later.
Thanks,
Kazu
>
> Thanks
> Lianbo
>
> --- a/memory.c
>> +++ b/memory.c
>> @@ -8457,6 +8457,11 @@ dump_kmeminfo(void)
>> get_slabs = nr_slab;
>> if (dump_vm_stat("NR_SLAB_UNRECLAIMABLE",
>> &nr_slab, 0))
>> get_slabs += nr_slab;
>> + } else if (dump_vm_stat("NR_SLAB_RECLAIMABLE_B",
&nr_slab,
>> 0)) {
>> + /* 5.9 and later */
>> + get_slabs = nr_slab;
>> + if (dump_vm_stat("NR_SLAB_UNRECLAIMABLE_B",
>> &nr_slab, 0))
>> + get_slabs += nr_slab;
>> }
>> }
>>
>> Thanks,
>> Kazu
>>
>>