On 05/29/2020 03:23 AM, Andrew Morton wrote:
> On Wed, 27 May 2020 10:25:18 +0800 Bibo Mao <maobibo@xxxxxxxxxxx> wrote:
>
>> If two threads concurrently fault at the same page, the thread that
>> won the race updates the PTE and its local TLB. For now, the other
>> thread gives up, simply does nothing, and continues.
>>
>> It could happen that this second thread triggers another fault, whereby
>> it only updates its local TLB while handling the fault. Instead of
>> triggering another fault, let's directly update the local TLB of the
>> second thread. Function update_mmu_tlb is used here to update local
>> TLB on the second thread, and it is defined as empty on other arches.
>>
>> --- a/mm/memory.c
>> +++ b/mm/memory.c
>> @@ -2752,6 +2752,7 @@ static vm_fault_t wp_page_copy(struct vm_fault *vmf)
>> new_page = old_page;
>> page_copied = 1;
>> } else {
>> + update_mmu_tlb(vma, vmf->address, vmf->pte);
>> mem_cgroup_cancel_charge(new_page, memcg, false);
>> }
>>
>
> When applying your patches on top of the -mm tree's changes, the above
> hunk didn't apply. The entire `else' block was removed by
> https://ozlabs.org/~akpm/mmotm/broken-out/mm-memcontrol-convert-anon-and-file-thp-to-new-mem_cgroup_charge-api.patch
>
> I assumed that dropping this hunk was appropriate. Please check?
yes, that is appropriate. Sorry to bother you, originally I should format the
patch based on mm-tree.
