On Fri 08-05-20 14:31:03, Johannes Weiner wrote:
[...]
> diff --git a/mm/memory.c b/mm/memory.c
> index 832ee914cbcf..93900b121b6e 100644
> --- a/mm/memory.c
> +++ b/mm/memory.c
> @@ -3125,9 +3125,20 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
> page = alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma,
> vmf->address);
> if (page) {
> + int err;
> +
> __SetPageLocked(page);
> __SetPageSwapBacked(page);
> set_page_private(page, entry.val);
> +
> + /* Tell memcg to use swap ownership records */
> + SetPageSwapCache(page);
> + err = mem_cgroup_charge(page, vma->vm_mm,
> + GFP_KERNEL, false);
> + ClearPageSwapCache(page);
> + if (err)
> + goto out_page;
err would be a return value from try_charge and that can be -ENOMEM. Now
we almost never return ENOMEM for GFP_KERNEL single page charge. Except
for async OOM handling (oom_disabled v1). So this needs translation to
VM_FAULT_OOM.
I am not an expert on the swap code so I might have missed some subtle
issues but the rest of the patch seems reasonable to me.
--
Michal Hocko
SUSE Labs
