Re: [PATCH 3/9] userfaultfd/shmem: support minor fault registration for shmem
From: Peter Xu <peterx@redhat.com>
Date: 2021-04-09 16:50:34
Also in:
linux-fsdevel, linux-kselftest, linux-mm, lkml
On Thu, Apr 08, 2021 at 04:43:21PM -0700, Axel Rasmussen wrote: [...]
quoted hunk ↗ jump to hunk
@@ -1820,16 +1821,27 @@ static int shmem_getpage_gfp(struct inode *inode, pgoff_t index, page = pagecache_get_page(mapping, index, FGP_ENTRY | FGP_HEAD | FGP_LOCK, 0); - if (xa_is_value(page)) { + swapped = xa_is_value(page); + if (swapped) { error = shmem_swapin_page(inode, index, &page, sgp, gfp, vma, fault_type); if (error == -EEXIST) goto repeat; - *pagep = page; - return error; + if (error) + return error; } + if (page && vma && userfaultfd_minor(vma)) { + unlock_page(page); + put_page(page); + *fault_type = handle_userfault(vmf, VM_UFFD_MINOR); + return 0; + }
If we need to consider swapping for UFFDIO_CONTINUE later (as Hugh pointed out previously, which looks the right thing to do), it's indeed a bit awkward to swapin here. Maybe move this chunk to right after pagecache_get_page() returns? Then no need to touch the rest.
+ + if (swapped) + return 0; + if (page) hindex = page->index; if (page && sgp == SGP_WRITE) -- 2.31.1.295.g9ea45b61b8-goog
-- Peter Xu