Re: [PATCH 05/18] MM: reclaim mustn't enter FS for SWP_FS_OPS swap-space
From: Christoph Hellwig <hch@infradead.org>
Date: 2021-12-21 08:43:24
Also in:
linux-mm, lkml
On Fri, Dec 17, 2021 at 10:48:22AM +1100, NeilBrown wrote:
+static bool test_may_enter_fs(struct page *page, gfp_t gfp_mask)
+{
+ if (gfp_mask & __GFP_FS)
+ return true;
+ if (!PageSwapCache(page) || !(gfp_mask & __GFP_IO))
+ return false;
+ /* We can "enter_fs" for swap-cache with only __GFP_IO
+ * providing this isn't SWP_FS_OPS.
+ * ->flags can be updated non-atomicially (scan_swap_map_slots),
+ * but that will never affect SWP_FS_OPS, so the data_race
+ * is safe.
+ */
+ return !data_race(page_swap_info(page)->flags & SWP_FS_OPS);Nit: the normal kernel comment style uses an empty /* line to start the comment.
quoted hunk ↗ jump to hunk
@@ -1514,8 +1529,7 @@ static unsigned int shrink_page_list(struct list_head *page_list, if (!sc->may_unmap && page_mapped(page)) goto keep_locked; - may_enter_fs = (sc->gfp_mask & __GFP_FS) || - (PageSwapCache(page) && (sc->gfp_mask & __GFP_IO)); + may_enter_fs = test_may_enter_fs(page, sc->gfp_mask); /* * The number of dirty pages determines if a node is marked@@ -1683,7 +1697,8 @@ static unsigned int shrink_page_list(struct list_head *page_list, goto activate_locked_split; } - may_enter_fs = true; + may_enter_fs = test_may_enter_fs(page, + sc->gfp_mask);
Now that may_enter_fs is always reset using test_may_enter_fs, do we even need the local variable?