diff options
| author | Hugh Dickins <hugh@veritas.com> | 2008-11-19 15:36:37 -0800 |
|---|---|---|
| committer | Linus Torvalds <torvalds@linux-foundation.org> | 2008-11-19 18:49:59 -0800 |
| commit | 63eb6b93ce725e4c5f38fc85dd703d49465b03cb (patch) | |
| tree | 5370a3080b8f20ae27904a3c36646f79b26c6665 /mm | |
| parent | bda8550deed96687f29992d711a88ea21cff4d26 (diff) | |
| download | lwn-63eb6b93ce725e4c5f38fc85dd703d49465b03cb.tar.gz lwn-63eb6b93ce725e4c5f38fc85dd703d49465b03cb.zip | |
vmscan: let GFP_NOFS go to swap again
In the past, GFP_NOFS (but of course not GFP_NOIO) was allowed to reclaim
by writing to swap. That got partially broken in 2.6.23, when may_enter_fs
initialization was moved up before the allocation of swap, so its
PageSwapCache test was failing the first time around,
Fix it by setting may_enter_fs when add_to_swap() succeeds with
__GFP_IO. In fact, check __GFP_IO before calling add_to_swap():
allocating swap we're not ready to use just increases disk seeking.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Cc: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Diffstat (limited to 'mm')
| -rw-r--r-- | mm/vmscan.c | 3 |
1 files changed, 3 insertions, 0 deletions
diff --git a/mm/vmscan.c b/mm/vmscan.c index c141b3e78071..f83a7ed5c6c4 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -623,6 +623,8 @@ static unsigned long shrink_page_list(struct list_head *page_list, * Try to allocate it some swap space here. */ if (PageAnon(page) && !PageSwapCache(page)) { + if (!(sc->gfp_mask & __GFP_IO)) + goto keep_locked; switch (try_to_munlock(page)) { case SWAP_FAIL: /* shouldn't happen */ case SWAP_AGAIN: @@ -634,6 +636,7 @@ static unsigned long shrink_page_list(struct list_head *page_list, } if (!add_to_swap(page, GFP_ATOMIC)) goto activate_locked; + may_enter_fs = 1; } #endif /* CONFIG_SWAP */ |