| Message ID | 20220115144150.1195590-1-shakeelb@google.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | mpage: remove ineffective __GFP_HIGH flag | expand |
FYI, I have a patch removing mpage_alloc entirely as part of a bio allocation refactoring series about to be sent out. So while your analysis here is correct I'd prefer to hold this as the issue goes away entirely.
On Sun, Jan 16, 2022 at 11:15 PM Christoph Hellwig <hch@infradead.org> wrote: > > FYI, I have a patch removing mpage_alloc entirely as part of a bio > allocation refactoring series about to be sent out. So while your > analysis here is correct I'd prefer to hold this as the issue goes away > entirely. SGTM
diff --git a/fs/mpage.c b/fs/mpage.c index 87f5cfef6caa..477ccc3f3ac3 100644 --- a/fs/mpage.c +++ b/fs/mpage.c @@ -609,7 +609,7 @@ static int __mpage_writepage(struct page *page, struct writeback_control *wbc, goto out; } bio = mpage_alloc(bdev, blocks[0] << (blkbits - 9), - BIO_MAX_VECS, GFP_NOFS|__GFP_HIGH); + BIO_MAX_VECS, GFP_NOFS); if (bio == NULL) goto confused;
Since the commit 8a5c743e308dd ("mm, memcg: use consistent gfp flags during readahead") mpage_alloc is clearing all the flag bits other than bits in GFP_KERNEL internally. Simply remove __GFP_HIGH from the call to mpage_alloc as it is cleared by it. Signed-off-by: Shakeel Butt <shakeelb@google.com> --- fs/mpage.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)