linux-kernel - [PATCH 5.6 094/194] gfs2: More gfs2_find

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <20200518173539.965917703@linuxfoundation.org>
Date:   Mon, 18 May 2020 19:36:24 +0200
From:   Greg Kroah-Hartman <gregkh@...uxfoundation.org>
To:     linux-kernel@...r.kernel.org
Cc:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        stable@...r.kernel.org, Andreas Gruenbacher <agruenba@...hat.com>,
        Bob Peterson <rpeterso@...hat.com>,
        Sasha Levin <sashal@...nel.org>
Subject: [PATCH 5.6 094/194] gfs2: More gfs2_find_jhead fixes

From: Andreas Gruenbacher <agruenba@...hat.com>

[ Upstream commit aa83da7f47b26c9587bade6c4bc4736ffa308f0a ]

It turns out that when extending an existing bio, gfs2_find_jhead fails to
check if the block number is consecutive, which leads to incorrect reads for
fragmented journals.

In addition, limit the maximum bio size to an arbitrary value of 2 megabytes:
since commit 07173c3ec276 ("block: enable multipage bvecs"), if we just keep
adding pages until bio_add_page fails, bios will grow much larger than useful,
which pins more memory than necessary with barely any additional performance
gains.

Fixes: f4686c26ecc3 ("gfs2: read journal in large chunks")
Cc: stable@...r.kernel.org # v5.2+
Signed-off-by: Andreas Gruenbacher <agruenba@...hat.com>
Signed-off-by: Bob Peterson <rpeterso@...hat.com>
Signed-off-by: Sasha Levin <sashal@...nel.org>
---
 fs/gfs2/lops.c | 19 ++++++++++++-------
 1 file changed, 12 insertions(+), 7 deletions(-)

diff --git a/fs/gfs2/lops.c b/fs/gfs2/lops.c
index c090d5ad3f221..3a020bdc358cd 100644
--- a/fs/gfs2/lops.c
+++ b/fs/gfs2/lops.c
@@ -259,7 +259,7 @@ static struct bio *gfs2_log_alloc_bio(struct gfs2_sbd *sdp, u64 blkno,
 	struct super_block *sb = sdp->sd_vfs;
 	struct bio *bio = bio_alloc(GFP_NOIO, BIO_MAX_PAGES);
 
-	bio->bi_iter.bi_sector = blkno << (sb->s_blocksize_bits - 9);
+	bio->bi_iter.bi_sector = blkno << sdp->sd_fsb2bb_shift;
 	bio_set_dev(bio, sb->s_bdev);
 	bio->bi_end_io = end_io;
 	bio->bi_private = sdp;
@@ -505,7 +505,7 @@ int gfs2_find_jhead(struct gfs2_jdesc *jd, struct gfs2_log_header_host *head,
 	unsigned int bsize = sdp->sd_sb.sb_bsize, off;
 	unsigned int bsize_shift = sdp->sd_sb.sb_bsize_shift;
 	unsigned int shift = PAGE_SHIFT - bsize_shift;
-	unsigned int readahead_blocks = BIO_MAX_PAGES << shift;
+	unsigned int max_bio_size = 2 * 1024 * 1024;
 	struct gfs2_journal_extent *je;
 	int sz, ret = 0;
 	struct bio *bio = NULL;
@@ -533,12 +533,17 @@ int gfs2_find_jhead(struct gfs2_jdesc *jd, struct gfs2_log_header_host *head,
 				off = 0;
 			}
 
-			if (!bio || (bio_chained && !off)) {
+			if (!bio || (bio_chained && !off) ||
+			    bio->bi_iter.bi_size >= max_bio_size) {
 				/* start new bio */
 			} else {
-				sz = bio_add_page(bio, page, bsize, off);
-				if (sz == bsize)
-					goto block_added;
+				sector_t sector = dblock << sdp->sd_fsb2bb_shift;
+
+				if (bio_end_sector(bio) == sector) {
+					sz = bio_add_page(bio, page, bsize, off);
+					if (sz == bsize)
+						goto block_added;
+				}
 				if (off) {
 					unsigned int blocks =
 						(PAGE_SIZE - off) >> bsize_shift;
@@ -564,7 +569,7 @@ int gfs2_find_jhead(struct gfs2_jdesc *jd, struct gfs2_log_header_host *head,
 			off += bsize;
 			if (off == PAGE_SIZE)
 				page = NULL;
-			if (blocks_submitted < blocks_read + readahead_blocks) {
+			if (blocks_submitted < 2 * max_bio_size >> bsize_shift) {
 				/* Keep at least one bio in flight */
 				continue;
 			}
-- 
2.20.1