linux-kernel - [PATCH v7 5/9] blk-throttle: improve handling of re-entered bio for bps limit

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: <20220802140415.2960284-6-yukuai1@huaweicloud.com>
Date:   Tue,  2 Aug 2022 22:04:11 +0800
From:   Yu Kuai <yukuai1@...weicloud.com>
To:     tj@...nel.org, mkoutny@...e.com, axboe@...nel.dk,
        ming.lei@...hat.com
Cc:     cgroups@...r.kernel.org, linux-block@...r.kernel.org,
        linux-kernel@...r.kernel.org, yukuai3@...wei.com,
        yukuai1@...weicloud.com, yi.zhang@...wei.com
Subject: [PATCH v7 5/9] blk-throttle: improve handling of re-entered bio for bps limit

From: Yu Kuai <yukuai3@...wei.com>

Currently, if new slice is started while bio is re-entered, such bio
will count multiple times for bps limit. Since 'bytes_skipped'
represents how many bytes will be skipped while dispatching bios, which
can handle that case, increasing it instead of decreasing 'bytes_disp'.

Signed-off-by: Yu Kuai <yukuai3@...wei.com>
---
 block/blk-throttle.c | 8 ++++----
 block/blk-throttle.h | 4 +++-
 2 files changed, 7 insertions(+), 5 deletions(-)

diff --git a/block/blk-throttle.c b/block/blk-throttle.c
index 621402cf2576..237668328e98 100644
--- a/block/blk-throttle.c
+++ b/block/blk-throttle.c
@@ -2183,14 +2183,14 @@ bool __blk_throtl_bio(struct bio *bio)
 		 * Splited bios can be re-entered because iops limit should be
 		 * counted again, however, bps limit should not. Since bps limit
 		 * will be counted again while dispatching it, compensate the
-		 * over-accounting here. Noted that compensation can fail if
-		 * new slice is started.
+		 * over-accounting here. Since decrement of 'bytes_disp' can't
+		 * handle the case that new slice is started, increase
+		 * 'bytes_skipped' instead.
 		 */
 		if (bio_flagged(bio, BIO_THROTTLED)) {
 			unsigned int bio_size = throtl_bio_data_size(bio);
 
-			if (tg->bytes_disp[rw] >= bio_size)
-				tg->bytes_disp[rw] -= bio_size;
+			tg->bytes_skipped[rw] += bio_size;
 			if (tg->last_bytes_disp[rw] >= bio_size)
 				tg->last_bytes_disp[rw] -= bio_size;
 		}
diff --git a/block/blk-throttle.h b/block/blk-throttle.h
index 0163aa9104c3..c9545616ba12 100644
--- a/block/blk-throttle.h
+++ b/block/blk-throttle.h
@@ -121,7 +121,9 @@ struct throtl_grp {
 	 * bytes/io are waited already in previous configuration, and they will
 	 * be used to calculate wait time under new configuration.
 	 *
-	 * Number of bytes will be skipped in current slice
+	 * Number of bytes will be skipped in current slice. In addition, this
+	 * field will help to handle re-entered bio for bps limit, see details
+	 * in __blk_throtl_bio().
 	 */
 	uint64_t bytes_skipped[2];
 	/* Number of bio will be skipped in current slice */
-- 
2.31.1