[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <20240410193642.1303741-1-leitao@debian.org>
Date: Wed, 10 Apr 2024 12:36:41 -0700
From: Breno Leitao <leitao@...ian.org>
To: Tejun Heo <tj@...nel.org>,
Josef Bacik <josef@...icpanda.com>,
Jens Axboe <axboe@...nel.dk>
Cc: leit@...a.com,
cgroups@...r.kernel.org (open list:CONTROL GROUP - BLOCK IO CONTROLLER (BLKIO)),
linux-block@...r.kernel.org (open list:CONTROL GROUP - BLOCK IO CONTROLLER (BLKIO)),
linux-kernel@...r.kernel.org (open list)
Subject: [PATCH] blk-iocost: Fix shift-out-of-bounds in iocg_kick_delay()
When running iocg_kick_delay(), iocg->delay_at could be way behind "now",
which causes a huge tdelta difference.
The tdelta value is used to shift some bits around, raising the
following UBSAN splat:
UBSAN: shift-out-of-bounds in block/blk-iocost.c:1366:23
Debugging this, these are some values I got in my machine with 6.9-rc3.
now = 3626064295
iocg->delay_at = 3275794093
Fix this by validating that the shift if valid, otherwise bail out,
similarly to commit 2a427b49d029 ("blk-iocost: Fix an UBSAN
shift-out-of-bounds warning")
Signed-off-by: Breno Leitao <leitao@...ian.org>
---
block/blk-iocost.c | 9 +++++++--
1 file changed, 7 insertions(+), 2 deletions(-)
diff --git a/block/blk-iocost.c b/block/blk-iocost.c
index 9a85bfbbc45a..398fe19db4ca 100644
--- a/block/blk-iocost.c
+++ b/block/blk-iocost.c
@@ -1347,7 +1347,7 @@ static bool iocg_kick_delay(struct ioc_gq *iocg, struct ioc_now *now)
{
struct ioc *ioc = iocg->ioc;
struct blkcg_gq *blkg = iocg_to_blkg(iocg);
- u64 tdelta, delay, new_delay;
+ u64 tdelta, delay, new_delay, shift;
s64 vover, vover_pct;
u32 hwa;
@@ -1362,8 +1362,13 @@ static bool iocg_kick_delay(struct ioc_gq *iocg, struct ioc_now *now)
/* calculate the current delay in effect - 1/2 every second */
tdelta = now->now - iocg->delay_at;
+
+ shift = div64_u64(tdelta, USEC_PER_SEC);
+ if (shift >= BITS_PER_LONG)
+ return false;
+
if (iocg->delay)
- delay = iocg->delay >> div64_u64(tdelta, USEC_PER_SEC);
+ delay = iocg->delay >> shift;
else
delay = 0;
--
2.43.0
Powered by blists - more mailing lists