]> www.infradead.org Git - users/willy/xarray.git/commit
blk_iocost: fix more out of bound shifts
authorKonstantin Ovsepian <ovs@ovs.to>
Thu, 22 Aug 2024 15:41:36 +0000 (08:41 -0700)
committerJens Axboe <axboe@kernel.dk>
Mon, 26 Aug 2024 20:13:22 +0000 (14:13 -0600)
commit9bce8005ec0dcb23a58300e8522fe4a31da606fa
tree8f998ee9db991307e2f6ed07422e65aead5a5148
parent87599eddc25ac03647ab76221523c6485e7594b1
blk_iocost: fix more out of bound shifts

Recently running UBSAN caught few out of bound shifts in the
ioc_forgive_debts() function:

UBSAN: shift-out-of-bounds in block/blk-iocost.c:2142:38
shift exponent 80 is too large for 64-bit type 'u64' (aka 'unsigned long
long')
...
UBSAN: shift-out-of-bounds in block/blk-iocost.c:2144:30
shift exponent 80 is too large for 64-bit type 'u64' (aka 'unsigned long
long')
...
Call Trace:
<IRQ>
dump_stack_lvl+0xca/0x130
__ubsan_handle_shift_out_of_bounds+0x22c/0x280
? __lock_acquire+0x6441/0x7c10
ioc_timer_fn+0x6cec/0x7750
? blk_iocost_init+0x720/0x720
? call_timer_fn+0x5d/0x470
call_timer_fn+0xfa/0x470
? blk_iocost_init+0x720/0x720
__run_timer_base+0x519/0x700
...

Actual impact of this issue was not identified but I propose to fix the
undefined behaviour.
The proposed fix to prevent those out of bound shifts consist of
precalculating exponent before using it the shift operations by taking
min value from the actual exponent and maximum possible number of bits.

Reported-by: Breno Leitao <leitao@debian.org>
Signed-off-by: Konstantin Ovsepian <ovs@ovs.to>
Acked-by: Tejun Heo <tj@kernel.org>
Link: https://lore.kernel.org/r/20240822154137.2627818-1-ovs@ovs.to
Signed-off-by: Jens Axboe <axboe@kernel.dk>
block/blk-iocost.c