Re: [rbtree_test_init] watchdog: BUG: soft lockup - CPU#0 stuck for 22s! [swapper:1]

From: Fengguang Wu
Date: Tue Nov 21 2017 - 21:15:38 EST

On Tue, Nov 21, 2017 at 01:59:13PM -0800, Andrew Morton wrote:
On Tue, 21 Nov 2017 19:10:22 +0800 Fengguang Wu <fengguang.wu@xxxxxxxxx> wrote:


FYI this happens in mainline kernel 4.14.0-03453-ge37e0ee.
It at least dates back to v4.1 .

It occurs in 2 out of 2 boots.

[ 140.043368] watchdog: BUG: soft lockup - CPU#0 stuck for 22s! [swapper:1]
[ 140.044140] Call Trace:
[ 140.044140] __asan_loadN+0x19/0x21
[ 140.044140] ftrace_likely_update+0x32/0xc3:
ftrace_likely_update at kernel/trace/trace_branch.c:209
[ 140.044140] ? augment_recompute+0x14b/0x14b:
augment_callbacks_rotate at lib/rbtree_test.c:97
[ 140.044140] __rb_insert_augmented+0x8d/0x639:
__rb_insert at lib/rbtree.c:110
(inlined by) __rb_insert_augmented at lib/rbtree.c:495
[ 140.044140] rbtree_test_init+0xbfb/0x1407:
rb_insert_augmented_cached at include/linux/rbtree_augmented.h:70
(inlined by) insert_augmented_cached at lib/rbtree_test.c:148
(inlined by) rbtree_test_init at lib/rbtree_test.c:371
[ 140.044140] ? glob_init+0x23f/0x23f:
rbtree_test_init at lib/rbtree_test.c:245

Does that kernel include 0b548e33e6cb2bf ("lib/rbtree-test: lower
default params")? If so, perhaps some cond_resched() sprinkling is

Ah I just tested 0b548e33e6cb2bf and it's fine there. 0b548e33e6cb2bf
is merged after e37e0ee0190 in this report, so it indicates a fixed
error. Sorry for the noise!