Re: [PATCH] kfence: Avoid stalling work queue task without allocations

From: Anders Roxell
Date: Tue Nov 10 2020 - 18:23:39 EST


On Tue, 10 Nov 2020 at 14:53, Marco Elver <elver@xxxxxxxxxx> wrote:
>
> To toggle the allocation gates, we set up a delayed work that calls
> toggle_allocation_gate(). Here we use wait_event() to await an
> allocation and subsequently disable the static branch again. However, if
> the kernel has stopped doing allocations entirely, we'd wait
> indefinitely, and stall the worker task. This may also result in the
> appropriate warnings if CONFIG_DETECT_HUNG_TASK=y.
>
> Therefore, introduce a 1 second timeout and use wait_event_timeout(). If
> the timeout is reached, the static branch is disabled and a new delayed
> work is scheduled to try setting up an allocation at a later time.
>
> Note that, this scenario is very unlikely during normal workloads once
> the kernel has booted and user space tasks are running. It can, however,
> happen during early boot after KFENCE has been enabled, when e.g.
> running tests that do not result in any allocations.
>
> Link: https://lkml.kernel.org/r/CADYN=9J0DQhizAGB0-jz4HOBBh+05kMBXb4c0cXMS7Qi5NAJiw@xxxxxxxxxxxxxx
> Reported-by: Anders Roxell <anders.roxell@xxxxxxxxxx>
> Signed-off-by: Marco Elver <elver@xxxxxxxxxx>
> ---
> mm/kfence/core.c | 6 +++++-
> 1 file changed, 5 insertions(+), 1 deletion(-)
>
> diff --git a/mm/kfence/core.c b/mm/kfence/core.c
> index 9358f42a9a9e..933b197b8634 100644
> --- a/mm/kfence/core.c
> +++ b/mm/kfence/core.c
> @@ -592,7 +592,11 @@ static void toggle_allocation_gate(struct work_struct *work)
> /* Enable static key, and await allocation to happen. */
> atomic_set(&allocation_gate, 0);
> static_branch_enable(&kfence_allocation_key);
> - wait_event(allocation_wait, atomic_read(&allocation_gate) != 0);
> + /*
> + * Await an allocation. Timeout after 1 second, in case the kernel stops
> + * doing allocations, to avoid stalling this worker task for too long.
> + */
> + wait_event_timeout(allocation_wait, atomic_read(&allocation_gate) != 0, HZ);
>
> /* Disable static key and reset timer. */
> static_branch_disable(&kfence_allocation_key);
> --
> 2.29.2.222.g5d2a92d10f8-goog
>

I gave them a spin on next-20201105 [1] and on next-20201110 [2].

I eventually got to a prompt on next-20201105.
However, I got to this kernel panic on the next-20201110:

[...]
[ 1514.089966][ T1] Testing event system initcall: OK
[ 1514.806232][ T1] Running tests on all trace events:
[ 1514.857835][ T1] Testing all events:
[ 1525.503262][ C0] hrtimer: interrupt took 10902600 ns
[ 1623.861452][ C0] BUG: workqueue lockup - pool cpus=0 node=0
flags=0x0 nice=0 stuck for 65s!
[...]
[ 7823.104349][ T28] Tainted: G W
5.10.0-rc3-next-20201110-00008-g8dc06700529d #3
[ 7833.206491][ T28] "echo 0 >
/proc/sys/kernel/hung_task_timeout_secs" disables this message.
[ 7840.750700][ T28] task:kworker/0:1 state:D stack:26640 pid:
1872 ppid: 2 flags:0x00000428
[ 7875.642531][ T28] Workqueue: events toggle_allocation_gate
[ 7889.178334][ T28] Call trace:
[ 7897.066649][ T28] __switch_to+0x1cc/0x1e0
[ 7905.326856][ T28] 0xffff00000f7077b0
[ 7928.354644][ T28] INFO: lockdep is turned off.
[ 7934.022572][ T28] Kernel panic - not syncing: hung_task: blocked tasks
[ 7934.032039][ T28] CPU: 0 PID: 28 Comm: khungtaskd Tainted: G
W 5.10.0-rc3-next-20201110-00008-g8dc06700529d #3
[ 7934.045586][ T28] Hardware name: linux,dummy-virt (DT)
[ 7934.053677][ T28] Call trace:
[ 7934.060276][ T28] dump_backtrace+0x0/0x420
[ 7934.067635][ T28] show_stack+0x38/0xa0
[ 7934.091277][ T28] dump_stack+0x1d4/0x278
[ 7934.098878][ T28] panic+0x304/0x5d8
[ 7934.114923][ T28] check_hung_uninterruptible_tasks+0x5e4/0x640
[ 7934.123823][ T28] watchdog+0x138/0x160
[ 7934.131561][ T28] kthread+0x23c/0x260
[ 7934.138590][ T28] ret_from_fork+0x10/0x18
[ 7934.146631][ T28] Kernel Offset: disabled
[ 7934.153749][ T28] CPU features: 0x0240002,20002004
[ 7934.161476][ T28] Memory Limit: none
[ 7934.171272][ T28] ---[ end Kernel panic - not syncing: hung_task:
blocked tasks ]---


Cheers,
Anders
[1] https://people.linaro.org/~anders.roxell/output-next-20201105-test.log
[2] https://people.linaro.org/~anders.roxell/output-next-20201110-test.log