Re: [PATCH v2] workqueue: fix state-dump console deadlock
From: John Ogness <john.ogness@linutronix.de>
Date: 2021-10-06 19:09:13
Also in:
lkml, stable
On 2021-10-06, Johan Hovold [off-list ref] wrote:
Console drivers often queue work while holding locks also taken in their console write paths, something which can lead to deadlocks on SMP when dumping workqueue state (e.g. sysrq-t or on suspend failures). For serial console drivers this could look like: CPU0 CPU1 ---- ---- show_workqueue_state(); lock(&pool->lock); <IRQ> lock(&port->lock); schedule_work(); lock(&pool->lock); printk(); lock(console_owner); lock(&port->lock); where workqueues are, for example, used to push data to the line discipline, process break signals and handle modem-status changes. Line disciplines and serdev drivers can also queue work on write-wakeup notifications, etc. Reworking every console driver to avoid queuing work while holding locks also taken in their write paths would complicate drivers and is neither desirable or feasible. Instead use the deferred-printk mechanism to avoid printing while holding pool locks when dumping workqueue state.
When I introduced the printk_deferred_enter/exit functions, I kind of expected patches like this to start showing up. The functions make it really convenient to establish general sections of console print deferring. When we move to kthread-printers, all printk calls will be deferred automatically. However, that is only during normal operation. The various printk_deferred sites may still be significant and will continue to have special meaning during startup and shutdown states, when the kthreads will not be active and direct printing will exist as it is now. FWIW, I am OK with this patch. It will be re-evaluated once we have kthread-printers, but I suspect even then it will remain.
Note that there are a few WARN_ON() assertions in the workqueue code
which could potentially also trigger a deadlock. Hopefully the ongoing
printk rework will provide a general solution for this eventually.
This was originally reported after a lockdep splat when executing
sysrq-t with the imx serial driver.
Fixes: 3494fc30846d ("workqueue: dump workqueues on sysrq-t")
Cc: stable@vger.kernel.org # 4.0
Reported-by: Fabio Estevam <redacted>
Tested-by: Fabio Estevam <redacted>
Signed-off-by: Johan Hovold <johan@kernel.org>Reviewed-by: John Ogness <john.ogness@linutronix.de>
quoted hunk ↗ jump to hunk
--- Changes in v2 - defer printing also of worker pool state (Peter Mladek) - add Fabio's tested-by tag kernel/workqueue.c | 18 ++++++++++++++++-- 1 file changed, 16 insertions(+), 2 deletions(-)diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 33a6b4a2443d..1b3eb1e9531f 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c@@ -4830,8 +4830,16 @@ void show_workqueue_state(void) for_each_pwq(pwq, wq) { raw_spin_lock_irqsave(&pwq->pool->lock, flags); - if (pwq->nr_active || !list_empty(&pwq->inactive_works)) + if (pwq->nr_active || !list_empty(&pwq->inactive_works)) { + /* + * Defer printing to avoid deadlocks in console + * drivers that queue work while holding locks + * also taken in their write paths. + */ + printk_deferred_enter(); show_pwq(pwq); + printk_deferred_exit(); + } raw_spin_unlock_irqrestore(&pwq->pool->lock, flags); /* * We could be printing a lot from atomic context, e.g.@@ -4849,7 +4857,12 @@ void show_workqueue_state(void) raw_spin_lock_irqsave(&pool->lock, flags); if (pool->nr_workers == pool->nr_idle) goto next_pool; - + /* + * Defer printing to avoid deadlocks in console drivers that + * queue work while holding locks also taken in their write + * paths. + */ + printk_deferred_enter(); pr_info("pool %d:", pool->id); pr_cont_pool_info(pool); pr_cont(" hung=%us workers=%d",@@ -4864,6 +4877,7 @@ void show_workqueue_state(void) first = false; } pr_cont("\n"); + printk_deferred_exit(); next_pool: raw_spin_unlock_irqrestore(&pool->lock, flags); /*-- 2.32.0