linux-kernel - Re: [PATCH 0/2] Fix BUG: Invalid wait context in hrtimer

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Tue, 26 Jan 2021 18:08:04 +0000
From:   Qais Yousef <qais.yousef@....com>
To:     Peter Zijlstra <peterz@...radead.org>
Cc:     Sergey Senozhatsky <sergey.senozhatsky@...il.com>,
        Petr Mladek <pmladek@...e.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        Steven Rostedt <rostedt@...dmis.org>,
        John Ogness <john.ogness@...utronix.de>,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH 0/2] Fix BUG: Invalid wait context in hrtimer_interrupt()

On 01/26/21 17:23, Peter Zijlstra wrote:
> On Wed, Jan 27, 2021 at 12:58:33AM +0900, Sergey Senozhatsky wrote:
> > On (21/01/26 14:59), Qais Yousef wrote:
> 
> > > 	# [67628.388606] hrtimer: interrupt took 304720 ns
> > > 	[67628.393546]
> > > 	[67628.393550] =============================
> > > 	[67628.393554] [ BUG: Invalid wait context ]
> > > 	[67628.393557] 5.11.0-rc3-00019-g86be331946f7 #37 Not tainted
> > > 	[67628.393560] -----------------------------
> > > 	[67628.393563] sugov:0/192 is trying to lock:
> > > 	[67628.393566] ffff000800b1d898 (&port_lock_key){-.-.}-{3:3}, at: pl011_console_write+0x138/0x218
> > > 	[67628.393581] other info that might help us debug this:
> > > 	[67628.393584] context-{2:2}
> > > 	[67628.393586] 4 locks held by sugov:0/192:
> > > 	[67628.393589]  #0: ffff0008059cb720 (&sg_policy->work_lock){+.+.}-{4:4}, at: sugov_work+0x58/0x88
> > > 	[67628.393603]  #1: ffff800015446f20 (prepare_lock){+.+.}-{4:4}, at: clk_prepare_lock+0x34/0xb0
> > > 	[67628.393618]  #2: ffff8000152aaa60 (console_lock){+.+.}-{0:0}, at: vprintk_emit+0x12c/0x310
> > > 	[67628.393632]  #3: ffff8000152aab88 (console_owner){-.-.}-{0:0}, at: console_unlock+0x190/0x6d8
> 
> > > Did I miss something?
> > 
> > printk() is not permitted to sleep/schedule/etc and it never does.
> > Generally it should be OK to call it from IRQ (module recursion paths).
> 
> The report is that it is trying to acquire spin_lock() while holding
> raw_spin_lock(), which is an invalid lock nesting.
> 
> Note that this is CONFIG_PROVE_RAW_LOCK_NESTING=y which specifically
> checks for this.
> 
> On current (mainline) kernel configs this is not yet a problem, but the
> moment we do land PREEMPT_RT this order will be problematic.

I should have dug more into the history of printk() and the meaning of the
splat. Sorry for the noise.

Looking at v5.10.8-rt24 the following fix is applied in RT

	https://git.kernel.org/pub/scm/linux/kernel/git/rt/linux-rt-devel.git/commit/drivers/tty/serial/amba-pl011.c?h=linux-5.10.y-rt&id=008cc77aff249e830e5eb90b7ae3a6784597b8cf

which is what John suggested.

Looking at the locks held

> > > 	[67628.393589]  #0: ffff0008059cb720 (&sg_policy->work_lock){+.+.}-{4:4}, at: sugov_work+0x58/0x88
> > > 	[67628.393603]  #1: ffff800015446f20 (prepare_lock){+.+.}-{4:4}, at: clk_prepare_lock+0x34/0xb0

These two are mutexes.

> > > 	[67628.393618]  #2: ffff8000152aaa60 (console_lock){+.+.}-{0:0}, at: vprintk_emit+0x12c/0x310

This is a semaphore.

> > > 	[67628.393632]  #3: ffff8000152aab88 (console_owner){-.-.}-{0:0}, at: console_unlock+0x190/0x6d8

I think this is acquired by console_lock_spinning_enable() which has acquiring
syntax I'm not familiar with. console_owner_lock is defined as RAW_SPINLOCK, so
regardless of how it is acquired, it must be the problem.

Looks like John has reworked this code in RT too. So maybe this is just a red
herring after all..

	https://git.kernel.org/pub/scm/linux/kernel/git/rt/linux-rt-devel.git/commit/kernel/printk/printk.c?h=linux-5.10.y-rt&id=0097798fd99948d3ffea535005eee7eb3b14fd06

Thanks

--
Qais Yousef