linux-kernel - Re: [PATCH] vt: Fix sleeping functions called from atomic context

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, 16 Nov 2021 16:35:07 +0100
From:   "Fabio M. De Francesco" <fmdefrancesco@...il.com>
To:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>
Cc:     Jiri Slaby <jirislaby@...nel.org>,
        Max Filippov <jcmvbkbc@...il.com>,
        David Sterba <dsterba@...e.com>,
        Bhaskar Chowdhury <unixbhaskar@...il.com>,
        Tetsuo Handa <penguin-kernel@...ove.sakura.ne.jp>,
        Igor Matheus Andrade Torrente <igormtorrente@...il.com>,
        nick black <dankamongmen@...il.com>,
        linux-kernel@...r.kernel.org,
        syzbot+5f47a8cea6a12b77a876@...kaller.appspotmail.com,
        Marco Elver <elver@...gle.com>
Subject: Re: [PATCH] vt: Fix sleeping functions called from atomic context

On Tuesday, November 16, 2021 3:58:44 PM CET Greg Kroah-Hartman wrote:
> On Tue, Nov 16, 2021 at 03:49:37PM +0100, Fabio M. De Francesco wrote:
> > Fix two sleeping functions called from atomic context by doing immediate
> > return to the caller if !preemptible() evaluates 'true'. Remove two
> > in_interrupt() tests because they are not suited for being used here.
> > 
> > Since functions do_con_write() and con_flush_chars() might sleep in
> > console_lock(), it must be assured that they are never executed in
> > atomic contexts.
> > 
> > This issue is reported by Syzbot which notices that they are executed
> > while holding spinlocks and with interrupts disabled. Actually Syzbot
> > emits a first report and then, after fixing do_con_write(), a second
> > report for the same problem in con_flush_chars() because these functions
> > are called one after the other by con_write().
> > 
> > Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
> > Reported-by: syzbot+5f47a8cea6a12b77a876@...kaller.appspotmail.com
> > Suggested-by: Marco Elver <elver@...gle.com>
> > Signed-off-by: Fabio M. De Francesco <fmdefrancesco@...il.com>
> > ---
> >  drivers/tty/vt/vt.c | 4 ++--
> >  1 file changed, 2 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/tty/vt/vt.c b/drivers/tty/vt/vt.c
> > index 7359c3e80d63..508f8a56d361 100644
> > --- a/drivers/tty/vt/vt.c
> > +++ b/drivers/tty/vt/vt.c
> > @@ -2902,7 +2902,7 @@ static int do_con_write(struct tty_struct *tty, 
const unsigned char *buf, int co
> >  	struct vt_notifier_param param;
> >  	bool rescan;
> >  
> > -	if (in_interrupt())
> > +	if (!preemptible())
> >  		return count;
> 
> Very odd, what code is calling these functions to trigger this check?

This is the call trace reported by Syzbot (https://syzkaller.appspot.com/bug?
id=fe5a4d5a2482bd73064db5de5d28e024f1e2a387):

Call Trace:
 <TASK>
 __dump_stack lib/dump_stack.c:88 [inline]
 dump_stack_lvl+0xcd/0x134 lib/dump_stack.c:106
 __might_resched.cold+0x222/0x26b kernel/sched/core.c:9539
 console_lock+0x17/0x80 kernel/printk/printk.c:2522
 do_con_write+0x10f/0x1e40 drivers/tty/vt/vt.c:2908
 con_write+0x21/0x40 drivers/tty/vt/vt.c:3295
 n_hdlc_send_frames+0x24b/0x490 drivers/tty/n_hdlc.c:290
 tty_wakeup+0xe1/0x120 drivers/tty/tty_io.c:534
 __start_tty drivers/tty/tty_io.c:806 [inline]
 __start_tty+0xfb/0x130 drivers/tty/tty_io.c:799
 n_tty_ioctl_helper+0x299/0x2d0 drivers/tty/tty_ioctl.c:880

	^^^^^^^^^^
n_tty_ioctl_helper() disabled interrupts via spin_lock_irq(&tty->flow.lock).

 n_hdlc_tty_ioctl+0xd2/0x340 drivers/tty/n_hdlc.c:633
 tty_ioctl+0xc69/0x1670 drivers/tty/tty_io.c:2814
 vfs_ioctl fs/ioctl.c:51 [inline]
 __do_sys_ioctl fs/ioctl.c:874 [inline]
 __se_sys_ioctl fs/ioctl.c:860 [inline]
 __x64_sys_ioctl+0x193/0x200 fs/ioctl.c:860
 do_syscall_x64 arch/x86/entry/common.c:50 [inline]
 do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80
 entry_SYSCALL_64_after_hwframe+0x44/0xae

> Shouldn't the caller be fixed instead?

Maybe that the caller has no need to disable IRQs, but I cannot yet answer to 
this particular question.

> What changed to suddenly cause this to show up?

Commit c545b66c6922 ("tty: Serialize tcflow() with other tty flow control 
changes") introduced a call to spin_lock_irq() for command "TCOON", just 
before calling __start_tty().

Thanks,

Fabio M. De Francesco

> Given that this check has been here for a _very_ long time, changing it
> now without finding out the root cause is probably not a good idea.
> 
> thanks,
> 
> greg k-h
>