linux-rt-users.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
To: Esben Haabendal <esben.haabendal@gmail.com>
Cc: "John Ogness" <john.ogness@linutronix.de>,
	"André Pribil" <a.pribil@beck-ipc.com>,
	"linux-rt-users@vger.kernel.org" <linux-rt-users@vger.kernel.org>
Subject: Re: tty latency and RT
Date: Wed, 11 Jul 2018 10:09:57 +0200	[thread overview]
Message-ID: <20180711080957.f6txdmzrrrrdm7ig@linutronix.de> (raw)
In-Reply-To: <87pnzwag3n.fsf@gmail.com>

On 2018-07-09 11:55:08 [+0200], Esben Haabendal wrote:
> I am using the following patch.
> Not sure if it is worth proposing it for mainline inclusion, though.
> RFC:

| ======================================================
| WARNING: possible circular locking dependency detected
| 4.16.18-rt9+ #187 Not tainted
| ------------------------------------------------------
| sshd/3205 is trying to acquire lock:
|  (&buf->lock){+.+.}, at: [<        (ptrval)>] flush_to_ldisc+0x1e/0xa0
| 
| but task is already holding lock:
|  (&ldata->output_lock){+.+.}, at: [<        (ptrval)>] n_tty_write+0x12a/0x480
| 
| which lock already depends on the new lock.
|
| 
| the existing dependency chain (in reverse order) is:
| 
| -> #2 (&ldata->output_lock){+.+.}:
|        _mutex_lock+0x26/0x40
|        n_tty_write+0x12a/0x480
|        tty_write+0x1b3/0x320
|        redirected_tty_write+0x9a/0xb0
|        do_iter_write+0x159/0x1a0
|        vfs_writev+0x93/0x110
|        do_writev+0x5f/0xf0
|        SyS_writev+0xb/0x10
|        do_syscall_64+0x73/0x220
|        entry_SYSCALL_64_after_hwframe+0x42/0xb7
| 
| -> #1 (&tty->termios_rwsem){++++}:
|        down_write+0x39/0x50
|        n_tty_flush_buffer+0x19/0xf0
|        tty_buffer_flush+0x71/0x90
|        tty_ldisc_flush+0x1d/0x40
|        tty_port_close_start.part.5+0xa0/0x1b0
|        tty_port_close+0x29/0x60
|        uart_close+0x26/0x70
|        tty_release+0xfc/0x4f0
|        __fput+0xf1/0x200
|        ____fput+0x9/0x10
|        task_work_run+0x8b/0xc0
|        exit_to_usermode_loop+0xbc/0xc0
|        do_syscall_64+0x21b/0x220
|        entry_SYSCALL_64_after_hwframe+0x42/0xb7
| 
| -> #0 (&buf->lock){+.+.}:
|        lock_acquire+0x95/0x240
|        _mutex_lock+0x26/0x40
|        flush_to_ldisc+0x1e/0xa0
|        tty_flip_buffer_push+0x28/0x40
|        pty_write+0x4e/0x60
|        n_tty_write+0x1ae/0x480
|        tty_write+0x1b3/0x320
|        __vfs_write+0x35/0x160
|        vfs_write+0xc1/0x1c0
|        SyS_write+0x53/0xc0
|        do_syscall_64+0x73/0x220
|        entry_SYSCALL_64_after_hwframe+0x42/0xb7
| 
| other info that might help us debug this:
|
| Chain exists of:
|   &buf->lock --> &tty->termios_rwsem --> &ldata->output_lock
|
|  Possible unsafe locking scenario:
|
|        CPU0                    CPU1
|        ----                    ----
|   lock(&ldata->output_lock);
|                                lock(&tty->termios_rwsem);
|                                lock(&ldata->output_lock);
|   lock(&buf->lock);
| 
|  *** DEADLOCK ***
|
| 4 locks held by sshd/3205:
|  #0:  (&tty->ldisc_sem){++++}, at: [<        (ptrval)>] ldsem_down_read+0x2d/0x40
|  #1:  (&tty->atomic_write_lock){+.+.}, at: [<        (ptrval)>] tty_write_lock+0x19/0x50
|  #2:  (&o_tty->termios_rwsem/1){++++}, at: [<        (ptrval)>] n_tty_write+0x9a/0x480
|  #3:  (&ldata->output_lock){+.+.}, at: [<        (ptrval)>] n_tty_write+0x12a/0x480
| 
| stack backtrace:
| CPU: 7 PID: 3205 Comm: sshd Not tainted 4.16.18-rt9+ #187
| Call Trace:
|  dump_stack+0x70/0xa7
|  print_circular_bug.isra.37+0x1d8/0x1e6
|  __lock_acquire+0x1284/0x1340
|  ? flush_to_ldisc+0x1e/0xa0
|  lock_acquire+0x95/0x240
|  ? lock_acquire+0x95/0x240
|  ? flush_to_ldisc+0x1e/0xa0
|  _mutex_lock+0x26/0x40
|  ? flush_to_ldisc+0x1e/0xa0
|  flush_to_ldisc+0x1e/0xa0
|  tty_flip_buffer_push+0x28/0x40
|  pty_write+0x4e/0x60
|  n_tty_write+0x1ae/0x480
|  ? do_wait_intr_irq+0xc0/0xc0
|  tty_write+0x1b3/0x320
|  ? process_echoes+0x60/0x60
|  __vfs_write+0x35/0x160
|  ? SYSC_newfstat+0x44/0x70
|  vfs_write+0xc1/0x1c0
|  SyS_write+0x53/0xc0
|  do_syscall_64+0x73/0x220
|  entry_SYSCALL_64_after_hwframe+0x42/0xb7
| RIP: 0033:0x7f0cbc297134
| RSP: 002b:00007ffc66125a38 EFLAGS: 00000246 ORIG_RAX: 0000000000000001
| RAX: ffffffffffffffda RBX: 0000000000000052 RCX: 00007f0cbc297134
| RDX: 0000000000000052 RSI: 00005639334fbf80 RDI: 0000000000000001
| RBP: 00005639334fbf80 R08: 0000000000000000 R09: 00007f0cbc564090
| R10: 0000000000000000 R11: 0000000000000246 R12: 00007f0cbc564760
| R13: 0000000000000052 R14: 00007f0cbc55f760 R15: 0000000000000052

Sebastian

  parent reply	other threads:[~2018-07-11  8:13 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-07-06  8:51 tty latency and RT André Pribil
2018-07-06  9:16 ` John Ogness
2018-07-09  9:55   ` Esben Haabendal
2018-07-09 11:37     ` André Pribil
2018-07-11  8:09     ` Sebastian Andrzej Siewior [this message]
2018-07-13  8:33       ` Esben Haabendal
2018-07-18 16:49         ` Sebastian Andrzej Siewior

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20180711080957.f6txdmzrrrrdm7ig@linutronix.de \
    --to=bigeasy@linutronix.de \
    --cc=a.pribil@beck-ipc.com \
    --cc=esben.haabendal@gmail.com \
    --cc=john.ogness@linutronix.de \
    --cc=linux-rt-users@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).