On Fri, Jan 11, 2019 at 01:33:09PM -0500, Nicolas Pitre wrote:
> I use Linux with the help of a braille display and the brltty daemon. It
> turns out that the latest mainline kernel I can work with comes from
> commit 231f8fd0cc. Anything past that and I lose the ability to read the
> console barely a few seconds after the system has booted as brltty is
> thrown a wrench and the braille display becomes completely inoperable.
>
> Things get somewhat better with commit c96cf923a9 as brltty is not
> longer incapacitated, but some programs would randomly crash. Even the
> very first login attempt won't work as I soon as I hit enter after my
> user name the password prompt is skipped over, just like if the enter
> key had been hit twice. Then lynx (the text web browser) would crash as
> soon as I switch the virtual console with LeftAlt+FN. Mind you, this
> isn't easy to perform bisection in those conditions.
>
> And the worst commit i.e. 83d817f410 is marked for stable! :-(
>
> Some interaction with brltty must be at play here otherwise such
> breakage would never have survived up to the mainline kernel.
>
> As far as latest mainline is concerned, I managed to reproduce at least
> one of the unwelcome behavior change (hoping that's all there is to this
> issue) with a very simple test case so you won't have to learn braille
> to debug this:
>
> # from any vt, make sure tty40 is allocated and empty
> openvt -c 40 -f -- true
>
> # open it and wait on read()
> cat /dev/tty40
>
> # from a second vt, simply open tty40 again
> true < /dev/tty40
>
> # come back to the first vt and watch cat bailing out with EAGAIN.
>
> Please fix.
Please try the patch below, it was just queued up to my tree and should
resolve the issue. If not, please let us know.
thanks,
greg k-h
>From d3736d82e8169768218ee0ef68718875918091a0 Mon Sep 17 00:00:00 2001
From: Dmitry Safonov <d...@arista.com>
Date: Wed, 9 Jan 2019 01:17:40 +0000
Subject: tty: Don't hold ldisc lock in tty_reopen() if ldisc present
Try to get reference for ldisc during tty_reopen().
If ldisc present, we don't need to do tty_ldisc_reinit() and lock the
write side for line discipline semaphore.
Effectively, it optimizes fast-path for tty_reopen(), but more
importantly it won't interrupt ongoing IO on the tty as no ldisc change
is needed.
Fixes user-visible issue when tty_reopen() interrupted login process for
user with a long password, observed and reported by Lukas.
Fixes: c96cf923a98d ("tty: Don't block on IO when ldisc change is pending")
Fixes: 83d817f41070 ("tty: Hold tty_ldisc_lock() during tty_reopen()")
Cc: Jiri Slaby <jsl...@suse.com>
Reported-by: Lukas F. Hartmann <lu...@mntmn.com>
Tested-by: Lukas F. Hartmann <lu...@mntmn.com>
Cc: stable <sta...@vger.kernel.org>
Signed-off-by: Dmitry Safonov <d...@arista.com>
Signed-off-by: Greg Kroah-Hartman <gre...@linuxfoundation.org>
---
drivers/tty/tty_io.c | 20 +++++++++++++-------
1 file changed, 13 insertions(+), 7 deletions(-)
diff --git a/drivers/tty/tty_io.c b/drivers/tty/tty_io.c
index bfe9ad85b362..23c6fd238422 100644
--- a/drivers/tty/tty_io.c
+++ b/drivers/tty/tty_io.c
@@ -1256,7 +1256,8 @@ static void tty_driver_remove_tty(struct tty_driver
*driver, struct tty_struct *
static int tty_reopen(struct tty_struct *tty)
{
struct tty_driver *driver = tty->driver;
- int retval;
+ struct tty_ldisc *ld;
+ int retval = 0;
if (driver->type == TTY_DRIVER_TYPE_PTY &&
driver->subtype == PTY_TYPE_MASTER)
@@ -1268,13 +1269,18 @@ static int tty_reopen(struct tty_struct *tty)
if (test_bit(TTY_EXCLUSIVE, &tty->flags) && !capable(CAP_SYS_ADMIN))
return -EBUSY;
- retval = tty_ldisc_lock(tty, 5 * HZ);
- if (retval)
- return retval;
+ ld = tty_ldisc_ref_wait(tty);
+ if (ld) {
+ tty_ldisc_deref(ld);
+ } else {
+ retval = tty_ldisc_lock(tty, 5 * HZ);
+ if (retval)
+ return retval;
- if (!tty->ldisc)
- retval = tty_ldisc_reinit(tty, tty->termios.c_line);
- tty_ldisc_unlock(tty);
+ if (!tty->ldisc)
+ retval = tty_ldisc_reinit(tty, tty->termios.c_line);
+ tty_ldisc_unlock(tty);
+ }
if (retval == 0)
tty->count++;
--
2.20.1
