releases/4.6.1/tty-n_gsm-fix-false-positive-warn_on.patch

   1 From d175feca89a1c162f60f4e3560ca7bc9437c65eb Mon Sep 17 00:00:00 2001
   2 From: Jiri Slaby <jslaby@suse.cz>
   3 Date: Tue, 22 Mar 2016 18:09:51 +0100
   4 Subject: TTY: n_gsm, fix false positive WARN_ON
   5
   6 From: Jiri Slaby <jslaby@suse.cz>
   7
   8 commit d175feca89a1c162f60f4e3560ca7bc9437c65eb upstream.
   9
  10 Dmitry reported, that the current cleanup code in n_gsm can trigger a
  11 warning:
  12 WARNING: CPU: 2 PID: 24238 at drivers/tty/n_gsm.c:2048 gsm_cleanup_mux+0x166/0x6b0()
  13 ...
  14 Call Trace:
  15 ...
  16  [<ffffffff81247ab9>] warn_slowpath_null+0x29/0x30 kernel/panic.c:490
  17  [<ffffffff828d0456>] gsm_cleanup_mux+0x166/0x6b0 drivers/tty/n_gsm.c:2048
  18  [<ffffffff828d4d87>] gsmld_open+0x5b7/0x7a0 drivers/tty/n_gsm.c:2386
  19  [<ffffffff828b9078>] tty_ldisc_open.isra.2+0x78/0xd0 drivers/tty/tty_ldisc.c:447
  20  [<ffffffff828b973a>] tty_set_ldisc+0x1ca/0xa70 drivers/tty/tty_ldisc.c:567
  21  [<     inline     >] tiocsetd drivers/tty/tty_io.c:2650
  22  [<ffffffff828a14ea>] tty_ioctl+0xb2a/0x2140 drivers/tty/tty_io.c:2883
  23 ...
  24
  25 But this is a legal path when open fails to find a space in the
  26 gsm_mux array and tries to clean up. So make it a standard test
  27 instead of a warning.
  28
  29 Reported-by: "Dmitry Vyukov" <dvyukov@google.com>
  30 Cc: Alan Cox <alan@linux.intel.com>
  31 Link: http://lkml.kernel.org/r/CACT4Y+bHQbAB68VFi7Romcs-Z9ZW3kQRvcq+BvHH1oa5NcAdLA@mail.gmail.com
  32 Fixes: 5a640967 ("tty/n_gsm.c: fix a memory leak in gsmld_open()")
  33 Signed-off-by: Jiri Slaby <jslaby@suse.cz>
  34 Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
  35
  36 ---
  37  drivers/tty/n_gsm.c |    4 +++-
  38  1 file changed, 3 insertions(+), 1 deletion(-)
  39
  40 --- a/drivers/tty/n_gsm.c
  41 +++ b/drivers/tty/n_gsm.c
  42 @@ -2045,7 +2045,9 @@ static void gsm_cleanup_mux(struct gsm_m
  43                 }
  44         }
  45         spin_unlock(&gsm_mux_lock);
  46 -       WARN_ON(i == MAX_MUX);
  47 +       /* open failed before registering => nothing to do */
  48 +       if (i == MAX_MUX)
  49 +               return;
  50
  51         /* In theory disconnecting DLCI 0 is sufficient but for some
  52            modems this is apparently not the case. */