linux-stable.git/drivers/tty, branch v4.14.41

earlycon: Use a pointer table to fix __earlycon_table stride

2018-05-01T19:58:24+00:00

commit dd709e72cb934eefd44de8d9969097173fbf45dc upstream.

Commit 99492c39f39f ("earlycon: Fix __earlycon_table stride") tried to fix
__earlycon_table stride by forcing the earlycon_id struct alignment to 32
and asking the linker to 32-byte align the __earlycon_table symbol.  This
fix was based on commit 07fca0e57fca92 ("tracing: Properly align linker
defined symbols") which tried a similar fix for the tracing subsystem.

However, this fix doesn't quite work because there is no guarantee that
gcc will place structures packed into an array format.  In fact, gcc 4.9
chooses to 64-byte align these structs by inserting additional padding
between the entries because it has no clue that they are supposed to be in
an array.  If we are unlucky, the linker will assign symbol
"__earlycon_table" to a 32-byte aligned address which does not correspond
to the 64-byte aligned contents of section "__earlycon_table".

To address this same problem, the fix to the tracing system was
subsequently re-implemented using a more robust table of pointers approach
by commits:
 3d56e331b653 ("tracing: Replace syscall_meta_data struct array with pointer array")
 654986462939 ("tracepoints: Fix section alignment using pointer array")
 e4a9ea5ee7c8 ("tracing: Replace trace_event struct array with pointer array")

Let's use this same "array of pointers to structs" approach for
EARLYCON_TABLE.

Fixes: 99492c39f39f ("earlycon: Fix __earlycon_table stride")
Signed-off-by: Daniel Kurtz 
Suggested-by: Aaron Durbin 
Reviewed-by: Rob Herring 
Tested-by: Guenter Roeck 
Reviewed-by: Guenter Roeck 
Cc: stable 
Signed-off-by: Greg Kroah-Hartman

tty: Use __GFP_NOFAIL for tty_ldisc_get()

2018-05-01T19:58:14+00:00

commit bcdd0ca8cb8730573afebcaae4138f8f4c8eaa20 upstream.

syzbot is reporting crashes triggered by memory allocation fault injection
at tty_ldisc_get() [1]. As an attempt to handle OOM in a graceful way, we
have tried commit 5362544bebe85071 ("tty: don't panic on OOM in
tty_set_ldisc()"). But we reverted that attempt by commit a8983d01f9b7d600
("Revert "tty: don't panic on OOM in tty_set_ldisc()"") due to reproducible
crash. We should spend resource for finding and fixing race condition bugs
rather than complicate error paths for 2 * sizeof(void *) bytes allocation
failure.

[1] https://syzkaller.appspot.com/bug?id=489d33fa386453859ead58ff5171d43772b13aa3

Signed-off-by: Tetsuo Handa 
Reported-by: syzbot 
Cc: Michal Hocko 
Cc: Vegard Nossum 
Cc: Dmitry Vyukov 
Cc: Jiri Slaby 
Cc: Peter Hurley 
Cc: One Thousand Gnomes 
Cc: Linus Torvalds 
Cc: stable 
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Greg Kroah-Hartman

tty: Avoid possible error pointer dereference at tty_ldisc_restore().

2018-05-01T19:58:13+00:00

commit 598c2d41ff44889dd8eced4f117403e472158d85 upstream.

syzbot is reporting crashes [1] triggered by memory allocation failure at
tty_ldisc_get() from tty_ldisc_restore(). While syzbot stops at WARN_ON()
due to panic_on_warn == true, panic_on_warn == false will after all trigger
an OOPS by dereferencing old->ops->num if IS_ERR(old) == true.

We can simplify tty_ldisc_restore() as three calls (old->ops->num, N_TTY,
N_NULL) to tty_ldisc_failto() in addition to avoiding possible error
pointer dereference.

If someone reports kernel panic triggered by forcing all memory allocations
for tty_ldisc_restore() to fail, we can consider adding __GFP_NOFAIL for
tty_ldisc_restore() case.

[1] https://syzkaller.appspot.com/bug?id=6ac359c61e71d22e06db7f8f88243feb11d927e7

Reported-by: syzbot+40b7287c2dc987c48c81@syzkaller.appspotmail.com
Signed-off-by: Tetsuo Handa 
Cc: Greg Kroah-Hartman 
Cc: Jiri Slaby 
Cc: Dmitry Vyukov 
Cc: Johannes Weiner 
Cc: Alan Cox 
Cc: Christoph Hellwig 
Cc: Michal Hocko 
Cc: stable 
Signed-off-by: Greg Kroah-Hartman

tty: n_gsm: Fix DLCI handling for ADM mode if debug & 2 is not set

2018-05-01T19:58:13+00:00

commit b2d89ad9c9682e795ed6eeb9ed455789ad6cedf1 upstream.

At least on droid 4 with control channel in ADM mode, there is no response
to Modem Status Command (MSC). Currently gsmtty_modem_update() expects to
have data in dlci->modem_rx unless debug & 2 is set. This means that on
droid 4, things only work if debug & 2 is set.

Let's fix the issue by ignoring empty dlci->modem_rx for ADM mode. In
the AMD mode, CMD_MSC will never respond and gsm_process_modem() won't
get called to set dlci->modem_rx.

And according to ts_127010v140000p.pdf, MSC is only relevant if basic
option is chosen, so let's test for that too.

Fixes: ea3d8465ab9b ("tty: n_gsm: Allow ADM response in addition to UA for control dlci")
Cc: linux-serial@vger.kernel.org
Cc: Alan Cox 
Cc: Dan Williams 
Cc: Jiri Prchal 
Cc: Jiri Slaby 
Cc: Marcel Partap 
Cc: Merlijn Wajer 
Cc: Michael Nazzareno Trimarchi 
Cc: Michael Scott 
Cc: Pavel Machek 
Cc: Peter Hurley 
Cc: Russ Gorby 
Cc: Sascha Hauer 
Cc: Sebastian Reichel 
Signed-off-by: Tony Lindgren 
Cc: stable 
Signed-off-by: Greg Kroah-Hartman

tty: n_gsm: Fix long delays with control frame timeouts in ADM mode

2018-05-01T19:58:13+00:00

commit e9ec22547986dd32c5c70da78107ce35dbff1344 upstream.

Commit ea3d8465ab9b ("tty: n_gsm: Allow ADM response in addition to UA for
control dlci") added support for DLCI to stay in Asynchronous Disconnected
Mode (ADM). But we still get long delays waiting for commands to other
DLCI to complete:

--> 5) C: SABM(P)
Q>  0) C: UIH(F)
Q>  0) C: UIH(F)
Q>  0) C: UIH(F)
...

This happens because gsm_control_send() sets cretries timer to T2 that is
by default set to 34. This will cause resend for T2 times for the control
frame. In ADM mode, we will never get a response so the control frame, so
retries are just delaying all the commands.

Let's fix the issue by setting DLCI_MODE_ADM flag after detecting the ADM
mode for the control DLCI. Then we can use that in gsm_control_send() to
set retries to 1. This means the control frame will be sent once allowing
the other end at an opportunity to switch from ADM to ABM mode.

Note that retries will be decremented in gsm_control_retransmit() so
we don't want to set it to 0 here.

Fixes: ea3d8465ab9b ("tty: n_gsm: Allow ADM response in addition to UA for control dlci")
Cc: linux-serial@vger.kernel.org
Cc: Alan Cox 
Cc: Dan Williams 
Cc: Jiri Prchal 
Cc: Jiri Slaby 
Cc: Marcel Partap 
Cc: Merlijn Wajer 
Cc: Michael Nazzareno Trimarchi 
Cc: Michael Scott 
Cc: Pavel Machek 
Cc: Peter Hurley 
Cc: Russ Gorby 
Cc: Sascha Hauer 
Cc: Sebastian Reichel 
Signed-off-by: Tony Lindgren 
Cc: stable 
Signed-off-by: Greg Kroah-Hartman

tty: Don't call panic() at tty_ldisc_init()

2018-05-01T19:58:13+00:00

commit 903f9db10f18f735e62ba447147b6c434b6af003 upstream.

syzbot is reporting kernel panic [1] triggered by memory allocation failure
at tty_ldisc_get() from tty_ldisc_init(). But since both tty_ldisc_get()
and caller of tty_ldisc_init() can cleanly handle errors, tty_ldisc_init()
does not need to call panic() when tty_ldisc_get() failed.

[1] https://syzkaller.appspot.com/bug?id=883431818e036ae6a9981156a64b821110f39187

Signed-off-by: Tetsuo Handa 
Reported-by: syzbot 
Cc: Greg Kroah-Hartman 
Cc: Jiri Slaby 
Cc: stable 
Signed-off-by: Greg Kroah-Hartman

tty: serial: exar: Relocate sleep wake-up handling

2018-04-26T09:02:11+00:00

[ Upstream commit c7e1b4059075c9e8eed101d7cc5da43e95eb5e18 ]

Exar sleep wake-up handling has been done on a per-channel basis by
virtue of INT0 being accessible from each channel's address space. I
believe this was initially done out of necessity, but now that Exar
devices have their own driver, we can do things more efficiently by
registering a dedicated INT0 handler at the PCI device level.

I see this change providing the following benefits:

    1. If more than one port is active, eliminates the redundant bus
       cycles for reading INT0 on every interrupt.
    2. This note associated with hooking in the per-channel handler in
       8250_port.c is resolved:
        /* Fixme: probably not the best place for this */

Cc: Matt Schulte 
Signed-off-by: Aaron Sierra 
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

tty: make n_tty_read() always abort if hangup is in progress

2018-04-24T07:36:21+00:00

commit 28b0f8a6962a24ed21737578f3b1b07424635c9e upstream.

A tty is hung up by __tty_hangup() setting file->f_op to
hung_up_tty_fops, which is skipped on ttys whose write operation isn't
tty_write().  This means that, for example, /dev/console whose write
op is redirected_tty_write() is never actually marked hung up.

Because n_tty_read() uses the hung up status to decide whether to
abort the waiting readers, the lack of hung-up marking can lead to the
following scenario.

 1. A session contains two processes.  The leader and its child.  The
    child ignores SIGHUP.

 2. The leader exits and starts disassociating from the controlling
    terminal (/dev/console).

 3. __tty_hangup() skips setting f_op to hung_up_tty_fops.

 4. SIGHUP is delivered and ignored.

 5. tty_ldisc_hangup() is invoked.  It wakes up the waits which should
    clear the read lockers of tty->ldisc_sem.

 6. The reader wakes up but because tty_hung_up_p() is false, it
    doesn't abort and goes back to sleep while read-holding
    tty->ldisc_sem.

 7. The leader progresses to tty_ldisc_lock() in tty_ldisc_hangup()
    and is now stuck in D sleep indefinitely waiting for
    tty->ldisc_sem.

The following is Alan's explanation on why some ttys aren't hung up.

 http://lkml.kernel.org/r/20171101170908.6ad08580@alans-desktop

 1. It broke the serial consoles because they would hang up and close
    down the hardware. With tty_port that *should* be fixable properly
    for any cases remaining.

 2. The console layer was (and still is) completely broken and doens't
    refcount properly. So if you turn on console hangups it breaks (as
    indeed does freeing consoles and half a dozen other things).

As neither can be fixed quickly, this patch works around the problem
by introducing a new flag, TTY_HUPPING, which is used solely to tell
n_tty_read() that hang-up is in progress for the console and the
readers should be aborted regardless of the hung-up status of the
device.

The following is a sample hung task warning caused by this issue.

  INFO: task agetty:2662 blocked for more than 120 seconds.
        Not tainted 4.11.3-dbg-tty-lockup-02478-gfd6c7ee-dirty #28
  "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
      0  2662      1 0x00000086
  Call Trace:
   __schedule+0x267/0x890
   schedule+0x36/0x80
   schedule_timeout+0x23c/0x2e0
   ldsem_down_write+0xce/0x1f6
   tty_ldisc_lock+0x16/0x30
   tty_ldisc_hangup+0xb3/0x1b0
   __tty_hangup+0x300/0x410
   disassociate_ctty+0x6c/0x290
   do_exit+0x7ef/0xb00
   do_group_exit+0x3f/0xa0
   get_signal+0x1b3/0x5d0
   do_signal+0x28/0x660
   exit_to_usermode_loop+0x46/0x86
   do_syscall_64+0x9c/0xb0
   entry_SYSCALL64_slow_path+0x25/0x25

The following is the repro.  Run "$PROG /dev/console".  The parent
process hangs in D state.

  #include 
  #include 
  #include 
  #include 
  #include 
  #include 
  #include 
  #include 
  #include 
  #include 
  #include 
  #include 

  int main(int argc, char **argv)
  {
	  struct sigaction sact = { .sa_handler = SIG_IGN };
	  struct timespec ts1s = { .tv_sec = 1 };
	  pid_t pid;
	  int fd;

	  if (argc < 2) {
		  fprintf(stderr, "test-hung-tty /dev/$TTY\n");
		  return 1;
	  }

	  /* fork a child to ensure that it isn't already the session leader */
	  pid = fork();
	  if (pid < 0) {
		  perror("fork");
		  return 1;
	  }

	  if (pid > 0) {
		  /* top parent, wait for everyone */
		  while (waitpid(-1, NULL, 0) >= 0)
			  ;
		  if (errno != ECHILD)
			  perror("waitpid");
		  return 0;
	  }

	  /* new session, start a new session and set the controlling tty */
	  if (setsid() < 0) {
		  perror("setsid");
		  return 1;
	  }

	  fd = open(argv[1], O_RDWR);
	  if (fd < 0) {
		  perror("open");
		  return 1;
	  }

	  if (ioctl(fd, TIOCSCTTY, 1) < 0) {
		  perror("ioctl");
		  return 1;
	  }

	  /* fork a child, sleep a bit and exit */
	  pid = fork();
	  if (pid < 0) {
		  perror("fork");
		  return 1;
	  }

	  if (pid > 0) {
		  nanosleep(&ts1s, NULL);
		  printf("Session leader exiting\n");
		  exit(0);
	  }

	  /*
	   * The child ignores SIGHUP and keeps reading from the controlling
	   * tty.  Because SIGHUP is ignored, the child doesn't get killed on
	   * parent exit and the bug in n_tty makes the read(2) block the
	   * parent's control terminal hangup attempt.  The parent ends up in
	   * D sleep until the child is explicitly killed.
	   */
	  sigaction(SIGHUP, &sact, NULL);
	  printf("Child reading tty\n");
	  while (1) {
		  char buf[1024];

		  if (read(fd, buf, sizeof(buf)) < 0) {
			  perror("read");
			  return 1;
		  }
	  }

	  return 0;
  }

Signed-off-by: Tejun Heo 
Cc: Alan Cox 
Cc: stable@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman

tty: n_gsm: Allow ADM response in addition to UA for control dlci

2018-04-12T10:32:19+00:00

[ Upstream commit ea3d8465ab9b3e01be329ac5195970a84bef76c5 ]

Some devices have the control dlci stay in ADM mode instead of the UA
mode. This can seen at least on droid 4 when trying to open the ts
27.010 mux port. Enabling n_gsm debug mode shows the control dlci
always respond with DM to SABM instead of UA:

# modprobe n_gsm debug=0xff
# ldattach -d GSM0710 /dev/ttyS0 &
gsmld_output: 00000000: f9 03 3f 01 1c f9
--> 0) C: SABM(P)
gsmld_receive: 00000000: f9 03 1f 01 36 f9
<-- 0) C: DM(P)
...
$ minicom -D /dev/gsmtty1
minicom: cannot open /dev/gsmtty1: No error information
$ strace minicom -D /dev/gsmtty1
...
open("/dev/gsmtty1", O_RDWR|O_NOCTTY|O_NONBLOCK|O_LARGEFILE) = -1 EL2HLT

Note that this is different issue from other n_gsm -EL2HLT issues such
as timeouts when the control dlci does not respond at all.

The ADM mode seems to be a quite common according to "RF Wireless World"
article "GSM Issue-UE sends SABM and gets a DM response instead of
UA response":

  This issue is most commonly observed in GSM networks where in UE sends
  SABM and expects network to send UA response but it ends up receiving
  DM response from the network. SABM stands for Set asynchronous balanced
  mode, UA stands for Unnumbered Acknowledge and DA stands for
  Disconnected Mode.

  An RLP entity can be in one of two modes:
  - Asynchronous Balanced Mode (ABM)
  - Asynchronous Disconnected Mode (ADM)

Currently Linux kernel closes the control dlci after several retries
in gsm_dlci_t1() on DM. This causes n_gsm /dev/gsmtty ports to produce
error code -EL2HLT when trying to open them as the closing of control
dlci has already set gsm->dead.

Let's fix the issue by allowing control dlci stay in ADM mode after the
retries so the /dev/gsmtty ports can be opened and used. It seems that
it might take several attempts to get any response from the control
dlci, so it's best to allow ADM mode only after the SABM retries are
done.

Note that for droid 4 additional patches are needed to mux the ttyS0
pins and to toggle RTS gpio_149 to wake up the mdm6600 modem are also
needed to use n_gsm. And the mdm6600 modem needs to be powered on.

Cc: linux-serial@vger.kernel.org
Cc: Alan Cox 
Cc: Jiri Prchal 
Cc: Jiri Slaby 
Cc: Marcel Partap 
Cc: Michael Scott 
Cc: Peter Hurley 
Cc: Russ Gorby 
Cc: Sascha Hauer 
Cc: Sebastian Reichel 
Signed-off-by: Tony Lindgren 
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

vt: change SGR 21 to follow the standards

2018-04-08T12:26:33+00:00

commit 65d9982d7e523a1a8e7c9af012da0d166f72fc56 upstream.

ECMA-48 [1] (aka ISO 6429) has defined SGR 21 as "doubly underlined"
since at least March 1984.  The Linux kernel has treated it as SGR 22
"normal intensity" since it was added in Linux-0.96b in June 1992.
Before that, it was simply ignored.  Other terminal emulators have
either ignored it, or treat it as double underline now.  xterm for
example added support in its 304 release (May 2014) [2] where it was
previously ignoring it.

Changing this behavior shouldn't be an issue:
- It isn't a named capability in ncurses's terminfo database, so no
  script is using libtinfo/libcurses to look this up, or using tput
  to query & output the right sequence.
- Any script assuming SGR 21 will reset intensity in all terminals
  already do not work correctly on non-Linux VTs (including running
  under screen/tmux/etc...).
- If someone has written a script that only runs in the Linux VT, and
  they're using SGR 21 (instead of SGR 22), the output should still
  be readable.

imo it's important to change this as the Linux VT's non-conformance
is sometimes used as an argument for other terminal emulators to not
implement SGR 21 at all, or do so incorrectly.

[1]: https://www.ecma-international.org/publications/standards/Ecma-048.htm
[2]: https://github.com/ThomasDickey/xterm-snapshots/commit/2fd29cb98d214cb536bcafbee00bc73b3f1eeb9d

Signed-off-by: Mike Frysinger 
Cc: stable 
Signed-off-by: Greg Kroah-Hartman