Skip to content

R
raspberrypi-kernel

openeuler / raspberrypi-kernel

通知 13
Star 1
Fork 0
R
raspberrypi-kernel
切换分支/标签
  1. 18 6月, 2013 1 次提交
  3. 21 5月, 2013 1 次提交
    • P
      tty/vt: Fix vc_deallocate() lock order · 421b40a6
      由 Peter Hurley 提交于 
      Now that the tty port owns the flip buffers and i/o is allowed
from the driver even when no tty is attached, the destruction
of the tty port (and the flip buffers) must ensure that no
outstanding work is pending.

Unfortunately, this creates a lock order problem with the
console_lock (see attached lockdep report [1] below).

For single console deallocation, drop the console_lock prior
to port destruction. When multiple console deallocation,
defer port destruction until the consoles have been
deallocated.

tty_port_destroy() is not required if the port has not
been used; remove from vc_allocate() failure path.

[1] lockdep report from Dave Jones <davej@redhat.com>

 ======================================================
 [ INFO: possible circular locking dependency detected ]
 3.9.0+ #16 Not tainted
 -------------------------------------------------------
 (agetty)/26163 is trying to acquire lock:
 blocked:  ((&buf->work)){+.+...}, instance: ffff88011c8b0020, at: [<ffffffff81062065>] flush_work+0x5/0x2e0

 but task is already holding lock:
 blocked:  (console_lock){+.+.+.}, instance: ffffffff81c2fde0, at: [<ffffffff813bc201>] vt_ioctl+0xb61/0x1230

 which lock already depends on the new lock.

 the existing dependency chain (in reverse order) is:

 -> #1 (console_lock){+.+.+.}:
        [<ffffffff810b3f74>] lock_acquire+0xa4/0x210
        [<ffffffff810416c7>] console_lock+0x77/0x80
        [<ffffffff813c3dcd>] con_flush_chars+0x2d/0x50
        [<ffffffff813b32b2>] n_tty_receive_buf+0x122/0x14d0
        [<ffffffff813b7709>] flush_to_ldisc+0x119/0x170
        [<ffffffff81064381>] process_one_work+0x211/0x700
        [<ffffffff8106498b>] worker_thread+0x11b/0x3a0
        [<ffffffff8106ce5d>] kthread+0xed/0x100
        [<ffffffff81601cac>] ret_from_fork+0x7c/0xb0

 -> #0 ((&buf->work)){+.+...}:
        [<ffffffff810b349a>] __lock_acquire+0x193a/0x1c00
        [<ffffffff810b3f74>] lock_acquire+0xa4/0x210
        [<ffffffff810620ae>] flush_work+0x4e/0x2e0
        [<ffffffff81065305>] __cancel_work_timer+0x95/0x130
        [<ffffffff810653b0>] cancel_work_sync+0x10/0x20
        [<ffffffff813b8212>] tty_port_destroy+0x12/0x20
        [<ffffffff813c65e8>] vc_deallocate+0xf8/0x110
        [<ffffffff813bc20c>] vt_ioctl+0xb6c/0x1230
        [<ffffffff813b01a5>] tty_ioctl+0x285/0xd50
        [<ffffffff811ba825>] do_vfs_ioctl+0x305/0x530
        [<ffffffff811baad1>] sys_ioctl+0x81/0xa0
        [<ffffffff81601d59>] system_call_fastpath+0x16/0x1b

 other info that might help us debug this:

 [ 6760.076175]  Possible unsafe locking scenario:

        CPU0                    CPU1
        ----                    ----
   lock(console_lock);
                                lock((&buf->work));
                                lock(console_lock);
   lock((&buf->work));

  *** DEADLOCK ***

 1 lock on stack by (agetty)/26163:
  #0: blocked:  (console_lock){+.+.+.}, instance: ffffffff81c2fde0, at: [<ffffffff813bc201>] vt_ioctl+0xb61/0x1230
 stack backtrace:
 Pid: 26163, comm: (agetty) Not tainted 3.9.0+ #16
 Call Trace:
  [<ffffffff815edb14>] print_circular_bug+0x200/0x20e
  [<ffffffff810b349a>] __lock_acquire+0x193a/0x1c00
  [<ffffffff8100a269>] ? sched_clock+0x9/0x10
  [<ffffffff8100a269>] ? sched_clock+0x9/0x10
  [<ffffffff8100a200>] ? native_sched_clock+0x20/0x80
  [<ffffffff810b3f74>] lock_acquire+0xa4/0x210
  [<ffffffff81062065>] ? flush_work+0x5/0x2e0
  [<ffffffff810620ae>] flush_work+0x4e/0x2e0
  [<ffffffff81062065>] ? flush_work+0x5/0x2e0
  [<ffffffff810b15db>] ? mark_held_locks+0xbb/0x140
  [<ffffffff8113c8a3>] ? __free_pages_ok.part.57+0x93/0xc0
  [<ffffffff810b15db>] ? mark_held_locks+0xbb/0x140
  [<ffffffff810652f2>] ? __cancel_work_timer+0x82/0x130
  [<ffffffff81065305>] __cancel_work_timer+0x95/0x130
  [<ffffffff810653b0>] cancel_work_sync+0x10/0x20
  [<ffffffff813b8212>] tty_port_destroy+0x12/0x20
  [<ffffffff813c65e8>] vc_deallocate+0xf8/0x110
  [<ffffffff813bc20c>] vt_ioctl+0xb6c/0x1230
  [<ffffffff810aec41>] ? lock_release_holdtime.part.30+0xa1/0x170
  [<ffffffff813b01a5>] tty_ioctl+0x285/0xd50
  [<ffffffff812b00f6>] ? inode_has_perm.isra.46.constprop.61+0x56/0x80
  [<ffffffff811ba825>] do_vfs_ioctl+0x305/0x530
  [<ffffffff812b04db>] ? selinux_file_ioctl+0x5b/0x110
  [<ffffffff811baad1>] sys_ioctl+0x81/0xa0
  [<ffffffff81601d59>] system_call_fastpath+0x16/0x1b

Cc: Dave Jones <davej@redhat.com>
Signed-off-by: NPeter Hurley <peter@hurleysoftware.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      421b40a6
  5. 22 11月, 2012 1 次提交
  7. 27 7月, 2012 1 次提交
    • R
      vt: fix race in vt_waitactive() · a7b12929
      由 Rabin Vincent 提交于 
      pm_restore_console() is called from the suspend/resume path, and this
calls vt_move_to_console(), which calls vt_waitactive().

There's a race in this path which causes the process which requests the
suspend to sleep indefinitely waiting for an event which already
happened:

P1                                      P2
 vt_move_to_console()
  set_console()
    schedule_console_callback()
  vt_waitactive()
    check n == fg_console +1
                                       console_callback()
                                         switch_screen()
                                         vt_event_post() // no waiters

    vt_event_wait() // forever

Fix the race by ensuring we're registered for the event before we check
if it's already completed.
Signed-off-by: NRabin Vincent <rabin.vincent@stericsson.com>
Acked-by: NAlan Cox <alan@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      a7b12929
  9. 15 6月, 2012 1 次提交
    • R
      vt: fix race in vt_waitactive() · 2fc46915
      由 Rabin Vincent 提交于 
      pm_restore_console() is called from the suspend/resume path, and this
calls vt_move_to_console(), which calls vt_waitactive().

There's a race in this path which causes the process which requests the
suspend to sleep indefinitely waiting for an event which already
happened:

P1                                      P2
 vt_move_to_console()
  set_console()
    schedule_console_callback()
  vt_waitactive()
    check n == fg_console +1
                                       console_callback()
                                         switch_screen()
                                         vt_event_post() // no waiters

    vt_event_wait() // forever

Fix the race by ensuring we're registered for the event before we check
if it's already completed.
Signed-off-by: NRabin Vincent <rabin.vincent@stericsson.com>
Acked-by: NAlan Cox <alan@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      2fc46915
  11. 25 4月, 2012 1 次提交
  13. 09 3月, 2012 4 次提交
  15. 25 2月, 2012 1 次提交
  17. 04 2月, 2012 1 次提交
  19. 03 2月, 2012 1 次提交
  21. 20 4月, 2011 4 次提交
  23. 02 3月, 2011 1 次提交
  25. 18 2月, 2011 3 次提交
  27. 26 1月, 2011 1 次提交
    • T
      console: rename acquire/release_console_sem() to console_lock/unlock() · ac751efa
      由 Torben Hohn 提交于 
      The -rt patches change the console_semaphore to console_mutex.  As a
result, a quite large chunk of the patches changes all
acquire/release_console_sem() to acquire/release_console_mutex()

This commit makes things use more neutral function names which dont make
implications about the underlying lock.

The only real change is the return value of console_trylock which is
inverted from try_acquire_console_sem()

This patch also paves the way to switching console_sem from a semaphore to
a mutex.

[akpm@linux-foundation.org: coding-style fixes]
[akpm@linux-foundation.org: make console_trylock return 1 on success, per Geert]
Signed-off-by: NTorben Hohn <torbenh@gmx.de>
Cc: Thomas Gleixner <tglx@tglx.de>
Cc: Greg KH <gregkh@suse.de>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
      ac751efa
  29. 05 11月, 2010 1 次提交
  31. 28 10月, 2010 1 次提交
  33. 29 8月, 2010 1 次提交
    • A
      Input: use PIT_TICK_RATE in vt beep ioctl · 2c4e9671
      由 Arnd Bergmann 提交于 
      The KIOCSOUND and KDMKTONE ioctls are based on the CLOCK_TICK_RATE,
which is architecture and sometimes configuration specific.

In practice, most user applications assume that it is actually defined
as the i8253 PIT base clock of 1193182 Hz, which is true on some
architectures but not on others.

This patch makes the vt code use the PIT frequency on all
architectures, which is much more well-defined.  It will change the
behavior of user applications sending the beep ioctl on all
architectures that define CLOCK_TICK_RATE different from
PIT_TICK_RATE.

The original breakage was introduced in commit bcc8ca09 "Adapt
drivers/char/vt_ioctl.c to non-x86".  Hopefully, reverting this change
will make the frequency correct in more cases than it will make it
incorrect.
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Acked-by: NAlan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>
      2c4e9671
  35. 11 8月, 2010 3 次提交
  37. 05 6月, 2010 1 次提交
  39. 19 3月, 2010 1 次提交
    • G
      Revert "tty: Add a new VT mode which is like VT_PROCESS but doesn't require a... · 87a6aca5
      由 Greg Kroah-Hartman 提交于 
      Revert "tty: Add a new VT mode which is like VT_PROCESS but doesn't require a VT_RELDISP ioctl call"

This reverts commit eec9fe7d.

Ari writes as the reason this should be reverted:
	The problems with this patch include:
	1. There's at least one subtlety I overlooked - switching
	between X servers (i.e. from one X VT to another) still requires
	the cooperation of both X servers. I was assuming that KMS
	eliminated this.
	2. It hasn't been tested at all (no X server patch exists which
	uses the new mode).

As he was the original author of the patch, I'll revert it.

Cc: Ari Entlich <atrigent@ccs.neu.edu>
Cc: Alan Cox <alan@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>
      87a6aca5
  41. 03 3月, 2010 1 次提交
    • A
      tty: Add a new VT mode which is like VT_PROCESS but doesn't require a VT_RELDISP ioctl call · eec9fe7d
      由 Ari Entlich 提交于 
      This new VT mode (VT_PROCESS_AUTO) does everything that VT_PROCESS does
except that it doesn't wait for a VT_RELDISP ioctl before switching
away from a VT with that mode.

If the X server eventually uses this new mode, debugging and crash
recovery should become easier. This is because even when currently in
the VT of a frozen X server it would still be possible to switch out
by doing SysRq-r and then CTRL-<number of a text vt>, sshing in and
doing chvt <number of a text vt>, or any other method of VT switching.
The general concensus on #xorg-devel seems to be that it should be
safe to use this with X now that we have KMS.

This also moves the VT_ACKACQ define to a more appropriate place,
for clarity's sake.
Signed-off-by: NAri Entlich <atrigent@ccs.neu.edu>
Acked-by: NAlan Cox <alan@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>
      eec9fe7d
  43. 20 11月, 2009 1 次提交
  45. 15 10月, 2009 1 次提交
  47. 02 10月, 2009 1 次提交
  49. 20 9月, 2009 5 次提交
  51. 13 7月, 2009 1 次提交