Signed-off-by: NArnd Bergmann <arnd@arndb.de>

This replaces the use of the BKL in the FAT family of filesystems with the
existing superblock lock instead.

The code already appears to do mostly proper locking with its own private
spinlocks (and mutexes), but while the BKL could possibly have been
dropped entirely, converting it to use the superblock lock (which is just
a regular mutex) is the conservative thing to do.

As a per-filesystem mutex, it not only won't have any of the possible
latency issues related to the BKL, but the lock is obviously private to
the particular filesystem instance and will thus not cause problems for
entirely unrelated users like the BKL can.
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
Cc: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Signed-off-by: NStephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

All in-kernel char device open() functions now either have their own
lock_kernel() calls or clearly do not need one.
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

I stared at this code for a while and almost deleted it before
understanding crept into my slow brain.  Hopefully this makes life easier
for the next person to happen on it.
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

All of the open() functions which don't need the BKL on their face may
still depend on its acquisition to serialize opens against driver
initialization.  So make those functions acquire then release the BKL to be
on the safe side.
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

A number of driver functions are so obviously trivial that they do not need
the big kernel lock - at least not overtly.  It turns out that the
acquisition of the BKL in driver open() functions can perform a sort of
poor-hacker's serialization function, delaying the open operation until the
driver is certain to have completed its initialization.  Add a simple
cycle_kernel_lock() function for these cases to make it clear that there is
no need to *hold* the BKL, just to be sure that we can acquire it.
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Put explicit lock_kernel() calls into videodev_open().  That function
itself seems OK, but one never knows about all the open() functions
provided by underlying video drivers.
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Add explicit lock_kernel() calls to usbdev_open()
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Add explicit lock_kernel() calls to mon_bin_open()
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Add explicit lock_kernel() calls to printer_open()
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Put explicit lock_kernel() calls in adb_open().  The fact that
adb_release() already has them suggests this is necessary.
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Put explicit lock_kernel() calls into maint_open().
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Put explicit lock_kernel() calls into capi_open()
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Add lock_kernel() calls to ch_open(), though the existing locking looks
adequate.
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Add lock_kernel() calls to adpt_open()
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Add explicit lock_kernel() calls to vfc_open().
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Put explicit lock_kernel() calls into bpp_open().  It has locking, but I'm
not convinced it won't race with ioctl().
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Add explicit lock_kernel calls to phantom_open().
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Be extra-cautious and protect the remaining open() functions.
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Put explicit lock_kernel calls into dsp56k_open().

[Stupid missing label error fixed]
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Put explicit lock_kernel() calls into raw_open(), even though the existing
locking looks adequate.
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Put explicit lock_kernel calls into tlclk_open()
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Add explicit lock_kernel() calls to hwicap_open() even though the existing
locking looks adequate.
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Add explicit BKL to vcs_open().
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Add the BKL to spidev_open(), even though the existing locking looks
adequate.
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

This documents the fact that somebody looked at the relevant open()
functions and concluded that, due to their trivial nature, no locking was
needed.
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Parts of the serial code actually BUG() if we don't do this.

Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Signed-off-by: NJonathan Corbet <corbet@lwn.net>

usb_open() is protected by a down_read(&minor_rwsem), but I'm not sure I
trust it to protect everything including subsidiary open() functions.
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Signed-off-by: NJonathan Corbet <corbet@lwn.net>

Signed-off-by: NJonathan Corbet <corbet@lwn.net>

->release() already has explicit lock_kernel() calls...
Signed-off-by: NJonathan Corbet <corbet@lwn.net>