My previous commit left the file empty and present in the
Makefile, which is a bit dirty and caused problems with
'make distclean', as pointed out by David Howells.
This hopefully cleans it up the right way.
Signed-off-by: Arnd Bergmann <[email protected]>
Acked-by: David Howells <[email protected]>
Acked-by: John Linn <[email protected]>
The zynq platform will never have board files other than the
device tree one, so there is no point splitting it from common.c.
This makes the code more compact.
Signed-off-by: Arnd Bergmann <[email protected]>
Acked-by: John Linn <[email protected]>
Lockdep found a locking inconsistency in the mkiss_close function:
> kernel: [ INFO: inconsistent lock state ]
> kernel: 2.6.39.1 #3
> kernel: ---------------------------------
> kernel: inconsistent {IN-SOFTIRQ-R} -> {SOFTIRQ-ON-W} usage.
> kernel: ax25ipd/2813 [HC0[0]:SC0[0]:HE1:SE1] takes:
> kernel: (disc_data_lock){+++?.-}, at: [<ffffffffa018552b>] mkiss_close+0x1b/0x90 [mkiss]
> kernel: {IN-SOFTIRQ-R} state was registered at:
The message hints that disc_data_lock is aquired with softirqs disabled,
but does not itself disable softirqs, which can in rare circumstances
lead to a deadlock.
The same problem is present in the 6pack driver, this patch fixes both
by using write_lock_bh instead of write_lock.
Reported-by: Bernard F6BVP <[email protected]>
Tested-by: Bernard F6BVP <[email protected]>
Signed-off-by: Arnd Bergmann <[email protected]>
Acked-by: Ralf Baechle<[email protected]>
Cc: [email protected]
Signed-off-by: David S. Miller <[email protected]>
board-generic.c now contains a reference to omap3_timer, but depends
only on ARCH_OMAP2, not on ARCH_OMAP3, which controls that symbol.
omap2_timer seems to be more appropriate anyway, so use that instead.
Signed-off-by: Arnd Bergmann <[email protected]>
Acked-by: Tony Lindgren <[email protected]>
GDB's interrupt.exp test cases currenly fail on ARM. The problem is how do_signal
handled restarting interrupted system calls:
The entry.S assembler code determines that we come from a system call; and that
information is passed as "syscall" parameter to do_signal. That routine then
calls get_signal_to_deliver [*] and if a signal is to be delivered, calls into
handle_signal. If a system call is to be restarted either after the signal
handler returns, or if no handler is to be called in the first place, the PC
is updated after the get_signal_to_deliver call, either in handle_signal (if
we have a handler) or at the end of do_signal (otherwise).
Now the problem is that during [*], the call to get_signal_to_deliver, a ptrace
intercept may happen. During this intercept, the debugger may change registers,
including the PC. This is done by GDB if it wants to execute an "inferior call",
i.e. the execution of some code in the debugged program triggered by GDB.
To this purpose, GDB will save all registers, allocate a stack frame, set up
PC and arguments as appropriate for the call, and point the link register to
a dummy breakpoint instruction. Once the process is restarted, it will execute
the call and then trap back to the debugger, at which point GDB will restore
all registers and continue original execution.
This generally works fine. However, now consider what happens when GDB attempts
to do exactly that while the process was interrupted during execution of a to-be-
restarted system call: do_signal is called with the syscall flag set; it calls
get_signal_to_deliver, at which point the debugger takes over and changes the PC
to point to a completely different place. Now get_signal_to_deliver returns
without a signal to deliver; but now do_signal decides it should be restarting
a system call, and decrements the PC by 2 or 4 -- so it now points to 2 or 4
bytes before the function GDB wants to call -- which leads to a subsequent crash.
To fix this problem, two things need to be supported:
- do_signal must be able to recognize that get_signal_to_deliver changed the PC
to a different location, and skip the restart-syscall sequence
- once the debugger has restored all registers at the end of the inferior call
sequence, do_signal must recognize that *now* it needs to restart the pending
system call, even though it was now entered from a breakpoint instead of an
actual svc instruction
This set of issues is solved on other platforms, usually by one of two
mechanisms:
- The status information "do_signal is handling a system call that may need
restarting" is itself carried in some register that can be accessed via
ptrace. This is e.g. on Intel the "orig_eax" register; on Sparc the kernel
defines a magic extra bit in the flags register for this purpose.
This allows GDB to manage that state: reset it when doing an inferior call,
and restore it after the call is finished.
- On s390, do_signal transparently handles this problem without requiring
GDB interaction, by performing system call restarting in the following
way: first, adjust the PC as necessary for restarting the call. Then,
call get_signal_to_deliver; and finally just continue execution at the
PC. This way, if GDB does not change the PC, everything is as before.
If GDB *does* change the PC, execution will simply continue there --
and once GDB restores the PC it saved at that point, it will automatically
point to the *restarted* system call. (There is the minor twist how to
handle system calls that do *not* need restarting -- do_signal will undo
the PC change in this case, after get_signal_to_deliver has returned, and
only if ptrace did not change the PC during that call.)
Because there does not appear to be any obvious register to carry the
syscall-restart information on ARM, we'd either have to introduce a new
artificial ptrace register just for that purpose, or else handle the issue
transparently like on s390. The patch below implements the second option;
using this patch makes the interrupt.exp test cases pass on ARM, with no
regression in the GDB test suite otherwise.
Cc: [email protected]
Signed-off-by: Ulrich Weigand <[email protected]>
Signed-off-by: Arnd Bergmann <[email protected]>
Signed-off-by: Russell King <[email protected]>
The BKL removal in appletalk introduced a use-after-free problem,
where atalk_destroy_socket frees a sock, but we still release
the socket lock on it.
An easy fix is to take an extra reference on the sock and sock_put
it when returning from atalk_release.
Signed-off-by: Arnd Bergmann <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Out of the three files accessing the device list,
one uses a mutex, one uses the BKL and one does
not have any locking. That is of course pointless,
so let's make all of them use the same mutex,
and get rid of one more BKL user.
Signed-off-by: Arnd Bergmann <[email protected]>
Signed-off-by: Palash Bandyopadhyay <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The file is required by the specialix driver, which was
moved to drivers/staging/.
Signed-off-by: Arnd Bergmann <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This replaces all instances of lock_kernel in the
IPX code with lock_sock. As far as I can tell, this
is safe to do, because there is no global state
that needs to be locked in IPX, and the code does
not recursively take the lock or sleep indefinitely
while holding it.
Compile-tested only.
Signed-off-by: Arnd Bergmann <[email protected]>
Acked-by: David S. Miller <[email protected]>
Cc: Arnaldo Carvalho de Melo <[email protected]>
Cc: [email protected]
This changes appletalk to use lock_sock instead of
lock_kernel for serialization. I tried to make sure
that we don't hold the socket lock during sleeping
functions, but I did not try to prove whether the
locks are necessary in the first place.
Compile-tested only.
Signed-off-by: Arnd Bergmann <[email protected]>
Acked-by: David S. Miller <[email protected]>
Cc: Arnaldo Carvalho de Melo <[email protected]>
Cc: David Miller <[email protected]>
Cc: [email protected]
This replaces all instances of lock_kernel in x25
with lock_sock, taking care to release the socket
lock around sleeping functions (sock_alloc_send_skb
and skb_recv_datagram). It is not clear whether
this is a correct solution, but it seem to be what
other protocols do in the same situation.
Includes a fix suggested by Eric Dumazet.
Signed-off-by: Arnd Bergmann <[email protected]>
Acked-by: David S. Miller <[email protected]>
Tested-by: Andrew Hendry <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: Eric Dumazet <[email protected]>
This introduces a new per-superblock mutex in UFS to replace
the big kernel lock. I have been careful to avoid nested
calls to lock_ufs and to get the lock order right with
respect to other mutexes, in particular lock_super.
I did not make any attempt to prove that the big kernel
lock is not needed in a particular place in the code,
which is very possible.
The mutex has a significant performance impact, so it is only
used on SMP or PREEMPT configurations.
As Nick Piggin noticed, any allocation inside of the lock
may end up deadlocking when we get to ufs_getfrag_block
in the reclaim task, so we now use GFP_NOFS.
Signed-off-by: Arnd Bergmann <[email protected]>
Tested-by: Nick Bowler <[email protected]>
Cc: Evgeniy Dushistov <[email protected]>
Cc: Nick Piggin <[email protected]>
This removes the BKL in hpfs in a rather awful
way, by making the code only work on uniprocessor
systems without kernel preemption, as suggested
by Andi Kleen.
The HPFS code probably has close to zero remaining
users on current kernels, all archeological uses of
the file system can probably be done with the significant
restrictions.
The hpfs_lock/hpfs_unlock functions are left in the
code, sincen Mikulas has indicated that he is still
interested in fixing it in a better way.
Signed-off-by: Arnd Bergmann <[email protected]>
Acked-by: Andi Kleen <[email protected]>
Cc: Mikulas Patocka <[email protected]>
Cc: [email protected]
usbip has its own infrastructure for managing kernel
threads, similar to kthread. By changing it to use
the standard functions, we can simplify the code
and get rid of one of the last BKL users at the
same time.
Includes changes suggested by Max Vozeler.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: Greg Kroah-Hartman <[email protected]>
Cc: Takahiro Hirofuchi <[email protected]>
Cc: Max Vozeler <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
These were missed the last time I cleaned this up
globally, because of code moving around or new code
getting merged.
Signed-off-by: Arnd Bergmann <[email protected]>
According to Russell King, adfs was written to not require the big
kernel lock, and all inode updates are done under adfs_dir_lock.
All other metadata in adfs is read-only and does not require locking.
The use of the BKL is the result of various pushdowns from the VFS
operations.
Signed-off-by: Arnd Bergmann <[email protected]>
Acked-by: Russell King <[email protected]>
Cc: Stuart Swales <[email protected]>
SMP i810 systems were practically nonexistent and the configuration
was not officially supported by Intel at the time when Pentium-III
was common.
With this change, it is still possible to build a distribution kernel
that has support for SMP and includes the i810 driver without the BKL.
As a precaution, check for the theoretical SMP case at run time and
refuse to load the driver.
We also need to disable CONFIG_PREEMPT builds for this driver.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: [email protected]
Signed-off-by: Dave Airlie <[email protected]>
This driver is one of the last users of the big kernel
lock, which is going away. All the hardware supported
by this driver also works with the newer i915 driver,
and recent X.org releases only work with that driver
anyway.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: Chris Wilson <[email protected]>
Cc: [email protected]
Signed-off-by: Dave Airlie <[email protected]>
For all I know, Appletalk is dead, the only reasonable
use right now would be nostalgia, and that can be served
well enough by old kernels. The code is largely not
in a bad shape, but it still uses the big kernel lock,
and nobody seems motivated to change that.
FWIW, the last release of MacOS that supported Appletalk
was MacOS X 10.5, made in 2007, and it has been abandoned
by Apple with 10.6. Using TCP/IP instead of Appletalk has
been supported since MacOS 7.6, which was released in
1997 and is able to run on most of the legacy hardware.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: Arnaldo Carvalho de Melo <[email protected]>
Cc: [email protected]
Signed-off-by: Greg Kroah-Hartman <[email protected]>
smbfs got moved to staging in 2.6.37, so we can
finally remove it in the 2.6.39 merge window.
All users should by now have migrated to cifs.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: [email protected]
Cc: Jeff Layton <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
autofs3 was moved to staging in 2.6.37, so we can
remove it in the 2.6.39 merge window. If we have
a reason to bring it back after that, this patch
can get reverted.
Signed-off-by: Arnd Bergmann <[email protected]>
About-fscking-timed-by: H. Peter Anvin <[email protected]>
Cc: Ian Kent <[email protected]>
Cc: [email protected]
Signed-off-by: Greg Kroah-Hartman <[email protected]>
There is nothing that the BKL can possibly
protect here, so just remove it.
Cc: Ross Cohen <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
The stradis driver is on its way out, but it should still be marked
correctly as depending on the big kernel lock. It could easily be
changed to not require it if someone decides to revive the driver and
port it to v4l2 in the process.
Signed-off-by: Arnd Bergmann <[email protected]>
Reported-by: Sedat Dilek <[email protected]>
Cc: Nathan Laredo <[email protected]>
Acked-by: Greg Kroah-Hartman <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Lock_kernel is gone from the code, so the comments should be updated,
too. nfsd now uses lock_flocks instead of lock_kernel to protect
against posix file locks.
Signed-off-by: Arnd Bergmann <[email protected]>
Acked-by: J. Bruce Fields <[email protected]>
Cc: [email protected]
Signed-off-by: Linus Torvalds <[email protected]>
The big kernel lock has been removed from all these files at some point,
leaving only the #include.
Remove this too as a cleanup.
Signed-off-by: Arnd Bergmann <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
All of the hard problems for BKL removal appear to be solved in the
v4l-dvb/master tree. This removes the BKL from the various open
functions that do not need it, or only use it to protect an
open count.
The zoran driver is nontrivial in this regard, so I introduce
a new mutex that locks both the open/release and the ioctl
functions. Someone with access to the hardware can probably
improve that by using the existing lock in all cases.
Finally, all drivers that still use the locked version of the
ioctl function now get called under a new mutex instead of
the BKL.
Signed-off-by: Arnd Bergmann <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The preempt count logic tries to take the BKL into account, which breaks
when CONFIG_BKL is not set.
Use the same preempt_count offset that we use without CONFIG_PREEMPT
when CONFIG_BKL is disabled.
Signed-off-by: Arnd Bergmann <[email protected]>
Reported-and-tested-by: Kirill A. Shutemov <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Nothing depends on lock_flocks using the BKL
any more, so we can do the switch over to
a private spinlock.
Signed-off-by: Arnd Bergmann <[email protected]>
As suggested by Christoph Hellwig, this moves allocation
of new file locks out of generic_setlease into the
callers, nfs4_open_delegation and fcntl_setlease in order
to allow GFP_KERNEL allocations when lock_flocks has
become a spinlock.
Signed-off-by: Arnd Bergmann <[email protected]>
Acked-by: J. Bruce Fields <[email protected]>
lockd should use lock_flocks() instead of lock_kernel()
to lock against posix locks accessing the i_flock list.
This is a prerequisite to turning lock_flocks into a
spinlock.
Signed-off-by: Arnd Bergmann <[email protected]>
Acked-by: J. Bruce Fields <[email protected]>
All uses of the BKL in freevxfs were the result of a pushdown into
code that doesn't really need it. As Christoph points out, this
is a read-only file system, which eliminates most of the races in
readdir/lookup.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: Christoph Hellwig <[email protected]>
All uses of the BKL in qnx4 were the result of a pushdown into
code that doesn't really need it. As Christoph points out, this
is a read-only file system, which eliminates most of the races in
readdir/lookup.
Signed-off-by: Arnd Bergmann <[email protected]>
Acked-by: Anders Larsen <[email protected]>
Cc: Christoph Hellwig <[email protected]>
With all the patches we have queued in the BKL removal tree, only a
few dozen modules are left that actually rely on the BKL, and even
there are lots of low-hanging fruit. We need to decide what to do
about them, this patch illustrates one of the options:
Every user of the BKL is marked as 'depends on BKL' in Kconfig,
and the CONFIG_BKL becomes a user-visible option. If it gets
disabled, no BKL using module can be built any more and the BKL
code itself is compiled out.
The one exception is file locking, which is practically always
enabled and does a 'select BKL' instead. This effectively forces
CONFIG_BKL to be enabled until we have solved the fs/lockd
mess and can apply the patch that removes the BKL from fs/locks.c.
Signed-off-by: Arnd Bergmann <[email protected]>
The bt8xx driver only uses the big kernel lock in its dst_ca_ioctl
function and never to serialize against other code, so we can
trivially replace it with a private mutex.
Signed-off-by: Arnd Bergmann <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The dvb core only uses the big kernel lock in the open
and ioctl functions, which means it can be replaced with
a dvb specific mutex. Fortunately, all the ioctl functions
go through dvb_usercopy, so we can move the serialization
in there.
Signed-off-by: Arnd Bergmann <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The dabusb device driver is sufficiently serialized using
its own mutex, no need for the big kernel lock here
in addition.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: Mauro Carvalho Chehab <[email protected]>
The sunrpc cache_ioctl function does not need the big kernel lock
because it uses its own queue_lock already.
rpc_pipe_ioctl apparently should be using i_lock like the other
operations on the pipe file descriptor do.
Signed-off-by: Arnd Bergmann <[email protected]>
According to Jens, this code does not need the BKL at all,
it is sufficiently serialized by bd_mutex.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: Jens Axboe <[email protected]>
Cc: Steven Rostedt <[email protected]>
The big kernel lock is going away, so make sure
that if it is disabled by Kconfig, we do not
try to validate it, which would result in
compile errors.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: Thomas Gleixner <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Arjan van de Ven <[email protected]>
Cc: Andrew Morton <[email protected]>
The dvb core only uses the big kernel lock in the open
and ioctl functions, which means it can be replaced with
a dvb specific mutex. Fortunately, all the ioctl functions
go through dvb_usercopy, so we can move the serialization
in there.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: Mauro Carvalho Chehab <[email protected]>
Cc: [email protected]
The bt8xx driver only uses the big kernel lock in its dst_ca_ioctl
function and never to serialize against other code, so we can
trivially replace it with a private mutex.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: [email protected]
Cc: Mauro Carvalho Chehab <[email protected]>
This driver already has a global mutex, so let's just
use that in the open function instead of the BKL.
It may not even be needed there, but this patch should
have the smallest impact.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: Mark Gross <[email protected]>
Three uml device drivers still use the big kernel lock,
but all of them can be safely converted to using
a per-driver mutex instead. Most likely this is not
even necessary, so after further review these can
and should be removed as well.
The exec system call no longer requires the BKL either,
so remove it from there, too.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: Jeff Dike <[email protected]>
Cc: [email protected]
All operations in the pci procfs ioctl functions are
atomic, so no lock is needed here.
Also add a compat_ioctl method, since all the commands
are compatible in 32 bit mode.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: Jesse Barnes <[email protected]>
Cc: Tejun Heo <[email protected]>
Cc: [email protected]
Signed-off-by: Jesse Barnes <[email protected]>
All file operations now have an explicit .llseek
operation pointer, so we can change the default
action for future code.
This makes changes the default from default_llseek
to no_llseek, which always returns -ESPIPE if
a user tries to seek on a file without a .llseek
operation.
The name of the default_llseek function remains
unchanged, if anyone thinks we should change it,
please speak up.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Cc: Al Viro <[email protected]>
Cc: [email protected]
There are currently 191 users of default_llseek.
Nine of these are in device drivers that use the
big kernel lock. None of these ever touch
file->f_pos outside of llseek or file_pos_write.
Consequently, we never rely on the BKL
in the default_llseek function and can
replace that with i_mutex, which is also
used in generic_file_llseek.
Signed-off-by: Arnd Bergmann <[email protected]>
All file_operations should get a .llseek operation so we can make
nonseekable_open the default for future file operations without a
.llseek pointer.
The three cases that we can automatically detect are no_llseek, seq_lseek
and default_llseek. For cases where we can we can automatically prove that
the file offset is always ignored, we use noop_llseek, which maintains
the current behavior of not returning an error from a seek.
New drivers should normally not use noop_llseek but instead use no_llseek
and call nonseekable_open at open time. Existing drivers can be converted
to do the same when the maintainer knows for certain that no user code
relies on calling seek on the device file.
The generated code is often incorrectly indented and right now contains
comments that clarify for each added line why a specific variant was
chosen. In the version that gets submitted upstream, the comments will
be gone and I will manually fix the indentation, because there does not
seem to be a way to do that using coccinelle.
Some amount of new code is currently sitting in linux-next that should get
the same modifications, which I will do at the end of the merge window.
Many thanks to Julia Lawall for helping me learn to write a semantic
patch that does all this.
===== begin semantic patch =====
// This adds an llseek= method to all file operations,
// as a preparation for making no_llseek the default.
//
// The rules are
// - use no_llseek explicitly if we do nonseekable_open
// - use seq_lseek for sequential files
// - use default_llseek if we know we access f_pos
// - use noop_llseek if we know we don't access f_pos,
// but we still want to allow users to call lseek
//
@ open1 exists @
identifier nested_open;
@@
nested_open(...)
{
<+...
nonseekable_open(...)
...+>
}
@ open exists@
identifier open_f;
identifier i, f;
identifier open1.nested_open;
@@
int open_f(struct inode *i, struct file *f)
{
<+...
(
nonseekable_open(...)
|
nested_open(...)
)
...+>
}
@ read disable optional_qualifier exists @
identifier read_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
expression E;
identifier func;
@@
ssize_t read_f(struct file *f, char *p, size_t s, loff_t *off)
{
<+...
(
*off = E
|
*off += E
|
func(..., off, ...)
|
E = *off
)
...+>
}
@ read_no_fpos disable optional_qualifier exists @
identifier read_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
@@
ssize_t read_f(struct file *f, char *p, size_t s, loff_t *off)
{
... when != off
}
@ write @
identifier write_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
expression E;
identifier func;
@@
ssize_t write_f(struct file *f, const char *p, size_t s, loff_t *off)
{
<+...
(
*off = E
|
*off += E
|
func(..., off, ...)
|
E = *off
)
...+>
}
@ write_no_fpos @
identifier write_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
@@
ssize_t write_f(struct file *f, const char *p, size_t s, loff_t *off)
{
... when != off
}
@ fops0 @
identifier fops;
@@
struct file_operations fops = {
...
};
@ has_llseek depends on fops0 @
identifier fops0.fops;
identifier llseek_f;
@@
struct file_operations fops = {
...
.llseek = llseek_f,
...
};
@ has_read depends on fops0 @
identifier fops0.fops;
identifier read_f;
@@
struct file_operations fops = {
...
.read = read_f,
...
};
@ has_write depends on fops0 @
identifier fops0.fops;
identifier write_f;
@@
struct file_operations fops = {
...
.write = write_f,
...
};
@ has_open depends on fops0 @
identifier fops0.fops;
identifier open_f;
@@
struct file_operations fops = {
...
.open = open_f,
...
};
// use no_llseek if we call nonseekable_open
////////////////////////////////////////////
@ nonseekable1 depends on !has_llseek && has_open @
identifier fops0.fops;
identifier nso ~= "nonseekable_open";
@@
struct file_operations fops = {
... .open = nso, ...
+.llseek = no_llseek, /* nonseekable */
};
@ nonseekable2 depends on !has_llseek @
identifier fops0.fops;
identifier open.open_f;
@@
struct file_operations fops = {
... .open = open_f, ...
+.llseek = no_llseek, /* open uses nonseekable */
};
// use seq_lseek for sequential files
/////////////////////////////////////
@ seq depends on !has_llseek @
identifier fops0.fops;
identifier sr ~= "seq_read";
@@
struct file_operations fops = {
... .read = sr, ...
+.llseek = seq_lseek, /* we have seq_read */
};
// use default_llseek if there is a readdir
///////////////////////////////////////////
@ fops1 depends on !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier readdir_e;
@@
// any other fop is used that changes pos
struct file_operations fops = {
... .readdir = readdir_e, ...
+.llseek = default_llseek, /* readdir is present */
};
// use default_llseek if at least one of read/write touches f_pos
/////////////////////////////////////////////////////////////////
@ fops2 depends on !fops1 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier read.read_f;
@@
// read fops use offset
struct file_operations fops = {
... .read = read_f, ...
+.llseek = default_llseek, /* read accesses f_pos */
};
@ fops3 depends on !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier write.write_f;
@@
// write fops use offset
struct file_operations fops = {
... .write = write_f, ...
+ .llseek = default_llseek, /* write accesses f_pos */
};
// Use noop_llseek if neither read nor write accesses f_pos
///////////////////////////////////////////////////////////
@ fops4 depends on !fops1 && !fops2 && !fops3 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier read_no_fpos.read_f;
identifier write_no_fpos.write_f;
@@
// write fops use offset
struct file_operations fops = {
...
.write = write_f,
.read = read_f,
...
+.llseek = noop_llseek, /* read and write both use no f_pos */
};
@ depends on has_write && !has_read && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier write_no_fpos.write_f;
@@
struct file_operations fops = {
... .write = write_f, ...
+.llseek = noop_llseek, /* write uses no f_pos */
};
@ depends on has_read && !has_write && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier read_no_fpos.read_f;
@@
struct file_operations fops = {
... .read = read_f, ...
+.llseek = noop_llseek, /* read uses no f_pos */
};
@ depends on !has_read && !has_write && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
@@
struct file_operations fops = {
...
+.llseek = noop_llseek, /* no read or write fn */
};
===== End semantic patch =====
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: Julia Lawall <[email protected]>
Cc: Christoph Hellwig <[email protected]>
This driver uses ioremap on regular memory to get an uncached mapping,
which causes problems on ARMv6 and higher due to aliasing with the
cached linar kernel mapping.
Make sure this gets fixed before the driver graduates from staging.
Cc: Felipe Contreras <[email protected]>
Signed-off-by: Arnd Bergmann <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
smbfs has been scheduled for removal in 2.6.27, so
maybe we can now move it to drivers/staging on the
way out.
smbfs still uses the big kernel lock and nobody
is going to fix that, so we should be getting
rid of it soon.
This removes the 32 bit compat mount and ioctl
handling code, which is implemented in common fs
code, and moves all smbfs related files into
drivers/staging/smbfs.
Signed-off-by: Arnd Bergmann <[email protected]>
Acked-by: Jeff Layton <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Nobody appears to be interested in fixing autofs3 bugs
any more and it uses the BKL, which is going away.
Move this to staging for retirement. Unless someone
complains until 2.6.38, we can remove it for good.
The include/linux/auto_fs.h header file is still used
by autofs4, so it remains in place.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: Ian Kent <[email protected]>
Cc: [email protected]
Cc: "H. Peter Anvin" <[email protected]>
Acked-by: H. Peter Anvin <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This marks up the code where sparse complains in most cases.
Most of the changes are in the ioctl handling code, which
gets __user annotations, finding one unchecked user access.
The rest is mostly about marking functions static when they
are only used in one file.
Signed-off-by: Arnd Bergmann <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This removes all warnings I get on a 64 bit build except
for those that look unfixable, where we convert a pointer
to a 32 bit integer and change its byte order!
Signed-off-by: Arnd Bergmann <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
In the original code, address space annotations are missing,
which hides a possible unchecked user pointer access.
Two functions use a lot of stack space.
Extern declarations are all in the wrong place, which leads
to type differences between caller and callee in some cases.
Signed-off-by: Arnd Bergmann <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This lets us see clearer when stuff breaks.
Most of the changes are fixes for casts between int and pointer
that don't work on 64 bit.
The ioctl function uses a large amount of stack, which gets
fixed by allocating the buffer dynamically.
Signed-off-by: Arnd Bergmann <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
As a preparation for changing the default behaviour of llseek to no_llseek,
every file_operations structure should have a .llseek operation.
There are three new instances in staging now, which can all be changed
into no_llseek explicitly since the devices do not need to seek.
Add nonseekable_open where appropriate, to prevent pread/pwrite as well.
Signed-off-by: Arnd Bergmann <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
The block device drivers have all gained new lock_kernel
calls from a recent pushdown, and some of the drivers
were already using the BKL before.
This turns the BKL into a set of per-driver mutexes.
Still need to check whether this is safe to do.
file=$1
name=$2
if grep -q lock_kernel ${file} ; then
if grep -q 'include.*linux.mutex.h' ${file} ; then
sed -i '/include.*<linux\/smp_lock.h>/d' ${file}
else
sed -i 's/include.*<linux\/smp_lock.h>.*$/include <linux\/mutex.h>/g' ${file}
fi
sed -i ${file} \
-e "/^#include.*linux.mutex.h/,$ {
1,/^\(static\|int\|long\)/ {
/^\(static\|int\|long\)/istatic DEFINE_MUTEX(${name}_mutex);
} }" \
-e "s/\(un\)*lock_kernel\>[ ]*()/mutex_\1lock(\&${name}_mutex)/g" \
-e '/[ ]*cycle_kernel_lock();/d'
else
sed -i -e '/include.*\<smp_lock.h\>/d' ${file} \
-e '/cycle_kernel_lock()/d'
fi
Signed-off-by: Arnd Bergmann <[email protected]>
All these files use the big kernel lock in a trivial
way to serialize their private file operations,
typically resulting from an earlier semi-automatic
pushdown from VFS.
None of these drivers appears to want to lock against
other code, and they all use the BKL as the top-level
lock in their file operations, meaning that there
is no lock-order inversion problem.
Consequently, we can remove the BKL completely,
replacing it with a per-file mutex in every case.
Using a scripted approach means we can avoid
typos.
These drivers do not seem to be under active
maintainance from my brief investigation. Apologies
to those maintainers that I have missed.
file=$1
name=$2
if grep -q lock_kernel ${file} ; then
if grep -q 'include.*linux.mutex.h' ${file} ; then
sed -i '/include.*<linux\/smp_lock.h>/d' ${file}
else
sed -i 's/include.*<linux\/smp_lock.h>.*$/include <linux\/mutex.h>/g' ${file}
fi
sed -i ${file} \
-e "/^#include.*linux.mutex.h/,$ {
1,/^\(static\|int\|long\)/ {
/^\(static\|int\|long\)/istatic DEFINE_MUTEX(${name}_mutex);
} }" \
-e "s/\(un\)*lock_kernel\>[ ]*()/mutex_\1lock(\&${name}_mutex)/g" \
-e '/[ ]*cycle_kernel_lock();/d'
else
sed -i -e '/include.*\<smp_lock.h\>/d' ${file} \
-e '/cycle_kernel_lock()/d'
fi
Signed-off-by: Arnd Bergmann <[email protected]>
This prepares the removal of the big kernel lock from the
file locking code. We still use the BKL as long as fs/lockd
uses it and ceph might sleep, but we can flip the definition
to a private spinlock as soon as that's done.
All users outside of fs/lockd get converted to use
lock_flocks() instead of lock_kernel() where appropriate.
Based on an earlier patch to use a spinlock from Matthew
Wilcox, who has attempted this a few times before, the
earliest patch from over 10 years ago turned it into
a semaphore, which ended up being slower than the BKL
and was subsequently reverted.
Someone should do some serious performance testing when
this becomes a spinlock, since this has caused problems
before. Using a spinlock should be at least as good
as the BKL in theory, but who knows...
Signed-off-by: Arnd Bergmann <[email protected]>
Acked-by: Matthew Wilcox <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Cc: Trond Myklebust <[email protected]>
Cc: "J. Bruce Fields" <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: Miklos Szeredi <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: John Kacur <[email protected]>
Cc: Sage Weil <[email protected]>
Cc: [email protected]
Cc: [email protected]
The BKL in ocfs2/dlmfs is used in put_super, fill_super and remount_fs
that are all three protected by the superblocks s_umount rw_semaphore.
The use in ocfs2_control_open is evidently unrelated and the function
is protected by ocfs2_control_lock.
Therefore it is safe to remove the BKL entirely.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: Mark Fasheh <[email protected]>
Cc: Joel Becker <[email protected]>
The BKL is only used in put_super and fill_super, which are both protected
by the superblocks s_umount rw_semaphore. Therefore it is safe to remove
the BKL entirely.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: Phillip Lougher <[email protected]>
The BKL is only used in put_super, fill_super and remount_fs that are all
three protected by the superblocks s_umount rw_semaphore. Therefore it is
safe to remove the BKL entirely.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: David Woodhouse <[email protected]>
The BKL is only used in fill_super, which is protected by the superblocks
s_umount rw_semaphorei, and in fasync, which does not do anything that
could require the BKL. Therefore it is safe to remove the BKL entirely.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: Dustin Kirkland <[email protected]>
Cc: Tyler Hicks <[email protected]>
Cc: [email protected]
The BKL is only used in put_super and fill_super, which are both protected
by the superblocks s_umount rw_semaphore. Therefore it is safe to remove
the BKL entirely.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: [email protected]
Cc: David Howells <[email protected]>
The BKL is only used in fill_super, which is protected by the superblocks
s_umount rw_semaphore. Therefore it is safe to remove the BKL entirely.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: David Brownell <[email protected]>
Cc: [email protected]
autofs4 uses the BKL only to guard its ioctl operations.
This can be trivially converted to use a mutex, as we have
done with most device drivers before.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: "H. Peter Anvin" <[email protected]>
Cc: Ian Kent <[email protected]>
As in other file systems, we can replace the big kernel lock
with a private mutex in isofs. This means we can now access
multiple file systems concurrently, but it also means that
we serialize readdir and lookup across sleeping operations
which previously released the big kernel lock. This should
not matter though, as these operations are in practice
serialized through the hardware access.
The isofs_get_blocks functions now does not take any lock
any more, it used to recursively get the BKL. After looking
at the code for hours, I convinced myself that it was never
needed here anyway, because it only reads constant fields
of the inode and writes to a buffer head array that is
at this time only visible to the caller.
The get_sb and fill_super operations do not need the locking
at all because they operate on a file system that is either
about to be created or to be destroyed but in either case
is not visible to other threads.
Signed-off-by: Arnd Bergmann <[email protected]>
The lock_kernel in fat_put_super is not needed because
it only protects the super block itself and we know that
no other thread can reach it because we are about to
kfree the object.
In the two fill_super functions, this converts the locking
to use lock_super like elsewhere in the fat code. This
is probably not needed either, but is consistent and puts
us on the safe side.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: OGAWA Hirofumi <[email protected]>
Cc: Jan Blunck <[email protected]>
The i810 and i830 device drivers may replace their file operations
on an open file descriptor. My previous patch to move the BKL
out of the common DRM code into these drivers only caught the
default file operations, not the ones that actually end up being
used.
Found while trying to come up with a way to kill the BKL for
good in these drivers.
Signed-off-by: Arnd Bergmann <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
All uses of the big kernel lock in the cris architecture
are for ioctl and open functions of character device drivers,
which can be trivially converted to a per-driver mutex.
Most of these are probably unnecessary, so it may make sense
to audit them and eventually remove the extra mutex introduced
by this patch.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: Mikael Starvik <[email protected]>
Cc: Jesper Nilsson <[email protected]>
Cc: [email protected]
All uses of the BKL on alpha are totally bogus, nothing
is really protected by this. Remove the remaining users
so we don't have to mark alpha as 'depends on BKL'.
Signed-off-by: Arnd Bergmann <[email protected]>
Cc: Richard Henderson <[email protected]>
Cc: Ivan Kokshaysky <[email protected]>
Cc: Matt Turner <[email protected]>
Cc: [email protected]
Arnd Bergmann