coredump: sanitize the setting of signal->group_exit_code

Now that the coredumping process can be SIGKILL'ed, the setting of
->group_exit_code in do_coredump() can race with complete_signal() and
SIGKILL or 0x80 can be "lost", or wait(status) can report status ==
SIGKILL | 0x80.

But the main problem is that it is not clear to me what should we do if
binfmt->core_dump() succeeds but SIGKILL was sent, that is why this patch
comes as a separate change.

This patch adds 0x80 if ->core_dump() succeeds and the process was not
killed.  But perhaps we can (should?) re-set ->group_exit_code changed by
SIGKILL back to "siginfo->si_signo |= 0x80" in case when core_dumped == T.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Tested-by: Mandeep Singh Baines <msb@chromium.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Neil Horman <nhorman@redhat.com>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: Roland McGrath <roland@hack.frob.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

coredump: ensure that SIGKILL always kills the dumping thread

prepare_signal() blesses SIGKILL sent to the dumping process but this
signal can be "lost" anyway.  The problems is, complete_signal() sees
SIGNAL_GROUP_EXIT and skips the "kill them all" logic.  And even if the
dumping process is single-threaded (so the target is always "correct"),
the group-wide SIGKILL is not recorded in task->pending and thus
__fatal_signal_pending() won't be true.  A multi-threaded case has even
more problems.

And even ignoring all technical details, SIGNAL_GROUP_EXIT doesn't look
right to me.  This coredumping process is not exiting yet, it can do a lot
of work dumping the core.

With this patch the dumping process doesn't have SIGNAL_GROUP_EXIT, we set
signal->group_exit_task instead.  This makes signal_group_exit() true and
thus this should equally close the races with exit/exec/stop but allows to
kill the dumping thread reliably.

Notes:
- It is not clear what should we do with ->group_exit_code
  if the dumper was killed, see the next change.

- we need more (hopefully straightforward) changes to ensure
  that SIGKILL actually interrupts the coredump. Basically we
  need to check __fatal_signal_pending() in dump_write() and
  dump_seek().

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Tested-by: Mandeep Singh Baines <msb@chromium.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Neil Horman <nhorman@redhat.com>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: Roland McGrath <roland@hack.frob.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

coredump: only SIGKILL should interrupt the coredumping task

There are 2 well known and ancient problems with coredump/signals, and a
lot of related bug reports:

- do_coredump() clears TIF_SIGPENDING but of course this can't help
  if, say, SIGCHLD comes after that.

  In this case the coredump can fail unexpectedly. See for example
  wait_for_dump_helper()->signal_pending() check but there are other
  reasons.

- At the same time, dumping a huge core on the slow media can take a
  lot of time/resources and there is no way to kill the coredumping
  task reliably. In particular this is not oom_kill-friendly.

This patch tries to fix the 1st problem, and makes the preparation for the
next changes.

We add the new SIGNAL_GROUP_COREDUMP flag set by zap_threads() to indicate
that this process dumps the core.  prepare_signal() checks this flag and
nacks any signal except SIGKILL.

Note that this check tries to be conservative, in the long term we should
probably treat the SIGNAL_GROUP_EXIT case equally but this needs more
discussion.  See marc.info/?l=linux-kernel&m=120508897917439

Notes:
- recalc_sigpending() doesn't check SIGNAL_GROUP_COREDUMP.
  The patch assumes that dump_write/etc paths should never
  call it, but we can change it as well.

- There is another source of TIF_SIGPENDING, freezer. This
  will be addressed separately.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Tested-by: Mandeep Singh Baines <msb@chromium.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Neil Horman <nhorman@redhat.com>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: Roland McGrath <roland@hack.frob.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

kmod: remove call_usermodehelper_fns()

This function suffers from not being able to determine if the cleanup is
called in case it returns -ENOMEM.  Nobody is using it anymore, so let's
remove it.

Signed-off-by: Lucas De Marchi <lucas.demarchi@profusion.mobi>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: David Howells <dhowells@redhat.com>
Cc: James Morris <james.l.morris@oracle.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Tejun Heo <tj@kernel.org>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

usermodehelper: split remaining calls to call_usermodehelper_fns()

These are the only users of call_usermodehelper_fns().  This function
suffers from not being able to determine if the cleanup is called.  Even
if in this places the cleanup pointer is NULL, convert them to use the
separate call_usermodehelper_setup() + call_usermodehelper_exec()
functions so we can remove the _fns variant.

Signed-off-by: Lucas De Marchi <lucas.demarchi@profusion.mobi>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: David Howells <dhowells@redhat.com>
Cc: James Morris <james.l.morris@oracle.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Tejun Heo <tj@kernel.org>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

coredump: remove trailling whitespace

Signed-off-by: Lucas De Marchi <lucas.demarchi@profusion.mobi>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: David Howells <dhowells@redhat.com>
Cc: James Morris <james.l.morris@oracle.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Tejun Heo <tj@kernel.org>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

KEYS: split call to call_usermodehelper_fns()

Use call_usermodehelper_setup() + call_usermodehelper_exec() instead of
calling call_usermodehelper_fns().  In case there's an OOM in this last
function the cleanup function may not be called - in this case we would
miss a call to key_put().

Signed-off-by: Lucas De Marchi <lucas.demarchi@profusion.mobi>
Cc: Oleg Nesterov <oleg@redhat.com>
Acked-by: David Howells <dhowells@redhat.com>
Acked-by: James Morris <james.l.morris@oracle.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Tejun Heo <tj@kernel.org>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

kmod: split call to call_usermodehelper_fns()

Use call_usermodehelper_setup() + call_usermodehelper_exec() instead of
calling call_usermodehelper_fns().  In case the latter returns -ENOMEM the
cleanup function may had not been called - in this case we would not free
argv and module_name.

Signed-off-by: Lucas De Marchi <lucas.demarchi@profusion.mobi>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: David Howells <dhowells@redhat.com>
Cc: James Morris <james.l.morris@oracle.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Tejun Heo <tj@kernel.org>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

usermodehelper-export-_exec-and-_setup-functions-fix

export call_usermodehelper_setup() to modules

Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: David Howells <dhowells@redhat.com>
Cc: James Morris <james.l.morris@oracle.com>
Cc: Lucas De Marchi <lucas.demarchi@profusion.mobi>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Tejun Heo <tj@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

usermodehelper: export call_usermodehelper_exec() and call_usermodehelper_setup()

call_usermodehelper_setup() + call_usermodehelper_exec() need to be called
instead of call_usermodehelper_fns() when the cleanup function needs to be
called even when an ENOMEM error occurs.  In this case using
call_usermodehelper_fns() the user can't distinguish if the cleanup
function was called or not.

Signed-off-by: Lucas De Marchi <lucas.demarchi@profusion.mobi>
Reviewed-by: Oleg Nesterov <oleg@redhat.com>
Cc: David Howells <dhowells@redhat.com>
Cc: James Morris <james.l.morris@oracle.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Tejun Heo <tj@kernel.org>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

selftest: add a test case for PTRACE_PEEKSIGINFO

* Dump signals from process-wide and per-thread queues with
  different sizes of buffers.
* Check error paths for buffers with restricted permissions. A part of
  buffer or a whole buffer is for read-only.
* Try to get nonexistent signal.

Signed-off-by: Andrew Vagin <avagin@openvz.org>
Cc: Roland McGrath <roland@redhat.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: David Howells <dhowells@redhat.com>
Cc: Dave Jones <davej@redhat.com>
Cc: "Michael Kerrisk (man-pages)" <mtk.manpages@gmail.com>
Cc: Pavel Emelyanov <xemul@parallels.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Pedro Alves <palves@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

ptrace: add ability to retrieve signals without removing from a queue (v4)

This patch adds a new ptrace request PTRACE_PEEKSIGINFO.

This request is used to retrieve information about pending signals
starting with the specified sequence number.  Siginfo_t structures are
copied from the child into the buffer starting at "data".

The argument "addr" is a pointer to struct ptrace_peeksiginfo_args.
struct ptrace_peeksiginfo_args {
u64 off; /* from which siginfo to start */
u32 flags;
s32 nr; /* how may siginfos to take */
};

"nr" has type "s32", because ptrace() returns "long", which has 32 bits on
i386 and a negative values is used for errors.

Currently here is only one flag PTRACE_PEEKSIGINFO_SHARED for dumping
signals from process-wide queue.  If this flag is not set, signals are
read from a per-thread queue.

The request PTRACE_PEEKSIGINFO returns a number of dumped signals.  If a
signal with the specified sequence number doesn't exist, ptrace returns
zero.  The request returns an error, if no signal has been dumped.

Errors:
EINVAL - one or more specified flags are not supported or nr is negative
EFAULT - buf or addr is outside your accessible address space.

A result siginfo contains a kernel part of si_code which usually striped,
but it's required for queuing the same siginfo back during restore of
pending signals.

This functionality is required for checkpointing pending signals.  Pedro
Alves suggested using it in "gdb" to peek at pending signals.  gdb already
uses PTRACE_GETSIGINFO to get the siginfo for the signal which was already
dequeued.  This functionality allows gdb to look at the pending signals
which were not reported yet.

The prototype of this code was developed by Oleg Nesterov.

Signed-off-by: Andrew Vagin <avagin@openvz.org>
Cc: Roland McGrath <roland@redhat.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: David Howells <dhowells@redhat.com>
Cc: Dave Jones <davej@redhat.com>
Cc: "Michael Kerrisk (man-pages)" <mtk.manpages@gmail.com>
Cc: Pavel Emelyanov <xemul@parallels.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Pedro Alves <palves@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

Documentation: update nfs option in filesystem/vfat.txt

Add descriptions about 'stale_rw' and 'nostale_ro' nfs options in
filesystem/vfat.txt

Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Ravishankar N <ravi.n1@samsung.com>
Signed-off-by: Amit Sahrawat <a.sahrawat@samsung.com>
Acked-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Acked-by: Rob Landley <rob@landley.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

fat (exportfs): rebuild directory-inode if fat_dget()

This patch enables rebuilding of directory inodes which are not present in
the cache.This is done by traversing the disk clusters to find the
directory entry of the parent directory and using its i_pos to build the
inode.

The traversal is done by fat_scan_logstart() which is similar to
fat_scan() but matches i_pos values instead of names.fat_scan_logstart()
needs an inode parameter to work, for which a dummy inode is created by
it's caller fat_rebuild_parent().  This dummy inode is destroyed after the
traversal completes.

All this is done  only if the nostale_ro nfs mount option is specified.

Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Ravishankar N <ravi.n1@samsung.com>
Signed-off-by: Amit Sahrawat <a.sahrawat@samsung.com>
Acked-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

fat (exportfs): rebuild inode if ilookup() fails

If the cache lookups fail,use the i_pos value to find the directory entry
of the inode and rebuild the inode.Since this involves accessing the FAT
media, do this only if the nostale_ro nfs mount option is specified.

Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Ravishankar N <ravi.n1@samsung.com>
Signed-off-by: Amit Sahrawat <a.sahrawat@samsung.com>
Acked-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

fat: restructure export_operations

Define two nfs export_operation structures,one for 'stale_rw' mounts and
the other for 'nostale_ro'.  The latter uses i_pos as a basis for encoding
and decoding file handles.

Also, assign i_pos to kstat->ino.  The logic for rebuilding the inode is
added in the subsequent patches.

Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Ravishankar N <ravi.n1@samsung.com>
Signed-off-by: Amit Sahrawat <a.sahrawat@samsung.com>
Acked-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

fat: introduce a helper fat_get_blknr_offset()

Introduce helper function to get the block number and offset for a given
i_pos value.  Use it in __fat_write_inode() now and later on in nfs.c

Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Ravishankar N <ravi.n1@samsung.com>
Signed-off-by: Amit Sahrawat <a.sahrawat@samsung.com>
Acked-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

fat: move fat_i_pos_read to fat.h

Move fat_i_pos_read to fat.h so that it can be called from nfs.c in the
subsequent patches to encode the file handle.

Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Ravishankar N <ravi.n1@samsung.com>
Signed-off-by: Amit Sahrawat <a.sahrawat@samsung.com>
Acked-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

fat: introduce 2 new values for the -o nfs mount option

This patchset eliminates the client side ESTALE errors when a FAT
partition exported over NFS has its dentries evicted from the cache.  The
idea is to find the on-disk location_'i_pos' of the dirent of the inode
that has been evicted and use it to rebuild the inode.

This patch:

Provide two possible values 'stale_rw' and 'nostale_ro' for the -o nfs
mount option.The first one allows all file operations but does not reduce
ESTALE errors on memory constrained systems.  The second one eliminates
ESTALE errors but mounts the filesystem as read-only.  Not specifying a
value defaults to 'stale_rw'.

Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Ravishankar N <ravi.n1@samsung.com>
Signed-off-by: Amit Sahrawat <a.sahrawat@samsung.com>
Acked-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

hfsplus-fix-warnings-in-fs-hfsplus-bfindc-in-function-hfs_find_1st_rec_by_cnid-fix

make the workaround more explicit

Cc: Vyacheslav Dubeyko <slava@dubeyko.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

hfsplus: fix warnings in fs/hfsplus/bfind.c

fs/hfsplus/bfind.c: In function 'hfs_find_1st_rec_by_cnid':
(1) include/uapi/linux/swab.h:60:2: warning: 'search_cnid' may be used uninitialized in this function [-Wmaybe-uninitialized]
(2) include/uapi/linux/swab.h:60:2: warning: 'cur_cnid' may be used uninitialized in this function [-Wmaybe-uninitialized]

Signed-off-by: Vyacheslav Dubeyko <slava@dubeyko.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

drivers/rtc/rtc-max8907.c: remove redundant code

Remove unnecessary goto statements to simplify the code.
devm_request_threaded_irq returns 0 upon success, hence explicit return 0
is not necessary.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Cc: Chiwoong Byun <woong.byun@samsung.com>
Cc: Jonghwa Lee <jonghwa3.lee@samsung.com>
Cc: Laxman dewangan <ldewangan@nvidia.com>
Cc: Venu Byravarasu <vbyravarasu@nvidia.com>
Cc: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

drivers/rtc/rtc-max77686.c: fix incorrect return value on error

'ret' was not initialized to error code before returning. While
at it also remove some redundant code and cleanup.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Cc: Chiwoong Byun <woong.byun@samsung.com>
Cc: Jonghwa Lee <jonghwa3.lee@samsung.com>
Cc: Laxman dewangan <ldewangan@nvidia.com>
Cc: Venu Byravarasu <vbyravarasu@nvidia.com>
Cc: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

drivers/rtc/rtc-max8997.c: fix incorrect return value on error

'ret' was being returned without initializing it to error code.  While at
it also remove duplicate return statement.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Cc: Jonghwa Lee <jonghwa3.lee@samsung.com>
Cc: Chiwoong Byun <woong.byun@samsung.com>
Cc: Laxman dewangan <ldewangan@nvidia.com>
Cc: Venu Byravarasu <vbyravarasu@nvidia.com>
Cc: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

drivers/rtc/rtc-tps65910.c: fix incorrect return value on error

'ret' was not initialized to correct error value before returning.  Since
'irq' is also being tested for 0, we cannot return irq itself as it means
function is success even though we are returning before completing the
probe.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Acked-by: Venu Byravarasu <vbyravarasu@nvidia.com>
Cc: Chiwoong Byun <woong.byun@samsung.com>
Cc: Jonghwa Lee <jonghwa3.lee@samsung.com>
Cc: Laxman dewangan <ldewangan@nvidia.com>
Cc: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

drivers/rtc/rtc-tps6586x.c: remove incorrect use of rtc_device_unregister

Device managed functions do not need explicit freeing/unregistering.
Moreover in this case it was done using non-device managed function which
is incorrect.  Hence remove it.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Cc: Laxman dewangan <ldewangan@nvidia.com>
Cc: Chiwoong Byun <woong.byun@samsung.com>
Cc: Jonghwa Lee <jonghwa3.lee@samsung.com>
Cc: Venu Byravarasu <vbyravarasu@nvidia.com>
Cc: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

drivers/rtc/rtc-88pm80x.c: add CONFIG_PM_SLEEP to suspend/resume functions

Add CONFIG_PM_SLEEP to suspend/resume functions to fix the following build
warning when CONFIG_PM_SLEEP is not selected.  This is because sleep PM
callbacks defined by SIMPLE_DEV_PM_OPS are only used when the
CONFIG_PM_SLEEP is enabled.

drivers/rtc/rtc-88pm80x.c:238:12: warning: 'pm80x_rtc_suspend' defined but not used [-Wunused-function]
drivers/rtc/rtc-88pm80x.c:243:12: warning: 'pm80x_rtc_resume' defined but not used [-Wunused-function]

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

drivers/rtc/rtc-ds1374.c: add CONFIG_PM_SLEEP to suspend/resume functions

Add CONFIG_PM_SLEEP to suspend/resume functions to fix the following build
warning when CONFIG_PM_SLEEP is not selected.  This is because sleep PM
callbacks defined by SIMPLE_DEV_PM_OPS are only used when the
CONFIG_PM_SLEEP is enabled.

drivers/rtc/rtc-ds1374.c:413:12: warning: 'ds1374_suspend' defined but not used [-Wunused-function]
drivers/rtc/rtc-ds1374.c:422:12: warning: 'ds1374_resume' defined but not used [-Wunused-function]

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

drivers/rtc/rtc-palmas.c: add dt support

Add of_device_id table for Palma RTC to be enable the driver from DT file.

The driver can be registered from DT file as:
palmas: tps65913@58 {
:::::::::::
palmas_rtc {
compatible = "ti,palmas-rtc";
interrupt-parent = <&palmas>;
interrupts = <8 0>;
};
};

Signed-off-by: Laxman Dewangan <ldewangan@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

drivers/rtc/rtc-omap.c: update to devm_* API

Update the code to use devm_* API so that driver core will manage
resources.

Signed-off-by: Vishwanathrao Badarkhe, Manish <manishv.b@ti.com>
Tested-by: Sekhar Nori <nsekhar@ti.com>
Cc: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-tps80031: switch to using SIMPLE_DEV_PM_OPS

Switch to using the SIMPLE_DEV_PM_OPS macro to declare the driver's
pm_ops.  It reduces code size.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-tps65910: switch to using SIMPLE_DEV_PM_OPS

Switch to using the SIMPLE_DEV_PM_OPS macro to declare the driver's
pm_ops.  It reduces code size.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-tps6586x: switch to using SIMPLE_DEV_PM_OPS

Switch to using the SIMPLE_DEV_PM_OPS macro to declare the driver's
pm_ops.  It reduces code size.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-wm8350: switch to using SIMPLE_DEV_PM_OPS

Switch to using the SIMPLE_DEV_PM_OPS macro to declare the driver's
pm_ops.  It reduces code size.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-sh: switch to using SIMPLE_DEV_PM_OPS

Switch to using the SIMPLE_DEV_PM_OPS macro to declare the driver's
pm_ops.  It reduces code size.  Also, CONFIG_PM_SLEEP is added to prevent
build warning when CONFIG_PM_SLEEP is not selected.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-sa1100: switch to using SIMPLE_DEV_PM_OPS

Switch to using the SIMPLE_DEV_PM_OPS macro to declare the driver's
pm_ops.  It reduces code size.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-rc5t583: switch to using SIMPLE_DEV_PM_OPS

Switch to using the SIMPLE_DEV_PM_OPS macro to declare the driver's
pm_ops.  It reduces code size.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-pxa: switch to using SIMPLE_DEV_PM_OPS

Switch to using the SIMPLE_DEV_PM_OPS macro to declare the driver's
pm_ops.  It reduces code size.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-mxc: switch to using SIMPLE_DEV_PM_OPS

Switch to using the SIMPLE_DEV_PM_OPS macro to declare the driver's
pm_ops.  It reduces code size.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-at91rm9200: switch to using SIMPLE_DEV_PM_OPS

Switch to using the SIMPLE_DEV_PM_OPS macro to declare the driver's
pm_ops.  It reduces code size.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-x1205: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-wm8350: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-tile: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-test: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-sun4v: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-starfire: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-rx8581: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-rx4581: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-rv3029c2: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-rs5c313: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-rc5t583: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-r9701: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-ps3: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-max6902: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-max6900: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-m48t86: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-m41t94: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-m41t93: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-ls1x: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: hid-sensor-time: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-generic: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-em3027: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-efi: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-ds3234: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-ds1672: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-ds1302: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-dm355evm: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-bq32k: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-au1xxx: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-ab3100: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup paths
simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-fm3130: use dev_dbg() instead of pr_debug()

dev_dbg() is preferred to pr_debug().

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-ds1307: use dev_dbg() instead of pr_debug()

dev_dbg() is preferred to pr_debug().

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

drivers/rtc/rtc-s3c.c: convert s3c_rtc to dev_pm_ops

Instead of using legacy suspend/resume methods, using newer dev_pm_ops
structure allows better control over power management.  Also, 'wake_en'
variable is moved, because it is only used when CONFIG_PM_SLEEP is
enabled.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Cc: Grant Likely <grant.likely@secretlab.ca>
Cc: Rob Herring <rob.herring@calxeda.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-palmas: use devm_request_threaded_irq()

Use devm_request_threaded_irq() to make cleanup paths more simple.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

drivers/rtc/rtc-ds1307.c: long block operations bugfix

The rtc-ds1307 driver does not properly handle block operations bigger
than 32 bytes in either of the two modes supported (SMbus native, or
emulated if not supported by the SMbus platform driver).

It also does not properly handle userland-supplied input (block operation
length) through sysfs and may suffer a type of buffer overrun.

The driver has been modified with proper input validation, buffer sizes,
and now splits block transfers bigger than 32 bytes into separate
transfers.

Explanation : Buffer size allocated is I2C_SMBUS_BLOCK_MAX which equals to
32 as per the SMbus spec.  Reads and write may be up to 56 bytes (to the
NVRAM).  This patch allocated a 255 byte buffer, the maximum allowable
(address is an u8).  It's not only a buffer problem, SMbus only supports
up to 32 bytes transfer at once, so it's needed to split bigger transfers.

Patch successfully tested on 3.2.27; cleanly applies on 3.7-rc4.

[akpm@linux-foundation.org: rework code to avoid 80-column overflows]
Signed-off-by: Bertrand Achard <ba@cykian.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

drivers/rtc/rtc-tegra.c: fix build warning

Fix the below build warning:

drivers/rtc/rtc-tegra.c: In function 'tegra_rtc_probe':
drivers/rtc/rtc-tegra.c:353:33: warning: passing argument 1 of 'devm_rtc_device_register' from incompatible pointer type [enabled by default]
include/linux/rtc.h:136:27: note: expected 'struct device *' but argument is of type 'const char *'
drivers/rtc/rtc-tegra.c:353:33: warning: passing argument 2 of 'devm_rtc_device_register' from incompatible pointer type [enabled by default]
include/linux/rtc.h:136:27: note: expected 'const char *' but argument is of type 'struct device *'

This warning was introduced with patch
commit 8a96d445e921b0d881421767fb6f7c5b6fb91b3a
drivers/rtc/rtc-tegra.c: use managed rtc_device_register()

Signed-off-by: Laxman Dewangan <ldewangan@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

drivers/rtc/rtc-tegra.c: use managed rtc_device_register()

Use devm_rtc_device_register() for registering RTC device.  This will
reduce the code for unregistering RTC device in cleanup path and remove
the implementation of remove callback of platform driver.

Signed-off-by: Laxman Dewangan <ldewangan@nvidia.com>
Reviewed-by: Thierry Reding <thierry.reding@avionic-design.de>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Cc: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

drivers/rtc/rtc-tegra.c: set irq name as device name

The irq name of tegra rtc shows as "rtc alarm" which actually does not
reflect the name related to driver.

Passing the device name to have the irq names with driver name.

Signed-off-by: Laxman Dewangan <ldewangan@nvidia.com>
Reviewed-by: Thierry Reding <thierry.reding@avionic-design.de>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Cc: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

drivers/rtc/rtc-tegra.c: use struct dev_pm_ops for power management

Make the Tegra RTC controller driver define its PM callbacks through a
struct dev_pm_ops object rather than by using legacy PM hooks in struct
platform_driver.

Signed-off-by: Laxman Dewangan <ldewangan@nvidia.com>
Reviewed-by: Thierry Reding <thierry.reding@avionic-design.de>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Cc: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

drivers/rtc/rtc-tegra.c: protect suspend/resume callbacks with CONFIG_PM_SLEEP

CONFIG_PM doesn't actually enable any of the PM callbacks, it only allows
to enable CONFIG_PM_SLEEP and CONFIG_PM_RUNTIME.  This means if CONFIG_PM
is used to protect system sleep callbacks then it may end up unreferenced
if only runtime PM is enabled.  Hence protecting sleep callbacks with
CONFIG_PM_SLEEP.

Signed-off-by: Laxman Dewangan <ldewangan@nvidia.com>
Reviewed-by: Thierry Reding <thierry.reding@avionic-design.de>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Cc: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

drivers-rtc-rtc-pxac-fix-set-time-sync-time-issue-fix

The "fix set time sync time issue" adds calls to udelay(), but
doesn't add the include file. End result is build breakage:

drivers/rtc/rtc-pxa.c: In function 'pxa_rtc_set_time':
drivers/rtc/rtc-pxa.c:267:2: error: implicit declaration of function 'udelay' [-Werror=implicit-function-declaration]

Signed-off-by: Olof Johansson <olof@lixom.net>
Cc: Leo Song <liangs@marvell.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

drivers/rtc/rtc-pxa.c: fix set time sync time issue

Fix set time and sync time issue, add some delay when set pxa rtc timer
according to spec

Signed-off-by: Leo Song <liangs@marvell.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

below is the warning reported with the gcc (gcc (GCC) 4.7.2 20121109 (Red
Hat 4.7.2-8))

with make W=1

drivers/rtc/rtc-ds1286.c: In function `ds1286_read_alarm':
drivers/rtc/rtc-ds1286.c:273:16: warning: variable `cmd' set but not used [-Wunused-but-set-variable]

fixed by removing the variable

Signed-off-by: Devendra Naga <devendra.aaru@gmail.com>
Cc: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Cc: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-wm831x: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup
paths simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-vt8500: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup
paths simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-tx4939: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup
paths simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-tps80031: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup
paths simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-tps65910: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup
paths simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-tps6586x: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup
paths simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-stk17ta8: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup
paths simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-spear: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup
paths simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-snvs: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup
paths simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-s3c: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup
paths simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-pcf8523: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup
paths simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-palmas: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup
paths simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-mxc: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup
paths simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-mv: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup
paths simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-max8997: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup
paths simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-max8907: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup
paths simpler.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

rtc: rtc-max77686: use devm_rtc_device_register()

devm_rtc_device_register() is device managed and makes cleanup
paths simpler. Also, this patch uses devm_request_threaded_irq().

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>