This patch combines Greg Bank's dprintk() work with the existing dynamic
printk patchset, we are now calling it 'dynamic debug'.

The new feature of this patchset is a richer /debugfs control file interface,
(an example output from my system is at the bottom), which allows fined grained
control over the the debug output. The output can be controlled by function,
file, module, format string, and line number.

for example, enabled all debug messages in module 'nf_conntrack':

echo -n 'module nf_conntrack +p' > /mnt/debugfs/dynamic_debug/control

to disable them:

echo -n 'module nf_conntrack -p' > /mnt/debugfs/dynamic_debug/control

A further explanation can be found in the documentation patch.
Signed-off-by: NGreg Banks <gnb@sgi.com>
Signed-off-by: NJason Baron <jbaron@redhat.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Commit 86c9508eb1c0ce5aa07b5cf1d36b60c54efc3d7a
"sysfs: don't block indefinitely for unmapped files" in linux-next
crashes the PowerMac G5 when X starts up.  It's caught out by the way
powerpc's pci_mmap of legacy_mem uses shmem_zero_setup(), substituting
a new vma->vm_file whose private_data no longer points to the bin_buffer
(substitution done because some versions of X crash if that mmap fails).

The fix to this is straightforward: the original vm_file is fput() in
that case, so this mmap won't block sysfs at all, so just don't switch
over to bin_vm_ops if vm_file has changed.

But more fixes made before realizing that was the problem:-

It should not be an error if bin_page_mkwrite() finds no underlying
page_mkwrite().

Check that a file already mmap'ed has the same underlying vm_ops
_before_ pointing vma->vm_ops at bin_vm_ops.

If the file being mmap'ed is a shmem/tmpfs file, don't fail the mmap
on CONFIG_NUMA=y, just because that has a set_policy and get_policy:
provide bin_set_policy, bin_get_policy and bin_migrate.
Signed-off-by: NHugh Dickins <hugh@veritas.com>
Acked-by: NEric Biederman <ebiederm@aristanetworks.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Right now, the kobject_uevent code blocks for each uevent that's being
generated, due to using (for hystoric reasons) UHM_WAIT_EXEC as flag to
call_usermode_helper().  Specifically, the effect is that each uevent
that is being sent causes the code to wake up keventd, then block until
keventd has processed the work. Needless to say, this happens many times
during the system boot.

This patches changes that to UHN_NO_WAIT (brilliant name for a constant
btw) so that we only schedule the work to fire the uevent message, but
do not wait for keventd to process the work.

This removes one of the bottlenecks during boot; each one of them is
only a small effect, but the sum of them does add up.

[Note, distros that need this are broken, they should be setting
CONFIG_UEVENT_HELPER_PATH to "", that way this code path will never be
excuted at all -- gregkh]
Signed-off-by: NArjan van de Ven <arjan@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

The only way for a sysfs attribute to remove itself (without
deadlock) is to use the sysfs_schedule_callback() interface.

Vegard Nossum discovered that a poorly written sysfs ->store
callback can repeatedly schedule remove callbacks on the same
device over and over, e.g.

	$ while true ; do echo 1 > /sys/devices/.../remove ; done

If the 'remove' attribute uses the sysfs_schedule_callback API
and also does not protect itself from concurrent accesses, its
callback handler will be called multiple times, and will
eventually attempt to perform operations on a freed kobject,
leading to many problems.

Instead of requiring all callers of sysfs_schedule_callback to
implement their own synchronization, provide the protection in
the infrastructure.

Now, sysfs_schedule_callback will only allow one scheduled
callback per kobject. On subsequent calls with the same kobject,
return -EAGAIN.

This is a short term fix. The long term fix is to allow sysfs
attributes to remove themselves directly, without any of this
callback hokey pokey.

[cornelia.huck@de.ibm.com: s390 ccwgroup bits]

Reported-by: vegard.nossum@gmail.com
Signed-off-by: NAlex Chiang <achiang@hp.com>
Acked-by: NCornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

dpm_list currently relies on the fact that child devices will
be registered after their parents to get a correct suspend
order. Using device_move() however destroys this assumption, as
an already registered device may be moved under a newly registered
one.

This patch adds a new argument to device_move(), allowing callers
to specify how dpm_list should be adapted.
Signed-off-by: NCornelia Huck <cornelia.huck@de.ibm.com>
Acked-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

do some cleanup on drivers/base/sys.c
Signed-off-by: NZhenwen Xu <helight.xu@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch implements uevent suppress in kobject and removes it
from struct device, based on the following ideas:

1,Uevent sending should be one attribute of kobject, so suppressing it
in kobject layer is more natural than in device layer. By this way,
we can do it for other objects embedded with kobject.

2,It may save several bytes for each instance of struct device.(On my
omap3(32bit ARM) based box, can save 8bytes per device object)

This patch also introduces dev_set|get_uevent_suppress() helpers to
set and query uevent_suppress attribute in case to help kobject
as private part of struct device in future.

[This version is against the latest driver-core patch set of Greg,please
ignore the last version.]
Signed-off-by: NMing Lei <tom.leiming@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

During bootup performance tracing I noticed many occurrences of
vca* device creation and removal, leading to the usual userspace
uevent processing, which are, in this case, rather pointless.

A simple test showing the kernel timing (not including all the
work userspace has to do), gives us these numbers:
  $ time for i in `seq 1000`; do echo a > /dev/tty2; done
  real    0m1.142s
  user    0m0.015s
  sys     0m0.540s

If we move the hook for the vcs* driver core devices from the
tty "binding" to the vc allocation/deallocation, which is what
the vcs* devices represent, we get the following numbers:
  $ time for i in `seq 1000`; do echo a > /dev/tty2; done
  real    0m0.152s
  user    0m0.030s
  sys     0m0.072s

Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: NKay Sievers <kay.sievers@vrfy.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

We will remove platform_data field from struct device until
all platform devices pass its specific data from platfom_device
and all platform drivers use platform specific data passed by
platform_device->platform_data. This kind of conversion will
need a long time, for thousands of files is affected.

To make the conversion easily, we allow platform specific data
passed by struct device or struct platform_device and platform
driver may use it from struct device or struct platform_device.
Signed-off-by: NMing Lei <tom.leiming@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch moves platform_data from struct device into
struct platform_device, based on the two ideas:

1. Now all platform_driver is registered by platform_driver_register,
   which makes probe()/release()/... of platform_driver passed parameter
   of platform_device *, so platform driver can get platform_data from
   platform_device;

2. Other kind of devices do not need to use platform_data, we can
   decrease size of device if moving it to platform_device.

Taking into consideration of thousands of files to be fixed and they
can't be finished in one night(maybe it will take a long time), so we
keep platform_data in device to allow two kind of cases coexist until
all platform devices pass its platfrom data from
platform_device->platform_data.

All patches to do this kind of conversion are welcome.
Signed-off-by: NMing Lei <tom.leiming@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Modify sysfs bin files so that we can remove the bin file while they are
still mapped.  When the kobject is removed we unmap the bin file and
arrange for future accesses to the mapping to receive SIGBUS.

Implementing this prevents a nasty DOS when pci devices are hot plugged
and unplugged.  Where if any of their resources were mmaped the kernel
could not free up their pci resources or release their pci data
structures.

[akpm@linux-foundation.org: remove unused var]
Signed-off-by: NEric W. Biederman <ebiederm@aristanetworks.com>
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Acked-by: NTejun Heo <tj@kernel.org>
Cc: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Nothing outside of the driver core should ever touch knode_bus, so
move it out of the public eye.

Cc: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Nothing outside of the driver core should ever touch knode_driver, so
move it out of the public eye.

Cc: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Nothing outside of the driver core should ever touch klist_children, or
knode_parent, so move them out of the public eye.

Cc: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This is to be used to move things out of struct device that no code
outside of the driver core should ever touch.

Cc: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch removes 100ms polling for driver_probe_done in
wait_for_device_probe(), and uses wait_event() instead.
Removing polling in fs initialization may lead to
a faster boot.

This patch also changes the return type of wait_for_device_done()
from int to void.

This patch is against Arjan's patch in linux-next tree.
Signed-off-by: NMing Lei <tom.leiming@gmail.com>
Acked-by: NCornelia Huck <cornelia.huck@de.ibm.com>
Reviewed-by: NArjan van de Ven <arjan@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

The sysfs_dirent serves as both an inode and a directory entry
for sysfs.  To prevent the sysfs inode numbers from being freed
prematurely hold a reference to sysfs_dirent from the sysfs inode.

[akpm@linux-foundation.org: add comment]
Signed-off-by: NEric W. Biederman <ebiederm@aristanetworks.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Cc: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

sysfs: sysfs_add_one WARNs with full path to duplicate filename

As a debugging aid, it can be useful to know the full path to a
duplicate file being created in sysfs.

We now will display warnings such as:

	sysfs: cannot create duplicate filename '/foo'

when attempting to create multiple files named 'foo' in the sysfs
root, or:

	sysfs: cannot create duplicate filename '/bus/pci/slots/5/foo'

when attempting to create multiple files named 'foo' under a
given directory in sysfs.

The path displayed is always a relative path to sysfs_root. The
leading '/' in the path name refers to the sysfs_root mount
point, and should not be confused with the "real" '/'.

Thanks to Alex Williamson for essentially writing sysfs_pathname.

Cc: Alex Williamson <alex.williamson@hp.com>
Signed-off-by: NAlex Chiang <achiang@hp.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

If the bus_type is not registerd, driver_register to that bus will cause oops.

I found this bug when test built-in usb serial drivers (ie. aircable driver)
with 'nousb' cmdline params.

In this patch:
1. set the bus->p=NULL when bus_register failed and unregisterd.
2. if bus->p is NULL, driver_register BUG_ON will be triggered.
Signed-off-by: NDave Young <hidave.darkstar@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

The patch from Ming Lei <tom.leiming@gmail.com> entitled:
    platform driver: fix incorrect use of 'platform_bus_type' with 'struct devic
introduced the following warnings on m68k, as `dev' is now a `struct
platform_device *' instead of a `struct device *':

| drivers/scsi/a4000t.c:64: warning: passing argument 3 of 'NCR_700_detect' from incompatible pointer type
| drivers/scsi/mvme16x_scsi.c:67: warning: passing argument 3 of 'NCR_700_detect' from incompatible pointer type
| drivers/scsi/bvme6000_scsi.c:61: warning: passing argument 3 of 'NCR_700_detect' from incompatible pointer type

I think the below is missing (untested on real hardware).
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>
Cc: Ming Lei <tom.leiming@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch fixes the bug reported in
	http://bugzilla.kernel.org/show_bug.cgi?id=11681.

"Lots of device drivers register a 'struct device_driver' with
the '.bus' member set to '&platform_bus_type'. This is wrong,
since the platform_bus functions expect the 'struct device_driver'
to be wrapped up in a 'struct platform_driver' which provides
some additional callbacks (like suspend_late, resume_early).
The effect may be that platform_suspend_late() uses bogus data
outside the device_driver struct as a pointer pointer to the
device driver's suspend_late() function or other hard to
reproduce failures."(Lothar Wassmann)
Signed-off-by: NMing Lei <tom.leiming@gmail.com>
Acked-by: NHenrique de Moraes Holschuh <hmh@hmh.eng.br>
Acked-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

If a UIO memory region does not start on a page boundary but straddles one,
the number of actual pages that overlap the memory region may be calculated
incorrectly because the offset isn't taken into account.  If userspace sets
the mmap length to offset+size, it may fail with -EINVAL if UIO thinks it's
trying to allocate too many pages.
Signed-off-by: NIan Abbott <abbotti@mev.co.uk>
Cc: Hans J. Koch <hjk@linutronix.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

UIO driver for the Adrienne Electronics Corporation PCI time code
device.

This device differs from other UIO devices since it uses I/O ports instead of
memory mapped I/O. In order to make it possible for UIO to work with this
device a utility, uioport, can be used to read and write the ports.

uioport is designed to be a setuid program and checks the permissions of
the /dev/uio* node and if the user has write permissions it will use
iopl and out*/in* to access the device.

[1] git clone git://ifup.org/philips/uioport.gitSigned-off-by: NBrandon Philips <brandon@ifup.org>
Signed-off-by: NHans J. Koch <hjk@linutronix.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

If a UIO device has several memory mappings, it can be difficult for userspace
to find the right one. The situation becomes even worse if the UIO driver can
handle different versions of a card that have different numbers of mappings.
Benedikt Spranger has such cards and pointed this out to me. Thanks, Bene!

To address this problem, this patch adds "name" sysfs attributes for each
mapping. Userspace can use these to clearly identify each mapping. The name
string is optional. If a driver doesn't set it, an empty string will be
returned, so this patch won't break existing drivers.

The same problem exists for port region information, so a "name" attribute is
added there, too.
Signed-off-by: NHans J. Koch <hjk@linutronix.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Now platform_device is being widely used on SoC processors where the
peripherals are attached to the system bus, which is simple enough.

However, silicon IPs for these SoCs are usually shared heavily across
a family of processors, even products from different companies.  This
makes the original simple driver name based matching insufficient, or
simply not straight-forward.

Introduce a module id table for platform devices, and makes it clear
that a platform driver is able to support some shared IP and handle
slight differences across different platforms (by 'driver_data').
Module alias is handled automatically when a MODULE_DEVICE_TABLE()
is defined.

To not disturb the current platform drivers too much, the matched id
entry is recorded and can be retrieved by platform_get_device_id().
Signed-off-by: NEric Miao <eric.miao@marvell.com>
Cc: Kay Sievers <kay.sievers@vrfy.org>
Cc: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This helps the code look more consistent and cleaner.
Signed-off-by: NEric Miao <eric.miao@marvell.com>
Cc: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch moves bus->match out from driver_probe_device and
does not hold device lock to check the match between a device
and a driver.

The idea has been verified by the commit 6cd49586,
which leads to a faster boot. But the commit 6cd49586 has
the following drawbacks: 1),only does the quick check in
the path of __driver_attach->driver_probe_device, not in other
paths; 2),for a matched device and driver, check the same match
twice. It is a waste of cpu ,especially for some drivers with long
device id table (eg. usb-storage driver).

This patch adds a helper of driver_match_device to check the match
in all paths, and testes the match only once.
Signed-off-by: NMing Lei <tom.leiming@gmail.com>
Acked-by: NCornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

sysfs_get_inode ultimately calls sysfs_count_nlink when the a
directory inode is fectched.  sysfs_count_nlink needs to be
called under the sysfs_mutex to guard against the unlikely
but possible scenario that the root directory is changing
as we are counting the number entries in it, and just in
general to be consistent.
Signed-off-by: NEric W. Biederman <ebiederm@aristanetworks.com>
Acked-by: NTejun Heo <tj@kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

SYSFS_MAGIC has been added into magic.h, so only use that definition
in magic.h to avoid potential consistency problem.
Signed-off-by: NQinghuang Feng <qhfeng.kernel@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Now that all users of bus_id is gone, we can remove it from struct
device.
Signed-off-by: NKay Sievers <kay.sievers@vrfy.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Replace references to bus_id with dev_name() to fix fhci driver build break.

drivers/usb/host/fhci-hcd.c:586: error: struct device has no member named bus_id
drivers/usb/host/fhci-hcd.c:653: error: struct device has no member named bus_id
drivers/usb/host/fhci-dbg.c:111: error: struct device has no member named bus_id
Signed-off-by: NSachin Sant <sachinp@in.ibm.com>
Acked-by: NAnton Vorontsov <avorontsov@ru.mvista.com>
Cc: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

These simple debug statments should be using dev_dbg() instead of
accessing bus_id directly (or they should use device_name).

As bus_id is going away, this patch is necessary.
Acked-by: NGeoff Levand <geoffrey.levand@am.sony.com>
Cc: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Acked-by: NGreg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: NKay Sievers <kay.sievers@vrfy.org>

Cc: inaky.perez-gonzalez@intel.com
Cc: linux-wimax@intel.com
Acked-by: NGreg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: NKay Sievers <kay.sievers@vrfy.org>

Cc: sameo@openedhand.com
Acked-by: NGreg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: NKay Sievers <kay.sievers@vrfy.org>

Cc: James.Bottomley@HansenPartnership.com
Acked-by: NGreg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: NKay Sievers <kay.sievers@vrfy.org>

Cc: Michael Buesch <mb@bu3sch.de>
Acked-by: NGreg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: NKay Sievers <kay.sievers@vrfy.org>

Acked-by: NGreg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: NKay Sievers <kay.sievers@vrfy.org>

Cc: Dan Williams <dan.j.williams@intel.com>
Acked-by: NGreg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: NKay Sievers <kay.sievers@vrfy.org>

Acked-by: NGreg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: NKay Sievers <kay.sievers@vrfy.org>