openwrt/staging/blogic.git
11 years agostaging: ozwpan: Remove unneeded variable initializer
Rupesh Gujare [Thu, 15 Aug 2013 11:00:21 +0000 (12:00 +0100)]
staging: ozwpan: Remove unneeded variable initializer

We are assigning value to hport before returning, there is
no need to initialize it.

Signed-off-by: Rupesh Gujare <rupesh.gujare@atmel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: dgnc: removes proc code
Lidza Louina [Thu, 15 Aug 2013 14:28:22 +0000 (10:28 -0400)]
staging: dgnc: removes proc code

This patch removes the dgnc_proc.c and dgnc_proc.h files and all
references to proc functions in dgnc_driver.c. This also removes proc.h
from the include headers in driver.c, mgmt.c and sysfs.c and proc.o
from the Makefile.

Drivers now use sysfs instead of proc.

Signed-off-by: Lidza Louina <lidza.louina@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agodrivers/staging/comedi/drivers/das16.c: needs slab.h
Andrew Morton [Tue, 13 Aug 2013 20:55:48 +0000 (13:55 -0700)]
drivers/staging/comedi/drivers/das16.c: needs slab.h

alpha allmodconfig:

drivers/staging/comedi/drivers/das16.c: In function 'das16_detach':
drivers/staging/comedi/drivers/das16.c:1234: error: implicit declaration of function 'kfree'

Cc: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: vt6656: rxtx.c s_uGetRTSCTSDuration return endian u16
Malcolm Priestley [Tue, 13 Aug 2013 19:17:11 +0000 (20:17 +0100)]
staging: vt6656: rxtx.c s_uGetRTSCTSDuration return endian u16

The return value of uGetRTSCTSDuration always needs to be
corrected to a u16 little endian corrected value.

Some places the endian correction is done on the next line

Change uGetRTSCTSDuration return u16 little endian corrected value.

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: vt6656: rxtx.c s:_uGetDataDuration simplify structure.
Malcolm Priestley [Tue, 13 Aug 2013 18:59:31 +0000 (19:59 +0100)]
staging: vt6656: rxtx.c s:_uGetDataDuration simplify structure.

A value is only returned when bNeedAck is true.

Only when byDurType == DATADUR_B is different.

Remove switch statement and simplify with if structure.

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: vt6656: rxtx.h dead code typedef union tagUTX_BUFFER_CONTAINER
Malcolm Priestley [Tue, 13 Aug 2013 18:52:08 +0000 (19:52 +0100)]
staging: vt6656: rxtx.h dead code typedef union tagUTX_BUFFER_CONTAINER

This is part of the Remote NDIS message format with patch

vt6656: rxtx.c: s_vGetFreeContext use single tx memset.

there is not need for this structure and its members.

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: vt6656: rxtx.h always pack BEACON_BUFFER/TX_BUFFER
Malcolm Priestley [Tue, 13 Aug 2013 18:47:08 +0000 (19:47 +0100)]
staging: vt6656: rxtx.h always pack BEACON_BUFFER/TX_BUFFER

All structure members of BEACON_BUFFER/TX_BUFFER should be packed.

Only the first 4 members of these structures are live.

The forth member is referenced at run-time by packed structures.
typedef struct tagSTxBufHead
typedef struct tagSTxShortBufHead
in desc.h

If these structures are not packed the alignment of these members
will be wrong.

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: ozwpan: Separate success & failure case for oz_hcd_pd_arrived()
Dan Carpenter [Tue, 13 Aug 2013 17:29:26 +0000 (18:29 +0100)]
staging: ozwpan: Separate success & failure case for oz_hcd_pd_arrived()

This patch separates success & failure block along with fixing
following issues:-

1. The way oz_hcd_pd_arrived() looks now it's easy to think we free "ep" but
actually we do this spaghetti thing of setting it to NULL on success.

2. It is hard to read it because there are unlocks scattered throughout.

3. Currently we set "ep" to NULL on the success path and then test it and or
free it. In current code you have to scroll to the start of the function
to read code.

Original patch was submitted by Dan here :-
http://driverdev.linuxdriverproject.org/pipermail/driverdev-devel/2013-August/040113.html

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Rupesh Gujare <rupesh.gujare@atmel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: ozwpan: Swap arguments of oz_ep_alloc() to match kmalloc()
Rupesh Gujare [Tue, 13 Aug 2013 17:29:25 +0000 (18:29 +0100)]
staging: ozwpan: Swap arguments of oz_ep_alloc() to match kmalloc()

Swap arguments of oz_ep_alloc() to match kmalloc() for better readability.

Signed-off-by: Rupesh Gujare <rupesh.gujare@atmel.com>
Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: ozwpan: Remove unneeded initializers
Rupesh Gujare [Tue, 13 Aug 2013 17:29:24 +0000 (18:29 +0100)]
staging: ozwpan: Remove unneeded initializers

Remove variable initialization wherever it is not required.

Signed-off-by: Rupesh Gujare <rupesh.gujare@atmel.com>
Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: ozwpan: Make oz_hcd_pd_departed() take a struct pointer.
Rupesh Gujare [Tue, 13 Aug 2013 17:29:23 +0000 (18:29 +0100)]
staging: ozwpan: Make oz_hcd_pd_departed() take a struct pointer.

oz_hcd_pd_departed() takes struct oz_port pointer instead of
void *, change function declaration to avoid ambiguity.

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Rupesh Gujare <rupesh.gujare@atmel.com>
Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: ozwpan: Make oz_hcd_pd_arrived() return a struct pointer
Rupesh Gujare [Tue, 13 Aug 2013 17:29:22 +0000 (18:29 +0100)]
staging: ozwpan: Make oz_hcd_pd_arrived() return a struct pointer

oz_hcd_pd_arrived returns struct oz_port *, change function
declaration to avoid ambiguity.

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Rupesh Gujare <rupesh.gujare@atmel.com>
Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: ozwpan: Remove unnecessary pointer check.
Rupesh Gujare [Tue, 13 Aug 2013 17:29:21 +0000 (18:29 +0100)]
staging: ozwpan: Remove unnecessary pointer check.

We are already checking "ep" earlier in function. Do not
need to check again.

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Rupesh Gujare <rupesh.gujare@atmel.com>
Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: ozwpan: Fix coding style.
Rupesh Gujare [Tue, 13 Aug 2013 17:24:24 +0000 (18:24 +0100)]
staging: ozwpan: Fix coding style.

Put spaces around math operations.

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Rupesh Gujare <rupesh.gujare@atmel.com>
Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: ozwpan: Simply if condition
Rupesh Gujare [Tue, 13 Aug 2013 17:24:23 +0000 (18:24 +0100)]
staging: ozwpan: Simply if condition

Making code simpler for readability.

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Rupesh Gujare <rupesh.gujare@atmel.com>
Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: ozwpan: Add a blank line between functions & declarations.
Rupesh Gujare [Tue, 13 Aug 2013 17:24:22 +0000 (18:24 +0100)]
staging: ozwpan: Add a blank line between functions & declarations.

This patch adds a blank line between global declarations &
functions for readability.

Signed-off-by: Rupesh Gujare <rupesh.gujare@atmel.com>
Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: ozwpan: Add a blank line between declaraction and code.
Rupesh Gujare [Tue, 13 Aug 2013 17:24:21 +0000 (18:24 +0100)]
staging: ozwpan: Add a blank line between declaraction and code.

This patch adds blank line between declaration &
code for readability.

Signed-off-by: Rupesh Gujare <rupesh.gujare@atmel.com>
Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: dgnc: removes read_cnt, real_raw, rawreadok and buf
Lidza Louina [Tue, 13 Aug 2013 19:49:58 +0000 (15:49 -0400)]
staging: dgnc: removes read_cnt, real_raw, rawreadok and buf

This patch removes the use of read_cnt, real_raw, buf and rawreadok. The
variable buf is never used in the code. The variables rawreadok read_cnt
and real_raw don't exist in the new API. Reading the data raw is no
longer supported by the tty layer.

Signed-off-by: Lidza Louina <lidza.louina@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: dgnc: fixes ioctl param list
Lidza Louina [Tue, 13 Aug 2013 19:49:57 +0000 (15:49 -0400)]
staging: dgnc: fixes ioctl param list

The declaration for the ioctl function has changed. The previous version
of this declaration took struct file *file as a parameter and the new
one does not. This patch removes that parameter.

Signed-off-by: Lidza Louina <lidza.louina@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: dgcn: removes unnecessary commands in ioctl
Lidza Louina [Tue, 13 Aug 2013 19:49:56 +0000 (15:49 -0400)]
staging: dgcn: removes unnecessary commands in ioctl

The commands TIOCGETP, TCGETS, and TCGETA are not supposed to be seen by
the ioctl. This patch removes the switch cases for these commands.

Signed-off-by: Lidza Louina <lidza.louina@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: dgnc: dereferences ts var in dgnc_tty_close()
Lidza Louina [Tue, 13 Aug 2013 19:49:55 +0000 (15:49 -0400)]
staging: dgnc: dereferences ts var in dgnc_tty_close()

The value tty->termios needed to be dereferenced to be assigned to the
variable ts.

Signed-off-by: Lidza Louina <lidza.louina@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: dgnc: fixes tty_port and tty_struct params
Lidza Louina [Tue, 13 Aug 2013 19:49:54 +0000 (15:49 -0400)]
staging: dgnc: fixes tty_port and tty_struct params

The functions tty_flip_buffer_push, tty_insert_flip_string,
tty_insert_flip_char and tty_buffer_request_room now require a struct of
type tty_port instead of struct tty_struct. This patch makes those
changes.

Signed-off-by: Lidza Louina <lidza.louina@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: dgnc: fixes termios error
Lidza Louina [Tue, 13 Aug 2013 19:49:53 +0000 (15:49 -0400)]
staging: dgnc: fixes termios error

This patch fixes this error: invalid type argument of ‘->’ (have ‘struct ktermios’).

There were changes in the tty layer's API. Access to the termios flags changed from tty->termios->*flag* to tty->termios.*flag*.

Signed-off-by: Lidza Louina <lidza.louina@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: dgnc: removes Makefile.inc
Lidza Louina [Tue, 13 Aug 2013 19:49:51 +0000 (15:49 -0400)]
staging: dgnc: removes Makefile.inc

The Makefile contains the variables needed for the driver to build.
The Makefile.inc isn't needed.

Signed-off-by: Lidza Louina <lidza.louina@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: vt6656: rxtx.c s_uFillDataHead remove dead variables.
Malcolm Priestley [Wed, 7 Aug 2013 20:31:48 +0000 (21:31 +0100)]
staging: vt6656: rxtx.c s_uFillDataHead remove dead variables.

As result of patches
staging: vt6656: rxtx.c: s_uGetDataDuration remove First Frag or Mid Frag dead code
staging: vt6656: rxtx.c: s_uGetDataDuration remove dead variables.

variables uFragIdx, cbLastFragmentSize, uMACfragNum are dead.

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: vt6656: rxtx.c: s_uGetDataDuration remove dead variables.
Malcolm Priestley [Wed, 7 Aug 2013 20:28:45 +0000 (21:28 +0100)]
staging: vt6656: rxtx.c: s_uGetDataDuration remove dead variables.

As result of patch
staging: vt6656: rxtx.c s_uGetDataDuration remove First Frag or Mid Frag dead code.

cbFrameLength, wRate, uFragIdx, cbLastFragmentSize, uMACfragNum, byFBOption are
dead variables.

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: vt6656: rxtx.c: s_uGetDataDuration remove First Frag or Mid Frag dead code.
Malcolm Priestley [Wed, 7 Aug 2013 20:26:12 +0000 (21:26 +0100)]
staging: vt6656: rxtx.c: s_uGetDataDuration remove First Frag or Mid Frag dead code.

The driver always calls s_uGetDataDuration with;
uFragIdx = 0, cbLastFragmentSize = 0 and uMACfragNum = 1

uFragIdx == (uMACfragNum-1) is always true and bLastFrag = 1.

Remove dead First Frag or Mid Frag if/else and variables.

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agozram: don't grab mutex in zram_slot_free_noity
Minchan Kim [Mon, 12 Aug 2013 06:13:56 +0000 (15:13 +0900)]
zram: don't grab mutex in zram_slot_free_noity

[1] introduced down_write in zram_slot_free_notify to prevent race
between zram_slot_free_notify and zram_bvec_[read|write]. The race
could happen if somebody who has right permission to open swap device
is reading swap device while it is used by swap in parallel.

However, zram_slot_free_notify is called with holding spin_lock of
swap layer so we shouldn't avoid holing mutex. Otherwise, lockdep
warns it.

This patch adds new list to handle free slot and workqueue
so zram_slot_free_notify just registers slot index to be freed and
registers the request to workqueue. If workqueue is expired,
it holds mutex_lock so there is no problem any more.

If any I/O is issued, zram handles pending slot-free request
caused by zram_slot_free_notify right before handling issued
request because workqueue wouldn't be expired yet so zram I/O
request handling function can miss it.

Lastly, when zram is reset, flush_work could handle all of pending
free request so we shouldn't have memory leak.

NOTE: If zram_slot_free_notify's kmalloc with GFP_ATOMIC would be
failed, the slot will be freed when next write I/O write the slot.

[1] [57ab0485, zram: use zram->lock to protect zram_free_page()
    in swap free notify path]

* from v2
  * refactoring

* from v1
  * totally redesign

Cc: Nitin Gupta <ngupta@vflare.org>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: stable@vger.kernel.org
Signed-off-by: Minchan Kim <minchan@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agozram: fix invalid memory access
Minchan Kim [Mon, 12 Aug 2013 06:13:55 +0000 (15:13 +0900)]
zram: fix invalid memory access

[1] tried to fix invalid memory access on zram->disk but it didn't
fix properly because get_disk failed during module exit path.

Actually, we don't need to reset zram->disk's capacity to zero
in module exit path so that this patch introduces new argument
"reset_capacity" on zram_reset_divice and it only reset it when
reset_store is called.

[1] 6030ea9b,  zram: avoid invalid memory access in zram_exit()

Cc: Nitin Gupta <ngupta@vflare.org>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: stable@vger.kernel.org
Signed-off-by: Minchan Kim <minchan@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: dwc2: reorder some kernel doc comments and struct members
Paul Zimmerman [Sun, 11 Aug 2013 19:50:17 +0000 (12:50 -0700)]
staging: dwc2: reorder some kernel doc comments and struct members

Reorder the kernel doc comments for 'struct dwc2_core_params' to
match the ordering in the struct itself. Reorder the members of
'struct dwc2_qh' (and its kerneldoc comments) to minimize the
amount of structure padding.

Signed-off-by: Paul Zimmerman <paulz@synopsys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agoStaging: imx-drm: imx-tve.c Fixed 80 character line coding style issue
Andreas Werner [Sun, 11 Aug 2013 15:20:23 +0000 (17:20 +0200)]
Staging: imx-drm: imx-tve.c Fixed 80 character line coding style issue

Fixed a coding style issue of 80 character per line.

Signed-off-by: Andreas Werner <wernerandy@gmx.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: cxt1e1: Remove #define SBE_INCLUDE_SYMBOLS
Shaun Laing [Fri, 9 Aug 2013 13:54:24 +0000 (07:54 -0600)]
staging: cxt1e1: Remove #define SBE_INCLUDE_SYMBOLS

Removed the unneeded SBE_INCLUDE_SYMBOLS #define, and the associated STATIC
#define, and replaced all occurances of STATIC with 'static'.  This was in
response to sparse warnings of the form "symbol 'XYZ' was not declared. Should
it be static?".

Removed a function prototype (musycc_init_port) as adding the 'static'
declaration produced a new gcc warning. (musycc_init_port is only declared if
SBE_WAN256T3_ENABLE is set)

Signed-off-by: Shaun Laing <shaun@xresource.ca>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agoStaging: zram: zram_drv.c: Fixed Error of trailing whitespace
Kumar Gaurav [Thu, 8 Aug 2013 18:23:24 +0000 (23:53 +0530)]
Staging: zram: zram_drv.c: Fixed Error of trailing whitespace

Fixed by removing trailing whitespace

Signed-off-by: Kumar Gaurav <kumargauravgupta3@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agoStaging: rtl8192e: rtllib_rx: checking NULL value after doing dev_alloc_skb
Iker Pedrosa [Thu, 8 Aug 2013 10:56:44 +0000 (12:56 +0200)]
Staging: rtl8192e: rtllib_rx: checking NULL value after doing dev_alloc_skb

Checking the return of dev_alloc_skb as stated in the following bug:
https://bugzilla.kernel.org/show_bug.cgi?id=60401

Reported-by: RUC_Soft_Sec rucsoftsec@gmail.com
Cc: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Iker Pedrosa <ikerpedrosam@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: comedi: drivers: use comedi_dio_insn_config() for simple cases
H Hartley Sweeten [Tue, 6 Aug 2013 16:32:33 +0000 (09:32 -0700)]
staging: comedi: drivers: use comedi_dio_insn_config() for simple cases

Convert the drivers with simple, per channel programmable i/o, to use the
comedi_dio_insn_config() helper function.

All of these pass a 'mask' of '0' to comedi_dio_insn_config() this causes
the per channel mask to be used to configure the i/o direction.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: comedi: skel: use comedi_dio_insn_config()
H Hartley Sweeten [Tue, 6 Aug 2013 16:31:59 +0000 (09:31 -0700)]
staging: comedi: skel: use comedi_dio_insn_config()

Convert this driver to use the comedi_dio_insn_config() helper function.

Tidy up the comments to reflect the new code.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: comedi: drivers: introduce comedi_dio_insn_config()
H Hartley Sweeten [Tue, 6 Aug 2013 16:31:35 +0000 (09:31 -0700)]
staging: comedi: drivers: introduce comedi_dio_insn_config()

DIO subdevices always handle the INSN_CONFIG_DIO_{INPUT,OUTPUT} instructions
to configure the DIO channels. They also always handle the INSN_CONFIG_DIO_QUERY
instruction to query the configuration of a DIO channel.

Introduce a helper function to handle the (*insn_config) boilerplate for
comedi DIO subdevices. This function has the same paramters as (*insn_config)
functions with an additional parameter to allow the caller to pass a 'mask'
value for grouped DIO channels.

This function returns:

  0 if the instruction was successful but requires additional handling by
  the caller (INSN_CONFIG_DIO_{INPUT,OUTPUT}

  insn->n if the instruction was handled (INSN_CONFIG_DIO_QUERY)

  -EINVAL for all unhandled instructions

The caller is responsible for actually configuring the hardware based on
the configuration (s->io_bits).

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agoRevert "staging: comedi: skel: use comedi_dio_insn_config()"
Greg Kroah-Hartman [Mon, 12 Aug 2013 22:17:17 +0000 (15:17 -0700)]
Revert "staging: comedi: skel: use comedi_dio_insn_config()"

This reverts commit 42ef678bef336a83fd0ae0b03a56c0a93665a18b.

It's incorrect as well... time for more coffee...

Cc: H Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agoRevert "staging: comedi: core: introduce comedi_dio_insn_config()"
Greg Kroah-Hartman [Mon, 12 Aug 2013 22:16:24 +0000 (15:16 -0700)]
Revert "staging: comedi: core: introduce comedi_dio_insn_config()"

This reverts commit 4f76463d3b8f8cc0cac5bb292ec766848f3f4fa1.

I applied an incorrect version here as well :(

Cc: H Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: comedi: drivers: use comedi_dio_insn_config() for complex cases
H Hartley Sweeten [Tue, 6 Aug 2013 16:33:05 +0000 (09:33 -0700)]
staging: comedi: drivers: use comedi_dio_insn_config() for complex cases

Convert the drivers with complex, port programmable i/o, to use the
comedi_dio_insn_config() helper function.

All of these drivers have some sort of 'port' programmable i/o where multiple
i/o channels are configured as a group. The 'mask' associated with the group
is passed to comedi_dio_insn_config() so that all the channels are configured.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: lustre: re-use kbasename() common helper
Andy Shevchenko [Mon, 12 Aug 2013 11:39:22 +0000 (14:39 +0300)]
staging: lustre: re-use kbasename() common helper

To get the name of a file from a pathname we may use kbasename() helper.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agoRevert "staging: comedi: drivers: use comedi_dio_insn_config() for complex cases"
Greg Kroah-Hartman [Mon, 12 Aug 2013 22:14:18 +0000 (15:14 -0700)]
Revert "staging: comedi: drivers: use comedi_dio_insn_config() for complex cases"

This reverts commit f21c53945cb95f66faa9636af5f23cb00ba73019.

I applied the wrong patch :(

Cc: H Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: zcache: delete it
Greg Kroah-Hartman [Mon, 12 Aug 2013 22:07:49 +0000 (15:07 -0700)]
staging: zcache: delete it

zcache is obsolete and not used anymore, Bob Liu has rewritten it and is
submitting it for inclusion through the main -mm tree, as it should have
been done in the first place...

Cc: Bob Liu <lliubbo@gmail.com>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Kyungmin Park <kmpark@infradead.org>
Cc: Wanpeng Li <liwanp@linux.vnet.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: comedi: drivers: use comedi_dio_insn_config() for complex cases
H Hartley Sweeten [Mon, 5 Aug 2013 22:01:25 +0000 (15:01 -0700)]
staging: comedi: drivers: use comedi_dio_insn_config() for complex cases

Convert the drivers with complex, port programmable i/o, to use the
comedi_dio_insn_config() helper function.

All of these drivers have some sort of 'port' programmable i/o where multiple
i/o channels are configured as a group. The 'mask' associated with the group
is passed to comedi_dio_insn_config() so that all the channels are configured.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: comedi: skel: use comedi_dio_insn_config()
H Hartley Sweeten [Mon, 5 Aug 2013 22:00:36 +0000 (15:00 -0700)]
staging: comedi: skel: use comedi_dio_insn_config()

Convert this driver to use the comedi_dio_insn_config() helper function.

Tidy up the comments to reflect the new code.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: comedi: core: introduce comedi_dio_insn_config()
H Hartley Sweeten [Mon, 5 Aug 2013 22:00:05 +0000 (15:00 -0700)]
staging: comedi: core: introduce comedi_dio_insn_config()

DIO subdevices always handle the INSN_CONFIG_DIO_{INPUT,OUTPUT} instructions
to configure the dio channels. They also always handle the INSN_CONFIG_DIO_QUERY
instruction to query the configuration of a dio channel.

Introduce a helper function to handle the (*insn_config) boilerplate for
comedi DIO subdevices. This function has the same parameters as (*insn_config)
functions with an additional parameter to allow the caller to pass a 'mask'
value for grouped dio channels.

This function returns:

  0 if the instruction was successful but requires additional handling by
  the caller (INSN_CONFIG_DIO_{INPUT,OUTPUT}

  insn->n if the instruction was handled (INSN_CONFIG_DIO_QUERY)

  -EINVAL for all unhandled instructions

The caller is responsible for actually configuring the hardware based on
the configuration (s->io_bits).

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: vt6656: desc.h remove typedef struct tagSCTSData
Malcolm Priestley [Mon, 5 Aug 2013 21:12:42 +0000 (22:12 +0100)]
staging: vt6656: desc.h remove typedef struct tagSCTSData

Replace structures with struct ieee80211_cts data

typedef struct tagSCTSData include extra packing wReserved
not in struct ieee80211_cts

This is corrected by adding reserved2 to the structures.

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: vt6656: desc.h remove typedef struct tagSRTSData
Malcolm Priestley [Mon, 5 Aug 2013 21:08:05 +0000 (22:08 +0100)]
staging: vt6656: desc.h remove typedef struct tagSRTSData

Replace structures with struct ieee80211_rts data.

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: vt6656: rxtx.c: s_vGetFreeContext use single tx memset.
Malcolm Priestley [Mon, 5 Aug 2013 20:09:14 +0000 (21:09 +0100)]
staging: vt6656: rxtx.c: s_vGetFreeContext use single tx memset.

memset to 0 structure _USB_SEND_CONTEXT pContext->Data to its
size of MAX_TOTAL_SIZE_WITH_ALL_HEADERS as defined
in device.h

This removes the need for memset in the TX path.

It also allows the removal of dead structure
TX_BUFFER_CONTAINER BufferHeader;

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: ozwpan: Return correct hub status.
Rupesh Gujare [Mon, 5 Aug 2013 17:40:15 +0000 (18:40 +0100)]
staging: ozwpan: Return correct hub status.

Fix a bug where we were not returning correct hub status
for 8th port.

Signed-off-by: Rupesh Gujare <rupesh.gujare@atmel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: ozwpan: Reset port configuration number.
Rupesh Gujare [Mon, 5 Aug 2013 17:40:14 +0000 (18:40 +0100)]
staging: ozwpan: Reset port configuration number.

Make sure that we reset port configuration no. when PD departs.

Signed-off-by: Rupesh Gujare <rupesh.gujare@atmel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agoStaging: winbond: mds_f: deleted extern function
Iker Pedrosa [Mon, 5 Aug 2013 09:58:32 +0000 (11:58 +0200)]
Staging: winbond: mds_f: deleted extern function

Deleted an extern function that wasn't used in this driver.

Signed-off-by: Iker Pedrosa <ikerpedrosam@gmail.com>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agoStaging: winbond: mds_f: fixed lines over 80 characters
Iker Pedrosa [Mon, 5 Aug 2013 09:58:31 +0000 (11:58 +0200)]
Staging: winbond: mds_f: fixed lines over 80 characters

Fixed some coding style issues.

Signed-off-by: Iker Pedrosa <ikerpedrosam@gmail.com>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agoStaging: winbond: mds: fixed some lines over 80 characters
Iker Pedrosa [Mon, 5 Aug 2013 09:58:30 +0000 (11:58 +0200)]
Staging: winbond: mds: fixed some lines over 80 characters

Fixed some coding style issues

Signed-off-by: Iker Pedrosa <ikerpedrosam@gmail.com>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agoStaging: winbond: wb35reg: avoided use of extern function
Iker Pedrosa [Mon, 5 Aug 2013 09:58:29 +0000 (11:58 +0200)]
Staging: winbond: wb35reg: avoided use of extern function

Prototype of a function added to the header to avoid the use of extern.

Signed-off-by: Iker Pedrosa <ikerpedrosam@gmail.com>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agoStaging: winbond: wb35reg: white space deleted
Iker Pedrosa [Mon, 5 Aug 2013 09:58:28 +0000 (11:58 +0200)]
Staging: winbond: wb35reg: white space deleted

White space deleted before a semicolon.

Signed-off-by: Iker Pedrosa <ikerpedrosam@gmail.com>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agoStaging: xgifb: Fixed a warning of Space before semicolon
Kumar Gaurav [Sun, 4 Aug 2013 18:41:02 +0000 (00:11 +0530)]
Staging: xgifb: Fixed a warning of Space before semicolon

Fixed issue by removing Space before semicolon

Signed-off-by: Kumar Gaurav <kumargauravgupta3@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agoStaging: olpc_dcon: updating MAINTAINERS
Jens Frederich [Sun, 4 Aug 2013 18:57:43 +0000 (20:57 +0200)]
Staging: olpc_dcon: updating MAINTAINERS

Andres Salomon and Chris Ball are not longer at OLPC.
Daniel Drake is the new OLPC go-to guy. And I wanna
also be CC'd on DCON and other OLPC patches.

Signed-off-by: Jens Frederich <jfrederich@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: ozwpan: Fixes crash due to invalid port aceess.
Rupesh Gujare [Mon, 5 Aug 2013 17:40:12 +0000 (18:40 +0100)]
staging: ozwpan: Fixes crash due to invalid port aceess.

This patch fixes kernel crash issue, when we receive URB request
after de-enumerating device.

Signed-off-by: Rupesh Gujare <rupesh.gujare@atmel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: ozwpan: Fix farewell report.
Rupesh Gujare [Mon, 5 Aug 2013 11:28:33 +0000 (12:28 +0100)]
staging: ozwpan: Fix farewell report.

This patch fix following issues reported by Dan:-

1) There is no check limiting the size to 32 and it could be up to
   253 bytes.
2) Use defines instead of magic numbers.
3) The oz_farewell struct is supposed to be a variable length struct
   but the variable part is put in the middle.  It doesn't make any
   sense to put the length of the variable size array after then end
   of the array because we can never find it again!  Put the
   variable size array at the end.  Make it a zero length array.
   u8 len;
   u8 report[0];
4) In oz_add_farewell() we do this:

f = kmalloc(sizeof(struct oz_farewell) + len - 1, GFP_ATOMIC);

    The "- 1" refers to sizeof(f->report) but because it was a magic
    number then it was missed when the sizeof(f->report) changed.
5) In [patch 6/6] we set the ->len member.  But because it is at the
   end of a variable length array with no limit check the remote
   attacker can just rewrite it using the memcpy() on the next line.

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Rupesh Gujare <rupesh.gujare@atmel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agostaging: ozwpan: Fix build warning.
Rupesh Gujare [Mon, 5 Aug 2013 11:14:04 +0000 (12:14 +0100)]
staging: ozwpan: Fix build warning.

This patch fixes following build warning.

drivers/built-in.o: In function `oz_hcd_heartbeat':
>> (.text+0x30aadd): undefined reference to `__divdi3'
   drivers/built-in.o: In function `oz_hcd_heartbeat':
>> (.text+0x30ac85): undefined reference to `__divdi3'

Reported-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Rupesh Gujare <rupesh.gujare@atmel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agoMerge 3.11-rc5 into staging-next
Greg Kroah-Hartman [Mon, 12 Aug 2013 17:22:49 +0000 (10:22 -0700)]
Merge 3.11-rc5 into staging-next

We want the staging fixes in here as well.

Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agoLinux 3.11-rc5
Linus Torvalds [Mon, 12 Aug 2013 01:04:20 +0000 (18:04 -0700)]
Linux 3.11-rc5

11 years agoMerge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Linus Torvalds [Sun, 11 Aug 2013 23:32:26 +0000 (16:32 -0700)]
Merge tag 'scsi-fixes' of git://git./linux/kernel/git/jejb/scsi

Pull SCSI fixes from James Bottomley:
 "This is three bug fixes: An fnic warning caused by sleeping under a
  lock, a major regression with our updated WRITE SAME/UNMAP logic which
  caused tons of USB devices (and one RAID card) to cease to function
  and a megaraid_sas firmware initialisation problem which causes kdump
  failures"

* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
  [SCSI] Don't attempt to send extended INQUIRY command if skip_vpd_pages is set
  [SCSI] fnic: BUG: sleeping function called from invalid context during probe
  [SCSI] megaraid_sas: megaraid_sas driver init fails in kdump kernel

11 years agoMerge branch 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc
Linus Torvalds [Sun, 11 Aug 2013 19:12:39 +0000 (12:12 -0700)]
Merge branch 'merge' of git://git./linux/kernel/git/benh/powerpc

Pull powerpc fixes from Ben Herrenschmidt:
 "This includes small series from Michael Neuling to fix a couple of
  nasty remaining problems with the new Power8 support, also targeted at
  stable 3.10, without which some new userspace accessible registers
  aren't properly context switched, and in some case, can be clobbered
  by the user of transactional memory.

  Along with that, a few slightly more minor things, such as a missing
  Kconfig option to enable handling of denorm exceptions when not
  running under a hypervisor (or userspace will randomly crash when
  hitting denorms with the vector unit), some nasty bugs in the new
  pstore oops code, and other simple bug fixes worth having in now.

  Note: I picked up the two powerpc KVM fixes as Alex Graf asked me to
  handle KVM bits while he is on vacation.  However I'll let him decide
  whether they should go to -stable or not when he is back"

* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc:
  powerpc/tm: Fix context switching TAR, PPR and DSCR SPRs
  powerpc: Save the TAR register earlier
  powerpc: Fix context switch DSCR on POWER8
  powerpc: Rework setting up H/FSCR bit definitions
  powerpc: Fix hypervisor facility unavaliable vector number
  powerpc/kvm/book3s_pr: Return appropriate error when allocation fails
  powerpc/kvm: Add signed type cast for comparation
  powerpc/eeh: Add missing procfs entry for PowerNV
  powerpc/pseries: Add backward compatibilty to read old kernel oops-log
  powerpc/pseries: Fix buffer overflow when reading from pstore
  powerpc: On POWERNV enable PPC_DENORMALISATION by default

11 years agoMerge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Linus Torvalds [Sun, 11 Aug 2013 19:11:33 +0000 (12:11 -0700)]
Merge tag 'for-linus' of git://git./virt/kvm/kvm

Pull s390 kvm fixes from Paolo Bonzini:
 "Two fixes for s390"

* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
  KVM: s390: fix pfmf non-quiescing control handling
  KVM: s390: move kvm_guest_enter,exit closer to sie

11 years agoMerge branch 'i2c/for-current' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa...
Linus Torvalds [Sun, 11 Aug 2013 19:10:47 +0000 (12:10 -0700)]
Merge branch 'i2c/for-current' of git://git./linux/kernel/git/wsa/linux

Pull i2c fixes from Wolfram Sang:
 "Some driver bugfixes for the I2C subsystem"

* 'i2c/for-current' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:
  i2c: mv64xxx: Document the newly introduced allwinner compatible
  i2c: Fix Kontron PLD prescaler calculation
  i2c: i2c-mxs: Use DMA mode even for small transfers

11 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux...
Linus Torvalds [Sat, 10 Aug 2013 22:21:47 +0000 (15:21 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/mason/linux-btrfs

Pull btrfs fixes from Chris Mason:
 "These are assorted fixes, mostly from Josef nailing down xfstests
  runs.  Zach also has a long standing fix for problems with readdir
  wrapping f_pos (or ctx->pos)

  These patches were spread out over different bases, so I rebased
  things on top of rc4 and retested overnight"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs:
  btrfs: don't loop on large offsets in readdir
  Btrfs: check to see if root_list is empty before adding it to dead roots
  Btrfs: release both paths before logging dir/changed extents
  Btrfs: allow splitting of hole em's when dropping extent cache
  Btrfs: make sure the backref walker catches all refs to our extent
  Btrfs: fix backref walking when we hit a compressed extent
  Btrfs: do not offset physical if we're compressed
  Btrfs: fix extent buffer leak after backref walking
  Btrfs: fix a bug of snapshot-aware defrag to make it work on partial extents
  btrfs: fix file truncation if FALLOC_FL_KEEP_SIZE is specified

11 years agoMerge tag 'nfs-for-3.11-4' of git://git.linux-nfs.org/projects/trondmy/linux-nfs
Linus Torvalds [Sat, 10 Aug 2013 22:20:37 +0000 (15:20 -0700)]
Merge tag 'nfs-for-3.11-4' of git://git.linux-nfs.org/projects/trondmy/linux-nfs

Pull NFS client bugfixes from Trond Myklebust:

 - Stable patch for lockd to fix Oopses due to inappropriate calls to
   utsname()->nodename

 - Stable patches for sunrpc to fix Oopses on shutdown when using
   AF_LOCAL sockets with rpcbind

 - Fix memory leak and error checking issues in nfs4_proc_lookup_mountpoint

 - Fix a regression with the sync mount option failing to work for nfs4
   mounts

 - Fix a writeback performance issue when doing cache invalidation

 - Remove an incorrect call to nfs_setsecurity in nfs_fhget

* tag 'nfs-for-3.11-4' of git://git.linux-nfs.org/projects/trondmy/linux-nfs:
  NFSv4: Fix up nfs4_proc_lookup_mountpoint
  NFS: Remove unnecessary call to nfs_setsecurity in nfs_fhget()
  NFSv4: Fix the sync mount option for nfs4 mounts
  NFS: Fix writeback performance issue on cache invalidation
  SUNRPC: If the rpcbind channel is disconnected, fail the call to unregister
  SUNRPC: Don't auto-disconnect from the local rpcbind socket
  LOCKD: Don't call utsname()->nodename from nlmclnt_setlockargs

11 years agoMerge branch 'for-3.11' of git://linux-nfs.org/~bfields/linux
Linus Torvalds [Sat, 10 Aug 2013 22:19:58 +0000 (15:19 -0700)]
Merge branch 'for-3.11' of git://linux-nfs.org/~bfields/linux

Pull nfsd fixes from Bruce Fields:
 "Some fixes for a 4.1 feature that in retrospect probably should have
  waited for 3.12....  But it appears to be working now"

* 'for-3.11' of git://linux-nfs.org/~bfields/linux:
  nfsd: Fix SP4_MACH_CRED negotiation in EXCHANGE_ID
  nfsd4: Fix MACH_CRED NULL dereference

11 years agoMerge tag 'sound-3.11' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound
Linus Torvalds [Sat, 10 Aug 2013 20:00:56 +0000 (13:00 -0700)]
Merge tag 'sound-3.11' of git://git./linux/kernel/git/tiwai/sound

Pull sound fixes from Takashi Iwai:
 "A couple of USB-audio fixes that should also go to stable kernels"

* tag 'sound-3.11' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
  ALSA: usb-audio: do not trust too-big wMaxPacketSize values
  ALSA: 6fire: fix DMA issues with URB transfer_buffer usage

11 years agoMerge tag 'staging-3.11-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh...
Linus Torvalds [Sat, 10 Aug 2013 16:00:51 +0000 (09:00 -0700)]
Merge tag 'staging-3.11-rc5' of git://git./linux/kernel/git/gregkh/staging

Pull staging driver fixes from Greg KH:
 "Here are 3 small fixes for staging/IIO drivers for 3.11-rc5.  Nothing
  huge, two IIO driver fixes, and a zcache fix.  All of these have been
  in linux-next for a while"

* tag 'staging-3.11-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging:
  staging: zcache: fix "zcache=" kernel parameter
  iio: ti_am335x_adc: Fix wrong samples received on 1st read
  iio:trigger: Fix use_count race condition

11 years agoMerge tag 'usb-3.11-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb
Linus Torvalds [Sat, 10 Aug 2013 16:00:21 +0000 (09:00 -0700)]
Merge tag 'usb-3.11-rc5' of git://git./linux/kernel/git/gregkh/usb

Pull USB fixes from Greg KH:
 "Here are 3 small USB fixes for 3.11-rc5.

  One is a fix that the ChromeOS developers ran into on some Intel
  hardware, one is a build fix, and the last is a MAINTAINERS update to
  help people figure out where to send USB network driver patches.

  All of these have been in linux-next for a while"

* tag 'usb-3.11-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb:
  MAINTAINERS: Add separate section for USB NETWORKING DRIVERS
  usb: xhci: add missing dma-mapping.h includes
  usb: core: don't try to reset_device() a port that got just disconnected

11 years agobtrfs: don't loop on large offsets in readdir
Zach Brown [Thu, 11 Jul 2013 23:19:42 +0000 (16:19 -0700)]
btrfs: don't loop on large offsets in readdir

When btrfs readdir() hits the last entry it sets the readdir offset to a
huge value to stop buggy apps from breaking when the same name is
returned by readdir() with concurrent rename()s.

But unconditionally setting the offset to INT_MAX causes readdir() to
loop returning any entries with offsets past INT_MAX.  It only takes a
few hours of constant file creation and removal to create entries past
INT_MAX.

So let's set the huge offset to LLONG_MAX if the last entry has already
overflowed 32bit loff_t.   Without large offsets behaviour is identical.
With large offsets 64bit apps will work and 32bit apps will be no more
broken than they currently are if they see large offsets.

Signed-off-by: Zach Brown <zab@redhat.com>
Signed-off-by: Josef Bacik <jbacik@fusionio.com>
Signed-off-by: Chris Mason <chris.mason@fusionio.com>
11 years agoBtrfs: check to see if root_list is empty before adding it to dead roots
Josef Bacik [Thu, 25 Jul 2013 19:11:47 +0000 (15:11 -0400)]
Btrfs: check to see if root_list is empty before adding it to dead roots

A user reported a panic when running with autodefrag and deleting snapshots.
This is because we could end up trying to add the root to the dead roots list
twice.  To fix this check to see if we are empty before adding ourselves to the
dead roots list.  Thanks,

Signed-off-by: Josef Bacik <jbacik@fusionio.com>
Signed-off-by: Chris Mason <chris.mason@fusionio.com>
11 years agoBtrfs: release both paths before logging dir/changed extents
Josef Bacik [Mon, 22 Jul 2013 16:54:30 +0000 (12:54 -0400)]
Btrfs: release both paths before logging dir/changed extents

The ceph guys tripped over this bug where we were still holding onto the
original path that we used to copy the inode with when logging.  This is based
on Chris's fix which was reported to fix the problem.  We need to drop the paths
in two cases anyway so just move the drop up so that we don't have duplicate
code.  Thanks,

Cc: stable@vger.kernel.org
Signed-off-by: Josef Bacik <jbacik@fusionio.com>
Signed-off-by: Chris Mason <chris.mason@fusionio.com>
11 years agoBtrfs: allow splitting of hole em's when dropping extent cache
Josef Bacik [Thu, 11 Jul 2013 14:34:59 +0000 (10:34 -0400)]
Btrfs: allow splitting of hole em's when dropping extent cache

I noticed while running multi-threaded fsync tests that sometimes fsck would
complain about an improper gap.  This happens because we fail to add a hole
extent to the file, which was happening when we'd split a hole EM because
btrfs_drop_extent_cache was just discarding the whole em instead of splitting
it.  So this patch fixes this by allowing us to split a hole em properly, which
means that added holes actually get logged properly and we no longer see this
fsck error.  Thankfully we're tolerant of these sort of problems so a user would
not see any adverse effects of this bug, other than fsck complaining.  Thanks,

Signed-off-by: Josef Bacik <jbacik@fusionio.com>
Signed-off-by: Chris Mason <chris.mason@fusionio.com>
11 years agoBtrfs: make sure the backref walker catches all refs to our extent
Josef Bacik [Fri, 5 Jul 2013 18:03:47 +0000 (14:03 -0400)]
Btrfs: make sure the backref walker catches all refs to our extent

Because we don't mess with the offset into the extent for compressed we will
properly find both extents for this case

[extent a][extent b][rest of extent a]

but because we already added a ref for the front half we won't add the inode
information for the second half.  This causes us to leak that memory and not
print out the other offset when we do logical-resolve.  So fix this by calling
ulist_add_merge and then add our eie to the existing entry if there is one.
With this patch we get both offsets out of logical-resolve.  With this and the
other 2 patches I've sent we now pass btrfs/276 on my vm with compress-force=lzo
set.  Thanks,

Signed-off-by: Josef Bacik <jbacik@fusionio.com>
Signed-off-by: Chris Mason <chris.mason@fusionio.com>
11 years agoBtrfs: fix backref walking when we hit a compressed extent
Josef Bacik [Fri, 5 Jul 2013 17:58:19 +0000 (13:58 -0400)]
Btrfs: fix backref walking when we hit a compressed extent

If you do btrfs inspect-internal logical-resolve on a compressed extent that has
been partly overwritten it won't find anything.  This is because we try and
match the extent offset we've searched for based on the extent offset in the
data extent entry.  However this doesn't work for compressed extents because the
offsets are for the uncompressed size, not the compressed size.  So instead only
do this check if we are not compressed, that way we can get an actual entry for
the physical offset rather than nothing for compressed.  Thanks,

Signed-off-by: Josef Bacik <jbacik@fusionio.com>
Signed-off-by: Chris Mason <chris.mason@fusionio.com>
11 years agoBtrfs: do not offset physical if we're compressed
Josef Bacik [Fri, 5 Jul 2013 17:52:51 +0000 (13:52 -0400)]
Btrfs: do not offset physical if we're compressed

xfstest btrfs/276 was freaking out on slower boxes partly because fiemap was
offsetting the physical based on the extent offset.  This is perfectly fine with
uncompressed extents, however the extent offset is into the uncompressed area,
not the compressed.  So we can return a physical value that isn't at all within
the area we have allocated on disk.  Fix this by returning the start of the
extent if it is compressed no matter what the offset.  Thanks,

Signed-off-by: Josef Bacik <jbacik@fusionio.com>
Signed-off-by: Chris Mason <chris.mason@fusionio.com>
11 years agoBtrfs: fix extent buffer leak after backref walking
Liu Bo [Wed, 3 Jul 2013 06:40:44 +0000 (14:40 +0800)]
Btrfs: fix extent buffer leak after backref walking

commit 47fb091fb787420cd195e66f162737401cce023f(Btrfs: fix unlock after free on rewinded tree blocks)
takes an extra increment on the reference of allocated dummy extent buffer, so now we
cannot free this dummy one, and end up with extent buffer leak.

Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
Reviewed-by: Jan Schmidt <list.btrfs@jan-o-sch.net>
Signed-off-by: Josef Bacik <jbacik@fusionio.com>
Signed-off-by: Chris Mason <chris.mason@fusionio.com>
11 years agoBtrfs: fix a bug of snapshot-aware defrag to make it work on partial extents
Liu Bo [Mon, 1 Jul 2013 14:13:26 +0000 (22:13 +0800)]
Btrfs: fix a bug of snapshot-aware defrag to make it work on partial extents

For partial extents, snapshot-aware defrag does not work as expected,
since
a) we use the wrong logical offset to search for parents, which should be
   disk_bytenr + extent_offset, not just disk_bytenr,
b) 'offset' returned by the backref walking just refers to key.offset, not
   the 'offset' stored in btrfs_extent_data_ref which is
   (key.offset - extent_offset).

The reproducer:
$ mkfs.btrfs sda
$ mount sda /mnt
$ btrfs sub create /mnt/sub
$ for i in `seq 5 -1 1`; do dd if=/dev/zero of=/mnt/sub/foo bs=5k count=1 seek=$i conv=notrunc oflag=sync; done
$ btrfs sub snap /mnt/sub /mnt/snap1
$ btrfs sub snap /mnt/sub /mnt/snap2
$ sync; btrfs filesystem defrag /mnt/sub/foo;
$ umount /mnt
$ btrfs-debug-tree sda (Here we can check whether the defrag operation is snapshot-awared.

This addresses the above two problems.

Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
Signed-off-by: Josef Bacik <jbacik@fusionio.com>
Signed-off-by: Chris Mason <chris.mason@fusionio.com>
11 years agobtrfs: fix file truncation if FALLOC_FL_KEEP_SIZE is specified
Jie Liu [Fri, 28 Jun 2013 05:15:52 +0000 (13:15 +0800)]
btrfs: fix file truncation if FALLOC_FL_KEEP_SIZE is specified

Create a small file and fallocate it to a big size with
FALLOC_FL_KEEP_SIZE option, then truncate it back to the
small size again, the disk free space is not changed back
in this case. i.e,

total 4
-rw-r--r-- 1 root root 512 Jun 28 11:35 test

Filesystem      Size  Used Avail Use% Mounted on
....
/dev/sdb1       8.0G   56K  7.2G   1% /mnt

-rw-r--r-- 1 root root 512 Jun 28 11:35 /mnt/test

Filesystem      Size  Used Avail Use% Mounted on
....
/dev/sdb1       8.0G  5.1G  2.2G  70% /mnt

Filesystem      Size  Used Avail Use% Mounted on
....
/dev/sdb1       8.0G  5.1G  2.2G  70% /mnt

With this fix, the truncated up space is back as:
Filesystem      Size  Used Avail Use% Mounted on
....
/dev/sdb1       8.0G   56K  7.2G   1% /mnt

Signed-off-by: Jie Liu <jeff.liu@oracle.com>
Signed-off-by: Josef Bacik <jbacik@fusionio.com>
Signed-off-by: Chris Mason <chris.mason@fusionio.com>
11 years agoMerge tag 'pm+acpi-3.11-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael...
Linus Torvalds [Fri, 9 Aug 2013 22:07:19 +0000 (15:07 -0700)]
Merge tag 'pm+acpi-3.11-rc5' of git://git./linux/kernel/git/rafael/linux-pm

Pull ACPI and power management fixes from Rafael Wysocki:

 - ACPI-based memory hotplug stopped working after a recent change,
   because it's not possible to associate sufficiently many "physical"
   devices with one ACPI device object due to an artificial limit.  Fix
   from Rafael J Wysocki removes that limit and makes memory hotplug
   work again.

 - A change made in 3.9 uncovered a bug in the ACPI processor driver
   preventing NUMA nodes from being put offline due to an ordering
   issue.  Fix from Yasuaki Ishimatsu changes the ordering to make
   things work again.

 - One of the recent ACPI video commits (that hasn't been reverted so
   far) uncovered a bug in the code handling quirky BIOSes that caused
   some Asus machines to boot with backlight completely off which made
   it quite difficult to use them afterward.  Fix from Felipe Contreras
   improves the quirk to cover this particular case correctly.

 - A cpufreq user space interface change made in 3.10 inadvertently
   renamed the ignore_nice_load sysfs attribute to ignore_nice which
   resulted in some confusion.  Fix from Viresh Kumar changes the name
   back to ignore_nice_load.

 - An initialization ordering change made in 3.9 broke cpufreq on
   loongson2 boards.  Fix from Aaro Koskinen restores the correct
   initialization ordering there.

 - Fix breakage resulting from a mistake made in 3.9 and causing the
   detection of some graphics adapters (that were detected correctly
   before) to fail.  There are two objects representing the same PCIe
   port in the affected systems' ACPI tables and both appear as
   "enabled" and we are expected to guess which one to use.  We used to
   choose the right one before by pure luck, but when we tried to
   address another similar corner case, the luck went away.  This time
   we try to make our guessing a bit more educated which is reported to
   work on those systems.

 - The /proc/acpi/wakeup interface code is missing some locking which
   may lead to breakage if that file is written or read during hotplug
   of wakeup devices.  That should be rare but still possible, so it's
   better to start using the appropriate locking there.

* tag 'pm+acpi-3.11-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  ACPI: Try harder to resolve _ADR collisions for bridges
  cpufreq: rename ignore_nice as ignore_nice_load
  cpufreq: loongson2: fix regression related to clock management
  ACPI / processor: move try_offline_node() after acpi_unmap_lsapic()
  ACPI: Drop physical_node_id_bitmap from struct acpi_device
  ACPI / PM: Walk physical_node_list under physical_node_lock
  ACPI / video: improve quirk check in acpi_video_bqc_quirk()

11 years agoMerge tag 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck...
Linus Torvalds [Fri, 9 Aug 2013 22:06:17 +0000 (15:06 -0700)]
Merge tag 'hwmon-for-linus' of git://git./linux/kernel/git/groeck/linux-staging

Pull hwmon fix from Guenter Roeck:
 "Fix bug in adt7470 driver which causes it to fail writing fan speed
  limits"

* tag 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging:
  hwmon: (adt7470) Fix incorrect return code check

11 years agoMerge branch 'v4l_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab...
Linus Torvalds [Fri, 9 Aug 2013 22:04:09 +0000 (15:04 -0700)]
Merge branch 'v4l_for_linus' of git://git./linux/kernel/git/mchehab/linux-media

Pull media fixes from Mauro Carvalho Chehab:
 "Some driver fixes (em28xx, coda, usbtv, s5p, hdpvr and ml86v7667) and
  a fix for media DocBook"

* 'v4l_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media:
  [media] em28xx: fix assignment of the eeprom data
  [media] hdpvr: fix iteration over uninitialized lists in hdpvr_probe()
  [media] usbtv: fix dependency
  [media] usbtv: Throw corrupted frames away
  [media] usbtv: Fix deinterlacing
  [media] v4l2: added missing mutex.h include to v4l2-ctrls.h
  [media] DocBook: upgrade media_api DocBook version to 4.2
  [media] ml86v7667: fix compile warning: 'ret' set but not used
  [media] s5p-g2d: Fix registration failure
  [media] media: coda: Fix DT driver data pointer for i.MX27
  [media] s5p-mfc: Fix input/output format reporting

11 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid
Linus Torvalds [Fri, 9 Aug 2013 18:53:06 +0000 (11:53 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/jikos/hid

Pull HID fix from Jiri Kosina:
 "Revert of a patch which breaks enumeration workaround in
  hid-logitech-dj"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid:
  Revert "HID: hid-logitech-dj: querying_devices was never set"

11 years agoMerge tag 'fbdev-fixes-3.11-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 9 Aug 2013 18:52:34 +0000 (11:52 -0700)]
Merge tag 'fbdev-fixes-3.11-rc5' of git://git./linux/kernel/git/tomba/linux

Pull fbdev fixes from Tomi Valkeinen:
 - omapdss: compilation fix and DVI fix for PandaBoard
 - mxsfb: fix colors when using 18bit LCD bus

* tag 'fbdev-fixes-3.11-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/tomba/linux:
  ARM: OMAP: dss-common: fix Panda's DVI DDC channel
  video: mxsfb: fix color settings for 18bit data bus and 32bpp
  OMAPDSS: analog-tv-connector: compile fix

11 years agoMerge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux
Linus Torvalds [Fri, 9 Aug 2013 18:51:29 +0000 (11:51 -0700)]
Merge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux

Pull drm fixes from Dave Airlie:
 "Mostly radeon, more fixes for dynamic power management which is is off
  by default for this release anyways, but there are a large number of
  testers, so I'd like to keep merging the fixes.

  Otherwise, radeon UVD fixes affecting suspend/resume regressions, i915
  regression fixes, one for your mac mini, ast, mgag200, cirrus ttm fix
  and one regression fix in the core"

* 'drm-fixes' of git://people.freedesktop.org/~airlied/linux: (25 commits)
  drm: Don't pass negative delta to ktime_sub_ns()
  drm/radeon: make missing smc ucode non-fatal
  drm/radeon/dpm: require rlc for dpm
  drm/radeon/cik: use a mutex to properly lock srbm instanced registers
  drm/radeon: remove unnecessary unpin
  drm/radeon: add more UVD CS checking
  drm/radeon: stop sending invalid UVD destroy msg
  drm/radeon: only save UVD bo when we have open handles
  drm/radeon: always program the MC on startup
  drm/radeon: fix audio dto calculation on DCE3+ (v3)
  drm/radeon/dpm: disable sclk ss on rv6xx
  drm/radeon: fix halting UVD
  drm/radeon/dpm: adjust power state properly for UVD on SI
  drm/radeon/dpm: fix spread spectrum setup (v2)
  drm/radeon/dpm: adjust thermal protection requirements
  drm/radeon: select audio dto based on encoder id for DCE3
  drm/radeon: properly handle pm on gpu reset
  drm/i915: do not disable backlight on vgaswitcheroo switch off
  drm/i915: Don't call encoder's get_config unless encoder is active
  drm/i915: avoid brightness overflow when doing scale
  ...

11 years agodlm: kill the unnecessary and wrong device_close()->recalc_sigpending()
Oleg Nesterov [Fri, 9 Aug 2013 15:19:13 +0000 (17:19 +0200)]
dlm: kill the unnecessary and wrong device_close()->recalc_sigpending()

device_close()->recalc_sigpending() is not needed, sigprocmask() takes
care of TIF_SIGPENDING correctly.

And without ->siglock it is racy and wrong, it can wrongly clear
TIF_SIGPENDING and miss a signal.

But even with this patch device_close() is still buggy:

  1. sigprocmask() should not be used, we have set_task_blocked(),
     but this is minor.

  2. We should never block SIGKILL or SIGSTOP, and this is what
     the code tries to do.

  3. This can't protect against SIGKILL or SIGSTOP anyway. Another
     thread can do signal_wake_up(), say, do_signal_stop() or
     complete_signal() or debugger.

  4. sigprocmask(SIG_BLOCK, allsigs) doesn't necessarily clears
     TIF_SIGPENDING, say, freezing() or ->jobctl.

  5. device_write() looks equally wrong by the same reason.

Looks like, this tries to protect some wait_event_interruptible() logic
from signals, it should be turned into uninterruptible wait.  Or we need
to implement something like signals_stop/start for such a use-case.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agoRevert "HID: hid-logitech-dj: querying_devices was never set"
Jiri Kosina [Fri, 9 Aug 2013 09:34:19 +0000 (11:34 +0200)]
Revert "HID: hid-logitech-dj: querying_devices was never set"

This reverts commit 407a2c2a4d85100c8c67953e4bac2f4a6c942335.

Explanation provided by Benjamin Tissoires:

Commit "HID: hid-logitech-dj, querying_devices was never set" activate
a flag which guarantees that we do not ask the receiver for too many
enumeration. When the flag is set, each following enumeration call is
discarded (the usb request is not forwarded to the receiver). The flag
is then released when the driver receive a pairing information event,
which normally follows the enumeration request.
However, the USB3 bug makes the driver think the enumeration request
has been forwarded to the receiver. However, it is actually not the
case because the USB stack returns -EPIPE. So, when a new unknown
device appears, the workaround consisting in asking for a new
enumeration is not working anymore: this new enumeration is discarded
because of the flag, which is never reset.

A solution could be to trigger a timeout before releasing it, but for
now, let's just revert the patch.

Reported-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
Tested-by: Sune Mølgaard <sune@molgaard.org>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
11 years agopowerpc/tm: Fix context switching TAR, PPR and DSCR SPRs
Michael Neuling [Fri, 9 Aug 2013 07:29:31 +0000 (17:29 +1000)]
powerpc/tm: Fix context switching TAR, PPR and DSCR SPRs

If a transaction is rolled back, the Target Address Register (TAR), Processor
Priority Register (PPR) and Data Stream Control Register (DSCR) should be
restored to the checkpointed values before the transaction began.  Any changes
to these SPRs inside the transaction should not be visible in the abort
handler.

Currently Linux doesn't save or restore the checkpointed TAR, PPR or DSCR.  If
we preempt a processes inside a transaction which has modified any of these, on
process restore, that same transaction may be aborted we but we won't see the
checkpointed versions of these SPRs.

This adds checkpointed versions of these SPRs to the thread_struct and adds the
save/restore of these three SPRs to the treclaim/trechkpt code.

Without this if any of these SPRs are modified during a transaction, users may
incorrectly see a speculated SPR value even if the transaction is aborted.

Signed-off-by: Michael Neuling <mikey@neuling.org>
Cc: <stable@vger.kernel.org> [v3.10]
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
11 years agopowerpc: Save the TAR register earlier
Michael Neuling [Fri, 9 Aug 2013 07:29:30 +0000 (17:29 +1000)]
powerpc: Save the TAR register earlier

This moves us to save the Target Address Register (TAR) a earlier in
__switch_to.  It introduces a new function save_tar() to do this.

We need to save the TAR earlier as we will overwrite it in the transactional
memory reclaim/recheckpoint path.  We are going to do this in a subsequent
patch which will fix saving the TAR register when it's modified inside a
transaction.

Signed-off-by: Michael Neuling <mikey@neuling.org>
Cc: <stable@vger.kernel.org> [v3.10]
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
11 years agopowerpc: Fix context switch DSCR on POWER8
Michael Neuling [Fri, 9 Aug 2013 07:29:29 +0000 (17:29 +1000)]
powerpc: Fix context switch DSCR on POWER8

POWER8 allows the DSCR to be accessed directly from userspace via a new SPR
number 0x3 (Rather than 0x11.  DSCR SPR number 0x11 is still used on POWER8 but
like POWER7, is only accessible in HV and OS modes).  Currently, we allow this
by setting H/FSCR DSCR bit on boot.

Unfortunately this doesn't work, as the kernel needs to see the DSCR change so
that it knows to no longer restore the system wide version of DSCR on context
switch (ie. to set thread.dscr_inherit).

This clears the H/FSCR DSCR bit initially.  If a process then accesses the DSCR
(via SPR 0x3), it'll trap into the kernel where we set thread.dscr_inherit in
facility_unavailable_exception().

We also change _switch() so that we set or clear the H/FSCR DSCR bit based on
the thread.dscr_inherit.

Signed-off-by: Michael Neuling <mikey@neuling.org>
Cc: <stable@vger.kernel.org> [v3.10]
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
11 years agopowerpc: Rework setting up H/FSCR bit definitions
Michael Neuling [Fri, 9 Aug 2013 07:29:28 +0000 (17:29 +1000)]
powerpc: Rework setting up H/FSCR bit definitions

This reworks the Facility Status and Control Regsiter (FSCR) config bit
definitions so that we can access the bit numbers.  This is needed for a
subsequent patch to fix the userspace DSCR handling.

HFSCR and FSCR bit definitions are the same, so reuse them.

Signed-off-by: Michael Neuling <mikey@neuling.org>
Cc: <stable@vger.kernel.org> [v3.10]
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
11 years agopowerpc: Fix hypervisor facility unavaliable vector number
Michael Neuling [Fri, 9 Aug 2013 07:29:27 +0000 (17:29 +1000)]
powerpc: Fix hypervisor facility unavaliable vector number

Currently if we take hypervisor facility unavaliable (from 0xf80/0x4f80) we
mark it as an OS facility unavaliable (0xf60) as the two share the same code
path.

The becomes a problem in facility_unavailable_exception() as we aren't able to
see the hypervisor facility unavailable exceptions.

Below fixes this by duplication the required macros.

Signed-off-by: Michael Neuling <mikey@neuling.org>
Cc: <stable@vger.kernel.org> [v3.10]
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
11 years agopowerpc/kvm/book3s_pr: Return appropriate error when allocation fails
Thadeu Lima de Souza Cascardo [Wed, 17 Jul 2013 15:10:29 +0000 (12:10 -0300)]
powerpc/kvm/book3s_pr: Return appropriate error when allocation fails

err was overwritten by a previous function call, and checked to be 0. If
the following page allocation fails, 0 is going to be returned instead
of -ENOMEM.

Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
11 years agopowerpc/kvm: Add signed type cast for comparation
Chen Gang [Mon, 22 Jul 2013 06:32:35 +0000 (14:32 +0800)]
powerpc/kvm: Add signed type cast for comparation

'rmls' is 'unsigned long', lpcr_rmls() will return negative number when
failure occurs, so it need a type cast for comparing.

'lpid' is 'unsigned long', kvmppc_alloc_lpid() return negative number
when failure occurs, so it need a type cast for comparing.

Signed-off-by: Chen Gang <gang.chen@asianux.com>
Acked-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
11 years agopowerpc/eeh: Add missing procfs entry for PowerNV
Mike Qiu [Wed, 7 Aug 2013 02:25:14 +0000 (22:25 -0400)]
powerpc/eeh: Add missing procfs entry for PowerNV

The procfs entry for global statistics has been missed on PowerNV
platform and the patch is going to add that.

Signed-off-by: Mike Qiu <qiudayu@linux.vnet.ibm.com>
Acked-by: Gavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
11 years agopowerpc/pseries: Add backward compatibilty to read old kernel oops-log
Aruna Balakrishnaiah [Thu, 8 Aug 2013 17:04:00 +0000 (22:34 +0530)]
powerpc/pseries: Add backward compatibilty to read old kernel oops-log

Older kernels has just length information in their header. Handle it
while reading old kernel oops log from pstore.

Applies on top of powerpc/pseries: Fix buffer overflow when reading from pstore

Signed-off-by: Aruna Balakrishnaiah <aruna@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>