linux.git/drivers/staging, branch v3.8-rc4

Merge tag 'staging-3.8-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging

2013-01-14T17:08:38+00:00

Pull staging fixes from Greg Kroah-Hartman:
 "Here are a number of small fixes to staging drivers for your 3.8-rc3
  tree.

  Well, the omapdrm fixes aren't really "small" but they were waiting on
  a number of other drm patches to go in through the drm tree, and got
  delayed by my vacation over the holidays.  They are totally
  self-contained, everyone involved have acked them, and they fix issues
  that people have been having with the driver.

  Other than that one, it's a bunch of tiny bugfixes for a number of
  reported issues.

  Signed-off-by: Greg Kroah-Hartman "

* tag 'staging-3.8-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging: (36 commits)
  staging: zram: fix invalid memory references during disk write
  staging: tidspbridge: use prepare/unprepare on dsp clocks
  staging: tidspbridge: Fix build breakage due to splitting CM functions.
  staging: comedi: comedi_test: fix race when cancelling command
  staging: comedi: Kconfig: COMEDI_NI_AT_A2150 should select COMEDI_FC
  staging: comedi: prevent auto-unconfig of manually configured devices
  staging: comedi: fix minimum AO period for NI 625x and NI 628x
  staging: vme_pio2: fix oops on module unloading
  staging: speakup: avoid out-of-range access in synth_add()
  staging: speakup: avoid out-of-range access in synth_init()
  staging: rtl8192e: Fix failure to check pci_map_single()
  staging: rtl8187se: Fix failure to check pci_map_single()
  staging: drm/imx: fix double free bug in error path
  staging: drm/imx: several bug fixes
  staging: drm/imx: check return value of ipu_reset()
  staging: drm/omap: fix flags in dma buf exporting
  staging: drm/omap: use omapdss low level API
  staging/fwserial: Update TODO file per reviewer comments
  staging/fwserial: Limit tx/rx to 1394-2008 spec maximum
  staging/fwserial: Refine Kconfig help text
  ...

staging: zram: fix invalid memory references during disk write

2013-01-14T03:40:02+00:00

Fixes a bug introduced by commit c8f2f0db1 ("zram: Fix handling
of incompressible pages") which caused invalid memory references
during disk write. Invalid references could occur in two cases:
 - Incoming data expands on compression: In this case, reference was
made to kunmap()'ed bio page.
 - Partial (non PAGE_SIZE) write with incompressible data: In this
case, reference was made to a kfree()'ed buffer.

Fixes bug 50081:
https://bugzilla.kernel.org/show_bug.cgi?id=50081

Signed-off-by: Nitin Gupta 
Cc: stable 
Reported-by: Mihail Kasadjikov 
Reported-by: Tomas M 
Reviewed-by: Minchan Kim 
Signed-off-by: Greg Kroah-Hartman

drm/prime: drop reference on imported dma-buf come from gem

2013-01-08T04:41:53+00:00

Increasing ref counts of both dma-buf and gem for imported dma-buf come from gem
makes memory leak. release function of dma-buf cannot be called because f_count
of dma-buf increased by importing gem and gem ref count cannot be decrease
because of exported dma-buf.

So I add dma_buf_put() for imported gem come from its own gem into each drivers
having prime_import and prime_export capabilities. With this, only gem ref
count is increased if importing gem exported from gem of same driver.

Signed-off-by: Seung-Woo Kim 
Signed-off-by: Kyungmin.park 
Cc: Inki Dae 
Cc: Daniel Vetter 
Cc: Rob Clark 
Cc: Alex Deucher 
Cc: stable 
Signed-off-by: Dave Airlie

staging: tidspbridge: use prepare/unprepare on dsp clocks

2013-01-07T23:03:56+00:00

This solves runtime failures while trying to enable WDT3 related
functionality on firmware load, however it does affect other clocks
controlled by the driver. Seen on 3.8-rc1.

CCF provides clk_prepare and clk_unprepare for enable and disable
operations respectively, this needs to be called in the correct
order while handling clocks.

Code path to enable/disable dsp clocks can still be reached from an
atomic context, hence we can't use clk_prepare_enable and
clk_disable_unprepare yet.

Signed-off-by: Omar Ramirez Luna 
Signed-off-by: Greg Kroah-Hartman

staging: tidspbridge: Fix build breakage due to splitting CM functions.

2013-01-07T23:02:37+00:00

Commit ff4ae5d (ARM: OMAP2+: CM/hwmod: split CM functions into OMAP2, OMAP3-specific files)
resulted in a build breakage for tidspbridge driver.

   ...
   CC [M]  drivers/staging/tidspbridge/core/tiomap3430.o
   staging/tidspbridge/core/tiomap3430.c: In function ‘bridge_brd_start’:
   staging/tidspbridge/core/tiomap3430.c:550:24: error: ‘OMAP3430_CM_AUTOIDLE_PLL’
     undeclared (first use in this function)
   make[3]: *** [drivers/staging/tidspbridge/core/tiomap3430.o] Error 1
   ...

Fix this by including the appropriate header file.

Signed-off-by: Enric Balletbo i Serra 
Signed-off-by: Greg Kroah-Hartman

staging: comedi: comedi_test: fix race when cancelling command

2013-01-07T22:48:25+00:00

Éric Piel reported a kernel oops in the "comedi_test" module.  It was a
NULL pointer dereference within `waveform_ai_interrupt()` (actually a
timer function) that sometimes occurred when a running asynchronous
command is cancelled (either by the `COMEDI_CANCEL` ioctl or by closing
the device file).

This seems to be a race between the caller of `waveform_ai_cancel()`
which on return from that function goes and tears down the running
command, and the timer function which uses the command.  In particular,
`async->cmd.chanlist` gets freed (and the pointer set to NULL) by
`do_become_nonbusy()` in "comedi_fops.c" but a previously scheduled
`waveform_ai_interrupt()` timer function will dereference that pointer
regardless, leading to the oops.

Fix it by replacing the `del_timer()` call in `waveform_ai_cancel()`
with `del_timer_sync()`.

Signed-off-by: Ian Abbott 
Reported-by: Éric Piel 
Cc: stable 
Signed-off-by: Greg Kroah-Hartman

staging: comedi: Kconfig: COMEDI_NI_AT_A2150 should select COMEDI_FC

2013-01-07T22:48:24+00:00

The 'ni_at_a2150' module links to `cfc_write_to_buffer` in the
'comedi_fc' module, so selecting 'COMEDI_NI_AT_A2150' in the kernel config
needs to also select 'COMEDI_FC'.

Signed-off-by: Ian Abbott 
Cc: stable  # 3.6+
Signed-off-by: Greg Kroah-Hartman

staging: comedi: prevent auto-unconfig of manually configured devices

2013-01-07T22:22:37+00:00

When a low-level comedi driver auto-configures a device, a `struct
comedi_dev_file_info` is allocated (as well as a `struct
comedi_device`) by `comedi_alloc_board_minor()`.  A pointer to the
hardware `struct device` is stored as a cookie in the `struct
comedi_dev_file_info`.  When the low-level comedi driver
auto-unconfigures the device, `comedi_auto_unconfig()` uses the cookie
to find the `struct comedi_dev_file_info` so it can detach the comedi
device from the driver, clean it up and free it.

A problem arises if the user manually unconfigures and reconfigures the
comedi device using the `COMEDI_DEVCONFIG` ioctl so that is no longer
associated with the original hardware device.  The problem is that the
cookie is not cleared, so that a call to `comedi_auto_unconfig()` from
the low-level driver will still find it, detach it, clean it up and free
it.

Stop this problem occurring by always clearing the `hardware_device`
cookie in the `struct comedi_dev_file_info` whenever the
`COMEDI_DEVCONFIG` ioctl call is successful.

Signed-off-by: Ian Abbott 
Cc: stable 
Signed-off-by: Greg Kroah-Hartman

staging: comedi: fix minimum AO period for NI 625x and NI 628x

2013-01-07T22:16:49+00:00

The minimum period was set to 357 ns, while the divider for these boards is 50
ns. This prevented to output at maximum speed as ni_ao_cmdtest() would return
357 but would not accept it.

Not sure why it was set to 357 ns (this was done before the git history,
which starts 5 years ago). My guess is that it comes from reading the
specification stating a 2.8 MHz rate (~ 357 ns). The latest
specification states a 2.86 MHz rate (~ 350 ns), which makes a lot
more sense.

Tested on a pci-6251.

Signed-off-by: Éric Piel 
Acked-By: Ian Abbott 
Cc: stable 
Signed-off-by: Greg Kroah-Hartman

staging: vme_pio2: fix oops on module unloading

2013-01-07T21:56:24+00:00

This patch forbids loading vme_pio2 module without specifing "num_bus" parameter.
Otherwise on module unloading pio2_exit() calls vme_unregister_driver() for not
registered pio2_driver.

Signed-off-by: Konstantin Khlebnikov 
Cc: Manohar Vanga 
Acked-by: Martyn Welch 
Signed-off-by: Greg Kroah-Hartman