Philipp Zabel [Fri, 21 Jun 2013 12:52:17 +0000 (14:52 +0200)]
staging: drm/imx: use generic irqchip
This depends on the patch "genirq: Generic chip: Add linear irq domain support"
and removes the custom IPU irq_chip and irq_domain_ops. Instead, the generic
irq chip implementation is reused.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de> Acked-by: Grant Likely <grant.likely@secretlab.ca> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Philipp Zabel [Fri, 21 Jun 2013 08:57:20 +0000 (10:57 +0200)]
staging: drm/imx: ipuv3-crtc: immediately update crtc->fb in ipu_page_flip
Since commit 8cf1e9811471f2910fa38dc1b28e1789080ba961
("drm: Add consistency check for page-flipping") drm_mode_page_flip_ioctl
contains a WARN_ON that triggers if the .page_flip callback didn't update
the crtc->fb pointer to the new framebuffer immediately.
The IPU can request up to four pixels per access, which gives four
times the bandwidth compared to what the driver currently assumes.
After correcting this, we have to increase safety margins for
bandwidth requirement calculations.
Philipp Zabel [Fri, 21 Jun 2013 08:27:39 +0000 (10:27 +0200)]
staging: drm/imx: convert IPU irq driver to irq_domain_add_linear
The IPU has a lot of interrupts. Instead of allocating descs for all
of them, register a linear irq domain and create mappings as needed.
This was listed in the TODO as a prerequisite to move the IPU driver
out of staging.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Move zram sysfs code to zram drv and remove zram_sysfs.c
file. This gives ability to make static a number of previously
exported zram functions, used from zram sysfs, e.g. internal zram
zram_meta_alloc/free(). We also can drop zram_drv wrapper
functions, used from zram sysfs:
e.g. zram_reset_device()/__zram_reset_device() pair.
v2: as suggested by Greg K-H, move MODULE description to the
bottom of the file.
Marc Dietrich [Fri, 21 Jun 2013 08:28:55 +0000 (10:28 +0200)]
staging: nvec: move device tree parsing to its own function
This moves device tree parsing to its own function so more stuff can
be cleanly added in the future.
Signed-off-by: Leon Romanovsky <leon@leon.nu> Signed-off-by: Marc Dietrich <marvin24@gmx.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Marc Dietrich [Fri, 21 Jun 2013 08:28:54 +0000 (10:28 +0200)]
staging: nvec: remove instantiating via platform device
Tegra has been converted to support device tree only. Remove support
for instantiating via platform device.
Signed-off-by: Leon Romanovsky <leon@leon.nu> Signed-off-by: Marc Dietrich <marvin24@gmx.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Dan Carpenter [Mon, 24 Jun 2013 16:12:09 +0000 (19:12 +0300)]
staging: vt6655: info leak in ioctl
The SCmdLinkStatus struct has a couple holes. There is one between
->bLink and ->wBSSType, and another between ->abySSID and ->uChannel.
I've added a memset() to initialize the struct to zero. Since we don't
need to clear abySSID any more so I've removed that memset. It was
wrong anyway: abySSID has "SSID_MAXLEN + 2" (34) bytes, not
"WLAN_SSID_MAXLEN + 1" (33).
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Paul Zimmerman [Mon, 24 Jun 2013 18:34:23 +0000 (11:34 -0700)]
staging: dwc2: fix thinko in dwc2_hc_set_even_odd_frame()
I screwed up the sense of this if() statement while porting our
vendor driver to create the dwc2 driver. This caused frame overrun
errors on periodic transfers when there were other transfers
active in the same (micro)frame.
With this fix, the dwc2 driver now works on the Raspberry Pi
platform even with the USB Ethernet controller enabled, where
before that would cause all USB devices to stop working.
Thanks to Ray Jui and Jerry Lin at Broadcom for tracking this down.
Reported-by: Ray Jui <rjui@broadcom.com> Signed-off-by: Paul Zimmerman <paulz@synopsys.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Arnd Bergmann [Fri, 21 Jun 2013 19:55:35 +0000 (21:55 +0200)]
staging: line6: avoid __sync_fetch_and_{and,or}
__sync_fetch_and_and and __sync_fetch_and_or are functions that are provided
by gcc and depending on the target architecture may be implemented in libgcc,
which is not always available in the kernel. This leads to a build failure
on ARMv5:
drivers/built-in.o: In function `line6_pcm_release':
:(.text+0x3bfe80): undefined reference to `__sync_fetch_and_and_4'
drivers/built-in.o: In function `line6_pcm_acquire':
:(.text+0x3bff30): undefined reference to `__sync_fetch_and_or_4'
To work around this, we can use the kernel-provided cmpxchg macro.
Build-tested only.
Signed-off-by: Arnd Bergmann <arnd@arndb.de> Cc: Markus Grabner <grabner@icg.tugraz.at> Acked-by: Randy Dunlap <rdunlap@infradead.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging: comedi: pcmuio: cleanup DIO subdevice (*insn_{bits,config})
Use the pcmuio_{read,write}() helpers to read/write all 24 channels
instead of handling the digital I/O as three separate ports. This
simplifies both functions with minimal overhead.
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>
staging: comedi: pcmuio: remove 'iobases' from subdevice private data
The 'asic' and 'port' associated with a given subdevice can easily be
found based on the subdevice 'index'. With that information we can
then calculate the correct iobase and register offset needed to read/
write the 8-bit ports.
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>
staging: comedi: pcmuio: simplify 'n_subdevs' and 's->n_chan'
The number of subdevices needed by the driver is based on the
'num_asics' reported by the boardinfo and each subdevice always
has 24 channels. Simplify the attach a bit.
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>
This member of the subdevice private data is always initialize to
'0' due to the 'thisasic_chanct' always being zero when it is set
during the attach. Simplify the driver a bit by removing it.
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>
This member of the subdevice private data is always initialize to
s->n_chan if the subdevice supports interrupts and is only used
in functions that can be called by the interrupt subdevice.
Simplyfy the driver a bit by removing it and just using s->n_chan.
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>
This member of the subdevice private data is always initialize to
'0' due to the 'byte_no' always being zero when it is set during
the attach. Simplify the driver a bit by removing it.
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>
All the I/O ports are left unlocked in the driver so the 'pagelock'
in the private data is not necessary. The paranoia sanity checks
are also unnecessary, Remove them.
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>
staging: comedi: have comedi_set_spriv() allocate the memory
As suggested by Ian Abbott, comedi_set_spriv() can only be used to
set the subdevice->private pointer to something that can be kfree()'d.
Rename the function to comedi_alloc_spriv() and have it kzalloc() the
memory as well as set the private pointer. This saves a function call
in the drivers and avoids the possibility of incorrectly calling
comedi_set_spriv() for some pointer that is not meant to be kfree()'d.
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>
None of the addi-data drivers that still use the "common" code have
ttl digital i/o. Remove the unnecessary boardinfo as well as the
subdevice initialization 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>
staging: comedi: addi-data: remove unused vars from boardinfo
Due to the on-going cleanup of the addi-data drivers, the boardinfo
used in the "common" code has a number of variables that are not
used. Remove the cruft.
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>
staging: comedi: addi-data: simplify the PCI bar reading
All of the remaining addi-data drivers that use the "common" code
either do not have an eeprom or the PCI controller chip is not a
PLX PCI 9054. Knowing this we can simplify the common code that
reads the PCI bars to get the iobase addresses.
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>
staging: comedi: addi-data: remove unused vars from private data
Due to the on-going cleanup of the addi-data drivers, the private
data used in the "common" code has a number of variables that
either are not used at all or or set but never used.
Remove all of them from the private data and the unnecessary use
in the drivers.
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>
Xenia Ragiadakou [Wed, 19 Jun 2013 01:58:07 +0000 (04:58 +0300)]
staging: rtl8192u: fix comments in r819xU_phy.c
This patch fixes the comments in r819xU_phy.c by:
- replacing "// ..." with "/* .... */"
- removing unnecessary comments, the dates and names of
developers from comments
- fixing some inconsistent comments
- fixing some typos
- fixing alignment issues
Xenia Ragiadakou [Wed, 19 Jun 2013 01:58:05 +0000 (04:58 +0300)]
staging: rtl8192u: remove variable 'Ret' in r819xU_phy.c
This patch removes variable 'Ret' from rtl8192_QueryBBReg()
since its value is returned immediately after it is
assigned. The name 'Ret', anyway, does not give any
insight and the function description comment is sufficiently
informative regarding the returned value.
Chad Williamson [Tue, 18 Jun 2013 03:32:05 +0000 (22:32 -0500)]
Staging: silicom: move more assignments out of if conditions
Remove more assignments from if-statement conditions in bpctl_mod.c,
resolving checkpatch.pl errors. Those that remain need more attention
than I'm presently prepared to give them.
Signed-off-by: Chad Williamson <chad@dahc.us> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Chad Williamson [Tue, 18 Jun 2013 03:32:04 +0000 (22:32 -0500)]
Staging: silicom: move assignments out of if conditions
Remove a bunch of assignments from if-statement conditions in
bpctl_mod.c, resolving checkpatch.pl errors. (This isn't all of
them, but the patch is getting rather long...)
Signed-off-by: Chad Williamson <chad@dahc.us> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Xenia Ragiadakou [Tue, 18 Jun 2013 02:29:41 +0000 (05:29 +0300)]
staging: rtl8192u: rename general variables in r819xU_phy.c
This patch renames the following variables:
- dwRegAddr and RegAddr into reg_addr
- dwData and Data into data
- BitShift into bitshift
- Offset into offset
- NewOffset into new_offset
- Bandwidth into bandwidth
The renaming was done to remove camel case and
unnecessary 'dw' prefix from the above variable names.
Xenia Ragiadakou [Tue, 18 Jun 2013 02:29:40 +0000 (05:29 +0300)]
staging: rtl8192u: rename *BitMask variables in r819xU_phy.c
This patch renames BitMask and dwBitMask variables into
bitmask to remove camel case.
dwBitMask in the given context can be renamed into simply
bitmask since its number of bits (dw->32) can be deduced
by its type which is u32.
Xenia Ragiadakou [Tue, 18 Jun 2013 02:29:39 +0000 (05:29 +0300)]
staging: rtl8192u: replace *Value variables in r819xU_phy.c
This patch renames the variables OriginalValue, NewValue,
Original_Value, New_Value and retValue to reg, and
the variables bRegValue and dwRegValue to reg_u8 and
reg_u32, respectively. This is done primarily in
order to reduce the variable name length and consequently
line length.
Also, it replaces the pair of Original and New variables
with one variable.
Markus Niebel [Tue, 18 Jun 2013 01:23:58 +0000 (09:23 +0800)]
staging: imx_drm: imx_ldb: fix 18 bit format
Since 18 bit is supported as datawidth in device tree it should be
supported in driver. Beside the LDB channel the IPU channel has also
to be configured to use BGR666.
Xenia Ragiadakou [Sat, 15 Jun 2013 04:29:02 +0000 (07:29 +0300)]
staging: rtl8192u: fix braces in r819xU_phy.c
This patch fixes the brace position and the whitespace
around braces.
That corrects the following checkpatch errors:
ERROR: that open brace { should be on the previous line
ERROR: else should follow close brace '}'
ERROR: space required after that close brace '}'
Also, it removes completely an if(0) statement,
instead of correcting its braces.
Xenia Ragiadakou [Sat, 15 Jun 2013 04:29:01 +0000 (07:29 +0300)]
staging: rtl8192u: fix bitshift calculation in r819xU_phy.c
This patch uses the ffs linux built-in routine in the
computation of the bitshift to take advantage of
architecture specific instructions that implement this
computation more efficiently.
Tobias Polzer [Thu, 13 Jun 2013 21:00:35 +0000 (23:00 +0200)]
staging: csr: Fix obvious logic error in staging/csr
In uf_process_ma_packet_req() (staging/csr/unifi_pdu_processing.c), the
test for
(staInfo->timSet != CSR_WIFI_TIM_RESET) || (staInfo->timSet !=
CSR_WIFI_TIM_RESETTING)
obviously makes no sense and causes GCC to emit a warning if
-Wlogical-op is enabled.
As suggested in the referenced bug report, || was replaced by &&.
References: https://bugzilla.kernel.org/show_bug.cgi?id=46571 Reported-By: David Binderman <dcb314@hotmail.com> Signed-off-by: Tobias Polzer <tobias.polzer@fau.de> Signed-off-by: Dominik Paulus <dominik.paulus@fau.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Laura Lawniczak [Sun, 9 Jun 2013 21:11:30 +0000 (23:11 +0200)]
staging: keucr: removed unnecessary variables and comments
blen and bn were only used in commented code fragments.
since comments should be informative and not for storing old
code, both, commented code and variables, were removed.
Signed-off-by: Laura Lawniczak <laura.lawniczak@googlemail.com> Signed-off-by: Johannes Schilling <of82ecuq@cip.cs.fau.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging: comedi: addi_apci_3xxx: only allocate needed subdevices
The number of subdevices needed by a given board supported by this
driver can cary from 2 (for the apci3500) to a maximum of 5. Currently
this driver always allocates 7 subdevices and sets the subdevice type
to COMEDI_SUBD_UNUSED for the ones that are not needed.
Calculate the actual number of needed subdevices from the boardinfo
and only allocate and setup the ones that are used.
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>
staging: comedi: addi_apci_3xxx: fix the analog input subdevice
The analog input subdevice support functions in hwdrv_apci3xxx.c
do not follow the comedi API.
The (*insn_config) function overrides the INSN_CONFIG_DIO_INPUT
instruction as an internal APCI3XXX_CONFIGURATION instruction. The
APCI3XXX_CONFIGURATION instruction requires 4 data parameters but
the comedi core sanity checks the INSN_CONFIG_DIO_INPUT instruction
which only has 1 data parameter. Because of this the (*insn_config)
function can never be called.
The (*insn_read) function is supposed to do "one-shot" or "software-
triggered" reads and return the data. The function in hwdrv_apci3xxx.c
does do this but it also is used to optionally start a "hardware-
triggered" conversion. Hardware-triggered conversions should be done
with the comedi_async command functions.
Delete the hwrdv_apci3xxx.c file and fix the analog input subdevice
in the driver by:
1) add a new (*insn_read) function for "one-shot" reads
2) implement the (*do_cmdtest) and (*do_cmd) functions for
"hardware-triggered" asyncronous reads
3) fix the interrupt handler to return the read data via the
comedi_async buffer
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>
staging: comedi: addi_apci_3xxx: stub in analog input command support
The analog input subdevice in this driver is broken. The attach of
this driver requests the irq and sets the dev->read_subdev to the
analog input subdevice but it does not setup the subdevice to support
commands.
Stub in the command support functions so we can fix the subdevice.
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>
staging: comedi: addi_apci_3xxx: remove 'b_SingleDiff' private data
The analog reference desired for the acquisition is stored in the
insn->chanspec that is passed to the (*insn_read) function. There is
no need to use a private data variable to hold the desired analog
reference that is set in the (*insn_config).
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>
The number of differential analog input channels is always half the
number of single-ended channels, 'ai_n_chan', or all the channels
if the board does not support single-ended channels. We can use the
'ai_subdev_flags' boardinfo to determine if the board does not support
single-ended channels.
To simplify the boardinfo, remove the 'i_NbrAiChannelDiff' and just
use 'ai_n_chan' to specify the number of channels.
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>
projects / karo-tx-linux.git / log