Recent changes to use the builtin min functions [1] introduced
type checking which wasn't present before. This resulted in
"comparision of distinct pointer types lacks a cast" build
warnings on non X86 architectures [2,3].
The call to min() which resulted in this warning took the result
of kiblnd_rd_frag_size(), which returned a __u32, and the
variable 'resid', which is an int. 'resid' is inside a while
loop which is only entered if it is positive. Casting it as a
__u32 can be perormed without a loss of data or change in
functionality.
Tristan Lelong [Sun, 18 Jan 2015 07:44:49 +0000 (23:44 -0800)]
staging: lustre: remove kmalloc from fld_proc_hash_seq_write
This patch simplifies the fld_proc_hash_seq_write function
by removing the dynamic memory allocation.
The longest fh_name used so far in lustre is 4 characters.
We use a 8 bytes variable to be on the safe side.
staging: comedi: vmk80xx: tidy up vmk80xx_cnt_insn_config()
Tidy up this (*insn_config) function to follow the normal format in
comedi drivers.
INSN_CONFIG_RESET instructions do not have any extra parameters (insn->n is
always 1) so the for loop used to write the packet doesn't make any sense.
Remove it and just write the single packet.
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: vmk80xx: GPCT_RESET is not an (*insn_config) instruction
The data[0] parameter to (*insn_config) functions is the "configuration instruction"
that should be handled. These are defined by the enum configuration_ids in comedi.h.
This driver is currently checking the data[0] value to be INSN_CONFIG_RESET or
GPCT_RESET in order to reset a counter channel. GPCT_RESET is defined as 0x0001
which would match the configuration instruction INSN_CONFIG_DIO_OUTPUT. That doesn't
make any sense for a counter.
Remove GPCT_RESET from the insn_cmd test.
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: ni_labpc_common: move comedi_alloc_devpriv() to common code
The ni_labpc_common module is used by the ni_labpc, ni_labpc_pci, and ni_labpc_cs
drivers to provide the common code support. Those drivers currently all call
comedi_alloc_devpriv() to allocate the private data before calling the common
attach function.
For aesthetics, move the private data allocation into the common 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: adv_pci1710: transfer all ai samples in one step
The two step process to transfer the ai samples in pci1710_handle_fifo() doesn't
make any sense. If the async buffer does not have enough room for the samples the
core will set the async event COMEDI_CB_OVERFLOW which will cause the async command
to cancel. Splitting the transfer doesn't add any value.
Transfer all the samples in one step to simplify the 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: adv_pci1710: do comedi_handle_events() in common code patch
The pci1710_handle_every_sample() and pci1710_handle_fifo() helpers, called by
the interrupt handler, both have multiple comedi_handle_events() calls. Remove
these and just do a single comedi_handle_events() at the end of the interrupt
handler.
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: adv_pci1710: clarify the 'act_chanlist'
This driver saves the channel list of a scan in the private data and uses
that list to check analog input samples for data dropout.
Currently the channel numbers are shifted 12 bits so that they match the
channel number in the samples that are read from the board. All of the
shifts make the driver a bit harder to follow.
Store the channel numbers directly to the 'act_chanlist' and shift the
sample value instead when checking for the dropout. Add a comment for
clarity.
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>
1) determine if the board is a pci1720 during the attach of the driver
and when reseting the board
2) determine is the board is not a pci1713 when checking for analog input
data dropout
There is also an unnecessary use of the 'cardtype' when canceling an analog
input async command.
Remove the 'cardtype' member and add two bit-field flags, 'is_pci1713' and
'is_pci1720'. Refactor the driver to use the new flags.
Remove the unnecessary cardtype handling in pci171x_ai_cancel().
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 boardinfo is not necessary. All the boards with analog outputs have
the same range options except for the pci1720 board. That board type is
already handled specially during the driver attach.
Remove the boardinfo and add the subdevice range_table initialization in
the cardtype switch.
Rename the range tables to clarify that they are "analog output" ranges.
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: adv_pci1710: tidy up analog input subdevice init
For aesthetics, add some whitespace to the analog input subdevice init and
rename the (*insn_read) support function. Also remove the unnecessary comments.
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: adv_pci1710: tidy up digital input and output subdevice init
For aesthetics, add some whitespace to the digital input and output subdevice
init and rename the support functions. Also remove the unnecessary comment.
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: adv_pci1710: use comedi_subdevice 'readback'
Use the new comedi_subdevice 'readback' member and the core provided
(*insn_read) for the readback of the analog output subdevice channels.
The board is "reset" prior to the subdevice init. Part of this reset
sets all the analog output channels to 0V. Move the initialization of
the 'readback' values after the 'readback' member has been allocated.
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: adv_pci1710: change boardinfo 'n_aochan' to 'has_ao'
Some of the boards supported by this driver have analog outputs. The pci1720
board has 4 analog output channels and the others have 2.
For aesthetics, change the 'n_aochan' member to a bit-field flag 'has_ao' and
refactor the board attach accordingly. Remove the unnecessary initialization
of the subdevice 'len_chanlist'. That member is only used by subdevices that
support async commands.
Also, remove the unnecessary 'number of channels' check in pci171x_reset().
The boards that used this function, and have analog outputs, always have 2
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>
staging: comedi: adv_pci1710: change boardinfo 'fifo_half_size' to 'has_large_fifo'
The boards supported by this driver have a 4K or 1K FIFO that is used when reading
analog input samples when running an async command. The maximum number of samples
in the FIFO is half the FIFO size due to the 2 bytes/sample (12-bit resolution).
For aesthetics, change the 'fifo_half_size' member to a bit-field flag 'has_large_fifo'
and add a new member to the private data to hold the 'max_samples' available in the
FIFO. Refactor the board attach and pci1710_handle_fifo() accordingly.
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: adv_pci1710: change boardinfo 'n_aichand' to 'has_diff_ai'
The 'n_aichand' member of the boardinfo is actually a flag indicating that the
board supports differential analog inputs. The number of channels is always
half the single-ended number of channels.
For aesthetics, change the 'n_aichand' member to a bit-field flag 'has_diff_ai'.
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 driver can acquire analog input samples during the async command with
DMA, by using the FIFO, or sample-by-sample using the End-Of-Conversion
interrupt. All three methods do the following sequence:
1) check for channel dropout
2) add the sample to the async buffer
3) advance the channel dropout detection and detect the end of the command
Merge this sequence into a new helper function.
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>
Hardware errors should be reported with the COMEDI_CB_ERROR event. This event
will cause the async command to cancel. It's not necessary to also set the
COMEDI_CB_EOA event. Remove these events.
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: das16: use COMEDI_CB_CANCEL_MASK to see if command is running
In das16_interrupt(), use COMEDI_CB_CANCEL_MASK to determine if the async
command is still running and the dma needs to be re-enabled. This will
cause the driver not re-enable the dma if the async buffer overflows
(COMEDI_CB_OVERFLOW), a hardware error occurs (COMEDI_CB_ERROR), or the
command completes (COMEDI_CB_EOA).
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_1500: rewrite the subdevice support functions
This driver is a mess. It violates the comedi API so much that I doubt
anything actually works.
Drop the addi-data/hwdrv_apci1500.c file and rewrite the subdevice support
functions.
This board has 16 digital inputs (subdevice 0) and 16 digital outputs
(subdevice 1).
It also has three 16-bit timer/counters provided by a Z8536 CIO chip
(subdevice 2). The Z8536 chip is also used to support pattern match
interrupt detection of the first 14 digital input 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>
staging: comedi: aio_iiro_16: return input state in async command sample
Modify the sample data returned by the async command to include the current
state of the digital inputs. Otherwise the command needs to be canceled in
order for the user to do an (*insn_bits) operation to check the digital
inputs.
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>
Ian Abbott [Mon, 19 Jan 2015 14:47:27 +0000 (14:47 +0000)]
staging: comedi: cb_pcidas64: fix incorrect AI range code handling
The hardware range code values and list of valid ranges for the AI
subdevice is incorrect for several supported boards. The hardware range
code values for all boards except PCI-DAS4020/12 is determined by
calling `ai_range_bits_6xxx()` based on the maximum voltage of the range
and whether it is bipolar or unipolar, however it only returns the
correct hardware range code for the PCI-DAS60xx boards. For
PCI-DAS6402/16 (and /12) it returns the wrong code for the unipolar
ranges. For PCI-DAS64/Mx/16 it returns the wrong code for all the
ranges and the comedi range table is incorrect.
Change `ai_range_bits_6xxx()` to use a look-up table pointed to by new
member `ai_range_codes` of `struct pcidas64_board` to map the comedi
range table indices to the hardware range codes. Use a new comedi range
table for the PCI-DAS64/Mx/16 boards (and the commented out variants).
Chase Southwood [Tue, 20 Jan 2015 02:42:49 +0000 (20:42 -0600)]
staging: comedi: das1800: prefer kmalloc_array over kmalloc with multiply
Checkpatch doesn't like kmalloc with multiply very much:
drivers/staging/comedi/drivers/das1800.c:1377: WARNING: Prefer kmalloc_array over kmalloc with multiply
So this patch swaps that use out for kmalloc_array instead.
comedi: rtd520: if condition with no effect - if identical to else
The if and the else branch code are identical - so the condition has no
effect on the effective code - this patch removes the condition and the
duplicated code.
Signed-off-by: Nicholas Mc Guire <der.herr@hofr.at> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Marc Dietrich [Tue, 20 Jan 2015 22:01:28 +0000 (23:01 +0100)]
staging: nvec: specify a platform-device base id
commit 6e3f62f079 (mfd: core: Fix platform-device id generation) modified
the computation of the mfd cell id. Negative numbers forbit the specification
of cell ids as we do. Fix this for now by specifying a base of 0 instead. In
the long run, this may be changed to automatic cell ids (base -2).
Reported-by: Misha Komarovskiy <zombah@gmail.com> Signed-off-by: Marc Dietrich <marvin24@gmx.de> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Merge tag 'iio-for-3.20a_take2' of git://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio into staging-testing
Jonathan writes:
First round of IIO new drivers, cleanups and functionality for the 3.20 cycle take 2
Updated pull request with Daniel's fix on top for the power management
Kconfig changes that had snuck in since last update of the IIO tree
worked it's way through from mainline.
Original pull message
New device support
* jsa1212 proxmity / ambient light sensor
* SM08500 supported added to the kxcjk-1013 accelerometer driver
* KMX61 Accelerometer/Magnetometer. This took a somewhat rocky path
being first merged, then reverted for a rewrite after a discussion of
how to support additional functionality and finally being merged prior
to some last reviews coming in, with resultant follow up patches.
* Freescale mma9551l driver (minor follow up warning supression patch).
* Semtech SX9500 proximity device driver.
* ak8975 gains support for ak09911 and ak09912 and drop the standalone driver
for the ak09911.
New functionality
* Dummy driver gains some virtual registers making it more flexible.
* IIO_ACTIVITY channel types, with modifiers running, walking etc. This is
to support on chip motion clasifiers. As such it is in the form of a
confidence percentage. The only devices so far only do binary decisions
but this gives us room when other devices give more nuanced clasification.
* IIO_EV_DIR_NONE type for events where there is no obvious direction.
First case is step detection.
* IIO_STEPS channel type for pedometers.
* ENABLE mask element used to control turning on counting types such as
the pedometer that need a 'start point'.
* INSTANCE event type to support things that happen once.
* info element for height calibration (used in various motion estimation
algorithms). Note heigh tof use
* dummy driver demonstration of the use of all the new bits above.
* event monitor support for the new events.
* inv_mpu6050 gains an i2c mux to allow bypassing the device to access
additional devices connected on the other side of it. Note that in
Windows these are handled by firmware on the device and not exposed
directly.
* inv_mpu6050 gains ACPI enumeration.
* inkern interface gains iio_write_channel_raw to allow in kernel users
of DAC functionality via a simple wrapper.
* Document input current readings in the ABI docs.
* Add an error message when we get an out of range error in device tree
processing for the in kernel interfaces. Basically a device tree debugging
aid.
* Add a sanity check that a scan index for a channel is unique during
registration. There to help catch bugs as this should never happen
in a bug free driver.
Cleanups and fixlets
A rework of buffer registration from Lars - a precursor to some other
upcoming new stuff (a few patches from others rolled in here as well).
* Ensure all drivers register the same channels for the device and buffer.
* Move buffer registration into the core rather than using the old
two step approach. Now we have simple ways of using a unified set channels
for both without requiring channels be exposed by both interface, this
removes a fair bit of boilerplate.
* Stop sca3000 and ad5933 (both in staging) enabling buffer channels by
default. It has long be convention in IIO to startup with no channels
enabled and leave it up to userspace to say what goes in the buffer.
Getting rid of these allows us to drop export of iio_scan_mask_set.
* Drop get_bytes_per_datum from iio_buffer_access_funcs as not been used
for a while.
* Allocate standard buffer attributes in the core rather than in every
driver with a buffer.
* Make the length attribute read only when a driver is not able to set
the length.
* Drop the get_length callback for buffers as it is already available in
struct iio_buffer.
* Drop an unused arguement form iio_kfifo_allocate and add devm allocator
for it.
* some kconfig entries gain anotation with the resulting module name.
* Fix a resulting compile issue in dummy driver due to a stub taking
wrong parameters as a result of the above rework.
* Fix an off by 2 error in copying the core assigned buffer attributes.
Other cleanups,
* Trivial space before comma fixups.
* ak8975 fixlets - none critical. Rework to allow more device support.
* Drop unnecessary sizeof(u8) calls.
* bmp280 - refactor the compensation code to reduce copy operations and
code length. A second patch futher optimized this and performed some
other minor cleanups.
* kxcjk-1013 - various power control cleanups to avoid unnecessary enable
/ disable of device. Make sure it is only controlled at all if CONFIG_PM
is enabled. Also som cleanups of error paths.
* Small cleanups in adf4530 driver - pointless message and unnecessary braces.
* Clarifiy the proximity ABI docs to make it clear it should get bigger
as we move futher away.
* Drop a misleading comment form industrialio-core.c
* Trivial white space cleanups.
* sca3000 looses an unused debug function.
* Fix char unsigned ordering in ad8366
* Increase the sleep time in ad9523 to make it predictable (value didn't
really matter so make it more than 20 msecs)
* mxs-lradc touchscreen property cleanups in device tree are fixed to ensure
the meet all the 'interesting' documentation.
* A couple of cleanups for the staging ad5933 driver to avoid unnecessary
conversion to a processed temperature vlaue in kernel and remove
platform data form the state structure as not needed after probe.
* Fix a wrong scale factor in the docs.
Misc
* Add IIO include files to the maintainers entry.
Daniel Baluta [Fri, 9 Jan 2015 09:39:30 +0000 (11:39 +0200)]
iio kmx61 / PM: Replace CONFIG_PM_RUNTIME with CONFIG_PM
After commit b2b49ccbdd54 (PM: Kconfig: Set PM_RUNTIME if PM_SLEEP is
selected) PM_RUNTIME is always set if PM is set, so #ifdef blocks
depending on CONFIG_PM_RUNTIME may now be changed to depend on
CONFIG_PM.
Replace CONFIG_PM_RUNTIME with CONFIG_PM everywhere under
drivers/iio/imu/kmx61.c
Signed-off-by: Daniel Baluta <daniel.baluta@intel.com> Signed-off-by: Jonathan Cameron <jic23@kernel.org>
Linus Torvalds [Sun, 18 Jan 2015 06:00:40 +0000 (18:00 +1200)]
Merge tag 'armsoc-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc
Pull ARM SoC fixes from Olof Johansson:
"We've been sitting on our fixes branch for a while, so this batch is
unfortunately on the large side.
A lot of these are tweaks and fixes to device trees, fixing various
bugs around clocks, reg ranges, etc. There's also a few defconfig
updates (which are on the late side, no more of those).
All in all the diffstat is bigger than ideal at this time, but nothing
in here seems particularly risky"
* tag 'armsoc-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (31 commits)
reset: sunxi: fix spinlock initialization
ARM: dts: disable CCI on exynos5420 based arndale-octa
drivers: bus: check cci device tree node status
ARM: rockchip: disable jtag/sdmmc autoswitching on rk3288
ARM: nomadik: fix up leftover device tree pins
ARM: at91: board-dt-sama5: add phy_fixup to override NAND_Tree
ARM: at91/dt: sam9263: Add missing clocks to lcdc node
ARM: at91: sama5d3: dt: correct the sound route
ARM: at91/dt: sama5d4: fix the timer reg length
ARM: exynos_defconfig: Enable LM90 driver
ARM: exynos_defconfig: Enable options for display panel support
arm: dts: Use pmu_system_controller phandle for dp phy
ARM: shmobile: sh73a0 legacy: Set .control_parent for all irqpin instances
ARM: dts: berlin: correct BG2Q's SM GPIO location.
ARM: dts: berlin: add broken-cd and set bus width for eMMC in Marvell DMP DT
ARM: dts: berlin: fix io clk and add missing core clk for BG2Q sdhci2 host
ARM: dts: Revert disabling of smc91x for n900
ARM: dts: imx51-babbage: Fix ULPI PHY reset modelling
ARM: dts: dra7-evm: fix qspi device tree partition size
ARM: omap2plus_defconfig: use CONFIG_CPUFREQ_DT
...
Linus Torvalds [Sun, 18 Jan 2015 03:29:11 +0000 (15:29 +1200)]
Merge tag 'clk-fixes-for-linus' of git://git.linaro.org/people/mike.turquette/linux
Pull clock driver fixes from Mike Turquette:
"Small number of fixes for clock drivers and a single null pointer
dereference fix in the framework core code.
The driver fixes vary from fixing section mismatch warnings to
preventing machines from hanging (and preventing developers from
crying)"
* tag 'clk-fixes-for-linus' of git://git.linaro.org/people/mike.turquette/linux:
clk: fix possible null pointer dereference
Revert "clk: ppc-corenet: Fix Section mismatch warning"
clk: rockchip: fix deadlock possibility in cpuclk
clk: berlin: bg2q: remove non-exist "smemc" gate clock
clk: at91: keep slow clk enabled to prevent system hang
clk: rockchip: fix rk3288 cpuclk core dividers
clk: rockchip: fix rk3066 pll lock bit location
clk: rockchip: Fix clock gate for rk3188 hclk_emem_peri
clk: rockchip: add CLK_IGNORE_UNUSED flag to fix rk3066/rk3188 USB Host
Linus Torvalds [Sun, 18 Jan 2015 03:26:52 +0000 (15:26 +1200)]
Merge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Pull SCSI fixes from James Bottomley:
"This is one fix for a Multiqueue sleeping in invalid context problem
and a MAINTAINER file update for Qlogic"
* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
scsi: ->queue_rq can't sleep
MAINTAINERS: Update maintainer list for qla4xxx
Fabian Holler [Wed, 7 Jan 2015 09:39:54 +0000 (10:39 +0100)]
staging: android: remove space after casts
This patch fixes the following checkpatch.pl warnings:
ashmem.c:753: CHECK:SPACING: No space is necessary after a cast
ashmem.c:756: CHECK:SPACING: No space is necessary after a cast
ashmem.c:777: CHECK:SPACING: No space is necessary after a cast
Fabian Holler [Wed, 7 Jan 2015 09:39:53 +0000 (10:39 +0100)]
staging: android: remove blank lines after open braces
This patch fixes the following checkpatch.pl warnings:
ashmem.c:552: CHECK:BRACES: Blank lines aren't necessary after an open brace '{'
ashmem.c:801: CHECK:BRACES: Blank lines aren't necessary after an open brace
projects / karo-tx-linux.git / log