[media] v4l: omap4iss: Signal fatal errors to the vb2 queue
When a fatal error occurs in the pipeline signal it to the vb2 queue
with a call to vb2_queue_error(). The queue will then take care to
return -EIO when preparing buffers, remove the driver-specific code that
now duplicates that check.
The OMAP4 camera support depends on I2C and VIDEO_V4L2, both
of which can be loadable modules. This causes build failures
if we want the camera driver to be built-in.
This can be solved by turning the option into "tristate",
which unfortunately causes another problem, because the
driver incorrectly calls a platform-internal interface
for omap4_ctrl_pad_readl/omap4_ctrl_pad_writel.
Instead, this patch just forbids the invalid configurations
and ensures that the driver can only be built if all its
dependencies are built-in.
When a fatal error occurs that render the device unusable, the only
options for a driver to signal the error condition to userspace is to
set the V4L2_BUF_FLAG_ERROR flag when dequeuing buffers and to return an
error from the buffer prepare handler when queuing buffers.
The buffer error flag indicates a transient error and can't be used by
applications to detect fatal errors. Returning an error from vb2_qbuf()
is thus the only real indication that a fatal error occurred. However,
this is difficult to handle for multithreaded applications that requeue
buffers from a thread other than the control thread. In particular the
poll() call in the control thread will not notify userspace of the
error.
This patch adds an explicit mechanism to report fatal errors to
userspace. Drivers can call the vb2_queue_error() function to signal a
fatal error. From this moment on, buffer preparation will return -EIO to
userspace, and vb2_poll() will set the POLLERR flag and return
immediately. The error flag is cleared when cancelling the queue, either
at stream off time (through vb2_streamoff) or when releasing the queue
with vb2_queue_release().
[media] v4l: vb2: Don't return POLLERR during transient buffer underruns
The V4L2 specification states that
"When the application did not call VIDIOC_QBUF or VIDIOC_STREAMON yet
the poll() function succeeds, but sets the POLLERR flag in the revents
field."
The vb2_poll() function sets POLLERR when the queued buffers list is
empty, regardless of whether this is caused by the stream not being
active yet, or by a transient buffer underrun.
Bring the implementation in line with the specification by returning
POLLERR if no buffer has been queued only when the queue is not
streaming. Buffer underruns during streaming are not treated specially
anymore and just result in poll() blocking until the next event.
Laurent Pinchart [Sat, 31 May 2014 00:45:48 +0000 (21:45 -0300)]
[media] v4l: vsp1: uds: Fix scaling of alpha layer
Pixel color components can be scaled using either bilinear interpolation
or a multitap filter. The multitap filter provides better results, but
can't be selected when the alpha layer need to be scaled down by more
than 1/2.
Disable alpha scaling when the input has a fixed alpha value, and
program the UDS to output a fixed alpha value in that case. This ensures
the multitap filter will be used whenever possible.
Laurent Pinchart [Tue, 27 May 2014 23:35:36 +0000 (20:35 -0300)]
[media] v4l: vsp1: bru: Support premultiplied alpha at the BRU inputs
Adjust the BRU blending formula to avoid the multiplication by alpha
when the corresponding input format is premultiplied. As this requires
access to the RPFs connected to the BRU inputs from the BRU module,
store pointers to the RPFs in the BRU structure when validating the
pipeline.
Laurent Pinchart [Mon, 26 May 2014 22:57:21 +0000 (19:57 -0300)]
[media] v4l: vsp1: Switch to XRGB formats
The driver ignores the alpha component on output video nodes and
hardcodes the alpha component to 0 on capture video nodes. Make this
explicit by exposing XRGB formats.
Compatibility with existing userspace applications is handled by
selecting the XRGB format corresponding to the requested old RGB format.
Laurent Pinchart [Wed, 28 May 2014 16:10:33 +0000 (13:10 -0300)]
[media] v4l: vsp1: wpf: Clear RPF to WPF association at stream off time
The VSP1 stores the video pipelines' input (RPF) to output (WPF)
mappings in a WPF register. An RPF must never be associated with
multiple WPFs, even if all of those WPFs but one are unused, otherwise
the hardware won't function properly.
The driver doesn't ensure this correctly as it never clears the
mappings. An RPF used with one WPF and later with a different WPF will
lead to malfunction, as it will be associated with two WPFs. Clear the
mappings at stream off time to fix this.
Laurent Pinchart [Sat, 31 May 2014 23:30:11 +0000 (20:30 -0300)]
[media] v4l: vsp1: sru: Make the intensity controllable during streaming
The control value is currently stored in the SRU structure by the
control set handler and written to the hardware at stream on time,
making control set during streaming ineffective. Fix it by writing to
the registers from within the control set handler.
Laurent Pinchart [Sat, 31 May 2014 13:40:51 +0000 (10:40 -0300)]
[media] v4l: vsp1: Setup control handler automatically at stream on time
When setting a control directly on a subdev node the VSP1 driver doesn't
guarantee that the device is powered on. This leads to crashes when the
control handlers writes to hardware registers. One easy way to fix this
is to ensure that the device gets powered on when a subdev node is
opened. However, this consumes power unnecessarily, as there's no need
to power the device on when setting formats on the pipeline.
Furthermore, control handler setup at entity init time suffers from the
same problem as the device isn't powered on easier.
Fix this by extend the entity base object to setup the control handler
automatically when starting the stream. Entities must then skip writing
to registers in the set control handler when not streaming, which can be
tested with the new vsp1_entity_is_streaming() helper function.
Laurent Pinchart [Wed, 28 May 2014 15:49:13 +0000 (12:49 -0300)]
[media] v4l: vsp1: Cleanup video nodes at removal time
Video nodes created and initialized in the RPF and WPF init code paths
are never unregistered, and the related resources (videobuf alloc
context and media entity) never released.
Fix this by storing a pointer to the vsp1_video object in vsp1_entity
and calling vsp1_video_cleanup() from vsp1_entity_destroy(). This also
allows simplifying the init error code paths.
Laurent Pinchart [Mon, 23 Jun 2014 20:15:02 +0000 (17:15 -0300)]
[media] v4l: vsp1: Fix pipeline stop timeout
If the pipeline was already stopped when stopping the stream, no
frame end interrupt will be generated and the driver will time out
waiting for the pipeline to stop.
Fix this by setting the pipeline state to STOPPED when the pipeline is
idle waiting for frames to process, and to STOPPING at stream stop time
only when the pipeline is currently RUNNING.
Laurent Pinchart [Mon, 23 Jun 2014 19:57:22 +0000 (16:57 -0300)]
[media] v4l: vsp1: Release buffers at stream stop
videobuf2 expects no buffer to be owned by the driver when the
stop_stream queue operation returns. As the vsp1 driver fails to do so,
a warning is generated at stream top time.
Fix this by releasing all buffers queued on the IRQ queue in the
stop_stream operation handler and marking them as erroneous.
Laurent Pinchart [Mon, 23 Jun 2014 21:40:49 +0000 (18:40 -0300)]
[media] v4l: vsp1: Fix routing cleanup when stopping the stream
Commit d9b45ed3d8b75e8cf38c8cd1563c29217eecba27 ("v4l: vsp1: Support
multi-input entities") reworked pipeline routing configuration and
introduced a bug by writing to the entities routing registers without
first checking whether the entity had a routing register. This results
in overwriting the value at offset 0 of the device register space when
stopping the stream.
Fix this by skipping routing register write for entities without a
routing register.
Laurent Pinchart [Mon, 23 Jun 2014 21:00:22 +0000 (18:00 -0300)]
[media] v4l: vb2: Fix stream start and buffer completion race
videobuf2 stores the driver streaming state internally in the queue in
the start_streaming_called variable. The state is set right after the
driver start_stream operation returns, and checked in the
vb2_buffer_done() function, typically called from the frame completion
interrupt handler. A race condition exists if the hardware finishes
processing the first frame before the start_stream operation returns.
Fix this by setting start_streaming_called to 1 before calling the
start_stream operation, and resetting it to 0 if the operation fails.
Cc: stable@vger.kernel.org # for v3.15 and up Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com> Reviewed-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
Laurent Pinchart [Tue, 27 May 2014 12:41:05 +0000 (09:41 -0300)]
[media] v4l: Support extending the v4l2_pix_format structure
The v4l2_pix_format structure has no reserved field. It is embedded in
the v4l2_framebuffer structure which has no reserved fields either, and
in the v4l2_format structure which has reserved fields that were not
previously required to be zeroed out by applications.
To allow extending v4l2_pix_format, inline it in the v4l2_framebuffer
structure, and use the priv field as a magic value to indicate that the
application has set all v4l2_pix_format extended fields and zeroed all
reserved fields following the v4l2_pix_format field in the v4l2_format
structure.
The availability of this API extension is reported to userspace through
the new V4L2_CAP_EXT_PIX_FORMAT capability flag. Just checking that the
priv field is still set to the magic value at [GS]_FMT return wouldn't
be enough, as older kernels don't zero the priv field on return.
To simplify the internal API towards drivers zero the extended fields
and set the priv field to the magic value for applications not aware of
the extensions.
Laurent Pinchart [Wed, 21 May 2014 22:16:01 +0000 (19:16 -0300)]
[media] DocBook: media: Document ALPHA_COMPONENT control usage on output devices
Extend the V4L2_CID_ALPHA_COMPONENT control for use on output devices,
to set the alpha component value when the output format doesn't have an
alpha channel.
Laurent Pinchart [Thu, 22 May 2014 01:07:55 +0000 (22:07 -0300)]
[media] v4l: Add ARGB and XRGB pixel formats
The existing RGB pixel formats are ill-defined in respect to their alpha
bits and their meaning is driver dependent. Create new standard ARGB and
XRGB variants with clearly defined meanings and make the existing
variants deprecated.
The new pixel formats 4CC values have been selected to match the DRM
4CCs for the same in-memory formats.
Hans Verkuil [Thu, 17 Jul 2014 15:31:23 +0000 (12:31 -0300)]
[media] Fix 64-bit division fall-out from 64-bit control ranges
Commit 0ba2aeb6dab80920edd9cf5b93b1ea4d6913b8f3 increased the internal control ranges
to 64 bit, but that caused problems in drivers that use the minimum/maximum/step/default_value
control values in a division or modulus operations since not all architectures support
those natively.
Luckily, in almost all cases it is possible to just cast to 32 bits (the control value
is known to be 32 bits, so it is safe to cast). Only in v4l2-ctrls.c was it necessary to
use do_div in one function.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
Hans Verkuil [Tue, 10 Jun 2014 10:39:04 +0000 (07:39 -0300)]
[media] go7007: add motion detection support
This patch adds motion detection support to the go7007 driver using the new
motion detection controls, events.
The global motion detection works fine, but the regional motion detection
support probably needs more work. There seems to be some interaction between
regions that makes setting correct thresholds difficult. The exact meaning of
the thresholds isn't entirely clear either.
I do not have any documentation, the only information I have is the custom code
in the driver and a modet.c application.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
Hans Verkuil [Mon, 5 May 2014 13:16:49 +0000 (10:16 -0300)]
[media] DocBook media: improve control section
Improve the control section:
- Clarify the handling of private controls
- Explain the V4L2_CTRL_FLAG_INACTIVE flag
- Remove obsolete text regarding missing control event (we have them
today) and the incorrect V4L2_CTRL_FLAG_DISABLED reference.
- Add a code example on how to enumerate over user controls.
Hans Verkuil [Sat, 18 Jan 2014 09:06:01 +0000 (06:06 -0300)]
[media] v4l2-ctrls: return elem_size instead of strlen
When getting a string and the size given by the application is too
short return the max length the string can have (elem_size) instead
of the string length + 1. That makes more sense.
Hans Verkuil [Fri, 17 Jan 2014 11:25:26 +0000 (08:25 -0300)]
[media] v4l2-ctrls: compare values only once
When setting a control the control's new value is compared to the current
value twice: once by new_to_cur(), once by cluster_changed(). Not a big
deal when dealing with simple values, but it can be a problem when dealing
with compound types or arrays. So fix this: cluster_changed() sets the
has_changed flag, which is used by new_to_cur() instead of having to do
another compare.
Hans Verkuil [Sun, 27 Apr 2014 06:26:30 +0000 (03:26 -0300)]
[media] v4l2-ctrls: rewrite copy routines to operate on union v4l2_ctrl_ptr
In order to implement array support and (for the future) configuration stores
we need to have more generic copy routines that all operate on the v4l2_ctrl_ptr
union. So instead of e.g. using ctrl->cur.string it uses ptr.p_char. This makes
e.g. cur_to_user generic so it can be used to copy any v4l2_ctrl_ptr value to
userspace, not just the (hardcoded) current value.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
Hans Verkuil [Sun, 27 Apr 2014 06:22:17 +0000 (03:22 -0300)]
[media] v4l2-ctrls: create type_ops
Since compound controls can have non-standard types we need to be able to do
type-specific checks etc. In order to make that easy type operations are added.
There are four operations:
- equal: check if two values are equal
- init: initialize a value
- log: log the value
- validate: validate a new value
The v4l2_ctrl struct adds p_new and p_cur unions at the end of the struct.
This union provides a standard way of accessing control types through a pointer,
which greatly simplifies internal control processing.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
Hans Verkuil [Thu, 12 Jun 2014 10:54:16 +0000 (07:54 -0300)]
[media] v4l2-ctrls: add support for compound types
This patch implements initial support for compound types.
The changes are fairly obvious: basic support for is_ptr types, the
type_is_int function is replaced by a is_int bitfield, and
v4l2_query_ext_ctrl is added.
Note that this patch does not yet add support for N-dimensional
arrays, that comes later. So v4l2_query_ext_ctrl just sets elems to
1 and nr_of_dims and dims[] are all zero.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
Hans Verkuil [Thu, 27 Mar 2014 14:25:45 +0000 (11:25 -0300)]
[media] videodev2.h: add initial support for compound controls
Compound controls are controls that can be used for compound and array
types. This allows for more compound data structures to be used with the
control framework.
The existing V4L2_CTRL_FLAG_NEXT_CTRL flag will only enumerate non-compound
controls, so a new V4L2_CTRL_FLAG_NEXT_COMPOUND flag is added to enumerate
compound controls. Set both flags to enumerate any control (compound or not).
Compound control types will start at V4L2_CTRL_COMPOUND_TYPES. In addition, any
control that uses the new 'ptr' field or the existing 'string' field will have
flag V4L2_CTRL_FLAG_HAS_PAYLOAD set.
While not strictly necessary, adding that flag makes life for applications
a lot simpler. If the flag is not set, then the control value is set
through the value or value64 fields of struct v4l2_ext_control, otherwise
a pointer points to the value.
Hans Verkuil [Wed, 16 Apr 2014 12:41:25 +0000 (09:41 -0300)]
[media] v4l2-ctrls: increase internal min/max/step/def to 64 bit
While VIDIOC_QUERYCTRL is limited to 32 bit min/max/step/def values
for controls, the upcoming VIDIOC_QUERY_EXT_CTRL isn't. So increase
the internal representation to 64 bits in preparation.
Because of these changes the msi3101 driver has been modified slightly
to fix a formatting issue (%d becomes %lld), vivi had to be modified
as well to cope with the new 64-bit min/max values and the PIXEL_RATE
control in a few sensor drivers required proper min/max/def values.
Cc: Michael Krufky <mkrufky@linuxtv.org> Signed-off-by: Olli Salonen <olli.salonen@iki.fi> Reviewed-by: Antti Palosaari <crope@iki.fi> Signed-off-by: Antti Palosaari <crope@iki.fi> Reviewed-by: Michael Ira Krufky <mkrufky@linuxtv.org> Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
Merge tag 'ext4_for_linus_stable' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4
Pull ext4 bugfixes from Ted Ts'o:
"More bug fixes for ext4 -- most importantly, a fix for a bug
introduced in 3.15 that can end up triggering a file system corruption
error after a journal replay.
It shouldn't lead to any actual data corruption, but it is scary and
can force file systems to be remounted read-only, etc"
* tag 'ext4_for_linus_stable' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4:
ext4: fix potential null pointer dereference in ext4_free_inode
ext4: fix a potential deadlock in __ext4_es_shrink()
ext4: revert commit which was causing fs corruption after journal replays
ext4: disable synchronous transaction batching if max_batch_time==0
ext4: clarify ext4_error message in ext4_mb_generate_buddy_error()
ext4: clarify error count warning messages
ext4: fix unjournalled bg descriptor while initializing inode bitmap
Merge tag 'clk-fixes-for-linus' of git://git.linaro.org/people/mike.turquette/linux
Pull clock driver fixes from Mike Turquette:
"This batch of fixes is for a handful of clock drivers from Allwinner,
Samsung, ST & TI. Most of them are of the "this hardware won't work
without this fix" variety, including patches that fix platforms that
did not boot under certain configurations. Other fixes are the result
of changes to the clock core introduced in 3.15 that had subtle
impacts on the clock drivers.
There are no fixes to the clock framework core in this pull request"
* tag 'clk-fixes-for-linus' of git://git.linaro.org/people/mike.turquette/linux:
clk: spear3xx: Set proper clock parent of uart1/2
clk: spear3xx: Use proper control register offset
clk: qcom: HDMI source sel is 3 not 2
clk: sunxi: fix devm_ioremap_resource error detection code
clk: s2mps11: Fix double free corruption during driver unbind
clk: ti: am43x: Fix boot with CONFIG_SOC_AM33XX disabled
clk: exynos5420: Remove aclk66_peric from the clock tree description
clk/exynos5250: fix bit number for tv sysmmu clock
clk: s3c64xx: Hookup SPI clocks correctly
clk: samsung: exynos4: Remove SRC_MASK_ISP gates
clk: samsung: add more aliases for s3c24xx
clk: samsung: fix several typos to fix boot on s3c2410
clk: ti: set CLK_SET_RATE_NO_REPARENT for ti,mux-clock
clk: ti: am43x: Fix boot with CONFIG_SOC_AM33XX disabled
clk: ti: dra7: return error code in failure case
clk: ti: apll: not allocating enough data
Merge tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc
Pull ARM SoC fixes from Olof Johansson:
"This week's arm-soc fixes:
- Another set of OMAP fixes
* Clock fixes
* Restart handling
* PHY regulators
* SATA hwmod data for DRA7
+ Some trivial fixes and removal of a bit of dead code
- Exynos fixes
* A bunch of clock fixes
* Some SMP fixes
* Exynos multi-core timer: register as clocksource and fix ftrace.
+ a few other minor fixes
There's also a couple more patches, and at91 fix for USB caused by
common clock conversion, and more MAINTAINERS entries for shmobile.
We're definitely switching to only regression fixes from here on out,
we've been a little less strict than usual up until now"
* tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (26 commits)
ARM: at91: at91sam9x5: add clocks for usb device
ARM: EXYNOS: Register cpuidle device only on exynos4210 and 5250
ARM: dts: Add clock property for mfc_pd in exynos5420
clk: exynos5420: Add IDs for clocks used in PD mfc
ARM: EXYNOS: Add support for clock handling in power domain
ARM: OMAP2+: Remove non working OMAP HDMI audio initialization
ARM: imx: fix shared gate clock
ARM: dts: Update the parent for Audss clocks in Exynos5420
ARM: EXYNOS: Update secondary boot addr for secure mode
ARM: dts: Fix TI CPSW Phy mode selection on IGEP COM AQUILA.
ARM: dts: am335x-evmsk: Enable the McASP FIFO for audio
ARM: dts: am335x-evm: Enable the McASP FIFO for audio
ARM: OMAP2+: Make GPMC skip disabled devices
ARM: OMAP2+: create dsp device only on OMAP3 SoCs
ARM: dts: dra7-evm: Make VDDA_1V8_PHY supply always on
ARM: DRA7/AM43XX: fix header definition for omap44xx_restart
ARM: OMAP2+: clock/dpll: fix _dpll_test_fint arithmetics overflow
ARM: DRA7: hwmod: Add SYSCONFIG for usb_otg_ss
ARM: DRA7: hwmod: Fixup SATA hwmod
ARM: OMAP3: PRM/CM: Add back macros used by TI DSP/Bridge driver
...
Merge branch 'fixes' of git://ftp.arm.linux.org.uk/~rmk/linux-arm
Pull ARM fixes from Russell King:
"Another round of fixes for ARM:
- a set of kprobes fixes from Jon Medhurst
- fix the revision checking for the L2 cache which wasn't noticed to
have been broken"
* 'fixes' of git://ftp.arm.linux.org.uk/~rmk/linux-arm:
ARM: l2c: fix revision checking
ARM: kprobes: Fix test code compilation errors for ARMv4 targets
ARM: kprobes: Disallow instructions with PC and register specified shift
ARM: kprobes: Prevent known test failures stopping other tests running
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux-m68k
Pull m68k fixes from Geert Uytterhoeven:
"Summary:
- Fix for a boot regression introduced in v3.16-rc1,
- Fix for a build issue in -next"
Christoph Hellwig questioned why mach_random_get_entropy should be
exported to modules, and Geert explains that random_get_entropy() is
called by at least the crypto layer and ends up using it on m68k. On
most other architectures it just uses get_cycles() (which is typically
inlined and doesn't need exporting),
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux-m68k:
m68k: Export mach_random_get_entropy to modules
m68k: Fix boot regression on machines with RAM at non-zero
Merge branch 'parisc-3.16-5' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux
Pull parisc fixes from Helge Deller:
"The major patch in here is one which fixes the fanotify_mark() syscall
in the compat layer of the 64bit parisc kernel. It went unnoticed so
long, because the calling syntax when using a 64bit parameter in a
32bit syscall is quite complex and even worse, it may be even
different if you call syscall() or the glibc wrapper. This patch
makes the kernel accept the calling convention when called by the
glibc wrapper.
The other two patches are trivial and remove unused headers, #includes
and adds the serial ports of the fastest C8000 workstation to the
parisc-kernel internal hardware database"
* 'parisc-3.16-5' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux:
parisc: drop unused defines and header includes
parisc: fix fanotify_mark() syscall on 32bit compat kernel
parisc: add serial ports of C8000/1GHz machine to hardware database
parisc: fix fanotify_mark() syscall on 32bit compat kernel
On parisc we can not use the existing compat implementation for fanotify_mark()
because for the 64bit mask parameter the higher and lower 32bits are ordered
differently than what the compat function expects from big endian
architectures.
Specifically:
It finally turned out, that on hppa we end up with different assignments
of parameters to kernel arguments depending on if we call the glibc
wrapper function
int fanotify_mark (int __fanotify_fd, unsigned int __flags,
uint64_t __mask, int __dfd, const char *__pathname);
or directly calling the syscall manually
syscall(__NR_fanotify_mark, ...)
Reason is, that the syscall() function is implemented as C-function and
because we now have the sysno as first parameter in front of the other
parameters the compiler will unexpectedly add an empty paramenter in
front of the u64 value to ensure the correct calling alignment for 64bit
values.
This means, on hppa you can't simply use syscall() to call the kernel
fanotify_mark() function directly, but you have to use the glibc
function instead.
This patch fixes the kernel in the hppa-arch specifc coding to adjust
the parameters in a way as if userspace calls the glibc wrapper function
fanotify_mark().
projects / linux-beck.git / log