diff options
author | Linus Torvalds <torvalds@linux-foundation.org> | 2020-10-16 15:02:21 -0700 |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2020-10-16 15:02:21 -0700 |
commit | 54a4c789ca8091ab8fcd70285caeee2c5bc62997 (patch) | |
tree | 1bc7d8b12a9456505fb2cbe708f1c2aabbe69039 /Documentation/userspace-api/media/v4l/func-read.rst | |
parent | 93f3d8f54a22eaa5ae4ec269615729c4f9b1cf1e (diff) | |
parent | 3e2ac9798e13ad1f52d735ea2ea1d252cb140ae5 (diff) |
Merge tag 'docs/v5.10-1' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media
Pull documentation updates from Mauro Carvalho Chehab:
"A series of patches addressing warnings produced by make htmldocs.
This includes:
- kernel-doc markup fixes
- ReST fixes
- Updates at the build system in order to support newer versions of
the docs build toolchain (Sphinx)
After this series, the number of html build warnings should reduce
significantly, and building with Sphinx 3.1 or later should now be
supported (although it is still recommended to use Sphinx 2.4.4).
As agreed with Jon, I should be sending you a late pull request by the
end of the merge window addressing remaining issues with docs build,
as there are a number of warning fixes that depends on pull requests
that should be happening along the merge window.
The end goal is to have a clean htmldocs build on Kernel 5.10.
PS. It should be noticed that Sphinx 3.0 is not currently supported,
as it lacks support for C domain namespaces. Such feature, needed in
order to document uAPI system calls with Sphinx 3.x, was added only on
Sphinx 3.1"
* tag 'docs/v5.10-1' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media: (75 commits)
PM / devfreq: remove a duplicated kernel-doc markup
mm/doc: fix a literal block markup
workqueue: fix a kernel-doc warning
docs: virt: user_mode_linux_howto_v2.rst: fix a literal block markup
Input: sparse-keymap: add a description for @sw
rcu/tree: docs: document bkvcache new members at struct kfree_rcu_cpu
nl80211: docs: add a description for s1g_cap parameter
usb: docs: document altmode register/unregister functions
kunit: test.h: fix a bad kernel-doc markup
drivers: core: fix kernel-doc markup for dev_err_probe()
docs: bio: fix a kerneldoc markup
kunit: test.h: solve kernel-doc warnings
block: bio: fix a warning at the kernel-doc markups
docs: powerpc: syscall64-abi.rst: fix a malformed table
drivers: net: hamradio: fix document location
net: appletalk: Kconfig: Fix docs location
dt-bindings: fix references to files converted to yaml
memblock: get rid of a :c:type leftover
math64.h: kernel-docs: Convert some markups into normal comments
media: uAPI: buffer.rst: remove a left-over documentation
...
Diffstat (limited to 'Documentation/userspace-api/media/v4l/func-read.rst')
-rw-r--r-- | Documentation/userspace-api/media/v4l/func-read.rst | 39 |
1 files changed, 18 insertions, 21 deletions
diff --git a/Documentation/userspace-api/media/v4l/func-read.rst b/Documentation/userspace-api/media/v4l/func-read.rst index 56b255c595e1..e6f6ac4bed77 100644 --- a/Documentation/userspace-api/media/v4l/func-read.rst +++ b/Documentation/userspace-api/media/v4l/func-read.rst @@ -1,4 +1,5 @@ .. SPDX-License-Identifier: GFDL-1.1-no-invariants-or-later +.. c:namespace:: V4L .. _func-read: @@ -11,7 +12,6 @@ Name v4l2-read - Read from a V4L2 device - Synopsis ======== @@ -19,15 +19,13 @@ Synopsis #include <unistd.h> - .. c:function:: ssize_t read( int fd, void *buf, size_t count ) - :name: v4l2-read Arguments ========= ``fd`` - File descriptor returned by :ref:`open() <func-open>`. + File descriptor returned by :c:func:`open()`. ``buf`` Buffer to be filled @@ -38,48 +36,48 @@ Arguments Description =========== -:ref:`read() <func-read>` attempts to read up to ``count`` bytes from file +:c:func:`read()` attempts to read up to ``count`` bytes from file descriptor ``fd`` into the buffer starting at ``buf``. The layout of the data in the buffer is discussed in the respective device interface -section, see ##. If ``count`` is zero, :ref:`read() <func-read>` returns zero +section, see ##. If ``count`` is zero, :c:func:`read()` returns zero and has no other results. If ``count`` is greater than ``SSIZE_MAX``, the result is unspecified. Regardless of the ``count`` value each -:ref:`read() <func-read>` call will provide at most one frame (two fields) +:c:func:`read()` call will provide at most one frame (two fields) worth of data. -By default :ref:`read() <func-read>` blocks until data becomes available. When -the ``O_NONBLOCK`` flag was given to the :ref:`open() <func-open>` +By default :c:func:`read()` blocks until data becomes available. When +the ``O_NONBLOCK`` flag was given to the :c:func:`open()` function it returns immediately with an ``EAGAIN`` error code when no data -is available. The :ref:`select() <func-select>` or -:ref:`poll() <func-poll>` functions can always be used to suspend +is available. The :c:func:`select()` or +:c:func:`poll()` functions can always be used to suspend execution until data becomes available. All drivers supporting the -:ref:`read() <func-read>` function must also support :ref:`select() <func-select>` and -:ref:`poll() <func-poll>`. +:c:func:`read()` function must also support :c:func:`select()` and +:c:func:`poll()`. Drivers can implement read functionality in different ways, using a single or multiple buffers and discarding the oldest or newest frames once the internal buffers are filled. -:ref:`read() <func-read>` never returns a "snapshot" of a buffer being filled. +:c:func:`read()` never returns a "snapshot" of a buffer being filled. Using a single buffer the driver will stop capturing when the application starts reading the buffer until the read is finished. Thus only the period of the vertical blanking interval is available for reading, or the capture rate must fall below the nominal frame rate of the video standard. -The behavior of :ref:`read() <func-read>` when called during the active picture +The behavior of :c:func:`read()` when called during the active picture period or the vertical blanking separating the top and bottom field depends on the discarding policy. A driver discarding the oldest frames keeps capturing into an internal buffer, continuously overwriting the previously, not read frame, and returns the frame being received at the -time of the :ref:`read() <func-read>` call as soon as it is complete. +time of the :c:func:`read()` call as soon as it is complete. A driver discarding the newest frames stops capturing until the next -:ref:`read() <func-read>` call. The frame being received at :ref:`read() <func-read>` +:c:func:`read()` call. The frame being received at :c:func:`read()` time is discarded, returning the following frame instead. Again this implies a reduction of the capture rate to one half or less of the nominal frame rate. An example of this model is the video read mode of -the bttv driver, initiating a DMA to user memory when :ref:`read() <func-read>` +the bttv driver, initiating a DMA to user memory when :c:func:`read()` is called and returning when the DMA finished. In the multiple buffer model drivers maintain a ring of internal @@ -94,14 +92,13 @@ the driver with the :ref:`VIDIOC_G_PARM <VIDIOC_G_PARM>` and however. The discarding policy is not reported and cannot be changed. For minimum requirements see :ref:`devices`. - Return Value ============ On success, the number of bytes read is returned. It is not an error if this number is smaller than the number of bytes requested, or the amount of data required for one frame. This may happen for example because -:ref:`read() <func-read>` was interrupted by a signal. On error, -1 is +:c:func:`read()` was interrupted by a signal. On error, -1 is returned, and the ``errno`` variable is set appropriately. In this case the next read will start at the beginning of a new frame. Possible error codes are: @@ -129,5 +126,5 @@ EIO communicate with a remote device (USB camera etc.). EINVAL - The :ref:`read() <func-read>` function is not supported by this driver, not + The :c:func:`read()` function is not supported by this driver, not on this device, or generally not on this type of device. |