The pool orphaning function was colling internal _stop() virtual function
implementation. This is not thread safe, as a private lock inside the buffer
pool is supposed to be held. Fix this by keeping delayed _stop() and orphaning
the GstV4L2Allocator instead (REQBUFS(0)).
Then, protect the orphaned boolean with the object lock for the case a buffer
is being released after we have orphaned the buffer. That would otherwise
cause a QBUF to happen while the queue is no longer owned by the buffer pool.
This boolean is otherwise used and set from the streaming lock, or after
threads have been stopped (final cleanup).
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/648>
On streamon, we need to resurrect (queue back) some buffers, as during
flushign seek we'd endup with an empty queued. We initially started with
resurrecting as many as we could without blocking, but that miss-behaved with
dynamic CREATE_BUFS, causing the pool to grow dramatically. This was limited
by the number of allocated buffers, but this still tried to resurrect too many
buffers for the first run, as activating the pool will queued buffers.
In this patch, we calculte the missing detal in the queue and only try and
resurrect that amount of buffers.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/651>
There are often only two buffers queued in the kernel so no new buffers are
requested.
With every qbuf, the kernel receives a new DMABUF for the specified index.
This most likely differs from the last DMABUF and the old cached entry is
released. This results in a lot of map/unmap overhead if the kernel driver
needs a mapping for the buffer.
With a larger queue, it's quite likely, that both old and new DMABUFs are
also mapped for another index. So the map/unmap is skipped, because the
mapping is reference counted.
The corresponding allocated buffers don't contain any actual memory, so
allocating them is quite cheep. So the log message is updated to clarify
this.
When using this mode each frame is split in two fields, each one being
transferred using its own buffer.
This is implemented with the V4L2_FIELD_ALTERNATE field format in v4l2.
This mode is enabled using a caps filter such as
"v4l2src ! video/x-raw\(format:Interlaced\)"
Here are the main changes related to this feature:
- use the INTERLACED caps feature with this mode.
- in this mode both fields of a given frame have the same sequence/offset
so adjust the algorithm checking for lost field/frame accordingly.
- double pool's min number of buffers as each frame requires 2 buffers.
Fix#504
Co-authored-by: Zeeshan Ali <zeenix@collabora.co.uk>
Before we do streamon, we queue all capture buffers by calling
resurrect. When the driver supports CREATE_BUFS, this would lead
to buffers being allocated till the maximum of 32 is reached.
Instead, we now save the number of allocated buffers and queue this
amount.
Allocator resources cannot be freed when a buffer pool is orphaned
while its buffers are in use. They should, however, be freed once those
buffers are no longer needed. This patch disposes of any buffers
belonging to an orphaned pool as they are released, and makes sure
that the allocator is cleaned up when the last buffer is returned.
When trying to orphan a buffer pool, successfully return and unref
the pool when the pool is either successfully stopped or orphaned.
Indicate failure and leave the pool untouched otherwise.
Now that the v4l2allocator allows orphaning the V4L2 buffer queue, add
support for orphaning in the v4l2bufferpool. gst_v4l2_buffer_pool_orphan
can be used as a replacement for gst_v4l2_buffer_pool_stop, without
having to wait for buffers to be returned to the pool.
The preparation code imports the buffer, doing bunch of
validation. Only queue the buffer in the driver if the
importation worked. This way we don't rely on the driver
to validate.
https://bugzilla.gnome.org/show_bug.cgi?id=583890
This change has no effect. We will need to acquire a buffer from the
pool later in order to validate / adapt with the video alignment for
the downstream buffers.
https://bugzilla.gnome.org/show_bug.cgi?id=583890
When the pool is started, we allocate and release buffer, expecting
the pool release-buffer handler to queue them. Though, as we rely
on release function, there is no direct way to detect that this
process didn't work.
To check this, validate that the number of queued buffer is the same
as the number of allocated buffers. This allow returning an error
when buffer importation was refused by the driver.
https://bugzilla.gnome.org/show_bug.cgi?id=583890
This will avoid sending EOS on v4l2src when a driver sends an empty
buffers. This case would be a bug in the driver, but yet the camera
should keep running.
This also removes the check for corrupted buffers, as this check is
already done later.
https://bugzilla.gnome.org/show_bug.cgi?id=794842
We were calling gst_v4l2_is_buffer_valid() before and inside
gst_v4l2_buffer_pool_qbuf() as we needed to access the group. The second
check failed since the writability of the buffer get inherited from the
GstMemory, which lead to pipeline failure. As we cannot avoid the extra
ref, it would be racy otherwise, just pass the group to _dbuf() so it
does not have to call gst_v4l2_is_buffer_valid() again.
https://bugzilla.gnome.org/show_bug.cgi?id=796692
Drop truncated frames regardless if they have the ERROR flag or not.
Truncated frame causes video frame map failure in many elements
including cluttersink, glupload etc.
In this patch we use a non-blocking poll in order to return all input
buffers (buffers from v4l2-output queue). This prevent holding too long
on upstreaming buffer in importing.
https://bugzilla.gnome.org/show_bug.cgi?id=794904
In MPLANE mode, the driver may set data_offset, which represent some
padding at the start of the buffer used internally. This portion of the
data need to be skipped, though it is included in bytesused.
This patch removes frame size sanity check as the method used will no
longer work. This check was simply there to help detect broken kernel
drivers. It would be re-implement by estimating the plane size, which is
not totally trivial and may be too much work for a simple debug check.
https://bugzilla.gnome.org/show_bug.cgi?id=733501
The purpose of being able to flush the buffer pool is only to
unlock any blocked operation. Doing streamoff/streamon had the
side effect of turning off and on the camera. As we do a flush_start
/ flush_stop sequence when shutting down, that would cause a really
quick sequence of streamoff/streamon/streamoff/close which was
causing some cameras to stop working.
https://bugzilla.gnome.org/show_bug.cgi?id=783945
Whenever we import from downstream pool (userptr or dmabuf-import), we
should copy over the flags and timestamp, otherwise downstream will not
get proper synchronization or will not be able to notice frames that has
corruption in it.
https://bugzilla.gnome.org/show_bug.cgi?id=785680
Removes the FIXME/Question in the buffer pool and add a ref to the
element in the GstAllocator too. This ref is strictly required to keep
the GstV4l2Object structure around.
The library has started preventing a lot of interesting use cases,
like CREATE_BUFS, DMABuf, usage of TRY_FMT. As the libv4l2 is totally
inactive and not maintained, we decided to disable it. As a convenience
we added a run-time environment that let you enable it for testing.
GST_V4L2_USE_LIBV4L2=1
This of course only works if you have enabled libv4l2 at build time.
First step of a larger cleanup, all function from v4l2_calls are in fact
methods on GstV4l2Object. This split makes the code really confusing.
This also remove no longer unused macros.
When resurrecting a buffer, the subsequent free call can result
in the group-released handler being called again, which causes
a recursive loop. This patch blocks the signal handler during
the time that it executes, ensuring that the loop will not occur.
https://bugzilla.gnome.org/show_bug.cgi?id=759292
This is needed for V4L2_OUTPUT interface, and is harmless of
V4L2_CAPTURE interfaces. This will fix timestamp in cases like:
v4l2src io-mode=dmabuf ! v4l2videoNenc output-io-mode=dmabuf-import ! ...
Same apply for userptr.
https://bugzilla.gnome.org/show_bug.cgi?id=781119
These can be called from different threads and both manipulate the
pool->buffers array. Lock them properly and let flush_stop move the
array contents into a temporary array on the stack to avoid having
to call release_buffer under the object lock.
https://bugzilla.gnome.org/show_bug.cgi?id=775015
If the pool is inactive, it is guaranteed to also be flushing, so the
following check will return GST_FLOW_FLUSHING anyway.
This can happen if a v4l2src is blocking on DQBUF in create and is sent
an EOS event on another thread. In that case the pool is set to
flushing/inactive without locking, the v4l2src is unblocked, and may
call pool_process with a valid buffer on the already inactive pool.
https://bugzilla.gnome.org/show_bug.cgi?id=775014
The same physical device can export multiple devices. In
this case, the capabilities field now contains a union of
all caps available from all exported V4L2 devices alongside
a V4L2_CAP_DEVICE_CAPS flag that should be used to decide
what capabilities to consider. In our case, we need the
ones from the exported device we are using.
https://bugzilla.gnome.org/show_bug.cgi?id=768195
