- 05 Dec, 2009 40 commits
-
-
Hans de Goede authored
The gscpa core was not resubmitting URB's when the URB status was an error, this means we will loose URB's (potentially all stopping the stream) in case of sporadic USB issues. I've seen this with an Aiptek stv0680 based PenCam connected through an USB 2.0 hub. Signed-off-by:
Hans de Goede <hdegoede@redhat.com> Signed-off-by:
Mauro Carvalho Chehab <mchehab@redhat.com>
-
Hans de Goede authored
This patch adds a new subdriver to gspca supporting cams with the stv0680 bridge (replacing the old in kernel v4l1 driver). Many thanks to Hans Verkuil for providing me with one of the 2 cams used in testing this new sub driver. Signed-off-by:
-
Hans de Goede authored
gspca w9968cf: slight tweak to vstart and hstart to get rid of black borders in 640x480 mode. Signed-off-by:
-
Hans de Goede authored
gspca w9968cf: Add support for JPEG compression, this enables much higher framerates at 320x240 / 352x288 and also allows for 640x480 mode for cams which can do this. The w9968cf uses planar JPEG, which libv4l until now did not support, so this requires atleast version 0.6.3 of libv4l. And something important I forgot to mention in my earlier w9968cf commits: Many thanks to Hans Verkuil for giving me a w9968cf based cam, which has allowed me to develop the gspca w9968cf support. Signed-off-by:
-
Hans de Goede authored
gspca-w9968cf: Don't add an extra packet to the buffer after the EOF, this makes the frame data size bigger then it should be which makes gstreamer discard it. This patch also fixes an unrelated compiler warning. Signed-off-by:
-
Hans de Goede authored
Cache sensor regs to avoid unnecessary slow i2c reads / writes, this speeds up sd_start a bit with most bridges and a lot (from 5 seconds down to 0.3 seconds) with W996xCF cams, as this avoids very slow bit bang IO over USB i2c reads. Signed-off-by:
-
Hans de Goede authored
This patch adds support to gspca for the Winbond W9967CF and W9968CF camera's. This is mostly a port of the existing v4l1 driver to gspca (making it v4l2). But this also features fixes to the bitbanging i2c code (send a nack not an ack after reading the last byte of a transfer), which gets rid of the weird errors which were being seen there, and of the smbus_refresh() hack to get around these errors. Also the vstart settings have been tweaked to work with different frequency filter settings. Signed-off-by:
-
Richard Röjfors authored
This patch adds support for TEF6862 Car Radio Enhanced Selectivity Tuner. It's implemented as a subdev, supporting checking signal strength and setting and getting frequency. Signed-off-by:
Richard Röjfors <richard.rojfors@mocean-labs.com> Signed-off-by:
-
Richard Röjfors authored
If the I2C device provides an interrupt it is registered and the standard is updated via interrupts rather than polling. Since I2C communication is needed, the interrupt handler fires off a work which will check the new standard, and store it in the internal structure. To handle mutual exclusion a mutex is introduced. Signed-off-by:
-
Richard Röjfors authored
This patch defines the probe and remove function as __devinit and __devexit. Signed-off-by:
-
Richard Röjfors authored
Support for settings the input standard of the ADV7180. When the input standard is set there is no use to ask the chip for standard, therefore it is cached in the driver. Signed-off-by:
-
Richard Röjfors authored
This patch adds support to the ADV7180 driver to check the input status. Since the status is held in the same register as the input standard a small restructuring of the code is done to reuse the code for reading the register Signed-off-by:
-
Danny Wood authored
V4L/DVB (13168): Add support for Asus Europa Hybrid DVB-T card (SAA7134 SubVendor ID: 0x1043 Device ID: 0x4847) Adds the device IDs and driver linking to allow the Asus Europa DVB-T card to operate with these drivers. The device has a SAA7134 chipset with a TD1316 Hybrid Tuner. All inputs work on the card including switching between DVB-T and Analogue TV, there is also no IR with this card. [mchehab@redhat.com: CodingStyle fixes] Signed-off-by:
Danny Wood <danwood76@gmail.com> Signed-off-by:
-
Julia Lawall authored
The definition of tvaudio_tuner_ops initializes the s_tuner field twice. It appears that the second case should initialize the g_tuner field. The semantic match that finds this problem is as follows: (http://coccinelle.lip6.fr/) // <smpl> @r@ identifier I, s, fld; position p0,p; expression E; @@ struct I s =@p0 { ... .fld@p = E, ...}; @s@ identifier I, s, r.fld; position r.p0,p; expression E; @@ struct I s =@p0 { ... .fld@p = E, ...}; @script:python@ p0 << r.p0; fld << r.fld; ps << s.p; pr << r.p; @@ if int(ps[0].line)!=int(pr[0].line) or int(ps[0].column)!=int(pr[0].column): cocci.print_main(fld,p0) // </smpl> Signed-off-by:
Julia Lawall <julia@diku.dk> Signed-off-by:
-
Tuukka Toivonen authored
When the image size (bytesperline*height) is not multiple of PAGE_SIZE, v4l2 rounded the required buffer size to be multiple of PAGE_SIZE. This prevented user space to store images directly into userptr buffers which were not multiple of PAGE_SIZE. This constraint is removed. The start address is still assumed to be required page-aligned, ie., when v4l2 allocates mmap buffers, the offset between different buffers is page-aligned. Signed-off-by:
Tuukka Toivonen <tuukka.o.toivonen@nokia.com> Signed-off-by:
-
Mauro Carvalho Chehab authored
DVBS2API introduced two new ioctls that allows changing/getting details about newer video standards like DVB-S2 and ISDB-T. There are lots of thing to do in this chapter, but, for now, just add a session there, in order to allow make to not complain about undefined symbol. With this patch, all ioctls at the frontend.h API were somehow documented. Signed-off-by: -
Mauro Carvalho Chehab authored
We'll need to add the documentation for FE_[GET|SET]_PROPERTY that is not specific to isdb-t. Instead of doing it at frontend.xml, it is better to have a separate session for it. Also, isdbt.xml assumes that those ioctls were already introduced, and documents several exisiting properties. So, better to just rename it and add the documentation for it, together with each existing property there. Signed-off-by: -
Mauro Carvalho Chehab authored
While this ioctl should not be used, and is a good canditate for Documentation/feature-removal-schedule.txt, while it is on kernel, it needs to be documented. Signed-off-by: -
Mauro Carvalho Chehab authored
Signed-off-by: -
Mauro Carvalho Chehab authored
Adds linux/dvb/frontend.h header file at DocBook, and creates cross-references for the frontend ioctls. Signed-off-by: -
Mauro Carvalho Chehab authored
Rename all sessions id's for frontend ioctls to match ioctl name and add the proper cross-link references for those ioctls. This also helps to identify what ioctls are missing specs. Signed-off-by: -
Mauro Carvalho Chehab authored
Signed-off-by: -
Laurent Pinchart authored
The default streaming control timeout was found by Ondrej Zary to be too low for some Logitech webcams. With kernel 2.6.22 and newer they would timeout during initialization unles the audio function was initialized before the video function. Add a module parameter to set the streaming control timeout and increase the default value from 1000ms to 3000ms to fix the above problem. Thanks to Ondrej Zary for investigating the issue and providing an initial patch. Signed-off-by:
Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by:
-
Laurent Pinchart authored
At least one 5986:0241 webcam model includes vendor-specific descriptors at the end of its streaming interface descriptors. Print an information UVC_TRACE_DESCR message and try to continue parsing the descriptors rather than bailing out with an error. Signed-off-by:
-
Ming Lei authored
This patch sets stream->urb_size as psize*npackets before calling uvc_alloc_urb_buffers, which may fix a possible failure of usb_buffer_free in case usb_buffer_alloc returns NULL. The patch is based on the ideas below: 1,If usb_buffer_alloc can't allocate a buffer sucessfully, uvc_free_urb_buffers will be called to free the allocated buffers, and stream->urb_size is required to be passed to usb_buffer_free; 2,uvc_free_urb_buffers can reset stream->urb_size. This patch is against linux-v2.6.31-next-20090926. Signed-off-by:
Ming Lei <tom.leiming@gmail.com> Acked-by:
-
Laurent Pinchart authored
The uvcvideo driver has a driver-wide lock and a reference count to protect against a disconnect/open race. Now that videodev handles the race itself, reference-counting in the driver can be removed. Signed-off-by:
-
Laurent Pinchart authored
Dismiss privileges on the file handle when calling VIDIOC_REQBUFS with a buffer count of 0. This allows applications to release the streaming permissions on the file handle without closing it. Signed-off-by:
-
Laurent Pinchart authored
The VIDIOC_TRY_FMT ioctl doesn't interfere with video streaming and thus should succeed on an unprivileged file handle. Signed-off-by:
-
Laurent Pinchart authored
Use the trace level for messages related to video URB buffers allocation and alternate setting selection. Signed-off-by:
-
Laurent Pinchart authored
Return minimum, maximum and step set to 0 without querying the hardware. Signed-off-by:
-
Hans de Goede authored
There were various "if (sensor == foo)" pieces of code inside ov519.c, move these bits to the sensor specific init functions, where they belong. Signed-off-by:
-
Hans de Goede authored
gscpa_ov519: add support for the ovfx2 bridge Signed-off-by:
-
Hans de Goede authored
This patch adds support for the OV2610 and OV3610 sensors to the gspca ov519 subdriver. This is a preparation patch for adding ovfx2 bridge support. Signed-off-by:
-
Hans de Goede authored
gspca_ov519 cleanup sensor detection Signed-off-by:
-
Hans de Goede authored
The CIF sensor type 1 exposure setting got clamped at 300, as settings below 300 do not work well (and do work for the other sensors). This patch scales the 0-4095 range to 300-4095 instead of ignoring changes between 300, avoiding have a part of the control range where nothing happens. This is esp. important for software autogain as done by libv4l. Signed-off-by:
-
Hans de Goede authored
While checking all gspca sub drivers pkt_scan functions for a bug I found in 1 of them (and after checking also in another), I noticed a bug in the gspca_sq905c work queue function, once it has decided to start discarding a frame because the application is not reading fast enough (and thus returning buffers to fill fast enough), it never stops discarding. This patch fixes this by simply completely removing the "discarding" variable, if we need to discard the current frame because there is no buffer to store it, the "frame" pointer will be NULL, so that is all we need to check. Signed-off-by:
-
Hans de Goede authored
While checking all gspca sub drivers pkt_scan functions for a bug I found in 1 of them (and after checking also in another), I noticed a bug in the gspca_jeilinj work queue function, once it has decided to start discard a frame because the application is not reading fast enough (and thus returning buffers to fill fast enough), it never stops discarding. This patch fixes this by simply completely removing the "discarding" variable, if we need to discard the current frame because there is no buffer to store it, the "frame" pointer will be NULL, so that is all we need to check. I've also moved the gspca_get_i_frame() call and the writing of the jpg header to the buffer to after the first usb_bulk_msg() call, as we don't need it before that, and that will give the app slightly more time to queue a buffer for us to fill. Signed-off-by:
-
Theodore Kilgore authored
Improve (and also simplify :) gspca_mr97310a: VGA sensor type detection. As it was still failing on some machines (not with some cams, but on some machines interesting enough). Signed-off-by:
Theodore Kilgore <kilgota@banach.math.auburn.edu> Signed-off-by:
-
Hans de Goede authored
-Remove use of unneeded discarding variable -Cleanup locking to only take usb_lock around access to the control endpoint, by no longer taking the lock around the bulk transfer (which takes most of the time) we can remove the msleep(1) which was needed to give the gspca core a chance to grab the usb_lock to signal us to stop. Signed-off-by:
-
Theodore Kilgore authored
This patch adds controls for vga cams with sensor type 0, in order to correctly report the present controls, the probing of the sensor type has been moved from sd_start to sd_config, since this made the sensor type probing unreliable the detection method was changed. Note this requires the camera to enter streaming mode, so sd_config now briefly makes the camera stream. Signed-off-by:
-
