- 04 Aug, 2012 40 commits
-
-
Antti Palosaari authored
Merge files dvb_usb_init.c and dvb_usb_dvb.c to dvb_usb_core.c. It is still under 1000 lines of code after all the optimization so put it one file. Signed-off-by:
Antti Palosaari <crope@iki.fi> Signed-off-by:
Mauro Carvalho Chehab <mchehab@redhat.com>
-
Antti Palosaari authored
Signed-off-by:
-
Antti Palosaari authored
We can live easily without it so remove it, make code and binary few bytes smaller. >From struct dvb_usb_device variable int num_adapters_initialized. Signed-off-by:
-
Antti Palosaari authored
Merge get_firmware_name() to identify_state(). It is wise to resolve firmware name in that routine as it does decision wether or not to load firmware at all. It is one very rarely needed callback less. Signed-off-by:
-
Antti Palosaari authored
Piggypag TS type callback to USB stream callback and change callback name slightly to fit better. Both of those are rather rare callback and has a relation. Transport Stream, TS, is input stream and USB stream is output stream of DVB USB bridge. Signed-off-by:
-
Antti Palosaari authored
We dont need that pointer as we can use container_of() macro thus remove it. Signed-off-by:
-
Antti Palosaari authored
Signed-off-by:
-
Antti Palosaari authored
Signed-off-by:
-
Antti Palosaari authored
Signed-off-by:
-
Antti Palosaari authored
Signed-off-by:
-
Antti Palosaari authored
We can live without it easily. Also USB data stream complete callback checks for feedcount seems to be quite redundant so remove those also. Signed-off-by:
-
Antti Palosaari authored
Signed-off-by:
-
Antti Palosaari authored
There was synchronization problem when streaming was stopped. Frontend was ran down before stream which causes problems. Use mutex to synchronization. Now it looks like that: LOCK start frontend UNLOCK LOCK start streaming [...] stop streaming UNLOCK LOCK stop frontend UNLOCK Signed-off-by:
-
Antti Palosaari authored
Signed-off-by:
-
Antti Palosaari authored
Signed-off-by:
-
Antti Palosaari authored
Signed-off-by:
-
Antti Palosaari authored
Initial / default USB stream configuration is now set always as a dvb_usb_device_properties. Dynamic configuration is done top of that if callback .get_usb_stream_config() exists. Default values are set when callback is called so callback needs only change values that are different than default. In addition of that define two macros DVB_USB_STREAM_BULK() and DVB_USB_STREAM_ISOC() for filling struct usb_data_stream_properties Signed-off-by:
-
Antti Palosaari authored
Signed-off-by:
-
Antti Palosaari authored
Give DVB adapter as a parameter for callback. MAC address is adapter property and it could be (at least in theory) there is devices having multiple adapters and all has own MAC. Also add error handling. If callback fails error is returned which leads device unregister. Signed-off-by:
-
Antti Palosaari authored
We can live easily without that variable. Signed-off-by:
-
Antti Palosaari authored
Signed-off-by:
-
Antti Palosaari authored
Remove state from struct dvb_usb_device as we can live without. Add some debug logs. Signed-off-by:
-
Antti Palosaari authored
Signed-off-by:
-
Antti Palosaari authored
It did not removed adapter on error case as .num_adapters_initialized was set after success adapter initialization. It should be safe to try remove all possible adapters in any-case. Signed-off-by:
-
Antti Palosaari authored
USB core will call disconnect and remove driver for us as device will disconnect itself. After that it is loaded again as a new device but it is warm and no firmware loading needed. Reported-by:
Malcolm Priestley <tvboxspy@gmail.com> Signed-off-by:
-
Malcolm Priestley authored
The first being this patch, no return value from dvb_usb_download_firmware causes system wide dead lock with COLD disconnect as system attempts to continue to warm state. Signed-off-by:
-
Antti Palosaari authored
Signed-off-by:
-
Antti Palosaari authored
It is more than nice to have all those frontend pointers when attaching tuner to the frontend in question. Make that possible attaching tuners after dvb_register_frontend() Signed-off-by:
-
Antti Palosaari authored
Small changes like log writings. No functionality changes. Signed-off-by:
-
Antti Palosaari authored
Signed-off-by:
-
Antti Palosaari authored
It is only three functions so move it to the main. Signed-off-by:
-
Antti Palosaari authored
Signed-off-by:
-
Antti Palosaari authored
Signed-off-by:
-
Antti Palosaari authored
Do not check active frontend as it could be already shutdown. Add some more debugs. It turns out that device is put sleep earlier than streaming is stopped in some cases. That is because streaming is running different task and there is no locking. Maybe some locking could be good idea to force stop streaming before device is shut down. I can guess there could be problems in someday cases like chip is sleep but streaming control is requested after that... Signed-off-by:
-
Antti Palosaari authored
Change logging. Some minor text and error number changes. Signed-off-by:
-
Antti Palosaari authored
Signed-off-by:
-
Antti Palosaari authored
Use pointers to properties as device and adapter properties are constant. No need to embed and copy those structures. Signed-off-by:
-
Antti Palosaari authored
Signed-off-by:
-
Antti Palosaari authored
Let optimize that code to the text segment. Signed-off-by:
-
Antti Palosaari authored
Move frontend_attach, tuner_attach, frontend_ctrl and streaming_ctrl from adapter to device. Signed-off-by:
-
