<!-- Some styling for better description lists --><style type='text/css'>dt { font-weight: bold;float: left;display:inline;margin-right: 1em} dd { display:block; margin-left: 2em}</style>

   jacky_: what's the proper way to handle input standard changing in tv decoder?
   hverkuil: <u>jacky_</u>: inside a driver or in an application?
   jacky_: Should driver stop streaming, or notify user?
   hverkuil: ah, only notify.
   jacky_: <u>hverkuil</u>: driver
   hverkuil: http://hverkuil.home.xs4all.nl/spec/media.html#vidioc-dqevent
   <br> http://hverkuil.home.xs4all.nl/spec/media.html#vidioc-subscribe-event
   <br> V4L2_EVENT_SOURCE_CHANGE
   jacky_: <u>hverkuil</u>: Should the driver reallocate dma buffer when standard chaning?
   hverkuil: no, just keep streaming. It's up to the application to stop, reallocate buffers and restart streaming.
   <br> I should say, 'keep streaming if possible'.
   <br> there may be hardware constraints that prevent further streaming, although all hardware I've seen can cope with such changes.
   <br> Obviously, once the standard changes the contents of frames may be bad.
   <br> BTW, are you talking about SDTV?
   jacky_: but, if input format change from NTSC to PAL, current buffer size is not enought
   hverkuil: The DMA is still setup for the old standard, so no buffer overflow will take place. Obviously, the contents of the buffers will be wrong.
   <br> It's up to the application to act on the source change event and stop streaming, detect the new standard and reallocate buffers, and start up again.
   <br> Note that NTSC to PAL or vice versa changes only happen during testing with a signal generator or something like that. Normal use (i.e. watching TV or videos) will never cause such changes.
   jacky_: <u>hverkuil</u>: so, the process will be: 1) input standard changed; 2) driver detected and notify application; 3) application stop streaming; 4) application ioctl s_std for reallocation dma buffer; 4) application restart streaming video; right?
   hverkuil: So I am not aware of any applications that act on that event, they all assume that you setup the standard once and leave it.
   <br> yes, that's the ideal case.
   <br> But nobody does that, since it's not something that ever happens during normal use.
   <br> None of the current set of SDTV drivers sends that event.
   wh_: i got a rtl2i32u+r820t DVB-T usb dongle , does't it have driver for 2.6.x version kernel ?
   jacky_: <u>hverkuil</u>: my DVD can changing output standard on the fly, :)
   ***: jacky_ has quit IRC (Quit: jacky_)
   <br> jacky_ has quit IRC (Quit: jacky_)
   hverkuil: <u>wh_</u>: no, it doesn't. You can try these instructions http://linuxtv.org/wiki/index.php/How_to_Obtain,_Build_and_Install_V4L-DVB_Device_Drivers to build the driver for such an old kernel.
   <br> But really, it's much better if you can upgrade if at all possible.
   wh_: hverkuil, thx, i will try it out
   hverkuil: <u>wh_</u>: make a copy of /lib/modules/&lt;version&gt;/kernel/drivers/media first, Just in case things are messed up afterwards.
   wh_: hverkuil, yes
   <br> hverkuil, i did the same things before as the link you give to me , but that also not for 2.6.x kernel .
   <br> hverkuil, it works for 3.5.x kernel on my laptop
   <br> but i want install it to my embed board which is 2.6.x kernel
   hverkuil: <u>wh_</u>: which 2.6 kernel?
   wh_: hverkuil, 2.6.34
   hverkuil: You need 2.6.36 at minimum. rtl2832 needs kernel support for i2c muxes which appeared in 2.6.36.
   Lewis: hi folks
   <br> is there a known fix for the current build failures (also showing up in the automated build) regarding smiapp-core.c ?
   hverkuil: <u>sailus</u>: can you make a patch for smiapp? It's your code :-)
   Lewis: haha, as an interim is there a workaround?
   hverkuil: You can hack include/linux/of.h and move the of_read_number and of_read_number inlines out of the #ifdef CONFIG_OF
   <br> that should work.
   Lewis: ok thanks, i'll take a look
   ***: wh_ has quit IRC (Quit: wh_)
   sailus: <u>hverkuil</u>: Did you receive my e-mail on the topic?
   <br> There's a patch for it which I think I'll need to send as a fix.
   Lewis: Would it be possible to get hold of that patch sailus?
   hverkuil: <u>sailus</u>: I did see the emails, but didn't really read them :-)
   <br> ah, OK, you posted a patch already. I'll test this tomorrow.
   <br> (compile-test, that is)