[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#229537: linux-kernel-headers: Building v4l?2? programs, general ioctl problem.



Thank you for your reply.  I guess the bug here is not that I can't build this one program.  It's
that ioctl information needs to be taken from the right header files.  As time goes on this
problem will get more and more important.  I wonder how vary old 2.2 or 2.1 kernels deal with the
outdatad ioctls?

I know that this is a Linus pet peve but the API was cahnge somewhere along the way.  Unless I'm
wrong with what my problem is as it could be a gcc or libc thing.  I don't know how to do furthe
testing I have read the source of the app and the bttv driver and I conclude that the apps data is
not being correctly unpacked by the driver.  Since I can't use all the native headers and get a
succsesfull build this would mean that there is something vary wrong with linux-kernel-headers.

mike

--- Daniel Jacobowitz <dan@debian.org> wrote:
> On Sat, Jan 24, 2004 at 09:35:17PM -0600, Mike Mestnik wrote:
> > Package: linux-kernel-headers
> > Version: 2.5.999-test7-bk-15
> > Severity: normal
> > 
> > I'm trying to build nvrec(a v4l2 program).  It builds fine but dose not get
> > the correct ioctl(headers).  I know the bigger issue is binary
> > compatability, however I don't think there are enuff ppl using v4l that it
> > will get fixed.  I have tryed copying header files from 2.6 with varable
> > results.  If it builds I get the same problem or I get struct time problems.
> > I have reported the struct time problem previously and it is fixed untill I
> > break it.
> > 
> > The meat.
> > paladin:/usr/src/real/nvrec/nvrec-20020812# divx4rec -o tmp.avi -F 20
> > Starting nvrec.c, version 20020812 (nvrec-20020812).
> > nvcore_parse_file(617): Error opening "/root/.nvrec.conf"
> > v4l1_core_init(323): Size: 384x288 from /dev/v4l/video0
> > v4l1_core_init(393): Set window: Invalid argument
> > main(65): nvrec: initialising capture device: Invalid argument
> > 
> > The 393 is the result of not using the right header during build.  ioctl
> > numbers and struct's may have changed.  I have not the time to find ought
> > exactly why the API has change.
> 
> If you can't tell us what changed, we can't fix it...  I will check the
> latest 2.6 kernel headers to see if this has been fixed since test7.
> 
> -- 
> Daniel Jacobowitz
> MontaVista Software                         Debian GNU/Linux Developer


__________________________________
Do you Yahoo!?
Yahoo! SiteBuilder - Free web site building tool. Try it!
http://webhosting.yahoo.com/ps/sb/



Reply to: