[linux-dvb] Re : Inclusion of STB0899 support in kernel

manu eallaud at yahoo.fr
Thu Aug 28 02:23:03 CEST 2008


Le 27.08.2008 19:07:26, Hans Werner a écrit :
> > Hi,
> > 
> > On Wed, Aug 27, 2008 at 10:19:06PM +0200, johan vdp wrote:
> > > 
> > > What happened to STB0899 support?
> > > Building a kernel, patching it (with some luck), building
> applications
> > with patches to match the driver, it is simply too much work.
> > > (It might be fun once, but starts to be annoying when automatic
> package
> > updates start to fail, the next day.)
> > > Having support in the kernel, will lead to applications that
> > 'automatically' start to support it.
> > >  
> > > Back in the days that 'multiproto' development was still alive 
> and
> > buzzing. The STB0899 cards looked like having the best support.
> > > I have followed these lists for some time and opted to buy a
> STB0899
> > based DVB-S2 tuner card because the outlook for support was good;
> THEN.
> > > NOW it must have been two years, and it is still not merged into
> the
> > kernel.
> > 
> > Same here. I am using multiproto for a long time now. Personally 
> for
> me
> > it works great, but it is really bad, that it is still not included
> in
> > the kernel.
> > 
> > Manu Abrahams hg tree is "out of date" and it is impossible to
> compile
> > it with a recent kernel without patching it.
> > 
> > Hopefully Igor M. Liplianin has created an hg tree, which merges 
> all
> the
> > different dvb-trees in one single tree
> > (http://liplianindvb.sourceforge.net/hg/liplianindvb).
> > This really helps in compiling multiproto, but this is not the
> solution.
> > 
> > What is required to get multiproto into the kernel?
> > Where are the problems?
> > 
> > Regards, Artem
> 
> I absolutely agree. Please see the messages I sent on this recently
> http://linuxtv.org/pipermail/linux-dvb/2008-August/028084.html
> 
> For the reasons mentioned above it is a terrible waste of effort to
> maintain
> code outside the kernel. Because multiproto (or equivalent code) has
> not
> been merged into the kernel (after 2 years and 4.5 months in
> development) there
> is a *lot* of out-of-kernel code supporting DVB-S2 cards, which of
> course are
> amongst the most popular, capable and important cards on the market.
> It is 
> not OK to expect everyone (including end users) to clone hg trees and
> update
> and patch them repeatedly. All driver code should be heading towards
> the kernel,
> preferably sooner than later. Anything else is pointless.
> 
> I think Linux DVB is in a state of unacknowledged crisis.  
> 
> The code works, but the kernel merging of multiproto or another way 
> of
> supporting
> DVB-S2 cards seems to be completely paralysed.  What is required to
> get it done?
> It must eventually be accomplished.
> 
> OT, even the simplest userspace code is sorely lacking. Simple yet
> important tools such
> as szap, and scan don't even have version numbers. Distros don't 
> agree
> on whether
> the package is called dvb-apps or dvb-utils and have had to create
> their own code
> versioning. How many szap patches are floating around?

May I add that the support of certain cards (especially TT 3200) has 
some issues and that this state of affairs just makes it even harder to 
fix (see my numerous posts about it).
I would love to hear from the maintainers about what needs to be done 
(and please forget the history so that we can probably avoid another 
long and sterile flamewar), things must go forward from now.
Hope that someone who can do something about this can explain to us how 
we can help.
Bye
Manu






More information about the linux-dvb mailing list