[linux-dvb] Mixing kernel and v4l-dvb sources
chris at malarky.org
Thu Aug 3 01:46:42 CEST 2006
On Wed, 2006-08-02 at 21:59 +0100, Andrew de Quincey wrote:
> On Wednesday 02 August 2006 17:43, Chris Malarky wrote:
> > Hi,
> > On Tue, 2006-08-01 at 22:17 +0100, Andrew de Quincey wrote:
> > > What I usually do is:
> > >
> > > 1) Install the kernel sources
> > > 2) hg update in v4l-dvb
> > > 3) in v4l-dvb, do make kernel-links KDIR=/path/to/kernel
> > >
> > > I too get the patch faillure at this stage, but I've just been ignoring
> > > it :)
> > >
> > > This'll hook the DVB drivers into the kernel, replacing the DVB stuff
> > > there already.. all you need to do now is cd to the kernel directory, and
> > > do a normal kernel configure/compilation. The HG drivers will have been
> > > patched into the kernel so you should see 'em pop up in there.
> > OK, so I've successfully compiled both 126.96.36.199 and 188.8.131.52 using
> > latest v4l-dvb sources. Runs fine with older cards (Nova-S and
> > Nova-SE), but I'm having problems with my TT S-1401.
> > The card is recognised when the budget module loads up, but I'm getting
> > a 'No tda826x found!' from frontend_init().
> > A snippet of dmesg:
> > [17179578.432000] saa7146: register extension 'budget_ci dvb'.
> > [17179578.512000] saa7146: register extension 'budget dvb'.
> > [17179578.532000] ACPI: PCI Interrupt Link [LNKB] enabled at IRQ 5
> > [17179578.536000] PCI: setting IRQ 5 as level-triggered
> > [17179578.536000] ACPI: PCI Interrupt 0000:00:14.0[A] -> Link [LNKB] ->
> > GSI 5 (level, low) -> IRQ 5
> > [17179578.544000] saa7146: found saa7146 @ mem ceaa6000 (revision 1, irq
> > 5) (0x13c2,0x1018).
> > [17179578.548000] saa7146 (0): dma buffer size 192512
> > [17179578.556000] DVB: registering new adapter (TT-Budget-S-1401 PCI).
> > [17179578.596000] adapter has MAC addr = 00:d0:5c:09:2f:7a
> > [17179578.620000] frontend_init: No tda826x found!
> > [17179578.624000] DVB: registering frontend 0 (Philips TDA10086
> > DVB-S)...
> > [17179583.536000] ACPI: PCI Interrupt Link [LNKC] enabled at IRQ 10
> > [17179583.548000] PCI: setting IRQ 10 as level-triggered
> > [17179583.548000] ACPI: PCI Interrupt 0000:00:11.5[C] -> Link [LNKC] ->
> > GSI 10 (level, low) -> IRQ 10
> > [17179583.560000] PCI: Setting latency timer of device 0000:00:11.5 to
> > 64
> > The frontend does get registered and I can run dvbtune against it but it
> > just sits there polling.
> > I tried re-loading the tda10086 and tda826x modules with debugging on
> > but didn't see anything extra.
> > I'm assuming I'm not missing something stupid like firmware (I didn't
> > see any mention of it anywhere in the code).
> > Do I perhaps need a later kernel version?
> Hmm, weird, if the tda10086 works, then so should the other chip... and no
> firmware is needed either you're right.
> Hmm, in budget.c, try varying the i2c address:
> if (dvb_attach(tda826x_attach, budget->dvb_frontend,
> 0x60, &budget->i2c_adap, 0) == NULL)
> the tuner was 0x60 on both of my cards - but it could be 61,62, or 63.
> also, maybe change the msleeps before it to 10msecs...
No joy. I tried 0x60 through to 0x63, with sleeps of both 1 and 10ms.
I modified the printk() statement each time to make sure I was
definitely picking up the re-compiled budget.ko.
As you say, it's weird that tda10086 is working but not tda826x. If the
card and demodulator weren't being recognised then I'd have (more) cause
to doubt my build of the modules.
The only thing I can suggest right now is that we swap some cards. If I
send you one of ours you can verify your code against it, and likewise
if you send me back one of yours I can do the same with my build. I
have a sneaking suspicion that our original two may have come direct
from Germany. Can you give me the hardware revision from your boards?
I'll check mine in the morning and let you know.
chris at malarky dot org
More information about the linux-dvb