I'm attemoting to get the Voladj patch applied to the 1.03 kernel, and have ran into a snag. It appears that most of my problems are in empeg/kernel/include/asm-arm/arch-sa1100/empeg.h. In the recent patch, this is added

#define EMPEG_DSP_GRAB_OUTPUT _IOR(EMPEG_DSP_MAGIC, 2, int)

The Voladj also tries to add a define line here. I tried this:

#define EMPEG_DSP_VOLADJ _IOW(EMPEG_DSP_MAGIC, 3, int)

but it didn't seem to work. Has anyone gotten the Voladj into 1.03?

Huh? I must have missed a patch. I'll look into this tomorrow.

It ought to work with the ioctl as 3, but you would have to change any clients that hard code the ioctl number, if you are actually using the ioctl.

Richard.

Beyond your selector, was there anything else that referred to it by number specificially? I thought I searched but I might have missed something. (The problem with working on that type of thing between doing real work)

Beyond your selector, was there anything else that referred to it by number specificially?

Frank wrote a command-line Ioctl selector for Volad that's used in my "Picker and Voladj For Dummies" package. I don't totally understand how it works under the hood, perhaps that's related to what you're referring to?

___________
Tony Fabris

I did recompile Frank's code with no problems. (Once I realised I had the old broken code first anyhow... .

I don't believe any other part of the voladj patch needed to be changed for the one adjustment I made to the ioctl number. But for some reason when I applied all the code to where it needed to go and compiled the kernel, the code was just not working. I didn't save all the kernel compilation output, so I don't know if I simply missed a non-critical error that would point me to what is wrong.

For now, it's probably just best to wait for the final version of 1.03.

(I guess thats one benefit to not being an alpha tester. No constant need to apply all the patches you use on a regular basis to new kernel code. I still need to go in and move all my custom stuff off the player partition though, so I can one day apply a new upgrade file. I would have imagined I would have learned this back from the beta days on the MK1.)

It turns out I did miss some kernel patches. It would be nice if the guys would release a full snapshot of their kernel source again sometime. It's easy to miss one of the patches, when you have to go all the way from 11d. :)

Did you make clean after fixing the voladj patch? I guess weird things could happen if the empeg_audio3 code was using the wrong ioctl numbers??? I'm a bit paranoid about that kind of thing, I make clean lots. I haven't been using 1.03 very much (the Mk1 uses my old head unit for the tuner, and I don't have a tuner for the Mk2). I've just built a v1.03beta4 kernel, and hopefully I'll get time to try it with 1.03beta6 today.

Richard.

I'll try it again later. I'm pretty sure I had untarred my patched but never compiled kernel backup, but who knows. I was still working out some kinks. I will download all the new kernel patches that have come out and try it again. One day I might upgrade to the 1.03 beta

Wait, should voladj be patching empeg_audio2.c or empeg_audio3.c? That may be my problem...

I think that if there is an empeg_audio3.c, then the others are probably being ignored. I'd forgotten that I hadn't publicly published this patch yet. I'll send you a copy now, and put it with the other voladj stuff soon (a day or two). I haven't tested it on a Mk1 and I haven't actually used the different ioctl number, but at the very least the player functions, and the volume adjustment is working on my (tunerless) Mk2.

Richard.

I received it and will try it tomorrow. (I don't feel like trying to compile a kernel correctly this tired).

Thank you for all the work on this. I think I speak for many people here when I say thanks.