linux-apus-devel

build problems of linux-apus cvs

[<ma...@e-...>]

I grabbed latest version today from cvs with following commands:

export CVSROOT=3D":pserver:ano...@cv...:/cvsr=
oot/linux-apus"
cvs -z3 co 2.4

Then configured it with "make menuconfig", and compiled it with "make dep=
"
and "make".

First problem is in the mem.c. At the end of this file, there is a call t=
o =

lp_m68k_init(), but "grep -r lp_m68k_init *" shows that this function isn=
't
defined anywhere!? I commented it for now to avoid linking problem.

Compiled kernel didn't work. My system freezed after bootstrap loaded
bootstrap_pup and inhibited drives. I suspected a configuration problem a=
nd
compared .config file of 2.4.3 release on the sf.net page with my .config=

(attached to my mail). Differences are minimal, I'll try again with "make=

oldconfig" later though. I don't want to do this since config of 2.4.3 is=

bloated (heck, it even has khttpd) but maybe it will solve the problem.

I wonder how the 2.4.3 binary is compiled. Am I doing something wrong whi=
le
downloading kernel and get older version or something? Or am i doing a
mistake while configuring it?


-- =

Gurer Ozen (email: ma...@e-..., jid: ma...@ja...)

Re: build problems of linux-apus cvs

[Glenn H. <gh...@c2...>]

GürerÖzen wrote:
> 
> I grabbed latest version today from cvs with following commands:
> 
> export CVSROOT=":pserver:ano...@cv...:/cvsroot/linux-apus"
> cvs -z3 co 2.4
> 
> Then configured it with "make menuconfig", and compiled it with "make dep"
> and "make".
> 
> First problem is in the mem.c. At the end of this file, there is a call to
> lp_m68k_init(), but "grep -r lp_m68k_init *" shows that this function isn't
> defined anywhere!? I commented it for now to avoid linking problem.

disable CONFIG_M68K_PRINTER (or something like that) in .config, and try
again.

Shouldn't the m68k printer option be removed from apus ?


-glenn

Re: build problems of linux-apus cvs

[Michel <mic...@ii...>]

Glenn Hisdal wrote:
> 
> GürerÖzen wrote:
> >
> > I grabbed latest version today from cvs with following commands:
> >
> > export CVSROOT=":pserver:ano...@cv...:/cvsroot/linux-apus"
> > cvs -z3 co 2.4
> >
> > Then configured it with "make menuconfig", and compiled it with "make dep"
> > and "make".
> >
> > First problem is in the mem.c. At the end of this file, there is a call to
> > lp_m68k_init(), but "grep -r lp_m68k_init *" shows that this function
> > isn't defined anywhere!? I commented it for now to avoid linking problem.
> 
> disable CONFIG_M68K_PRINTER (or something like that) in .config, and try
> again.
> 
> Shouldn't the m68k printer option be removed from apus ?

I build it as a module for our binaries, why should I remove it?


-- 
Earthling Michel Dänzer (MrCooper)    \   Debian GNU/Linux (powerpc) developer
CS student, Free Software enthusiast   \        XFree86 and DRI project member

Re: build problems of linux-apus cvs

[Glenn H. <gh...@c2...>]

Michel Dänzer wrote:
> 
> Glenn Hisdal wrote:
> >
> > GürerÖzen wrote:
> > >
> > > I grabbed latest version today from cvs with following commands:
> > >
> > > export CVSROOT=":pserver:ano...@cv...:/cvsroot/linux-apus"
> > > cvs -z3 co 2.4
> > >
> > > Then configured it with "make menuconfig", and compiled it with "make dep"
> > > and "make".
> > >
> > > First problem is in the mem.c. At the end of this file, there is a call to
> > > lp_m68k_init(), but "grep -r lp_m68k_init *" shows that this function
> > > isn't defined anywhere!? I commented it for now to avoid linking problem.
> >
> > disable CONFIG_M68K_PRINTER (or something like that) in .config, and try
> > again.
> >
> > Shouldn't the m68k printer option be removed from apus ?
> 
> I build it as a module for our binaries, why should I remove it?

Oh, sorry then...
iirc. it don't work if you try to compile it directly into the kernel.
The function lp_m68k_init is in the file lp_m68k.c in the 2.2 tree, but
doesn't exist in 2.4.
It's been a while since I tried it, so maybe something has changed.

I know this option had to be enabled in 2.2 to make printing work, but
iirc. it doesn't work in 2.4. 

- glenn

Re: build problems of linux-apus cvs

[Michel <mic...@ii...>]

Glenn Hisdal wrote:

> > > Shouldn't the m68k printer option be removed from apus ?
> >
> > I build it as a module for our binaries, why should I remove it?
>=20
> iirc. it don't work if you try to compile it directly into the kernel.
> The function lp_m68k_init is in the file lp_m68k.c in the 2.2 tree, but
> doesn't exist in 2.4.
> It's been a while since I tried it, so maybe something has changed.
>=20
> I know this option had to be enabled in 2.2 to make printing work, but
> iirc. it doesn't work in 2.4.

If that's true then we should either fix or remove it of course.


--=20
Earthling Michel D=E4nzer (MrCooper)    \   Debian GNU/Linux (powerpc) de=
veloper
CS student, Free Software enthusiast   \        XFree86 and DRI project m=
ember

Re: build problems of linux-apus cvs

[Roman Z. <zi...@li...>]

Hi,

"GürerÖzen" wrote:

> First problem is in the mem.c. At the end of this file, there is a call to
> lp_m68k_init(), but "grep -r lp_m68k_init *" shows that this function isn't
> defined anywhere!? I commented it for now to avoid linking problem.

It's obsolete, it's replaced with the parallel drivers under
drivers/parport.

> I wonder how the 2.4.3 binary is compiled. Am I doing something wrong while
> downloading kernel and get older version or something? Or am i doing a
> mistake while configuring it?

I guess it's CONFIG_RTC, in the precompiled kernel this is a module, so
it does no harm.

bye, Roman

Re: build problems of linux-apus cvs

[<ma...@e-...>]

RZ>> First problem is in the mem.c. At the end of this file, there is a call
RZ>> to lp_m68k_init(), but "grep -r lp_m68k_init *" shows that this

RZ> It's obsolete, it's replaced with the parallel drivers under
RZ> drivers/parport.

ok.

RZ> I guess it's CONFIG_RTC, in the precompiled kernel this is a module, so
RZ> it does no harm.

It was the generic pm2 gfx card option, don't know why i enabled it.

I disabled CONFIG_RTC. Now I get "cannot read/set hardware clock"
messages each time I bootstrap or halt linux.

Other than that 2.4.3 is working very well on my system. Sound started
to work again and printer seems like working ok (haven't tested it much
though).


-- 
Gurer Ozen (email: ma...@e-..., jid: ma...@ja...)

Re: build problems of linux-apus cvs

[Roman Z. <zi...@li...>]

"GürerÖzen" wrote:

> RZ> I guess it's CONFIG_RTC, in the precompiled kernel this is a module, so
> RZ> it does no harm.
> 
> It was the generic pm2 gfx card option, don't know why i enabled it.

Hmm, it's not needed, but it also shouldn't hurt.

> I disabled CONFIG_RTC. Now I get "cannot read/set hardware clock"
> messages each time I bootstrap or halt linux.

CONFIG_GEN_RTC is still enabled? That's actually the driver that takes
care of this.

> Other than that 2.4.3 is working very well on my system. Sound started
> to work again and printer seems like working ok (haven't tested it much
> though).

That would be strange, I haven't seen any important change here, but if
it works... :)

bye, Roman

Re: build problems of linux-apus cvs

[Michel <mic...@ii...>]

Roman Zippel wrote:
> 
> "GürerÖzen" wrote:
> 
> > RZ> I guess it's CONFIG_RTC, in the precompiled kernel this is a module,
> > so RZ> it does no harm.
> >
> > It was the generic pm2 gfx card option, don't know why i enabled it.
> 
> Hmm, it's not needed, but it also shouldn't hurt.

But it does currently. For some reason the pm2fb generic PCI code isn't happy
with our PCI implementation...


-- 
Earthling Michel Dänzer (MrCooper)    \   Debian GNU/Linux (powerpc) developer
CS student, Free Software enthusiast   \        XFree86 and DRI project member