2003-09-17 18:29:01

by Jeffrey W. Baker

[permalink] [raw]
Subject: 2.4.23-pre4 compile failure in hw_random.c and aic7xxx on amd64

hw_random.c: In function `via_init':
hw_random.c:433: error: `MSR_VIA_RNG' undeclared (first use in this function)
hw_random.c:433: error: (Each undeclared identifier is reported only once
hw_random.c:433: error: for each function it appears in.)
hw_random.c: In function `via_cleanup':
hw_random.c:459: error: `MSR_VIA_RNG' undeclared (first use in this function)

The strange bit is that I didn't even have Intel/AMD/VIA hardware rng
configured, only AMD 768/8??? rng support. So it seems like some sort
of bug in oldconfig. After removing CONFIG_HW_RANDOM I can build again.

I also still have compile failures in drivers/scsi/aic7xxx due to
Werror.

-jwb


2003-09-17 18:36:05

by Jeffrey W. Baker

[permalink] [raw]
Subject: Re: 2.4.23-pre4 compile failure in hw_random.c and aic7xxx on amd64

On Wed, 2003-09-17 at 11:28, Jeffrey W. Baker wrote:
> hw_random.c: In function `via_init':
> hw_random.c:433: error: `MSR_VIA_RNG' undeclared (first use in this function)
> hw_random.c:433: error: (Each undeclared identifier is reported only once
> hw_random.c:433: error: for each function it appears in.)
> hw_random.c: In function `via_cleanup':
> hw_random.c:459: error: `MSR_VIA_RNG' undeclared (first use in this function)
>
> The strange bit is that I didn't even have Intel/AMD/VIA hardware rng
> configured, only AMD 768/8??? rng support. So it seems like some sort
> of bug in oldconfig. After removing CONFIG_HW_RANDOM I can build again.
>
> I also still have compile failures in drivers/scsi/aic7xxx due to
> Werror.

Follow-up:

fs/fs.o(.text+0x23ed7): In function `interrupts_open':
: undefined reference to `show_interrupts'

This appears to be defined on ppc64 and i386, but not x86_64. Possibly
#ifndef CONFIG_X86 confusion because x86_64 sets CONFIG_X86_64 and
CONFIG_X86. Not sure how to fix.

-jwb


2003-09-17 19:09:16

by Andi Kleen

[permalink] [raw]
Subject: Re: 2.4.23-pre4 compile failure in hw_random.c and aic7xxx on amd64

"Jeffrey W. Baker" <[email protected]> writes:

> On Wed, 2003-09-17 at 11:28, Jeffrey W. Baker wrote:
> > hw_random.c: In function `via_init':
> > hw_random.c:433: error: `MSR_VIA_RNG' undeclared (first use in this function)
> > hw_random.c:433: error: (Each undeclared identifier is reported only once
> > hw_random.c:433: error: for each function it appears in.)
> > hw_random.c: In function `via_cleanup':
> > hw_random.c:459: error: `MSR_VIA_RNG' undeclared (first use in this function)
> >
> > The strange bit is that I didn't even have Intel/AMD/VIA hardware rng
> > configured, only AMD 768/8??? rng support. So it seems like some sort
> > of bug in oldconfig. After removing CONFIG_HW_RANDOM I can build again.
> >
> > I also still have compile failures in drivers/scsi/aic7xxx due to
> > Werror.
>
> Follow-up:
>
> fs/fs.o(.text+0x23ed7): In function `interrupts_open':
> : undefined reference to `show_interrupts'
>
> This appears to be defined on ppc64 and i386, but not x86_64. Possibly
> #ifndef CONFIG_X86 confusion because x86_64 sets CONFIG_X86_64 and
> CONFIG_X86. Not sure how to fix.

2.4 has gotten to the old habit again of breaking other architectures
faster than they can be fixed ;-/ It's already fixed in the x86-64.org CVS.

I have a bigger patchkit that I plan to merge to Marcelo soon, but
didn't have time to clean it up for him yet.

For a very raw patch (still some uglities etc.) you can
use ftp://ftp.x86-64.org/pub/linux/v2.4/x86_64-2.4.23pre4-0.bz2
Also note that ext3 seems to be broken currently (at least since 2.4.22)

-andi