[vox-tech] Canonical collection of kernel compile error codes?

[Peter Jay Salzman]

richard,

sometimes a kernel compile bombs.  at some point, you get so much
experience, you know what to do in certain situations.  like unresolved
symbols tend to be errors in the depends system, so building that
particular subsystem directly into the kernel will often help.

however....

On Tue 29 Jun 04,  6:23 PM, Richard Burkhart <richard at khanfusion.net> said:
> Hi all;
> 
> I tried my first un-supervised kernel compile (at least my first one that
> did more than bomb out after 15 seconds of scrolling.)
> 
> I'm working with 2.4.26 kernel source -- patched up from 2.4.20, because I
> have to keep existing Win4Lin patches.  I ran each patch, and then ran 'make
> oldconfig' each time to apply the new changes and bring my config file
> forward into the next patch.

that's probably the problem.

> on an AMD K6-2, the compile probably takes all night).
> 
> The compile bombed out at the point listed below.
> 
> My questions:
>  a) Any suggestions, from the errors listed below, of what I did wrong?

i don't think you did anything wrong.  my guess is that the win4lin
patch needs to be updated.

>  b) Is there a site listing kernel compile error codes, with suggestions on
> how to fix things?  I've searched on google, trying to find the error, and
> turned up nothing.

these aren't kernel compile error codes.  they're gcc error messages.
you can get the same types of messages with programs you write yourself.
clearly there's a symbol called "priority" which should be available
during the compilation of vmscan.c but for some reason it isn't.

i _seriously_ doubt the problem is with anything you've done.

i would suggest building the kernel without the win4lin patches and see
if that compiles.  if so, apply the win4lin patch and see if it
compiles.

pete

> Thanks!
> 
> _____cut/paste_____
> gcc -D__KERNEL__ -I/home/src/linux/include -Wall -Wstrict-prototypes -Wno-tr
> igraphs -O2 -fno-strict-aliasing -fno-common -fomit-frame-pointer -pipe -mpr
> eferred-stack-boundary=2 -march=k6   -nostdinc -iwithprefix
> include -DKBUILD_BASENAME=swap  -c -o swap.o swap.c
> gcc -D__KERNEL__ -I/home/src/linux/include -Wall -Wstrict-prototypes -Wno-tr
> igraphs -O2 -fno-strict-aliasing -fno-common -fomit-frame-pointer -pipe -mpr
> eferred-stack-boundary=2 -march=k6   -nostdinc -iwithprefix
> include -DKBUILD_BASENAME=vmscan  -c -o vmscan.o vmscan.c
> vmscan.c: In function `swap_out':
> vmscan.c:339: error: `priority' undeclared (first use in this function)
> vmscan.c:339: error: (Each undeclared identifier is reported only once
> vmscan.c:339: error: for each function it appears in.)
> vmscan.c:339: error: `gfp_mask' undeclared (first use in this function)
> make[3]: *** [vmscan.o] Error 1
> make[3]: Leaving directory `/home/src/linux/mm'
> make[2]: *** [first_rule] Error 2
> make[2]: Leaving directory `/home/src/linux/mm'
> make[1]: *** [_dir_mm] Error 2
> make[1]: Leaving directory `/home/src/linux'
> make: *** [stamp-build] Error 2
> minerva:/usr/src# ls
> _____end cut/paste_____
> 
> 
> 
> _______________________________________________
> vox-tech mailing list
> vox-tech at lists.lugod.org
> http://lists.lugod.org/mailman/listinfo/vox-tech

-- 
In theory, theory and practise are the same.  In practise, they aren't.
GPG Instructions: http://www.dirac.org/linux/gpg
GPG Fingerprint: B9F1 6CF3 47C4 7CD8 D33E 70A9 A3B9 1945 67EA 951D