kernel fails to compile?

Message

kinkos · Post by **kinkos** » Mon Apr 04, 2005 11:45 pm

Hey all, i have an HP 9000 K220 and i'm *trying* to get this thing running. CUrrently, the only liveCD i was able to get to boot with reiserfs doesn not have SMP, so everything takes ages to compile (instead of using all four precessors) anywho, i tried a stage 1 and switched to stage 3 when that failed (wiped clean and fresh stage 3) and i'm trying to compile the kernel for SMP before i install anything else. here's the error i get:

Code: Select all

  CC      fs/vfat/namei.o
  LD      fs/vfat/vfat.o
  LD      fs/vfat/built-in.o
  LD      fs/built-in.o
  CC [M]  fs/binfmt_som.o
fs/binfmt_som.c: In function `create_som_tables':
fs/binfmt_som.c:94: warning: cast from pointer to integer of different size
fs/binfmt_som.c:94: warning: cast from pointer to integer of different size
fs/binfmt_som.c:100: warning: cast from pointer to integer of different size
fs/binfmt_som.c:100: warning: cast from pointer to integer of different size
fs/binfmt_som.c:220:2: #error "Fix security hole before enabling me"
fs/binfmt_som.c: In function `load_som_binary':
fs/binfmt_som.c:257: error: `STACK_TOP' undeclared (first use in this function)
fs/binfmt_som.c:257: error: (Each undeclared identifier is reported only once
fs/binfmt_som.c:257: error: for each function it appears in.)
make[1]: *** [fs/binfmt_som.o] Error 1
make: *** [fs] Error 2

I checked EVERY option under file systems in menuconfig and no longer have anything compiled in as modules, so i'm a bit stumped. Also, shouldn't the kernel compile cleanly? here's the pertinent system info:

Code: Select all

livecd linux # USE="-doc" emerge -pv hppa-sources

These are the packages that I would merge, in order:

Calculating dependencies ...done!
[ebuild   R   ] sys-kernel/hppa-sources-2.6.11_p2  -build -doc -symlink 0 kB 

Total size of downloads: 0 kB

i thought doc would be a good use flag, but libxml refuses to compile (which is needed for doc's in the kernel) so i turned it off for this.

Code: Select all

livecd linux # emerge info
Portage 2.0.51-r2 (default-linux/hppa/2004.3, gcc-3.3.5, glibc-2.3.4.20040808-r1, 2.4.26-pa7-livecd32 parisc)
=================================================================
System uname: 2.4.26-pa7-livecd32 parisc PA7200 (PCX-T')
Gentoo Base System version 1.4.16
Autoconf: sys-devel/autoconf-2.59-r5
Automake: sys-devel/automake-1.8.5-r1
Binutils: sys-devel/binutils-2.14.90.0.7-r4
Headers:  sys-kernel/linux-headers-2.6.8.1-r2
Libtools: sys-devel/libtool-1.5.2-r5
ACCEPT_KEYWORDS="hppa"
AUTOCLEAN="yes"
CFLAGS="-O3 -pipe -march=1.1"
CHOST="hppa1.1-unknown-linux-gnu"
COMPILER=""
CONFIG_PROTECT="/etc /usr/kde/2/share/config /usr/kde/3/share/config /usr/share/config /var/qmail/control"
CONFIG_PROTECT_MASK="/etc/gconf /etc/terminfo /etc/env.d"
CXXFLAGS="-O1 -pipe"
DISTDIR="/usr/portage/distfiles"
FEATURES="autoaddcvs autoconfig ccache distlocks sandbox sfperms strict"
GENTOO_MIRRORS="http://mirror.datapipe.net/gentoo http://gentoo.osuosl.org http://www.ibiblio.org/pub/Linux/distributions/gentoo"
MAKEOPTS="-j5"
PKGDIR="/usr/portage/packages"
PORTAGE_TMPDIR="/var/tmp"
PORTDIR="/usr/portage"
PORTDIR_OVERLAY=""
SYNC="rsync://rsync.gentoo.org/gentoo-portage"
USE="apache2 berkdb bitmap-fonts crypt doc foomaticdb fortran ftp gdbm gif gpm imlib ipv6 jpeg libwww mysql ncurses nls ntpl ntplonly offensive perl php pic png python readline samba session spell ssl tcpd truetype-fonts type1-fonts xml2 zlib"

Code: Select all

livecd linux # gcc -v
Reading specs from /usr/lib/gcc-lib/hppa1.1-unknown-linux-gnu/3.3.5/specs
Configured with: /var/tmp/portage/gcc-3.3.5-r1/work/gcc-3.3.5/configure --enable-version-specific-runtime-libs --prefix=/usr --bindir=/usr/hppa1.1-unknown-linux-gnu/gcc-bin/3.3.5 --includedir=/usr/lib/gcc-lib/hppa1.1-unknown-linux-gnu/3.3.5/include --datadir=/usr/share/gcc-data/hppa1.1-unknown-linux-gnu/3.3.5 --mandir=/usr/share/gcc-data/hppa1.1-unknown-linux-gnu/3.3.5/man --infodir=/usr/share/gcc-data/hppa1.1-unknown-linux-gnu/3.3.5/info --with-gxx-include-dir=/usr/lib/gcc-lib/hppa1.1-unknown-linux-gnu/3.3.5/include/g++-v3 --host=hppa1.1-unknown-linux-gnu --disable-altivec --enable-nls --without-included-gettext --with-system-zlib --disable-checking --disable-werror --disable-libunwind-exceptions --disable-multilib --disable-libgcj --enable-languages=c,c++,f77 --enable-shared --enable-threads=posix --enable-__cxa_atexit --enable-clocale=gnu
Thread model: posix
gcc version 3.3.5  (Gentoo Linux 3.3.5-r1, pie-8.7.7.1)

any ideas?

cartpullerjack · Post by **cartpullerjack** » Tue Apr 05, 2005 12:54 am

Test your RAM?

kinkos · Post by **kinkos** » Tue Apr 05, 2005 12:58 am

Good idea. However, i got this machine about a week ago and haven't the slightest idea on how to go about doing that... i know there's memtest86, but thats obviously not gonna work on my hppa machine... suggestions?

kinkos · Post by **kinkos** » Tue Apr 05, 2005 5:58 am

I'm starting to think maybe some of the kernel isn't so hot... Also keep in mind i'm using an older compiler (3.3.5) which may not like the new kernel's code (2.6.11_p2). I found that under Executable File Formats, SOM is HP-UX compatibility for executables, removing that module obviously got rid of the error, but i hit some scsi problems. I'll just keep reconfiguring until i get all my hardware working, and remove all the drivers that wont compile.

HPRichard · Post by **HPRichard** » Tue Apr 05, 2005 10:36 pm

gcc 3.3.5 should not be an problem with hppa-sources-2.6.11_p2, they compiled fine on my HP 9000 735, as you can read here http://forums.gentoo.org/viewtopic-t-289278.html (with the exception of the HVD SCSI controller, which is not supported at least on 735 and 755, no idea if it works on K class machines). A good hint for compiling kernels is always to start with a minimal set of functionality, especially on strange systems. My suggestion is to try a kernel with only the absolutely neede options for running the system (and perhaps SMP for faster compiling, but your machine is twice as fast as mine according to SPECInt95/SPECfp95).

If you want to test the RAM, the sysadmins for HP machines on my university told me that there exists a RAM testing program from HP which runs under HP-UX, so if you did not throw away the HP-UX installation on that machine, you may be in luck or can obtain that tool from HP. I have absolutely no idea how this tool is called, nor do I have any other information related to this tool.

kinkos · Post by **kinkos** » Tue Apr 05, 2005 11:30 pm

Nah, i already trashed HP-UX, but i have the installation media if armageddon arrives and i decide to go closed source. Thanks for the tip though

I finally managed to get everything compiled and the kernel booting (although the console didnt work at first, i needed DPC under character devices i believe). New problem is the SMP. The bootable cd was a 2.4 kernel, now i've got a 2.6 kernel with SMP and all that jazz, but when it boots up, it gives off some error like Processor 3 is stuck, (when doing some stuff with the processors) then a couple more lines down it starts spitting out SMP function timed out errors, with a retry counter. I'd paste output, but haven't the slightest idea of how to capture it. So i figured Processor 3 might be bad, used the "bios" to disable the third one, now it complains about processor 1 being stuck. To be honest, i only know that one processor works, since only one was in there when it was donated to me. I bought two others off ebay, and the last one came from the same source as the working HP 9000. One of the guys down the hall said he smelled burning silicon, but i havent smelled it and didnt notice anything that suggested burning, but there was an odor (i figured it was the janitors or something). the smell could or could not be related, and while it was compiling the air coming off was warm, but not hot, so i *think* it's being cooled adaquately. Sorry for the rambling, but i don't suppose you have any experience with any of this? Right now i have it turned off, and if i feel adventurous, i may try to re-seat all the processors later.

One thing to note, according to the bios all four processors are idle and in good working order. *shrugs*

~kinkos

HPRichard · Post by **HPRichard** » Thu Apr 07, 2005 2:31 pm

The SMP problem may have three causes:
- faulty CPU(s)
- faulty motherboard
- faulty kernel
but I cannot decide which is the case here. There are some people havin Linux running on a K220, you can find them here http://hwdb.parisc-linux.org/view.php?t ... &name=K220, just email them, they should habe experiences with that machines. Another idea are the mailing lists at http://www.parisc-linux.org/mailing-lists/index.html, these guys should know an answer to nearly any question regarding Linux on PA-RISC machines, this is the mailing list for the development of the PA-RISC Linux kernel...

If you want to re-seat the CPUs, then my suggestion is (not because I am an expert in K-class machines, but out of common sense) that at first you test one CPU at a time, and only in the socket where the first CPU worked from the beginning. This way you can find or exclude faulty CPUs. Then make your way through the 2CPU possibilities ect...