l4-hackers July 2000

l4-hackers@os.inf.tu-dresden.de

8 participants
8 discussions

Sigma0 bug?
by Christian Stueble 25 Jul '00

25 Jul '00

-----BEGIN PGP SIGNED MESSAGE----- Hello again, Last week I had some problems using fiasco on a 196 MB machine. Suzuki told me to modify l4/pkg/rmgr/server/src/memmap.h, but this does not help. (I set MEM_MAX to 256MB and the RMGR told me: "WARNING: All memory above 256 is wasted", on a 196MB machine!!). I noticed that sigma0 also defines MEM_MAX, and after setting it to 256MB, too, it works now. But, isn't it wrong that sigma0 returns memory pages up to RMGR's MEM_MAX value, even if only 196MB are available? Chris -- -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Christian Stueble......................stueble(a)cs.uni-sb.de PubKey[BF7104F5].......fp=8678C5D3CAD9CD8C F1DDB8EC202F116A To be or not to be is true... (apocrypha of George Boole) -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- -----BEGIN PGP SIGNATURE----- Version: PGPfreeware 5.0i for non-commercial use MessageID: wEaiE9zGKl756mb1+UDwYiIm1DkcNdUa iQCVAwUBOX2vsfyJjJy/cQT1AQGyxQQArkMBqWKaIOditHQIuoh+so+7mL0GK3i7 uPYdvQ5Ys6QDw12xhR1zhtPBA1KZcg4DE5CHzqiTY4RfKMx9X5wovIOVu1thujUv Z1OyvqjNTbCwinOoijq+yH53W7hz3aNA+ljbuJ8JghQE5lbFlHjJMkl5aFLwX93u knN6nHmSqPA= =e4UE -----END PGP SIGNATURE-----

1 0

L4 and grub difficulties
by James Riordan 20 Jul '00

20 Jul '00

I have been trying to launch L4 (Fiasco) and the hello server from a version of grub other than that supplied on the grub-ext2fs-floppy.gz image. When I take grub from the gnu server, it compiles but when I boot I get hung at: RMGR: loading (hd1,1)/fiasco/sigma0 RMGR: detected new-style DD-L4 RMGR: starting (hd1,1)/fiasco/main -nokdb proto=0x10128c Welcome to Fiasco! DD-L4/x86 microkernel (c) 1998 TU Dresden - Jul 4 2000 SIGMA0: Hello! RMGR: Hi there! RMGR running on L4/Pentium (hang) When I take grub from the CVS server at Dresen, I fail in compilation: gcc -DHAVE_CONFIG_H -I. -I. -I.. -I../stage1 -Wall -Wmissing-prototypes -Wunused -Os -fno-builtin -nostdinc -DFSYS_EXT2FS=1 -DFSYS_FAT=1 -DFSYS_FFS=1 -DFSYS_MINIX=1 -g -c -o pre_stage2_exec-asm.o asm.S /tmp/cc0xSXW3.s: Assembler messages: /tmp/cc0xSXW3.s:1189: Error: operands given don't match any known 386 instruction /tmp/cc0xSXW3.s:2109: Error: operands given don't match any known 386 instruction this is using the latest versions of gcc and binutils (to the best of my knowledge) montepulciano 502 # gcc -v Reading specs from /usr/local/lib/gcc-lib/i686-pc-linux-gnu/egcs-2.91.66/specs gcc version egcs-2.91.66 19990314 (egcs-1.1.2 release) montepulciano 503 # as --version GNU assembler 2.9.5 Might anybody have any advice on how to proceed? Thanks, james -- Dr. James Riordan email: rij(a)zurich.ibm.com IBM Forschungslaboratorium notes: James Riordan/Zurich/IBM Säumerstrasse 4 Tel +41 (0) 1 724 89 81 CH-8803 Rüschlikon, Switzerland Fax +41 (0) 1 724 89 53

4 9

L4-Linux bug
by Christian Stueble 20 Jul '00

20 Jul '00

-----BEGIN PGP SIGNED MESSAGE----- Hello, the file linux22/kernel/setup.c includes the "../arch/l4-i386/include/fake_interrupt.h" but I think it should include <l4linux/fake_interrupt.h>. Right? Chris -- -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Christian Stueble......................stueble(a)cs.uni-sb.de PubKey[BF7104F5].......fp=8678C5D3CAD9CD8C F1DDB8EC202F116A To be or not to be is true... (apocrypha of George Boole) -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- -----BEGIN PGP SIGNATURE----- Version: PGPfreeware 5.0i for non-commercial use MessageID: uFwTpm7e3g3EJ1l+XzVtymp/RbDe9Usn iQCVAwUBOXQZ2/yJjJy/cQT1AQHkZAQArm3523HYKSy3RbLiBFygHXyiC9PIDN8J O7jHy6edu+xotsW531deas8aWPKSakumU5UeHqGDi8FLNAsvvLj9k1nHivpq0F55 sfGmrJFaLDsWYneinB29qFJYnmWy1YkH2CymVFuzRnMPxN8IWEZwD1hkH9bQdjCo eDc/W9iNelI= =UdXf -----END PGP SIGNATURE-----

2 1

Re: L4Linux hangs on boot
by Jean Wolter 19 Jul '00

19 Jul '00

James Riordan <rij(a)zurich.ibm.com> writes: > Continuing to try to get L4Linux to boot, I run into the problem that > it cannot open an initial console: > > Partition check: > hda: hda1 hda2 hda3 hda4 > hdb: hdb1 hdb2 hdb3 > VFS: Mounted root (ext2 filesystem) readonly. > Freeing unused kernel memory: 36k freed > Warning: unable to open an initial console > > and then goes into an infinite loop of: > > Might anybody have any pointers or suggestions of what to try? Can you send me an image of your boot disk (or if you boot over the network a tar file containing all your binaries)? I would like to try to reproduce the problem here. Jean -- I get up each morning, gather my wits. Pick up the paper, read the obits. if I'm not there I know I'm not dead. So I eat a good breakfast and go back to bed. Peete Seeger

2 1

L4Linux hangs on boot
by James Riordan 19 Jul '00

19 Jul '00

Continuing to try to get L4Linux to boot, I run into the problem that it cannot open an initial console: Partition check: hda: hda1 hda2 hda3 hda4 hdb: hdb1 hdb2 hdb3 VFS: Mounted root (ext2 filesystem) readonly. Freeing unused kernel memory: 36k freed Warning: unable to open an initial console and then goes into an infinite loop of: root: rcvd 4772a0, 4772a0, src: a.0 unknown request from a.0 root: rcvd 4772a0, 4772a0, src: a.0 unknown request from a.0 Actually this is the message of what happens under vmware; when I run natively, it seems to be the same but the messages go by too quickly to read. Might anybody have any pointers or suggestions of what to try? BTW: The build is from the modules "grub l4 linux22 oskit10" from the CVS server on 11 Jul 200. james -- Dr. James Riordan email: rij(a)zurich.ibm.com IBM Forschungslaboratorium notes: James Riordan/Zurich/IBM Säumerstrasse 4 Tel +41 (0) 1 724 89 81 CH-8803 Rüschlikon, Switzerland Fax +41 (0) 1 724 89 53

1 0

RH6.2 and Fiasco
by James Riordan 13 Jul '00

13 Jul '00

I'm still not having a great deal getting Fiasco + hello running. I opted for the slash and burn technique and reinstalled RedHat 6.2, downloaded fresh from the cvs server (10 July 2000) and tried to build from there. The only snag that I hit that I was not sure about was when compiling oskit10, the file oskit10/freebsd/src/lib/libc/i386/gen/alloca.S did not want to compile. There was, however, a comment in oskit mailing list suggesting that a fix ("jmp %edx " changes to "jmp *%edx"). I've appended a script at the end of this mail that make the other small changes that I needed to build. Now that it is built, it crashes at boot time: RMGR: loading (hd1,1)/fiasco/sigma0 RMGR: detected new-style DD-L4 RMGR: starting (hd1,1)/fiasco/main -nokdb proto=0x10128c Welcome to Fiasco! DD-L4/x86 microkernel (c) 1998 TU Dresden - Jul 13 2000 KDB: init L4KD -BP----------------------------------------------EIP: f00011c9 (hang) Any further suggestions? james -- Dr. James Riordan email: rij(a)zurich.ibm.com IBM Forschungslaboratorium notes: James Riordan/Zurich/IBM Säumerstrasse 4 Tel +41 (0) 1 724 89 81 CH-8803 Rüschlikon, Switzerland Fax +41 (0) 1 724 89 53 O / ----------- X Cut Here ----------------------------------------------- O \ #!/bin/bash -nolineediting # # This is a build script for the Fiasco distribution on a generic RedHat 6.2 platform # James Riordan (rij(a)zurich.ibm.com) # # easier to see in color but "alias comment=echo" will also work #function comment () {echo "[01;34m"$*" [0m"} alias comment=echo function backup () { if [ ! -f $1.orig ]; then cp $1{,.orig}; fi } BUILD_ROOT=$PWD export L4_DIR=${BUILD_ROOT}/l4 if [ -d grub -a -d l4 -a -d oskit -a -d oskit10 ]; then comment all directories found else comment you seem to be in the wrong directory exit -1 fi comment making oskit cd oskit ./configure comment edit oskit/flux/config.h to get rid of HAVE_CODE16 if [ ! -f flux/config.h.orig ]; then cp flux/config.h flux/config.h.orig; fi cat << EOF | ed flux/config.h 1 /CODE16 d w q EOF make # making oskit comment making oskit10 cd $BUILD_ROOT/oskit10 ./configure comment edit oskit/config.h to get rid of HAVE_CODE16 backup oskit/config.h cat << EOF | ed oskit/config.h 1 /CODE16 d w q EOF comment edit Makeconf to get rid of have_code16 backup Makeconf cat << EOF | ed Makeconf 1 /code16 s/have_code16// w q EOF comment removing in2000 from oskit/dev/linux_scsi.h backup oskit/dev/linux_scsi.h cat << EOF | ed oskit/dev/linux_scsi.h 1 /in2000 d w q EOF comment changing freebsd/src/lib/libc/i386/gen/alloca.S backup freebsd/src/lib/libc/i386/gen/alloca.S cat << EOF | ed freebsd/src/lib/libc/i386/gen/alloca.S 1 /jmp s/%/*%/ w q EOF make exit comment making fiasco cd $BUILD_ROOT/l4 comment killing pkg/names/man/Makefile as we do not have sdc backup pkg/names/man/Makefile echo all: > pkg/names/man/Makefile comment killing pkg/l4util/examples/perf/Makefile comment as we do not need it and it does not work backup pkg/l4util/examples/perf/Makefile echo all: > pkg/l4util/examples/perf/Makefile comment change egcc to gcc in kernel/fiasco/src/Makefile backup kernel/fiasco/src/Makefile cat << EOF | ed kernel/fiasco/src/Makefile 1 /egcc /CC s/egcc/gcc w q EOF comment create dependency files I am not sure how they are supposed comment to be created but this seems to work cat > Makedeps <<EOF ifeq (\${L4_DIR},) L4DIR = ../../.. else L4DIR = \${L4_DIR} endif ifeq (\${OSKIT_DIR},) OSKITDIR = \$(L4DIR)/../oskit else OSKITDIR = \${OSKIT_DIR} endif include \$(L4DIR)/Makeconf EOF allow_null_glob_expansion=t pushd kernel/fiasco/src/ /bin/rm .*.d for file in *.cc *.c *.S *.cpp; do CC=gcc make -f ${L4_DIR}/Makedeps .$file.d; done popd make pushd pkg/log make popd make

2 1

Re: L4 and grub difficulties
by Lars Reuther 12 Jul '00

12 Jul '00

Christian Stueble <stueble(a)cs.uni-sb.de> writes: > Another question: > I think oskit10 is required now to compile the pkg package, but some > subdirectories also use oskit/flux/c/*.h include files. Do I really need both > oskit versions or should the makefiles be modified to use oskit10/oskit/c/*.h? Which packages? All packages available in our remote CVS should still use the old OSKit (version 0.6), except the Fiasco kernel. We are currently working on a support package for the latest OSKit, but aren't finished yet. Lars -- Dipl.-Inf. Lars Reuther Dresden University of Technology Department of Computer Science Operating Systems Research Group Phone: +49 (351) 463-8401 Fax: +49 (351) 463-8284 Email: reuther(a)os.inf.tu-dresden.de

1 0

Re: Problems compiling Fiasco (OSKIT)
by Pete Buechler 07 Jul '00

07 Jul '00

On Fri, 07 Jul 2000, Christian Stueble wrote: > I had the same problems. Use an older compiler. The > gcc-2.7.2.3 works for me. To compile Fiasco, > use e.g. egcs-2.91.60. I used egcs-2.91.66 for Fiasco and gcc-2.7.2.3 for oskit. Then I had to follow the directions by Edmund GRIMLEY EVANS <edmundo(a)rano.org> - remove HAVE_CODE16 from oskit/config.h - remove have_code16 from Makeconf - remove in2000 from oskit/dev/linux_scsi.h Then I had all kinds of path problems in the Makefiles, having to search for various paths and add them to the include files, plus moving header files around. And when the dust all settled, Fiasco came up and printed some stuff, but the hello server never managed to print hello to my screen. -- Pete Buechler : SuSE Linux Developer Work e-mail : peterb(a)suse.com Work web page : http://www.suse.com/~peterb Personal e-mail : peter.buechler(a)home.com

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

l4-hackers July 2000