- l4-hackers - example.com Site

Multiple Test Environments
by David Denny 21 Sep '03

21 Sep '03

Hello L4 There seemed to be a sentiment that your group might appreciate having the Fiasco installation process tested under a variety of Linux distributions. I've had to swap the machine I'd originally acquired for this project, because of video issues with Fiasco (onboard card, Compaq mobo, say no more), so I have a bit more disk space and with careful use of the server I can probably compile/install for between 4 and 6 distributions. Those I have in mind are Debian (as a basic sync point with yourselves). SuSE as an RPM distro, Slackware because it's different, and the two DIY distros LFS and Gentoo. What I envisage is coming back to yourselves _only_ if I have Debian issues, and for the other distros I'm happy to hack my way through and send you back the completed "Expedition Log" for each of them. That is, if you'd like to have something as a resource for other external users to use. I would not expect to ask for help outside Debian. My time is limited so I would tend to go for a major hacking event, from time to time, rather than keeping up with things as they happen At The Centre as it were. I run other free *nixes but my feeling is that at present, as an evaluation envelope, Linux is sufficient for your needs. Looking forward to this, rather

1 0

RE: build error
by "Wen, Jun (溫軍 IAC-N)" 11 Sep '03

11 Sep '03

Hi, thanks,now I am using i386-elf-gcc,but also has some error. would you please give a advise which compiler and which enviroment could compile them pass? i am using fiasco-1.0.tar.gz and fiasco-1.0-oskit.tar.gz BRS john -----Original Message----- From: Volkmar Uhlig [mailto:volkmar@ira.uka.de] Sent: Wednesday, September 10, 2003 11:59 AM To: Wen, Jun (? ? IAC-N) Cc: l4-hackers(a)os.inf.tu-dresden.de Subject: RE: build error Using the native cygwin compiler will quite likely make you pulling your hair out. Try building a cross compiler for linux and elf and use that instead. The PE standard the cygwin gcc spits out conflicts with many assumptions in asm code and linker files, such as naming conventions of sections, symbols, etc. - Volkmar

1 0

RE: build error
by Volkmar Uhlig 10 Sep '03

10 Sep '03

Using the native cygwin compiler will quite likely make you pulling your hair out. Try building a cross compiler for linux and elf and use that instead. The PE standard the cygwin gcc spits out conflicts with many assumptions in asm code and linker files, such as naming conventions of sections, symbols, etc. - Volkmar

1 0

RE: build error
by "Wen, Jun (溫軍 IAC-N)" 10 Sep '03

10 Sep '03

Hi thanks for reply quickly.follow advise it shows below. BRS john make -I../src -f ../src/Makefile \ srcdir=../src "all" Reading specs from /usr/lib/gcc-lib/i686-pc-cygwin/3.2/specs make[1]: Entering directory `/cygdrive/f/fiasco-1.0/l4/kernel/fiasco/build' ... Making context-ia32-ux.o g++ -v -c -MD -I../src/types/ia32 -I../src/types -I../src/drivers/ia32 -I../src/ drivers -I../src/kern/ia32 -I../src/kern/shared -I../src/kern -I../src/boot -I.. /src/lib/libk/ia32 -I../src/lib/libk -I../src/lib/kern/include -I../src/lib/amm -I../src/lib/lmm -I../src/lib/gzip -I../src/lib/disasm -I../src/lib/minilibc/ia3 2/include -I../src/lib/minilibc/include -I. -Iauto -fno-rtti -fno-exceptions - march=i586 -fomit-frame-pointer -mpreferred-stack-boundary=2 -fno-defer-pop -fr eg-struct-return -g -Wall -W -finline-limit=2000 -O3 auto/context-ia32-ux. cc - o context-ia32-ux.o Reading specs from /usr/lib/gcc-lib/i686-pc-cygwin/3.2/specs Configured with: /netrel/src/gcc-3.2-3/configure --enable-languages=c,c++,f77,ja va --enable-libgcj --enable-threads=posix --with-system-zlib --enable-nls --with out-included-gettext --enable-interpreter --disable-sjlj-exceptions --disable-ve rsion-specific-runtime-libs --enable-shared --build=i686-pc-linux --host=i686-pc -cygwin --target=i686-pc-cygwin --enable-haifa --prefix=/usr --exec-prefix=/usr --sysconfdir=/etc --libdir=/usr/lib --includedir=/nonexistent/include --libexecd ir=/usr/sbin Thread model: posix gcc version 3.2 20020927 (prerelease) /usr/lib/gcc-lib/i686-pc-cygwin/3.2/cc1plus.exe -v -I../src/types/ia32 -I.. /src /types -I../src/drivers/ia32 -I../src/drivers -I../src/kern/ia32 -I../src/kern/s hared -I../src/kern -I../src/boot -I../src/lib/libk/ia32 -I../src/lib/libk -I../ src/lib/kern/include -I../src/lib/amm -I../src/lib/lmm -I../src/lib/gzip -I../sr c/lib/disasm -I../src/lib/minilibc/ia32/include -I../src/lib/minilibc/include -I . -Iauto -MD context-ia32-ux.d -MQ context-ia32-ux.o -D__GNUC__=3 -D__GNUC_MINOR __=2 -D__GNUC_PATCHLEVEL__=0 -D__GXX_ABI_VERSION=102 -D_X86_=1 -D_X86_=1 -Asyste m=winnt -D__OPTIMIZE__ -D__STDC_HOSTED__=1 -Acpu=i386 -Amachine=i386 -Di386 -D__ i386 -D__i386__ -D__i586 -D__i586__ -D__pentium -D__pentium__ -D__tune_i586__ -D __tune_pentium__ -D__stdcall=__attribute__((__stdcall__)) -D__fastcall=__attribu te__((__fastcall__)) -D__cdecl=__attribute__((__cdecl__)) -D_stdcall=__attribute __((__stdcall__)) -D_fastcall=__attribute__((__fastcall__)) -D_cdecl=__attribute __((__cdecl__)) -D__declspec(x)=__attribute__((x)) -D__i386__ -D__i386 -D__CYGWI N32__ -D__CYGWIN__ -Dunix -D__unix__ -D__unix -isystem /usr/lib/gcc-lib/i686-pc- cygwin/3.2/../../../../include/w32api -isystem /usr/lib/gcc-lib/i686-pc-cygwin/3 .2/../../../../i686-pc-cygwin/lib/../../include/w32api auto/context-ia32-ux. cc - D__GNUG__=3 -D__DEPRECATED -quiet -dumpbase context-ia32-ux.cc -march=i586 -mpre ferred-stack-boundary=2 -g -O3 -Wall -W -version -fno-rtti -fno-exceptions -fomi t-frame-pointer -fno-defer-pop -freg-struct-return -finline-limit=2000 -o /cygdr ive/c/DOCUME~1/INC970~1/LOCALS~1/Temp/ccNnjfaX.s GNU CPP version 3.2 20020927 (prerelease) (cpplib) (80386, BSD syntax) GNU C++ version 3.2 20020927 (prerelease) (i686-pc-cygwin) compiled by GNU C version 3.2 20020927 (prerelease). ignoring nonexistent directory "/usr/local/include" ignoring nonexistent directory "/usr/i686-pc-cygwin/include" ignoring duplicate directory "/usr/i686-pc-cygwin/lib/../../include/w32api" #include "..." search starts here: #include <...> search starts here: ../src/types/ia32 ../src/types ../src/drivers/ia32 ../src/drivers ../src/kern/ia32 ../src/kern/shared ../src/kern ../src/boot ../src/lib/libk/ia32 ../src/lib/libk ../src/lib/kern/include ../src/lib/amm ../src/lib/lmm ../src/lib/gzip ../src/lib/disasm ../src/lib/minilibc/ia32/include ../src/lib/minilibc/include . auto /usr/include/w32api /usr/include/c++/3.2 /usr/include/c++/3.2/i686-pc-cygwin /usr/include/c++/3.2/backward /usr/lib/gcc-lib/i686-pc-cygwin/3.2/include /usr/include End of search list. /usr/lib/gcc-lib/i686-pc-cygwin/3.2/../../../../i686-pc-cygwin/bin/as.exe --tra ditional-format -o context-ia32-ux.o /cygdrive/c/DOCUME~1/INC970~1/LOCALS~1/Temp /ccNnjfaX.s /cygdrive/c/DOCUME~1/INC970~1/LOCALS~1/Temp/ccNnjfaX.s: Assembler messages: /cygdrive/c/DOCUME~1/INC970~1/LOCALS~1/Temp/ccNnjfaX.s:826: Error: unknown pseud o-op: `.previous' make[1]: *** [context-ia32-ux.o] Error 1 make[1]: Leaving directory `/cygdrive/f/fiasco-1.0/l4/kernel/fiasco/build' make: *** [all] Error 2 -----Original Message----- From: Udo A. Steinberg [mailto:us15@os.inf.tu-dresden.de] Sent: Wednesday, September 10, 2003 9:24 AM To: l4-hackers(a)os.inf.tu-dresden.de Subject: Re: build error _______________________________________________ l4-hackers mailing list l4-hackers(a)os.inf.tu-dresden.de http://os.inf.tu-dresden.de/mailman/listinfo/l4-hackers

1 0

build error
by "Wen, Jun (溫軍 IAC-N)" 10 Sep '03

10 Sep '03

Hi, i am new to L4 and trying to build l4 using cygwin.but get some error like this in \fiasco-1.0\l4\kernel\fiasco\build just make ---------------------------------------------------------------------------- ----------------------- g++ -c -MD -I../src/types/ia32 -I../src/types -I../src/drivers/ia32 -I../src/dri vers -I../src/kern/ia32 -I../src/kern/shared -I../src/kern -I../src/boot -I../sr c/lib/libk/ia32 -I../src/lib/libk -I../src/lib/kern/include -I../src/lib/amm -I. ./src/lib/lmm -I../src/lib/gzip -I../src/lib/disasm -I../src/lib/minilibc/ia32/i nclude -I../src/lib/minilibc/include -I. -Iauto -fno-rtti -fno-exceptions -mar ch=i586 -fomit-frame-pointer -mpreferred-stack-boundary=2 -fno-defer-pop -freg- struct-return -g -Wall -W -finline-limit=2000 -O3 auto/context-ia32-ux.cc -o c ontext-ia32-ux.o /cygdrive/c/DOCUME~1/INC970~1/LOCALS~1/Temp/ccjaRJqc.s: Assembler messages: /cygdrive/c/DOCUME~1/INC970~1/LOCALS~1/Temp/ccjaRJqc.s:826: Error: unknown pseud o-op: `.previous' ---------------------------------------------------------------------------- ---------------------------- any suggestion will be appreciated! BRS john

2 1

Re: L4Linux boot stop "param_word bits 16-19 must be 0. Fix yourbindings!"
by Udo A. Steinberg 07 Sep '03

07 Sep '03

On Fri, 05 Sep 2003 20:45:53 +0100 David Denny (DD) wrote: DD> I've now got myself a sensible environment and I know which compiler is DD> which. I have jdb but not kdb. I have no info to help me debug the DD> "fix my bindings" message that I can interpret. KDB has been superseded by JDB. JDB is in fact the Fiasco kernel debugger. When entering the kernel debugger, you should get a line similar to this: ---------------------------------------------------ESP:201807a8 EIP:00006526 (6.00) jdb: 6.00 in this case is the thread that was active, when the debugger was entered (thread 0 of task 6). When the "fix your bindings" message appears, then note the task number (number before the dot). Compare it to the task numbers that "rmgr" or "loader" report when starting tasks, i.e. RMGR: loading "pingpong-ux" from 01d37000-01d6f72d to [ 00800000-0080f1c3 00810000-00960580 ] task #0x06 at entry 00800000 via trampoline page code 00070150 ^^^^^^^^^^ Task 6 is pingpong-ux or loader | vmlinuz.l4env,#f: Continue at l4env_init (00013350, libloader.s.so) ^^^^^^^^^^^^^^^^ Task 0x0f is L4Linux And voila, you'll find the offending program that needs fixing. DD> Is it possible that it's DD> the linux22 that needs checking out again, as this is what is loading just DD> prior to the message? Could well be. Only the task number will tell you for sure. Remote-CVS has been updated recently (Friday, Sep 5), so all the fixed bindings should have propagated to the wide world. DD> Furthermore, have you instructions on cvs checkout for packages? I've DD> only been to the website and my cvs's so far have been for the package DD> groups like linux22, hello, oskit, etc. You don't need to re-checkout packages. A simple "cvs update -dP" in any of your checked out directories will bring it up-to-date. Of course you still need to recompile the relevant packages afterwards. -Udo.

1 0

New cli/sti semantics in Fiasco 1.0?
by Jacob Gorm Hansen 30 Aug '03

30 Aug '03

hi, as far as I recall, Fiasco used to re-enable interrupts upon returning from a syscall, even if they had been disabled before the syscall. Some of my old code (my patched version of L4Linux22 which I am using with nomadbios) seems to have relied on this behaviour and now fails with a "page fault in cli mode" on the first intruction after an IPC-operation. Am I correct that the behaviour has been changed like this, or do I just have to configure my Fiasco differently? The old Fiasco I am using is from CVS about a year ago. thanks, Jacob

1 1

l4/pkg/con etc
by David Denny 30 Aug '03

30 Aug '03

Hello Hackers. Having got to the stage of compiling successfully my second (debian) hello+fiasco+glinux set, I'm now looking at cool next step (read: "easy wins for the temporally challenged"). UX and CON seem to be my candidates. OK, UX looks more like the world I'm used to, so I won't ask about that yet. But I /do/ have a Q. about the drops con, which has wider learning implications, so a few hints here won't go amiss. My guess is that to yield up a con binary, I cd to l4/pkg/con, "make", and /no/ that isn't it because it falls over... my predecessor instructions were per website, CVSco, compile hello and fiasco using gcc3.2, compile linux22 using gcc2.95. so: how do I compile "con" and other packages, please? (or: a generalised rule to cover this) ;-) TIA David D

2 1

L4Linux boot stop "param_word bits 16-19 must be 0. Fix your bindings!"
by David Denny 28 Aug '03

28 Aug '03

Hello Hackers booting l4linux stops with: KDB: thread_schedule: param_word bits 16-19 must be 0. Fix your bindings! My grub menu contains: +-------------------------------------------------------------------[0]---+ | 0: kernel (hd0,2)/denny/Dresden/rmgr -sigma0 | | 1: modaddr 0x02000000 | | 2: module (hd0,2)/denny/Dresden/main -nokdb -nowait -comspeed 9600 -com| | 3: module (hd0,2)/denny/Dresden/sigma0 | | 4: module (hd0,2)/denny/Dresden/glinux root=/dev/hda3 | | 5: module (hd0,2)/denny/Dresden/hello umm, pasting from the remote serial console.. line 2 ends in "-comport 1" OK let's reboot the big system & paste the whole thing. BTW your hackked grub is now sitting in the mbr of (hd0) AND it WORKS ! 8-) title failsafe chavin:/boot/grub # less menu.lst # Modified by YaST2. Last modification on Tue Aug 26 06:06:59 2003 color white/blue black/light-gray # default 0 # timeout 8 serial --unit=0 --speed=9600 --word=8 --parity=no --stop=1 #terminal --timeout=10 serial console terminal serial console title SuSE Linux 8.2 Kernel 2.4.20 kernel (hd0,2)/boot/vmlinuz root=/dev/hda3 vga=0x31a splash=silent showopts initrd (hd0,2)/boot/initrd title (Remote) SuSE Linux 8.2 Kernel 2.4.20 kernel (hd0,2)/boot/vmlinuz root=/dev/hda3 console=tty0 console=ttyS0,9600n8 initrd (hd0,2)/boot/initrd title Debian Linux testing/unstable Kernel 2.2.20 kernel (hd0,1)/boot/vmlinuz-2.2.20 root=/dev/hda2 splash=silent vga=0x31a title (Remote) Debian Linux testing/unstable Kernel 2.2.20 kernel (hd0,1)/boot/vmlinuz-2.2.20 root=/dev/hda2 console=tty0 console=ttyS0,9600n8 title floppy root (fd0) chainloader +1 title failsafe kernel (hd0,2)/boot/vmlinuz.shipped root=/dev/hda3 showopts ide=nodma apm=of f acpi=off vga=normal nosmp noapic maxcpus=0 3 initrd (hd0,2)/boot/initrd.shipped title L4Linux -+- Hello World kernel (hd0,2)/denny/Dresden/rmgr -sigma0 modaddr 0x02000000 module (hd0,2)/denny/Dresden/main -nokdb -nowait -comspeed 9600 -comport 1 module (hd0,2)/denny/Dresden/sigma0 module (hd0,2)/denny/Dresden/glinux root=/dev/hda3 module (hd0,2)/denny/Dresden/hello title L4Linux-------------------><8------------------------ Apols I have not gone the netboot road, as "all"! my boxes are totally different there are too many risks for cross compiling, I compile for the OS and box where it loads from. Host OS is SuSE8.2 with linux22 compiled under 2.95.3 and l4 + the rest under 3.2 (i.e. reverted), all compiles terminated OK, all libraries & dependencies found to the best of my ability. The same error has stopped me for some 24 hours; I have been round the houses tidying up all I can find here so I await your advice with eagerness. In the meantime, to get a bit nearer, I have the Debian system ready for a similar assault when I have my energy restored... TIA David Denny

3 3

Driver using DDE
by Chris Lambert 26 Aug '03

26 Aug '03

Hello, we try to implement drivers for DROPS using the DDE, but we only have the sounddriver described in the diploma thesis of Christian Helmuth. Are there other sample drivers (smaller) available for us to see how to use DDE? It's part of a project at Operating Systems Chair at University of Saarland. Chris Lambert Markus Pister Marc Schlickling

2 1