>Hello Tobias... I was hoping that you would jump in here. I would be thrilled >to have whatever info you could offer on the >powerbook. I didn't mean any >offence by my posts, I just wanted to figure out what changes had been >made.... you are >right: what is in the kernel is correct. i just wanted to >cross-check the kernel code against my actual machine and then >provide some >useful information with the specific pointers that were not in the code in >case anyone else wants to jump in >and check them out. I checked out all the >addresses in nbpmac_node.h in the kernel and In the original code that you did >>not fix up, there seem to be a few pointers to nothing for sound and floppy >left over from 7 years ago, followed by /* XXX >Where is it? */ with no >reference in the interrupt_info section. Those are what I am trying to help >with. You all know a lot >more about this than i do, but I have some time to >kill, so I did a bunch of research. Hi Alex, sorry for reacting like this. But I didn't want you to reimplement the old buggy interrupt management routines. And I know that would have been much work - for nothing. In drivers/block are three files that represent the attempt to implement a floppy driver: swim_asm.c swim_asm_ppc.S swim_driver.c They are ported from m68k Linux and the file swim_asm_ppc.S is a port of the original m68k assembler code to PPC assembler done with a tool program. swim_asm.c is a C rewrite of that code. swim_driver.c is the main driver file. You have to activate 'SWIM2 floppy support' in the kernel config in order to compile the driver. The driver does something with the drive head but nothing more. And there doesn't seem to be an interrupt assigned to the floppy controller.
>I am just trying help anyone passing through here wit as much information as I >can to see if i hit on anything new that >might be useful. My primary >concerns are getting sound to work and maybe getting the floppy to work and >maybe help with >the power-off issue... I have not looked into the SCSI at >all. I re-traced all the I/O in Macsbug, TMON and Macnosy based on >known >addresses from 68k-style machines working on a similar ROM, and then guessed >the rest and dissassembled the low >memory addresses to make sure... and >indeed confirmed that what is in the kernel now is correct... I am just >looking for a >way to get the other features working. I used MacsBug to trace and modify the hardware registers and to disassemble the drivers. >You are absolutely right that there is only one VIA, and the other is emulated >in the same manner as the RBV virtual VIA2 on >other machines. i know that >Apple had problems with that... so I am sure that is part of the issue. I >also know that on >some of those old machines, the 68k emulator was used to >switch the status register through the OS since the powerpc >itself could send >commands in supervisor mode... it had to clue-in the 68k emulator to do it, >and then execute some >instructions to translate from the 7 interrupt levels >of 68k to the single register of the powerpc... but that is all old Apple >>stuff that you all already know and is easily available in Apple's hardware >developer notes, as well as all the low mwmory >pointers to hardware addresses >from the ROM and very detailed notes about interrupt handling on the x100 >series. >I am also trying to figure out how the kernel polls the ADB/PMU timer since >the countdown is what seemed to keep the >machine from shutting off in >MacOS... I have not found it yet, so maybe I am looking for the wrong thing. The file drivers/macintosh/via-pmu.c is the PMU driver. I modified it to match more with the code from Mac OS X. In what occasions do you experience shutdowns? Are you using the latest kernels built by me? I only heard of shutdowns that had to do with PCMCIA cards. And I never heard anything of a PMU countdown. I know that the classic Mac OS uses polling to communicate with the PMU but the Linux kernel uses interrupts. That could be the cause for the problem because the PMU will shutdown the machine if the kernel doesn't respond in time or when it interrupts the transfer in order to respond to another interrupt. >I have run MkLinux... from the first developer release through the newest >revision. I have it installed on a machine right >now actually. You are >right... it is buggy, but my floppy and sound "work" and it did not randomly >shut the machine down >except when it crashed... which happened a lot... so I >was hoping to pull some code from them and the m68k people to see >if it >helps. I even have the MkLinux book if anyone needs any information from >that. The later kernels seem to fix the >powerbook ide timeout problem, but >you can't get them online anymore. I think I am running the last generic >kernel-9 >version something or other. My first install I tried an old kernel >through the CD drive and had to deal with hitting a cmnd->pwr NMI every few >seconds, so I know what you are talking about there... BIG pain in the butt! MkLinux in the latest CVS version on their web site uses PMU polling and I think that is the reason why it doesn't habe PMU problems - but PMU polling slows down the kernel very much because it has to wait until the PMU command is completed. Floppy and sound work in MkLinux??? Not in the versions that you can download in the internet and the MkLinux site says that floppy and sound don't work. If you have the sources for the drivers, go on and port them to Linux!!! >Anyway, I won't step on anyone's toes since I am no Linux guru, but if you >need it, I am good with old mac hardware, and I >have support on my old >machines for Apple's old MPW developer kits, lots of debuggers and can compile >and debug old mac >programs, and I am running machines with about 6 different >versions of MacOS classic as well as OSX. I am happy to help >with whatever >is needed. I myself don't know much about the internals Macintosh architecture but what I needed to know in order to improve the kernel for PowerBook. I'm happy that we finally have somebody who want to continue that work. I don't have many time anymore to deal with the kernel sources. And I don't have the PowerBook 1400 available to test patches. I hope you will success in the port of the sound and floppy driver. If you have any questions feel free to ask me - but I've already forgotten many things... Bye Tobias _______________________________________________________________________ Jetzt neu! Schützen Sie Ihren PC mit McAfee und WEB.DE. 3 Monate kostenlos testen. http://www.pc-sicherheit.web.de/startseite/?mc=022220 ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2005. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Nubus-pmac-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/nubus-pmac-users
