diff --git a/sys/arch/amiga/AMIGA-MMU b/sys/arch/amiga/AMIGA-MMU deleted file mode 100644 index a7fe635ce22a..000000000000 --- a/sys/arch/amiga/AMIGA-MMU +++ /dev/null @@ -1,25 +0,0 @@ -The HP300 version uses a kernel which is loaded into the range 0xfffxxx... but -is linked for range 0x000.. This causes the problem that with switching the -MMU to the mapped kernel, the PC becomes invalid. The HP version solves the -problem by mapping 1:1 the last page of physical memory into virtual memory. - -On the Amiga, there *is* memory at PA 0x0, so we might as well use it, causes -much less grief and weirdness in locore.s. However, since memory down there -is CHIPMEM, inherently slower than FASTMEM, and badly needed for framebuffer -storage space, we'll copy the kernel over to the first bank of FASTRAM, and -when enabling the MMU, the kernel will execute in FASTRAM, although the PC -stays the same. - -This strategy has the big advantage (compared to the solution in Amiga MACH) -that we can link the kernel absolutely to VA 0, for all memory models, since -we'll never execute the kernel in the FASTMEM space while the MMU is turned -off. - -This strategy also means, that we don't have to relocate any addresses while -bootstrapping the mmu!! - -Initialization of the MMU happens in amiga_init.c. This file is quite a mess, -I have generated it trying to understand what's happening in the hp300 locore.s -file. I think, it should be able to handle MMU initialization much cleaner -now that we don't have to think about relocation until the MMU is enabled. -If you need your daily bit of horror, take a look at amiga_init.c... diff --git a/sys/arch/amiga/BUGS b/sys/arch/amiga/BUGS deleted file mode 100644 index 37d86abff297..000000000000 --- a/sys/arch/amiga/BUGS +++ /dev/null @@ -1,33 +0,0 @@ -There are probably dozens of bugs, keep in mind this is the first release :-) - -Those I'm aware of: -- the console has some scrolling bug. This shows in the ircII client, where - incoming message lines just `scroll' in the last displayed line, instead - of scrolling the entire display. -- the dma code is currently broken. I'd really like to get this going again, - it's scribbling over innocent memory at the moment, and I think there might - be some bugs in the A3000 Service Manual describing some registers, we'll - see... -- nfsd and mountd crash regularly when going to multiuser mode... -- the console seems to have some problems displaying text when parity is - enabled. This shows in a distorted login-prompt (multiuser mode). -- the vt200 emulator is really far from perfect, and needs a lot more work - to be honestly called a vt200 (or vt320 even) emulator. -- 8bit characters are all displayed as ^@. Problems could lie in wrong tty - settings or the ite driver. -- although autoconfig information is passed into the kernel and the hardware - table is generated, no I/O-space is currently allocated in kernel VM for - boards. This will probably be one of the first things to fix or nobody is - able to access their boards under BSD ... -- sun-style disklabels are not yet supported. -- disklabels can't be written back to disk. You'll have to configure your - drives under amigados (with hdtoolbox), and just format the partitions - under BSD. I think this is a tolerable limitation. -- the clock runs much too fast. I have an idea the kernel might think in - 60Hz units, when the clock really runs at 100Hz, I'll look into this. -- there's currently no provision for reading the realtime-clock, so time - is always set using the last modification date of the mounted root - filesystem. -- severe crash after dumping to disk... - -to be continued... diff --git a/sys/arch/amiga/INSTALL b/sys/arch/amiga/INSTALL deleted file mode 100644 index 27ed808cab1f..000000000000 --- a/sys/arch/amiga/INSTALL +++ /dev/null @@ -1,121 +0,0 @@ -How to install BSD on your Amiga: --------------------------------- - -*Please* (re)read the README.amiga file, and make sure your system is -supported by the current kernel. - -Since BSD doesn't yet have a floppy driver, and you'll need a root fs -to start using commands (chicken and egg problem..) to make your own -fs, installation of the root fs is a bit hacky... - -vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv -> We copy a binary image (8M) of the root fs directly to the place on < -> your harddisk. This is an extremely dangerous thing to do, since if < -> you get the offsets wrong, you'll destroy data on other partitions of < -> the drive! < -^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ - -If you don't feel like risking the life of innocent data on your disk, -better stop installing now, and wait until installation gets safer in -the future!!! - - -Ok.. if you're reading on, start up hdtoolbox, and create some partitions -for use by BSD. You'll have to at least create two partitions: - -o a root partition: - Must have FileSystem: Custom File System - Identifier: 0x42534452 (`BSDR') - Reserved begin: 0 - end: 0 - Use custom boot code: NO (ie. no checkmark) - FS block size: 512 - You don't have to check Automount, the partition could only - confuse AmigaDOS. - -o a swap partition: - Must have FileSystem: Custom File System - Identifier: 0x42534453 (`BSDS') - Reserved begin: 0 - end: 0 - Use custom boot code: NO (ie. no checkmark) - FS block size: 512 - You don't have to check Automount, the partition could only - confuse AmigaDOS. - - -You might also add a /usr partition, for example on BSDD: - Must have FileSystem: Custom File System - Identifier: 0x42534444 (`BSDD') - Reserved begin: 0 - end: 0 - Use custom boot code: NO (ie. no checkmark) - FS block size: 512 - You don't have to check Automount, the partition might only - confuse AmigaDOS. - - -Other settings (like MaxTransfer and Mask) are ignored. - -Now, you'll have to find out about the exact block start address and length -in blocks of the BSDR fs (BSD will find out about the data of the other -partitions once it's started up). Go into "Change Drive Type" menu, and -write down the "Blocks per Cylinder" value it shows you for this drive. Then -go into the "Partition Drive" menu, select your BSDR partition, enable -"Advanced Options", and write down the "Start Cyl" and "End Cyl" values. - -Doing "Blocks per Cylinder" * "Start Cyl" gives you the block start number -of the BSDR partition, doing "Blocks per Cylinder" * ("End Cyl" - "Start Cyl") -results in the length of the partition in blocks. You'll need these values -to install the root fs onto the drive, so please write them down, and PLEASE -make sure you calculated them right, since if not, you'll scribble anywhere -on the drive, and that just might be.. (insert worst imaginations..). - -************************************************************************** -The rootfs has a length of ~8M, 16448 blocks. Your root fs must have at -least (!) the same amount of blocks. You don't have to make it exactly -16448 blocks though, a task that could even be impossible if the number of -blocks per cylinder on your drive is not a divisor of 16448. -************************************************************************** - -Now you're almost ready to do the installation. Unzip the distributed rootfs.gz -file (you'll need 8M of free space on your amigados partition for this!), and -then do: - - filetodev {START_NUMBER} 16448 rootfs scsi.device {UNIT} 1000 - -with: - {START_NUMBER}: above calculated block start number - {UNIT}: the scsi unit you're installing on, 0-6 - - -The filetodev program is included in the bffs11.lzh archive, and was written -by Chris Hooper (cdh@mtu.edu). - - - -How to start BSD ----------------- - -You start BSD with a loader program from AmigaDOS. This loader is called -`loadbsd', and it takes as parameter the kernel file, normally vmunix. - -So, - - loadbsd vmunix - -will (try to...) boot BSD. The kernel as distributed prefers to boot from -SCSI unit 6. It will probably boot from other units as well (can't test -this at the moment, since I don't want to repartition my other drives). If -you run into problems (ie. BSD says it can't mount the root filesystem), -you'll have to recompile the kernel. Please see the RECOMPILE file for -how to do this. - - -What about /usr ? ------------------ -For the first distribution of the kernel, I decided not to distribute -binaries for the /usr filesystem. Simply, because /usr is huge compared -to root, and a.out headers are likely to change in the future, rendering -a distributed /usr filesystem almost useless. - diff --git a/sys/arch/amiga/README.68040 b/sys/arch/amiga/README.68040 deleted file mode 100644 index 7575f85aa930..000000000000 --- a/sys/arch/amiga/README.68040 +++ /dev/null @@ -1,45 +0,0 @@ -This version of NetBSD now includes support for the 68040. This version -should support both the 68030 and 68040 processors with a single kernel. - -The config option FPSP will include the full Floating Point Support -Package to provide emulation of the floating point instructions not -implemented by the 68040. If the FPSP option is not used, then a -dummy file is included that will cause a panic if any FPSP routine -is called. This allows building a smaller kernel specifically for -the 68030. - -The FPSP is provided as an object file. The FPSP was ported to gas by -Eric Norum (eric@skatter.usask.ca) and is available by anonymous ftp from: - ftp.usask.ca /pub/software/fpsp/fpsp_gas.tar.gz -Note: Motorola still holds the title and all rights to the FPSP. - -Support for the Progressive Peripherals Incorporated Zeus SCSI interface -is also included. If the configuration file does not include support for -the A3000/A2091/GVP11 drivers, then including the "zeusscsi" driver will -configure the driver to use the sd (disk) and st (tape) devices. If any -of the A3000/A2091/GVP11 drivers are configured, then including the zeusscsi -driver will be configured to use a second set of scsi devices: rz (disk) -and tz (tape). The rz and tz devices have different major device numbers: - rz - 5 = block, 9 = character - tz - 8 = block, 23 = character -WARNING: since the PPI Zeus SCSI uses the level 6 interrupt, the kernel -must be compiled with the splbio() macro defined as spl6(). If not, -then then random kernel hangs will occur. One bad side effect of changing -splbio() is that it will also block the clock interrupts, which can -result in incorrect timing. The 53C710 driver (siop.c) also includes -support the the CSA Magnum 40, but this has not been tested yet. - -The 68040 MMU implementation is not optimal at this time. It was added -in a manner to fit in with the current 68030 MMU setup without needing -to completely rewrite the MMU routines. The existing 68030 MMU layout -was implemented using a two level table: a segment table, and a page -table. The 68040 MMU can only use a three level table. To simulate a -two level table, the top level descriptor table is set up and the full -level 2 descriptor table is allocated and used as the "segment" table. -This results in 9 pages, or 72K bytes, being allocated for the initial -page table for each process. - -The reboot command does not work for some reason. It appears to do the -reset instruction, but doesn't seem to do anything after that. - -Michael Hitch diff --git a/sys/arch/amiga/README.CHOPPS-CONSOLE b/sys/arch/amiga/README.CHOPPS-CONSOLE deleted file mode 100644 index 67d52f4f2a80..000000000000 --- a/sys/arch/amiga/README.CHOPPS-CONSOLE +++ /dev/null @@ -1,83 +0,0 @@ -The reason I chose 804, thats 709 plus my local version increments since -709. - -Ok first let me say that this is just a beta. There are known cosmetic -(small) bugs with some of the code. I felt it was necsesary to get -this stuff into the public ASAP. For instance the X window people -really need to be using the new graphics system. I am very concerned -that they have spent alot of time on code they will be re-writing now. -I asked them to drop me a note but no one replied. - -Anyway, there are few patches to the main amiga source tree all of -which were nec. and proper. Note however that I made a patch to -the main sources, sys/sys/ioctl.h had a bug (IOCBASECMD(x) was incorrect.) -my code depends on this being right so I included my change. - -I did not get to clean up my include messiness in amiga/cc_* stuff -or in grf_cc stuff, however I did switch the grf_* stuff over to -the BSD style of users including what they need. - -This release is a major update in the area of graphics. All parts -are not fully done yet but they work and work solidly. The interface -(grf) for monitors can be considered frozen (we may add cosmetics -later but nothing funcitonal.) So that people wishing to write -monitor code can do so now. - -If you are going to write a monitor you can use my code as an example -it may get complex and overwhelming though because of the complexity -of supporting the custom chips. So you may want to start of only -with the grf_* stuff and then consult the grf_cc stuff when needed. - -In the area of the ite. I have consolidated all the monitor dependent -files into a single ite_std.c file. This was possible due to the -new view device (more on that in a little bit) The ite now has -support for _underline_ and BOLD attributes as well as continuing -support of inverse as before. Becuase of these additions the font -programs need to be updated (I added the code in fontdumper and -changed the kernel_font.c.distrib to get you started.) The only -thing that gets added is kernel_font_(baseline|boldsmear) attributes. - -All known ite scrolling bugs and repeating bugs are fixed in this version. -(actually there is one, but I need to get this out.) - -The new ``/dev/view??'' are the character devices that replace the -old ``/dev/grf''. These new psuedo-devices represent a single monitor -screen. They go through the new grf interface in a portable way. -All programs will interface graphics on Amiga-BSD through these devices. -(ite does) - -I will shortly (a day or two) be releasing some example code on how -to program with the view?? devices. For now let me give you an -example on how to map the bitmap data into your process with the -view device. - -bitplane_data = mmap (0, size, PROT_READ|PROT_WRITE, MAP_FILE, fdesc, 0); - -``size'' is the number of bytes of bitmap data to map. ``fdesc'' is the -file descriptor returned from open(). - -There are some major bugs with mmap()'ing the 709 grf device. The largest -being the allowing of processes to quickly vm_fault the kernel. (one access). -These are gone now. - -Support for the custom hardware is now in the kernel, things such as a -``/dev/audio'' will probably be written. Its easy now. Any way, the ite -bell now goes through this interface. - -Some missing stuff: mouse pointer. sorry sprites are not retargetable* -X windows is however and when that is done you will get your pointer -back (exactly what was it doing there anyway?) - -A screen blanker. This is a very easy one I hope to have installed -into the system in a upcoming patch (maybe 20 lines of code needed) -Again I need to get this out now so it was skipped. - -I hope everyone enjoys the new code features. Now i can try and get -emacs working... - -Chris... - -(this was written after a 20 hour stint to try and finalize stuff please -forgive mispellings and improper grammar. thanks) --- -(* spites could be but I didn't make them.)