sdyer@bbncca.ARPA (Steve Dyer) (11/18/83)
I'm trying to get OS-9 to use all 40 tracks and both sides of my Remex floppy, and have it seek at the higher 6ms rate. (Standard R/S disk is single sided, 35 track, 30ms seek time.) It seems I've been partially successful--those with 40tk, fast step-rate drives will find this useful. I'm also looking for any help with the problem I describe below... First, I do the following (refer to page 39 of the OS-9 Technical Information booklet): OS9: debug (all numbers in hexadecimal) L d1 (my Remex is drive 1--use dN where N is your drive #) . .+14 (select step rate) = 3 (change from 0 [30ms] to 3 [6ms]) - (move back to verify change) . .+4 (select no of tracks) = 28 (change from 35 tracks [23] to 40 [28]) = 2 (change from 1 side to two sides) q (quit the debugger) A "format /d1" procedes fine until it gets past track 0x27 (which looks like the last track on the first side.) At that point, it gets read errors. Looks like there's some problem accessing the other side. For the record, I should state that I'm using a special drive cable from Frank Hogg Labs which allows me to use their FLEX with a double sided drive. The only difference (I think) with this cable and the standard R/S one is that the controller's "side select" signal is passed to the selected drive, rather than being used to select drive 3 (as in R/S DOS). Perhaps the current disk driver ignores side selection, since the R/S cable doesn't supply it? Anyone out there have any ideas or suggestions? Have fun with everything except double-sided drives, though... /Steve Dyer decvax!bbncca!sdyer P.S.: I haven't yet figured out how to save this change. Will keep you posted.
sjt@sb6.UUCP (11/19/83)
Since Steve Dyer obviously did not get my previous responce on using debug to change modules I will post this one as a followup. After you quit the debug module you have only to SAVE the module that you just modified. This will put an exact copy of the module "in memory" onto your current work directory. Then uset ATTR and verify to update the module. The exact sequence should be: verify u < module to update > new updated module attr new updated module e pe this sets execute and public execute After you have formatted a new disk use os9gen to build yourself a new boot module. It will probably be necessary to save all modules out of memory and include them one at a time. Read the book since there are several that you do not include in the os9boot. OS9, OS9P2, INIT, BOOT <--- don't include Frank Hogg also published a letter outlining the mods to make in the modules of OS 9 for the COCO. To paraphrase: change the CCDisk module and if you are game BOOT. a) make changes below as required 1. byte #1(old value) #2(new value) ==== ============ ============ 0007 81 82 Rev number 0008 D3 D0 checksum 01FE 13 10 speed 10=6ms 11=12ms 12=20ms 13=30ms 0204 22 08 delay 0205 2E 8B " 0210 04 41 make drive 2 back of d 0 0211 40 42 make d 3 back of d1 02DD 43 40 speed 40=6ms 41=12ms 42=20ms 43=30ms 02E9 03 00 speed 00=6ms 01=12ms 02=20ms 03=30ms 0308 52 43 crc 0309 4C BC crc 030A EB 6E crc If you make changes exactly as listed you SHOULD not have to do a verify, I do one every time since it is safer to do so. 2. save the module onto disk and then gen yourself a new boot( or since you did a mod to the Rev number just have your startup do a load newmodule each time you boot and this one will overlay the earlier version) I did modify the modules on the RS OS 9 and it does work! This procedure is required since the device descriptors on the RS are different from the real OS 9 (which is what I run since I do all my development on a SWTPC type system). I hope this is of some help as OS 9 is not easy to learn and is not as user friendly as it should be. Sidney Thompson
emjej@uokvax.UUCP (11/21/83)
#R:bbncca:-32100:uokvax:3500015:000:958 uokvax!emjej Nov 21 09:49:00 1983 Re side selection on the CoCo: the Frank Hogg method of using double-sided drives makes each side of a drive look like a separate one-sided drive. The RS disk controller uses the Western Digital 1793 controller chip, which doesn't have a side-select pin. Enough folks are nattering and grommishing (and rightly so, too) about the current disk driver module, which is tied to the RS disks (it *doesn't* look at the device descriptors, which would automatically take care of differing step rates, tracks per side, and the like!) that I think future releases will do things right (to the extent permitted by the RS controller). Speaking of the controller--it is a nasty bottleneck, what with programmed data transfer. If someone would come out with a controller that allowed DMA or kept a one-sector buffer on-cartridge, or *anything* just to keep from pestering the 6809 every time a byte is transferred, OS-9 would really fly on the CoCo. James Jones