Re: [cbm4linux-users] XMP1541 parallel Problems
Brought to you by:
cbm4linux
From: Michael K. <mic...@pu...> - 2003-05-26 19:49:32
|
Andreas, > Infact with SJD I have noticed several problems that are all on the behal= f of > SJD though. E.g. > > # cbmctrl status 8 > ............................................ > # cbmctrl dir 8 > cbmctrl: dir: No such device Uhm, dumb question: does a SJD-equipped drive work with a stock 64 kernel? > Since here it doesn't even respond to standard IO commands correctly I wo= nder > how at all it works out :) ... will check closer whenever I get my JD128 > back. > > Plain JD works rather well, just doesn't seem reliable. Actually, as I no= ted > on a previous mail, JD+XMP-Cable seems slightly faster than plain > ROM+XMP-Cable on SERIAL (1 and2) transfers. That's probably just because of the higher stepper frequency. No idea why it should be less reliable, though. > Trying JD+XMP with parallel modes caused the same type of error behaviour= I > used to have. Enabling -w warp mode completely bugs out. error 23 all ove= r > the place and no sector is transferred ok. > > Didn't take the time to check with all other serial modes yet. > Whenever I feel happy and my ordinary (Warp&Parallel) transfers works, I = may > start to dig around why the stuff doesn't work with SJD especially. > > Joe, regarding my previous posts, is there anything else I could try, or = have > you heard of similar problems before ? > I am pretty sure I tried about anything possible, without even disturbing= the > errors or influencing the transfer at all :( How about parallel turbo? > Even under heavy system load the transfer goes smooth for quite a few tra= cks, > bugs out, continues for some tracks, bugs out ... etc. > It does the whole process slower though. That's normal, since all user-space protocols rely on busy-loops. > The errors come up on the same disk > on completely random locations, cables are tripple checked and everything= =2E > All this connected with the adaptor from your cable shop I assume. (got i= t > from Protovision ppl) --- just hardwire - hacked to become XM instead of = XE. > > ANY guess ??? I'll try anything --- still regarding myself as someone kno= wing > a little about drives... :) Ok, wild guess: Is your drive speed ok? I hacked up a small drivemeter a while ago (hey, even with GUI!): http://a98.shuttle.de/~michael/rpm1541 (both binary + source, needs GTK+-1.2) It will trash track 41, so be sure to use a scratch disk. Good luck, --=20 Michael "Nicht alle Steuerelementpfade geben einen Wert zur=FCck" -- Microsoft C/C++ Version 12.00.8186 |