eesrajm@cc.brunel.ac.uk (Andrew J Michael) (10/06/90)
Having just had a good look at the AmigaMINIX sources, I'm not entirely surprised that there are problems with floppy disks. This is NOT meant as a criticism of AmigaMINIX (quite the opposite, it's VERY clever). It's more a comment on the Amiga's hardware design. From looking at the code, it seems that the Amiga doesn't understand sectors; it only knows about tracks, and raw ones at that. To simulate a normal MFM disk controller, some natty software in the AmigaMINIX kernel manufactures what looks like an MFM track, complete with sector gaps, headers and CRCs and gets the Amiga hardware to write it out to the floppy as a raw track. I don't know exactly how P-H are duplicating the floppies, but I can well see that there might be problems bearing this in mind. So for those who have one of more faulty Amiga MINIX floppies, try the following. It isn't guaranteed to work, but probably worth an attempt. Get someone with a genuine MFM floppy controller to copy the faulty disk(s) for you. The ideal here is an ST; either use something like procopy under TOS, or copy the disks under MINIX if you can read them under ST MINIX. If you can't find an ST, then an IBM with 720k floppies should work. You probably won't be able to copy them with diskcopy, but one of the many disk copying programs supplied for nefarious reasons should cope. The theory here is that the proper MFM controllers like the 1772 in an ST have pretty good error performance, certainly better than the software in AmigaMINIX. They should be able to recover data which is unreadable on the Amiga, and write it out in a proper error-free MFM format. If anyone trys this and either succeeds or fails, please let me know. If all else fails, you could try asking The MINIX Centre to re-write the floppy for you. Andy Michael -- Andy Michael (eesrajm@cc.brunel.ac.uk) " Emulation is the sincerest 85 Hawthorne Crescent form of pottery." West Drayton Middlesex - William Frend De Morgan UB7 9PA
jkurki@tukki.jyu.fi (Jouni Petri Kurki) (10/07/90)
(my first reply here... ) In article <1850@Terra.cc.brunel.ac.uk> eesrajm@cc.brunel.ac.uk (Andrew J Michael) writes: > >Having just had a good look at the AmigaMINIX sources, I'm not entirely >surprised that there are problems with floppy disks. This is NOT meant as >a criticism of AmigaMINIX (quite the opposite, it's VERY clever). It's more >a comment on the Amiga's hardware design. I haven't purchase Amiga minix yet, because I'm looking for ready made HD support (Commodore 2090A). I must admid that I probably can't do it right now myself. :-) As I have not seen Amiga Minix sources, I can't really know about this particular port of Minix, but I have read original book (Tanenbaum) and IMHO good floppydriver can be done with reasonable knowledge of Amiga hardware. >From looking at the code, it seems that the Amiga doesn't understand sectors; >it only knows about tracks, and raw ones at that. To simulate a normal MFM >disk controller, some natty software in the AmigaMINIX kernel manufactures >what looks like an MFM track, complete with sector gaps, headers and CRCs and >gets the Amiga hardware to write it out to the floppy as a raw track. Amiga hardware is a bit unique here. But there has been at least four commercial,and couple PD programs that can read/write PC formatted 720k floppies. Crossdos (Commercial, Consultronics) and MessyDos (PD, Olaf Siebert) even have MS-dos compatible filesystems for Amiga, so you can directly access all files from PC 720k disks. Besides full-track caching shoud make disk I/O quite a lot faster compared to single sector read/write. It needs more code than simple diskcontroller chip (as found in Pc) but it's allso much more flexible. I belive these troublsa are some minor features (bugs?) in Amiga-minix software, and will be fixed quite soon (I hope). Well... these are my thougths anyway... Jouni Kurki PS. Is there good English spellchecker in Minix? ;-) -- ============================================================================= Jouni Kurki Internet: jkurki@tukki.jyu.fi ( I need to write something clever here )
ast@cs.vu.nl (Andy Tanenbaum) (10/09/90)
In article <1850@Terra.cc.brunel.ac.uk> eesrajm@cc.brunel.ac.uk (Andrew J Michael) writes: >If anyone trys this [copying Amiga disks on an ST or PC] > and either succeeds or fails, please let me know. Please post the results to the net. It is now abundantly clear that they company that did the disk production munged the boot disk. Fortunately, there is a backup file, minix.img.bu on it, which may save the day. I am not yet sure how many other disks were badly copied. Part of the problem is that the company was instructed to copy disk 1 as an Amiga disk and the rest as PC 720K disks. This instruction was repeated many times in many media (phone, letter, FAX, ...) but they probably couldn't believe it, and did it their way instead. After all, they are the experts on disk copying, and I am merely a humble user. Andy Tanenbaum (ast@cs.vu.nl)
JAGBDED1%PANAM.BITNET@ricevm1.rice.edu (10/10/90)
Andy Tanenbaum writes: >In article <1850@Terra.cc.brunel.ac.uk> eesrajm@cc.brunel.ac.uk (Andrew J > Michael) writes: >>If anyone trys this [copying Amiga disks on an ST or PC] >> and either succeeds or fails, please let me know. > >Please post the results to the net. It is now abundantly clear that >they company that did the disk production munged the boot disk. Fortunately, >there is a backup file, minix.img.bu on it, which may save the day. I am >not yet sure how many other disks were badly copied. > I was able to duplicate the 720K floppies on an AT clone using Copy II PC. The P-H disks seem to be ok as far as that goes: they seem to be real 720K 9 sectors/track floppies. The boot floppy was easily duplicated with the Amiga. >Part of the problem is that the company was instructed to copy disk 1 as >an Amiga disk and the rest as PC 720K disks. This instruction was repeated >many times in many media (phone, letter, FAX, ...) but they probably couldn't >believe it, and did it their way instead. After all, they are the experts >on disk copying, and I am merely a humble user. > >Andy Tanenbaum (ast@cs.vu.nl) This is not the real problem though. The minix floppy driver is somehow corrupting every floppy it writes on. If you run minix without write protection, you soon end up with many unreadable floppies. This has serious implications: you can't do anything (like recompile,edit,copy) on an Amiga. The only people in a position to correct the floppy driver are the 1.5 ST people. Needless to say, I'm bathing in frustration over here. I can only hope that the problem can be resolved with a minimum of pain. Joe "AmigaMan" Gonzalez JAGBDED1@PANAM jagbded1%panam.bitnet@ricevm1.rice.edu Amiga: floppies without sector gaps, who would have guessed?
vodall@hpfcso.HP.COM (Bill Vodall) (10/11/90)
/ hpfcso:comp.os.minix / ast@cs.vu.nl (Andy Tanenbaum) / 5:14 pm Oct 8, 1990 / In article <1850@Terra.cc.brunel.ac.uk> eesrajm@cc.brunel.ac.uk (Andrew J Michael) writes: >If anyone trys this [copying Amiga disks on an ST or PC] > and either succeeds or fails, please let me know. Please post the results to the net. It is now abundantly clear that they company that did the disk production munged the boot disk. Fortunately, there is a backup file, minix.img.bu on it, which may save the day. I am not yet sure how many other disks were badly copied. Andy Tanenbaum (ast@cs.vu.nl) ---------- Yes. Thank you a thousand times to the person who put the minix.imp.bu on the BOOT disk. I received my Amiga Minix last night. Upon making a backup of the BOOT disk the dreaded bad track #13 showed up. After playing with it some I was able to copy the entire disk, except for "minix.img", into the AMIGA ram disk. I renamed the minix.img.bu to minix.img and built a new boot disk. Hurray! It works. And this is on an A1000! Now if only I can get the disk format program "BOOT:c/transfer" to work. It locks up the system whenever I run it. Bill "finally on Minix after all these years"
raymond@cs.vu.nl (=Raymond Michiels) (10/11/90)
JAGBDED1%PANAM.BITNET@ricevm1.rice.edu writes: > This is not the real problem though. The minix floppy driver is somehow >corrupting every floppy it writes on. If you run minix without write >protection, you soon end up with many unreadable floppies. OK, so it seems that are some minor (ahem) problems with AmigaMINIX. We are willing to fix the problems, but until now we don't have a clue as to what the problems are. We do know that all BOOT disks have a read error in the "minix.img" file, but fortunately enough we have included a backup of this file ("minix.img.bu"). Other problems that I have heard of are other unreadable disks (MINIX file systems) and, as Joe Gonzalez found out, that AmigaMINIX can't read back data it has written to disk. Since we have debugged MINIX before we shipped it to P-H we are wondering what might cause the abovementioned problems. To get a better idea of the situation I would like to have a small reports on people's experiences with AmigaMINIX: good and bad, from America and Europe, on A500's, and A2000's, etc, etc. When we know what is going on, you'll be the first to know. I have tried AmigaMINIX at several other places and haven't found any of the described bugs, so I really am curious as to what is happening. -Raymond. PS: Just in case you're wondering: I have personally recompiled AmigaMINIX from scratch on my own Amiga500 (1 disk driver, 1.5Mb) without any errors.
root@cca.ucsf.edu (Systems Staff) (10/12/90)
In article <7897@star.cs.vu.nl>, raymond@cs.vu.nl (=Raymond Michiels) writes: > > I have tried AmigaMINIX at several other places and haven't found any > of the described bugs, so I really am curious as to what is happening. > > -Raymond. > > PS: Just in case you're wondering: I have personally recompiled AmigaMINIX > from scratch on my own Amiga500 (1 disk driver, 1.5Mb) without any > errors. But did you do this testing on a copy as being shipped by P.H. to ordinary users? Thos Sumner Internet: thos@cca.ucsf.edu (The I.G.) UUCP: ...ucbvax!ucsfcgl!cca.ucsf!thos BITNET: thos@ucsfcca U.S. Mail: Thos Sumner, Computer Center, Rm U-76, UCSF San Francisco, CA 94143-0704 USA I hear nothing in life is certain but death and taxes -- and they're working on death. #include <disclaimer.std>
guineau@wjg.enet.dec.com (W. John Guineau) (10/12/90)
I've had the same problems as others with AmigaMINIX. The boot floppy was easy enought to fix. Now I have a 'corrupted' /usr floppy, system binaries #2 and my c compiler disk got hosed the first time I tried to recompile something and write it to the C compiler disk. Since my /usr disk has been bad since day 1, I am unable to access things like diskcopy, mkfs, fsck etc. transfer is real sensitive, and can't read some floppies I write under AmigaDOS. I find I need to format a floppy every time I want to transfer from AmigaDOS to Minix. Someone sent a copy of fsck to me so I could transfer it to minix and try to repair the /usr filesystem. I (finally) got it to transfer only to find that transfer on minix creates a file with NO execute permission and chmod is on the corrupted /usr disk... can you say 'frustration' ? :-) -- W. John Guineau guineau@wjg.enet.dec.com Digital Equipment Corporation Marlboro MA. 01752
guineau@wjg.enet.dec.com (W. John Guineau) (10/12/90)
In article <1990Oct12.074457@wjg.enet.dec.com>, guineau@wjg.enet.dec.com (W. John Guineau) writes: |> From: guineau@wjg.enet.dec.com (W. John Guineau) |> Newsgroups: comp.os.minix |> Subject: Re: Trouble with Amiga floppies - a possible solution |> Reply-To: guineau@wjg.enet.dec.com |> |> I've had the same problems as others with AmigaMINIX. |> |> The boot floppy was easy enought to fix. |> |> Now I have a 'corrupted' /usr floppy, system binaries #2 and my c compiler |> disk got hosed the first time I tried to recompile something and write it |> to the C compiler disk. |> |> Since my /usr disk has been bad since day 1, I am unable to access things like |> diskcopy, mkfs, fsck etc. |> |> transfer is real sensitive, and can't read some floppies I write under |> AmigaDOS. I find I need to format a floppy every time I want to transfer |> from AmigaDOS to Minix. |> |> Someone sent a copy of fsck to me so I could transfer it to minix and |> try to repair the /usr filesystem. I (finally) got it to transfer only to |> find that transfer on minix creates a file with NO execute permission |> and chmod is on the corrupted /usr disk... |> |> can you say 'frustration' ? :-) |> |> -- |> W. John Guineau guineau@wjg.enet.dec.com |> Digital Equipment Corporation |> Marlboro MA. 01752 |> Forgot to add my configuration: Amiga 2500/30 (28mhz Imtronics 68030/882) 4 meg 32 bit ram, 1 meg chip (the new fat agnus) GVP hardcard w/200 MB Conner Peripherals scsi disk 2 internal floppies I pull the Imtronics 030 and usually the GVP when booting Minix. -- W. John Guineau guineau@wjg.enet.dec.com Digital Equipment Corporation Marlboro MA. 01752
raymond@cs.vu.nl (=Raymond Michiels) (10/12/90)
>> PS: Just in case you're wondering: I have personally recompiled AmigaMINIX >> from scratch on my own Amiga500 (1 disk driver, 1.5Mb) without any >> errors. >But did you do this testing on a copy as being shipped by P.H. to >ordinary users? Yes: I did find the error on track 13 ("minix.img") but the other disks worked fine for me. -Raymond.
peter@ficc.ferranti.com (Peter da Silva) (10/13/90)
Sounds to me like the decision to make the Amiga port native rather than hosted is coming home to roost. When can I expect to run Minix as a task in an Intuition window? -- Peter da Silva. `-_-' +1 713 274 5180. 'U` peter@ferranti.com
lennox@minilove.diag.stratus.com (Craig Scott Lennox) (10/15/90)
In article <YMD6AKB@xds13.ferranti.com> peter@ficc.ferranti.com (Peter da Silva) writes: Sounds to me like the decision to make the Amiga port native rather than hosted is coming home to roost. When can I expect to run Minix as a task in an Intuition window? I, for one, am glad they did it the way they did. IMHO, it would be a hell of a lot easier (and fun) for a dedicated Amiga hacker to change AmigaMINIX to be a hosted task than the other way round. I'd hate to have to dig through all those hardware references.... -- | flame me at: lennox@minilove.diag.stratus.com, (Craig Scott Lennox) | |"Oh boy, virtual memory! Now I'm gonna make myself a REALLY BIG ram disk!" | | Disclaimer: My opinions are covered by section 2b of the Gnu Public | | License and thus do not belong to Stratus Computer. |