9fans - fans of the OS Plan 9 from Bell Labs
 help / color / mirror / Atom feed
* fixes in the works
@ 1995-09-06  7:48 Staffan
  0 siblings, 0 replies; 3+ messages in thread
From: Staffan @ 1995-09-06  7:48 UTC (permalink / raw)


In article <67661.sharris@fox.nstn.ca> sharris@fox.nstn.CA (sharris) writes:

   jmk@plan9.ATt.COM  <jmk@plan9.ATt.COM> wrote:

   >  > From: staffanu@tcs07.nada.kth.SE (Staffan Ulfberg)
   >  > Subject: Re: Installation problems:-(
   >  > ...
   >  > When b.com is started, it prints info about the hard drives it can
   >  > find on the screen; in my case:
   >  > 
   >  > >hd0: 1669260 sectors 854661120 bytes
   >  > >hd1: 128 sectors 280559616 bytes	(This is wrong!  The drive is 560MB.)
   >  > 					(Exactly half of "real capacity"?)

   DOS or the BIOS has a limitation where it can't handle anything > 1023 
   cylinders, so there's often something in the boot sector fiddles the
   numbers. e.g. with a 1200 cylinder 8 head drive, the BIOS pretends
   it's 600 and 16 to get around that limitation. Linux on large drives
   needs to have this nonsense overridden or it gets very confused.

   Sounds to me like Plan 9 & the DOS number-fiddling software don't
   like each other & Plan 9 is only seeing half the cylinders.

I also thought it might have to do with this, but not anymore... I
have deleted all partitions on the second drive, fiddled with the CMOS
setup options,... nothing helps.  (Also, I don't have a DOS driver to
use the >1024 cylinders - I use OS/2 :-)  Also, I have succesfully
installed Plan 9 on my first drive now, which is 850 MB.  I just think
my drive is incompatible.  Anyone else using a Samsung 560 MB?

Staffan

P.S.  I found some very informative documents describing how
partitoining, booting, 1024 cylinder boundaries, etc. work on a PC:

ftp://ftp.rahul.net/pub/lps/hard-disk/How_It_Works...

-- 
/D{def}def /d{.00017 add D}D /C{2 copy dup mul exch dup mul}D /g 150 string
D /y .29 D 150 150 8[.4 0 0 .4 -45 -90]{/x -1.2 D 0 1 149{x y /n 300 D{/n n
5 sub D C exch sub x add 3 1 roll 2 mul mul y add C add 4 gt n 5 eq or{exit
}if}loop pop pop g exch n put /x x d}for /y y d g}image showpage







^ permalink raw reply	[flat|nested] 3+ messages in thread

* fixes in the works
@ 1995-09-04 11:33 sharris
  0 siblings, 0 replies; 3+ messages in thread
From: sharris @ 1995-09-04 11:33 UTC (permalink / raw)


jmk@plan9.ATt.COM  <jmk@plan9.ATt.COM> wrote:

>  > From: staffanu@tcs07.nada.kth.SE (Staffan Ulfberg)
>  > Subject: Re: Installation problems:-(
>  > ...
>  > When b.com is started, it prints info about the hard drives it can
>  > find on the screen; in my case:
>  > 
>  > >hd0: 1669260 sectors 854661120 bytes
>  > >hd1: 128 sectors 280559616 bytes	(This is wrong!  The drive is 560MB.)
>  > 					(Exactly half of "real capacity"?)

DOS or the BIOS has a limitation where it can't handle anything > 1023 
cylinders, so there's often something in the boot sector fiddles the
numbers. e.g. with a 1200 cylinder 8 head drive, the BIOS pretends
it's 600 and 16 to get around that limitation. Linux on large drives
needs to have this nonsense overridden or it gets very confused.

Sounds to me like Plan 9 & the DOS number-fiddling software don't
like each other & Plan 9 is only seeing half the cylinders.
 --
 Sandy Harris
 sharris@fox.nstn.ca






^ permalink raw reply	[flat|nested] 3+ messages in thread

* fixes in the works
@ 1995-09-02 14:14 jmk
  0 siblings, 0 replies; 3+ messages in thread
From: jmk @ 1995-09-02 14:14 UTC (permalink / raw)


We have fixes in the works (with help from others) for the floppy disc
read and Adaptec 1542CF timeout problems. These will hopefully be available
for pick-up by ftp by this time next week. Source updates based on diffs
to the distribution source should also start appearing soon.

The hard disc partitioning problem will take longer to fix, but I've started
working on it. Unfortunately it means doing a crash-course in booting OS/2
and Linux.

Now, some specific questions:

  > From: pynq@ellis.uchicago.EDU (Jeremy Mathers)
  > Subject: Still can't get the demo version to install on any of my PCs.
  > ...
  >	3) On the laptop (386, 4 megs of memory), it got pretty far into
  >	   the installation, but then ran out of memory.  (Message "No
  >	   physical memory" scrolls by over and over and then the
  >	   machine hung.)

Installing the Plan 9 Distribution states 8Mb of memory is necessary. It might
work with 6Mb.

  > From: pp000199@pop3.INTerramp.COM (Atindra Chaturvedi)
  > Subject: Plan9 on IBM Thinkpad installation
  > ...
  > 1. Plan9 does not boot from 2.84M diskette drive.
  > 2. Using 1.4M diskette drive and 810M IDE drive Plan9 boots but hd0 shows 
  > up with zero heads and zero tracks but the right number of bytes. Later 
  > on in the installation, while copying files I get >Problem copying files 
  > and I/O errors etc.
  > 3. Using 1.4M diskette drive and 540M IDE drive Plan9 boots and hd0 shows 
  > the correct geometry. Later on I have the same problem as in item 2 
  > above.
  > ...
  > I had the same diskette drive boot problem with earlier Linux kernels. 
  > With the newer ones I have to enter the string "floppy=thinkpad" which 
  > configures the kernel for the "inverted DCL bit", after which everything 
  > works well. I suspect the same problem with Plan9 b.com !?!

We don't have a Thinkpad. It's possible that the floppy driver fixes mentioned above
will fix your problem with the 1.44Mb drive. Perhaps someone more familiar with
the Thinkpad hardware can help with the 'inverted DCL' fix.
When a hard disc geometry shows up with 0 heads and 0 tracks but the correct size
it indicates that the drive is in LBA mode (the current b.com does not print values
for heads and tracks in this case).

  > From: staffanu@tcs07.nada.kth.SE (Staffan Ulfberg)
  > Subject: Re: Installation problems:-(
  > ...
  > When b.com is started, it prints info about the hard drives it can
  > find on the screen; in my case:
  > 
  > >hd0: 1669260 sectors 854661120 bytes
  > >hd1: 128 sectors 280559616 bytes	(This is wrong!  The drive is 560MB.)
  > 					(Exactly half of "real capacity"?)
  > ...
  > erroneous.  Now, does anybody know: does b.com use the BIOS to read
  > the disk?  Is it worth trying to change the BIOS addressing of the
  > drive (NORMAL/LBA/LARGE are possible choices).  I already tried

Yes, this is clearly wrong. As mentioned above, these drives are in LBA mode
and the size is taken from the drive's response to an Ident command. Looking
in the driver I can see a path to messing this up if the drive subsequently
times-out doing a read of the supposedly last sector of the disc. I will check
on this.

The first thing b.com does is a BIOS call to set CGA mode. That's the only use
of the BIOS.






^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~1995-09-06  7:48 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
1995-09-06  7:48 fixes in the works Staffan
  -- strict thread matches above, loose matches on Subject: below --
1995-09-04 11:33 sharris
1995-09-02 14:14 jmk

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).