9fans - fans of the OS Plan 9 from Bell Labs
 help / color / mirror / Atom feed
* Re: [9fans] missing memory
@ 2002-01-21 18:04 jmk
  0 siblings, 0 replies; 4+ messages in thread
From: jmk @ 2002-01-21 18:04 UTC (permalink / raw)
  To: 9fans

[-- Attachment #1: Type: text/plain, Size: 316 bytes --]

IF you look int the ramscan routine you will see it reads NVRAM locations 0x18 and 0x17.
It would be interesting to know what is in there as that will limit the amount of physical
memory the routine scans.

Put
	*maxmem=0xFF79000
in your plan9.ini and see if the whole memory is detected (and can be used).


[-- Attachment #2: Type: message/rfc822, Size: 3895 bytes --]

From: iwanek@nttdata.co.jp
To: 9fans@cse.psu.edu
Subject: [9fans] missing memory
Date: Tue, 22 Jan 2002 02:07:48 +0900
Message-ID: <50F591D39557D511B12C0090274DCEBC03B6754D@noanet03.noanet.nttdata.co.jp>

Hello 9fans,

About a month ago, I reported a problem of getting the

    panic: exception/interrupt 14

from floppy installation. Not quite sure what I did differently (BIOS
power management setting, maybe), somehow I got through it. Now
I have new problem and would like to ask the list for help again.

My PC is a Dell Dimension 8200. This monster has a 1.6GHz
Pentium4, i850 chipset (danger here?) and 256MB RIMM. 9pcdisk
detects only 64MB of RAM. Setting MEMDEBUG to 1 in
/sys/src/9/pc/memory.c, I get

    maxmem 30000000 D0000000
    maxpa = F9DC -> 3F77000, maxpa1 = F9DC maxpa2 = 280
    00006000 00099C00 0009FC00	# rmapram
    001F9000 03E07000 04000000	# rmapram
    000D0000 00020000 000F0000	# rmapumb
    04000000 FC000000 00000000	# rmapupa

from memdebug(). I guess the range 04000000-10000000 should
be RAM but is classified as UPA (what is it anyway, and its size
FC000000?). Also I don't have rmapumbrw at all. Strange.

I read ramscan() code, and although many details of dealing with
MMU are beyond me, I can't tell why this doesn't work. If you
know anything about how to fix this problem, please let me know.

By the way, linux 2.4.something recognizes all 256MB. Here is
its idea of memory map:

    00000000000000000 - 00000000000A0000 (usable)
    000000000000F0000 - 0000000000100000 (reserved)
    00000000000100000 - 000000000FF77000 (usable)
    0000000000FF77000 - 000000000FF79000 (ACPI NVS)
    0000000000FF79000 - 0000000010000000 (reserved)
    000000000FEC00000 - 00000000FEC10000 (reserved)
    000000000FEE00000 - 00000000FEE10000 (reserved)
    000000000FFB00000 - 0000000100000000 (reserved)

Note that linux found out about this using BIOS's ACPI service
(INT 15h, AX=E820h). I think you can use this only from *real*
mode (horror...).

- kazumi

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [9fans] missing memory
  2002-01-21 18:30 jmk
@ 2002-01-21 18:38 ` Ronald G Minnich
  0 siblings, 0 replies; 4+ messages in thread
From: Ronald G Minnich @ 2002-01-21 18:38 UTC (permalink / raw)
  To: 9fans

I need to talk to you plan 9 guys so we can sure that you can parse
linuxbios tables (i.e. Plan 9booted from flash, the info is not in NVRAM,
there is no BIOS as you're used to, but there is an in-memory table from
LinuxBIOS) that we will be passing to the kernel. Once we get that code
right I need to know who to talk to.

ron



^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [9fans] missing memory
@ 2002-01-21 18:30 jmk
  2002-01-21 18:38 ` Ronald G Minnich
  0 siblings, 1 reply; 4+ messages in thread
From: jmk @ 2002-01-21 18:30 UTC (permalink / raw)
  To: 9fans

Oops, I just remembered the nvram locations are printed out in your message
by MEMDEBUG, it's 'maxpa' below.

    maxpa = F9DC -> 3F77000, maxpa1 = F9DC maxpa2 = 280

The convention is that this location is 0 if the physical memory size is
not known by the bios or is >64MB.

I guess we need to rethink how we do this, it's old code and hardware
moves on.


^ permalink raw reply	[flat|nested] 4+ messages in thread

* [9fans] missing memory
@ 2002-01-21 17:07 iwanek
  0 siblings, 0 replies; 4+ messages in thread
From: iwanek @ 2002-01-21 17:07 UTC (permalink / raw)
  To: 9fans

Hello 9fans,

About a month ago, I reported a problem of getting the

    panic: exception/interrupt 14

from floppy installation. Not quite sure what I did differently (BIOS
power management setting, maybe), somehow I got through it. Now
I have new problem and would like to ask the list for help again.

My PC is a Dell Dimension 8200. This monster has a 1.6GHz
Pentium4, i850 chipset (danger here?) and 256MB RIMM. 9pcdisk
detects only 64MB of RAM. Setting MEMDEBUG to 1 in
/sys/src/9/pc/memory.c, I get

    maxmem 30000000 D0000000
    maxpa = F9DC -> 3F77000, maxpa1 = F9DC maxpa2 = 280
    00006000 00099C00 0009FC00	# rmapram
    001F9000 03E07000 04000000	# rmapram
    000D0000 00020000 000F0000	# rmapumb
    04000000 FC000000 00000000	# rmapupa

from memdebug(). I guess the range 04000000-10000000 should
be RAM but is classified as UPA (what is it anyway, and its size
FC000000?). Also I don't have rmapumbrw at all. Strange.

I read ramscan() code, and although many details of dealing with
MMU are beyond me, I can't tell why this doesn't work. If you
know anything about how to fix this problem, please let me know.

By the way, linux 2.4.something recognizes all 256MB. Here is
its idea of memory map:

    00000000000000000 - 00000000000A0000 (usable)
    000000000000F0000 - 0000000000100000 (reserved)
    00000000000100000 - 000000000FF77000 (usable)
    0000000000FF77000 - 000000000FF79000 (ACPI NVS)
    0000000000FF79000 - 0000000010000000 (reserved)
    000000000FEC00000 - 00000000FEC10000 (reserved)
    000000000FEE00000 - 00000000FEE10000 (reserved)
    000000000FFB00000 - 0000000100000000 (reserved)

Note that linux found out about this using BIOS's ACPI service
(INT 15h, AX=E820h). I think you can use this only from *real*
mode (horror...).

- kazumi


^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2002-01-21 18:38 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2002-01-21 18:04 [9fans] missing memory jmk
  -- strict thread matches above, loose matches on Subject: below --
2002-01-21 18:30 jmk
2002-01-21 18:38 ` Ronald G Minnich
2002-01-21 17:07 iwanek

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).