Re: [NTLK] About System Patches

From: Paul Guyot (pguyot_at_kallisys.net)
Date: Thu Dec 27 2001 - 14:11:14 EST


Stephanie Maks <steph_at_maksystems.com> wrote:

>I was tinkering around with my MP2100 yesterday, doing things I probably
>shouldn't, and managed to blow away the system patch in it.

I had such a problem before, but I think it wiped the internal store as well.

>Formerly, it
>displayed in the memory info: 2.1 (717260) -1
>
>And after I messed it up, it displayed: 2.1 (717006) -0 in the same spot.

I think that "1" is the patch level (always 1 AFAIK). It might also
be the patch number zero-based (and 0 if you don't have any) as
717260 (and all 2.1 patches I had a look at) patch both the base ROM
and REX 1.

>I read in the FAQ about the system patches, and in the Readme that
>accompanied the 717260 download from Apple, it said that the MP2100 =
>shipped with 717260 installed. But I guess they actually just
>installed the patch at the factory, as opposed to having new ROM
>chips.

Exactly.

>Another thing I noticed which was interesting (to me at least) was that =
>the system RAM info was Higher under 717006 than it was 717260. With 717260 =
>the memory info slip displays something in the range of 3980K of System Ram.
>But under 717006 it said 4042K of System Ram (Or something in that area, I
>didn't write it down. It was over 4000 though.) So the System Patch =
>loads itself into DRAM I guess? Or writes into the Flash, but on
>restart loads into RAM?

I can see two causes for this difference.
First, when the system boots up, it loads the system patch into the
DRAM. It also modifies the MMU accordingly.
717260 and 710031 are 10 pages as follows (a page is 4 KB)
Base ROM:
2 pages of MMU
6 pages of patches
REX 1:
1 page of MMU
1 page of patches

So technically, having a system patch decreases the available memory
by 6 + 1 pages (28 KB).

However, the MMU pages for the system patch are probably in the ROM,
so it might require 10 pages in total.

Since MMU table pages cannot be seen while the MMU is active, I
suspect the RAM Installed total to not count them (and therefore
explain, with other factors such as system patches, why we don't have
4096).

The second cause is that the 717260 patch (in fact any patch >=
717246 I think) changes the way the Newton sees the memory. This is
what you noticed here:

>One other thing I have noticed, is that the heap displayed by Avi's =
>Backdrop and Dashboard has dropped from around 400K down to about
>150K. Though as I said, the indicated System Ram still shows the
>full 4MB.

This value is the NewtonScript heap. The 717246, 717260 and 710031
patches allocates more for this heap. I don't know exactly how and
how much, though, but it increases indeed from something around 150
KB to something around 400 KB. I think it changes other things in the
memory management.

>Now a question about the system patches. I have read that there was a
>patch from Apple numbered 717246 that was identical to the 717260
>patch, except that supposedly it displayed '(c) Newton Inc' instead
>of '(c) Apple
>Computers'. People appearantly wanted to use that patch instead of 260 so
>their Message Pad could say '(c)Newton Inc'. So here's my silly question.
>Where does it say (c) Apple Computers? Not in the memory info. Not
>in the extras drawer. I don't have the 717246 patch, though I've
>been scouring the
>web for it. I'm just wondering if it makes any difference, considering I
>can't figure out where the (c) Apple.. message is to begin with.

I think it's only in the login screen (put a password to have the
time to read it or use NPDS GIF Server private mode :).

If you have any other question regarding system patches, don't
hesitate to ask. I prefer to answer them than knowing that you're
trying to wipe it again <g>.

Paul

-- 
Home page: http://www.kallisys.com/
Newton-powered WebServer: http://newt.dyndns.org:8080/

-- This is the Newtontalk mailinglist - http://www.newtontalk.net To unsubscribe or manage: visit the above link or mailto:newtontalk-request_at_newtontalk.net?Subject=unsubscribe



This archive was generated by hypermail 2.1.2 : Wed Jan 02 2002 - 12:02:08 EST