|
From: Michael D. <Fre...@de...> - 2005-06-24 13:41:30
|
If something loads in upper memory with the NOEMS option and not without, it means that the EMS page frame eats up too much upper memory. Although upper memory totals may look large enough in a memory report, the EMS page frame often makes the amounts discontiguous. You need sufficient contiguous upper memory to load a single application. Moving the EMS page frame to a different block of 64K upper memory may also work, although it has rules about the alignments allowed. At 02:20 PM 6/24/2005 +0200, Bernd Blaauw wrote: >Hi Arachne list, > >EMM386 v2.03 has been released by Michael Devore and Tom Ehlert. >It's available from >ftp://ftp.devoresoftware.com/downloads/emm386/emms203.zip (source) >ftp://ftp.devoresoftware.com/downloads/emm386/emmx203.zip (executable) > >new versions are generally announced on the FreeDOS lists, >reachable from http://www.freedos.org/freedos/lists/ >(freedos-devel, freedos-kernel, freedos-users) > >Can you give more information about the testing setup? >*kernel revision (VER /R) >*shell version (VER /R) >*contents of config.sys (or fdconfig.sys) >*contents of autoexec.bat >*MEM /C directed to a textfile >*version of himem.exe (1.15 as mentioned) >*version of emm386.exe (1.15 as mentioned) > >might be an idea for me to make a batchfile which does all of this >automated :) > >regards, > >Bernd Blaauw > > >X-Account-Key: account1 >Return-Path: <cm...@ns...> >Received: from mx7.home.nl ([213.51.128.150]) by mail8-sh.home.nl > (InterMail vM.5.01.06.05 201-253-122-130-105-20030824) with ESMTP > id <200...@mx...> > for <bb...@ho...>; Fri, 24 Jun 2005 03:09:53 +0200 >Received: from mk-smarthost-4.mail.uk.tiscali.com ([212.74.114.40]:2466) > by mx7.home.nl with esmtp (Exim 4.30) > id 1Dlchc-0003CK-R4 > for bb...@ho...; Fri, 24 Jun 2005 03:09:52 +0200 >Received: from dial-80-47-199-100.th.lond.access.as9105.com >([80.47.199.100]:1025) > by mk-smarthost-4.mail.uk.tiscali.com with smtp (Exim 4.50) > id 1Dlcha-000O1J-40; Fri, 24 Jun 2005 02:09:51 +0100 >To: ara...@co... >Cc: bb...@ho... >Subject: LSPPP with FreeDOS EMM386.EXE >Date: Fri, 24 Jun 2005 01:44:12 -0400 >Reply-To: cm...@ns... >Message-ID: <s25...@ns...> >X-Mailer: NOS-BOX 2.05 >X-AtHome-MailScanner-Information: Neem contact op met su...@ho... voor >meer informatie >X-AtHome-MailScanner: Found to be clean > > >Hi Guys. > >Spent most of this evening (and into the wee-hours) identifying a problem >with what I thought was LSPPP versions 1.0 and 1.01 > >For many of you, this probably wont apply as its FreeDOS specific, e.g. >you can probably ignore this if your using the likes of DR-DOS etc. > >Its looking more like its an EMM386.EXE problem rather than an LSPPP one. >The version of EMM386.EXE I'm using shows up as V1.15 dated 19th Feb 2005 >(FreeDOS). > >When EMM=2048 is defined, then LSPPP wont load high despite having more >than enough free Upper Memory. However if NOEMS is defined, then LSPPP >loads high OK. e.g. > >device=emm386.exe X=TEST NOEMS /v then LSPPP loads high fine but >device=emm386.exe X=TEST EMM=2048 /v LSPPP doesn't load high > >Whilst LSPPP still loads into conventional memory, the problem is that >if your using a UKA_PPP/Yarn combination, then LSPPP is loaded ontop of >UKA_PPP when loaded into conventional memory, so when you switch to >Arachne, despite there being enough conventional memory after UKA_PPP >drops out, that memory is fragmented and results in Arachne having >insufficient memory to load (largest executable (contiguous) memory is >too low). > >I've tried posting to fre...@li... but my postings >bounce back (any ideas as to why this might happen? (I'm not that familiar >with News)). > >I've cc'd this to one of the guys listed in the emm386 /? page in the >hope that it might get passed on to the relevant parties. > >Regards. > >CJ |