xrayer

Hi, I just tested new bighdd package aslo with scandisk and defrag are working now I use NU2002 anyway but good work. Tested om 500GB SATA drive with 2x192GB FAT32 partitions

>GreyPhound I finally had time to read through entire thread from the beginning. Thaks for your effort with translation of Igor's technical article, I got some better view or windows memory management now. Spasibo! But it's still not clear why in some cases himemx.exe works and himem.sys not. Anyway I'm use this solutoin more than week and didn't found problem. I only sometims hit the limit of swapfile in ramdisk when etiting huge soundtracks in goldwawe, swapfile ran out of space and shoot windows down. If I would had set swap to HDD it could grow to bigger size. But from ramdisk it runs pretty fast! Unfortunatelly memory protection is poor so instead kernel should shoot the app exceeding memory it allow shoot off all system It depends on used program.

I'm not sure about it. When you have small clusters then FAT must be greater and seeking for the specified cluster will take long time when browsing large FAT. And when using the ramdisk for paging file the minimum allocated size is one page - 4kB so smaller clusters don't bring any advantage. Anyway for my ramdisk size and FA16 FS limitation the smallest cluster size is 16kB. I didn't checked what value was automatically set.

No, /c1 parameter only set cluster size I don't care about it let it set automatically. Parameter /t is important.

Hoho, it works So now I got 850MB ramdisk for swapfile c:\dos\xmsdsk.exe 850000 z: /t /y And thanks for thoughts about himem vs himemx. I don't know windows kernel magic how it can behave with different xms mgrs. EDIT: Ou, seems that windows cannot use xmsdsk ramdisk for swappping. I cannot see the drive letter in swap file setting dialog EDIT2: Aha, it can be redirected in system.ini now it works [386Enh] PagingDrive=Z:

Shortly 1) I upgraded io.sys patch to replace himem.sys by himem.exe so rename of the file is needed 2) I can confirm that /max=... parameter IS NOT really needed in config sys (so then I wondew why MS himem.dont work and this himemx works) 3) I did a finetune of MaxPhysPage and ended at value 488FF. Crazy is that when I set 1st itme edit 40000 to 48000 it booted but in VGA mode and display ed error of VGA configuration and open resolution change CPL. But whe I go back an increment MaxPhysPage with smaller steps and do rebooting between every change I was able to reach 488FF (over I got insufficent memory). I tried also xmsdsk again but always got VFAT error BSOD.

Good poit, can we display or log what will io.sys really call? Anyway renaming to himem.exe is the simplest way to be sure. Another thing is if /MAX= is needed at all, StarRiver claims he run without this parameter. Hm, I will need to do some more testing.

Hi, I wrote a small patch yesterday http://www.volny.cz/rayer/os/w98iopat.exe which fix IO.SYS to load himemx.exe instead of himem.sys for safe mode. I don't have problems with crashing in safemode (btw I know about mentioned fixed vcache.vxd). Just copy patch to root dir beside io.sys and run it (make backup before). I didn't much experimended how can be RAM increased, I just try a few values 1.5G 1.25G (physpage=60000 resp 50000) and it hangs during boot so I gave it up, some additional few megs + are not much important for me, I need stable system. I laso tried to eat excesive RAM with XMSDSK but it didn't work for me.

Currently, with only replacing the himem.sys its unable to boot safemode coz windoze then load himem.sys on its own. But I found the strings in IO.SYS and patched and then it run also in safemode. I will program some small patch utility to rewrite it. Installation of himemx.exe is quite esay, just download it, extract himemx.exe to windows directory and replace the line in config.sys to calling this driver. It has regular EXE stub so you can type "himemx.exe" in command line to get help, also there is included documentation and source code Thanks to Japheth maintaining it.

I use 500GB HDD few months and didn't find problems with DOS version of scandisk on large partitions (I have 2x192GB FAT32 there). Under DOS it's used INT13h BIOS extension - no any ESDI_506.PDR or other driver. So I think you don't need to wory about it. Just replace ESDI_506.PDR before windows first boot.

Problem (partially) solved! After some days of messing I finally found HimemX XMS manager freely available here http://japheth.de/Jemm.html which allows you to limit XMS size by /MAX= option. After I limit XMS to 1GB it boots and seems to work fine! Of course windows sees then only 1GB but it's acceptable for me, I don't need then any commercial patch. Here's my lines of system.ini: [386Enh] MaxPhysPage=40000 [vcache] MaxFileCache=261120 MinFileCache=32768 ChunkSize=4096

Problem (partially) solved! After some days of messing I finally found HimemX XMS manager freely available here http://japheth.de/Jemm.html which allows you to limit XMS size by /MAX= option. After I limit XMS to 1GB it boots and seems to work fine! Of course windows sees then only 1GB but it's acceptable for me, I don't need then any commercial patch. Here's my lines of system.ini: [386Enh] MaxPhysPage=40000 [vcache] MaxFileCache=261120 MinFileCache=32768 ChunkSize=4096

Problem solved. I used older driver distributed on Gigabyte installation CD. I will upload it on windows98.ic.cz EDIT: Working RTL8111 driver is available here: http://windows98.ic.cz/net/rtl8111.zip I also did a small update of intel chipset inf files (change in 945.inf), download here: http://windows98.ic.cz/chipset/intelinf.zip

Yes, this is now also my case of troubles: I still use Win98SE (multiboot with DOS, WXP, Linux) and I just upgraded RAM from 1GB to 2GB (my system is 945/ICH7 mb & 7600GS PCIE 256M VGA). I hadn't any problems under 98SE with 1G, as I know about VCACHE issue I limited it in system.ini: MaxFileCache=261120 MinFileCache=32768 ChunkSize=4096 ... MaxPhysPage=40000 As this works fine I didn't expected further problems with adding more mem. I just pull two 1G DIMMs and when 1st booting 98SE it hangs in text mode screen with blinking cursor. I tried safe mode but got insufficent memory message during boot. When I tried normal mode again, it went further but I got only VGA mode 640x480/16 with message "There is a problem with your display settings. The adapter type is incorrect, or the current settings do not work with your hardware [OK]". Then display dialog appeared. I tried to set some higher resolution, reboot but then it hanged again. I search a lot of topic on web and try adjusting vcache size and lowering physpage limit, even some patched vcache.vxd but nothing helped. I also found Loew's memory limit patch demo. 1st problem was that I use czech win98se so patch didn't applied. I had to seek some torrents to get eng. ISO. I backup my system and made clean win98se-eng installation. I needed to modify system.ini physpage to complette installation to solve insuficent memory error. System seemed to work OK with 2GB installed - seeing only 1GB. But after I installed ForceWare 82.69 drivers for my nVidia it become unbootable like my CZ version, hanging during boot. So then I tried Loew's parch 3.2 demo. Now it successfully patched the english files. I booted safe mode and it booted OK and even show up all 2GB. http://rayer.ic.cz/350d/WIN2GBPT.PNG Well but the problem is that in safe mode VGA drivers are not loaded so I cannot see if it will works together like in normal mode. I don't need to use whole 2GB under 98SE (1G is enough) I just want boot it in normal mode with display driver in usable resolution. I also tried to change VGA to my older 6600GT 128MB and same problem. It seems like VGA LFB memory mapping conflicts with something else. But I don't understand why. If I'm right, W9x use for this system area located C0000000-FFFFFFFF. My VGA use LFB at always at D0000000-DFFFFFFF despite I use 1GB or 2GB RAM it's in system area where it fitted well before. Do you think that Loew's patch (full) will solve it or is there other way to use it with limited memory? Can be size of memory limited at very low HW level? I mean e.g. program 945 MCH chipset to temporary disable one DIMM bank to reach same effect like when DIMM removed physically?

Hi, thanks for report. I can confirm that Win98SE works well on P31+ICH7 (GA-P31-DS3L). I used modified INF files by Peter and all HW components was detected. I don't have any /!\ in device manager. On prevoius MB I had one ! for LPC interface. But I have now other 2 problems: 1) I extend RAM 1GB -> 2GB even if I tweak Vcache size in system.ini I cannot boot windows in normal mode and safe mode. Normal mode hangs with blinking cursor on text scree. Safe mode will not boot due to not enough memory 2) Onboard 1Gbit ethernet Realtek 8111B. There's Win98SE driver on Realtek site but after I installed it, windows cannot boot. It hangs with message about corruption of memory protection in textmode. I can boot in safe mode an disable network controller in CPL. Then it boots in normal mode again. Interesting is that when I enable the network controller in CPL windows enables it without restart and I can use network normally unitl next reboot. Maybe I could try older driver.

Yes, I already had asked him.

I think foppy size is derived from its physical geometry, so you should write 80*18*2*512 = 1474560 instead meaningless 1.44*1024*1000 even if the result is same. Anyway I don't bother about MB and MiB, it's not serious difference compared e.g. to nowdays HDD sizes. From your experience you will know where is used MiBs and MBs. But sometimes I wonder too, e.g. I would expect that CF card will use MiBs because it contains flash chips which are surely 2^N sized but they probably save this capacity for badblock relocation and other hidden storage for CF controller.

Hm, interesting. Maybe shutdown problem is only releated to PCI-E graphics? I use this drivers without problems except the shutdown. When I switch to BearWindows Win9X universal VESA driver beta it shutdown normally so it seem the problem is in nvidia drivers.

When I look into control panel/SCSI adapters I can see driver is not running but runs standard ATAPI.SYS. But when i go to control panel/devices and seek for uniata it is set to run on startup and when I tried to startup manually it says driver allready running, so what is true? Probably is not running, fallback to atapi.sys. I tried contact author but no reply.

Hi, does anybody tried to install Win98 on mobos with new intel chipset P31/P35? The major advantage is 1333MHz FSB support for new 45nm Core 2 Duo CPUs. E.g. Gigabyte sells some nice and cheap mobos: http://www.czechcomputer.cz/product.jsp?artno=48787 This mobos combine P31 with well proven ICH7® which is not problem in Win98. I didn't studied differences of P31 and i945 but maybe very similar and INF would be possible to modify easily - add new PCI VID & DID. P35 is used in combination with ICH9® which hasn PATA port. This is done via Jmicron controller and I don't know if there are drivers for it.

Hi, I have installed NT 4.0 on my PC for long years and works fine but I run into trouble with last update of MB and HDD. Now I have mobo with ICH7 (using IDE mode) and new 500GB SATA HDD. I moved partitions from old HDD and all my OSes can boot except NT 4.0. For Win98SE I had to update esdi.pdr to have LBA48 support and works fine too. NT 4.0 just displays a BSOD saying unacessible boot device. But I found if I move them (and edit boot.ini of course) on slave HDD which is less 128GB they can boot. So I tried to install UniATA driver. It was installed successfully but when moved back to 500GB disk it still didn't boot I also tried to replace original NT .sys driver by uniata.sys but it hangs during boot. Anybody was successfull with >128G HDD?

I must give a warning for everybody: according to this intel's document it's clear that since ICH6 ISA DMA is not supported by chipset and then any ISA souncard will not play sound in DOS games, etc. under DOS or Windows 9x. Unfortunately I find this too late...

Yes but this is much aditional work messing with drivers. Pure DOS is da best for this kind of playing. IO port access is quite easy to get but much harder is access to MSR registers or physical memory access under NT-based OS. I wrote my own kernel mode driver for MSR access in mingw32 so I know what I'm talking about. If you are not familiar with windows kernel programming (like me) it's not a piece of cake. And even if it now works through a driver the IOCTL calls adds heavy overhead to this simple operations making it much slower than in DOS/Win9x.

I think there are extract.exe and makecab.exe (not sure of 2nd name) freely available allows you to extract and rebuild CAB files. But I did only extraction not rebuilding new cab.

Oh sure, I don't support this kind of hazards and agree that integrating patch into installation is a good idea. Just wanted to say that it should't be catastrophic if you know what you are doing. About recycled - I have disabled it and when use my favorite file manager to delete fiels even recycle.bin directories are not created. And I didn't try to delete anything from those partitions anyway. BTW does anybody know why Win98 don't like (hang at boot) SATA enhanced mode (enabled in BIOS to use more than 4 IDE drives [2 on PATA, 2 on SATA]). It works fine in DOS and safemode because then BIOS int 13h extension is used and take care about it.