VOICE Home Page: http://www.os2voice.org |
[Previous Page] [Next Page] [Features Index] editor@os2voice.org |
Editor's note: these tips are from OS/2-eComStation users and in some cases
can not be verified by myself. Please heed this as a warning that if you are not
sure about something, don't do it.
I have one of these on my Thinkpad A21p. I found that I needed to do a couple of things. In the cardbus readme, you will find that the Xircom is point enabled and you need to modify IBM2SS14.SYS (BASEDEV=IBM2SS14.SYS /IG0=1) to keep card services from trying to mess with it.
I also had to disable the resource checking for the card. That is one of the parameters you will see if you edit the card in MPTS. PROTOCOL.INI winds up looking like:
[CE3OS2_nif]
DriverName = XIRCOM$
IOADDRESS = 0x320
IRQ = 10
MODE = "IO"
NOCHECK
VERBOSE
I brought it up first on the Win98SE side, locked the settings for address and irq, and transferred those settings to the OS/2 side.
I posted a message yesterday about poor graphics performance with SDD on SMP systems, DIVE in particular. I was talking a bit to Markus yesterday, and he said the problem is that one cannot set the MTRR's for both cpu's on a SMP system. In particular, what one would want to do is to enable write-combining for both/all CPU's. However I thought that hey, this has to be possible, because afaik they do on Linux.. And rightly enough; there is a utility on hobbes (p6k7mtrr_v006.zip) that lets you set the MTRR's for Pentium (I/II/III/Celeron etc.) and K7 (Athlon/Duron/etc) CPU's from Ring3, through a special device driver. The utility has in it's latest versions got a -m parameter, which makes it try to set the MTRR's for _all_ cpu's in a SMP system.
And trust me on this: It works. DIVE performance is like quadrupled, and the system otherwise seems snappier aswell, especially Netscape. Not to mention that the problems I've had in the past with messed-up sound in FLASH animations are now gone (and FLASH performance is _much_ better).. This is like getting a new computer!
So.. Have a look at this, all you SMP people.. And for you, Scitech, maybe this is a (not _the_) way to go?
Best regards,
Eirik Overby
Btw, I noticed that in the WarpMedia FAQ/Readme they mention this too; but afaik they had no idea it would be so effective on SMP systems ;)
Find your PREFS.JS file and add the following line:
This should trick most servers into believing you are using Netscape 4.61user_pref("general.useragent.override", "Mozilla/4.61 [en] (OS/2; U)");
Two possible things:
1 - TCP 4.3 uses TCPCFG2.CMD to start the Java configuration noteboot. Check you are not using Java 1.3 but Java 1.1.7 or 1.1.8.
2 - Your CONFIG.SYS file should have an empty line at the end, otherwise the TCP/IP configuration notebook will not start or will start with empty pages...
I was having this problem today as well.
Using the patchymf program to change the chipset recognition for the driver allowed the driver to see the YMF724 chip in my sound card but the load of the driver with the "/V" switch showed an attempt to allocate IRO x0F rather than the IRQ assigned by the PCI BIOS
I moved the card to another slot and it worked. Check to make sure the PCI slot you are placing the card in is on PCI bus 0. Some machines (with more than 4 PCI slots) use PCI bus 0 and PCI bus 2 (bus 1 is the AGP slot) and the driver may have a problem with that(?). This was not the case with my machine but I have seen this happen with other cards when the driver does not handle the PCI bus number correctly.
Try searching the PS6 disk for the fonts and installing them into OS2 using the OS2 font installer. This took care of the font errors I received when I installed Turbo Tax 2000.
I agree with this, the existing Preview 1 may contains some problems that will carry over if upgraded...
But, thanks to LVM you can try an upgrade without losing anything (backups are STILL a good idea) if you have spare Harddrive space...
Clone boot partition: (for install, or testing)
Using a freshly booted system:
1. Start LVM
2. Create a new bootable volume, give it any letter (x:)
3. Close and save LVM. (reboot if asked)
4. Format X: in the filesystem of your boot vol. (HPFS)
5. Copy the contents of your existing system(C:) to X: XCOPY C:\* X:\ /H/O/T/S/E/R/V
6. Start LVM
7. in LVM, hide the current boot vol (C:)
8. Change the X: volume's letter to C:
9. Reboot and test C:
(it should be a clone of the other)
Test the upgrade against this "new" C:
to restore the old C: (if it doesn't work from bootmanager anyway)
1. Start LVM
2. hide/change letter of the new C:
3. assign C: to the old C: volume again.
4. Save and close LVM
5. Reboot.
A google search turned up at least three other threads on this topic (PMWIN.1500 and NS 4.61, going back to Feb-2000). This fragment sounds promising:
"IBM TECH advised me that when I had this problem, it was because the Watchcat DLL's were in front of the Netscape DLL's in the Libpath statement of config.sys. swapping them around totally fixed it. YMMV"
Hmmm. The first few items in my LIBPATH are: LIBPATH=E:\IBMLAN\NETLIB;E:\MUGLIB\DLL;.;D:\NS4\PROGRAM;... Make sure that you don't have some odd app's dir ahead of Netscape's dir, in your LIBPATH? And Lorne had this to say: "The program you ran is calling some routine in PMBIDI.DLL that is attempting to load a routine in PMWIN.DLL.
According to the information displayed by EXEHDR the ordinal entry point PMWIN.1500 is called WINSETLANGINFO and it is forwarded to PMMERGE.3500. The error occurs if the routine is not present in PMMERGE.DLL. Perhaps the version of PMMERGE you are running got back levelled somehow and is missing the routine. (The one on my machine has it, AFAICT)."
Whenever I have a hung app and use the focus change feature of Warp to regain control (repeated[CTL]-[Esc] alternating with[Alt]-[Tab] , I end up with the following "Swith To" window which only offers the option to hit[Enter] to end the application. Frequently though I don't want to end the application. By experimentation I discovered that hitting the[Esc] key will close the "Switch To" window and not close the application.