OS2 World.Com Forum

Subject  :  Apsoft - sspcic.sys -topic100
Author  :  melf mikaelelf@os2ug.se
Date  :  21 Mar, 2007 on 11:36
Is there somebody with experience of laptop Toshiba tecra 8100?

I'm trying to get pcmcia work on that computer. I'm using sspcic.sys which is supporting the topic100 socket. With the usual options (/VS:auto /AP:on /EC:auto /II:off /MX:off) os2dasd.dmd refuses to load and system just hangs there. Adding options /MH:0xFB5F /FS:0xFB5F make driver report sockets on boot up and os2dasd.dmd loads, but system crashes (trap06) just before desktop will be visible.

If I use the option /RA:off then the driver find sockets but in the same time says that it can't implement /RA:off. Anyway system starts up, sockets are found but system crashes (trap) when trying to open PnP for PCMCIA.

Well, why am I trying these options? The first one was from Apsofts site and is ment to resolve an irq resource conflict. Unfortunately I don't know how to modify these opitons to suit my own system. The other option I just tried, I really don't know why. Does it seem confused? Yes it is for me too!

From pcisniffer I get that socket base memory is 00000000h, that is the only fact I have.

My hope is that some sspcic skillful person know what to do or that someone has a tecra 8100 and knows the problem.


/Mikael

Subject  :  Re:Apsoft - sspcic.sys -topic100
Author  :  Radek hajek@vuv.cz
Date  :  21 Mar, 2007 on 12:09
Int 06h is an invalid operation exception. Possibilities:

- Wild run.
- The driver is using an instruction set not known to the Toshiba's CPU (for example, Pentium instructions on a 486).
- OS/2 is talking nonsense (less possible).

Therefore, what is the Toshiba's CPU and what CPU is required by the driver?


Subject  :  Re:Apsoft - sspcic.sys -topic100
Author  :  melf mikaelelf@os2ug.se
Date  :  21 Mar, 2007 on 14:47
Hello Radek,

The CPU is an Intel pentium II, but I can't find any specification in Apsofts papers that the driver don't work with a certain CPU. The driver support lots of sockets, both older and more recent ones. I think this driver has been existing for a lot of years and the last update was 2004. My driver should be from that year too. So I guess a PII should be of no problem. I use it right now on an AthlonXP 2600+.

But what about that os2dasd.dmd not loading? I'm using ecs beta4, ACPI disabled.

---
/Mikael


Subject  :  Re:Apsoft - sspcic.sys -topic100
Author  :  Radek hajek@vuv.cz
Date  :  21 Mar, 2007 on 15:49
PII should be more than enough so that it isn't a problem of an old CPU. It still can be a wild run - a usual result of joint effort of a corrupted stack an the 'ret' or 'iret' instruction. The instruction pops bogus (instead of a return address) from the stack and passes the control to the hell. An invalid opcode exception follows soon. This can also prevent other drivers from loading - it suffices that the faulty driver loads first and causes enough problems in the system.

Let us suppose the wild run. Then, the /RA:off option bypasses the location where the stack gets corrupted somehow, nevertheless, the stack gets corrupted later, when you try to open the PCMCIA.

How the stack can get corrupted? IMO, it can be a problem of a Toshiba hardware. The driver gets a request packet from the system and tries to do what it is wanted. The hardware can be "specific" and response in a way the driver does not expect. The driver gets confused and pushes something on the stack what it forgets to pop from the stack later. The problems are done.

If it is so then I see only one chance: playing with the driver parameters and trying to avoid the place where the stack gets corrupted. In the other words, prophesing from the parameter setting and the moment when the system crashes. Not a good perspective Sorry I cannot help more.


Subject  :  Re:Apsoft - sspcic.sys -topic100
Author  :  melf mikaelelf@os2ug.se
Date  :  21 Mar, 2007 on 16:52
Thanks Radek for your comments. I hope you are wrong, but it sounds logical. I will see for how long time I can stand trial and error with driver parameters .

---
/Mikael

Subject  :  Re:Apsoft - sspcic.sys -topic100
Author  :  melf mikaelelf@os2ug.se
Date  :  25 Mar, 2007 on 22:20
I have had no success so far, but I seem to have read a "d" for a "6", so the trap was "trap0d", which I think makes the perspective better?

---
/Mikael

Subject  :  Re:Apsoft - sspcic.sys -topic100
Author  :  Radek hajek@vuv.cz
Date  :  26 Mar, 2007 on 09:51
A bit better, but only a bit int 0Dh is an ordinary general protection fault. It can be caused by dozens of causes. The usual ones are: accessing memory not owned by the driver, trying to deallocate a NULL pointer, writing into a non-writable segment, accessing a segment without enough access rights, or accessing memory beyond the end of a segment. int 0Dh is one of the most frequent faults.

Therefore, trying to understand the cause of the fault is futile for "non-insiders", who don't know the structure of the driver. Only Apsoft can help here.


Subject  :  Re:Apsoft - sspcic.sys -topic100
Author  :  Glenn
Date  :  26 Mar, 2007 on 18:03
how much RAM do you have on the PC ? The APSoft driver may have some issues with more than 1GB of RAM. Usually it just does not load correctly, but a bug may hide another one...

Subject  :  Re:Apsoft - sspcic.sys -topic100
Author  :  melf mikaelelf@os2ug.se
Date  :  26 Mar, 2007 on 18:48
It is not much RAM, 128MB, thats all. But I understand from their support side and from what I've red on the web, that som computers and sockets needs "special" treatment. I have seen that sockets on Tecra have been hard to configure also for Linux people. So I guess it'll work with the right parameters.

---
/Mikael

Subject  :  Re:Apsoft - sspcic.sys -topic100
Author  :  melf mikaelelf@os2ug.se
Date  :  27 Mar, 2007 on 01:23
Well, holy shit! Sometimes the gods are with you. I just tested the ibmtos2.sys for toshiba laptops included in ecs. It had been disabled during installation so I thought it didn't work, but it really worked straight away, without any parameters. Sockets were found and installed, only thing was xwlan stayed in configuring mode. That seemed to depend on an irq conflict, so reserving irq 3 made it all work. Really can't believe my eyes!!!

Next thing is sound, its detected and installed but not working..but well a good thing is that graphics is fully supported by SNAP.

Written on the same Toshiba

---
/Mikael


Powered by UltraBoard 2000 <www.ub2k.com>