VOGONS


HWiNFO for DOS resurrected !

Topic actions

Reply 501 of 881, by feipoa

User metadata
Rank l33t++
Rank
l33t++

I'll be on this system for awhile still. I need to solve that floppy copy overwrite issue (which is due to a BIOS setting); need to test the system with a sound card (previous testing with BL3 on other boards failed when using sound, but this board seems happy with the BL3); need to try other BL3 adapters; need to try NT4/3.51; maybe try a DMA SCSI controller - those also sometimes cause issues with these upgrades.

It would be nice to setup a boxed system around the BL3, but the darn BL3 module covers the VLB slots, so I'd have to find or create a PGA-132 interposer board which rotates the pin locations by 180 degrees.

Plan your life wisely, you'll be dead before you know it.

Reply 502 of 881, by feipoa

User metadata
Rank l33t++
Rank
l33t++

Here's a photo of the testbed. Any idea how I can use a VLB graphics card in this system? Raising the CPU up with PGA sockets 15 cm is not an option!

PANDA_Buffalo_BL3_testbed.jpg
Filename
PANDA_Buffalo_BL3_testbed.jpg
File size
1.43 MiB
Views
1508 views
File license
CC-BY-4.0

Plan your life wisely, you'll be dead before you know it.

Reply 504 of 881, by hasat

User metadata
Rank Newbie
Rank
Newbie
Mumak wrote on 2020-05-09, 20:29:
hasat wrote on 2020-05-09, 13:14:
Mumak wrote on 2020-05-08, 18:06:
Thanks. Here a new build attached. This should recognize the graphics adapter as SiS 86C201 and hopefully fix the other bug too. […]
Show full quote

Thanks. Here a new build attached.
This should recognize the graphics adapter as SiS 86C201 and hopefully fix the other bug too.
Would be great if you could run it on the HP Vectra too. It will most probably work same as before (no need to make tests with and w/o EMM) , but I've added more information into the DBG file that could tell me how to fix the SIO hang.
HWINF610.ZIP

Here is new test on ASUS SP97-V motherboard with integrated video card in SiS chipset: name is OK, but reported video memory is wrong - HWINFO still reporting 2048 kB memory regardless of what is set in BIOS (1 MB, 2 MB or 4 MB).
VBE2 revision string is also still wrong: i discovered new thing, text in this field is changing depending on what is loaded in memory (for example i run HWINFO from DOS prompt and from Volkov Commander). Two examples are in video.log file. Very interesting bug! 😁

Thanks. I've got an update for the VBE revision issue, but don't know how else to report the shared video memory size. Will post once you provide new data for the HP Vectra.

Think i found, how to detect shared memory size for integrated VGA. I read datasheet for SiS 5598 chipset and found here registers for setting memory size. Can you look at this? Page 151 in this PDF http://66.113.161.23/~mR_Slug/pub/datasheets/ … /SIS/5597-8.pdf

BTW on this web is a lot of datasheets for old chipsets.

Reply 505 of 881, by Mumak

User metadata
Rank Oldbie
Rank
Oldbie
hasat wrote on 2020-05-21, 16:14:
Mumak wrote on 2020-05-09, 20:29:
hasat wrote on 2020-05-09, 13:14:

Here is new test on ASUS SP97-V motherboard with integrated video card in SiS chipset: name is OK, but reported video memory is wrong - HWINFO still reporting 2048 kB memory regardless of what is set in BIOS (1 MB, 2 MB or 4 MB).
VBE2 revision string is also still wrong: i discovered new thing, text in this field is changing depending on what is loaded in memory (for example i run HWINFO from DOS prompt and from Volkov Commander). Two examples are in video.log file. Very interesting bug! 😁

Thanks. I've got an update for the VBE revision issue, but don't know how else to report the shared video memory size. Will post once you provide new data for the HP Vectra.

Think i found, how to detect shared memory size for integrated VGA. I read datasheet for SiS 5598 chipset and found here registers for setting memory size. Can you look at this? Page 151 in this PDF http://66.113.161.23/~mR_Slug/pub/datasheets/ … /SIS/5597-8.pdf

BTW on this web is a lot of datasheets for old chipsets.

Thanks, this is interesting. Have you maybe verified if that register also matches the BIOS setting?

Reply 506 of 881, by ALEKS

User metadata
Rank Newbie
Rank
Newbie

Hi Martin,

I've been using HWiNFO since 1998 or so and I always liked it. You have some good programming skills!
However, I recently downloaded HWiNFO v6.1.1 and since I am using an unusual computer, please check this:

hwinfo.jpg
Filename
hwinfo.jpg
File size
140.75 KiB
Views
1382 views
File license
Public domain

The controller display appears to be weird. The main computer is an ISA single board computer, running on a DIY system backplane. Also the IDE controller is my own design and implementation.
I suppose there is no way of detecting an IDE interface adapter that simply buffers the ISA BUS while decoding certain addresses and additional signals to address the drives.
But the idea is that the Intel on-board controller is disabled in BIOS. In my opinion, it should not appear if I have a different master controller installed in the system.

Here's my DIY controller (2 x IDE interfaces, and some other peripherals which, by the way, are detected correctly):

isa-io-interface-pcba12.jpg
Filename
isa-io-interface-pcba12.jpg
File size
283.59 KiB
Views
1382 views
File license
GPL-2.0-or-later

And the full report:

Filename
HWINFO.LOG
File size
64.49 KiB
Downloads
48 downloads
File license
Public domain

Also I've spotted a minor unusual thing, while running 4 compact-flash cards on my controller, and using XTIDE, an inexistent SCSI controller is detected by HWiNFO.
But I'll try to take pictures and describe the issues when I'll have some free time.

Also, somebody mentioned that NSSI has some extra features that HWiNFO doesn't have. What I'd like to see in HWiNFO is a list of adapter card ROMs, if present. I programmed a variety of option ROMs and even my IDE controller has two self-programmed ROMs. I've written my own utilities to program and read-back ROMs but it would've been nice to see a list of ROMs (Adaptec cards, Network cards, Cache controllers, other weird or DIY hardware with embedded ROM firmware). Searching for ROMs is pretty simple to implement and if you consider this a good addition, I'd be pleased to see it done.

Thanks for all the great work!
A.

TX486DLC / 40 MHz | 32 Mb RAM | 16-bit ISA Backplane | Tseng Labs ET4000/W32i 2 Mb | I/O Interface | Audio Interface | PC Speaker Driver | Signal View Interface
3.5" & 5.25" FDD | 4 x 512 Mb CF | HP 82341D Interface | Intel EtherExpress 16

Reply 507 of 881, by Mumak

User metadata
Rank Oldbie
Rank
Oldbie
ALEKS wrote on 2020-06-07, 07:15:
Hi Martin, […]
Show full quote

Hi Martin,

I've been using HWiNFO since 1998 or so and I always liked it. You have some good programming skills!
However, I recently downloaded HWiNFO v6.1.1 and since I am using an unusual computer, please check this:

hwinfo.jpg

The controller display appears to be weird. The main computer is an ISA single board computer, running on a DIY system backplane. Also the IDE controller is my own design and implementation.
I suppose there is no way of detecting an IDE interface adapter that simply buffers the ISA BUS while decoding certain addresses and additional signals to address the drives.
But the idea is that the Intel on-board controller is disabled in BIOS. In my opinion, it should not appear if I have a different master controller installed in the system.

Here's my DIY controller (2 x IDE interfaces, and some other peripherals which, by the way, are detected correctly):

isa-io-interface-pcba12.jpg

And the full report:

HWINFO.LOG

Also I've spotted a minor unusual thing, while running 4 compact-flash cards on my controller, and using XTIDE, an inexistent SCSI controller is detected by HWiNFO.
But I'll try to take pictures and describe the issues when I'll have some free time.

Also, somebody mentioned that NSSI has some extra features that HWiNFO doesn't have. What I'd like to see in HWiNFO is a list of adapter card ROMs, if present. I programmed a variety of option ROMs and even my IDE controller has two self-programmed ROMs. I've written my own utilities to program and read-back ROMs but it would've been nice to see a list of ROMs (Adaptec cards, Network cards, Cache controllers, other weird or DIY hardware with embedded ROM firmware). Searching for ROMs is pretty simple to implement and if you consider this a good addition, I'd be pleased to see it done.

Thanks for all the great work!
A.

Like you assumed, detecting ISA/VLB IDE controllers is not easy and if they don't feature advanced registers or some special ones that would allow detection, it's not possible.
Also the link between a controller and drive is not easy to determine, in many cases this is not possible to do precisely, so HWiNFO makes certain assumptions that sometimes don't work properly.

HWiNFO does detect OROMs. In case of PCI devices this should be universal and work well. But for ISA/VLB only certain known ones are detected. If you can send me dumps of OROMs that are not detected, I will add them.

Reply 508 of 881, by Grzyb

User metadata
Rank Oldbie
Rank
Oldbie
Mumak wrote on 2020-06-07, 07:45:

HWiNFO does detect OROMs. In case of PCI devices this should be universal and work well. But for ISA/VLB only certain known ones are detected.

Isn't there some standard procedure the main BIOS uses to detect and initialize ROMs with BIOS extensions?

Żywotwór planetarny, jego gnijące błoto, jest świtem egzystencji, fazą wstępną, i wyłoni się z krwawych ciastomózgowych miedź miłująca...

Reply 509 of 881, by ALEKS

User metadata
Rank Newbie
Rank
Newbie

The procedure is to check the standard mapped upper memory area in 2 K increments for existing code. Each Option ROM has a header that, amongst other things (jump to code start, and Option ROM identifier) specifies the ROM size in 512 bytes blocks. Then you need to perform an 8-bit checksum of all data contained in that ROM and compare against the last byte of the whole 512 * ROM_BLOCKS. If the values match, then we can assume we found a valid ROM.
This is what the BIOS does when scanning the ROM area during POST.

That procedure applies to ISA ROMs. I am not sure about where PCI ROMs are mapped though, as I never programmed one of those.

I don't know how HWiNFO does the detection though.

TX486DLC / 40 MHz | 32 Mb RAM | 16-bit ISA Backplane | Tseng Labs ET4000/W32i 2 Mb | I/O Interface | Audio Interface | PC Speaker Driver | Signal View Interface
3.5" & 5.25" FDD | 4 x 512 Mb CF | HP 82341D Interface | Intel EtherExpress 16

Reply 510 of 881, by Mumak

User metadata
Rank Oldbie
Rank
Oldbie

HWiNFO does the same standard test and it should work for all OROMs since ISA days up-to latest Legacy (Non-UEFI) systems.
This is no problem, but once you find an OROM you don't know to which device it belongs to when it's Non-PCI. So one needs to do further device-specific checks to determine that. That's why I asked for the dumps of such OROMs.

Reply 511 of 881, by ALEKS

User metadata
Rank Newbie
Rank
Newbie

I am attaching my own ROM for the controller above. It is an 8 K ROM.
However at the moment it is not completely finished. I still need to finalize the INT 13h drive accessing interrupt service routines and I still have about 50% of the ROM capacity available.
The ROM BIOS text-mode UI is done already. Detection of drives is also done.

The string I/O-IF ROM #0 (padded left right by single null character, and starts at 7th hex byte) in the ROM header, however, is final, both in terms of offset and characters. If you could relate this to the ISA I/O Interface name, it would be great.

Either I don't find any reference to this ROM in HWiNFO at the moment, or the program does not detect it. In my system, it is mapped at D000:0000.
If I look under ISA and other devices section, I can see only the Intel 82595 Ethernet Adapter. Which by the way is not based on 82595 but on 82586 LAN coprocessor.

Thanks,

Attachments

TX486DLC / 40 MHz | 32 Mb RAM | 16-bit ISA Backplane | Tseng Labs ET4000/W32i 2 Mb | I/O Interface | Audio Interface | PC Speaker Driver | Signal View Interface
3.5" & 5.25" FDD | 4 x 512 Mb CF | HP 82341D Interface | Intel EtherExpress 16

Reply 512 of 881, by Mumak

User metadata
Rank Oldbie
Rank
Oldbie
ALEKS wrote on 2020-06-07, 15:50:
I am attaching my own ROM for the controller above. It is an 8 K ROM. However at the moment it is not completely finished. I sti […]
Show full quote

I am attaching my own ROM for the controller above. It is an 8 K ROM.
However at the moment it is not completely finished. I still need to finalize the INT 13h drive accessing interrupt service routines and I still have about 50% of the ROM capacity available.
The ROM BIOS text-mode UI is done already. Detection of drives is also done.

The string I/O-IF ROM #0 (padded left right by single null character, and starts at 7th hex byte) in the ROM header, however, is final, both in terms of offset and characters. If you could relate this to the ISA I/O Interface name, it would be great.

Either I don't find any reference to this ROM in HWiNFO at the moment, or the program does not detect it. In my system, it is mapped at D000:0000.
If I look under ISA and other devices section, I can see only the Intel 82595 Ethernet Adapter. Which by the way is not based on 82595 but on 82586 LAN coprocessor.

Thanks,

Thanks. I will use the "I/O-IF ROM" string for recognition. How should I officially call the adapter? 😀

Reply 513 of 881, by ALEKS

User metadata
Rank Newbie
Rank
Newbie

Thanks, Martin!

I named it simply: ISA I/O Interface
And model number: 2486-IOIF-301
(just for the fun of it, the project is on my site at: http://www.alexandrugroza.ro/microelectronics … face/index.html)

Since I won't be designing any other IDE or I/O controller anytime soon, we can say it's a final name and model.

Thanks again!

PS1: I won't be able to test the new HWiNFO version since tomorrow I will be leaving for my summer vacation. After mid-June I can check again.
PS2: I've noticed that the history text file is not available with HWiNFO anymore. I remember one summer night back in 2000 or so, I read all that file. It took me some (many) hours. But it provided me some insights on what was done in HWiNFO 4.x.y (I can't remember the exact version number) up until that point.

Last edited by ALEKS on 2020-09-13, 06:10. Edited 1 time in total.

TX486DLC / 40 MHz | 32 Mb RAM | 16-bit ISA Backplane | Tseng Labs ET4000/W32i 2 Mb | I/O Interface | Audio Interface | PC Speaker Driver | Signal View Interface
3.5" & 5.25" FDD | 4 x 512 Mb CF | HP 82341D Interface | Intel EtherExpress 16

Reply 514 of 881, by deksar

User metadata
Rank Member
Rank
Member

ibm PS/1 Model 2168 Tower PC gives Fatal Error.

Ran it as:
"Shutdown and reboot into MS-DOS" (Under Windows 95) menu.

Update:
Restarted PC, re-tried it again, passed the into screen. However it hangs under "Peripherals Information".
Keyboard unresponsive. CTRL+ALT+DEL doesn't work.

Attachments

Reply 515 of 881, by Mumak

User metadata
Rank Oldbie
Rank
Oldbie
deksar wrote on 2020-07-27, 09:41:
ibm PS/1 Model 2168 Tower PC gives Fatal Error. […]
Show full quote

ibm PS/1 Model 2168 Tower PC gives Fatal Error.

Ran it as:
"Shutdown and reboot into MS-DOS" (Under Windows 95) menu.

Update:
Restarted PC, re-tried it again, passed the into screen. However it hangs under "Peripherals Information".
Keyboard unresponsive. CTRL+ALT+DEL doesn't work.

Try to disable the "Check Keyboard Controller and Super IO" option in HWiNFO before entering that screen if that will help.
It would be great if you could attach the HWiNFO LOG and Debug File (use -d switch to enable Debug Mode).

Reply 516 of 881, by ALEKS

User metadata
Rank Newbie
Rank
Newbie

There is another bug that manifests if, for instance, I launch HWiNFO without being in its directory but by using a command of this type:

hwinfo\hwinfo.exe

(".exe" part is optional in MS-DOS as the OS will search for ".bat", ".com", ".exe" files of that name anyway)

It will quickly throw up an error as pictured. This can be easily fixed by inferring the path to the database file directly from the commandline.

hwinfo-error.jpg
Filename
hwinfo-error.jpg
File size
350.36 KiB
Views
1105 views
File license
Fair use/fair dealing exception

TX486DLC / 40 MHz | 32 Mb RAM | 16-bit ISA Backplane | Tseng Labs ET4000/W32i 2 Mb | I/O Interface | Audio Interface | PC Speaker Driver | Signal View Interface
3.5" & 5.25" FDD | 4 x 512 Mb CF | HP 82341D Interface | Intel EtherExpress 16

Reply 517 of 881, by RayeR

User metadata
Rank Oldbie
Rank
Oldbie

It's not unusual issue. If you want to run HWI from bat without CD you can try utility INDIR 2.1 that launches a program from specified dir like you would do CD but current path is not affected.

Gigabyte GA-P67-DS3-B3, Core i7-2600K @4,5GHz, 8GB DDR3, 128GB SSD, GTX970(GF7900GT), SB Audigy + YMF724F + DreamBlaster combo + LPC2ISA

Reply 518 of 881, by mbliss11

User metadata
Rank Newbie
Rank
Newbie

Hey Mumak - thanks for all of your hard work! I was wondering if you could help with a troubleshooting issue that HWINFO is strangely fixing for me. I have an old Optiplex GL5100 and a Voodoo3 PCI running in a pure DOS environment. For some reason the board/bios will not detect the Voodoo3's VBE support correctly until I run HWINFO and check out the video details. The moment I do this it works perfectly. Without doing so I cannot for the life of me get the Voodoo3 to work correctly in VESA modes. I was able to find someone with a Voodoo5 that was experiencing the same issue and stumbled upon HWINFO as well:

Voodoo5 Pure DOS VESA/VBE 3.0 Issues

Could you shed any light on what HWINFO is doing to pull the card details? I have tried a number of different vesa tools and fixes to no avail and this is the only thing that works and is rock solid.

Reply 519 of 881, by Mumak

User metadata
Rank Oldbie
Rank
Oldbie
mbliss11 wrote on 2020-08-10, 23:50:

Hey Mumak - thanks for all of your hard work! I was wondering if you could help with a troubleshooting issue that HWINFO is strangely fixing for me. I have an old Optiplex GL5100 and a Voodoo3 PCI running in a pure DOS environment. For some reason the board/bios will not detect the Voodoo3's VBE support correctly until I run HWINFO and check out the video details. The moment I do this it works perfectly. Without doing so I cannot for the life of me get the Voodoo3 to work correctly in VESA modes. I was able to find someone with a Voodoo5 that was experiencing the same issue and stumbled upon HWINFO as well:

Voodoo5 Pure DOS VESA/VBE 3.0 Issues

Could you shed any light on what HWINFO is doing to pull the card details? I have tried a number of different vesa tools and fixes to no avail and this is the only thing that works and is rock solid.

I'm afraid it's very difficult to determine what exactly is causing this. HWiNFO does lots of low-level checks by accessing various registers and BIOS calls and I assume one of them is causing this unintentional fix in VESA support. But which one exactly would require running dozens of tests with tracing the code on such a system. This is almost impossible to determine remotely.