< SOLVED > Pci siig problem

good day,

I'm trying to set up my system pico system, but seems to have problem with my brand SIIG PCI card reference P005-62,

I can not run the diagnostic software with pico systems.

My question is simple to someone else worked with this type of PCI and has had a similar problem,

as I can know that protocol handles this board?

Elson Jon tells me:

then it may be necessary to figure out the offset
to trick pcisetup into configuring it at the correct address. Your parport
board apparently does not use the standard offset of 0x402 between the
main port and the config register.


thanks,

Hi

This post should really be in 'Driver Boards', where Jon Elson is a moderator, as it is specific to Pico stuff.
(for some reason I can no longer move posts - only delete them - JT or Andy could you oblige)

I have to assume that your PCI card is a parport card? You have not said.

I also have to assume that you are trying to set it to EPP mode? This is only because the address in the port register which sets the mode is usually at 0x402 above the port.

I did once use a SIIG card for Mach and it worked fine

I had all sorts of problems getting it to work with linuxcnc and can't remember if I ever got it into EPP mode.
I have a feeling I didn't need EPP and the problems were to do with getting it to work in 'input' mode with pins 2-9 as input pins

The pcisetup program Jon refers to is a commandline program that sets the address specified to EPP mode by setting it to 0x80.
It takes the port address as an argument and adds 0x402

If you read the below thread you will see why you should talk to Jon, he had the same problem himself recently and the thread does not make it clear exactly what the offset is, although it appears to be 0x882 above the port if I am reading it correctly

sourceforge.net/mailarchive/forum.php?th...forum_name=emc-users

You will need to do a lspci -vv (2 x V not 1 x W) and post it to show the port and register addresses for the card, as he will need this info

regards

Regards,

Thanks for your answer.

but do not understand, my card is parport, if I'm trying to place it in EPP mode,

What is strange to me is that I bought 3 cards and none of them work. this system really works?

I just tested on three different computers, even one that brings the integrated port and nothing.

I will continue proving to see him.


thank you very much,

jorge viloria

As ArcEye said, please do a lspci -vv and post the results here,
otherwise it will be difficult to help.

Heres how:

In a terminal window type:

lspci -vv > lspci.txt

and post the file lspci.txt here

thanks

jlviloria wrote:

thanks

05:00.0 Parallel controller: Siig Inc CyberParallel (1-port) (prog-if 02)
Subsystem: Siig Inc CyberParallel (1-port)
Control: I/O+ Mem+ BusMaster- SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
Status: Cap+ 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
Interrupt: pin A routed to IRQ 18
Region 0: I/O ports at d010
Region 1: I/O ports at d000
Capabilities: <access denied>
Kernel driver in use: parport_pc
Kernel modules: parport_pc

So, the problem here is that this board does not follow the old standard of having the
ECR at an offset of +0x402 above the data register. In fact, it is below the data register.
My dumb pcisetup program assumes the 0x402 offset, so to use it, it has to be
fooled. I think the ECR will be at 0xD002, so the value to supply to that program would be
d002 - 402 = cc00

So, you should try running the command like this :
sudo ./pcisteup cc00

and then try running the diagnostic program, supplying it with the d010 value,
as in :
sudo ./univpwmdiags d010 bus

and see what you get. Apparently the ppmc driver does this correctly, as
Mr. Viloria reports that works at some level. I don't know if he has connected
than just the UPC board to the parallel port so far.

Jon

Jon thanks for your answer,

look today I did the following:



1) UPC connected to 12VDC and parallel port.
2) open LinuxCNC (PPMC sample settings)
3) run the diags bus and Commtest
4) do not work well, showing errors and no display bus 1 and 2 only shows "unknown"
5) close LinuxCNC
6) run diags bus. 1 and 2 the same "unknown"
7) run diags Commtest and showed me 0 errors

but unfortunately I can not determine if it is right or wrong

tomorrow I'll do what you tell me.


thank you very much for your patience and collaboration,

Jorge

jlviloria wrote:

Jon thanks for your answer,

look today I did the following:



1) UPC connected to 12VDC and parallel port.
2) open LinuxCNC (PPMC sample settings)

Hmmm, you should be using the univpwm sample config for the Universal PWM Controller.
It should exit if you try the ppmc sample config, as that is different hardware.

3) run the diags bus and Commtest

OK, I should have been more clear, you can't run the diags AND LinuxCNC at the same time.
They interfere with each other accessing the board.

4) do not work well, showing errors and no display bus 1 and 2 only shows "unknown"
5) close LinuxCNC
6) run diags bus. 1 and 2 the same "unknown"
7) run diags Commtest and showed me 0 errors

WHAT? That sounds wrong. OH, note that one of your boards is set to address 1, the other
is set to address zero. For the diags to talk to it, the DIP switch position 10 needs to be on.
The PPMC driver knows how to enumerate the boards, so it will work with any setting,
but the diagnostic is very simple. BUT, if commtest shows zero errors, that sounds
like something must be working right!

Jon

good did what you told me and the results are the same attachment files to show you that I get.

that parport use your, or better than parport're absolutely do not have problems, it is only assembled and ready.

as suggestions or advice you should sell complete kits with parport included.

we tested three computers one with built-in parallel port in the bios set EPP mode address 378 and nothing.
It may also be that I is doing something wrong, but I should be more common, simple.


in bus.txt did your procedure and ran diags bus and Commtest

in bus1.txt LinuxCNC opened and then closed. my bus and ran the diags Commtest,

Commtest both let him run for over 20 minutes and 0 errors

in bus3.txt run several times bus and that was the result

thanks

jorge

Jon, PCW, Arceye

not know if this is worth, I think so, but this goes beyond my knowledge,

marc.info/?l=linux-kernel&m=93855197710528&w=2

regards

Jorge Viloria