7i43 not running

Hi
I have a seeedstudio x86 and just for fun i try to connect my old 7i43-4U
I can have the same parport card generating signals, using the 0xe070 address
loadrt hal_parport cfg="0xe070" .... Works like a charm.

NOW
Trying to run the 7i43 is not succesfull. I used Pncconf for it with looks very nice nowadays.
So I do lspci:

---

02:00.0 Serial controller: Oxford Semiconductor Ltd OX16PCI954 (Quad 16950 UART) function 0 (Uart) (prog-if 06 [16950])
Subsystem: Oxford Semiconductor Ltd OX16PCI954 (Quad 16950 UART) function 0 (Uart)
Flags: medium devsel, IRQ 22
I/O ports at e040
Memory at a1102000 (32-bit, non-prefetchable) [size=4K]
I/O ports at e020
Memory at a1101000 (32-bit, non-prefetchable) [size=4K]
Capabilities: <access denied>
Kernel driver in use: serial

02:00.1 Parallel controller: Oxford Semiconductor Ltd OX16PCI954 (Quad 16950 UART) function 1 (parallel port) (prog-if 01 [BiDir])
Subsystem: Oxford Semiconductor Ltd OX16PCI954 (Quad 16950 UART) function 1 (parallel port)
Flags: medium devsel, IRQ 22
I/O ports at e070
I/O ports at e060
I/O ports at e000
Memory at a1100000 (32-bit, non-prefetchable) [size=4K]
Capabilities: <access denied>
Kernel driver in use: parport_pc
Kernel modules: parport_pc

03:00.0 Ethernet controller: Intel Corporation I211 Gigabit Network Connection (rev 03)
Subsystem: Intel Corporation I211 Gigabit Network Connection
Flags: bus master, fast devsel, latency 0, IRQ 22
Memory at a1400000 (32-bit, non-prefetchable) [size=128K]
I/O ports at d000 [disabled]
Memory at a1420000 (32-bit, non-prefetchable) [size=16K]
Capabilities: <access denied>
Kernel driver in use: igb
Kernel modules: igb

04:00.0 Ethernet controller: Intel Corporation I211 Gigabit Network Connection (rev 03)
Subsystem: Intel Corporation I211 Gigabit Network Connection
Flags: bus master, fast devsel, latency 0, IRQ 23
Memory at a1300000 (32-bit, non-prefetchable) [size=128K]
I/O ports at c000 [disabled]
Memory at a1320000 (32-bit, non-prefetchable) [size=16K]
Capabilities: <access denied>
Kernel driver in use: igb
Kernel modules: igb

Then going the mesaflash road ( also with the newest compiled pre version) i see no sucess:

cnc@SeeedCNC:~/mesaflash$ sudo ./mesaflash --epp --print-pd --device 7i43 --addr 0xe070
failed to clear EPP Timeout!
EPP timeout on data cycle of read(addr=0x0000, size=4)
failed to clear EPP Timeout!
EPP timeout on data cycle of read(addr=0x0100, size=4)
failed to clear EPP Timeout!
Segmentation fault
cnc@SeeedCNC:~/mesaflash$

What might I do wrong?
Mike

I suspect the parallel port is not in EPP mode

Setting the mode requires access to the "high" EPP address
which is normally the base address + 0x400 but may be
something entirely different. You may be able to find the "high"
address by trying the other I/O addresses with the "--addr_hi"
mesaflash option

thx pcw
I read the chips doc and yes this Oxford chip have epp and these registers are normaly 0x400 higher. But i get very strange results.

I use this line in a normal hal/ini config.

loadrt hm2_7i43 ioaddr=0xE070 ioaddr_hi=0xE470 epp_wide=1 config="firmware=hm2/7i43-4/SVST4_4.BIT num_encoders=4 num_pwmgens=1 num_stepgens=4"

Any address on ioaddr=0xE070 ( tried them in 10 increments from 0xE000 ) generates the error like:
hm2/hm2_7i43.0: failed to reset fpga, aborting hm2_register
hm2_7i43.0: board at (ioaddr=0xE060, ioaddr_hi=0xE470, epp_wide ON) not found!
hm2_7i43: rtapi_app_main: Input/output error (-5)
AND shows ioaddr_hi=0xE470 ( ore any other number i write in the ini file).

BUT when i write the probably correct ioaddr=0xE070 in the ini commandline:

loadrt hm2_7i43 ioaddr=0xE070 ioaddr_hi=0xE470 epp_wide=1 config="firmware=hm2/7i43-4/SVST4_4.BIT num_encoders=4 num_pwmgens=1 num_stepgens=4"
THEN
the error response comes back with nulled ioaddr_hi
hm2_7i43.0: board at (ioaddr=0xE070, ioaddr_hi=0x0000, epp_wide ON) not found!

VERY STRANGE

This card has to my best knowledge no parameters from bios or so on.
Wasn't there a trick "disable" it to linux and give complete control to Linuxcnc?
Mike

Maybe I did not explain this as well I as could
The base address of 0xE070 is correct but its likely
the the high address is wrong

So I would try the other I/O addresses as the address high

You might also need epp_wide=0

Thx pcw
You explained very well. And i also tried the epp_wide switch.
But me invers was not clear enough.
If i set
The base adress wrong and the predicted epp correct then
I get in the error line a result with given epp adress
The base adress correct and (doesend matter what value) the predicted epp 0xE470
I get in the error line a result with ioaddr_hi always set =0000

So there is some difference in working with the correct base address.
Have you an idea what is "common" if not 0x400 register "distance"
Mike

I read controlled dmesg witch confirms recognition of the base_address:
[ 5.423364] ppdev: user-space parallel port driver
[ 5.441430] parport_pc 0000:02:00.1: enabling device (0000 -> 0003)
[ 5.441577] PCI parallel port detected: 1415:9513, I/O at 0xe070(0x0), IRQ 22
[ 5.442419] parport0: PC-style at 0xe070, irq 22 [PCSPP]
[ 5.533704] lp0: using parport0 (interrupt-driven).

Mike

"Have you an idea what is "common" if not 0x400 register "distance"

There is no standard, I think it has to be one of the other addresses
(+0x400 is not possible because there is no I/O region listed with a 0xE470 address)

The port starts out as SPP and you need to do the correct setup register magic
to set EPP mode (The BIOS will not do this if its a plug-in card)
Most hardware has the setup registers at base + 0x400 but yours does not
(so setting a ioaddr_high of 0xE470 makes no sense since 1. that's the default
that LinuxCNC will use if not specified and 2. its not an available address for your card)
have you tried 0xE060 or 0xE000 as ioaddr_high?

I agree to your weirdness remarks
This board is plugged in on the M2-PCI-E port via adapter. I believe that the seeedstudio mb alows to do quite a bunch of modifications on those two M2-plugs..
But the doc is simply not good and most of those choosable options..., Me, I never heard of most of them!
I did the parport test first and yes, I can see on an old 25-lines led-based tester that the output pins move signals. Yes, I'm to lazy to do cabling to real world or do oscilloscoping. My old LED tester shows result witch 97%, fair enough for my IT WORKS as normal parport. Not tuned latency is not good, but I do not wan't to tune for that case.
My main application will be 7i76e.
P.S.
Just have that old 7i43-4u laying around and will invest a bit of time to see if it is usable. ( Had a minor operation, now laying a bit around and use some time to remote via my business notebook into that tiny MB.

Thx to both ouf you.
Will try tomorrow with the low adresses.

But as you wrote pcw, also a lspci -vv does not show more.

And i learned it's good to make additionaly a power cycle.

Will make few tries, if not sucessfull i will look for another card.
Thereafter i will just give up as I'm not forced to use...
Had the idea to try one day to combine the 7i43 with some tmc2209.
I'm impressed how smooth and silent they run on my 3dprinter.
Thx again for your help
M7ke