Serial MUX support for Nova and K-Class systems.

Hello all,

I just committed support for the Serial MUX driver as linux-2.4.19-pa5.

This patch has been tested on my K460 and has been working well. The driver uses the same /dev/ttyB0 device the PDC Software console that Christoph Plattner wrote and is available under the Non-standard serial ports section.

Currently the driver only supports the Eole CAP/MUX (A Direct) device since that is all I have to test with. Additional devices can be added under this driver as people with systems have a chance to test.

I tried to document the driver with kerneldoc, but it would be nice if someone would check over my comment as I had to google/guess what most of the ioctls and tty flags were.

The committed patch can be found at:

        ftp://ftp.parisc-linux.org/patches/serial_mux_v0.1.diff

Feedback greatly appreciated!

Thanks,

- Ryan

_______________________________________________
parisc-linux mailing list
[email protected]
http://lists.parisc-linux.org/mailman/listinfo/parisc-linux

Hi Christoph,

I'm working on solving the problem why we do not find the MUX device. The ESIEE team has generously given me access to an E35 so I could poke around the HP-UX iotree and see what I could find.

This is the summary of what I have found:

name_buffer instance hw_addr parent sibling child

"bc"               1      56 0x43b3a0 0xc15240 0xc15040
  "scsi1"          0      52 0xc15000 0xe74140 0xc15080
  ""              -1      53 0xc15000 0xc15200        0
  "mux2"           0      56 0xc15000        0        0
"bus_adapter"      0      60 0x43b3a0 0xe74240 0xc15280
  "lan2"           0       6 0xc15240        0        0
"processor"        0      62 0x43b3a0 0xe74280        0
"memory"           0      63 0x43b3a0        0        0

The MUX sits at: 56/56 with HPA: 0xfffb8000

Quoted 1 line

From your boot logs posted to the mailing list:

---- walk_native_bus: testing hpa=0xfffb8000, i=56
---- walk_native_bus: -- device not processed, mod = i !
------ alloc_pa_dev: entry -------------------------- find_parisc_device: entry, mod=56

-------------------------- find_parisc_device:   bc[0]=-1
-------------------------- find_parisc_device:   bc[1]=-1
-------------------------- find_parisc_device:   bc[2]=-1
-------------------------- find_parisc_device:   bc[3]=-1
-------------------------- find_parisc_device:   bc[4]=-1
-------------------------- find_parisc_device:   bc[5]=-1

-- walk_lower_bus: entry
-- walk_lower_bus: dev: hpa=0xfffb8000, io_io_low=0xfff00000, io_io_high=0xfff80000

You also found a device at 0xfffb8000 which I am assuming is the MUX. (I need to get back on the E35 at the ESIEE to verify my thoughts.)

The problem we are running into here is that 0xfffb8000 is on the Central BUS, not under the 56/ bus as the console wants to be in the HP-UX iotree. It appears to me that the bc and the MUX exist at the same HPA and that is what is causing the confusion in the parisc-linux I/O tree.

The other thing that is curious to me is the "" devices under the bc between the scsi1 and the mux2 device.

I think I can get this problem solved once I get access to the ESIEE E35 and am able to boot some test kernels on the system to play with.

As for the scsi driver. I have the bare-bones (ie. does not work) of both the skunk (8-bit) and the wizard (16-bit) drivers started from a while back. I think you found a reference to the SPIFI chip that we could use in the driver to further progress the drivers and hopefully get them talking.

The document is the IOACD document I believe. What are you looking for? Parts of the document are on the parisc-linux website, but if that is not enough info, I'll try to find it for you in the IOACD.

Thanks!

- Ryan

Quoted 64 lines · Christoph Plattner

Hallo Ryan,

after that long time, I did the first experiment with your new MUX driver ...

... and it fails on my E55.

But NO WONDER ! I already have the problem finding any devices on the I/O interface card, and I had no time since weeks now, to do further steps.

In general I have the problem in basic stuff and in docu, special for the I/O scanning via PDC in more details. In the normal PDC/IO there is not enough docu for that. I was informed, that there is one more docu handling this, bt I forgot the name, something like "AIO" or so.

How we can go one step forward, as I also want to start (in slowly steps) the SCSI driver ... ?

Bye,
Christoph

Quoted 23 lines · Ryan Bradetich

Hello all,

I just committed support for the Serial MUX driver as linux-2.4.19-pa5.

This patch has been tested on my K460 and has been working well. The driver uses the same /dev/ttyB0 device the PDC Software console that Christoph Plattner wrote and is available under the Non-standard serial ports section.

Currently the driver only supports the Eole CAP/MUX (A Direct) device since that is all I have to test with. Additional devices can be added under this driver as people with systems have a chance to test.

I tried to document the driver with kerneldoc, but it would be nice if someone would check over my comment as I had to google/guess what most of the ioctls and tty flags were.

The committed patch can be found at:

        ftp://ftp.parisc-linux.org/patches/serial_mux_v0.1.diff

Feedback greatly appreciated!

Quoted 8 lines

Thanks,

- Ryan

_______________________________________________
parisc-linux mailing list
[email protected]
http://lists.parisc-linux.org/mailman/listinfo/parisc-linux

--
-------------------------------------------------------

private:      [email protected]
company:      [email protected]