[parisc-linux] MUX driver on E55

[parisc-linux] MUX driver on E55

[Christoph Plattner]

Hallo Ryan !

Congratulation, your MUX driver on port 0 is really
working on my machine !!

What do you miss to support the other lines ?
Have you already knowledge of the data structure ?
Can I help you here in the code ?

Another point: you mentioned the skeleton of the
SCSI SPIFI driver. Where can I find that ? I only
found a NetBSD code, which I will use as functional
template, but I do not know, if this would be enough.
Can you give me a hint in this point ?

Christoph


-- 
-------------------------------------------------------
private:	christoph.plattner@gmx.at
company:	christoph.plattner@alcatel.at

[parisc-linux] Re: MUX driver on E55

[Ryan Bradetich]

On Fri, 2002-09-13 at 15:59, Christoph Plattner wrote:
> Hallo Ryan !
> 
> Congratulation, your MUX driver on port 0 is really
> working on my machine !!

Glad to hear that!

> What do you miss to support the other lines ?

I am not sure, I've never tried to hook up additional
lines.  I do have an 8-port MDP somewhere, I'll look
for it here shortly, and see if I can hook it up.

> Have you already knowledge of the data structure ?

Just what I have added into the driver.  If there is 
something that we are obviously missing, I can go back
and reference the HP-UX mux driver and see what is
missing.

> Can I help you here in the code ?

Please! :)  I do not use multiple lines currently ...
I only use the line 0, so any testing, optimizing, etc
is greatly appreciated!

> Another point: you mentioned the skeleton of the
> SCSI SPIFI driver. Where can I find that ? I only
> found a NetBSD code, which I will use as functional
> template, but I do not know, if this would be enough.
> Can you give me a hint in this point ?

I have it on my local system.  It is _very_ ugly at this
point.  I have got enough to query information from the 
spifi chip, register an interrupt, etc... but currently
I can not get the spifi chip to actually generate an
interrupt.

I would be happy to send you a copy when I get back next
monday (unless you catch my in the next hour or so), but
I'm not ready to check it into CVS yeah until I get rid
of the some ugly hardcoded hacks, global variables, etc.
Right now the driver is really coded to gather info from
the ESIEE E35 that I have been testing on.  Once again,
I do not have any docs on the chipset, I am using both
the NetBSD and HP-UX drivers as reference. :(

The driver should be easy to modify for your specific E55,
but not sure if you are interested in that level of work,
messyness now.  I promise it will get much better :)

Thanks,

- Ryan


> Christoph
> 
> 
> -- 
> -------------------------------------------------------
> private:	christoph.plattner@gmx.at
> company:	christoph.plattner@alcatel.at
> 

[parisc-linux] Re: MUX driver on E55

[Christoph Plattner]

Hello,

if it is possible, send the code to me (perhpas I can
have a look @ week end.

Is it possible to geht the HP-UX driver codes for those
to things ? MUX + SCSI ?

I was offered to sign a NDA, but then nothing happens ...
It would be more easy for me to read such original HP-UX
code. Perhaps you see a possibility to send it also.

Have a nice week end

Christoph



Ryan Bradetich wrote:
> 
> On Fri, 2002-09-13 at 15:59, Christoph Plattner wrote:
> > Hallo Ryan !
> >
> > Congratulation, your MUX driver on port 0 is really
> > working on my machine !!
> 
> Glad to hear that!
> 
> > What do you miss to support the other lines ?
> 
> I am not sure, I've never tried to hook up additional
> lines.  I do have an 8-port MDP somewhere, I'll look
> for it here shortly, and see if I can hook it up.
> 
> > Have you already knowledge of the data structure ?
> 
> Just what I have added into the driver.  If there is
> something that we are obviously missing, I can go back
> and reference the HP-UX mux driver and see what is
> missing.
> 
> > Can I help you here in the code ?
> 
> Please! :)  I do not use multiple lines currently ...
> I only use the line 0, so any testing, optimizing, etc
> is greatly appreciated!
> 
> > Another point: you mentioned the skeleton of the
> > SCSI SPIFI driver. Where can I find that ? I only
> > found a NetBSD code, which I will use as functional
> > template, but I do not know, if this would be enough.
> > Can you give me a hint in this point ?
> 
> I have it on my local system.  It is _very_ ugly at this
> point.  I have got enough to query information from the
> spifi chip, register an interrupt, etc... but currently
> I can not get the spifi chip to actually generate an
> interrupt.
> 
> I would be happy to send you a copy when I get back next
> monday (unless you catch my in the next hour or so), but
> I'm not ready to check it into CVS yeah until I get rid
> of the some ugly hardcoded hacks, global variables, etc.
> Right now the driver is really coded to gather info from
> the ESIEE E35 that I have been testing on.  Once again,
> I do not have any docs on the chipset, I am using both
> the NetBSD and HP-UX drivers as reference. :(
> 
> The driver should be easy to modify for your specific E55,
> but not sure if you are interested in that level of work,
> messyness now.  I promise it will get much better :)
> 
> Thanks,
> 
> - Ryan
> 
> > Christoph
> >
> >
> > --
> > -------------------------------------------------------
> > private:      christoph.plattner@gmx.at
> > company:      christoph.plattner@alcatel.at
> >

-- 
-------------------------------------------------------
private:	christoph.plattner@gmx.at
company:	christoph.plattner@alcatel.at

[parisc-linux] Re: SCSI/MUX driver on E55, NDA for sorce code

[Christoph Plattner]

Yes, I would be interested in HP-UX source code access,
if there are no costs I have to pay. I would sign a NDA.

The status of SCSI driver you sent me:
The machine does not really access the hardware. Also
if I patch the data structures or the "dev->hpa" pointer,
the driver reacts the same strange way.

The status of MUX: It is not working to access to offset
0x80 for the lines. The machine dies. I think this problem
can be solved very easy, looking in the HP-UX driver code.

I will do some futher code instrumentation on the SPIFI
driver. Then I will reread the driver of you and NetBSD
to have an idea, which access does fail here. Or do we
have a wrong hardware access anyway ?

Good start of the next week

Christoph



Ryan Bradetich wrote:
> 
> Thomas and Christoph,
> 
> Here is the current diff I have (I told you it was ugly! :))
> 
> I can not send you the HP-UX code.  If you are interested
> in pursuing the NDA to get access to the drivers, I can approach
> Alan Meyer (LSO Lab manager).  I am in IT, not the R&D lab so
> I do not have much insight in what they are doing :(
> 
> Hope this diff helps, and together hopefully we can get this to
> work and committed!
> 
> thanks,
> 
> - Ryan
> 
> On Fri, 2002-09-13 at 18:25, Christoph Plattner wrote:
> > Hello,
> >
> > if it is possible, send the code to me (perhpas I can
> > have a look @ week end.
> >
> > Is it possible to geht the HP-UX driver codes for those
> > to things ? MUX + SCSI ?
> >
> > I was offered to sign a NDA, but then nothing happens ...
> > It would be more easy for me to read such original HP-UX
> > code. Perhaps you see a possibility to send it also.
> >
> > Have a nice week end
> >
> > Christoph
> >
> >
> >
> > Ryan Bradetich wrote:
> > >
> > > On Fri, 2002-09-13 at 15:59, Christoph Plattner wrote:
> > > > Hallo Ryan !
> > > >
> > > > Congratulation, your MUX driver on port 0 is really
> > > > working on my machine !!
> > >
> > > Glad to hear that!
> > >
> > > > What do you miss to support the other lines ?
> > >
> > > I am not sure, I've never tried to hook up additional
> > > lines.  I do have an 8-port MDP somewhere, I'll look
> > > for it here shortly, and see if I can hook it up.
> > >
> > > > Have you already knowledge of the data structure ?
> > >
> > > Just what I have added into the driver.  If there is
> > > something that we are obviously missing, I can go back
> > > and reference the HP-UX mux driver and see what is
> > > missing.
> > >
> > > > Can I help you here in the code ?
> > >
> > > Please! :)  I do not use multiple lines currently ...
> > > I only use the line 0, so any testing, optimizing, etc
> > > is greatly appreciated!
> > >
> > > > Another point: you mentioned the skeleton of the
> > > > SCSI SPIFI driver. Where can I find that ? I only
> > > > found a NetBSD code, which I will use as functional
> > > > template, but I do not know, if this would be enough.
> > > > Can you give me a hint in this point ?
> > >
> > > I have it on my local system.  It is _very_ ugly at this
> > > point.  I have got enough to query information from the
> > > spifi chip, register an interrupt, etc... but currently
> > > I can not get the spifi chip to actually generate an
> > > interrupt.
> > >
> > > I would be happy to send you a copy when I get back next
> > > monday (unless you catch my in the next hour or so), but
> > > I'm not ready to check it into CVS yeah until I get rid
> > > of the some ugly hardcoded hacks, global variables, etc.
> > > Right now the driver is really coded to gather info from
> > > the ESIEE E35 that I have been testing on.  Once again,
> > > I do not have any docs on the chipset, I am using both
> > > the NetBSD and HP-UX drivers as reference. :(
> > >
> > > The driver should be easy to modify for your specific E55,
> > > but not sure if you are interested in that level of work,
> > > messyness now.  I promise it will get much better :)
> > >
> > > Thanks,
> > >
> > > - Ryan
> > >
> > > > Christoph
> > > >
> > > >
> > > > --
> > > > -------------------------------------------------------
> > > > private:      christoph.plattner@gmx.at
> > > > company:      christoph.plattner@alcatel.at
> > > >
> >
> > --
> > -------------------------------------------------------
> > private:      christoph.plattner@gmx.at
> > company:      christoph.plattner@alcatel.at
> >
> 
>   ------------------------------------------------------------------------
>                   Name: spifi.patch
>    spifi.patch    Type: text/x-patch
>               Encoding: quoted-printable

-- 
-------------------------------------------------------
private:	christoph.plattner@gmx.at
company:	christoph.plattner@alcatel.at

Re: [parisc-linux] Re: SCSI/MUX driver on E55, NDA for sorce code

[Matthew Wilcox]

On Sun, Sep 15, 2002 at 08:11:07PM +0200, Christoph Plattner wrote:
> The status of MUX: It is not working to access to offset
> 0x80 for the lines. The machine dies. I think this problem
> can be solved very easy, looking in the HP-UX driver code.

Not so sure...

common/sio/mux2.c:

for (loop = 0; loop < ports; loop++, ttyp++) {
        HW.hpa                  = MUX_DATA.hpa + 0x800 + (loop * 0x80);
        HW.io_reg_ptr           = (io_reg_type *)(HW.hpa + 0x30);
}

-- 
Revolutions do not require corporate support.

Re: [parisc-linux] Re: SCSI/MUX driver on E55, NDA for sorce code

[Christoph Plattner]

Thanks for that !

But I think, my Linux code dose the same !!!

	info->iomem_base = (void *)(hpa + MUX_OFFSET + 
				    (line * MUX_LINE_OFFSET));

This info->iomem_base is used by all accesses, the 
registers of i/o are defined by Ryan as:

#define IO_COMMAND_REG_OFFSET 0x30
#define IO_STATUS_REG_OFFSET 0x34
#define IO_DATA_REG_OFFSET 0x3c
#define IO_DCOUNT_REG_OFFSET 0x40
#define IO_UCOUNT_REG_OFFSET 0x44
#define IO_FIFOS_REG_OFFSET 0x48

So the offset of 0x30 is inside those macros.

An example of an access:

		data = __raw_readl((unsigned long)info->iomem_base 
				   + IO_DATA_REG_OFFSET);

... reading characters from the input buffers.

Any further idea ?

Ryan, you mentioned problems concerning missing locks, but
I cannot think about this. Those devices should be controlled
independent, am I wrong ?

I hope, we will solve this all (all means including SPIFI
SCSI driver too !).

Christoph



Matthew Wilcox wrote:
> 
> On Sun, Sep 15, 2002 at 08:11:07PM +0200, Christoph Plattner wrote:
> > The status of MUX: It is not working to access to offset
> > 0x80 for the lines. The machine dies. I think this problem
> > can be solved very easy, looking in the HP-UX driver code.
> 
> Not so sure...
> 
> common/sio/mux2.c:
> 
> for (loop = 0; loop < ports; loop++, ttyp++) {
>         HW.hpa                  = MUX_DATA.hpa + 0x800 + (loop * 0x80);
>         HW.io_reg_ptr           = (io_reg_type *)(HW.hpa + 0x30);
> }
> 
> --
> Revolutions do not require corporate support.
> _______________________________________________
> parisc-linux mailing list
> parisc-linux@lists.parisc-linux.org
> http://lists.parisc-linux.org/mailman/listinfo/parisc-linux

-- 
-------------------------------------------------------
private:	christoph.plattner@gmx.at
company:	christoph.plattner@alcatel.at

Re: [parisc-linux] Re: SCSI/MUX driver on E55, NDA for sorce code

[Ryan Bradetich]

On Sun, 2002-09-15 at 17:37, Christoph Plattner wrote:
> Thanks for that !
> 
> But I think, my Linux code dose the same !!!
> 
> 	info->iomem_base = (void *)(hpa + MUX_OFFSET + 
> 				    (line * MUX_LINE_OFFSET));
> 
> This info->iomem_base is used by all accesses, the 
> registers of i/o are defined by Ryan as:
> 
> #define IO_COMMAND_REG_OFFSET 0x30
> #define IO_STATUS_REG_OFFSET 0x34
> #define IO_DATA_REG_OFFSET 0x3c
> #define IO_DCOUNT_REG_OFFSET 0x40
> #define IO_UCOUNT_REG_OFFSET 0x44
> #define IO_FIFOS_REG_OFFSET 0x48
> 
> So the offset of 0x30 is inside those macros.
> 
> An example of an access:
> 
> 		data = __raw_readl((unsigned long)info->iomem_base 
> 				   + IO_DATA_REG_OFFSET);
> 
> ... reading characters from the input buffers.
> 
> Any further idea ?
> 
> Ryan, you mentioned problems concerning missing locks, but
> I cannot think about this. Those devices should be controlled
> independent, am I wrong ?

Locks was just an idea of why it might now work.  Nothing suggested that
yet.  I found the MPD, now I just need to make a cable for it to hook up
to a different system and test with.  Hopefully will get this done this
week and I'll take a look at it :)

> I hope, we will solve this all (all means including SPIFI
> SCSI driver too !).
> 
> Christoph