Status of linuxppc support for Galileo GT64260

Status of linuxppc support for Galileo GT64260

[mod+linuxppc-dev]

I'm working with older linuxppc sources
that run on a box designed around
the Galileo GT64260 and I'm thinking
about merging forward to the current
linuxppc code base, so I'm wondering:
what's the current status of the code
that got checked in as the official
PPClinux code for the Galileo GT64260?

I'm asking because I recall seeing
messages on these lists saying that
code had been checked in with interrupt
problems, PCI problems, no integrated-Enet
driver and generally in "needs work"
condition.  I notice that there is now a
(strangely familiar) driver included for
the integrated Enet, but (except for the
recent exchange about how the GT64260 and
VT82C686B don't play so nice together) I
don't find much 64260-related discussion
here except one terse followup message
stating that PCI works.

So, is anybody using the current code?
Would I be wrong to interpret the fact
that it wouldn't compile when I tried it
today as indication that not much has
happened since the initial checkin by
MontaVista?  Are the interrupt problems
alluded to at the time of that initial
checkin (could somebody characterize
them for me?) still an issue?  Does
the PCI support seem solid?  Anything
in particular I need to watch out for?

Thanks.
          Michael O'Donnell
        Mission Critical Linux


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Status of linuxppc support for Galileo GT64260

[Mark A. Greer]

mod+linuxppc-dev@missioncriticallinux.com wrote:

> I'm asking because I recall seeing
> messages on these lists saying that
> code had been checked in with interrupt
> problems, PCI problems, no integrated-Enet
> driver and generally in "needs work"
> condition.  I notice that there is now a
> (strangely familiar) driver included for
> the integrated Enet,

Yes, its the same driver the galileo originally wrote that mclx, your
company, moved to 2.4 AFAICT.  If you wrote it, fine, change the
copyright and author.  BTW, its *still* needs work so feel free.


> but (except for the
> recent exchange about how the GT64260 and
> VT82C686B don't play so nice together) I
> don't find much 64260-related discussion
> here except one terse followup message
> stating that PCI works.

There have been a few pushes to 64260 related code since the original
push.  If you want to keep up on it, do regular pulls and look at the
code and/or get on the linuxppc-commit mailing list and look for pushes
that touch files you care about.

> So, is anybody using the current code?

Yes, me--at least the code of a week or so ago.  Others are to if they
wish to speak up.

> Would I be wrong to interpret the fact
> that it wouldn't compile when I tried it
> today as indication that not much has

> happened since the initial checkin by
> MontaVista?

My src tree is a week or so out of date but it compiles just fine.
Not compiling is news to me.  Why don't you fix it and send a patch? ;)

> Are the interrupt problems
> alluded to at the time of that initial
> checkin (could somebody characterize
> them for me?) still an issue?

You can still lose PCI interrupts with the 64260 (see errata) which is a
real problem if you have an nfs-mounted root filesystem.  The 64260A is
stable for me except after a while something happens and my bdi can no
longer talk to the processor and the system appears to hang.

>  Does
> the PCI support seem solid?

It was last time I tested.

> Anything
> in particular I need to watch out for?

No mpsc/uart driver yet.  That's being worked on but if you have one that
works well with the 64260A's errata, feel free to pass it along.  I know
there is at least one person working on a dual 7450 + 64260A combo.
Don't think its up with both cpu's yet.

Mark


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Status of linuxppc support for Galileo GT64260

[Tom Rini]

On Tue, Dec 04, 2001 at 06:32:12PM -0500, mod+linuxppc-dev@MissionCriticalLinux.com wrote:

[snip]
> Would I be wrong to interpret the fact
> that it wouldn't compile when I tried it
> today as indication that not much has
> happened since the initial checkin by
> MontaVista?

No, just that people don't always test all of the changes when they do a
commit, and that the people with the hardware and write access aren't
staying up to date (in an effort to limit bugs one can assume).

--
Tom Rini (TR1265)
http://gate.crashing.org/~trini/

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/