linuxppc-dev.lists.ozlabs.org archive mirror
 help / color / mirror / Atom feed
* 8260 io and caches
@ 2000-05-01 16:38 Arto Vuori
  2000-05-01 16:59 ` Dan Malek
  2000-05-01 17:19 ` Neil Russell
  0 siblings, 2 replies; 4+ messages in thread
From: Arto Vuori @ 2000-05-01 16:38 UTC (permalink / raw)
  To: linuxppc-dev


Hi all,

I'm currently trying to get linux running on EST SBC8260 board. I had
some problems with serial ports. Initially it just sent some garbage. I
found that serial port driver doesn't initialize BRG division factor in
SCCR register even though it assumes that it should be set to 0. I
modified my bootloader to initialize it and output looks now much
better.

There still seems to be some problems with caches. Everything is just
fine with caches disabled, but if i enable caches output doesn't look
correct. Adding some flush_dcache_range() calls to uart.c seems to fix
that problem, but shouldn't rx & tx buffers be allocated from some non
cacheable region??

Thanks

Arto

--
Arto Vuori
a-mail: avuori@ssh.com
mobile:	+358 40 754 5223

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

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: 8260 io and caches
  2000-05-01 16:38 8260 io and caches Arto Vuori
@ 2000-05-01 16:59 ` Dan Malek
  2000-05-01 17:19 ` Neil Russell
  1 sibling, 0 replies; 4+ messages in thread
From: Dan Malek @ 2000-05-01 16:59 UTC (permalink / raw)
  To: Arto Vuori; +Cc: linuxppc-dev


Arto Vuori wrote:

> I'm currently trying to get linux running on EST SBC8260 board. I had
> some problems with serial ports.

OK.....

> There still seems to be some problems with caches.

What version of processor silicon do you have?


	-- Dan

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

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: 8260 io and caches
  2000-05-01 16:38 8260 io and caches Arto Vuori
  2000-05-01 16:59 ` Dan Malek
@ 2000-05-01 17:19 ` Neil Russell
  2000-05-01 21:53   ` Dan Malek
  1 sibling, 1 reply; 4+ messages in thread
From: Neil Russell @ 2000-05-01 17:19 UTC (permalink / raw)
  To: Arto Vuori, linuxppc-dev


>From what I can see, the uart driver in the current linux source is written
for the 860, with no mods for the 8260.  The 8260 has some different bits
in the RFCR/TFCR parameter RAM fields.  There is a GBL bit (bit 2) that
enables cache snooping.

I'm looking at the 2.3.99-pre5 sources, at about line 2519 in uart.c:

	up->smc_rfcr = SMC_EB;
	up->smc_tfcr = SMC_EB;

change them to:

	up->smc_rfcr = SMC_EB | 0x20;
	up->smc_tfcr = SMC_EB | 0x20;

and see what happens.  Of course, you probably want to add a SMC_GBL
#define to do this properly.


Neil.


On Mon, May 01, 2000 at 07:38:56PM +0300, Arto Vuori wrote:
> I'm currently trying to get linux running on EST SBC8260 board. I had
> some problems with serial ports. Initially it just sent some garbage. I
> found that serial port driver doesn't initialize BRG division factor in
> SCCR register even though it assumes that it should be set to 0. I
> modified my bootloader to initialize it and output looks now much
> better.
>
> There still seems to be some problems with caches. Everything is just
> fine with caches disabled, but if i enable caches output doesn't look
> correct. Adding some flush_dcache_range() calls to uart.c seems to fix
> that problem, but shouldn't rx & tx buffers be allocated from some non
> cacheable region??

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

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: 8260 io and caches
  2000-05-01 17:19 ` Neil Russell
@ 2000-05-01 21:53   ` Dan Malek
  0 siblings, 0 replies; 4+ messages in thread
From: Dan Malek @ 2000-05-01 21:53 UTC (permalink / raw)
  To: Neil Russell; +Cc: Arto Vuori, linuxppc-dev


Neil Russell wrote:
>
> >From what I can see, the uart driver in the current linux source is written
> for the 860, with no mods for the 8260.

Oh, there are some mods, but I did start with the 8xx code :-).

> ....  The 8260 has some different bits
> in the RFCR/TFCR parameter RAM fields.  There is a GBL bit (bit 2) that
> enables cache snooping.

Oooops.  Sorry about that.  I was playing with caches (and not) and
I guess I didn't get everything checked in properly.  Make sure you
also make the same cache changes to the SCC Ethernet.

Don't consider this software ready to run.  I have more to check
in and I am still making lots of changes.

> .....  Of course, you probably want to add a SMC_GBL
> #define to do this properly.

It was supposed to be there...........


	-- Dan

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

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2000-05-01 21:53 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2000-05-01 16:38 8260 io and caches Arto Vuori
2000-05-01 16:59 ` Dan Malek
2000-05-01 17:19 ` Neil Russell
2000-05-01 21:53   ` Dan Malek

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).