Hi,

Today I've got the Allied-Telesis RE2000 and the Fujitsu FMV-181,
and also got datasheet of MB86960 and MB86965.
I'll work to clean up mb86960 drivers (in my spare time :-)

cpg@aladdin.de wrote:

> What is "FE_D7_IDENT_NICE"? Is it that nice?

According to MB86960 manual, NICE is an alias of MB86960, which
stands for "Network Interface Controller with Encoder/decorder."
I think "FE_D7_IDENT_86960" is more readable, and current
FE_D7_IDENT_NICE in mb86960reg.h is wrong (it should be 0x00).

> Ok, I've added them. The only place one of the old devices was used,
> was in mb86860.c (FE_D7_IDENT_EC):

Maybe "EC" means "EtherCoupler," which is also a nickname of MB86965.

> My data sheet (for the 86967) says that bits 6 and 7 of DLCR7
> shouldn't be written. It says "Writing to these bits is invalid." I
> take this as "Writing _ones_ to these bits is invalid."
> So the above line is at least suspect.

On MB86960, these bits have a different function, encoder/decorder
configuration. Maybe we should have definitions like:

#define FE_D7_ED_NORMAL		0x00	/* Normal NICE */
#define FE_D7_ED_MON		0x40	/* NICE + Monitor */
#define FE_D7_ED_BYPASS		0x80	/* Encoder/Decorder Bypass */
#define FE_D7_ED_TEST		0xc0	/* Encoder/Decorder Test */

I'll update mb86960reg.h for these bits.
(I'm not sure that "ED_TEST" is appropriate settings, though..)

> >>>	/* Is this really needs to be done here? XXX */
> >>>	/* Turn the "master interrupt control" flag of ASIC on. */
> >>>	bus_space_write_1(memt, memh, FE_FMV3, FE_FMV3_ENABLE_FLAG);
> 
> Removed (and it still works :-)

Good. (I've already cleaned up some code around if_fereg.h)

>>> file dev/ic/ate_subr.c mbe_g2bus
> This didn't work, config reported an error about this file used twice.

Ah, I see, it actually has problem. Sorry.

> I did this change in conf/files:
 :
> -file   dev/ic/ate_subr.c               ate
> +file   dev/ic/ate_subr.c               ate | mbe_g2bus
 :
> I don't know if this is correct and doesn't break other things...

Hmm, it isn't good thing to add a MD definition in MI conf/files.
It might cause error on other ports.

Ok, I'll merge dev/ic/ate_subr.c into dev/ic/mb86960.c
since RE2000 seems to have only the MB86965 chip and some SRAM/TTLs
so it would not have any Allied-Telesis specific feature.
---
Izumi Tsutsui
tsutsui@ceres.dti.ne.jp