Re: gem driver for sun4m

[Erik LaBine <erik%eriklabine.com@localhost>]

On Saturday, July 18, 2026 10:19:40 AM Eastern Daylight Time Jason Thorpe 
wrote:
> > On Jul 18, 2026, at 2:56 AM, Julian Coleman <jdc%coris.org.uk@localhost> wrote:
> > 
> > When I added the SBus GEM support, I also couldn't find any documentation
> > 
> > for the best values - the only information I have is the:
> >  Specification for Sbus GEM Gigabit Ethernet ASIC (100-5544-03)
> 
> I have a copy of "Specification for Sbus GEM Gigabit Ethernet ASIC
> (100-5544-03)”.  The table that describes the SBus configuration register
> is on page 70, section 3.1.2.2, Table 3-2.

That was the document I was trying to track down, actually; it was mentioned 
in the comments of if_gem_sbus.c, but the sun.com link was long dead. Although 
just now I did some further searching for that exact title and managed to 
track down a copy on archive.org: https://ia601700.us.archive.org/7/items/
SPARCDocs/gem_sbus-1.pdf

> [2:0] System Maximum Burst Support -- Tells if the host system supports 4, 
8, or 16 word maximum burst size
> [3] System Extended -- When set means system is in 64 bit extended SBus 
mode; when not set means system is in 32 bit SBus mode
> ...
> [9] Parity Enable -- Parity enable, enables parity checking

By omitting the system extended bit, I suppose it's probably operating with 
shorter words than on sun4u. I spent quite a while wondering why BSIZE64 
seemed fine, when the sun4m documentation for SBus says the maximum is a 32 
byte burst.

It's also possible parity checking could be enabled if the platform supports 
it, rather than making that architecture dependent. I might have to look into 
that.

> Anyway, I’ve checked in the changes and have teed them up for inclusion in
> NetBSD 11.1, as well.
> 
> Thanks for submitting them, Erik!
> 
> -- thorpej

My pleasure! This has been a weekend project for me for a few months off-and-
on, I was just glad to get it into a stable state so it could see the light of 
day.

-Erik