At 02:44 PM 2/10/2004, Kate Feng wrote:
>Matt Thomas wrote:
>
> > On Thursday, October 2, 2003, at 10:07 AM, Kate Feng wrote:
> >
> > > Hello,
> > >
> > > I am looking for BSD device drivesr written for the Marvell GT64260B
> > > system
> > > controller.  I found some traces talking about support for the Marvell
> > > GT64260
> > > Evaluation Board.  However, I  could not find the software.  I am new
> > > to
> > >
> > > netBSD.....Sorry.
> >
> > Support for it is in NetBSD-current:
> >
> > sys/arch/evbppc/ev64260
> > sys/arch/powerpc/marvell
> > sys/dev/marvell
> > --
>
>Are those codes fully tested , especially for the 10/100 M
>ethernet on the GT64260 ?
>One particular file I looked at is the sys/dev/marvell/if_gfe.c.

The datasheet is wrong.  It took me a week of analysis to figure
out the exact algorithm.

>in the   gfe_hash_entry_op(),
>
>he = ((uint64_t) eaddr[5] << 43) |
>       ((uint64_t) eaddr[4] << 35) |
>       ((uint64_t) eaddr[3] << 27) |
>       ((uint64_t) eaddr[2] << 19) |
>       ((uint64_t) eaddr[1] << 11) |
>       ((uint64_t) eaddr[0] <<  3) |
>       HSH_PRIO_INS(prio) | HSH_V | HSH_R;
>
>However, based on  the datasheet I have , I thought it
>should be :
>
>he = (((unsigned long long) eaddr[0]) << 43) |
>        (((unsigned long long) eaddr[1]) << 35) |
>        (((unsigned long long) eaddr[2]) << 27) |
>        (((unsigned long long) eaddr[3]) << 19) |
>        (((unsigned long long) eaddr[4]) << 11) |
>        (((unsigned long long) eaddr[5]) <<  3) |
>        ((unsigned long long) HSH_PRIO_INS(prio) | HSH_V | HSH_R);
>
>May datasheet says:
>
>The  bit 6:3 :MAC address[43:40]
>        bit 10:7 :MAC address[47:44]
>      bit 14:11 : MAC address[35:32]
>      bit 18:15: MAC address [39:36]
>
>............
>
>Any  pointer would be highly appreciated.
>
>Thanks,
>Kate

-- 
Matt Thomas                     email: matt@3am-software.com
3am Software Foundry              www: http://3am-software.com/bio/matt/
Cupertino, CA              disclaimer: I avow all knowledge of this message.