Current status of FreeBSD/MIPS

[Tim Wiess]

    Juli,
    I just posted a bunch of my mips errata for you to snag.
    http://www.zetaflops.net/~tim/errata

    R4000_3.0_2.2_MC_errata.ps.Z has a couple entries related to the
    ll/sc bugs. However it seems that issue was resolved in the R4400s
    since there is no mention of it in any of those docs.

    BTW, last time I checked, ftp.sgi.com still has all the old mips
    manuals and errata available.

    tim


> After many hours sitting and looking blankly at heisenbugs,
> printfs that would make data accesses work, etc., I finally
> sat down the other day and started looking for CPU errata.
> I found some mention of LL/SC errata that might affect my
> processor, and decided to replace atomic.h with a C version,
> and suddenly I'm getting to cpu0: messages, and beyond by
> some ways.  I've got about two large pages of SI_SUB listing
> to make things go through, which at this point means fixing
> up NetBSD's exception, etc., code to work, and getting stuff
> like interrupts working, and doing a RealPmap(tm) instead
> of just doing everything lazily out of KSEG0.  Here's what
> goes on, modulo debugging printfs and the like:
> 
> %%%
> Copyright (c) 1992-2003 The FreeBSD Project.
> Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994
>         The Regents of the University of California. All rights reserved.
> FreeBSD 5.0-CURRENT #376: Fri May 23 00:17:16 CDT 2003
>     jmallett at dalek:/usr/people/jmallett/mips.build/sgimips/usr/people/jmallet
t/p4/mips/sys/INDY
> MIPS R4400 CPU (0x460) Rev. 6.0 with MIPS R4010 FPC Rev. 0.0
> cpu0: 16KB/16B direct-mapped L1 Instruction cache, 48 TLB entries
> cpu0: 16KB/16B direct-mapped write-back L1 Data cache
> cpu0: 0KB/128B direct-mapped write-back L2 Unified cache
> machine: SGI-IP22
> ARCS memory = 786432 (768 KB)
> Loaded program memory = 1449984 (1416 KB)
> avail memory = 131964928 (125 MB)
> real memory = 134201344 (127 MB)
> %%%
> 
> I'd like a bit of help with one thing, if anyone could help...
> 
> MIPS redid their website, and in the process seems to have lost
> all their R4K docs.  This includes all the CPU errata from that
> era, and I've looked at some Google-Cached HTML of these things,
> but that's very hard to read, and I've had mixed results with
> doing that.  If someone has the errata PDFs and could send them
> to me, that would be really great.  And also, if someone knows
> much about the LL/SC errata, I can send off a copy of the atomic.h
> that I'd *like* to be doing, and you can maybe tell me what
> changes need made :)
> 
> Thanx much,
> juli.