lkml.org 
[lkml]   [1996]   [Oct]   [17]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    /
    Date
    From
    SubjectRe: Cyrix cpuid'n stuff...
    On Wed, 16 Oct 1996, Johnny Stenback wrote:

    > Yeah, but there's a way around that, there is (as posted in some
    > patches here a couple of times already) a way to determine if the cpu
    > is intel or cyrix before touching those ports are touched, it's done
    > by setting eflags to 0, then doing div xxx,xxx and then checking if
    > other than the 2:nd bit in eflags is 1, if it is then it's a some kind
    > of non-intel cpu (this isn't necessarily the exact way to do it but at
    > least it's in that direction). After this check those ports can be
    > safely (?) touched, I don't say that the cyrix way of configures these
    > registers isn't dumb but it should be doable without the possibility
    > to break other cpu:s...

    Yeah, the Cyrix changes flags in ways that other chips don't which
    makes a rough and ready test. It isn't that easy to add features,
    especially when you are the smaller player, casting in silicon,
    entering a volume market where you need compatibility with all kinds
    of existing crap code and can't a mistake. "Should" is not a good
    basis for fundamental design decisions :-).

    > > P.S. Someone was having problem with the magic number weren't they?
    > > You set the MAPEN to 1. You don't set each bit in MAPEN to 1...
    >
    > Yeah, it was me, I found out, guess if I feel STUPID. Hopefylly I will
    > learn (to read) from my mistakes...

    Don't worry. I did exactly the same thing. It seemed _obvious_ that
    since MAPEN was a set of bits in a byte that they were talking about
    setting bits... Nope... :-(

    Mike

    P.S. I have the patches for Cyrix recognition and use of the variable
    size paging mechanism (VSPM) working happily. When I roll them into
    a reasonably recent kernel I'll let Linus have them. (Oh, I didn't
    bother differentiating between IBM and Cyrix badged parts. It doesn't
    seem worth the trouble to me?)

    --
    .----------------------------------------------------------------------.
    | Mike Jagdis | Internet: mailto:mike@roan.co.uk |
    | Roan Technology Ltd. | |
    | 54A Peach Street, Wokingham | Telephone: +44 118 989 0403 |
    | RG40 1XG, ENGLAND | Fax: +44 118 989 1195 |
    `----------------------------------------------------------------------'


    \
     
     \ /
      Last update: 2005-03-22 13:37    [W:0.018 / U:209.680 seconds]
    ©2003-2016 Jasper Spaans. hosted at Digital OceanAdvertise on this site