Re: config.guess for ev68 [Was: Re: glibc-2.2.4pre2 success onalpha]

[Andreas Jaeger <aj at suse dot de>]

Andreas Jaeger <aj@suse.de> writes:

> Ben Elliston <bje@redhat.com> writes:
>
>>    > The cpu detect doesn't work for ev68, which has a different 
>>    > amask value (0x1307) from the ev67 (0x307), so 
>>    > scripts/config.guess returns plain alpha-unknown-linux-gnu.
>>    > I'd imagine that it would be better to default an ev68 to
>>    > host type alphaev67 until/whenever alphaev68 support is added.
>>    
>>    Please first send your patch to the config maintainers (CC'ed).  If
>>    they approve it, we'll import the new config.guess version in
>>    glibc.
>>
>> I'm going to discard this patch because the current version of
>> config.guess reworks all of this stuff -- please try it out and see
>> how it goes.  It uses /proc/cpuinfo, so if /proc/cpuinfo says the
>> right thing on the ev68, then it's a no-op.
>
> I just had a look at the current version and the patch still seems to
> apply.  Did you commit it already?

Ben,

I have to correct me: the first part of the patch is still needed -
the one for alpha:OSF1, the linux patch should work fine.

But as Ulrich asked: Why are you not using alphaev68?  Currently you
do:
    alpha:Linux:*:*)
        case `sed -n '/^cpu model/s/^.*: \(.*\)/\1/p' < /proc/cpuinfo` in
          EV5)   UNAME_MACHINE=alphaev5 ;;
          EV56)  UNAME_MACHINE=alphaev56 ;;
          PCA56) UNAME_MACHINE=alphapca56 ;;
          PCA57) UNAME_MACHINE=alphapca56 ;;
          EV6)   UNAME_MACHINE=alphaev6 ;;
          EV67)  UNAME_MACHINE=alphaev67 ;;
          EV68*) UNAME_MACHINE=alphaev67 ;;
        esac

I would change the last line to *68 - or is there a problem with this?

Ulrich, if we import this, we should add
sysdeps/alpha/alphaev68/Implies to glibc,

Andreas
-- 
 Andreas Jaeger
  SuSE Labs aj@suse.de
   private aj@arthur.inka.de
    http://www.suse.de/~aj