Subject: Re: "pmap_unwire: wiring ... didn't change!"
To: Izumi Tsutsui <tsutsui@ceres.dti.ne.jp>
From: Markus W Kilbinger <kilbi@rad.rwth-aachen.de>
List: port-cobalt
Date: 03/20/2005 22:09:51
>>>>> "Izumi" == Izumi Tsutsui <tsutsui@ceres.dti.ne.jp> writes:

    Izumi> How about this one? (adding cache flush before/after KSEG0 access)
    Izumi> [...]

I've just tested it togehther with chuq's "pmap_unwire: wiring ...
didn't change!" patch (is this still relevant/actual!?:

  http://mail-index.netbsd.org/port-mips/2005/02/12/0000.html

) on my qube2: This is the first time my several 100 mb data
copy/compare test script passes successfully, no data corruptions! :-)
... only a system hangup during the 4th run of that script (no further
system/kernel message for that :-/)

    Izumi> As discussed on port-mips this is not a "real" fix, but
    Izumi> IMHO this could be an acceptable compromise for current
    Izumi> pmap,

-> A very good compromise, no obvious performance loss! :-)

Any hint how to investigate the hangups further?

Markus.