Something's wrong with PowerPC SMP

[Michael Lorenz <macallan%netbsd.org@localhost>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello,

as the subject says - PowerPC SMP seems to have problems. My dual G4  
runs into UVM or filesystem related locking KASSERT()s as soon as  
there's concurrent disk IO, build.sh -j4 sooner or later triggers it.  
Building kernels usually works, building anything bigger pretty much  
always ends up in a KASSERT(curthread==whatever) panic, either in  
some vfs code or uvm_pagefault_internal(). ( sorry for the 10 finger  
copy, I can provide details and backtraces if necessary ).
The funny thing is:
- - no such problems with UP kernels, my G3 manages to build entire  
userlands without a hitch. Takes a while but it gets there.
- - no such problems on sparc64, my U60 manages just finished userlands  
for sparc64 and macppc with -j4 without trouble.
Therefore I think it's safe to assume that the cause is in the  
powerpc-specific SMP code. Probably the locking primitives? I'd  
appreciate input from someone with more lowlevel PowerPC clue than me.

have fun
Michael

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (Darwin)

iQEVAwUBSNbjlspnzkX8Yg2nAQIZlggAqN8C0XBQzHbGetM1KyCrP0D6sBSADRry
CmjU+FOEUiP/ur1RknneUf6M2MF8kHAPAMTaSHOlYPPjZdXrzEZTeWBw4fAe0yGX
CmcYgTjmuk5ywUfH47z94TZ6EuuQJpYZWrbJNlG0KMwU6X94PV1NymBkU7pOHu6i
S1m+TnVjjy5J0hjS07Ch6EEnAHAIr4Fcm5ZvtqbwPDTyj5+jNeKXMHopH6/EMCRv
WcnHOFd2epVptpL2uWE/5rQ7X9IuNH6NH+KL0fDXu/h30jZjaMpLHmQlZ6ytwUjT
yrguQat92JqFeat8FTZcwsXd7oJcRmSGoxn4cedYV2w9Vn6QOGADSA==
=T6Jn
-----END PGP SIGNATURE-----