On Tue, Sep 30, 2008 at 11:57:03PM +0900, Izumi Tsutsui wrote: > wrstuden%netbsd.org@localhost wrote: > > > > > Sorry, I forgot to mention I'm using a fixsa userland. So you'll need a > > > > fixsa libpthread. :-) > > > > > > Maybe it would be worth to run the regress tests with a plain 4.0 pthread, > > > and compare outputs on 4.0 and revivesa kernels. > > > > That would be fine. That would avoid the concern I had, which is that the > > 4.0 libpthread had problems. ;-) Since we're keeping it constant in the > > testing, the new kernel doesn't get blamed for a broken library. :-) > > - "make regress" in src/regress/lib/libpthread on stock 4.0 enviroment > with a revivesa kernel is passed successfully. Great! Thanks! > - "make test" in pkgsrc/lang/perl5 on the same environment > works similar with 4.0 kernel (fails in 4 thread tests as well). Great! > - An assertion caused by firefox3 on 4.0 environment I reporeted > can't be reproducible 100%. (some race condition?) Yes. I believe the problem is in delivering two or more unblocked upcalls at once. The second one comes out garbled. Last night, I looked at the code, and I can't see a substantive difference between the two. I'll try again tonight. I also have a test case program in mind. > - Maybe it's much better to do these tests on any SMP box > (I don't have MP machines) and amd64 (or any other ports). That would probably be good. I have tried the tests on an SMP box (VMware Fusion 2-CPU VM i386), but I have not tried amd64 or other ports. > - sysctl kernl.no_sa_support returns 1 even on revivesa. > Is it better to make it writable? Yes, thank you. I forgot about that. Take care, Bill
Attachment:
pgp6IPm_E0FuP.pgp
Description: PGP signature