Re: MP kernels are broken (RE: upgrading to 5.0.1)

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

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

Hello,

On Aug 17, 2009, at 10:44 PM, Larson, Timothy E. wrote:

> > How about problems with 5.0 MP?
>
> Looks like the same sorts of things.
>
> 0xd47aee20: at panic+0x228
> 0xd47aee70: at dbdma_start+0xac
> 0xd47aeea0: at wdc_obio_dma_start+0x14
> 0xd47aeeb0: at _wdc_ata_bio_start+0x830
> 0xd47aef00: at atabus_thread+0x2f4
> 0xd47aef40: at emptyidlespin+0x10
> 0xd47aef50: at _prop_dictionary_keysym32_pool+0xc9e2ede4
> Trap: kernel read DSI trap @ 0xe52a4884 by 0x4a5394 (DSISR  
> 0x40000000, err=14), lr 0x4a58d4
> Panic: trap
> Faulted in DDB; continuing...
> (and then it seems to loop)
>
> So I'm stuck using 1 CPU of a dual-proc box.  :(

Hmm, works here ( GigE with 2x 500MHz G4 ).
No clue what could be causing it - I have a similar setup as you do, 2  
disks at the ATA66 channel, a burner and a ZIP at the 2nd channel.  
What you could do is to disconnect the burner and the ZIP to see from  
which channel the panic comes. I also remember weird problems  
depending on how the devices are jumpered - do you have then set to  
cable select or master/slave? It might be worth trying the other way  
around.

have fun
Michael

PS: my G4 didn't run a single CPU kernel since we got things working  
again in the ppcoea-renovation branch, quite some time before 5.0 was  
branched.

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

iQEVAwUBSoo3QMpnzkX8Yg2nAQK/RggAoY+dLGOcrk1E+dhh96TJVWq7eFWWxYyp
Qe7kugMQxPTQcYc1tp25v0JfE0bc0o/Tu5AdOdIAUPHN/jfp2Qz2be+1EPNWTLcS
ZBfFmkgUf3u+z5JTg0dfLa79WcOBot7jsuSwJnygwW5CpEgi59mflTyj9QCPMlOw
yHMPwzEwG4hxIzXDkZUUY/pDr8HNDIUf6r92Fl37X3mS30tVspuLQutit/vKXsng
/u9sPaZCGKGnD5sgaW+lyVy6LmAvcE4vgghGNDf5rU9YSnGtwbsP3jrHgJkJt0DS
gI+61u5p6KgY2EJT9r4NRT1k7EsBCp7gI3Xyyp2JcghhGsIKQkp/Rg==
=R05v
-----END PGP SIGNATURE-----