[OpenAFS-port-darwin] 1.4.1rc1 on Tiger killing the kernel.

[Jon Shier]

	So I've been running the 1.4.1rc1 OpenAFS release for Tiger (10.4.3)  
for a few days now on about 20 machines and have had a half dozen  
occurrences of users trying to access their AFS space (through save  
dialogs, browsing it, there doesn't seem to be a way that always  
causes a problem) and the machine beachballs, starting with whatever  
app they used to access their space. Eventually, as they try to use  
other apps, they beachball too, telling me that the OpenAFS kext is  
killing the kernel outright. Over the next few days I'll try running  
some automated scripts and see if I can find a replicable scenario.
	Here is the relevant content of the console.log relating to this crash.

Nov  3 14:46:52 Molecular-Biology-1340-G5-PowerMac-11 kernel[0]:  
Starting AFS cache scan...[211] waiting for afs_osi_ctxtp
Nov  3 14:46:52 Molecular-Biology-1340-G5-PowerMac-11 kernel[0]:  
[212] waiting for afs_osi_ctxtp
Nov  3 14:46:52 Molecular-Biology-1340-G5-PowerMac-11 kernel[0]:  
[213] waiting for afs_osi_ctxtp
Nov  3 14:46:52 Molecular-Biology-1340-G5-PowerMac-11 kernel[0]:  
[211] waiting for afs_osi_ctxtp
Nov  3 14:46:52 Molecular-Biology-1340-G5-PowerMac-11 kernel[0]:  
[212] waiting for afs_osi_ctxtp
Nov  3 14:46:52 Molecular-Biology-1340-G5-PowerMac-11 kernel[0]:  
[213] waiting for afs_osi_ctxtp
Nov  3 14:46:52 Molecular-Biology-1340-G5-PowerMac-11 kernel[0]:  
[211] waiting for afs_osi_ctxtp
Nov  3 14:46:52 Molecular-Biology-1340-G5-PowerMac-11 kernel[0]:  
[212] waiting for afs_osi_ctxtp
Nov  3 14:46:52 Molecular-Biology-1340-G5-PowerMac-11 kernel[0]:  
[211] waiting for afs_osi_ctxtp
Nov  3 14:46:52 Molecular-Biology-1340-G5-PowerMac-11 kernel[0]:  
[213] waiting for afs_osi_ctxtp

As you can see, it looks like there's still some problems with the  
cache code on Tiger.


Jon Shier
Iowa State University