I'd like to hear if anyone has any thoughts/comments/etc on the code in the attachment. The name of the project is the "ethernet speaker system" (see google). Basically, what I'm thinking is that--using this code as a starting point--the MI audio system could be extended to provide allocation of virtual audio devices in the same way as ttys are allocated. In theory, it could even be implemented as a puffs filesystem (using symlinks for the multiple devices) for greater flexibility. Now ...

On Fri, 2 Mar 2007 22:00:25 +0100 The link above contains the following link to the paper: http://cm.bell-labs.com/sys/doc/lexnames.html That goes to a page that says that the document was not found. Do you have a more recent/active link? (Should update the netbsd projects page with it, too.) Cheers, -- Andrew

I'm trying to modload -s a lkm and I'm getting modload: /dev/ksyms: Device not configured (I need the symbol table loading for use by another lkm). This is GENERIC 3.1_STABLE with the lkm compiled up against the same source tree. The same lkm source and NetBSD tree code yields a loadable lkm elsewhere (not tried the same kernel, but GENERIC has "options LKM"). I can't see what's different except for this machine is using grub to dual boot with Linux. I'm doing a kernel load with "kernel ...

On Fri, 2 Mar 2007 20:49:31 +0000 (GMT) Hmm odd, I do see in GENERIC: pseudo-device ksyms # /dev/ksyms You also should have: cr--r--r-- 1 root wheel 101, 0 Oct 5 2004 /dev/ksyms However I wonder if this could be related (from ksyms(4) man page): BUGS No LKM modules can be loaded or unloaded while /dev/ksyms is open. Although the proper error shouldn't be "not configured" in this case normally, and I'm not sure what could be holding the device open if ...

So long as it can be done without an ABI change. -- thorpej

Sorry for having sent this with a bogus address yesterday. Also, i got another error plus those mentioned yesterday continued after the parity rebuild finished. Yesterday, I got the following error while rebuilding RAID parity. Even after the rebuild finished, the errors contiued, apperantly once per minute. ahd1: ahd_timeout ahd1: Timedout SCB already complete. Interrupts may not be functioning. After reboot, I got another one while reconstructing a hosed component: ahd1: Transmission ...

Yes. I would like to have a single callback mechanism based on continuations. Softintrs and callout should go away as an API. -- thorpej

Yeah, I've noticed it still doesn't work right. It's gone from non-functional to usable, but tcpdump still shows bogus packets with invalid checksums and performance is bad. I'll try the changes you suggested. -Tobias

Possibly, and also add the variant used by HFS+, if it is not exactly equivalent to any of the above. However, since these conversion tables are large, maybe we should limit ourselves to the IN need not be NUL-terminated, in which case OUT will not be eiter. Neither the pathname component passed to VOP_LOOKUP, nor the on disk structure of at leaset HFS+ are NUL-terminated, so it makes little Omitting outlen? Actually, I don't care much about the order of arguments. I patterned it ...