> I seem to remember a big stink a while back about unimplemented system calls, > including 103 & 119. This was fixed in later stable kernel snapshots. I'm > just guessing here, but it could be that libc6 2.0.95 is using more of these > syscalls than previous versions, so if you don't have the implementation from > the latest kernels you're sunk.
There are a few more syscalls in the newest code which won't report errors anymore but won't do anything, which will be in the next snapshot. The problem is presumably that glibc is testing for existance but not usefulness of syscalls and so it finds lots of which aren't implemented, when in reality it would be nice if it ignored them. (The reason is, or was, some things needed for SunOS and not for Linux) In some cases the messages are harmless but annoying, in others we've implemented things. I think those are fixed in what will be the next snapshot -D
