> Got this while writing my last message to Linux kernel :-/
> Seems to be not-fatal, but it is odd that kswapd caused it.
I should have posted this earlier; but better late than never.
It's not actually kswapd that causes it; kswapd just happened to be
the current process when shrink_dcache() was called. (Actually
shrink_dcache() is often caused by kswapd.)
The oops is caused because dentry->d_inode is -1 (rather than NULL or
a pointer to a struct inode), but I can't determine more than that
from the Oops.
pjm.