(Illustration by Gaich Muramatsu)
On Thu, Sep 21, 2000 at 04:01:42PM -0500, Corey R. Halpin wrote: > What do I make of these? > Sep 21 15:55:56 hobbes kernel: coda_pioctl: Venus returns: -107 for (0x7f000000,0x1,0x1) Venus returns ENOTCONN, I have seen this a couple of times, but haven't even looked at where it is coming from. > Sep 21 15:56:01 hobbes kernel: psdev_read: Venus read 8464 bytes of 8468 in message This is because there is probably some skew between a structure in the kernel and in venus. It is actually non-fatal, because venus doesn't really need all 8468 bytes. The latest test9-pre5 contains a fix for this, but has a nasty Oops. The patch for that is already sent off to Linus, so test9-pre6 or test9-final should be working like a charm. > I'm using cvs coda, linux-2.4.0-test8 with the reiserfs 3.6.17 patch Coda's kernel module accesses files on a very low level using generic_file_read/write/mmap. This is definitely broken for reiserfs on 2.2 kernels, and I am unsure if it would work reliably on 2.4. My best advise at the moment is to keep venus.cache in an ext2 formatted partition. JanReceived on 2000-09-22 11:43:08