After running successfully the FreeMiNT NFS client on ARAnyM, I tried to do the same thing with the FireBee.
mfro wrote:There must be something else going wrong with NFS on the Firebee.
Definitely. It absolutely does not work as expected, with huge timeouts.
First of all: there is an issue with my recent build of ping on the FireBee (both 68000 and ColdFire binaries). It is not able to resolve hostnames (while it works fine on ARAnyM). I tracked the problem down to gethostbyname(). This is currently discussed on the MiNT Mailing List. I suspect a MiNTLib regression.
As a workaround, the old ping binary from EasyMiNT works fine on the FireBee.
BlankVector wrote:You can find a ColdFire mount_nfs there:
That new build of mount_nfs does not work (both 68000 and ColdFire binaries).
It always fails with:
Code: Select all
do_nfs_mount: RPC: Timed out
./mount_nfs-new68k: could not do NFS mount
Again, as a workaround, the old mout_nfs from SpareMiNT works fine.
I suspect it is the same problem as ping, since the symptoms are the same. This will be discussed on the MiNT Mailing List, after the ping issue.
jfl wrote:I finally was able to mount an NFS share on my FireBee with the 68k mount_nfs from SpareMiNT and even if it is dead slow it kind of works.
Same for me. There are huge timeouts, up to 1 minute per operation. During that time, the client process hangs and nothing happens. Sometimes it is a bit faster, some other times operations completely fail after a timeout.
The situation is exactly the same with the ColdFire nfs.xfs or the 68000 one.
There is definitely something severely broken about NFS inside the ColdFire kernel.
On the other side, Samba and OpenSSH work perfectly on the FireBee, so there is no general network problem on the FireBee.