more weird bugs with mmap-ing via NFS

[Mikhail Teterin]

вівторок 21 березень 2006 17:56, Matthew Dillon Ви написали:
>     For UDP mounts, 65536 is too large (the UDP data length can
>     only be 65536 bytes.  For that matter, the *IP* packet itself can
>     not exceed 65535 bytes.  So 65536 will not work with a UDP mount.

Well, then the mount should fail with EINVALID or something, maybe, it just 
quietly reduces the size to the maximum? However, the problem is the same 
with 32K and 16K packets. You give good reasons for such sizes being slow, 
but this does not explain, why it does not work at all...

>     The second problem is related to the network driver.  The packet MTU
>     is 1500, which means, typically, a limit of around 1460-1480 payload
>     bytes per packet.

Yes, unfortunately, your 2004 patches, that allowed one to specify a different 
MTU for different hosts on the local network never made it to any tree :-(

>     I would stick to an NFS block size of 8K or 16K.  Frankly, there is
>     no real reason to use a larger block size.

The problem is about same with 32K and 16K packets. With 8K packets, the thing 
kind-of works (although trying to `systat -vm' still stalls disk access), but 
the outgoing traffic is over 20Mb/s on average -- MUCH more, than the writing 
program itself generates.

Thanks for your help. Yours,

	-mi