nfs: extended buffer length for eeproo 100

PATCH Chuck Lever 1 participant 1 message Jun 20, 2008 2 messages not in the archive

Subsystems (2)

NFS client NFS server

lore

Share & tools

lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

Chuck Lever

author

2008-06-20 16:51 UTC #2 lore raw

Quoted 18 lines · Guillaume Rousse

And old patch still present in mandriva nfs-utils package, from an unknown author

Signature · 4 lines

Guillaume Rousse
Moyens Informatiques - INRIA Futurs
Tel: 01 69 35 69 62
diff -uNr nfs-utils-0.2.1/support/nfs/rpcmisc.c nfs-utils-0.2.1-eepro100/support/nfs/rpcmisc.c

support/nfs/rpcmisc.c +0 -0

13:51:45 2001

support/nfs/rpcmisc.c +2 -1

@@ -176,7 +176,8 @@
 		int sblen, rblen;
 
 		/* 1024 for rpc & transport overheads */
-		sblen = rblen = socksz + 1024;
+		/* but, 2048 needed for eepro100 */
+		sblen = rblen = socksz + 2048;
 		if (setsockopt(s, SOL_SOCKET, SO_SNDBUF, &sblen, sizeof sblen) < 0

Quoted 2 lines

		    setsockopt(s, SOL_SOCKET, SO_RCVBUF, &rblen, sizeof rblen) < 0)
			xlog(L_ERROR, "setsockopt failed: %s\n", strerror(errno));

I can't see why user-space applications would need a larger socket buffer for a specific NIC.

However, this is moot because this whole section of code has been ifdef'd out since 2001. I think you can safely drop this one.

--
Chuck Lever
chuck[dot]lever[at]oracle[dot]com