On Tue, 1 Sep 2009 23:48:01 -0600
Grant Grundler [off-list ref] wrote:

On Mon, Aug 31, 2009 at 10:50:53PM -0700, Stephen Hemminger wrote:

quoted

Signed-off-by: Stephen Hemminger <redacted>
---
 drivers/net/tulip/de2104x.c     |    3 ++-
 drivers/net/tulip/de4x5.c       |   11 +++++------
 drivers/net/tulip/dmfe.c        |    5 +++--
 drivers/net/tulip/tulip_core.c  |    5 +++--
 drivers/net/tulip/uli526x.c     |    6 ++++--
 drivers/net/tulip/winbond-840.c |    4 ++--
 drivers/net/tulip/xircom_cb.c   |    6 ++++--
 7 files changed, 23 insertions(+), 17 deletions(-)

--- a/drivers/net/tulip/de2104x.c	2009-08-29 23:10:19.609527457 -0700
+++ b/drivers/net/tulip/de2104x.c	2009-08-29 23:12:54.117527675 -0700

@@ -599,7 +599,8 @@ next:
 		netif_wake_queue(de->dev);
 }
 
-static int de_start_xmit (struct sk_buff *skb, struct net_device *dev)
+static netdev_tx_t de_start_xmit (struct sk_buff *skb,
+					struct net_device *dev)

Stephen,
The patches look harmless and I'm inclined to ACK them...but google can't find
any context or explanation for "netdev_tx_t". URL to email which proposed
netdev_tx_t?

thanks,
grant

From patch series intro:

This is a little change over a lot of files. It changes the
return value of network device transmit from an integer to
an enum type (like irqreturn_t).  This allows compiler to warn
about stupid code that tries to return -ENOBUFS or other mistakes
like that. All the code that did that is gone, but it would
be better to try and do some decent type checking.

This patch converts all the x86 drivers (except staging).
Other drivers will work but produce a warning until converted.