The SKGE driver needs to return 0 from SkGeChangeMtu() on success.
Signed-off-by: John W. Linville <linville@xxxxxxxxxxxxx>
---
The proper sucessful return code for the change_mtu() method is zero.
For some reason, SkGeChangeMtu() is returning the new mtu value
instead. The comments would seem to indicate past problems, but the
current correct behaviour is clear.
drivers/net/sk98lin/skge.c | 14 +-------------
1 files changed, 1 insertion(+), 13 deletions(-)
--- 1.54/drivers/net/sk98lin/skge.c 2004-11-03 17:31:05 -05:00
+++ 1.55/drivers/net/sk98lin/skge.c 2004-11-18 11:12:36 -05:00
@@ -2849,19 +2849,7 @@
SkEventDispatcher(pAC, pAC->IoBase);
spin_unlock_irqrestore(&pAC->SlowPathLock, Flags);
- /*
- ** While testing this driver with latest kernel 2.5 (2.5.70), it
- ** seems as if upper layers have a problem to handle a successful
- ** return value of '0'. If such a zero is returned, the complete
- ** system hangs for several minutes (!), which is in acceptable.
- **
- ** Currently it is not clear, what the exact reason for this problem
- ** is. The implemented workaround for 2.5 is to return the desired
- ** new MTU size if all needed changes for the new MTU size where
- ** performed. In kernels 2.2 and 2.4, a zero value is returned,
- ** which indicates the successful change of the mtu-size.
- */
- return NewMtu;
+ return 0;
} /* SkGeChangeMtu */
|