Re: [PATCH] wireless: Fix ethtool stats and other ops.

[Stanislaw Gruszka <sgruszka@redhat.com>]

On Wed, Dec 05, 2012 at 10:02:00AM -0800, Ben Greear wrote:
> On 12/05/2012 09:57 AM, Bjørn Mork wrote:
> >Ben Greear <greearb@candelatech.com> writes:
> >
> >>net/core/dev.c now assigns a default ethtool ops, so
> >>the net/wireless/core.c check for existing ops is always true
> >>so the wireless ops would never be assigned.
> >>
> >>Simply remove the check for existing ops and always assign
> >>the wireless ops.
> >>
> >>Signed-off-by: Ben Greear <greearb@candelatech.com>
> >>---
> >>  net/wireless/core.c |    3 +--
> >>  1 files changed, 1 insertions(+), 2 deletions(-)
> >>
> >>diff --git a/net/wireless/core.c b/net/wireless/core.c
> >>index 4e6fe62..6309699 100644
> >>--- a/net/wireless/core.c
> >>+++ b/net/wireless/core.c
> >>@@ -863,8 +863,7 @@ static int cfg80211_netdev_notifier_call(struct notifier_block *nb,
> >>  		/* allow mac80211 to determine the timeout */
> >>  		wdev->ps_timeout = -1;
> >>
> >>-		if (!dev->ethtool_ops)
> >>-			dev->ethtool_ops = &cfg80211_ethtool_ops;
> >>+		dev->ethtool_ops = &cfg80211_ethtool_ops;
> >>
> >>  		if ((wdev->iftype == NL80211_IFTYPE_STATION ||
> >>  		     wdev->iftype == NL80211_IFTYPE_P2P_CLIENT ||
> >
> >
> >Won't this break drivers which for some reason have their own
> >ethtool_ops?
> 
> I guess it will.  What a mess.
> 
> Maybe we could assign individual method pointers in the ethtool_ops
> struct if it already exists (and if those pointers are NULL)?

We should probably assing ethtool_ops before wireless core will
call alloc_netdev.

Stanislaw