On Thu, Jun 05, 2008 at 05:46:58AM -0700, Dale Farnsworth wrote:

quoted text
> > The two 64-bit MIB registers (good_octets_received and
> > good_octets_sent) can trigger weird behavior if reads are done
> > from their 32-bit top parts.  This patch avoids reading the top
> > halves entirely.
> > 
> > Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
> > ---
> >  drivers/net/mv643xx_eth.c |    4 ++--
> >  1 files changed, 2 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/net/mv643xx_eth.c b/drivers/net/mv643xx_eth.c
> > index 0f8e79f..54c9c76 100644
> > --- a/drivers/net/mv643xx_eth.c
> > +++ b/drivers/net/mv643xx_eth.c
> > @@ -983,7 +983,7 @@ static void mib_counters_update(struct mv643xx_eth_private *mep)
> >  	struct mib_counters *p = &mep->mib_counters;
> >  
> >  	p->good_octets_received += mib_read(mep, 0x00);
> > -	p->good_octets_received += (u64)mib_read(mep, 0x04) << 32;
> > +//	p->good_octets_received += (u64)mib_read(mep, 0x04) << 32;
> >  	p->bad_octets_received += mib_read(mep, 0x08);
> >  	p->internal_mac_transmit_err += mib_read(mep, 0x0c);
> >  	p->good_frames_received += mib_read(mep, 0x10);
> > @@ -997,7 +997,7 @@ static void mib_counters_update(struct mv643xx_eth_private *mep)
> >  	p->frames_512_to_1023_octets += mib_read(mep, 0x30);
> >  	p->frames_1024_to_max_octets += mib_read(mep, 0x34);
> >  	p->good_octets_sent += mib_read(mep, 0x38);
> > -	p->good_octets_sent += (u64)mib_read(mep, 0x3c) << 32;
> > +//	p->good_octets_sent += (u64)mib_read(mep, 0x3c) << 32;
> >  	p->good_frames_sent += mib_read(mep, 0x40);
> >  	p->excessive_collision += mib_read(mep, 0x44);
> >  	p->multicast_frames_sent += mib_read(mep, 0x48);
> 
> I'd add a comment referring to the weird behavior.  Also, aren't
> C++-style comments discouraged?

ACK.  Let me pull out this patch for now and come up with a better one.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html