Jeff:
I will apply the process change to all future submittals.
I agree with your decision not to apply patch 5 "enable TSO".
Thanks,
ganesh
-------------------------------------------------
Ganesh Venkatesan
Network/Storage Division, Hillsboro, OR
-----Original Message-----
From: Jeff Garzik [mailto:jgarzik@xxxxxxxxx]
Sent: Thursday, May 27, 2004 10:48 AM
To: Venkatesan, Ganesh; Netdev
Subject: Re: [PATCH 2.4] e1000 0/7 driver update
ganesh.venkatesan@xxxxxxxxx wrote:
> Jeff:
>
> Patches against 2.4 tot for the latest version of e1000 (5.2.52). I
> verified that these patches apply*, the kernel builds fine and the
driver
> works as expected. Please apply these patches to the upstream kernel.
Two patch submission process changes:
1) include the "1/7" number _inside_ the "[PATCH ... 2.4]" brackets.
2) please CC netdev@xxxxxxxxxxx on patches. This allows for greater
peer review.
> * - p6/p7 apply with a -/+1 line offset. But this does not affect
> anything.
>
> p1.diff - clear auto-mdix mode when the link is forced to 10/100 mbps
> speeds on certain PHYs
> p2.diff - workaround for link LED staying ON even when there is no
link
> p3.diff - fix to determine link status correctly while using
fiber/serdes
> interface
> p4.diff - rewrite logic to estimate # of tx descriptors required for a
> given packet, shrink the span for which tx_lock is held
> p5.diff - enable TSO
> p6.diff - ethtool_ops support
> p7.diff - enable error logging
I applied all patches to 2.4 and 2.6 EXCEPT patch #5 "enable TSO".
Judging from your own comments, I am not confident that turning it on is
productive, with the lockups still occurring.
Jeff
|