On Mon, 2019-02-18 at 13:25 +0100, Greg KH wrote:
On Fri, Feb 08, 2019 at 03:46:26PM -0800, Jeff Kirsher wrote:
On Sat, 2019-02-02 at 12:28 +0100, Greg KH wrote:
On Fri, Feb 01, 2019 at 12:50:27PM -0800, Jeff Kirsher wrote:
From: Anirudh Venkataramanan anirudh.venkataramanan@intel.com
This is a backport of the mainline commit b36c598c999c ("ice: Updates to Tx scheduler code"). This change is required for the driver to work with the latest firmware. Without this patch, the driver fails probe.
Note, this is a big change. You can't have these systems just run the old firmware? The fact that you require a kernel change for newer firmware, isn't the best, you don't provide backwards compatibility somehow?
I'm all for adding new device ids and the like to stable kernels, but to backport new functionality just to get newer hardware to work on older kernels, is not the goal of the LTS kernel trees. Why can't you just have users use 4.20 or newer? There shouldn't be anything keeping them from doing that and just updating to the latest stable release every 4 months or so, right?
Sorry Greg, I am not ignoring you. I have been trying to get more information and reasoning from others within Intel which needed this support within the 4.19 kernel. I am hoping I will have a response for you by Monday.
It's been more than one "Monday" now, so I'm dropping this series from my patch queue.
That is fine.