On 12/16/24 05:53, Przemek Kitszel wrote:
On 12/13/24 04:00, Jakub Kicinski wrote:
quoted
On Wed, 11 Dec 2024 14:32:14 -0800 Tony Nguyen wrote:
quoted
From: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Print the ethtool stats and skb diagnostic information as part of Tx hang
devlink health dump.
Move the declarations of ethtool functions that devlink health uses out
to a new file: ice_ethtool_common.h
To utilize our existing ethtool code in this context, convert it to
non-static.
This is going too far, user space is fully capable of capturing this
data. It gets a netlink notification when health reporter flips to
a bad state.
It really pays to split your patches into trivial vs controversial ones.
not so trivial for git anyway...
Will it be fine to merge this series without patch 6 (and 3) then?
we will have to resend, so I will remove just ethtool stats part for now
sorry for the noise
Patches 2, 4 and 5 are dependency for another health reporters that
Konrad did:
https://lore.kernel.org/intel-wired-lan/20241211110357.196167-1- konrad.knitter@intel.com
quoted
I think Jiri worked on a daemon what could capture more
data from user space ? I may be misremembering...
We would love to read more on that, then with more knowledge revisit
what to do about our needs covered by this patch.
still interested ofc