ACK: [SRU][Focal][PATCH 0/2] Intel x710 LOMs do not work on Focal

Colin Ian King colin.king at canonical.com
Thu Sep 17 08:17:18 UTC 2020


On 16/09/2020 22:51, Jeff Lane wrote:
> BugLink: https://bugs.launchpad.net/bugs/1893956
> 
> [IMPACT]
> The Intel x710-TM4 is one of the latest 10GbE controllers from intel using the
> i40e driver. This particular 4 port comes in a 2x2 arrangement: 2x SFP+ and 2x
> RJ-45. This card is enabled in 5.4 via the inbox version of the i40e driver,
> and hwinfo does show both sides of the card but the kernel only sees the two
> SFP+ ports and cannot address or use the two copper ports.
> 
> These devices have up to 4 ports, but only two seem to work and this has been
> seen on at least two different systems using LOMs based on this.
> 
> This is currently blocking certification for one of our hardware partners.
> 
> [FIX]
> This list of Patches enables the latest X710 controller fully:
> 
> * 3df5b9a6a9ec i40e: enable X710 support
> * d80a476f4a12 i40e: Fix LED blinking flow for X710T*L devices
> 
> These are all clean cherry picks from 5.5 into 5.4
> 
> My tree with these can be found here:
> https://code.launchpad.net/~bladernr/ubuntu/+source/linux/+git/focal/+ref/1893956-update-i40e-intel-lom
> 
> [TESTCASE]
> 
> i40e should load successfully, all ports on Carlsville x710 LOM should be
> visible, can be configured, and can pass traffic at full speed.
> 
> Test kernels have been verified by the vendor reporting the issue that all
> network ports on the LOM are accessible and pass traffic at expected speeds
> once the fix is applied.
> 
> [REGRESSION RISK]
> Should be pretty low. This is a small patch set to a single driver that only
> enables a newer chipset.
> 
> Alice Michael (1):
>   i40e: enable X710 support
> 
> Damian Milosek (1):
>   i40e: Fix LED blinking flow for X710T*L devices
> 
>  drivers/net/ethernet/intel/i40e/i40e_common.c | 31 +++++++++++++++++--
>  drivers/net/ethernet/intel/i40e/i40e_devids.h |  2 ++
>  2 files changed, 30 insertions(+), 3 deletions(-)
> 

Clean cherry picks, been tested by vendor. Affects LEDs, so if this
breaks other devices I guess the regression is not a major showstopper.

Acked-by: Colin Ian King <colin.king at canonical.com>



More information about the kernel-team mailing list