Ubuntu 16.10 KVM: kernel 4.8 can not enable SRIOV for power systems

Bug #1624961 reported by bugproxy
8
This bug affects 1 person
Affects Status Importance Assigned to Milestone
linux (Ubuntu)
Fix Released
Undecided
Tim Gardner
Yakkety
Fix Released
Undecided
Tim Gardner

Bug Description

---Problem Description---
Using Ubuntu mainline kernel
http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.8-rc6/
if we try to enable SRIOV in Mellanox CX4 card it will fail.

# uname -r
4.8.0-040800rc6-generic
root@c158f2u09os:~# echo 7 > /sys/class/net/enP5p1s0f0/device/sriov_numvfs
-bash: echo: write error: Cannot allocate memory

logs show:
Sep 16 00:52:46 c158f2u09os kernel: [ 1419.778118] mlx5_core 0005:01:00.0: not enough MMIO resources for SR-IOV
Sep 16 00:52:46 c158f2u09os kernel: [ 1419.778238] mlx5_core 0005:01:00.0: enable sriov failed -12
Sep 16 00:52:46 c158f2u09os kernel: [ 1419.778604] mlx5_core 0005:01:00.0: mlx5_core_sriov_enable failed -12

---uname output---
4.8.0-040800rc6-generic

---Additional Hardware Info---
To recreate this you need power system and a Mellanox card that can be used for SRIOV.

Machine Type = P8

---Steps to Reproduce---
 with a Mellanox card like CX4 with SRIOV enabled.
just do this echo to the PF interface in this case was enP5p1s0f0.
echo 1 > /sys/class/net/enP5p1s0f0/device

If I take kernel 4.8 and I revert the following patch I can get it to work again:

http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/patch/arch/powerpc/platforms/powernv/pci-ioda.c?id=5958d19a143eb229e9ece20bd4c781ad41cb7d24

 From 5958d19a143eb229e9ece20bd4c781ad41cb7d24 Mon Sep 17 00:00:00 2001
 From: Benjamin Herrenschmidt <email address hidden>
 Date: Fri, 8 Jul 2016 15:55:43 +1000
 Subject: powerpc/pnv/pci: Fix incorrect PE reservation attempt on some 64-bit BARs

bugproxy (bugproxy)
tags: added: architecture-ppc64le bugnameltc-146447 severity-critical targetmilestone-inin1610
Changed in ubuntu:
assignee: nobody → Taco Screen team (taco-screen-team)
affects: ubuntu → linux (Ubuntu)
Revision history for this message
Leann Ogasawara (leannogasawara) wrote :

Can you also confirm this with an Ubuntu based v4.8-rc6 kernel (currently available in yakkety-proposed)

https://launchpad.net/ubuntu/+source/linux/4.8.0-11.12

Based on the bug description, it appears this is a regression which originates from upstream and was introduced in 4.8-rc2:

linux$ git describe --contains 5958d19a143eb229e9ece20bd4c781ad41cb7d24
v4.8-rc2~20^2~9

Upstream is now at v4.8-rc7. If the issue remains, please report this upstream as well.

Revision history for this message
Breno Leitão (breno-leitao) wrote :

Leann,

This issue is already fixed upstream with commit b79331a5eb9f96e4dfd216974581168ec4c8a4d4

From b79331a5eb9f96e4dfd216974581168ec4c8a4d4 Mon Sep 17 00:00:00 2001
From: Russell Currey <email address hidden>
Date: Wed, 14 Sep 2016 16:37:17 +1000
Subject: powerpc/powernv/pci: Fix m64 checks for SR-IOV and window alignment

Looking at the tags, it seems that it will make 4.8:

$ git tag --contains b79331a5eb9f96e4dfd216974581168ec4c8a4d4
v4.8-rc8

Revision history for this message
bugproxy (bugproxy) wrote : Comment bridged from LTC Bugzilla

------- Comment From <email address hidden> 2016-09-26 11:17 EDT-------
Hi Ubuntu
patch is upstream.
http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/patch/arch/powerpc/platforms/powernv?id=b79331a5eb9f96e4dfd216974581168ec4c8a4d4

Revision history for this message
Tim Gardner (timg-tpi) wrote :

Marking fix released since the bug fix commit is in 4.8-rc8 which will be included in the GA kernel.

Changed in linux (Ubuntu Yakkety):
assignee: Taco Screen team (taco-screen-team) → Tim Gardner (timg-tpi)
status: New → Fix Released
Revision history for this message
bugproxy (bugproxy) wrote :

------- Comment From <email address hidden> 2016-09-27 04:23 EDT-------
(In reply to comment #15)
> Marking fix released since the bug fix commit is in 4.8-rc8 which will be
> included in the GA kernel.

I am seeing the same issue reported in todays daily build. When will it make into daily builds?

------- Comment From <email address hidden> 2016-09-27 04:27 EDT-------
(In reply to comment #16)
> (In reply to comment #15)
> > Marking fix released since the bug fix commit is in 4.8-rc8 which will be
> > included in the GA kernel.
>
> I am seeing the same issue reported in todays daily build. When will it make
> into daily builds?

kernel level what I see today in 16.10 after upgrade:

4.8.0-17-generic

Revision history for this message
Tim Gardner (timg-tpi) wrote :

4.8.0-17-generic is based on 4.8-rc7. The next kernel upload will be based on 4.8-rc8 (which has the fix commit)

Revision history for this message
bugproxy (bugproxy) wrote :

------- Comment From <email address hidden> 2016-09-27 09:02 EDT-------
(In reply to comment #18)
> 4.8.0-17-generic is based on 4.8-rc7. The next kernel upload will be based
> on 4.8-rc8 (which has the fix commit)

would it be tomorrows daily build?

Revision history for this message
Tim Gardner (timg-tpi) wrote :

It'll get uploaded by Thursday Sept 29 at the latest.

Revision history for this message
bugproxy (bugproxy) wrote :

------- Comment From <email address hidden> 2016-09-30 07:15 EDT-------
(In reply to comment #20)
> It'll get uploaded by Thursday Sept 29 at the latest.

I still dont see daily build with kernel which has this bug fix commit. When it will be available?

Revision history for this message
bugproxy (bugproxy) wrote :

------- Comment From <email address hidden> 2016-10-07 09:49 EDT-------
*** Bug 146293 has been marked as a duplicate of this bug. ***

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.