[v3,net-next,00/10] Improve the taprio qdisc's relationship with its children

Message

Vladimir Oltean Aug. 1, 2023, 6:24 p.m. UTC

Changes in v3:
Fix ptp_mock compilation as module, fix small mistakes in selftests.

Changes in v2:
It was requested to add test cases for the taprio software and offload modes.
Those are patches 08 and 09.

That implies adding taprio offload support to netdevsim, which is patch 07.

In turn, that implies adding a PHC driver for netdevsim, which is patch 06.

v1 at:
https://lore.kernel.org/lkml/[e-mail]/

Original message:

Prompted by Vinicius' request to consolidate some child Qdisc
dereferences in taprio:
https://lore.kernel.org/netdev/[e-mail]/

I remembered that I had left some unfinished work in this Qdisc, namely
commit af7b29b1deaa ("Revert "net/sched: taprio: make qdisc_leaf() see
the per-netdev-queue pfifo child qdiscs"").

This patch set represents another stab at, essentially, what's in the
title. Not only does taprio not properly detect when it's grafted as a
non-root qdisc, but it also returns incorrect per-class stats.
Eventually, Vinicius' request is addressed too, although in a different
form than the one he requested (which was purely cosmetic).

Review from people more experienced with Qdiscs than me would be
appreciated. I tried my best to explain what I consider to be problems.
I am deliberately targeting net-next because the changes are too
invasive for net - they were reverted from stable once already.

Vladimir Oltean (10):
  net/sched: taprio: don't access q->qdiscs[] in unoffloaded mode during
    attach()
  net/sched: taprio: keep child Qdisc refcount elevated at 2 in offload
    mode
  net/sched: taprio: try again to report q->qdiscs[] to qdisc_leaf()
  net/sched: taprio: delete misleading comment about preallocating child
    qdiscs
  net/sched: taprio: dump class stats for the actual q->qdiscs[]
  net: ptp: create a mock-up PTP Hardware Clock driver
  net: netdevsim: use mock PHC driver
  net: netdevsim: mimic tc-taprio offload
  selftests/tc-testing: test that taprio can only be attached as root
  selftests/tc-testing: verify that a qdisc can be grafted onto a taprio
    class

 MAINTAINERS                                   |   7 +
 drivers/net/Kconfig                           |   1 +
 drivers/net/netdevsim/ethtool.c               |  11 ++
 drivers/net/netdevsim/netdev.c                |  38 +++-
 drivers/net/netdevsim/netdevsim.h             |   2 +
 drivers/ptp/Kconfig                           |  11 ++
 drivers/ptp/Makefile                          |   1 +
 drivers/ptp/ptp_mock.c                        | 175 ++++++++++++++++++
 include/linux/ptp_mock.h                      |  38 ++++
 net/sched/sch_taprio.c                        |  68 ++++---
 .../tc-testing/tc-tests/qdiscs/taprio.json    | 100 +++++++++-
 11 files changed, 423 insertions(+), 29 deletions(-)
 create mode 100644 drivers/ptp/ptp_mock.c
 create mode 100644 include/linux/ptp_mock.h

Comments

Paolo Abeni Aug. 3, 2023, 10:04 a.m. UTC | #1

Hi,

On Tue, 2023-08-01 at 21:24 +0300, Vladimir Oltean wrote:
> Changes in v3:
> Fix ptp_mock compilation as module, fix small mistakes in selftests.
> 
> Changes in v2:
> It was requested to add test cases for the taprio software and offload modes.
> Those are patches 08 and 09.
> 
> That implies adding taprio offload support to netdevsim, which is patch 07.
> 
> In turn, that implies adding a PHC driver for netdevsim, which is patch 06.
> 
> v1 at:
> https://lore.kernel.org/lkml/[e-mail]/
> 
> Original message:
> 
> Prompted by Vinicius' request to consolidate some child Qdisc
> dereferences in taprio:
> https://lore.kernel.org/netdev/[e-mail]/
> 
> I remembered that I had left some unfinished work in this Qdisc, namely
> commit af7b29b1deaa ("Revert "net/sched: taprio: make qdisc_leaf() see
> the per-netdev-queue pfifo child qdiscs"").
> 
> This patch set represents another stab at, essentially, what's in the
> title. Not only does taprio not properly detect when it's grafted as a
> non-root qdisc, but it also returns incorrect per-class stats.
> Eventually, Vinicius' request is addressed too, although in a different
> form than the one he requested (which was purely cosmetic).
> 
> Review from people more experienced with Qdiscs than me would be
> appreciated. I tried my best to explain what I consider to be problems.
> I am deliberately targeting net-next because the changes are too
> invasive for net - they were reverted from stable once already.

The series LGTM, modulo the minor comments from Kurt on patch 6/10. I
agree they can be handled with follow-up patches.

Keeping it a little longer on PW: it would be great if someone from the
tc crew could have a look!

Thanks!

Paolo