[net-next,v2,0/4] net: improve core queue API handling while device is down

Message

Jakub Kicinski Feb. 6, 2025, 10:56 p.m. UTC

The core netdev_rx_queue_restart() doesn't currently take into account
that the device may be down. The current and proposed queue API
implementations deal with this by rejecting queue API calls while
the device is down. We can do better, in theory we can still allow
devmem binding when the device is down - we shouldn't stop and start
the queues just try to allocate the memory. The reason we allocate
the memory is that memory provider binding checks if any compatible
page pool has been created (page_pool_check_memory_provider()).

Alternatively we could reject installing MP while the device is down
but the MP assignment survives ifdown (so presumably MP doesn't cease
to exist while down), and in general we allow configuration while down.

Previously I thought we need this as a fix, but gve rejects page pool
calls while down, and so did Saeed in the patches he posted. So this
series just makes the core act more sensibly but practically should
be a noop for now.

v2:
 - allow also mode 2 and 3 (patch 3 is new)
v1: https://lore.kernel.org/[e-mail]

Jakub Kicinski (4):
  net: refactor netdev_rx_queue_restart() to use local qops
  net: devmem: don't call queue stop / start when the interface is down
  net: page_pool: avoid false positive warning if NAPI was never added
  netdevsim: allow normal queue reset while down

 include/net/netdev_queues.h              |  4 +++
 net/core/dev.h                           | 12 ++++++++
 drivers/net/netdevsim/netdev.c           | 10 +++----
 net/core/netdev_rx_queue.c               | 37 +++++++++++++-----------
 net/core/page_pool.c                     |  7 ++---
 tools/testing/selftests/net/nl_netdev.py | 18 +++++++++++-
 6 files changed, 59 insertions(+), 29 deletions(-)

Comments

[e-mail] Feb. 8, 2025, 4:20 a.m. UTC | #1

Hello:

This series was applied to netdev/net-next.git (main)
by Jakub Kicinski <[e-mail]>:

On Thu,  6 Feb 2025 14:56:34 -0800 you wrote:
> The core netdev_rx_queue_restart() doesn't currently take into account
> that the device may be down. The current and proposed queue API
> implementations deal with this by rejecting queue API calls while
> the device is down. We can do better, in theory we can still allow
> devmem binding when the device is down - we shouldn't stop and start
> the queues just try to allocate the memory. The reason we allocate
> the memory is that memory provider binding checks if any compatible
> page pool has been created (page_pool_check_memory_provider()).
> 
> [...]

Here is the summary with links:
  - [net-next,v2,1/4] net: refactor netdev_rx_queue_restart() to use local qops
    https://git.kernel.org/netdev/net-next/c/1eb824d69f8d
  - [net-next,v2,2/4] net: devmem: don't call queue stop / start when the interface is down
    https://git.kernel.org/netdev/net-next/c/3e7efc3f4f03
  - [net-next,v2,3/4] net: page_pool: avoid false positive warning if NAPI was never added
    https://git.kernel.org/netdev/net-next/c/c1e00bc4be06
  - [net-next,v2,4/4] netdevsim: allow normal queue reset while down
    https://git.kernel.org/netdev/net-next/c/285b3f78eabd

You are awesome, thank you!