On Thu, Jul 17, 2025 at 05:21:44PM -0700, Joshua Hay wrote:
> This series fixes a stability issue in the flow scheduling Tx send/clean
> path that results in a Tx timeout.
> 
> The existing guardrails in the Tx path were not sufficient to prevent
> the driver from reusing completion tags that were still in flight (held
> by the HW).  This collision would cause the driver to erroneously clean
> the wrong packet thus leaving the descriptor ring in a bad state.
> 
> The main point of this refactor is to replace the flow scheduling buffer
> ring with a large pool/array of buffers.  The completion tag then simply
> is the index into this array.  The driver tracks the free tags and pulls
> the next free one from a refillq.  The cleaning routines simply use the
> completion tag from the completion descriptor to index into the array to
> quickly find the buffers to clean.
> 
> All of the code to support the refactor is added first to ensure traffic
> still passes with each patch.  The final patch then removes all of the
> obsolete stashing code.
> 
> ---
> v2:
> - Add a new patch "idpf: simplify and fix splitq Tx packet rollback
>   error path" that fixes a bug in the error path. It also sets up
>   changes in patch 4 that are necessary to prevent a crash when a packet
>   rollback occurs using the buffer pool.
> 
> v1:
> https://lore.kernel.org/intel-wired-lan/[email protected]/T/#maf9f464c598951ee860e5dd24ef8a451a488c5a0
> 
> Joshua Hay (6):
>   idpf: add support for Tx refillqs in flow scheduling mode
>   idpf: improve when to set RE bit logic
>   idpf: simplify and fix splitq Tx packet rollback error path
>   idpf: replace flow scheduling buffer ring with buffer pool
>   idpf: stop Tx if there are insufficient buffer resources
>   idpf: remove obsolete stashing code
> 
>  .../ethernet/intel/idpf/idpf_singleq_txrx.c   |  61 +-
>  drivers/net/ethernet/intel/idpf/idpf_txrx.c   | 723 +++++++-----------
>  drivers/net/ethernet/intel/idpf/idpf_txrx.h   |  87 +--
>  3 files changed, 356 insertions(+), 515 deletions(-)

Hi Joshua, all,

Perhaps it is not followed much anymore, but at least according to [1]
patches for stable should not be more than 100 lines, with context.

This patch-set is an order of magnitude larger.
Can something be done to create a more minimal fix?

[1] https://docs.kernel.org/process/stable-kernel-rules.html#stable-kernel-rules

Reply via email to