On Fri, Mar 06, 2026 at 07:22:28PM -0800, Jakub Kicinski wrote:
> This is an AI-generated review of your patch. The human sending this
> email has considered the AI review valid, or at least plausible.
>
> For local reproduction steps see
> https://netdev-ai.bots.linux.dev/ai-local.html
> ---
> net: mana: Expose hardware diagnostic info via debugfs
>
> This commit adds debugfs entries to expose hardware configuration and
> diagnostic information for the MANA driver. It creates device-level and
> port-level debugfs entries to aid in debugging initialization and runtime
> operations.
> Do these debugfs entries get recreated on every suspend/resume cycle?
No. Thankyou for pointing out Jakub.
I'll make this change in next version.
>
> Both mana_gd_query_max_resources() and mana_gd_verify_vf_version() are
> called from mana_gd_setup() during device initialization. On suspend,
> mana_gd_suspend() calls mana_gd_cleanup() which doesn't remove these
> debugfs entries. When mana_gd_resume() calls mana_gd_setup() again, it
> attempts to create the same debugfs entries a second time.
>
> This means debugfs_start_creating() in fs/debugfs/inode.c will print an
> error for each duplicate file:
>
> "'num_msix_usable' already exists in 'mana/<slot>/'"
> "'max_num_queues' already exists in 'mana/<slot>/'"
> "'gdma_protocol_ver' already exists in 'mana/<slot>/'"
> "'pf_cap_flags1' already exists in 'mana/<slot>/'"
>
> This produces 4 error messages in dmesg on every suspend/resume cycle,
> contradicting the commit message's goal of operating "without adding
> noise to dmesg."
>
> Should these device-level entries be removed during suspend (in
> mana_gd_cleanup or mana_gd_suspend), or should their creation be moved
> to mana_gd_probe where they're only created once?
Right. I will remove the entries during suspend and re-create them in
resume.
>
> [ ... ]
> --
> pw-bot: cr
