Re: [PATCH net v2] net: Drop the lock in skb_may_tx_timestamp()

[Sebastian Andrzej Siewior <bigeasy@linutronix.de>]

On 2026-02-26 09:11:16 [+0100], Kurt Kanzenbach wrote:
> Hi maintainers,
> 
> thanks for applying the patch.
> 
> What's your preferred solution for igb? Can we do the time stamping from
> IRQ context now [1] or should we switch to BH workqueue [2]?

We have this quote from Jakub [a]:

| But FWIW I can't stop myself from restating that networking core is not
| supposed to be called by drivers in hard IRQ context in general.
| Especially when it comes to anything that may involve user sockets.

which I kind of agree in terms of "easy" maintenance. The kfree_skb()
in the error case does not hurt as of today so does the socket
interaction.

Based on Paolo's [b], we can have this fix and insist on !hardirq
processing?

[a] https://lore.kernel.org/all/20260220132932.1ed45656@kernel.org/
[b] https://lore.kernel.org/all/55b24261-45cb-41f1-857f-f29527a1ee8d@redhat.com

> Thanks,
> Kurt
> 
> [1] - https://lore.kernel.org/intel-wired-lan/20260205-igb_irq_ts-v3-1-2efc7bc4b885@linutronix.de/
> [2] - https://lore.kernel.org/intel-wired-lan/20260211134436.1e623034@kernel.org/

Sebastian