Hey Carlos,

On Wed Oct 14 2026, Carlos Fangmeier wrote:
> On Tue, Jun 10 2026, Jakub Kicinski wrote:
>> This still looks racy. Please look at other drivers
>
> You were right. The intel drivers (ixgbe/i40e) disable the queue pair
> as a unit around the pool swap rather than gating TX piecemeal, which
> is the pattern my patch was missing.
>
> Kurt Kanzenbach's series "net: stmmac: Fix XSK crashes on stm32mp2"
> [1] takes that approach for the XSK pool enable/disable path and is
> further along in review, so I'll drop this patch in favour of it.
>
> [1] https://lore.kernel.org/netdev/20261013-stmmac_xsk_crashes-v3-0-50fca7790af2@linutronix.de/
>

Well, looking at the Sashiko reports, it needs some work though. I'll
put you on Cc for the next iteration.

Thanks,
Kurt