Jakub Kicinski [off-list ref] writes:
On Thu, 20 Aug 2026 15:01:59 +0200 Andrew Gaylard wrote:
quoted
spl2sw_tx_poll() runs in NAPI (softirq) context and acquires
comm->tx_lock with plain spin_lock(). The hardirq handler
spl2sw_ethernet_interrupt() acquires the same lock, so a hardirq on
the same CPU while the NAPI poll holds the lock would deadlock with
IRQs disabled. Use spin_lock_irqsave/spin_unlock_irqrestore.
Please explain how the issue was found and how the fix was tested.
I mean - this is an ancient driver, if you don't have a way to test the
fix you should not be sending the patch.
I'm testing on a Tibbo LTTP3G2 board:
https://tibbo.com/store/tps/ltpp3g2.html
This work is part of getting it to boot, getting it stable, adding a
couple of missing drivers, and sending the changes upstream.
I found the issue using lockdep. At least, I though I did, but I've not
been able to reproduce it, so I must have been mistaken.
Apologies. Please drop this patch.
--
Andrew