ath9k: fix race condition in enabling/disabling IRQs
authorFelix Fietkau <nbd@nbd.name>
Thu, 2 Feb 2017 09:14:52 +0000 (10:14 +0100)
committerKalle Valo <kvalo@qca.qualcomm.com>
Tue, 7 Feb 2017 09:00:25 +0000 (11:00 +0200)
commit3a5e969bb2f6692a256352649355d56d018d6b88
tree51dc21b1b0ddc1ee2236c819fcd7194dbf70a94f
parenta34d0a0da1abae46a5f6ebd06fb0ec484ca099d9
ath9k: fix race condition in enabling/disabling IRQs

The code currently relies on refcounting to disable IRQs from within the
IRQ handler and re-enabling them again after the tasklet has run.

However, due to race conditions sometimes the IRQ handler might be
called twice, or the tasklet may not run at all (if interrupted in the
middle of a reset).

This can cause nasty imbalances in the irq-disable refcount which will
get the driver permanently stuck until the entire radio has been stopped
and started again (ath_reset will not recover from this).

Instead of using this fragile logic, change the code to ensure that
running the irq handler during tasklet processing is safe, and leave the
refcount untouched.

Cc: stable@vger.kernel.org
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
drivers/net/wireless/ath/ath9k/ath9k.h
drivers/net/wireless/ath/ath9k/init.c
drivers/net/wireless/ath/ath9k/mac.c
drivers/net/wireless/ath/ath9k/mac.h
drivers/net/wireless/ath/ath9k/main.c