From: "Linus Lüssing" <linus.luessing@c0d3.blue>
To: Felix Fietkau <nbd@nbd.name>
Cc: "Kalle Valo" <kvalo@codeaurora.org>,
"Sujith Manoharan" <c_manoha@qca.qualcomm.com>,
ath9k-devel@qca.qualcomm.com, linux-wireless@vger.kernel.org,
"David S . Miller" <davem@davemloft.net>,
"Jakub Kicinski" <kuba@kernel.org>,
"John W . Linville" <linville@tuxdriver.com>,
"Felix Fietkau" <nbd@openwrt.org>,
"Simon Wunderlich" <sw@simonwunderlich.de>,
"Sven Eckelmann" <sven@narfation.org>,
netdev@vger.kernel.org, linux-kernel@vger.kernel.org,
"Linus Lüssing" <ll@simonwunderlich.de>
Subject: Re: [PATCH 3/3] ath9k: Fix potential hw interrupt resume during reset
Date: Wed, 15 Sep 2021 21:18:03 +0200 [thread overview]
Message-ID: <YUJGxZW1a+vlG335@sellars> (raw)
In-Reply-To: <255a49c7-d763-50d9-87e0-da22f4a9b053@nbd.name>
On Wed, Sep 15, 2021 at 11:48:55AM +0200, Felix Fietkau wrote:
>
> On 2021-09-14 21:25, Linus Lüssing wrote:
> > From: Linus Lüssing <ll@simonwunderlich.de>
> >
> > There is a small risk of the ath9k hw interrupts being reenabled in the
> > following way:
> >
> > 1) ath_reset_internal()
> > ...
> > -> disable_irq()
> > ...
> > <- returns
> >
> > 2) ath9k_tasklet()
> > ...
> > -> ath9k_hw_resume_interrupts()
> > ...
> >
> > 1) ath_reset_internal() continued:
> > -> tasklet_disable(&sc->intr_tq); (= ath9k_tasklet() off)
> >
> > By first disabling the ath9k interrupt there is a small window
> > afterwards which allows ath9k hw interrupts being reenabled through
> > the ath9k_tasklet() before we disable this tasklet in
> > ath_reset_internal(). Leading to having the ath9k hw interrupts enabled
> > during the reset, which we should avoid.
> I don't see a way in which interrupts can be re-enabled through the
> tasklet. disable_irq disables the entire PCI IRQ (not through ath9k hw
> registers), and they will only be re-enabled by the corresponding
> enable_irq call.
Ah, okay, then I think I misunderstood the previous fixes to the
ath9k interrupt shutdown during reset here. I had only tested the
following diff and assumed that it were not okay to have the ath9k
hw interrupt registers enabled within the spinlock'd section:
```
@@ -299,11 +299,23 @@ static int ath_reset_internal(struct ath_softc *sc, struct ath9k_channel *hchan)
__ath_cancel_work(sc);
disable_irq(sc->irq);
+ for (r = 0; r < 200; r++) {
+ msleep(5);
+
+ if (REG_READ(ah, AR_INTR_SYNC_CAUSE) ||
+ REG_READ(ah, AR_INTR_ASYNC_CAUSE)) {
+ break;
+ }
+ }
tasklet_disable(&sc->intr_tq);
tasklet_disable(&sc->bcon_tasklet);
spin_lock_bh(&sc->sc_pcu_lock);
+ if (REG_READ(ah, AR_INTR_SYNC_CAUSE) ||
+ REG_READ(ah, AR_INTR_ASYNC_CAUSE))
+ ATH_DBG_WARN(1, "%s: interrupts are enabled", __func__);
+
if (!sc->cur_chan->offchannel) {
fastcc = false;
caldata = &sc->cur_chan->caldata;
```
And yes, the general ath9k interrupt should still be disabled. Didn't
test for that but seems like it.
(And now I noticed that actually
ath_reset_internal()
-> ath_prepare_reset()
-> ath9k_hw_disable_interrupts()
-> ath9k_hw_kill_interrupts()
disables the ath9k hw interrupt registers before the
ath_reset_internal()->ath9k_hw_reset() call anyway.)
What would you prefer, should this patch just be dropped or should
I resend it without the "Fixes:" line as a cosmetic change? (e.g. to
have the order in reverse to reenablement at the end of
ath_reset_internal(), to avoid confusion in the future?)
Regards, Linus
next prev parent reply other threads:[~2021-09-15 19:18 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-09-14 19:25 [PATCH 0/3] ath9k: interrupt fixes on queue reset Linus Lüssing
2021-09-14 19:25 ` [PATCH 1/3] ath9k: add option to reset the wifi chip via debugfs Linus Lüssing
2021-10-05 14:27 ` Kalle Valo
2021-09-14 19:25 ` [PATCH 2/3] ath9k: Fix potential interrupt storm on queue reset Linus Lüssing
2021-09-14 19:25 ` [PATCH 3/3] ath9k: Fix potential hw interrupt resume during reset Linus Lüssing
2021-09-15 9:48 ` Felix Fietkau
2021-09-15 19:18 ` Linus Lüssing [this message]
2021-09-14 19:53 ` [PATCH 0/3] ath9k: interrupt fixes on queue reset Toke Høiland-Jørgensen
2021-09-15 9:23 ` Linus Lüssing
2021-10-05 14:12 ` Linus Lüssing
2021-10-05 14:24 ` Kalle Valo
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=YUJGxZW1a+vlG335@sellars \
--to=linus.luessing@c0d3.blue \
--cc=ath9k-devel@qca.qualcomm.com \
--cc=c_manoha@qca.qualcomm.com \
--cc=davem@davemloft.net \
--cc=kuba@kernel.org \
--cc=kvalo@codeaurora.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-wireless@vger.kernel.org \
--cc=linville@tuxdriver.com \
--cc=ll@simonwunderlich.de \
--cc=nbd@nbd.name \
--cc=nbd@openwrt.org \
--cc=netdev@vger.kernel.org \
--cc=sven@narfation.org \
--cc=sw@simonwunderlich.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).