Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Sender: Guenter Roeck <groeck7@gmail.com>
Date:   Wed, 24 Feb 2021 07:25:29 -0800
From:   Guenter Roeck <linux@roeck-us.net>
To:     Alexandre Belloni <alexandre.belloni@bootlin.com>
Cc:     Bruno Thomsen <bruno.thomsen@gmail.com>,
        linux-watchdog@vger.kernel.org, linux-rtc@vger.kernel.org,
        open list <linux-kernel@vger.kernel.org>,
        Lars Alex Pedersen <laa@kamstrup.com>,
        Bruno Thomsen <bth@kamstrup.com>,
        NXP Linux Team <linux-imx@nxp.com>
Subject: Re: watchdog: pcf2127: systemd fails on 5.11
Message-ID: <20210224152529.GA242356@roeck-us.net>
References: <CAH+2xPDs8f=bR7y5QYCpYpJTE1KJPfuiML1og3S9TfSFtOFBHw@mail.gmail.com>
 <20210222224325.GB177866@roeck-us.net>
 <CAH+2xPDDiUxKk3Y3R=fj0cOU+7vJRSC5yUb_XmfOUXnqoe+2Zg@mail.gmail.com>
 <YDZp/u+fO/8HX8qo@piout.net>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <YDZp/u+fO/8HX8qo@piout.net>
User-Agent: Mutt/1.9.4 (2018-02-28)
Precedence: bulk

On Wed, Feb 24, 2021 at 04:00:14PM +0100, Alexandre Belloni wrote:
> Hi,
> 
> On 24/02/2021 15:55:00+0100, Bruno Thomsen wrote:
> > You could be right about that, I don't think the watchdog feature should
> > be available for use if the alarm feature is enabled due to how CTRL2
> > register behaves.
> > 
> > The hardware I am testing on is a custom board, but it's actually
> > possible to get a Raspberry Pi module called RasClock that has
> > the chip.
> > 
> 
> I have an eval board for the PCF2127 (and PCF2129), the OM13513.
> 
> > I will test some locking around WD_VAL register access as that is used
> > in pcf2127_wdt_ping function.
> > 
> > My initial test shows that spin_lock_irqsave around regmap calls are not
> > a good idea as it result in:
> > BUG: scheduling while atomic: watchdog/70/0x00000002
> > BUG: scheduling while atomic: systemd/1/0x00000002
> > 
> 
> The issue is not only regmap but the fact that i2C and spi accesses are
> allowed to sleep.
> 
Correct, those would have to be mutexes.

Thanks,
Guenter