Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: Re: [PATCH v10 31/55] Input: atmel_mxt_ts - delay enabling IRQ when
 not using regulators
To:     Jiada Wang <jiada_wang@mentor.com>, nick@shmanahar.org,
        dmitry.torokhov@gmail.com, jikos@kernel.org,
        benjamin.tissoires@redhat.com, bsz@semihalf.com
Cc:     linux-input@vger.kernel.org, linux-kernel@vger.kernel.org,
        erosca@de.adit-jv.com, Andrew_Gabbasov@mentor.com
References: <20200331105051.58896-1-jiada_wang@mentor.com>
 <20200331105051.58896-32-jiada_wang@mentor.com>
From:   Dmitry Osipenko <digetx@gmail.com>
Message-ID: <46e0b0cf-63f1-4b46-dc3e-0e6610e72d75@gmail.com>
Date:   Sat, 11 Apr 2020 01:38:07 +0300
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101
 Thunderbird/68.5.0
MIME-Version: 1.0
In-Reply-To: <20200331105051.58896-32-jiada_wang@mentor.com>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

31.03.2020 13:50, Jiada Wang пишет:
> From: Nick Dyer <nick.dyer@itdev.co.uk>
> 
> The path of enabling the IRQ in the probe function is not safe in level
> triggered operation, if it was already powered up and there is a message
> waiting on the device (eg finger down) because the object table has not yet
> been read. This forces the ISR into a hard loop.
> 
> Delay enabling the interrupt until it is first needed.


> 
> -	if (data->use_retrigen_workaround) {
> +		/* Presence of data->irq means IRQ initialised */
> +		data->irq = data->client->irq;

IIRC, IRQ=0 could be a valid interrupt since this is a "virtual"
interrupt number.

...
>  	init_completion(&data->chg_completion);
> @@ -3826,26 +3842,22 @@ static int mxt_probe(struct i2c_client *client, const struct i2c_device_id *id)
>  		return error;
>  	}

What about just to tell the IRQ core not to enable the interrupt
handling during of the devm_request_threaded_irq()?

To achieve that, add this line here:

	irq_set_status_flags(client->irq, IRQ_NOAUTOEN);

> -	error = devm_request_threaded_irq(&client->dev, client->irq,
> -					  NULL, mxt_interrupt, IRQF_ONESHOT,
> -					  client->name, data);
...

Then the interrupt will be requested in the disabled state and it will
be enabled only after the first enable_irq() invocation.