linux-kernel - Re: [PATCH 2/2] mfd: max77620: Add pre/post irq handler before/after servicing interrupt

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Wed, 8 Jun 2016 21:12:24 +0530
From:	Laxman Dewangan <ldewangan@...dia.com>
To:	Lee Jones <lee.jones@...aro.org>
CC:	<broonie@...nel.org>, <gregkh@...uxfoundation.org>,
	<linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 2/2] mfd: max77620: Add pre/post irq handler before/after
 servicing interrupt

Hi Lee,

On Wednesday 08 June 2016 08:11 PM, Lee Jones wrote:
> On Fri, 20 May 2016, Laxman Dewangan wrote:
>
>> + * MAX77620 and MAX20024 has the following steps of the interrupt handling
>> + * for TOP interrupts:
>> + * 1. When interrupt occurs from PMIC, mask the PMIC interrupt by setting GLBLM.
>> + * 2. Read IRQTOP and service the interrupt.
>> + * 3. Once all interrupts has been checked and serviced, the interrupt service
>> + *    routine un-masks the hardware interrupt line by clearing GLBLM.
>> + */
>> +static int max77620_top_irq_chip_pre_irq_handler(void *irq_drv_data)
>> +{
>> +	struct max77620_chip *chip = irq_drv_data;
>> +	int ret;
>> +
>> +	ret = regmap_update_bits(chip->rmap, MAX77620_REG_INTENLBT,
>> +				 MAX77620_GLBLM_MASK, MAX77620_GLBLM_MASK);
>> +	if (ret < 0)
>> +		dev_err(chip->dev, "Failed to set GLBLM: %d\n", ret);
>> +
>> +	return ret;
>> +}
>> +
>> +static int max77620_top_irq_chip_post_irq_handler(void *irq_drv_data)
>> +{
>> +	struct max77620_chip *chip = irq_drv_data;
>> +	int ret;
>> +
>> +	ret = regmap_update_bits(chip->rmap, MAX77620_REG_INTENLBT,
>> +				 MAX77620_GLBLM_MASK, 0);
>> +	if (ret < 0)
>> +		dev_err(chip->dev, "Failed to reset GLBLM: %d\n", ret);
>> +
>> +	return ret;
>> +}
> This seems massively over compacted.  All you're effectively doing
> here is masking and unmasking the IRQs, which we do almost
> ubiquitously with interrupt controllers.  Can't you just call the
> functions "max77629_{un}mask_irqs()"?
>
>

Actually, per PMIC HW design, we need to toggle this bit on ISRs. Before 
reading the status, need to set 1 and then after handling it need to set 
0. This cannot be done by any other bit toggling or masking/unmasking 
interrupt controller interrupt.

This is hard requirement from the PMIC chip.