linux-kernel - Re: [PATCH] gpio: Request interrupts after IRQ is initialized

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]

Message-ID: <67df4178-5943-69d8-0d61-f533671a1248@amd.com>
Date:   Sun, 17 Apr 2022 23:34:43 -0500
From:   Mario Limonciello <mario.limonciello@....com>
To:     firew4lker <firew4lker@...il.com>,
        Linus Walleij <linus.walleij@...aro.org>,
        Bartosz Golaszewski <brgl@...ev.pl>,
        Andy Shevchenko <andy.shevchenko@...il.com>,
        Shreeya Patel <shreeya.patel@...labora.com>,
        "open list:GPIO SUBSYSTEM" <linux-gpio@...r.kernel.org>,
        open list <linux-kernel@...r.kernel.org>
Cc:     Basavaraj.Natikar@....com, Richard.Gong@....com,
        stable@...r.kernel.org
Subject: Re: [PATCH] gpio: Request interrupts after IRQ is initialized

On 4/17/22 07:24, firew4lker wrote:
> On 4/14/22 05:57, Mario Limonciello wrote:
>> commit 5467801f1fcb ("gpio: Restrict usage of GPIO chip irq members 
>> before
>> initialization") attempted to fix a race condition that lead to a NULL
>> pointer, but in the process caused a regression for _AEI/_EVT declared
>> GPIOs. This manifests in messages showing deferred probing while trying
>> to allocate IRQs like so:
>>
>> [    0.688318] amd_gpio AMDI0030:00: Failed to translate GPIO pin 
>> 0x0000 to IRQ, err -517
>> [    0.688337] amd_gpio AMDI0030:00: Failed to translate GPIO pin 
>> 0x002C to IRQ, err -517
>> [    0.688348] amd_gpio AMDI0030:00: Failed to translate GPIO pin 
>> 0x003D to IRQ, err -517
>> [    0.688359] amd_gpio AMDI0030:00: Failed to translate GPIO pin 
>> 0x003E to IRQ, err -517
>> [    0.688369] amd_gpio AMDI0030:00: Failed to translate GPIO pin 
>> 0x003A to IRQ, err -517
>> [    0.688379] amd_gpio AMDI0030:00: Failed to translate GPIO pin 
>> 0x003B to IRQ, err -517
>> [    0.688389] amd_gpio AMDI0030:00: Failed to translate GPIO pin 
>> 0x0002 to IRQ, err -517
>> [    0.688399] amd_gpio AMDI0030:00: Failed to translate GPIO pin 
>> 0x0011 to IRQ, err -517
>> [    0.688410] amd_gpio AMDI0030:00: Failed to translate GPIO pin 
>> 0x0012 to IRQ, err -517
>> [    0.688420] amd_gpio AMDI0030:00: Failed to translate GPIO pin 
>> 0x0007 to IRQ, err -517
>>
>> The code for walking _AEI doesn't handle deferred probing and so this 
>> leads
>> to non-functional GPIO interrupts.
>>
>> Fix this issue by moving the call to 
>> `acpi_gpiochip_request_interrupts` to
>> occur after gc->irc.initialized is set.
>>
>> Cc: Shreeya Patel <shreeya.patel@...labora.com>
>> Cc: stable@...r.kernel.org
>> Fixes: 5467801f1fcb ("gpio: Restrict usage of GPIO chip irq members 
>> before initialization")
>> Reported-by: Mario Limonciello <mario.limonciello@....com>
>> Link: 
>> https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Flore.kernel.org%2Flinux-gpio%2FBL1PR12MB51577A77F000A008AA694675E2EF9%40BL1PR12MB5157.namprd12.prod.outlook.com%2FT%2F%23u&amp;data=04%7C01%7Cmario.limonciello%40amd.com%7C96ec39c78488493fd5ca08da206d3c7b%7C3dd8961fe4884e608e11a82d994e183d%7C0%7C0%7C637857951204650754%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&amp;sdata=xZbNC%2F50JqlNwcTYAtGLn6z0%2FEPbfCKKOc%2BlZlMh0EQ%3D&amp;reserved=0 
>>
>> Signed-off-by: Mario Limonciello <mario.limonciello@....com>
>> ---
>>   drivers/gpio/gpiolib.c | 4 ++--
>>   1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c
>> index 085348e08986..b7694171655c 100644
>> --- a/drivers/gpio/gpiolib.c
>> +++ b/drivers/gpio/gpiolib.c
>> @@ -1601,8 +1601,6 @@ static int gpiochip_add_irqchip(struct gpio_chip 
>> *gc,
>>       gpiochip_set_irq_hooks(gc);
>> -    acpi_gpiochip_request_interrupts(gc);
>> -
>>       /*
>>        * Using barrier() here to prevent compiler from reordering
>>        * gc->irq.initialized before initialization of above
>> @@ -1612,6 +1610,8 @@ static int gpiochip_add_irqchip(struct gpio_chip 
>> *gc,
>>       gc->irq.initialized = true;
>> +    acpi_gpiochip_request_interrupts(gc);
>> +
>>       return 0;
>>   }
> 
> Tested-By:firew4lker@...il.com
> 
> This patch addresses the issue. Tested on a Lenovo T14 with AMD Ryzen 5 
> PRO 4650U with Radeon Graphics Graphics.
> 
> Without this patch the laptop is impossible to wake from S3 and S0ix.
> 

Thanks for testing it!

Linus Walleij,

As this is backported to 5.15.y, 5.16.y, 5.17.y and those all had point 
releases a bunch of people are hitting it now.  If you choose to adopt 
this patch instead of revert the broken one, you can add to the commit 
message too:

Link: https://gitlab.freedesktop.org/drm/amd/-/issues/1976