linux-kernel - Re: [PATCH] platform: finally disallow IRQ0 in platform_get

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: <f176d16c-000e-f886-754e-1fc222f80230@omp.ru>
Date:   Tue, 4 Jan 2022 15:23:50 +0300
From:   Sergey Shtylyov <s.shtylyov@....ru>
To:     Geert Uytterhoeven <geert@...ux-m68k.org>,
        Marc Zyngier <maz@...nel.org>
CC:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        "Rafael J. Wysocki" <rafael@...nel.org>,
        <linux-kernel@...r.kernel.org>,
        Thomas Gleixner <tglx@...utronix.de>
Subject: Re: [PATCH] platform: finally disallow IRQ0 in platform_get_irq() and
 its ilk

On 1/4/22 12:47 PM, Geert Uytterhoeven wrote:

[...]
>>> The commit a85a6c86c25b ("driver core: platform: Clarify that IRQ 0 is
>>> invalid") only calls WARN() when IRQ0 is about to be returned, however
>>> using IRQ0 is considered invalid (according to Linus) outside the arch/
>>> code where it's used by the i8253 drivers. Many driver subsystems treat
>>> 0 specially (e.g. as an indication of the polling mode by libata), so
>>> the users of platform_get_irq[_byname]() in them would have to filter
>>> out IRQ0 explicitly and this (quite obviously) doesn't scale...
>>> Let's finally get this straight and return -EINVAL instead of IRQ0!
>>>
>>> Fixes: a85a6c86c25b ("driver core: platform: Clarify that IRQ 0 is invalid")
>>> Signed-off-by: Sergey Shtylyov <s.shtylyov@....ru>
> 
>>> --- driver-core.orig/drivers/base/platform.c
>>> +++ driver-core/drivers/base/platform.c
>>> @@ -231,7 +231,8 @@ int platform_get_irq_optional(struct pla
>>>  out_not_found:
>>>       ret = -ENXIO;
>>>  out:
>>> -     WARN(ret == 0, "0 is an invalid IRQ number\n");
>>> +     if (WARN(!ret, "0 is an invalid IRQ number\n"))
>>> +             return -EINVAL;
>>>       return ret;
>>>  }
>>>  EXPORT_SYMBOL_GPL(platform_get_irq_optional);
>>> @@ -445,7 +446,8 @@ static int __platform_get_irq_byname(str
>>>
>>>       r = platform_get_resource_byname(dev, IORESOURCE_IRQ, name);
>>>       if (r) {
>>> -             WARN(r->start == 0, "0 is an invalid IRQ number\n");
>>> +             if (WARN(!r->start, "0 is an invalid IRQ number\n"))
>>> +                     return -EINVAL;
>>>               return r->start;
>>>       }
>>
>> Geert recently mentioned that a few architectures (such as sh?) still
>> use IRQ0 as something valid in limited cases.
> 
> https://lore.kernel.org/all/CAMuHMdUg3=q7gyaVHP0XcYUOo3PQUUv8Hc8wp5faVQ+bTBpg4A@mail.gmail.com
> 
> TL;DR: Probably only smsc911x Ethernet on the AP-SH4A-3A and
> AP-SH4AD-0A boards, which should trigger the warning since v5.8.

   Gr... indeed these use IRQ0 and should cause WARN. Do you have any idea how to avoid this?

>> From my PoV, this patch is fine, but please be prepared to fix things
>> in a couple of years when someone decides to boot a recent kernel on
>> their pet dinosaur. With that in mind:
>>
>> Acked-by: Marc Zyngier <maz@...nel.org>
> 
> TBH, I don't see much point in this patch, as the WARN() has been
> there since a while,

   Yet there's WARN() -- which (at the end of day) should be avoided.

> and the end goal is to return zero instead of
> -ENXIO for no interrupt, right?

   I don't care that much about platform_get_irq_optional() (Andy does),
I do care about its caller, platform_get_irq(). The end goal here is to
avoid WARN() and avoid having to handle IRQ0 in this function's callers.

> Gr{oetje,eeting}s,
> 
>                         Geert
[...]

MBR, Sergey