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: <a45b348e-3d98-2343-6648-b2ec817bb236@omp.ru>
Date:   Sun, 9 Jan 2022 14:31:39 +0300
From:   Sergey Shtylyov <s.shtylyov@....ru>
To:     Andy Shevchenko <andy.shevchenko@...il.com>
CC:     Andy Shevchenko <andriy.shevchenko@...el.com>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        "Rafael J. Wysocki" <rafael@...nel.org>,
        "Linux Kernel Mailing List" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] platform: finally disallow IRQ0 in platform_get_irq() and
 its ilk

On 1/5/22 1:09 PM, Andy Shevchenko wrote:

>>>>> Wanna help?
>>>>
>>>>    No, I'm afraid you're on your own here...
>>
>>    Tell me please, how far you've got with this by now?
>>    (I've already started to add the fixups to your patch -- unfortunately, this change has to be
>> done atomically, not piecemeal.)

   Now I'm doing "the 2nd pass" over the platform_get_irq_optional() patch, and I think I'm done
with the 2nd, platform_get_irq_byname_optional() patch... Expecting to post those 2 early next week.

> I just returned from vacation and will have another one soon, I don't
> think I will be doing much for the next couple of weeks.

   You've settled in well -- having (several!) vacations while I've been doing your homework for you
during my holidays. B-)

>>>>>> Fixes: a85a6c86c25b ("driver core: platform: Clarify that IRQ 0 is invalid")
>>>>>
>>>>> Not sure.
>>>>
>>>>    Why? It fixes gthe IRQ0 problem, so that you don't have to check for IRQ0 in many callers
>>>> (for the subsytems that treat 0 as s/th special, like polling mode)... If you have something
>>>> to improve, you can do that atop of this patch...
>>>
>>> Because first we need to fix all users of platform_get_irq_optional().
>>
>>    I still don't understand why your issue should be fixed 1st -- but I don't really care about
>> the order...
> 
> See my other comments on the discussion.
> The rough roadmap is:
> 1) check which drivers are still subjects of vIRQ0 which is retrieved
> via IRQ resource

   We have WARN() added for that -- which isn't even limited to the static IRQ resources...

> 2) fix them accordingly (for example, by transforming to IRQ domains)

   I think we may choose to do a quick workaround, with the IRQ domain transformation somewhat
deferred...

> 3) convert platform_get_irq() and Co (including optional variants) to
> follow the pattern
>  a) non-optional APIs never return 0
>  b) optional APIs return negative error, or positive vIRQ or 0 when
> IRQ not found

   Yeah, and that means that we should 1st convert your platform_do_get_irq() (I'm renaming it)to not return 0 on IRQ0 -- in order to avoid the ambiguity with the "IRQ not found" outcome.

> Alternatively you may put a big comment in the drivers first and use
> platform_get_resource() for retrieving IRQ0 without WARN(). Then they
> will be subject to fix later on.

  No -- that would be a step backward, I think...

MBR, Sergey