Re: [PATCH] Introduce the pkill_on_warn boot parameter

On 27.07.2022 19:30, Jann Horn wrote:
> On Wed, Jul 27, 2022 at 6:17 PM Alexey Khoroshilov
> <[email protected]> wrote:
>> On 01.10.2021 22:59, Linus Torvalds wrote:
>> Coming back to the discussion of WARN_ON()/pr_warn("WARNING:") semantics.
>>
>> We see a number of cases where WARNING is used to inform userspace that
>> it is doing something wrong, e.g.
>> https://elixir.bootlin.com/linux/v5.19-rc8/source/net/can/j1939/socket.c#L181
>> https://elixir.bootlin.com/linux/v5.19-rc8/source/drivers/video/fbdev/core/fbmem.c#L1023
>>
>> It is definitely useful, but it does not make sense in case of fuzzing
>> when the userspace should do wrong things and check if kernel behaves
>> correctly.
>>
>> As a result we have warnings with two different intentions:
>> - warn that something wrong happens in kernel, but we are able to continue;
>> - warn userspace that it is doing something wrong.
>>
>> During fuzzing we would like to report the former and to ignore the
>> latter. Are any ideas how these intentions can be recognized automatically?
> 
> https://elixir.bootlin.com/linux/v5.19-rc8/source/include/asm-generic/bug.h#L74
> says:
> 
>  * WARN(), WARN_ON(), WARN_ON_ONCE, and so on can be used to report
>  * significant kernel issues that need prompt attention if they should ever
>  * appear at runtime.
>  *
>  * Do not use these macros when checking for invalid external inputs
>  * (e.g. invalid system call arguments, or invalid data coming from
>  * network/devices), and on transient conditions like ENOMEM or EAGAIN.
>  * These macros should be used for recoverable kernel issues only.
>  * For invalid external inputs, transient conditions, etc use
>  * pr_err[_once/_ratelimited]() followed by dump_stack(), if necessary.
>  * Do not include "BUG"/"WARNING" in format strings manually to make these
>  * conditions distinguishable from kernel issues.
> 
> So if you see drivers intentionally using WARN() or printing
> "WARNING:" on codepaths that are reachable with bogus inputs from
> userspace, those codepaths should be fixed to log warnings in a
> different format.

Thank you, Jann!

I have missed that.

--
Alexey