Re: Linux 6.3.1

[Chris Clayton]

On 06/05/2023 01:52, Greg Kroah-Hartman wrote:
> On Tue, May 02, 2023 at 02:12:58PM +0100, Chris Clayton wrote:
>> Hi Greg
>>
>> On 02/05/2023 08:31, Greg Kroah-Hartman wrote:
>>> On Tue, May 02, 2023 at 08:43:47AM +0900, Greg Kroah-Hartman wrote:
>>>> On Tue, May 02, 2023 at 06:40:03AM +0900, Greg Kroah-Hartman wrote:
>>>>> On Mon, May 01, 2023 at 07:22:56AM +0100, Chris Clayton wrote:
>>>>>> 6.3.1 FTBFS thusly:
>>>>>
>>>>> What is "FTBFS"?
>>>>>
>>>>>> drivers/net/wireguard/timers.c: In function 'wg_expired_retransmit_handshake':
>>>>>> <command-line>: error: format '%d' expects argument of type 'int', but argument 6 has type 'long unsigned int'
>>>>>> [-Werror=format=]
>>>>>> <command-line>: note: in expansion of macro 'KBUILD_MODNAME'
>>>>>> ./include/linux/dynamic_debug.h:223:29: note: in expansion of macro 'pr_fmt'
>>>>>>   223 |                 func(&id, ##__VA_ARGS__);                       \
>>>>>>       |                             ^~~~~~~~~~~
>>>>>
>>>>> <snip>
>>>>>
>>>>>> There's a patch to drivers/net/wireguard/timers.c that fixes these errors and you can find it at
>>>>>> 2d4ee16d969c97996e80e4c9cb6de0acaff22c9f in Linus' tree.
>>>>>
>>>>> Thanks for this report, we'll queue it up soon.
>>>>
>>>> Odd, that commit is in 6.2 already, so how are you applying this to
>>>> 6.3.y?
>>>
>>> And this is only a gcc13 issue, right?  So it's not a regression, it's
>>> always been there, nothing new caused it in this release from what I can
>>> tell.
>>>
>>> Actually I don't see how this is an issue in 6.3.1 anyway, as again, the
>>> proposed fix you gave is in 6.2, so what really is happening here?
>>
>> You seem not to have seen the email I sent shortly after my original report. In that email I apologised for the noise
>> because, due to a caffeine deficiency, I had reported the problem against the wrong kernel version and that it actually
>> occurs 6.1.27. gcc13 spits it out as a warning, but kernel build system is setbup to treat warnings as errors, so the
>> build fails.
> 
> Ok, but this is not a new issue, right?  6.1.26 should have had the same
> problem from what I can tell.
> 

I don't know. I haven't tried to build 6.1.26 with gcc13. The commit message for the patch in Linus' tree says that this
is a new warning in gcc13.  The kernel build system configures gcc to treat warnings as errors so, with gcc13, the build
failed and I reported the failure in the initial email of this thread. Looking at that patch submission
(https://lore.kernel.org/all/20221213225208.3343692-2-Jason@xxxxxxxxx/), it appears not to have been tagged for or
copied to stable.

Once I'd realised that I had reported the problem against the wrong kernel version, I sent my apologetic email and sent
a new email reporting the problem against 6.1.27. In response to that (I assume), you've queued the patch for 6.1 stable
and for 5.15 and 5.10, so everything seems to be in order.

Chris


> greg k-h