On 5/14/2021 7:00 AM, Arnd Bergmann wrote:
> From: Arnd Bergmann <[email protected]>
>
> clang points out that an initcall funciton should return an 'int':
>
> kernel/kcsan/debugfs.c:274:15: error: returning 'void' from a function with incompatible result type 'int'
> late_initcall(kcsan_debugfs_init);
> ~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~
> include/linux/init.h:292:46: note: expanded from macro 'late_initcall'
> #define late_initcall(fn) __define_initcall(fn, 7)
>
> Fixes: e36299efe7d7 ("kcsan, debugfs: Move debugfs file creation out of early init")
> Signed-off-by: Arnd Bergmann <[email protected]>
For the record, this requires CONFIG_LTO_CLANG to be visible.
Reviewed-by: Nathan Chancellor <[email protected]>
> ---
> kernel/kcsan/debugfs.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/kernel/kcsan/debugfs.c b/kernel/kcsan/debugfs.c
> index c1dd02f3be8b..e65de172ccf7 100644
> --- a/kernel/kcsan/debugfs.c
> +++ b/kernel/kcsan/debugfs.c
> @@ -266,9 +266,10 @@ static const struct file_operations debugfs_ops =
> .release = single_release
> };
>
> -static void __init kcsan_debugfs_init(void)
> +static int __init kcsan_debugfs_init(void)
> {
> debugfs_create_file("kcsan", 0644, NULL, NULL, &debugfs_ops);
> + return 0;
> }
>
> late_initcall(kcsan_debugfs_init);
>
On Fri, May 14, 2021 at 11:29:18AM -0700, Nathan Chancellor wrote:
> On 5/14/2021 7:00 AM, Arnd Bergmann wrote:
> > From: Arnd Bergmann <[email protected]>
> >
> > clang points out that an initcall funciton should return an 'int':
> >
> > kernel/kcsan/debugfs.c:274:15: error: returning 'void' from a function with incompatible result type 'int'
> > late_initcall(kcsan_debugfs_init);
> > ~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~
> > include/linux/init.h:292:46: note: expanded from macro 'late_initcall'
> > #define late_initcall(fn) __define_initcall(fn, 7)
> >
> > Fixes: e36299efe7d7 ("kcsan, debugfs: Move debugfs file creation out of early init")
> > Signed-off-by: Arnd Bergmann <[email protected]>
>
> For the record, this requires CONFIG_LTO_CLANG to be visible.
>
> Reviewed-by: Nathan Chancellor <[email protected]>
Queued with the three Reviewed-by tags, thank you all!
Nathan, I lost the thread on exactly what it is that requires that
CONFIG_LTO_CLANG be visible. A naive reader might conclude that the
compiler diagnostic does not appear unless CONFIG_LTO_CLANG=y, but
that would be surprising (and yes, I have been surprised many times).
If you are suggesting that the commit log be upgraded, could you please
supply suggested wording?
Once this is nailed down (or by Wednesday if I hear no more), I will
rebase it to the bottom of the current kcsan stack, let it soak in -next
for a couple of days, then send to Linus as a fix for a regression.
Hopefully some time next week.
Thanx, Paul
> > ---
> > kernel/kcsan/debugfs.c | 3 ++-
> > 1 file changed, 2 insertions(+), 1 deletion(-)
> >
> > diff --git a/kernel/kcsan/debugfs.c b/kernel/kcsan/debugfs.c
> > index c1dd02f3be8b..e65de172ccf7 100644
> > --- a/kernel/kcsan/debugfs.c
> > +++ b/kernel/kcsan/debugfs.c
> > @@ -266,9 +266,10 @@ static const struct file_operations debugfs_ops =
> > .release = single_release
> > };
> > -static void __init kcsan_debugfs_init(void)
> > +static int __init kcsan_debugfs_init(void)
> > {
> > debugfs_create_file("kcsan", 0644, NULL, NULL, &debugfs_ops);
> > + return 0;
> > }
> > late_initcall(kcsan_debugfs_init);
> >
>
Hi Paul,
On 5/14/2021 12:36 PM, Paul E. McKenney wrote:
> On Fri, May 14, 2021 at 11:29:18AM -0700, Nathan Chancellor wrote:
>> On 5/14/2021 7:00 AM, Arnd Bergmann wrote:
>>> From: Arnd Bergmann <[email protected]>
>>>
>>> clang points out that an initcall funciton should return an 'int':
>>>
>>> kernel/kcsan/debugfs.c:274:15: error: returning 'void' from a function with incompatible result type 'int'
>>> late_initcall(kcsan_debugfs_init);
>>> ~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~
>>> include/linux/init.h:292:46: note: expanded from macro 'late_initcall'
>>> #define late_initcall(fn) __define_initcall(fn, 7)
>>>
>>> Fixes: e36299efe7d7 ("kcsan, debugfs: Move debugfs file creation out of early init")
>>> Signed-off-by: Arnd Bergmann <[email protected]>
>>
>> For the record, this requires CONFIG_LTO_CLANG to be visible.
>>
>> Reviewed-by: Nathan Chancellor <[email protected]>
>
> Queued with the three Reviewed-by tags, thank you all!
>
> Nathan, I lost the thread on exactly what it is that requires that
> CONFIG_LTO_CLANG be visible. A naive reader might conclude that the
> compiler diagnostic does not appear unless CONFIG_LTO_CLANG=y, but
> that would be surprising (and yes, I have been surprised many times).
> If you are suggesting that the commit log be upgraded, could you please
> supply suggested wording?
You can see my response to Marco here:
https://lore.kernel.org/r/[email protected]/
Maybe some improved wording might look like
clang with CONFIG_LTO_CLANG points out that an initcall function should
return an 'int' due to the changes made to the initcall macros in commit
3578ad11f3fb ("init: lto: fix PREL32 relocations"):
...
Arnd, do you have any objections?
Cheers,
Nathan