Re: Linux 6.18-rc6

["David Hildenbrand (Red Hat)" <david@kernel.org>]

On 17.11.25 18:28, Linus Torvalds wrote:
> On Mon, 17 Nov 2025 at 06:08, David Hildenbrand (Red Hat)
> <david@kernel.org> wrote:
>>
>> To not lose too much time, I just pushed the following patch to
>>
>> https://github.com/davidhildenbrand/linux.git zerotags
> 
> Hmm. Why isn't the fix for this simply this (intentionally
> whitespace-damaged - don't apply mindlessly) one-liner:
> 
>    --- a/include/linux/highmem.h
>    +++ b/include/linux/highmem.h
>    @@ -253,5 +253,6 @@ static inline void
> clear_highpage_kasan_tagged(struct page *page)
> 
>     static inline void tag_clear_highpage(struct page *page)
>     {
>    +     clear_highpage(page);
>     }
> 
> because even when the *real* tag_clear_highpage() triggers, it falls down to
> 
>          if (!system_supports_mte()) {
>                  clear_highpage(page);
>                  return;
>          }
> 
> so basically I think the fundamental bug here is that our fallback
> tag_clear_highpage() was just buggy and didn't do what it was supposed
> to do.
> 
> That one-liner would seem to be a lot simpler and more robust than
> making this configuration-dependent. Just make the fallback do the
> right thing - blammo, problem solved.
> 
> Am I missing something?

I had the same in mind for a second, but then I looked at 
kernel_init_pages() with the kasan_disable_current() handling and 
concluded that it's clearer to just disallow tag_clear_highpage() being 
abused in the first place and reduce the effective code footprint of 
post_alloc_hook().

-- 
Cheers

David