Re: [PATCH v2 1/3] kasan: Avoid sleepable page allocation from atomic context

On Wed, Apr 09, 2025 at 04:56:29PM +0200, Andrey Ryabinin wrote:

Hi Andrey,

...

...

...

...

• page = __get_free_page(GFP_KERNEL);

• page = __get_free_page(GFP_ATOMIC); if (!page)

I think a better way to fix this would be moving out allocation from atomic context. Allocate page prior to apply_to_page_range() call and pass it down to kasan_populate_vmalloc_pte().

I think the page address could be passed as the parameter to kasan_populate_vmalloc_pte().

We'll need to pass it as 'struct page **page' or maybe as pointer to some struct, e.g.: struct page_data { struct page *page; };

...

Thanks for the hint! I will try to implement that, but will likely start in two weeks, after I am back from vacation.

Not sure wether this version needs to be dropped.

Thanks!