Re: [PATCH] mm: move mask update out of the atomic context

[David Hildenbrand]

On 23.06.25 12:09, Dev Jain wrote:
> On 23/06/25 3:07 pm, Alexander Gordeev wrote:
> > On Mon, Jun 23, 2025 at 02:26:29PM +0530, Dev Jain wrote:
> > > On 23/06/25 1:34 pm, Alexander Gordeev wrote:
> > > > There is not need to modify page table synchronization mask
> > > > while apply_to_pte_range() holds user page tables spinlock.
> > > I don't get you, what is the problem with the current code?
> > > Are you just concerned about the duration of holding the
> > > lock?
> > Yes.
>
> Doesn't really matter but still a correct change:

Let's ask the real questions: who checks PGTBL_PTE_MODIFIED?

I see

if (mask & ARCH_PAGE_TABLE_SYNC_MASK)
	arch_sync_kernel_mappings(start, start + size);

And then

arch/arm/include/asm/page.h:#define ARCH_PAGE_TABLE_SYNC_MASK   PGTBL_PMD_MODIFIED
arch/x86/include/asm/pgtable-2level_types.h:#define ARCH_PAGE_TABLE_SYNC_MASK   PGTBL_PMD_MODIFIED
arch/x86/include/asm/pgtable-3level_types.h:#define ARCH_PAGE_TABLE_SYNC_MASK   PGTBL_PMD_MODIFIED


Which makes me wonder why we need PGTBL_PTE_MODIFIED at all? Is there some other check I am missing?

(same question regarding everything excepy PGTBL_PMD_MODIFIED, because that actually seems to be used)

--
Cheers,

David / dhildenb