Re: [PATCH v6 14/41] x86/mm: Introduce _PAGE_SAVED_DIRTY
From: "Edgecombe, Rick P" <rick.p.edgecombe@intel.com>
Date: 2023-02-21 20:08:49
Also in:
linux-arch, linux-doc, linux-mm, lkml
From: "Edgecombe, Rick P" <rick.p.edgecombe@intel.com>
Date: 2023-02-21 20:08:49
Also in:
linux-arch, linux-doc, linux-mm, lkml
On Tue, 2023-02-21 at 09:38 +0100, David Hildenbrand wrote:
Again, I'd not specialize on COW in all patches to much (IMHO, it creates more confusion than it actually helps for understanding what's happening) and just call it a read-only PTE that is dirty. Simple as that. And it's easy to see why that's problematic, because read-only PTEs that are dirty would be identified as shadow stack PTEs, which we want to work around. Again, just my 2 cents. I'm not an x86 maintainer ;)
Right, I see the point. Let's see if they have any opinion. There is a bit of a historical reason for the focus on COW. As you well know the dirty bit used to be important for that case. But I think it's still not a terrible example. It covers some typical cases, but yes we don't want to mislead the reader that it is a Cow only scenario.