On Fri, Aug 23, 2024 at 11:02:06PM GMT, Zhiguo Jiang wrote: > After CoWed by do_wp_page, the vma established a new mapping relationship > with the CoWed folio instead of the non-CoWed folio. However, regarding > the situation where vma->anon_vma and the non-CoWed folio's anon_vma are > not same, the avc binding relationship between them will no longer be > needed, so it is issue for the avc binding relationship still existing > between them. > > This patch will remove the avc binding relationship between vma and the > non-CoWed folio's anon_vma, which each has their own independent > anon_vma. It can also alleviates rmap overhead simultaneously. > > Signed-off-by: Zhiguo Jiang <justinjiang@xxxxxxxx> NACK (until fixed). This is broken (see below). I'm not seeing any numbers to back anything up here as to why we want to make changes to this incredibly sensitive code? Also anon_vma logic is very complicated and confusing, this commit message feels about 3 paragraphs too light. Under what circumstances will vma->anon_vma be different from folio_anon_vma(non_cowed_folio)? etc. Confusing topics strongly require explanations that help (somewhat) compensate. This is one of them. > --- > > -v2: > * Solve the kernel test robot noticed "WARNING" > Reported-by: kernel test robot <oliver.sang@xxxxxxxxx> > Closes: https://lore.kernel.org/oe-lkp/202408230938.43f55b4-lkp@xxxxxxxxx It doesn't. Saw a bunch of warning output in dmesg when running in qemu, bisected it to this commit. The below assert is being fired (did you build this kernel with CONFIG_DEBUG_VM?): VM_WARN_ON(anon_vma->num_children);
