On Thu, Jan 23, 2025 at 02:42:56PM -0500, Liam R. Howlett wrote:
> * Roman Gushchin <roman.gushchin@xxxxxxxxx> [250123 11:44]:
> > Commit b67fbebd4cf9 ("mmu_gather: Force tlb-flush VM_PFNMAP vmas")
> > added a forced tlbflush to tlb_vma_end(), which is required to avoid a
> > race between munmap() and unmap_mapping_range(). However it added some
> > overhead to other paths where tlb_vma_end() is used, but vmas are not
> > removed, e.g. madvise(MADV_DONTNEED).
> >
> > Fix this by moving the tlb flush out of tlb_end_vma() into
> > free_pgtables(), somewhat similar to the stable version of the
> > original commit: e.g. stable commit 895428ee124a ("mm: Force TLB flush
> > for PFNMAP mappings before unlink_file_vma()").
> >
> > Note, that if tlb->fullmm is set, no flush is required, as the whole
> > mm is about to be destroyed.
> >
> > ---
>
> Hugh didn't mean to add a ---, he meant to move the version info between
> the Cc list and the patch so that it's not in the git history.
Ah, ok, somehow it wasn't a problem previously for me, usually Andrew
was dropping the part while merging into the mm tree.
Thanks for clarifying.
>
> You can find examples on the ML.
>
> >
> > v3:
> > - added initialization of vma_pfn in __tlb_reset_range() (by Hugh D.)
> >
> > v2:
> > - moved vma_pfn flag handling into tlb.h (by Peter Z.)
> > - added comments (by Peter Z.)
> > - fixed the vma_pfn flag setting (by Hugh D.)
> >
>
>
> > Suggested-by: Jann Horn <jannh@xxxxxxxxxx>
> > Signed-off-by: Roman Gushchin <roman.gushchin@xxxxxxxxx>
>
> Link: some email discussion url lore.kernel.org..
It was based on some non-public discussion unfortunately, so no lore link.
Thank you!
