Re: [RFC PATCH 1/6] mmu_notifier: add event information to address invalidation v4

[Jerome Glisse]

On Thu, Sep 11, 2014 at 01:00:52PM +0300, Haggai Eran wrote:
> On 29/08/2014 22:10, j.glisse@xxxxxxxxx wrote:
> > + * - MMU_MUNMAP: the range is being unmapped (outcome of a munmap syscall or
> > + * process destruction). However, access is still allowed, up until the
> > + * invalidate_range_free_pages callback. This also implies that secondary
> > + * page table can be trimmed, because the address range is no longer valid.
> 
> I couldn't find the invalidate_range_free_pages callback. Is that a left over 
> from a previous version of the patch?
> 
> Also, I think that you have to invalidate the secondary PTEs of the range being 
> unmapped immediately, because put_page may be called immediately after the 
> invalidate_range_start returns.

This is because patchset was originaly on top of a variation of another
patchset :

https://lkml.org/lkml/2014/9/9/601

In which invalidate_range_free_pages was a function call right after cpu
page table is updated but before page are free. Hence the comment was
right if on top of that patchset but on top of master you are right this
comment is wrong.

Cheers,
Jérôme

> 
> Haggai
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/