Re: [RFC PATCH 0/5] madvise MADV_DOEXEC

[Steven Sistare]

On 7/27/2020 1:11 PM, Anthony Yznaga wrote:
> This patchset adds support for preserving an anonymous memory range across
> exec(3) using a new madvise MADV_DOEXEC argument.  The primary benefit for
> sharing memory in this manner, as opposed to re-attaching to a named shared
> memory segment, is to ensure it is mapped at the same virtual address in
> the new process as it was in the old one.  An intended use for this is to
> preserve guest memory for guests using vfio while qemu exec's an updated
> version of itself.  By ensuring the memory is preserved at a fixed address,
> vfio mappings and their associated kernel data structures can remain valid.
> In addition, for the qemu use case, qemu instances that back guest RAM with
> anonymous memory can be updated.

I forgot to mention, our use case is not just theoretical.  It has been implemented
and is pretty cool (but I am biased).  The pause time for the guest is in the
100 - 200 msec range.  We submitted qemu patches for review based on the MADV_DOEXEC
proposal.  In case you are curious:
  https://lore.kernel.org/qemu-devel/1596122076-341293-1-git-send-email-steven.sistare@xxxxxxxxxx/

- Steve