Re: [PATCH 0/3] migration: Error fixes and improvements

[Peter Xu <peterx@redhat.com>]

On Sat, Nov 15, 2025 at 09:34:57AM +0100, Markus Armbruster wrote:
> Maintainers decide what to take for 10.2, if anything.
> 
> Let me know if you'd prefer the "perhaps should take ownership" idea
> in PATCH 1's commit message.

I recall I had such patch previously, so I digged it out:

https://lore.kernel.org/all/20230705163502.331007-3-peterx@redhat.com/

I found that I dropped it in v3's post of that series, where I mentioned in
the change log that either way is not clean, so I dropped that until I
could have a better understanding:

https://lore.kernel.org/all/20231004220240.167175-1-peterx@redhat.com/

I think at that time I should have hit an use case where the caller forgot
to error_copy(), hence passing it in causing an UAF.  Then I thought memory
leak was better in error path comparing to UAF if the API was used wrong
either way.

But feel free to share your thoughts.  We can definitely revisit this.

I queued the series for this release, thanks Markus.

> 
> Markus Armbruster (3):
>   migration: Plug memory leaks after migrate_set_error()
>   migration: Use warn_reportf_err() where appropriate
>   migration/postcopy-ram: Improve error reporting after loadvm failure
> 
>  migration/cpr-exec.c     |  3 ++-
>  migration/multifd.c      |  6 ++++--
>  migration/postcopy-ram.c | 17 ++++++++---------
>  3 files changed, 14 insertions(+), 12 deletions(-)
> 
> -- 
> 2.49.0
> 
> 

-- 
Peter Xu