On Fri, Feb 14, 2025 at 10:59:11AM -0500, Laine Stump wrote:
On 2/14/25 9:21 AM, Andrea Bolognani wrote:
> You can restrict the check to just the vhost-user passt case, leaving
> other vhost-user cases alone, so that we don't have to worry about
> accidentally breaking existing configurations.
I will also send a separate patch that removes the "passt-only" aspect and
does it for vhost-user as well; since validation isn't run on existing
configs, and any existing config that had vhost-user but no shared memory
would have previously failed to function anyway, I don't see much danger in
validating it for plain vhost-user as well.
That sounds good, as long as it's a separate patch that this series
doesn't depend on. I'm highly confident that we can do that safely
for the passt case, slightly less so for the general case, and I
wouldn't want to hold up the former because of the latter.
> You need to respin to add the documentation anyway, don't
you?
If there weren't any other major problems, I wasn't planning to resend the
entire series for that, just add a "Patch 10/9".
Please do a quick resend even if the changes are minimal. That gives
reviewers the chance to take one last look before the series is
merged. And don't forget to update the release notes too ;)
--
Andrea Bolognani / Red Hat / Virtualization