On Wed, Oct 02, 2019 at 01:21:21PM -0700, David Miller wrote:
From: Kees Cook <redacted>
Date: Wed, 2 Oct 2019 11:19:16 -0700
quoted
On Thu, Sep 26, 2019 at 01:56:55PM -0700, Kees Cook wrote:
quoted
On Thu, Sep 26, 2019 at 01:06:01PM -0700, Linus Torvalds wrote:
quoted
(a) why didn't this use the already existing and well-named macro
that nobody really had issues with?
That was suggested, but other folks wanted the more accurate "member"
instead of "field" since a treewide change was happening anyway:
https://www.openwall.com/lists/kernel-hardening/2019/07/02/2
At the end of the day, I really don't care -- I just want to have _one_
macro. :)
quoted
(b) I see no sign of the networking people having been asked about
their preferences.
Yeah, that's entirely true. Totally my mistake; it seemed like a trivial
enough change that I didn't want to bother too many people. But let's
fix that now... Dave, do you have any concerns about this change of
FIELD_SIZEOF() to sizeof_member() (or if it prevails, sizeof_field())?
David, can you weight in on this? Are you okay with a mass renaming of
FIELD_SIZEOF() to sizeof_member(), as the largest user of the old macro
is in networking?
I have no objection to moving to sizeof_member().
Great; thank you!
Linus, are you still open to taking this series with Dave's buy-in? I'd
really hate to break it up since it's such a mechanical treewide
change. I'm also happy to wait until the next -rc1 window; whatever you
think is best here.
Thanks!
--
Kees Cook