On Sat, Jun 20, 2020 at 5:57 PM Kees Cook [off-list ref] wrote:

On Sat, Jun 20, 2020 at 09:03:34AM +0200, Sedat Dilek wrote:

quoted

On Sat, Jun 20, 2020 at 5:30 AM Kees Cook [off-list ref] wrote:

quoted

v2:
- more special-cased fixes
- add reviews
v1: https://lore.kernel.org/lkml/20200603233203.1695403-1-keescook@chromium.org (local)

Using uninitialized_var() is dangerous as it papers over real bugs[1]
(or can in the future), and suppresses unrelated compiler warnings
(e.g. "unused variable"). If the compiler thinks it is uninitialized,
either simply initialize the variable or make compiler changes.

As recommended[2] by[3] Linus[4], remove the macro.

Most of the 300 uses don't cause any warnings on gcc 9.3.0, so they're in
a single treewide commit in this series. A few others needed to actually
get cleaned up, and I broke those out into individual patches.

The tree is:
https://git.kernel.org/pub/scm/linux/kernel/git/kees/linux.git/log/?h=kspp/uninit/macro

-Kees

Hi Kees,

thanks for doing a v2 of your patchset.

As I saw Jason Yan providing some "uninitialized_var() macro" patches
to the MLs I pointen him to your tree "v1".

Thanks!

quoted

BTW, I have tested your "v1" against Linux v5.7 (see [1]) - just
yesterday with Linux v5.7.5-rc1.

Is it possible to have a v2 of this patchset on top od Linux v5.7 - if
you do not mind.

Since it's only going to be for post-v5.8, I'm fine skipping the v5.7
testing. Mainly I'm looking at v5.8 and linux-next.

Thanks for looking at it!

Thanks for the feedback.

"I knew you'd say that."
( Judge Dredd )

- Sedat -