Hi Wolfram,
> On August 8, 2018 at 9:59 AM Wolfram Sang <[email protected]>
> wrote:
>
> Use BIT macro to avoid shift-31-problem,
May I ask how exactly you spotted the "shift-31-problem" in
drivers/i2c/busses/i2c-rcar.c:
- visual code review?
- static analysis, special compiler flags?
- run-time testing with UBSAN=y?
The background of my question is seeing a lot of UBSAN errors caused
by (1 << 31) in U-Boot with UBSAN=y [1]. Based on my experiments, the
Linux UBSAN simply doesn't complain about (1 << 31). I traced the
difference to be caused by the -std=gnu{89,11} value passed to gcc,
which varies between U-Boot and Linux.
According to feedback from GCC community [2], with 'gcc -std=gnu89',
shifting into (not past) the sign bit is "defined behavior" which is why
UBSAN doesn't report this as an issue in Linux kernel. That makes me
curious about the background of this patch, since it might shed some
light onto how to further handle the (1 << 31) UBSAN warnings in
U-Boot.
Thank you very much for your feedback.
[1] https://patchwork.ozlabs.org/cover/962307/
[2] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87392
Best regards,
Eugeniu.
