[linux-next:master] BUILD REGRESSION 2dac75696c6da3c848daa118a729827541c89d33

Philip Li philip.li at intel.com
Thu Oct 19 23:23:31 UTC 2023


On Thu, Oct 19, 2023 at 05:32:05PM +0200, Heiko Carstens wrote:
> On Thu, Oct 19, 2023 at 04:07:35AM +0800, kernel test robot wrote:
> > arch/s390/include/asm/ctlreg.h:129:9: warning: array subscript 0 is outside array bounds of 'struct ctlreg[0]' [-Warray-bounds=]
> > arch/s390/include/asm/ctlreg.h:80:9: warning: array subscript 0 is outside array bounds of 'struct ctlreg[0]' [-Warray-bounds=]
> ...
> > |-- s390-defconfig
> > |   `-- arch-s390-include-asm-ctlreg.h:warning:array-subscript-is-outside-array-bounds-of-struct-ctlreg
> ...
> > s390                                defconfig   gcc  
> 
> I'm wondering how this warning can appear in the builds. array-bounds
> warnings are explicitly disabled, see init/Kconfig: CC_NO_ARRAY_BOUNDS. And
> as expected, if I compile the kernel with gcc, defconfig, and with or
> without W=1 the option -Wno-array-bounds is passed to the compiler.
> 
> And also as expected I do not see the above warnings.
> 
> So something is quite odd here.

Sorry about this Heiko, this is a bug in the bot that it wrongly ignores
the CC_NO_ARRAY_BOUNDS config and always test with -Warray-bounds. We
will fix this asap.

> 


More information about the amd-gfx mailing list