Is this a regression? Has this compiler ever been able to build this arch like this?
Yes. It is a regression with gcc-11.
stable rc Linux 5.13.14 with gcc-11 - powerpc - FAILED stable rc Linux 5.13.14 with gcc-10 - powerpc - PASSED
Ah, ok, and does 5.14 or newer work properly?
No. stable rc Linux 5.14.1-rc1 with gcc-11 - powerpc - FAILED stable rc Linux 5.14.1-rc1 with gcc-10 - powerpc - PASS
I will check Linux mainline and Linux next and get back to you.
Since we started building with gcc-11 last weekend and started noticing these build failures on stable-rc, mainline and next. and today's stable-rc review also failed with these combinations.
Linux next master with gcc-11 - powerpc-defconfig - FAILED Linux mainline master with gcc-11 - powerpc-defconfig - FAILED stable rc Linux 5.14.2-rc1 with gcc-11 - powerpc-defconfig - FAILED stable rc Linux 5.13.15-rc1 with gcc-11 - powerpc-defconfig - FAILED stable rc Linux 5.10.63-rc1 with gcc-11 - powerpc-defconfig - FAILED
FYI, The following powerpc config build PASS with gcc-11 - allnoconfig - mpc83xx_defconfig - tqm8xx_defconfig - maple_defconfig - cell_defconfig - ppc64e_defconfig - ppc6xx_defconfig
- Naresh