Mass-building defconfigs: many fail with assembler errors

[Segher Boessenkool]

Hi!

On Mon, Dec 12, 2022 at 10:51:17PM +0100, Jan-Benedict Glaw wrote:
> Is anybody else routinely building current Binutils + GCC, to try to
> build all the Linux defconfigs?

I do regularly build kernels for powerpc-linux, powerpc64-linux,
powerpc64le-linux; ppc6xx_defconfig and ppc64_defconfig and
ppc64le_defconfig.  Totally boring, but it frequently does not build.
Not as frequently as for the other Linux targets I build (32 total at
the moment).

> For PPC, a good number of those fail,
> and I probably don't understand PPC well enough to propose patches. Or
> did I pick wrongly targeted toolchains? Most of the time, my suspicion
> is that we're not giving the correct -m<cpu> flags in
> ./arch/powerpc/boot/?  (My setup for doing test builds is fairly automated, I
> can easily throw in patches for testing.)

Many of those use a 32-bit toolchain with a 64-bit kernel, or they
require some e500 specific config but not getting it (or the other way
around).

> 64-bit.config

> 	==> Why "-m32 -mcpu=powerpc"? Binutils/GCC are for --target=powerpc64-linux

Something in your config is forcing that.

> 85xx-64bit.config
> 	  powerpc64-linux-gcc -Wp,-MMD,arch/powerpc/kernel/vdso/.gettimeofday-64.o.d -nostdinc -I./arch/powerpc/include -I./arch/powerpc/include/generated  -I./include -I./arch/powerpc/include/uapi -I./arch/powerpc/include/generated/uapi -I./include/uapi -I./include/generated/uapi -include ./include/linux/compiler-version.h -include ./include/linux/kconfig.h -D__KERNEL__ -I ./arch/powerpc -DHAVE_AS_ATHIGH=1 -fmacro-prefix-map=./= -D__ASSEMBLY__ -fno-PIE -m64 -Wl,-a64 -mabi=elfv1 -Wa,-me500 -Wa,-me500mc -mabi=elfv1 -mbig-endian    -Wl,-soname=linux-vdso64.so.1 -D__VDSO64__ -s -c -o arch/powerpc/kernel/vdso/gettimeofday-64.o arch/powerpc/kernel/vdso/gettimeofday.S

e500mc is a 32-bit core.  This cannot fly.

> 85xx-hw.config
> 	  powerpc-linux-gcc -Wp,-MMD,arch/powerpc/kernel/.epapr_hcalls.o.d -nostdinc -I./arch/powerpc/include -I./arch/powerpc/include/generated  -I./include -I./arch/powerpc/include/uapi -I./arch/powerpc/include/generated/uapi -I./include/uapi -I./include/generated/uapi -include ./include/linux/compiler-version.h -include ./include/linux/kconfig.h -D__KERNEL__ -I ./arch/powerpc -fmacro-prefix-map=./= -D__ASSEMBLY__ -fno-PIE -m32 -Wl,-a32 -mcpu=powerpc -mbig-endian    -c -o arch/powerpc/kernel/epapr_hcalls.o arch/powerpc/kernel/epapr_hcalls.S 
> 	arch/powerpc/kernel/epapr_hcalls.S: Assembler messages:
> 	arch/powerpc/kernel/epapr_hcalls.S:24: Error: unrecognized opcode: `wrteei'

wrteei is a BookE instruction (not a PowerPC instruction), and something
specifically asked for just PowerPC.

> powernv_defconfig

> 	cc1: all warnings being treated as errors

Self-inflicted wound.  The warnings may reveal more, or they might show
something that GCC isn't as good at as you may want; all warnings have
false positives, that is why they are warnings and not errors.

> 	Compiler ICEs (during GIMPLE pass: ccp) in align.c:
> 
> 	  powerpc-linux-gcc -Wp,-MMD,arch/powerpc/kernel/.align.o.d -nostdinc -I./arch/powerpc/include -I./arch/powerpc/include/generated  -I./include -I./arch/powerpc/include/uapi -I./arch/powerpc/include/generated/uapi -I./include/uapi -I./include/generated/uapi -include ./include/linux/compiler-version.h -include ./include/linux/kconfig.h -include ./include/linux/compiler_types.h -D__KERNEL__ -I ./arch/powerpc -fmacro-prefix-map=./= -Wall -Wundef -Werror=strict-prototypes -Wno-trigraphs -fno-strict-aliasing -fno-common -fshort-wchar -fno-PIE -Werror=implicit-function-declaration -Werror=implicit-int -Werror=return-type -Wno-format-security -std=gnu11 -mbig-endian -m32 -msoft-float -pipe -ffixed-r2 -mmultiple -mno-readonly-in-sdata -mcpu=440 -mno-prefixed -mno-pcrel -mno-altivec -mno-vsx -mno-mma -fno-asynchronous-unwind-tables -mno-string -Wa,-m440 -mbig-endian -mstack-protector-guard=tls -mstack-protector-guard-reg=r2 -fno-delete-null-pointer-checks -Wno-frame-address -Wno-format-truncation -Wno-format-overflow -Wno-address-of-packed-member -O2 -fno-allow-store-data-races -Wframe-larger-than=1024 -fstack-protector-strong -Wno-main -Wno-unused-but-set-variable -Wno-unused-const-variable -Wno-dangling-pointer -fomit-frame-pointer -ftrivial-auto-var-init=zero -fno-stack-clash-protection -Wdeclaration-after-statement -Wvla -Wno-pointer-sign -Wcast-function-type -Wno-stringop-truncation -Wno-stringop-overflow -Wno-restrict -Wno-maybe-uninitialized -Wno-alloc-size-larger-than -Wimplicit-fallthrough=5 -fno-strict-overflow -fno-stack-check -fconserve-stack -Werror=date-time -Werror=incompatible-pointer-types -Werror=designated-init -Wno-packed-not-aligned -g -mstack-protector-guard-offset=1080 -Werror    -DKBUILD_MODFILE='"arch/powerpc/kernel/align"' -DKBUILD_BASENAME='"align"' -DKBUILD_MODNAME='"align"' -D__KBUILD_MODNAME=kmod_align -c -o arch/powerpc/kernel/align.o arch/powerpc/kernel/align.c  
> 	during GIMPLE pass: ccp
> 	arch/powerpc/kernel/align.c: In function '__copy_inst_from_kernel_nofault':
> 	arch/powerpc/kernel/align.c:364:1: internal compiler error: in maybe_register_def, at tree-into-ssa.cc:1948
> 	  364 | }
> 	      | ^
> 	0x19d8886 internal_error(char const*, ...)
> 	   ???:0
> 	0x7bb4fe fancy_abort(char const*, int, char const*)
> 	   ???:0
> 	0x1791bfe dom_walker::walk(basic_block_def*)
> 	   ???:0
> 	0xe94ec0 update_ssa(unsigned int)
> 	   ???:0
> 	0x103d6b9 execute_update_addresses_taken()
> 	   ???:0
> 	Please submit a full bug report, with preprocessed source (by using -freport-bug).
> 	Please include the complete backtrace with any bug report.
> 	See <https://gcc.gnu.org/bugs/> for instructions.
> 	make[3]: *** [scripts/Makefile.build:250: arch/powerpc/kernel/align.o] Error 1
> 	make[2]: *** [scripts/Makefile.build:500: arch/powerpc/kernel] Error 2
> 	make[1]: *** [scripts/Makefile.build:500: arch/powerpc] Error 2
> 	make: *** [Makefile:1992: .] Error 2
> 
> 	==> Should probably open a PR for this.

Yes please!


Segher