On Wed, Aug 31, 2022 at 01:46:09AM +0200, Ilya Leoshkevich wrote: > On Tue, 2022-08-30 at 18:46 +0200, Ilya Leoshkevich wrote: > > On Tue, 2022-08-30 at 15:48 +0200, Jiri Olsa wrote: > > > On Tue, Aug 30, 2022 at 12:25:25AM +0200, Daniel Borkmann wrote: > > > > On 8/26/22 8:46 PM, Jiri Olsa wrote: > > > > > hi, > > > > > as discussed [1] sending fix that moves bpf dispatcher function > > > > > of out > > > > > ftrace locations together with Peter's > > > > > HAVE_DYNAMIC_FTRACE_NO_PATCHABLE > > > > > dependency change. > > > > > > > > Looks like the series breaks s390x builds; BPF CI link: > > > > > > > > https://github.com/kernel-patches/bpf/runs/8079411784?check_suite_focus=true > > > > > > > > [...] > > > > CC net/xfrm/xfrm_state.o > > > > CC net/packet/af_packet.o > > > > {standard input}: Assembler messages: > > > > {standard input}:16055: Error: bad expression > > > > {standard input}:16056: Error: bad expression > > > > {standard input}:16057: Error: bad expression > > > > {standard input}:16058: Error: bad expression > > > > {standard input}:16059: Error: bad expression > > > > CC drivers/s390/char/raw3270.o > > > > CC net/ipv6/ip6_output.o > > > > [...] > > > > CC net/xfrm/xfrm_output.o > > > > CC net/ipv6/ip6_input.o > > > > {standard input}:16055: Error: invalid operands (*ABS* and > > > > *UND* > > > > sections) for `%' > > > > {standard input}:16056: Error: invalid operands (*ABS* and > > > > *UND* > > > > sections) for `%' > > > > {standard input}:16057: Error: invalid operands (*ABS* and > > > > *UND* > > > > sections) for `%' > > > > {standard input}:16058: Error: invalid operands (*ABS* and > > > > *UND* > > > > sections) for `%' > > > > {standard input}:16059: Error: invalid operands (*ABS* and > > > > *UND* > > > > sections) for `%' > > > > make[3]: *** [scripts/Makefile.build:249: net/core/filter.o] > > > > Error 1 > > > > make[2]: *** [scripts/Makefile.build:465: net/core] Error 2 > > > > make[2]: *** Waiting for unfinished jobs.... > > > > CC net/ipv4/tcp_fastopen.o > > > > [...] > > > > CC lib/percpu-refcount.o > > > > make[1]: *** [Makefile:1855: net] Error 2 > > > > CC lib/rhashtable.o > > > > make[1]: *** Waiting for unfinished jobs.... > > > > CC lib/base64.o > > > > [...] > > > > AR lib/built-in.a > > > > CC kernel/kheaders.o > > > > AR kernel/built-in.a > > > > make: *** [Makefile:353: __build_one_by_one] Error 2 > > > > Error: Process completed with exit code 2. > > > > > > > > > it does not break on my cross build with gcc 12, but I can > > > reproduce with gcc 8 (CI seems to be on gcc 9) > > > > > > the problem seems to be wrong assembler code with extra '%' > > > that's generated for patchable_function_entry(5) > > > > > > gcc 8 generates: > > > > > > .LPFE1: > > > nopr %%r0 > > > nopr %%r0 > > > nopr %%r0 > > > nopr %%r0 > > > nopr %%r0 > > > > > > and gcc 12 generates: > > > > > > .LPFE1: > > > nopr %r0 > > > nopr %r0 > > > nopr %r0 > > > nopr %r0 > > > nopr %r0 > > > > > > perhaps we need to upgrade gcc in CI? cc-ing Ilya, any idea? > > > > > > thanks, > > > jirka > > > > It's not obvious to me which gcc commit fixed this; I will bisect and > > find out. This will take some time. > > > > However, officially, the kernel must be buildable by gcc 5.1+. > > Whatever I find, it's unlikely that we'll be able to backport it > > that far. > > > > Therefore I think we need to find a way to conditionally > > do something else when using broken gccs. Or maybe just keep this > > x86-only after all. > > > > Best regards, > > Ilya > > FWIW, bisect points to > > https://gcc.gnu.org/git/?p=gcc.git;a=commit;h=45d06a4045bebc3dbaaf0b1c676f4e22b7c6aca1 great, thanks for doing that > > which makes perfect sense. Still, as I mentioned above, it's probably > worth tolerating brokens gccs instead of spending time backporting this > everywhere. And upgrading the CI machine will only paper over the > issue. > > At a closer look, it looks weird to me that we have > patchable_function_entry(5) in a common header. If this optimization > is ever implemented for another architecture, a different number will > be required. > > For simplicity, would it make sense to hide this under an #ifdef? > Something like this (untested): > > #ifdef CONFIG_X86 > #define BPF_DISPATCHER_ATTRIBUTES > __attribute__((patchable_function_entry(5))) > #else > #define BPF_DISPATCHER_ATTRIBUTES > #endif right, I think we can limit it directly to x86_64 like below jirka --- diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index f9920f1341c8..089c20cefd2b 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -284,6 +284,7 @@ config X86 select PROC_PID_ARCH_STATUS if PROC_FS select HAVE_ARCH_NODE_DEV_GROUP if X86_SGX imply IMA_SECURE_AND_OR_TRUSTED_BOOT if EFI + select HAVE_DYNAMIC_FTRACE_NO_PATCHABLE config INSTRUCTION_DECODER def_bool y diff --git a/include/linux/bpf.h b/include/linux/bpf.h index 9c1674973e03..4ab4b0a1beb8 100644 --- a/include/linux/bpf.h +++ b/include/linux/bpf.h @@ -924,7 +924,15 @@ int arch_prepare_bpf_dispatcher(void *image, s64 *funcs, int num_funcs); }, \ } +#ifdef CONFIG_X86_64 +#define BPF_DISPATCHER_ATTRIBUTES __attribute__((__no_instrument_function__)) \ + __attribute__((patchable_function_entry(5))) +#else +#define BPF_DISPATCHER_ATTRIBUTES +#endif + #define DEFINE_BPF_DISPATCHER(name) \ + BPF_DISPATCHER_ATTRIBUTES \ noinline __nocfi unsigned int bpf_dispatcher_##name##_func( \ const void *ctx, \ const struct bpf_insn *insnsi, \
