On 2022-12-05 13:56, Christophe Leroy wrote:
Le 05/12/2022 à 19:19, Michael Jeanson a écrit :
[Vous ne recevez pas souvent de courriers de mjeanson@efficios.com. Découvrez pourquoi ceci est important à https://aka.ms/LearnAboutSenderIdentification ]
On 2022-12-05 00:34, Michael Ellerman wrote:
Michael Jeanson mjeanson@efficios.com writes:
In v5.7 the powerpc syscall entry/exit logic was rewritten in C, on PPC64_ELF_ABI_V1 this resulted in the symbols in the syscall table changing from their dot prefixed variant to the non-prefixed ones.
Since ftrace prefixes a dot to the syscall names when matching them to build its syscall event list, this resulted in no syscall events being available.
Remove the PPC64_ELF_ABI_V1 specific version of arch_syscall_match_sym_name to have the same behavior across all powerpc variants.
This doesn't seem to work for me.
Event with it applied I still don't see anything in /sys/kernel/debug/tracing/events/syscalls
Did we break it in some other way recently?
cheers
I've just tried this change on top of v6.1-rc8 in qemu with a base config of 'corenet32_smp_defconfig' and these options on top:
CONFIG_FTRACE=y CONFIG_FTRACE_SYSCALLS=y
And I can trace syscalls with ftrace.
What kernel tree and config are you using?
If you are using a ppc32 config, CONFIG_PPC64_ELF_ABI_V1 won't be set, so it doesn't matter whether this change is there or not.
You should try corenet64_smp_defconfig if you want CONFIG_PPC64_ELF_ABI_V1 to be set.
You can also use ppc64_defconfig, that's a different platform but it also has CONFIG_PPC64_ELF_ABI_V1.
You are absolutely right, I used the wrong environment, I blame Monday morning. I tested again this time using 'corenet64_smp_defconfig' with the same options and syscall tracing with ftrace also works.
I double checked that /proc/config.gz contained CONFIG_PPC64_ELF_ABI_V1 to be sure.
Christophe
Thanks for looking into this.