Re: [PATCH 5/7] tracing: add config option for print arguments in ftrace

On Mon, 09 Sep 2024 10:16:55 +0200
Sven Schnelle [off-list ref] wrote:

Masami Hiramatsu (Google) [off-list ref] writes:

quoted

On Fri, 6 Sep 2024 11:36:11 +0800
Zheng Yejian [off-list ref] wrote:

quoted

On 2024/9/4 14:58, Sven Schnelle wrote:

quoted

Add a config option to disable/enable function argument
printing support during runtime.

Signed-off-by: Sven Schnelle <svens@linux.ibm.com>
---
  kernel/trace/Kconfig | 12 ++++++++++++
  1 file changed, 12 insertions(+)

diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig
index 721c3b221048..8b9b6cdf39ac 100644
--- a/kernel/trace/Kconfig
+++ b/kernel/trace/Kconfig

@@ -242,6 +242,18 @@ config FUNCTION_GRAPH_RETVAL
  	  enable it via the trace option funcgraph-retval.
  	  See Documentation/trace/ftrace.rst
  
+config FUNCTION_TRACE_ARGS
+	bool "Kernel Function Tracer Arguments"
+	depends on HAVE_FUNCTION_ARG_ACCESS_API
+	depends on DEBUG_INFO_BTF && BPF_SYSCALL

Nice feature!

Just a nit, DEBUG_INFO_BTF currently depends on BPF_SYSCALL,
so BPF_SYSCALL may not be necessary here. This feature
also doesn't seem to depend on bpf.

Indeed. Sven, you can check the PROBE_EVENTS_BTF_ARGS as
an example.

config PROBE_EVENTS_BTF_ARGS
        depends on HAVE_FUNCTION_ARG_ACCESS_API
        depends on FPROBE_EVENTS || KPROBE_EVENTS
        depends on DEBUG_INFO_BTF && BPF_SYSCALL
        bool "Support BTF function arguments for probe events"

Now i'm confused - Zheng wrote that DEBUG_INFO_BTF depends on
BPF_SYSCALL. I just verified that. So i could just remove BPF_SYSCALL
from the dependencies - but your example also has BPF_SYSCALL listed.

Ah, sorry for confusion. In this case, just need DEBUG_INFO_BTF.
Hmm, I think I also need to remove BPF_SYSCALL.

Thanks,

Regards
Sven

-- 
Masami Hiramatsu (Google) [off-list ref]