Hi Steven,

On Sun, 24 Oct 2021 22:42:15 -0400 Steven Rostedt [off-list ref] wrote:

On Sun, 24 Oct 2021 21:11:19 -0500
"Gustavo A. R. Silva" [off-list ref] wrote:

quoted

On Mon, Oct 25, 2021 at 12:54:46PM +1100, Stephen Rothwell wrote:  

quoted

quoted

--- a/include/linux/ftrace.h
+++ b/include/linux/ftrace.h

@@ -45,6 +45,7 @@ struct ftrace_ops;
 void arch_ftrace_ops_list_func(unsigned long ip, unsigned long parent_ip);
 #else
 # define FTRACE_FORCE_LIST_FUNC 0
+struct ftrace_regs;
 void arch_ftrace_ops_list_func(unsigned long ip, unsigned long parent_ip,
 			       struct ftrace_ops *op, struct ftrace_regs *fregs);
 #endif

-- 
2.33.0    

I am still applying this fix ...    

I will add this fix to the original patch for my tomorrow's -next tree.  

Hmm, I pushed my version to my for-next branch. Do you have this in the
next tree?

  34cdd18b8d245f3e901e5325 ("tracing: Use linker magic instead of recasting
  ftrace_ops_list_func()")

Which shouldn't need this fix.

Well, that is in your tree today, but (1) doesn't help with the build
failure in Gustavo's tree on its own (i.e. it creates a dependency
between Gustavo's tree and your tree) and (2) now I will get a conflict
between the two trees today.  Oh, well.

-- 
Cheers,
Stephen Rothwell