linux-next: manual merge of the ftrace tree with the arm64 tree

linux-next: manual merge of the ftrace tree with the arm64 tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 2256 bytes --]

Hi all,

Today's linux-next merge of the ftrace tree got a conflict in:

  include/asm-generic/vmlinux.lds.h

between commit:

  a1326b17ac03 ("module/ftrace: handle patchable-function-entry")

from the arm64 tree and commit:

  b83b43ffc6e4 ("fgraph: Fix function type mismatches of ftrace_graph_return using ftrace_stub")

from the ftrace tree.

I fixed it up (see below) and can carry the fix as necessary. This
is now fixed as far as linux-next is concerned, but any non trivial
conflicts should be mentioned to your upstream maintainer when your tree
is submitted for merging.  You may also want to consider cooperating
with the maintainer of the conflicting tree to minimise any particularly
complex conflicts.

-- 
Cheers,
Stephen Rothwell

diff --cc include/asm-generic/vmlinux.lds.h
index 63cedc3c0c77,0f358be551cd..000000000000
--- a/include/asm-generic/vmlinux.lds.h
+++ b/include/asm-generic/vmlinux.lds.h
@@@ -136,19 -110,30 +136,29 @@@
  #endif
  
  #ifdef CONFIG_FTRACE_MCOUNT_RECORD
 -#ifdef CC_USING_PATCHABLE_FUNCTION_ENTRY
 +/*
 + * The ftrace call sites are logged to a section whose name depends on the
 + * compiler option used. A given kernel image will only use one, AKA
 + * FTRACE_CALLSITE_SECTION. We capture all of them here to avoid header
 + * dependencies for FTRACE_CALLSITE_SECTION's definition.
 + */
+ /*
+  * Need to also make ftrace_graph_stub point to ftrace_stub
+  * so that the same stub location may have different protocols
+  * and not mess up with C verifiers.
+  */
 -#define MCOUNT_REC()	. = ALIGN(8);				\
 -			__start_mcount_loc = .;			\
 -			KEEP(*(__patchable_function_entries))	\
 -			__stop_mcount_loc = .;			\
 -			ftrace_graph_stub = ftrace_stub;
 -#else
  #define MCOUNT_REC()	. = ALIGN(8);				\
  			__start_mcount_loc = .;			\
  			KEEP(*(__mcount_loc))			\
 +			KEEP(*(__patchable_function_entries))	\
- 			__stop_mcount_loc = .;
+ 			__stop_mcount_loc = .;			\
+ 			ftrace_graph_stub = ftrace_stub;
 -#endif
  #else
- #define MCOUNT_REC()
+ # ifdef CONFIG_FUNCTION_TRACER
+ #  define MCOUNT_REC()	ftrace_graph_stub = ftrace_stub;
+ # else
+ #  define MCOUNT_REC()
+ # endif
  #endif
  
  #ifdef CONFIG_TRACE_BRANCH_PROFILING

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: linux-next: manual merge of the ftrace tree with the arm64 tree

[Mark Rutland]

On Fri, Nov 15, 2019 at 01:53:57PM +1100, Stephen Rothwell wrote:
> Hi all,
> 
> Today's linux-next merge of the ftrace tree got a conflict in:
> 
>   include/asm-generic/vmlinux.lds.h
> 
> between commit:
> 
>   a1326b17ac03 ("module/ftrace: handle patchable-function-entry")
> 
> from the arm64 tree and commit:
> 
>   b83b43ffc6e4 ("fgraph: Fix function type mismatches of ftrace_graph_return using ftrace_stub")
> 
> from the ftrace tree.
> 
> I fixed it up (see below) and can carry the fix as necessary. This
> is now fixed as far as linux-next is concerned, but any non trivial
> conflicts should be mentioned to your upstream maintainer when your tree
> is submitted for merging.  You may also want to consider cooperating
> with the maintainer of the conflicting tree to minimise any particularly
> complex conflicts.
> 
> -- 
> Cheers,
> Stephen Rothwell

Thanks Stephen!

Steven (Rostedt), what's your preference for handling this?

We can drop revert the arm64 change to vmlinux.lds.h for now (and I can
send it as a subsequent cleanup), or you could merge my
arm64/ftrace-with-regs branch and fix this up in the ftrace tree.

Thanks,
Mark.

> 
> diff --cc include/asm-generic/vmlinux.lds.h
> index 63cedc3c0c77,0f358be551cd..000000000000
> --- a/include/asm-generic/vmlinux.lds.h
> +++ b/include/asm-generic/vmlinux.lds.h
> @@@ -136,19 -110,30 +136,29 @@@
>   #endif
>   
>   #ifdef CONFIG_FTRACE_MCOUNT_RECORD
>  -#ifdef CC_USING_PATCHABLE_FUNCTION_ENTRY
>  +/*
>  + * The ftrace call sites are logged to a section whose name depends on the
>  + * compiler option used. A given kernel image will only use one, AKA
>  + * FTRACE_CALLSITE_SECTION. We capture all of them here to avoid header
>  + * dependencies for FTRACE_CALLSITE_SECTION's definition.
>  + */
> + /*
> +  * Need to also make ftrace_graph_stub point to ftrace_stub
> +  * so that the same stub location may have different protocols
> +  * and not mess up with C verifiers.
> +  */
>  -#define MCOUNT_REC()	. = ALIGN(8);				\
>  -			__start_mcount_loc = .;			\
>  -			KEEP(*(__patchable_function_entries))	\
>  -			__stop_mcount_loc = .;			\
>  -			ftrace_graph_stub = ftrace_stub;
>  -#else
>   #define MCOUNT_REC()	. = ALIGN(8);				\
>   			__start_mcount_loc = .;			\
>   			KEEP(*(__mcount_loc))			\
>  +			KEEP(*(__patchable_function_entries))	\
> - 			__stop_mcount_loc = .;
> + 			__stop_mcount_loc = .;			\
> + 			ftrace_graph_stub = ftrace_stub;
>  -#endif
>   #else
> - #define MCOUNT_REC()
> + # ifdef CONFIG_FUNCTION_TRACER
> + #  define MCOUNT_REC()	ftrace_graph_stub = ftrace_stub;
> + # else
> + #  define MCOUNT_REC()
> + # endif
>   #endif
>   
>   #ifdef CONFIG_TRACE_BRANCH_PROFILING

Re: linux-next: manual merge of the ftrace tree with the arm64 tree

[Steven Rostedt]

On Fri, 15 Nov 2019 12:25:14 +0000
Mark Rutland <mark.rutland@arm.com> wrote:

> Thanks Stephen!
> 
> Steven (Rostedt), what's your preference for handling this?
> 
> We can drop revert the arm64 change to vmlinux.lds.h for now (and I can
> send it as a subsequent cleanup), or you could merge my
> arm64/ftrace-with-regs branch and fix this up in the ftrace tree.

Or we can just make a note of this, and mention it in our pull requests
to Linus.

We could make a single branch where we merged the two branches, fix it
up (as I believe Stephen did it correctly) and tell Linus "here's the
fix up".

This is a common occurrence and Linus has no problems with this. As
long as he's given a heads up.

-- Steve

Re: linux-next: manual merge of the ftrace tree with the arm64 tree

[Catalin Marinas]

On Fri, Nov 15, 2019 at 08:48:27AM -0500, Steven Rostedt wrote:
> On Fri, 15 Nov 2019 12:25:14 +0000
> Mark Rutland <mark.rutland@arm.com> wrote:
> > Steven (Rostedt), what's your preference for handling this?
> > 
> > We can drop revert the arm64 change to vmlinux.lds.h for now (and I can
> > send it as a subsequent cleanup), or you could merge my
> > arm64/ftrace-with-regs branch and fix this up in the ftrace tree.
> 
> Or we can just make a note of this, and mention it in our pull requests
> to Linus.
> 
> We could make a single branch where we merged the two branches, fix it
> up (as I believe Stephen did it correctly) and tell Linus "here's the
> fix up".
> 
> This is a common occurrence and Linus has no problems with this. As
> long as he's given a heads up.

I'm fine sending Linus the pull request with the conflict as long as you
and Mark are ok with Stephen's resolution. The only nitpick I have is
that the two comments prior to MCOUNT_REC should be merged into a single
block (I guess we can leave this to Linus to sort out out).

-- 
Catalin

Re: linux-next: manual merge of the ftrace tree with the arm64 tree

[Steven Rostedt]

On Fri, 15 Nov 2019 13:59:24 +0000
Catalin Marinas <catalin.marinas@arm.com> wrote:

> I'm fine sending Linus the pull request with the conflict as long as you
> and Mark are ok with Stephen's resolution. The only nitpick I have is
> that the two comments prior to MCOUNT_REC should be merged into a single
> block (I guess we can leave this to Linus to sort out out).

Agreed,

I pushed up the merge conflict resolution to:

 git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace.git ftrace/aarch64-conflict

Feel free to point to that branch as a reference to the conflict
resolution.

-- Steve

Re: linux-next: manual merge of the ftrace tree with the arm64 tree

[Catalin Marinas]

On Fri, Nov 15, 2019 at 09:21:34AM -0500, Steven Rostedt wrote:
> On Fri, 15 Nov 2019 13:59:24 +0000
> Catalin Marinas <catalin.marinas@arm.com> wrote:
> 
> > I'm fine sending Linus the pull request with the conflict as long as you
> > and Mark are ok with Stephen's resolution. The only nitpick I have is
> > that the two comments prior to MCOUNT_REC should be merged into a single
> > block (I guess we can leave this to Linus to sort out out).
> 
> Agreed,
> 
> I pushed up the merge conflict resolution to:
> 
>  git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace.git ftrace/aarch64-conflict
> 
> Feel free to point to that branch as a reference to the conflict
> resolution.

Great. Thanks.

-- 
Catalin

linux-next: manual merge of the ftrace tree with the arm64 tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 2324 bytes --]

Hi all,

Today's linux-next merge of the ftrace tree got a conflict in:

  include/asm-generic/vmlinux.lds.h

between commit:

  a1326b17ac03 ("module/ftrace: handle patchable-function-entry")

from the arm64 tree and commits:

  b83b43ffc6e4 ("fgraph: Fix function type mismatches of ftrace_graph_return using ftrace_stub")
  46f9469247c6 ("ftrace: Rename ftrace_graph_stub to ftrace_stub_graph")

from the ftrace tree.

I fixed it up (see below) and can carry the fix as necessary. This
is now fixed as far as linux-next is concerned, but any non trivial
conflicts should be mentioned to your upstream maintainer when your tree
is submitted for merging.  You may also want to consider cooperating
with the maintainer of the conflicting tree to minimise any particularly
complex conflicts.

-- 
Cheers,
Stephen Rothwell

diff --cc include/asm-generic/vmlinux.lds.h
index 63cedc3c0c77,996db32c491b..000000000000
--- a/include/asm-generic/vmlinux.lds.h
+++ b/include/asm-generic/vmlinux.lds.h
@@@ -136,19 -110,30 +136,28 @@@
  #endif
  
  #ifdef CONFIG_FTRACE_MCOUNT_RECORD
 -#ifdef CC_USING_PATCHABLE_FUNCTION_ENTRY
  /*
 + * The ftrace call sites are logged to a section whose name depends on the
 + * compiler option used. A given kernel image will only use one, AKA
 + * FTRACE_CALLSITE_SECTION. We capture all of them here to avoid header
 + * dependencies for FTRACE_CALLSITE_SECTION's definition.
++ *
+  * Need to also make ftrace_stub_graph point to ftrace_stub
+  * so that the same stub location may have different protocols
+  * and not mess up with C verifiers.
   */
 -#define MCOUNT_REC()	. = ALIGN(8);				\
 -			__start_mcount_loc = .;			\
 -			KEEP(*(__patchable_function_entries))	\
 -			__stop_mcount_loc = .;			\
 -			ftrace_stub_graph = ftrace_stub;
 -#else
  #define MCOUNT_REC()	. = ALIGN(8);				\
  			__start_mcount_loc = .;			\
  			KEEP(*(__mcount_loc))			\
 +			KEEP(*(__patchable_function_entries))	\
- 			__stop_mcount_loc = .;
+ 			__stop_mcount_loc = .;			\
+ 			ftrace_stub_graph = ftrace_stub;
 -#endif
  #else
- #define MCOUNT_REC()
+ # ifdef CONFIG_FUNCTION_TRACER
+ #  define MCOUNT_REC()	ftrace_stub_graph = ftrace_stub;
+ # else
+ #  define MCOUNT_REC()
+ # endif
  #endif
  
  #ifdef CONFIG_TRACE_BRANCH_PROFILING

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]