[PATCH AUTOSEL 6.7 07/26] arm64: jump_label: use constraints "Si" instead of "i"

[PATCH AUTOSEL 6.7 07/26] arm64: jump_label: use constraints "Si" instead of "i"

[Sasha Levin]

From: Fangrui Song <maskray@google.com>

[ Upstream commit f9daab0ad01cf9d165dbbbf106ca4e61d06e7fe8 ]

The generic constraint "i" seems to be copied from x86 or arm (and with
a redundant generic operand modifier "c"). It works with -fno-PIE but
not with -fPIE/-fPIC in GCC's aarch64 port.

The machine constraint "S", which denotes a symbol or label reference
with a constant offset, supports PIC and has been available in GCC since
2012 and in Clang since 7.0. However, Clang before 19 does not support
"S" on a symbol with a constant offset [1] (e.g.
`static_key_false(&nf_hooks_needed[pf][hook])` in
include/linux/netfilter.h), so we use "i" as a fallback.

Suggested-by: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Fangrui Song <maskray@google.com>
Link: https://github.com/llvm/llvm-project/pull/80255 [1]
Acked-by: Mark Rutland <mark.rutland@arm.com>
Link: https://lore.kernel.org/r/20240206074552.541154-1-maskray@google.com
Signed-off-by: Will Deacon <will@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 arch/arm64/include/asm/jump_label.h | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

diff --git a/arch/arm64/include/asm/jump_label.h b/arch/arm64/include/asm/jump_label.h
index 48ddc0f45d228..b7716b215f91a 100644
--- a/arch/arm64/include/asm/jump_label.h
+++ b/arch/arm64/include/asm/jump_label.h
@@ -15,6 +15,10 @@
 
 #define JUMP_LABEL_NOP_SIZE		AARCH64_INSN_SIZE
 
+/*
+ * Prefer the constraint "S" to support PIC with GCC. Clang before 19 does not
+ * support "S" on a symbol with a constant offset, so we use "i" as a fallback.
+ */
 static __always_inline bool arch_static_branch(struct static_key * const key,
 					       const bool branch)
 {
@@ -23,9 +27,9 @@ static __always_inline bool arch_static_branch(struct static_key * const key,
 		 "	.pushsection	__jump_table, \"aw\"	\n\t"
 		 "	.align		3			\n\t"
 		 "	.long		1b - ., %l[l_yes] - .	\n\t"
-		 "	.quad		%c0 - .			\n\t"
+		 "	.quad		(%[key] - .) + %[bit0]  \n\t"
 		 "	.popsection				\n\t"
-		 :  :  "i"(&((char *)key)[branch]) :  : l_yes);
+		 :  :  [key]"Si"(key), [bit0]"i"(branch) :  : l_yes);
 
 	return false;
 l_yes:
@@ -40,9 +44,9 @@ static __always_inline bool arch_static_branch_jump(struct static_key * const ke
 		 "	.pushsection	__jump_table, \"aw\"	\n\t"
 		 "	.align		3			\n\t"
 		 "	.long		1b - ., %l[l_yes] - .	\n\t"
-		 "	.quad		%c0 - .			\n\t"
+		 "	.quad		(%[key] - .) + %[bit0]  \n\t"
 		 "	.popsection				\n\t"
-		 :  :  "i"(&((char *)key)[branch]) :  : l_yes);
+		 :  :  [key]"Si"(key), [bit0]"i"(branch) :  : l_yes);
 
 	return false;
 l_yes:
-- 
2.43.0

[PATCH AUTOSEL 6.7 25/26] gen_compile_commands: fix invalid escape sequence warning

[Sasha Levin]

From: Andrew Ballance <andrewjballance@gmail.com>

[ Upstream commit dae4a0171e25884787da32823b3081b4c2acebb2 ]

With python 3.12, '\#' results in this warning
    SyntaxWarning: invalid escape sequence '\#'

Signed-off-by: Andrew Ballance <andrewjballance@gmail.com>
Reviewed-by: Justin Stitt <justinstitt@google.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 scripts/clang-tools/gen_compile_commands.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/scripts/clang-tools/gen_compile_commands.py b/scripts/clang-tools/gen_compile_commands.py
index 5dea4479240bc..e4fb686dfaa9f 100755
--- a/scripts/clang-tools/gen_compile_commands.py
+++ b/scripts/clang-tools/gen_compile_commands.py
@@ -170,7 +170,7 @@ def process_line(root_directory, command_prefix, file_path):
     # escape the pound sign '#', either as '\#' or '$(pound)' (depending on the
     # kernel version). The compile_commands.json file is not interepreted
     # by Make, so this code replaces the escaped version with '#'.
-    prefix = command_prefix.replace('\#', '#').replace('$(pound)', '#')
+    prefix = command_prefix.replace(r'\#', '#').replace('$(pound)', '#')
 
     # Return the canonical path, eliminating any symbolic links encountered in the path.
     abs_path = os.path.realpath(os.path.join(root_directory, file_path))
-- 
2.43.0

Re: [PATCH AUTOSEL 6.7 07/26] arm64: jump_label: use constraints "Si" instead of "i"

[Mark Rutland]

On Thu, Feb 29, 2024 at 10:48:26AM -0500, Sasha Levin wrote:
> From: Fangrui Song <maskray@google.com>
> 
> [ Upstream commit f9daab0ad01cf9d165dbbbf106ca4e61d06e7fe8 ]

That upstream patch caused issues and was subsequently reverted in commit:

  a6b3eb304a82c29665a0ab947cfe276f6d29f523
  ("Revert "arm64: jump_label: use constraints "Si" instead of "i""")

Please drop this; it'll break stable in the same way, and it wasn't necessary
to backport in the first place.

As an aside, that revert has been upstream since v6.8-rc4, ~2.5 weeks ago; it's
unfortunate for the broken patch to be AUTOSEL'd now! It'd be good if something
could automatically check for reverts in mainline...

Mark.

> The generic constraint "i" seems to be copied from x86 or arm (and with
> a redundant generic operand modifier "c"). It works with -fno-PIE but
> not with -fPIE/-fPIC in GCC's aarch64 port.
> 
> The machine constraint "S", which denotes a symbol or label reference
> with a constant offset, supports PIC and has been available in GCC since
> 2012 and in Clang since 7.0. However, Clang before 19 does not support
> "S" on a symbol with a constant offset [1] (e.g.
> `static_key_false(&nf_hooks_needed[pf][hook])` in
> include/linux/netfilter.h), so we use "i" as a fallback.
> 
> Suggested-by: Ard Biesheuvel <ardb@kernel.org>
> Signed-off-by: Fangrui Song <maskray@google.com>
> Link: https://github.com/llvm/llvm-project/pull/80255 [1]
> Acked-by: Mark Rutland <mark.rutland@arm.com>
> Link: https://lore.kernel.org/r/20240206074552.541154-1-maskray@google.com
> Signed-off-by: Will Deacon <will@kernel.org>
> Signed-off-by: Sasha Levin <sashal@kernel.org>
> ---
>  arch/arm64/include/asm/jump_label.h | 12 ++++++++----
>  1 file changed, 8 insertions(+), 4 deletions(-)
> 
> diff --git a/arch/arm64/include/asm/jump_label.h b/arch/arm64/include/asm/jump_label.h
> index 48ddc0f45d228..b7716b215f91a 100644
> --- a/arch/arm64/include/asm/jump_label.h
> +++ b/arch/arm64/include/asm/jump_label.h
> @@ -15,6 +15,10 @@
>  
>  #define JUMP_LABEL_NOP_SIZE		AARCH64_INSN_SIZE
>  
> +/*
> + * Prefer the constraint "S" to support PIC with GCC. Clang before 19 does not
> + * support "S" on a symbol with a constant offset, so we use "i" as a fallback.
> + */
>  static __always_inline bool arch_static_branch(struct static_key * const key,
>  					       const bool branch)
>  {
> @@ -23,9 +27,9 @@ static __always_inline bool arch_static_branch(struct static_key * const key,
>  		 "	.pushsection	__jump_table, \"aw\"	\n\t"
>  		 "	.align		3			\n\t"
>  		 "	.long		1b - ., %l[l_yes] - .	\n\t"
> -		 "	.quad		%c0 - .			\n\t"
> +		 "	.quad		(%[key] - .) + %[bit0]  \n\t"
>  		 "	.popsection				\n\t"
> -		 :  :  "i"(&((char *)key)[branch]) :  : l_yes);
> +		 :  :  [key]"Si"(key), [bit0]"i"(branch) :  : l_yes);
>  
>  	return false;
>  l_yes:
> @@ -40,9 +44,9 @@ static __always_inline bool arch_static_branch_jump(struct static_key * const ke
>  		 "	.pushsection	__jump_table, \"aw\"	\n\t"
>  		 "	.align		3			\n\t"
>  		 "	.long		1b - ., %l[l_yes] - .	\n\t"
> -		 "	.quad		%c0 - .			\n\t"
> +		 "	.quad		(%[key] - .) + %[bit0]  \n\t"
>  		 "	.popsection				\n\t"
> -		 :  :  "i"(&((char *)key)[branch]) :  : l_yes);
> +		 :  :  [key]"Si"(key), [bit0]"i"(branch) :  : l_yes);
>  
>  	return false;
>  l_yes:
> -- 
> 2.43.0
> 

Re: [PATCH AUTOSEL 6.7 07/26] arm64: jump_label: use constraints "Si" instead of "i"

[Sasha Levin]

On Thu, Feb 29, 2024 at 03:56:40PM +0000, Mark Rutland wrote:
>On Thu, Feb 29, 2024 at 10:48:26AM -0500, Sasha Levin wrote:
>> From: Fangrui Song <maskray@google.com>
>>
>> [ Upstream commit f9daab0ad01cf9d165dbbbf106ca4e61d06e7fe8 ]
>
>That upstream patch caused issues and was subsequently reverted in commit:
>
>  a6b3eb304a82c29665a0ab947cfe276f6d29f523
>  ("Revert "arm64: jump_label: use constraints "Si" instead of "i""")
>
>Please drop this; it'll break stable in the same way, and it wasn't necessary
>to backport in the first place.

Will do.

>As an aside, that revert has been upstream since v6.8-rc4, ~2.5 weeks ago; it's
>unfortunate for the broken patch to be AUTOSEL'd now! It'd be good if something
>could automatically check for reverts in mainline...

Sorry for the noise - usually the fixes/revert check happens before
commits actually end up in the queue.

-- 
Thanks,
Sasha