linuxppc-dev.lists.ozlabs.org archive mirror
 help / color / mirror / Atom feed
* [PATCH] powerpc: avoid broken GCC __attribute__((optimize))
@ 2020-10-28  8:04 Ard Biesheuvel
  2020-10-28  8:19 ` Test Results: " snowpatch
                   ` (2 more replies)
  0 siblings, 3 replies; 5+ messages in thread
From: Ard Biesheuvel @ 2020-10-28  8:04 UTC (permalink / raw)
  To: linux-kernel
  Cc: Daniel Borkmann, Josh Poimboeuf, Peter Zijlstra, Randy Dunlap,
	Nick Desaulniers, Alexei Starovoitov, Arvind Sankar,
	Paul Mackerras, Geert Uytterhoeven, Thomas Gleixner,
	linuxppc-dev, Ard Biesheuvel, Kees Cook

Commit 7053f80d9696 ("powerpc/64: Prevent stack protection in early boot")
introduced a couple of uses of __attribute__((optimize)) with function
scope, to disable the stack protector in some early boot code.

Unfortunately, and this is documented in the GCC man pages [0], overriding
function attributes for optimization is broken, and is only supported for
debug scenarios, not for production: the problem appears to be that
setting GCC -f flags using this method will cause it to forget about some
or all other optimization settings that have been applied.

So the only safe way to disable the stack protector is to disable it for
the entire source file.

[0] https://gcc.gnu.org/onlinedocs/gcc/Common-Function-Attributes.html

Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Nick Desaulniers <ndesaulniers@google.com>
Cc: Arvind Sankar <nivedita@alum.mit.edu>
Cc: Randy Dunlap <rdunlap@infradead.org>
Cc: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Alexei Starovoitov <ast@kernel.org>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Kees Cook <keescook@chromium.org>
Fixes: 7053f80d9696 ("powerpc/64: Prevent stack protection in early boot")
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
---
Related discussion here:
https://lore.kernel.org/lkml/CAMuHMdUg0WJHEcq6to0-eODpXPOywLot6UD2=GFHpzoj_hCoBQ@mail.gmail.com/

TL;DR using __attribute__((optimize("-fno-gcse"))) in the BPF interpreter
causes the compiler to forget about -fno-asynchronous-unwind-tables passed
on the command line, resulting in unexpected .eh_frame sections in vmlinux.

 arch/powerpc/kernel/Makefile   | 3 +++
 arch/powerpc/kernel/paca.c     | 2 +-
 arch/powerpc/kernel/setup.h    | 6 ------
 arch/powerpc/kernel/setup_64.c | 2 +-
 4 files changed, 5 insertions(+), 8 deletions(-)

diff --git a/arch/powerpc/kernel/Makefile b/arch/powerpc/kernel/Makefile
index bf0bf1b900d2..fe2ef598e2ea 100644
--- a/arch/powerpc/kernel/Makefile
+++ b/arch/powerpc/kernel/Makefile
@@ -173,6 +173,9 @@ KCOV_INSTRUMENT_cputable.o := n
 KCOV_INSTRUMENT_setup_64.o := n
 KCOV_INSTRUMENT_paca.o := n
 
+CFLAGS_setup_64.o		+= -fno-stack-protector
+CFLAGS_paca.o			+= -fno-stack-protector
+
 extra-$(CONFIG_PPC_FPU)		+= fpu.o
 extra-$(CONFIG_ALTIVEC)		+= vector.o
 extra-$(CONFIG_PPC64)		+= entry_64.o
diff --git a/arch/powerpc/kernel/paca.c b/arch/powerpc/kernel/paca.c
index 0ad15768d762..fe70834d7283 100644
--- a/arch/powerpc/kernel/paca.c
+++ b/arch/powerpc/kernel/paca.c
@@ -208,7 +208,7 @@ static struct rtas_args * __init new_rtas_args(int cpu, unsigned long limit)
 struct paca_struct **paca_ptrs __read_mostly;
 EXPORT_SYMBOL(paca_ptrs);
 
-void __init __nostackprotector initialise_paca(struct paca_struct *new_paca, int cpu)
+void __init initialise_paca(struct paca_struct *new_paca, int cpu)
 {
 #ifdef CONFIG_PPC_PSERIES
 	new_paca->lppaca_ptr = NULL;
diff --git a/arch/powerpc/kernel/setup.h b/arch/powerpc/kernel/setup.h
index 2ec835574cc9..2dd0d9cb5a20 100644
--- a/arch/powerpc/kernel/setup.h
+++ b/arch/powerpc/kernel/setup.h
@@ -8,12 +8,6 @@
 #ifndef __ARCH_POWERPC_KERNEL_SETUP_H
 #define __ARCH_POWERPC_KERNEL_SETUP_H
 
-#ifdef CONFIG_CC_IS_CLANG
-#define __nostackprotector
-#else
-#define __nostackprotector __attribute__((__optimize__("no-stack-protector")))
-#endif
-
 void initialize_cache_info(void);
 void irqstack_early_init(void);
 
diff --git a/arch/powerpc/kernel/setup_64.c b/arch/powerpc/kernel/setup_64.c
index bb9cab3641d7..da447a62ea1e 100644
--- a/arch/powerpc/kernel/setup_64.c
+++ b/arch/powerpc/kernel/setup_64.c
@@ -283,7 +283,7 @@ void __init record_spr_defaults(void)
  * device-tree is not accessible via normal means at this point.
  */
 
-void __init __nostackprotector early_setup(unsigned long dt_ptr)
+void __init early_setup(unsigned long dt_ptr)
 {
 	static __initdata struct paca_struct boot_paca;
 
-- 
2.17.1


^ permalink raw reply related	[flat|nested] 5+ messages in thread

* Test Results: RE: powerpc: avoid broken GCC __attribute__((optimize))
  2020-10-28  8:04 [PATCH] powerpc: avoid broken GCC __attribute__((optimize)) Ard Biesheuvel
@ 2020-10-28  8:19 ` snowpatch
  2020-10-28  8:29 ` [PATCH] " Ard Biesheuvel
  2020-11-25 11:57 ` Michael Ellerman
  2 siblings, 0 replies; 5+ messages in thread
From: snowpatch @ 2020-10-28  8:19 UTC (permalink / raw)
  To: Ard Biesheuvel, linuxppc-dev

[-- Attachment #1: Type: text/plain, Size: 2316 bytes --]

Thanks for your contribution, unfortunately we've found some issues.

Your patch was successfully applied on branch powerpc/merge (8cb17737940b156329cb5210669b9c9b23f4dd56)

The test build-ppc64le reported the following: Build failed!

 Full log: https://openpower.xyz/job/snowpatch/job/snowpatch-linux-sparse/21048//artifact/linux/report.txt


Here's a preview of the log:

arch/powerpc/kernel/paca.c:244:25: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'setup_paca'
  244 | void __nostackprotector setup_paca(struct paca_struct *new_paca)
      |                         ^~~~~~~~~~
make[2]: *** [scripts/Makefile.build:283: arch/powerpc/kernel/paca.o] Error 1
make[1]: *** [scripts/Makefile.build:500: arch/powerpc/kernel] Error 2
make[1]: *** Waiting for unfinished jobs....
make: *** [Makefile:1799: arch/powerpc] Error 2
make: *** Waiting for unfinished jobs....


The test build-ppc64be reported the following: Build failed!

 Full log: https://openpower.xyz/job/snowpatch/job/snowpatch-linux-sparse/21049//artifact/linux/report.txt


Here's a preview of the log:

arch/powerpc/kernel/paca.c:244:25: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'setup_paca'
  244 | void __nostackprotector setup_paca(struct paca_struct *new_paca)
      |                         ^~~~~~~~~~
make[2]: *** [scripts/Makefile.build:283: arch/powerpc/kernel/paca.o] Error 1
make[1]: *** [scripts/Makefile.build:500: arch/powerpc/kernel] Error 2
make[1]: *** Waiting for unfinished jobs....
make: *** [Makefile:1799: arch/powerpc] Error 2
make: *** Waiting for unfinished jobs....


The test build-ppc64e reported the following: Build failed!

 Full log: https://openpower.xyz/job/snowpatch/job/snowpatch-linux-sparse/21050//artifact/linux/report.txt


Here's a preview of the log:

arch/powerpc/kernel/paca.c:244:25: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'setup_paca'
  244 | void __nostackprotector setup_paca(struct paca_struct *new_paca)
      |                         ^~~~~~~~~~
make[2]: *** [scripts/Makefile.build:283: arch/powerpc/kernel/paca.o] Error 1
make[1]: *** [scripts/Makefile.build:500: arch/powerpc/kernel] Error 2
make[1]: *** Waiting for unfinished jobs....
make: *** [Makefile:1799: arch/powerpc] Error 2
make: *** Waiting for unfinished jobs....




^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH] powerpc: avoid broken GCC __attribute__((optimize))
  2020-10-28  8:04 [PATCH] powerpc: avoid broken GCC __attribute__((optimize)) Ard Biesheuvel
  2020-10-28  8:19 ` Test Results: " snowpatch
@ 2020-10-28  8:29 ` Ard Biesheuvel
  2020-10-29 11:35   ` Michael Ellerman
  2020-11-25 11:57 ` Michael Ellerman
  2 siblings, 1 reply; 5+ messages in thread
From: Ard Biesheuvel @ 2020-10-28  8:29 UTC (permalink / raw)
  To: Linux Kernel Mailing List
  Cc: Daniel Borkmann, Josh Poimboeuf, Peter Zijlstra, Randy Dunlap,
	Nick Desaulniers, Alexei Starovoitov, Arvind Sankar,
	Paul Mackerras, Geert Uytterhoeven, Thomas Gleixner,
	open list:LINUX FOR POWERPC (32-BIT AND 64-BIT),
	Kees Cook

On Wed, 28 Oct 2020 at 09:04, Ard Biesheuvel <ardb@kernel.org> wrote:
>
> Commit 7053f80d9696 ("powerpc/64: Prevent stack protection in early boot")
> introduced a couple of uses of __attribute__((optimize)) with function
> scope, to disable the stack protector in some early boot code.
>
> Unfortunately, and this is documented in the GCC man pages [0], overriding
> function attributes for optimization is broken, and is only supported for
> debug scenarios, not for production: the problem appears to be that
> setting GCC -f flags using this method will cause it to forget about some
> or all other optimization settings that have been applied.
>
> So the only safe way to disable the stack protector is to disable it for
> the entire source file.
>
> [0] https://gcc.gnu.org/onlinedocs/gcc/Common-Function-Attributes.html
>
> Cc: Michael Ellerman <mpe@ellerman.id.au>
> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
> Cc: Paul Mackerras <paulus@samba.org>
> Cc: Nick Desaulniers <ndesaulniers@google.com>
> Cc: Arvind Sankar <nivedita@alum.mit.edu>
> Cc: Randy Dunlap <rdunlap@infradead.org>
> Cc: Josh Poimboeuf <jpoimboe@redhat.com>
> Cc: Thomas Gleixner <tglx@linutronix.de>
> Cc: Alexei Starovoitov <ast@kernel.org>
> Cc: Daniel Borkmann <daniel@iogearbox.net>
> Cc: Peter Zijlstra (Intel) <peterz@infradead.org>
> Cc: Geert Uytterhoeven <geert@linux-m68k.org>
> Cc: Kees Cook <keescook@chromium.org>
> Fixes: 7053f80d9696 ("powerpc/64: Prevent stack protection in early boot")
> Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
> ---
> Related discussion here:
> https://lore.kernel.org/lkml/CAMuHMdUg0WJHEcq6to0-eODpXPOywLot6UD2=GFHpzoj_hCoBQ@mail.gmail.com/
>
> TL;DR using __attribute__((optimize("-fno-gcse"))) in the BPF interpreter
> causes the compiler to forget about -fno-asynchronous-unwind-tables passed
> on the command line, resulting in unexpected .eh_frame sections in vmlinux.
>
>  arch/powerpc/kernel/Makefile   | 3 +++
>  arch/powerpc/kernel/paca.c     | 2 +-
>  arch/powerpc/kernel/setup.h    | 6 ------
>  arch/powerpc/kernel/setup_64.c | 2 +-
>  4 files changed, 5 insertions(+), 8 deletions(-)
>

FYI i was notified by one of the robots that I missed one occurrence
of __nostackprotector in arch/powerpc/kernel/paca.c

Let me know if I need to resend.


> diff --git a/arch/powerpc/kernel/Makefile b/arch/powerpc/kernel/Makefile
> index bf0bf1b900d2..fe2ef598e2ea 100644
> --- a/arch/powerpc/kernel/Makefile
> +++ b/arch/powerpc/kernel/Makefile
> @@ -173,6 +173,9 @@ KCOV_INSTRUMENT_cputable.o := n
>  KCOV_INSTRUMENT_setup_64.o := n
>  KCOV_INSTRUMENT_paca.o := n
>
> +CFLAGS_setup_64.o              += -fno-stack-protector
> +CFLAGS_paca.o                  += -fno-stack-protector
> +
>  extra-$(CONFIG_PPC_FPU)                += fpu.o
>  extra-$(CONFIG_ALTIVEC)                += vector.o
>  extra-$(CONFIG_PPC64)          += entry_64.o
> diff --git a/arch/powerpc/kernel/paca.c b/arch/powerpc/kernel/paca.c
> index 0ad15768d762..fe70834d7283 100644
> --- a/arch/powerpc/kernel/paca.c
> +++ b/arch/powerpc/kernel/paca.c
> @@ -208,7 +208,7 @@ static struct rtas_args * __init new_rtas_args(int cpu, unsigned long limit)
>  struct paca_struct **paca_ptrs __read_mostly;
>  EXPORT_SYMBOL(paca_ptrs);
>
> -void __init __nostackprotector initialise_paca(struct paca_struct *new_paca, int cpu)
> +void __init initialise_paca(struct paca_struct *new_paca, int cpu)
>  {
>  #ifdef CONFIG_PPC_PSERIES
>         new_paca->lppaca_ptr = NULL;
> diff --git a/arch/powerpc/kernel/setup.h b/arch/powerpc/kernel/setup.h
> index 2ec835574cc9..2dd0d9cb5a20 100644
> --- a/arch/powerpc/kernel/setup.h
> +++ b/arch/powerpc/kernel/setup.h
> @@ -8,12 +8,6 @@
>  #ifndef __ARCH_POWERPC_KERNEL_SETUP_H
>  #define __ARCH_POWERPC_KERNEL_SETUP_H
>
> -#ifdef CONFIG_CC_IS_CLANG
> -#define __nostackprotector
> -#else
> -#define __nostackprotector __attribute__((__optimize__("no-stack-protector")))
> -#endif
> -
>  void initialize_cache_info(void);
>  void irqstack_early_init(void);
>
> diff --git a/arch/powerpc/kernel/setup_64.c b/arch/powerpc/kernel/setup_64.c
> index bb9cab3641d7..da447a62ea1e 100644
> --- a/arch/powerpc/kernel/setup_64.c
> +++ b/arch/powerpc/kernel/setup_64.c
> @@ -283,7 +283,7 @@ void __init record_spr_defaults(void)
>   * device-tree is not accessible via normal means at this point.
>   */
>
> -void __init __nostackprotector early_setup(unsigned long dt_ptr)
> +void __init early_setup(unsigned long dt_ptr)
>  {
>         static __initdata struct paca_struct boot_paca;
>
> --
> 2.17.1
>

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH] powerpc: avoid broken GCC __attribute__((optimize))
  2020-10-28  8:29 ` [PATCH] " Ard Biesheuvel
@ 2020-10-29 11:35   ` Michael Ellerman
  0 siblings, 0 replies; 5+ messages in thread
From: Michael Ellerman @ 2020-10-29 11:35 UTC (permalink / raw)
  To: Ard Biesheuvel, Linux Kernel Mailing List
  Cc: Kees Cook, Daniel Borkmann, Peter Zijlstra, Randy Dunlap,
	Nick Desaulniers, Alexei Starovoitov, Arvind Sankar,
	Paul Mackerras, Josh Poimboeuf, Geert Uytterhoeven,
	Thomas Gleixner, open list:LINUX FOR POWERPC (32-BIT AND 64-BIT)

Ard Biesheuvel <ardb@kernel.org> writes:
> On Wed, 28 Oct 2020 at 09:04, Ard Biesheuvel <ardb@kernel.org> wrote:
>>
>> Commit 7053f80d9696 ("powerpc/64: Prevent stack protection in early boot")
>> introduced a couple of uses of __attribute__((optimize)) with function
>> scope, to disable the stack protector in some early boot code.
>>
>> Unfortunately, and this is documented in the GCC man pages [0], overriding
>> function attributes for optimization is broken, and is only supported for
>> debug scenarios, not for production: the problem appears to be that
>> setting GCC -f flags using this method will cause it to forget about some
>> or all other optimization settings that have been applied.
>>
>> So the only safe way to disable the stack protector is to disable it for
>> the entire source file.
>>
>> [0] https://gcc.gnu.org/onlinedocs/gcc/Common-Function-Attributes.html
>>
>> Cc: Michael Ellerman <mpe@ellerman.id.au>
>> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
>> Cc: Paul Mackerras <paulus@samba.org>
>> Cc: Nick Desaulniers <ndesaulniers@google.com>
>> Cc: Arvind Sankar <nivedita@alum.mit.edu>
>> Cc: Randy Dunlap <rdunlap@infradead.org>
>> Cc: Josh Poimboeuf <jpoimboe@redhat.com>
>> Cc: Thomas Gleixner <tglx@linutronix.de>
>> Cc: Alexei Starovoitov <ast@kernel.org>
>> Cc: Daniel Borkmann <daniel@iogearbox.net>
>> Cc: Peter Zijlstra (Intel) <peterz@infradead.org>
>> Cc: Geert Uytterhoeven <geert@linux-m68k.org>
>> Cc: Kees Cook <keescook@chromium.org>
>> Fixes: 7053f80d9696 ("powerpc/64: Prevent stack protection in early boot")
>> Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
>> ---
>> Related discussion here:
>> https://lore.kernel.org/lkml/CAMuHMdUg0WJHEcq6to0-eODpXPOywLot6UD2=GFHpzoj_hCoBQ@mail.gmail.com/
>>
>> TL;DR using __attribute__((optimize("-fno-gcse"))) in the BPF interpreter
>> causes the compiler to forget about -fno-asynchronous-unwind-tables passed
>> on the command line, resulting in unexpected .eh_frame sections in vmlinux.
>>
>>  arch/powerpc/kernel/Makefile   | 3 +++
>>  arch/powerpc/kernel/paca.c     | 2 +-
>>  arch/powerpc/kernel/setup.h    | 6 ------
>>  arch/powerpc/kernel/setup_64.c | 2 +-
>>  4 files changed, 5 insertions(+), 8 deletions(-)

Thanks for the patch.

> FYI i was notified by one of the robots that I missed one occurrence
> of __nostackprotector in arch/powerpc/kernel/paca.c
>
> Let me know if I need to resend.

That's fine I'll fix it up when applying.

With the existing code, with STACKPROTECTOR_STRONG=y, I see two
functions in setup_64.c that are triggering stack protection. One is
__init, and the other takes no parameters and is not easily reachable
from userspace, so I don't think losing the stack canary on either of
those is a concern.

I don't see anything in paca.c triggering stack protection.

I don't think there's any evidence this is causing a bug for us, so I'll
plan to put this in next for v5.11.

cheers

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH] powerpc: avoid broken GCC __attribute__((optimize))
  2020-10-28  8:04 [PATCH] powerpc: avoid broken GCC __attribute__((optimize)) Ard Biesheuvel
  2020-10-28  8:19 ` Test Results: " snowpatch
  2020-10-28  8:29 ` [PATCH] " Ard Biesheuvel
@ 2020-11-25 11:57 ` Michael Ellerman
  2 siblings, 0 replies; 5+ messages in thread
From: Michael Ellerman @ 2020-11-25 11:57 UTC (permalink / raw)
  To: linux-kernel, Ard Biesheuvel
  Cc: Kees Cook, Daniel Borkmann, Peter Zijlstra, Randy Dunlap,
	Nick Desaulniers, Alexei Starovoitov, Arvind Sankar,
	Paul Mackerras, Josh Poimboeuf, Geert Uytterhoeven,
	Thomas Gleixner, linuxppc-dev

On Wed, 28 Oct 2020 09:04:33 +0100, Ard Biesheuvel wrote:
> Commit 7053f80d9696 ("powerpc/64: Prevent stack protection in early boot")
> introduced a couple of uses of __attribute__((optimize)) with function
> scope, to disable the stack protector in some early boot code.
> 
> Unfortunately, and this is documented in the GCC man pages [0], overriding
> function attributes for optimization is broken, and is only supported for
> debug scenarios, not for production: the problem appears to be that
> setting GCC -f flags using this method will cause it to forget about some
> or all other optimization settings that have been applied.
> 
> [...]

Applied to powerpc/next.

[1/1] powerpc: Avoid broken GCC __attribute__((optimize))
      https://git.kernel.org/powerpc/c/a7223f5bfcaeade4a86d35263493bcda6c940891

cheers

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2020-11-25 12:14 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-10-28  8:04 [PATCH] powerpc: avoid broken GCC __attribute__((optimize)) Ard Biesheuvel
2020-10-28  8:19 ` Test Results: " snowpatch
2020-10-28  8:29 ` [PATCH] " Ard Biesheuvel
2020-10-29 11:35   ` Michael Ellerman
2020-11-25 11:57 ` Michael Ellerman

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).