objtool: Add fortify_panic as __noreturn function

PATCH In review Kees Cook 4 participants 7 messages Jun 15, 2017 objtool

lore

Share & tools

lore

Kees Cook

author

2017-06-15 02:18 UTC #0 lore raw

PATCH 1/1 objtool: Add fortify_panic as __noreturn function

CONFIG_FORTIFY_SOURCE implements fortify_panic() as a __noreturn function, so objtool needs to know about it too.

Signed-off-by Kees Cook <[email protected]>

Cc Josh Poimboeuf <[email protected]>

---
 tools/objtool/builtin-check.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

tools/objtool/builtin-check.c +2 -1

index 282a60368b14..5f66697fe1e0 100644
@@ -192,7 +192,8 @@ static int __dead_end_function(struct objtool_file *file, struct symbol *func,
 		"complete_and_exit",
 		"kvm_spurious_fault",
 		"__reiserfs_panic",
-		"lbug_with_loc"
+		"lbug_with_loc",
+		"fortify_panic",
 	};
 
 	if (func->bind == STB_WEAK)

Signature · 6 lines

2.7.4


-- 
Kees Cook
Pixel Security

Stephen Rothwell

2017-06-15 04:06 UTC #1 lore raw

Hi Kees,

Quoted 27 lines · On Wed, 14 Jun 2017 19:18:51 -0700 Kees Cook <[email protected]> wrote:

CONFIG_FORTIFY_SOURCE implements fortify_panic() as a __noreturn function, so objtool needs to know about it too.
Suggested-by Daniel Micay <[email protected]>
Signed-off-by Kees Cook <[email protected]>
Cc Josh Poimboeuf <[email protected]>
---
 tools/objtool/builtin-check.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
tools/objtool/builtin-check.c +2 -1
index 282a60368b14..5f66697fe1e0 100644
@@ -192,7 +192,8 @@ static int __dead_end_function(struct objtool_file *file, struct symbol *func,
 		"complete_and_exit",
 		"kvm_spurious_fault",
 		"__reiserfs_panic",
-		"lbug_with_loc"
+		"lbug_with_loc",
+		"fortify_panic",
 	};
 
 	if (func->bind == STB_WEAK)
Signature · 1 line
2.7.4

OK, I have added this to the end of linux-next today (in the part of Andrew's tree afte the linux-next merge). I also move the fortify string patch to after it (but I assume I will still need to revert it).

I willl drop this patch when it appears elsewhere in linux-next.

Signature · 2 lines

Cheers,
Stephen Rothwell

Stephen Rothwell

2017-06-15 04:26 UTC #2 lore raw

Hi all,

Quoted 26 lines · On Thu, 15 Jun 2017 14:05:22 +1000 Stephen Rothwell <[email protected]>…

Quoted 24 lines · On Wed, 14 Jun 2017 19:18:51 -0700 Kees Cook <[email protected]> wrote:
CONFIG_FORTIFY_SOURCE implements fortify_panic() as a __noreturn function, so objtool needs to know about it too.
Suggested-by Daniel Micay <[email protected]>
Signed-off-by Kees Cook <[email protected]>
Cc Josh Poimboeuf <[email protected]>
---
 tools/objtool/builtin-check.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
tools/objtool/builtin-check.c +2 -1
index 282a60368b14..5f66697fe1e0 100644
@@ -192,7 +192,8 @@ static int __dead_end_function(struct objtool_file *file, struct symbol *func,
 		"complete_and_exit",
 		"kvm_spurious_fault",
 		"__reiserfs_panic",
-		"lbug_with_loc"
+		"lbug_with_loc",
+		"fortify_panic",
 	};
 

Quoted 7 lines

Quoted 3 lines
tools/objtool/builtin-check.c +0 -0
 	if (func->bind == STB_WEAK)
Signature · 1 line
2.7.4  
OK, I have added this to the end of linux-next today (in the part of Andrew's tree afte the linux-next merge). I also move the fortify string patch to after it (but I assume I will still need to revert it).

So after that the errors (x86_64 allmodconfig build) are only:

In file included from include/linux/bitmap.h:8:0,

                 from include/linux/cpumask.h:11,
                 from arch/x86/include/asm/cpumask.h:4,
                 from arch/x86/include/asm/msr.h:10,
                 from arch/x86/include/asm/processor.h:20,
                 from arch/x86/include/asm/cpufeature.h:4,
                 from arch/x86/include/asm/thread_info.h:52,
                 from include/linux/thread_info.h:37,
                 from arch/x86/include/asm/preempt.h:6,
                 from include/linux/preempt.h:80,
                 from include/linux/spinlock.h:50,
                 from include/linux/mmzone.h:7,
                 from include/linux/gfp.h:5,
                 from arch/x86/power/hibernate_64.c:11:
In function 'memcpy',
    inlined from 'relocate_restore_code' at arch/x86/power/hibernate_64.c:150:2,
    inlined from 'swsusp_arch_resume' at arch/x86/power/hibernate_64.c:186:8:
include/linux/string.h:309:4: error: call to '__read_overflow2' declared with attribute error: detected read beyond size of object passed as 2nd parameter
    __read_overflow2();
    ^
In file included from include/linux/bitmap.h:8:0,
                 from include/linux/cpumask.h:11,
                 from arch/x86/include/asm/cpumask.h:4,
                 from arch/x86/include/asm/msr.h:10,
                 from arch/x86/include/asm/processor.h:20,
                 from arch/x86/include/asm/cpufeature.h:4,
                 from arch/x86/include/asm/thread_info.h:52,
                 from include/linux/thread_info.h:37,
                 from arch/x86/include/asm/preempt.h:6,
                 from include/linux/preempt.h:80,
                 from include/linux/spinlock.h:50,
                 from include/linux/mmzone.h:7,
                 from include/linux/gfp.h:5,
                 from include/linux/mm.h:9,
                 from kernel/kexec_file.c:15:
In function 'memcmp',
    inlined from 'kexec_load_purgatory' at kernel/kexec_file.c:900:6:
include/linux/string.h:348:4: error: call to '__read_overflow' declared with attribute error: detected read beyond size of object passed as 1st parameter
    __read_overflow();
    ^

Signature · 2 lines

Cheers,
Stephen Rothwell

Daniel Micay

2017-06-15 04:49 UTC #3 lore raw

Quoted 41 lines

So after that the errors (x86_64 allmodconfig build) are only:

In file included from include/linux/bitmap.h:8:0,

                 from include/linux/cpumask.h:11,
                 from arch/x86/include/asm/cpumask.h:4,
                 from arch/x86/include/asm/msr.h:10,
                 from arch/x86/include/asm/processor.h:20,
                 from arch/x86/include/asm/cpufeature.h:4,
                 from arch/x86/include/asm/thread_info.h:52,
                 from include/linux/thread_info.h:37,
                 from arch/x86/include/asm/preempt.h:6,
                 from include/linux/preempt.h:80,
                 from include/linux/spinlock.h:50,
                 from include/linux/mmzone.h:7,
                 from include/linux/gfp.h:5,
                 from arch/x86/power/hibernate_64.c:11:

In function 'memcpy',
inlined from 'relocate_restore_code' at
arch/x86/power/hibernate_64.c:150:2,
inlined from 'swsusp_arch_resume' at
arch/x86/power/hibernate_64.c:186:8:
include/linux/string.h:309:4: error: call to '__read_overflow2' declared with attribute error: detected read beyond size of object passed as 2nd parameter

    __read_overflow2();
    ^
In file included from include/linux/bitmap.h:8:0,
                 from include/linux/cpumask.h:11,
                 from arch/x86/include/asm/cpumask.h:4,
                 from arch/x86/include/asm/msr.h:10,
                 from arch/x86/include/asm/processor.h:20,
                 from arch/x86/include/asm/cpufeature.h:4,
                 from arch/x86/include/asm/thread_info.h:52,
                 from include/linux/thread_info.h:37,
                 from arch/x86/include/asm/preempt.h:6,
                 from include/linux/preempt.h:80,
                 from include/linux/spinlock.h:50,
                 from include/linux/mmzone.h:7,
                 from include/linux/gfp.h:5,
                 from include/linux/mm.h:9,
                 from kernel/kexec_file.c:15:

Quoted 8 lines

In function 'memcmp',
inlined from 'kexec_load_purgatory' at kernel/kexec_file.c:900:6:
include/linux/string.h:348:4: error: call to '__read_overflow' declared with attribute error: detected read beyond size of object passed as 1st parameter
    __read_overflow();
    ^

Kees has the remaining ones here now:

https://git.kernel.org/pub/scm/linux/kernel/git/kees/linux.git/log/?h=for-next/kspp

Not entirely sure what happened to the powerpc bits though.

Stephen Rothwell

2017-06-15 05:07 UTC #4 lore raw

Hi Daniel,

Quoted 51 lines · On Thu, 15 Jun 2017 00:49:21 -0400 Daniel Micay <[email protected]> wrote:

Quoted 42 lines
So after that the errors (x86_64 allmodconfig build) are only:
In file included from include/linux/bitmap.h:8:0,
                 from include/linux/cpumask.h:11,
                 from arch/x86/include/asm/cpumask.h:4,
                 from arch/x86/include/asm/msr.h:10,
                 from arch/x86/include/asm/processor.h:20,
                 from arch/x86/include/asm/cpufeature.h:4,
                 from arch/x86/include/asm/thread_info.h:52,
                 from include/linux/thread_info.h:37,
                 from arch/x86/include/asm/preempt.h:6,
                 from include/linux/preempt.h:80,
                 from include/linux/spinlock.h:50,
                 from include/linux/mmzone.h:7,
                 from include/linux/gfp.h:5,
                 from arch/x86/power/hibernate_64.c:11:
In function 'memcpy',
inlined from 'relocate_restore_code' at
arch/x86/power/hibernate_64.c:150:2,
inlined from 'swsusp_arch_resume' at
arch/x86/power/hibernate_64.c:186:8:
include/linux/string.h:309:4: error: call to '__read_overflow2' declared with attribute error: detected read beyond size of object passed as 2nd parameter
    __read_overflow2();
    ^
In file included from include/linux/bitmap.h:8:0,
                 from include/linux/cpumask.h:11,
                 from arch/x86/include/asm/cpumask.h:4,
                 from arch/x86/include/asm/msr.h:10,
                 from arch/x86/include/asm/processor.h:20,
                 from arch/x86/include/asm/cpufeature.h:4,
                 from arch/x86/include/asm/thread_info.h:52,
                 from include/linux/thread_info.h:37,
                 from arch/x86/include/asm/preempt.h:6,
                 from include/linux/preempt.h:80,
                 from include/linux/spinlock.h:50,
                 from include/linux/mmzone.h:7,
                 from include/linux/gfp.h:5,
                 from include/linux/mm.h:9,
                 from kernel/kexec_file.c:15:
Quoted 8 lines
In function 'memcmp',
inlined from 'kexec_load_purgatory' at kernel/kexec_file.c:900:6:
include/linux/string.h:348:4: error: call to '__read_overflow' declared with attribute error: detected read beyond size of object passed as 1st parameter
    __read_overflow();
    ^

Quoted 3 lines

Kees has the remaining ones here now:
https://git.kernel.org/pub/scm/linux/kernel/git/kees/linux.git/log/?h=for-next/kspp

Excellent, so hopefully tomorrow I won't need to revert (at which point the overnight builds (http://kisskb.ellerman.id.au/linux-next) will point out any more problems).

Quoted 1 line

Not entirely sure what happened to the powerpc bits though.

Maybe in the powerpc tree
(git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux.git#next).

Signature · 2 lines

Cheers,
Stephen Rothwell

Stephen Rothwell

Tested-by on v1 1/1

2017-06-15 05:05 UTC #5 lore raw

In reply to Kees Cook

Hi Kees,

Quoted 7 lines · On Wed, 14 Jun 2017 19:18:51 -0700 Kees Cook <[email protected]> wrote:

CONFIG_FORTIFY_SOURCE implements fortify_panic() as a __noreturn function, so objtool needs to know about it too.
Suggested-by Daniel Micay <[email protected]>
Signed-off-by Kees Cook <[email protected]>
Cc Josh Poimboeuf <[email protected]>

Tested-by Stephen Rothwell <[email protected]>

Signature · 2 lines

Cheers,
Stephen Rothwell

Josh Poimboeuf

2017-06-15 13:14 UTC #6 lore raw

In reply to Kees Cook

Quoted 15 lines · Kees Cook

CONFIG_FORTIFY_SOURCE implements fortify_panic() as a __noreturn function, so objtool needs to know about it too.
Suggested-by Daniel Micay <[email protected]>
Signed-off-by Kees Cook <[email protected]>
Cc Josh Poimboeuf <[email protected]>
---
 tools/objtool/builtin-check.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
tools/objtool/builtin-check.c +0 -0
index 282a60368b14..5f66697fe1e0 100644

tools/objtool/builtin-check.c static int __dead_end_function(struct objtool_file *file, struct symbol *func, · line 199

Quoted 9 lines

tools/objtool/builtin-check.c +2 -1

@@ -192,7 +192,8 @@ static int __dead_end_function(struct objtool_file *file, struct symbol *func,
 		"complete_and_exit",
 		"kvm_spurious_fault",
 		"__reiserfs_panic",
-		"lbug_with_loc"
+		"lbug_with_loc",
+		"fortify_panic",
 	};
 
 	if (func->bind == STB_WEAK)

Thanks, I'll take the patch.

Eventually I'd like to experiment with a GCC plugin to automatically detect these noreturn cases instead of the hard-coded list.

Signature · 1 line

Josh