[PATCH 1/2] scripts/recordmcount.pl: support data in text section on powerpc

Subsystems: the rest

STALE3860d

6 messages, 2 authors, 2016-01-13 · open the first message on its own page

[PATCH 1/2] scripts/recordmcount.pl: support data in text section on powerpc

From: Michael Ellerman <mpe@ellerman.id.au>
Date: 2016-01-12 12:14:39

From: Ulrich Weigand <redacted>

If a text section starts out with a data blob before the first
function start label, disassembly parsing doing in recordmcount.pl
gets confused on powerpc, leading to creation of corrupted module
objects.

This was not a problem so far since the compiler would never create
such text sections.  However, this has changed with a recent change
in GCC 6 to support distances of > 2GB between a function and its
assoicated TOC in the ELFv2 ABI, exposing this problem.

There is already code in recordmcount.pl to handle such data blobs
on the sparc64 platform.  This patch uses the same method to handle
those on powerpc as well.

Cc: stable@vger.kernel.org
Signed-off-by: Ulrich Weigand <redacted>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 scripts/recordmcount.pl | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

Steve can we get an ack for this one, to go via powerpc? cheers
diff --git a/scripts/recordmcount.pl b/scripts/recordmcount.pl
index 826470d7f000..96e2486a6fc4 100755
--- a/scripts/recordmcount.pl
+++ b/scripts/recordmcount.pl
@@ -263,7 +263,8 @@ if ($arch eq "x86_64") {
 
 } elsif ($arch eq "powerpc") {
     $local_regex = "^[0-9a-fA-F]+\\s+t\\s+(\\.?\\S+)";
-    $function_regex = "^([0-9a-fA-F]+)\\s+<(\\.?.*?)>:";
+    # See comment in the sparc64 section for why we use '\w'.
+    $function_regex = "^([0-9a-fA-F]+)\\s+<(\\.?\\w*?)>:";
     $mcount_regex = "^\\s*([0-9a-fA-F]+):.*\\s\\.?_mcount\$";
 
     if ($bits == 64) {
-- 
2.5.0

[PATCH 2/2] powerpc/module: Handle R_PPC64_ENTRY relocations

From: Michael Ellerman <mpe@ellerman.id.au>
Date: 2016-01-12 12:14:40

From: Ulrich Weigand <redacted>

GCC 6 will include changes to generated code with -mcmodel=large,
which is used to build kernel modules on powerpc64le.  This was
necessary because the large model is supposed to allow arbitrary
sizes and locations of the code and data sections, but the ELFv2
global entry point prolog still made the unconditional assumption
that the TOC associated with any particular function can be found
within 2 GB of the function entry point:

func:
	addis r2,r12,(.TOC.-func)@ha
	addi  r2,r2,(.TOC.-func)@l
	.localentry func, .-func

To remove this assumption, GCC will now generate instead this global
entry point prolog sequence when using -mcmodel=large:

	.quad .TOC.-func
func:
	.reloc ., R_PPC64_ENTRY
	ld    r2, -8(r12)
	add   r2, r2, r12
	.localentry func, .-func

The new .reloc triggers an optimization in the linker that will
replace this new prolog with the original code (see above) if the
linker determines that the distance between .TOC. and func is in
range after all.

Since this new relocation is now present in module object files,
the kernel module loader is required to handle them too.  This
patch adds support for the new relocation and implements the
same optimization done by the GNU linker.

Cc: stable@vger.kernel.org
Signed-off-by: Ulrich Weigand <redacted>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 arch/powerpc/include/uapi/asm/elf.h |  2 ++
 arch/powerpc/kernel/module_64.c     | 27 +++++++++++++++++++++++++++
 2 files changed, 29 insertions(+)
diff --git a/arch/powerpc/include/uapi/asm/elf.h b/arch/powerpc/include/uapi/asm/elf.h
index 59dad113897b..c2d21d11c2d2 100644
--- a/arch/powerpc/include/uapi/asm/elf.h
+++ b/arch/powerpc/include/uapi/asm/elf.h
@@ -295,6 +295,8 @@ do {									\
 #define R_PPC64_TLSLD		108
 #define R_PPC64_TOCSAVE		109
 
+#define R_PPC64_ENTRY		118
+
 #define R_PPC64_REL16		249
 #define R_PPC64_REL16_LO	250
 #define R_PPC64_REL16_HI	251
diff --git a/arch/powerpc/kernel/module_64.c b/arch/powerpc/kernel/module_64.c
index 68384514506b..59663af9315f 100644
--- a/arch/powerpc/kernel/module_64.c
+++ b/arch/powerpc/kernel/module_64.c
@@ -635,6 +635,33 @@ int apply_relocate_add(Elf64_Shdr *sechdrs,
 			 */
 			break;
 
+		case R_PPC64_ENTRY:
+			/*
+			 * Optimize ELFv2 large code model entry point if
+			 * the TOC is within 2GB range of current location.
+			 */
+			value = my_r2(sechdrs, me) - (unsigned long)location;
+			if (value + 0x80008000 > 0xffffffff)
+				break;
+			/*
+			 * Check for the large code model prolog sequence:
+		         *	ld r2, ...(r12)
+			 *	add r2, r2, r12
+			 */
+			if ((((uint32_t *)location)[0] & ~0xfffc)
+			    != 0xe84c0000)
+				break;
+			if (((uint32_t *)location)[1] != 0x7c426214)
+				break;
+			/*
+			 * If found, replace it with:
+			 *	addis r2, r12, (.TOC.-func)@ha
+			 *	addi r2, r12, (.TOC.-func)@l
+			 */
+			((uint32_t *)location)[0] = 0x3c4c0000 + PPC_HA(value);
+			((uint32_t *)location)[1] = 0x38420000 + PPC_LO(value);
+			break;
+
 		case R_PPC64_REL16_HA:
 			/* Subtract location pointer */
 			value -= (unsigned long)location;
-- 
2.5.0

Re: [PATCH 1/2] scripts/recordmcount.pl: support data in text section on powerpc

From: Steven Rostedt <rostedt@goodmis.org>
Date: 2016-01-12 15:42:29

On Tue, 12 Jan 2016 23:14:22 +1100
Michael Ellerman [off-list ref] wrote:
From: Ulrich Weigand <redacted>

If a text section starts out with a data blob before the first
function start label, disassembly parsing doing in recordmcount.pl
gets confused on powerpc, leading to creation of corrupted module
objects.

This was not a problem so far since the compiler would never create
such text sections.  However, this has changed with a recent change
in GCC 6 to support distances of > 2GB between a function and its
assoicated TOC in the ELFv2 ABI, exposing this problem.

There is already code in recordmcount.pl to handle such data blobs
on the sparc64 platform.  This patch uses the same method to handle
those on powerpc as well.

Cc: stable@vger.kernel.org
Signed-off-by: Ulrich Weigand <redacted>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 scripts/recordmcount.pl | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

Steve can we get an ack for this one, to go via powerpc? cheers
Acked-by: Steven Rostedt <rostedt@goodmis.org>

-- Steve
quoted hunk
diff --git a/scripts/recordmcount.pl b/scripts/recordmcount.pl
index 826470d7f000..96e2486a6fc4 100755
--- a/scripts/recordmcount.pl
+++ b/scripts/recordmcount.pl
@@ -263,7 +263,8 @@ if ($arch eq "x86_64") {
 
 } elsif ($arch eq "powerpc") {
     $local_regex = "^[0-9a-fA-F]+\\s+t\\s+(\\.?\\S+)";
-    $function_regex = "^([0-9a-fA-F]+)\\s+<(\\.?.*?)>:";
+    # See comment in the sparc64 section for why we use '\w'.
+    $function_regex = "^([0-9a-fA-F]+)\\s+<(\\.?\\w*?)>:";
     $mcount_regex = "^\\s*([0-9a-fA-F]+):.*\\s\\.?_mcount\$";
 
     if ($bits == 64) {

Re: [PATCH 1/2] scripts/recordmcount.pl: support data in text section on powerpc

From: Michael Ellerman <mpe@ellerman.id.au>
Date: 2016-01-12 22:43:29

On Tue, 2016-01-12 at 10:42 -0500, Steven Rostedt wrote:
On Tue, 12 Jan 2016 23:14:22 +1100
Michael Ellerman [off-list ref] wrote:
quoted
From: Ulrich Weigand <redacted>

If a text section starts out with a data blob before the first
function start label, disassembly parsing doing in recordmcount.pl
gets confused on powerpc, leading to creation of corrupted module
objects.

This was not a problem so far since the compiler would never create
such text sections.  However, this has changed with a recent change
in GCC 6 to support distances of > 2GB between a function and its
assoicated TOC in the ELFv2 ABI, exposing this problem.

There is already code in recordmcount.pl to handle such data blobs
on the sparc64 platform.  This patch uses the same method to handle
those on powerpc as well.

Cc: stable@vger.kernel.org
Signed-off-by: Ulrich Weigand <redacted>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 scripts/recordmcount.pl | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

Steve can we get an ack for this one, to go via powerpc? cheers
Acked-by: Steven Rostedt <rostedt@goodmis.org>
Thanks.

Re: [1/2] scripts/recordmcount.pl: support data in text section on powerpc

From: Michael Ellerman <mpe@ellerman.id.au>
Date: 2016-01-13 13:09:25

On Tue, 2016-12-01 at 12:14:22 UTC, Michael Ellerman wrote:
From: Ulrich Weigand <redacted>

If a text section starts out with a data blob before the first
function start label, disassembly parsing doing in recordmcount.pl
gets confused on powerpc, leading to creation of corrupted module
objects.

This was not a problem so far since the compiler would never create
such text sections.  However, this has changed with a recent change
in GCC 6 to support distances of > 2GB between a function and its
assoicated TOC in the ELFv2 ABI, exposing this problem.

There is already code in recordmcount.pl to handle such data blobs
on the sparc64 platform.  This patch uses the same method to handle
those on powerpc as well.

Cc: stable@vger.kernel.org
Signed-off-by: Ulrich Weigand <redacted>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Acked-by: Steven Rostedt <rostedt@goodmis.org>
Applied to powerpc next.

https://git.kernel.org/powerpc/c/2e50c4bef77511b42cc226865d

cheers

Re: [2/2] powerpc/module: Handle R_PPC64_ENTRY relocations

From: Michael Ellerman <mpe@ellerman.id.au>
Date: 2016-01-13 13:09:26

On Tue, 2016-12-01 at 12:14:23 UTC, Michael Ellerman wrote:
From: Ulrich Weigand <redacted>

GCC 6 will include changes to generated code with -mcmodel=large,
which is used to build kernel modules on powerpc64le.  This was
necessary because the large model is supposed to allow arbitrary
sizes and locations of the code and data sections, but the ELFv2
global entry point prolog still made the unconditional assumption
that the TOC associated with any particular function can be found
within 2 GB of the function entry point:

func:
	addis r2,r12,(.TOC.-func)@ha
	addi  r2,r2,(.TOC.-func)@l
	.localentry func, .-func

To remove this assumption, GCC will now generate instead this global
entry point prolog sequence when using -mcmodel=large:

	.quad .TOC.-func
func:
	.reloc ., R_PPC64_ENTRY
	ld    r2, -8(r12)
	add   r2, r2, r12
	.localentry func, .-func

The new .reloc triggers an optimization in the linker that will
replace this new prolog with the original code (see above) if the
linker determines that the distance between .TOC. and func is in
range after all.

Since this new relocation is now present in module object files,
the kernel module loader is required to handle them too.  This
patch adds support for the new relocation and implements the
same optimization done by the GNU linker.

Cc: stable@vger.kernel.org
Signed-off-by: Ulrich Weigand <redacted>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Applied to powerpc next.

https://git.kernel.org/powerpc/c/a61674bdfc7c2bf909c4010699

cheers
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help