[PATCH] [POWERPC] iSeries: fix section mismatch warnings

WARNING: vmlinux.o(.text+0x8124): Section mismatch: reference to .init.text:.iSeries_early_setup (between '.__start_initialization_iSeries' and '.__mmu_off')
WARNING: vmlinux.o(.text+0x8128): Section mismatch: reference to .init.text:.early_setup (between '.__start_initialization_iSeries' and '.__mmu_off')

Signed-off-by: Stephen Rothwell <redacted>
---
 arch/powerpc/kernel/head_64.S |    7 ++++---
 include/asm-powerpc/ppc_asm.h |   12 ++++++++++++
 2 files changed, 16 insertions(+), 3 deletions(-)

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au

diff --git a/arch/powerpc/kernel/head_64.S b/arch/powerpc/kernel/head_64.S
index 8cdd48e..1448af9 100644
--- a/arch/powerpc/kernel/head_64.S
+++ b/arch/powerpc/kernel/head_64.S

@@ -809,8 +809,9 @@ system_reset_iSeries:
 	mtmsrd	r24			/* RI on */
 	lhz	r24,PACAPACAINDEX(r13)	/* Get processor # */
 	cmpwi	0,r24,0			/* Are we processor 0? */
-	beq	.__start_initialization_iSeries	/* Start up the first processor */
-	mfspr	r4,SPRN_CTRLF
+	bne	1f
+	b	.__start_initialization_iSeries	/* Start up the first processor */
+1:	mfspr	r4,SPRN_CTRLF
 	li	r5,CTRL_RUNLATCH	/* Turn off the run light */
 	andc	r4,r4,r5
 	mtspr	SPRN_CTRLT,r4

@@ -1611,7 +1612,7 @@ _GLOBAL(generic_secondary_smp_init)
 #endif
 
 #ifdef CONFIG_PPC_ISERIES
-_STATIC(__start_initialization_iSeries)
+_INIT_STATIC(__start_initialization_iSeries)
 	/* Clear out the BSS */
 	LOAD_REG_IMMEDIATE(r11,__bss_stop)
 	LOAD_REG_IMMEDIATE(r8,__bss_start)

diff --git a/include/asm-powerpc/ppc_asm.h b/include/asm-powerpc/ppc_asm.h
index fa083d8..6532572 100644
--- a/include/asm-powerpc/ppc_asm.h
+++ b/include/asm-powerpc/ppc_asm.h

@@ -181,6 +181,18 @@ name: \
 	.type GLUE(.,name),@function; \
 GLUE(.,name):
 
+#define _INIT_STATIC(name) \
+	.section ".text.init.refok"; \
+	.align 2 ; \
+	.section ".opd","aw"; \
+name: \
+	.quad GLUE(.,name); \
+	.quad .TOC.@tocbase; \
+	.quad 0; \
+	.previous; \
+	.type GLUE(.,name),@function; \
+GLUE(.,name):
+
 #else /* 32-bit */
 
 #define _GLOBAL(n)	\

-- 
1.5.2.4

This part isnt clicking for me.. 
How does changing a "beq" to a "bne" over a "b" fit into changing
__start_initialization_iSeries static?  


-Will

Hi Will,

On Wed, 25 Jul 2007 11:55:31 -0500 Will Schmidt [off-list ref] wrote:

quoted

 	cmpwi	0,r24,0			/* Are we processor 0? */
-	beq	.__start_initialization_iSeries	/* Start up the first processor */
-	mfspr	r4,SPRN_CTRLF
+	bne	1f
+	b	.__start_initialization_iSeries	/* Start up the first processor */
+1:	mfspr	r4,SPRN_CTRLF
 	li	r5,CTRL_RUNLATCH	/* Turn off the run light */

This part isnt clicking for me.. 
How does changing a "beq" to a "bne" over a "b" fit into changing
__start_initialization_iSeries static?  

Because I moved __start_initialization_iSeries into another section, it
ends up too far away for a conditional branch so something adds a jump
table to the .text section and changes this branch to be via that table.
Unfortunately, the jump table ends up at the start of the .text and ruins
our carefully laid out kernel image.  By inverting the test I can turn
the branch into an unconditional one which has a larger possible offse
(effectively building the jump table manually).

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

On Thu, 2007-07-26 at 11:56 +1000, Stephen Rothwell wrote:

Hi Will,

On Wed, 25 Jul 2007 11:55:31 -0500 Will Schmidt [off-list ref] wrote:

quoted

quoted

 	cmpwi	0,r24,0			/* Are we processor 0? */
-	beq	.__start_initialization_iSeries	/* Start up the first processor */
-	mfspr	r4,SPRN_CTRLF
+	bne	1f
+	b	.__start_initialization_iSeries	/* Start up the first processor */
+1:	mfspr	r4,SPRN_CTRLF
 	li	r5,CTRL_RUNLATCH	/* Turn off the run light */

This part isnt clicking for me.. 
How does changing a "beq" to a "bne" over a "b" fit into changing
__start_initialization_iSeries static?  

Because I moved __start_initialization_iSeries into another section, it
ends up too far away for a conditional branch so something adds a jump
table to the .text section and changes this branch to be via that table.
Unfortunately, the jump table ends up at the start of the .text and ruins
our carefully laid out kernel image.  By inverting the test I can turn
the branch into an unconditional one which has a larger possible offse
(effectively building the jump table manually).

Gotcha, thanks for the clarification.  :-)

-Will

_______________________________________________
Linuxppc-dev mailing list
Linuxppc-dev@ozlabs.org
https://ozlabs.org/mailman/listinfo/linuxppc-dev