Skip to contentSkip to conversation

Fix section mismatches when compiling atlas and decstation defconfigs

MIPS Applied Shane McDonald 3 participants 5 messages Jul 5, 2008 – Jul 9, 2008 MIPS

lore
Share & tools
lore
#0 lore raw

From: Shane McDonald <[email protected]>

Section mismatches are reported when compiling the default Atlas configuration and the default Decstation configuration. This patch resolves those mismatches by defining affected functions with the __cpuinit attribute, rather than __init.

---
 arch/mips/mm/c-r3k.c   |    6 +++---
 arch/mips/mm/sc-rm7k.c |    4 ++--
 2 files changed, 5 insertions(+), 5 deletions(-)
arch/mips/mm/c-r3k.c +3 -3
diff -uprN -X orig/Documentation/dontdiff orig/arch/mips/mm/c-r3k.c patched/arch/mips/mm/c-r3k.c
@@ -26,7 +26,7 @@
 static unsigned long icache_size, dcache_size;		/* Size in bytes */
 static unsigned long icache_lsize, dcache_lsize;	/* Size in bytes */
 
-unsigned long __init r3k_cache_size(unsigned long ca_flags)
+unsigned long __cpuinit r3k_cache_size(unsigned long ca_flags)
 {
 	unsigned long flags, status, dummy, size;
 	volatile unsigned long *p;
@@ -61,7 +61,7 @@ unsigned long __init r3k_cache_size(unsi
 	return size * sizeof(*p);
 }
 
-unsigned long __init r3k_cache_lsize(unsigned long ca_flags)
+unsigned long __cpuinit r3k_cache_lsize(unsigned long ca_flags)
 {
 	unsigned long flags, status, lsize, i;
 	volatile unsigned long *p;
@@ -90,7 +90,7 @@ unsigned long __init r3k_cache_lsize(uns
 	return lsize * sizeof(*p);
 }
 
-static void __init r3k_probe_cache(void)
+static void __cpuinit r3k_probe_cache(void)
 {
 	dcache_size = r3k_cache_size(ST0_ISC);
 	if (dcache_size)
arch/mips/mm/sc-rm7k.c +2 -2
diff -uprN -X orig/Documentation/dontdiff orig/arch/mips/mm/sc-rm7k.c patched/arch/mips/mm/sc-rm7k.c
@@ -86,7 +86,7 @@ static void rm7k_sc_inv(unsigned long ad
 /*
  * This function is executed in uncached address space.
  */
-static __init void __rm7k_sc_enable(void)
+static __cpuinit void __rm7k_sc_enable(void)
 {
 	int i;
 
@@ -107,7 +107,7 @@ static __init void __rm7k_sc_enable(void
 	}
 }
 
-static __init void rm7k_sc_enable(void)
+static __cpuinit void rm7k_sc_enable(void)
 {
 	if (read_c0_config() & RM7K_CONF_SE)
 		return;
Applied
#1 lore raw
Quoted 1 line · Shane McDonald

From: Shane McDonald <[email protected]>

Btw, while it doesn't harm there is no need to insert this From: line into the patch if it's identical to the sender's email address as in your case. Where it's missing git will just pick the From: address from the email headers.

Quoted 6 lines

Section mismatches are reported when compiling the default Atlas configuration and the default Decstation configuration. This patch resolves those mismatches by defining affected functions with the __cpuinit attribute, rather than __init.

Thanks, applied.

Ralf

#2 lore raw
In reply to Shane McDonald
Quoted 8 lines · Shane McDonald

From: Shane McDonald <[email protected]>

Section mismatches are reported when compiling the default Atlas configuration and the default Decstation configuration. This patch resolves those mismatches by defining affected functions with the __cpuinit attribute, rather than __init.

I already had an earlier version of your patch applied so I have only applied the c-r4k.c part now with an adjusted comment.

Thanks,

Ralf

#3 lore raw
Quoted 9 lines · Ralf Baechle
Quoted 7 lines · Shane McDonald

From: Shane McDonald <[email protected]>

Section mismatches are reported when compiling the default Atlas configuration and the default Decstation configuration. This patch resolves those mismatches by defining affected functions with the __cpuinit attribute, rather than __init.

Quoted 4 lines
Quoted 1 line

I already had an earlier version of your patch applied so I have only applied the c-r4k.c part now with an adjusted comment.

Isn't Atlas doomed to death?

Thanks,
Dmitri

Quoted 5 lines

Thanks,

Ralf

#4 lore raw
Quoted 10 lines · Dmitri Vorobiev
Quoted 9 lines · Ralf Baechle
Quoted 7 lines · Shane McDonald

From: Shane McDonald <[email protected]>

Section mismatches are reported when compiling the default Atlas configuration and the default Decstation configuration. This patch resolves those mismatches by defining affected functions with the __cpuinit attribute, rather than __init.

Quoted 6 lines
Quoted 4 lines
Quoted 1 line

I already had an earlier version of your patch applied so I have only applied the c-r4k.c part now with an adjusted comment.

Isn't Atlas doomed to death?

Yes - but there is nothing in this patch that actually makes it Atlas or DECstation specific. It's simply fixing bugs in the CPU cache code.

Ralf