Skip to contentSkip to conversation

ARM architecture fixes

PATCH New Arnd Bergmann 3 participants 9 messages May 2, 2013

Subsystems (5)
lore
Share & tools
lore
#0 lore raw

cover ARM architecture fixes

Hi Russell,

These are four bug fixes that came out of my build testing. The first one replaces a patch I put into the patch tracker earlier but had to retract because it broke THUMB2_KERNEL.

The other three are for recent regressions coming from your tree.

Arnd

Arnd Bergmann (4):
ARM: compressed/head.S: work around new binutils warning
ARM: fix prototypes for arch_ioremap_caller functions
ARM: fix NOMMU __arm_ioremap prototype
ARM: highbank: fix build after cache flush change

 arch/arm/boot/compressed/Makefile      | 2 +-
 arch/arm/boot/compressed/head-sa1100.S | 1 +
 arch/arm/boot/compressed/head-shark.S  | 1 +
 arch/arm/boot/compressed/head.S        | 1 +
 arch/arm/mach-ebsa110/core.c           | 2 +-
 arch/arm/mach-highbank/hotplug.c       | 1 +
 arch/arm/mach-imx/mm-imx3.c            | 2 +-
 arch/arm/mach-iop13xx/io.c             | 2 +-
 arch/arm/mach-ixp4xx/common.c          | 2 +-
 arch/arm/mach-msm/common.h             | 2 +-
 arch/arm/mach-msm/io.c                 | 2 +-
 arch/arm/mm/nommu.c                    | 4 ++--
 12 files changed, 13 insertions(+), 9 deletions(-)
Signature · 1 line
1.8.1.2
#1 lore raw

PATCH 1/4 ARM: compressed/head.S: work around new binutils warning

In August 2012, Matthew Gretton-Dann checked a change into binutils labelled "Error on obsolete & warn on deprecated registers", apparently as part of ARMv8 support. Apparently, this was supposed to emit the message "Warning: This coprocessor register access is deprecated in ARMv8" when using certain mcr/mrc instructions and building for ARMv8. Unfortunately, the message that is actually emitted appears to be '(null)', which is less helpful in comparison.

Even more unfortunately, this is biting us on every single kernel build with a new gas, because arch/arm/boot/compressed/head.S and some other files in that directory are built with -march=all since kernel commit 80cec14a8 "[ARM] Add -march=all to assembly file build in arch/arm/boot/compressed" back in v2.6.28.

This patch reverts Russell's nice solution and instead marks the head.S file to be built for armv7-a, which fortunately lets us build all instructions in that file without warnings even on the broken binutils.

Without this patch, building anything results in:

arch/arm/boot/compressed/head.S: Assembler messages:
arch/arm/boot/compressed/head.S:565: Warning: (null) arch/arm/boot/compressed/head.S:676: Warning: (null) arch/arm/boot/compressed/head.S:698: Warning: (null) arch/arm/boot/compressed/head.S:722: Warning: (null) arch/arm/boot/compressed/head.S:726: Warning: (null) arch/arm/boot/compressed/head.S:957: Warning: (null) arch/arm/boot/compressed/head.S:996: Warning: (null) arch/arm/boot/compressed/head.S:997: Warning: (null) arch/arm/boot/compressed/head.S:1027: Warning: (null) arch/arm/boot/compressed/head.S:1035: Warning: (null) arch/arm/boot/compressed/head.S:1046: Warning: (null) arch/arm/boot/compressed/head.S:1060: Warning: (null) arch/arm/boot/compressed/head.S:1092: Warning: (null) arch/arm/boot/compressed/head.S:1094: Warning: (null) arch/arm/boot/compressed/head.S:1095: Warning: (null) arch/arm/boot/compressed/head.S:1102: Warning: (null) arch/arm/boot/compressed/head.S:1134: Warning: (null)

Cc Matthew Gretton-Dann <[email protected]>
Cc Russell King <[email protected]>
---
 arch/arm/boot/compressed/Makefile      | 2 +-
 arch/arm/boot/compressed/head-sa1100.S | 1 +
 arch/arm/boot/compressed/head-shark.S  | 1 +
 arch/arm/boot/compressed/head.S        | 1 +
 4 files changed, 4 insertions(+), 1 deletion(-)
arch/arm/boot/compressed/Makefile +1 -1
index 001a13a..592ede5 100644
@@ -128,7 +128,7 @@ KBUILD_CFLAGS = $(subst -pg, , $(ORIG_CFLAGS))
 endif
 
 ccflags-y := -fpic -mno-single-pic-base -fno-builtin -I$(obj)
-asflags-y := -Wa,-march=all -DZIMAGE
+asflags-y := -DZIMAGE
 
 # Supply kernel BSS size to the decompressor via a linker symbol.
 KBSS_SZ = $(shell $(CROSS_COMPILE)size $(obj)/../../../../vmlinux | \
arch/arm/boot/compressed/head-sa1100.S +1 -0
index 6179d94..3115e31 100644
@@ -11,6 +11,7 @@
 #include <asm/mach-types.h>
 
 		.section        ".start", "ax"
+		.arch	armv4
 
 __SA1100_start:
 
arch/arm/boot/compressed/head-shark.S +1 -0
index 089c560..92b5689 100644
@@ -18,6 +18,7 @@
 	
 		.section	".start", "ax"
 
+		.arch armv4
 		b	__beginning
 	
 __ofw_data:	.long	0				@ the number of memory blocks
arch/arm/boot/compressed/head.S +1 -0
index fe4d9c3..6938559 100644
@@ -11,6 +11,7 @@
 #include <linux/linkage.h>
 #include <asm/assembler.h>
 
+	.arch	armv7-a
 /*
  * Debugging stuff
  *
Signature · 1 line
1.8.1.2
#2 lore raw

PATCH 2/4 ARM: fix prototypes for arch_ioremap_caller functions

In reply to Arnd Bergmann

Patch "ARM: 7704/1: mm: Use phys_addr_t properly for ioremap functions" changes the prototype for the arch_ioremap_caller() function, a change which was evidently never tested, because all functions that get assigned to this pointer still have an unchanged prototype:

arch/arm/mach-ebsa110/core.c: In function 'ebsa110_init_early':
arch/arm/mach-ebsa110/core.c:130:22: warning: assignment from incompatible pointer type [enabled by default]

  arch_ioremap_caller = ebsa110_ioremap_caller;
                      ^
arch/arm/mach-iop13xx/io.c: In function 'iop13xx_init_early':
arch/arm/mach-iop13xx/io.c:88:22: warning: assignment from incompatible pointer type [enabled by default]
  arch_ioremap_caller = __iop13xx_ioremap_caller;
                      ^

This does the trivial change to all the respective functions to avoid the build warnings.

Cc Laura Abbott <[email protected]>
Cc Russell King <[email protected]>
---
 arch/arm/mach-ebsa110/core.c  | 2 +-
 arch/arm/mach-imx/mm-imx3.c   | 2 +-
 arch/arm/mach-iop13xx/io.c    | 2 +-
 arch/arm/mach-ixp4xx/common.c | 2 +-
 arch/arm/mach-msm/common.h    | 2 +-
 arch/arm/mach-msm/io.c        | 2 +-
 arch/arm/mm/nommu.c           | 2 +-
 7 files changed, 7 insertions(+), 7 deletions(-)
arch/arm/mach-ebsa110/core.c +1 -1
index b13cc74..8a53f34 100644
@@ -116,7 +116,7 @@ static void __init ebsa110_map_io(void)
 	iotable_init(ebsa110_io_desc, ARRAY_SIZE(ebsa110_io_desc));
 }
 
-static void __iomem *ebsa110_ioremap_caller(unsigned long cookie, size_t size,
+static void __iomem *ebsa110_ioremap_caller(phys_addr_t cookie, size_t size,
 					    unsigned int flags, void *caller)
 {
 	return (void __iomem *)cookie;
arch/arm/mach-imx/mm-imx3.c +1 -1
index e0e69a6..eed32ca 100644
@@ -65,7 +65,7 @@ static void imx3_idle(void)
 		: "=r" (reg));
 }
 
-static void __iomem *imx3_ioremap_caller(unsigned long phys_addr, size_t size,
+static void __iomem *imx3_ioremap_caller(phys_addr_t phys_addr, size_t size,
 					 unsigned int mtype, void *caller)
 {
 	if (mtype == MT_DEVICE) {
arch/arm/mach-iop13xx/io.c +1 -1
index 183dc8b..faaf7d4 100644
@@ -23,7 +23,7 @@
 
 #include "pci.h"
 
-static void __iomem *__iop13xx_ioremap_caller(unsigned long cookie,
+static void __iomem *__iop13xx_ioremap_caller(phys_addr_t cookie,
 	size_t size, unsigned int mtype, void *caller)
 {
 	void __iomem * retval;
arch/arm/mach-ixp4xx/common.c +1 -1
index 6600cff..d7223b3 100644
@@ -559,7 +559,7 @@ void ixp4xx_restart(char mode, const char *cmd)
  * fallback to the default.
  */
 
-static void __iomem *ixp4xx_ioremap_caller(unsigned long addr, size_t size,
+static void __iomem *ixp4xx_ioremap_caller(phys_addr_t addr, size_t size,
 					   unsigned int mtype, void *caller)
 {
 	if (!is_pci_memory(addr))
arch/arm/mach-msm/common.h +1 -1
index ce8215a..421cf77 100644
@@ -23,7 +23,7 @@ extern void msm_map_msm8x60_io(void);
 extern void msm_map_msm8960_io(void);
 extern void msm_map_qsd8x50_io(void);
 
-extern void __iomem *__msm_ioremap_caller(unsigned long phys_addr, size_t size,
+extern void __iomem *__msm_ioremap_caller(phys_addr_t phys_addr, size_t size,
 					  unsigned int mtype, void *caller);
 
 extern struct smp_operations msm_smp_ops;
arch/arm/mach-msm/io.c +1 -1
index 123ef9c..fd65b6d 100644
@@ -172,7 +172,7 @@ void __init msm_map_msm7x30_io(void)
 }
 #endif /* CONFIG_ARCH_MSM7X30 */
 
-void __iomem *__msm_ioremap_caller(unsigned long phys_addr, size_t size,
+void __iomem *__msm_ioremap_caller(phys_addr_t phys_addr, size_t size,
 				   unsigned int mtype, void *caller)
 {
 	if (mtype == MT_DEVICE) {
arch/arm/mm/nommu.c +1 -1
index dd3a6c6..e9f2e44 100644
@@ -95,7 +95,7 @@ void __iomem *__arm_ioremap(unsigned long phys_addr, size_t size,
 }
 EXPORT_SYMBOL(__arm_ioremap);
 
-void __iomem * (*arch_ioremap_caller)(unsigned long, size_t, unsigned int, void *);
+void __iomem * (*arch_ioremap_caller)(phys_addr_t, size_t, unsigned int, void *);
 
 void __iomem *__arm_ioremap_caller(unsigned long phys_addr, size_t size,
 				   unsigned int mtype, void *caller)
Signature · 1 line
1.8.1.2
#3 lore raw

PATCH 3/4 ARM: fix NOMMU __arm_ioremap prototype

In reply to Arnd Bergmann

The declaration of __arm_ioremap has changed to take a phys_addr_t as the first argument, while the definition for nommu has not changed. This adds the obvious change.

Cc Laura Abbott <[email protected]>
Cc Russell King <[email protected]>
---
 arch/arm/mm/nommu.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/arm/mm/nommu.c +1 -1
index e9f2e44..98303d9 100644
@@ -88,7 +88,7 @@ void __iomem *__arm_ioremap_pfn_caller(unsigned long pfn, unsigned long offset,
 	return __arm_ioremap_pfn(pfn, offset, size, mtype);
 }
 
-void __iomem *__arm_ioremap(unsigned long phys_addr, size_t size,
+void __iomem *__arm_ioremap(phys_addr_t phys_addr, size_t size,
 			    unsigned int mtype)
 {
 	return (void __iomem *)phys_addr;
Signature · 1 line
1.8.1.2
#4 lore raw

PATCH 3/4 ARM: fix NOMMU __arm_ioremap prototype

Quoted 3 lines · Arnd Bergmann

The declaration of __arm_ioremap has changed to take a phys_addr_t as the first argument, while the definition for nommu has not changed. This adds the obvious change.

I've now dropped the offending patch from Laura which created this problem (and from what it looks like other problems elsewhere).

It would be much better to have one patch which addresses this properly rather than one patch causing a problem and then lots of smaller patches fixing the problems caused by the first patch.

#5 lore raw

PATCH 4/4 ARM: highbank: fix build after cache flush change

In reply to Arnd Bergmann

Commit "ARM: cpu hotplug: remove majority of cache flushing from platforms" has removed the inclusion of asm/cacheflush.h from a number of files. Highbank still needs it because it calls flush_cache_louis().

Cc Russell King <[email protected]>
Cc Rob Herring <[email protected]>
---
 arch/arm/mach-highbank/hotplug.c | 1 +
 1 file changed, 1 insertion(+)
arch/arm/mach-highbank/hotplug.c +1 -0
index 80957f4..a019e4e 100644
@@ -14,6 +14,7 @@
  * this program.  If not, see <http://www.gnu.org/licenses/>.
  */
 #include <linux/kernel.h>
+#include <asm/cacheflush.h>
 
 #include "core.h"
 #include "sysregs.h"
Signature · 1 line
1.8.1.2
#6 lore raw

PATCH 4/4 ARM: highbank: fix build after cache flush change

Quoted 4 lines · Arnd Bergmann

Commit "ARM: cpu hotplug: remove majority of cache flushing from platforms" has removed the inclusion of asm/cacheflush.h from a number of files. Highbank still needs it because it calls flush_cache_louis().

AFAICT, the commit in question only showed up buried in a reply to another thread and was never cc'ed to the relevant platform maintainers...

Quoted 5 lines
Cc Russell King <[email protected]>
Cc Rob Herring <[email protected]>

---

I'd rather see the highbank changes dropped as this will all go away with PSCI in 3.11, but:

Rob

Quoted 16 lines
 arch/arm/mach-highbank/hotplug.c | 1 +
 1 file changed, 1 insertion(+)
arch/arm/mach-highbank/hotplug.c +1 -0
index 80957f4..a019e4e 100644
@@ -14,6 +14,7 @@
  * this program.  If not, see <http://www.gnu.org/licenses/>.
  */
 #include <linux/kernel.h>
+#include <asm/cacheflush.h>
 
 #include "core.h"
 #include "sysregs.h"
#7 lore raw

PATCH 4/4 ARM: highbank: fix build after cache flush change

Quoted 15 lines · Rob Herring
Quoted 4 lines · Arnd Bergmann

Commit "ARM: cpu hotplug: remove majority of cache flushing from platforms" has removed the inclusion of asm/cacheflush.h from a number of files. Highbank still needs it because it calls flush_cache_louis().

AFAICT, the commit in question only showed up buried in a reply to another thread and was never cc'ed to the relevant platform maintainers...

Quoted 5 lines
Cc Russell King <[email protected]>
Cc Rob Herring <[email protected]>

---

Quoted 4 lines

I'd rather see the highbank changes dropped as this will all go away with PSCI in 3.11, but:

I'll drop the removal of asm/cacheflush.h but not the rest of it; I believe your existing code to be in error with or without my changes.

#8 lore raw

PATCH 4/4 ARM: highbank: fix build after cache flush change

Quoted 16 lines · Russell King - ARM Linux
Quoted 15 lines · Rob Herring
Quoted 4 lines · Arnd Bergmann

Commit "ARM: cpu hotplug: remove majority of cache flushing from platforms" has removed the inclusion of asm/cacheflush.h from a number of files. Highbank still needs it because it calls flush_cache_louis().

AFAICT, the commit in question only showed up buried in a reply to another thread and was never cc'ed to the relevant platform maintainers...

Quoted 5 lines
Cc Russell King <[email protected]>
Cc Rob Herring <[email protected]>

---

Quoted 7 lines
Quoted 4 lines

I'd rather see the highbank changes dropped as this will all go away with PSCI in 3.11, but:

I'll drop the removal of asm/cacheflush.h but not the rest of it; I believe your existing code to be in error with or without my changes.

Actually, no I'm not even going to do that, because that is _wrong_. This is what my mach-highbank/hotplug.c looks like after my change:

#include <linux/kernel.h>

#include "core.h"
#include "sysregs.h"

extern void secondary_startup(void);

/*
* platform-specific code to shutdown a CPU
*
*/
void __ref highbank_cpu_die(unsigned int cpu)

{
        highbank_set_cpu_jump(cpu, phys_to_virt(0));
        highbank_set_core_pwr();

        cpu_do_idle();

        /* We should never return from idle */
        panic("highbank: cpu %d unexpectedly exit from shutdown\n", cpu);
}

hence my change is completely correct in removing asm/cacheflush.h.