[PATCH 0/5 v2] mm/memblock: Introduce memblock_first_region_size() helper

[PATCH 0/5 v2] mm/memblock: Introduce memblock_first_region_size() helper

[Alexander Kuleshov]

Some architectures (like s390, microblaze and etc...) require size
of the first memory region. This patch set provides new memblock_first_region_size()
helper for this case and change usage of memblock.memory.regions[0].size on
memblock_first_region_size() for the following architectures:

* s390
* microblaze
* arm
* unicore32

Changelog:

v2:

Added changes in the architectures to the patchset.

Alexander Kuleshov (5):
  mm/memblock: Introduce memblock_first_region_size() helper
  s390/setup: use memblock_first_region_size helper
  microblaze/mm: Use memblock_first_region_size() helper
  unicore32/mmu: use memblock_first_region_size() helper
  arm/mmu: Use memblock_first_region_size() helper

 arch/arm/mm/mmu.c         | 2 +-
 arch/microblaze/mm/init.c | 6 +++---
 arch/s390/kernel/setup.c  | 2 +-
 arch/unicore32/mm/mmu.c   | 2 +-
 include/linux/memblock.h  | 1 +
 mm/memblock.c             | 5 +++++
 6 files changed, 12 insertions(+), 6 deletions(-)

-- 
2.5.0

[PATCH 1/5 v2] mm/memblock: Introduce memblock_first_region_size() helper

[Alexander Kuleshov]

Some architectures (like s390, microblaze and etc...) require size
of the first memory region. This patch provides new memblock_first_region_size()
helper for this case.

Signed-off-by: Alexander Kuleshov <kuleshovmail@gmail.com>
---
 include/linux/memblock.h | 1 +
 mm/memblock.c            | 5 +++++
 2 files changed, 6 insertions(+)

diff --git a/include/linux/memblock.h b/include/linux/memblock.h
index cc4b019..8a481e5 100644
--- a/include/linux/memblock.h
+++ b/include/linux/memblock.h
@@ -319,6 +319,7 @@ phys_addr_t memblock_phys_mem_size(void);
 phys_addr_t memblock_mem_size(unsigned long limit_pfn);
 phys_addr_t memblock_start_of_DRAM(void);
 phys_addr_t memblock_end_of_DRAM(void);
+phys_addr_t memblock_first_region_size(void);
 void memblock_enforce_memory_limit(phys_addr_t memory_limit);
 int memblock_is_memory(phys_addr_t addr);
 int memblock_is_region_memory(phys_addr_t base, phys_addr_t size);
diff --git a/mm/memblock.c b/mm/memblock.c
index 87108e7..fb4b7ca 100644
--- a/mm/memblock.c
+++ b/mm/memblock.c
@@ -1463,6 +1463,11 @@ phys_addr_t __init_memblock memblock_end_of_DRAM(void)
 	return (memblock.memory.regions[idx].base + memblock.memory.regions[idx].size);
 }
 
+phys_addr_t __init_memblock memblock_first_region_size(void)
+{
+	return memblock.memory.regions[0].size;
+}
+
 void __init memblock_enforce_memory_limit(phys_addr_t limit)
 {
 	phys_addr_t max_addr = (phys_addr_t)ULLONG_MAX;
-- 
2.5.0

[PATCH 2/5 v2] s390/setup: use memblock_first_region_size helper

[Alexander Kuleshov]

The 0a11142e commit (mm/memblock: Introduce memblock_first_region_size()
helper) introduced memblock_first_region_size() helper for the getting
size of the first memblock region. Let's use it instead of directly access
to structure.

Signed-off-by: Alexander Kuleshov <kuleshovmail@gmail.com>
---
 arch/s390/kernel/setup.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/s390/kernel/setup.c b/arch/s390/kernel/setup.c
index ca070d2..69c3859 100644
--- a/arch/s390/kernel/setup.c
+++ b/arch/s390/kernel/setup.c
@@ -563,7 +563,7 @@ static void __init reserve_crashkernel(void)
 	if (rc || crash_size == 0)
 		return;
 
-	if (memblock.memory.regions[0].size < crash_size) {
+	if (memblock_first_region_size() < crash_size) {
 		pr_info("crashkernel reservation failed: %s\n",
 			"first memory chunk must be at least crashkernel size");
 		return;
-- 
2.5.0

[PATCH 4/5 v2] unicore32/mmu: use memblock_first_region_size() helper

[Alexander Kuleshov]

The 0a11142e commit (mm/memblock: Introduce memblock_first_region_size()
helper) introduced memblock_first_region_size() helper for the getting
size of the first memblock region. Let's use it instead of directly access
to structure.

Signed-off-by: Alexander Kuleshov <kuleshovmail@gmail.com>
---
 arch/unicore32/mm/mmu.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/unicore32/mm/mmu.c b/arch/unicore32/mm/mmu.c
index 4f5a532..8cd2d9d 100644
--- a/arch/unicore32/mm/mmu.c
+++ b/arch/unicore32/mm/mmu.c
@@ -314,7 +314,7 @@ static inline void prepare_page_table(void)
 	/*
 	 * Find the end of the first block of lowmem.
 	 */
-	end = memblock.memory.regions[0].base + memblock.memory.regions[0].size;
+	end = memblock.memory.regions[0].base + memblock_first_region_size();
 	if (end >= lowmem_limit)
 		end = lowmem_limit;
 
-- 
2.5.0

[PATCH 5/5 v2] arm/mmu: Use memblock_first_region_size() helper

[Alexander Kuleshov]

The 0a11142e commit (mm/memblock: Introduce memblock_first_region_size()
helper) introduced memblock_first_region_size() helper for the getting
size of the first memblock region. Let's use it instead of directly access
to structure.

Signed-off-by: Alexander Kuleshov <kuleshovmail@gmail.com>
---
 arch/arm/mm/mmu.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/arm/mm/mmu.c b/arch/arm/mm/mmu.c
index 870838a..fc60882 100644
--- a/arch/arm/mm/mmu.c
+++ b/arch/arm/mm/mmu.c
@@ -1176,7 +1176,7 @@ static inline void prepare_page_table(void)
 	/*
 	 * Find the end of the first block of lowmem.
 	 */
-	end = memblock.memory.regions[0].base + memblock.memory.regions[0].size;
+	end = memblock.memory.regions[0].base + memblock_first_region_size();
 	if (end >= arm_lowmem_limit)
 		end = arm_lowmem_limit;
 
-- 
2.5.0

Re: [PATCH 1/5 v2] mm/memblock: Introduce memblock_first_region_size() helper

[Andrew Morton]

On Tue,  1 Sep 2015 20:27:11 +0600 Alexander Kuleshov <kuleshovmail@gmail.com> wrote:

> Some architectures (like s390, microblaze and etc...) require size
> of the first memory region. This patch provides new memblock_first_region_size()
> helper for this case.
> 
> ...
>
> +phys_addr_t __init_memblock memblock_first_region_size(void)
> +{
> +	return memblock.memory.regions[0].size;
> +}
> +

Some callers call this from __init code, which is OK.

Other callers call it from an inlined function and I'm too lazy to work
out if all the callers of those callers are calling
memblock_first_region_size() from a compatible section.

So please either a) demonstrate that all the sectioning is correct (and
maintainable!) or b) simply inline memblock_first_region_size()...