memblock: clenup
v2 cover memblock: clenup
Changes in v2:
* remove the two round reduction patch
* add 129th memory block to memblock.reserved
* add memblock_reserve_many_may_conflict_check()
* two new patch at the end
Wei Yang (8):
memblock tests: reserve the 129th memory block at all possible
position
memblock tests: add memblock_reserve_many_may_conflict_check() mm/memblock: fix comment for memblock_isolate_range() mm/memblock: remove consecutive regions at once memblock tests: add memblock_overlaps_region_checks mm/memblock: return true directly on finding overlap region mm/memblock: use PAGE_ALIGN_DOWN to get pgend in free_memmap mm/memblock: default region's nid may be MAX_NUMNODES
mm/memblock.c | 36 +-- tools/include/linux/mm.h | 1 + tools/testing/memblock/tests/basic_api.c | 286 ++++++++++++++++++----- tools/testing/memblock/tests/common.c | 4 +- tools/testing/memblock/tests/common.h | 4 + 5 files changed, 262 insertions(+), 69 deletions(-)
Signature · 1 line
2.34.1
PATCH v2 1/8 memblock tests: reserve the 129th memory block at all possible position
In stead of add 129th memory block at the last position, let's try all possible position.
--- tools/testing/memblock/tests/basic_api.c | 121 ++++++++++++----------- 1 file changed, 65 insertions(+), 56 deletions(-)
tools/testing/memblock/tests/basic_api.c +65 -56
@@ -898,84 +898,93 @@ static int memblock_reserve_near_max_check(void)* memblock.memory.max, find a new valid memory as * reserved.regions. */ +/* Keep the gap so these memory region will not be merged. */ +#define MEMORY_BASE (idx) (SZ_128K + (MEM_SIZE * 2) * (idx))static int memblock_reserve_many_check (void ) { -int i; +int i, skip;void *orig_region;struct region r = { .base =SZ_16K , .size =SZ_16K , }; -phys_addr_t memory_base =SZ_128K ;phys_addr_t new_reserved_regions_size;PREFIX_PUSH (); -reset_memblock_regions (); -memblock_allow_resize (); +/* Reserve the 129th memory block for all possible positions*/ +for (skip =0 ; skip <INIT_MEMBLOCK_REGIONS +1 ; skip++) { +reset_memblock_regions (); +memblock_allow_resize (); -/* Add a valid memory region used by double_array(). */ -dummy_physical_memory_init (); -memblock_add (dummy_physical_memory_base (),MEM_SIZE ); +/* Add a valid memory region used by double_array(). */ +dummy_physical_memory_init (); +memblock_add (dummy_physical_memory_base (),MEM_SIZE ); -for (i =0 ; i <INIT_MEMBLOCK_REGIONS ; i++) { -/* Reserve some fakes memory region to fulfill the memblock. */ -memblock_reserve (memory_base,MEM_SIZE ); +for (i =0 ; i <INIT_MEMBLOCK_REGIONS +1 ; i++) { +if (i == skip) +continue ; -ASSERT_EQ (memblock.reserved.cnt, i +1 ); -ASSERT_EQ (memblock.reserved.total_size, (i +1 ) *MEM_SIZE ); +/* Reserve some fakes memory region to fulfill the memblock. */ +memblock_reserve (MEMORY_BASE (i),MEM_SIZE ); -/* Keep the gap so these memory region will not be merged. */ - memory_base +=MEM_SIZE *2 ; - } +if (i < skip) { +ASSERT_EQ (memblock.reserved.cnt, i +1 ); +ASSERT_EQ (memblock.reserved.total_size, (i +1 ) *MEM_SIZE ); + }else { +ASSERT_EQ (memblock.reserved.cnt, i); +ASSERT_EQ (memblock.reserved.total_size, i *MEM_SIZE ); + } + } - orig_region = memblock.reserved.regions; + orig_region = memblock.reserved.regions; -/* This reserve the 129 memory_region, and makes it double array. */ -memblock_reserve (memory_base,MEM_SIZE ); +/* This reserve the 129 memory_region, and makes it double array. */ +memblock_reserve (MEMORY_BASE (skip),MEM_SIZE ); -/* -* This is the memory region size used by the doubled reserved.regions, -* and it has been reserved due to it has been used. The size is used to -* calculate the total_size that the memblock.reserved have now. -*/ - new_reserved_regions_size =PAGE_ALIGN ((INIT_MEMBLOCK_REGIONS *2 ) * -sizeof (struct memblock_region )); -/* -* The double_array() will find a free memory region as the new -* reserved.regions, and the used memory region will be reserved, so -* there will be one more region exist in the reserved memblock. And the -* one more reserved region's size is new_reserved_regions_size. -*/ -ASSERT_EQ (memblock.reserved.cnt,INIT_MEMBLOCK_REGIONS +2 ); -ASSERT_EQ (memblock.reserved.total_size, (INIT_MEMBLOCK_REGIONS +1 ) *MEM_SIZE + - new_reserved_regions_size); -ASSERT_EQ (memblock.reserved.max,INIT_MEMBLOCK_REGIONS *2 ); +/* +* This is the memory region size used by the doubled reserved.regions, +* and it has been reserved due to it has been used. The size is used to +* calculate the total_size that the memblock.reserved have now. +*/ + new_reserved_regions_size =PAGE_ALIGN ((INIT_MEMBLOCK_REGIONS *2 ) * +sizeof (struct memblock_region )); +/* +* The double_array() will find a free memory region as the new +* reserved.regions, and the used memory region will be reserved, so +* there will be one more region exist in the reserved memblock. And the +* one more reserved region's size is new_reserved_regions_size. +*/ +ASSERT_EQ (memblock.reserved.cnt,INIT_MEMBLOCK_REGIONS +2 ); +ASSERT_EQ (memblock.reserved.total_size, (INIT_MEMBLOCK_REGIONS +1 ) *MEM_SIZE + + new_reserved_regions_size); +ASSERT_EQ (memblock.reserved.max,INIT_MEMBLOCK_REGIONS *2 ); -/* -* Now memblock_double_array() works fine. Let's check after the -* double_array(), the memblock_reserve() still works as normal. -*/ -memblock_reserve (r.base, r.size); -ASSERT_EQ (memblock.reserved.regions[0 ].base, r.base); -ASSERT_EQ (memblock.reserved.regions[0 ].size, r.size); +/* +* Now memblock_double_array() works fine. Let's check after the +* double_array(), the memblock_reserve() still works as normal. +*/ +memblock_reserve (r.base, r.size); +ASSERT_EQ (memblock.reserved.regions[0 ].base, r.base); +ASSERT_EQ (memblock.reserved.regions[0 ].size, r.size); -ASSERT_EQ (memblock.reserved.cnt,INIT_MEMBLOCK_REGIONS +3 ); -ASSERT_EQ (memblock.reserved.total_size, (INIT_MEMBLOCK_REGIONS +1 ) *MEM_SIZE + - new_reserved_regions_size + - r.size); -ASSERT_EQ (memblock.reserved.max,INIT_MEMBLOCK_REGIONS *2 ); +ASSERT_EQ (memblock.reserved.cnt,INIT_MEMBLOCK_REGIONS +3 ); +ASSERT_EQ (memblock.reserved.total_size, (INIT_MEMBLOCK_REGIONS +1 ) *MEM_SIZE + + new_reserved_regions_size + + r.size); +ASSERT_EQ (memblock.reserved.max,INIT_MEMBLOCK_REGIONS *2 ); -dummy_physical_memory_cleanup (); +dummy_physical_memory_cleanup (); -/* -* The current reserved.regions is occupying a range of memory that -* allocated from dummy_physical_memory_init(). After free the memory, -* we must not use it. So restore the origin memory region to make sure -* the tests can run as normal and not affected by the double array. -*/ - memblock.reserved.regions = orig_region; - memblock.reserved.cnt =INIT_MEMBLOCK_RESERVED_REGIONS ; +/* +* The current reserved.regions is occupying a range of memory that +* allocated from dummy_physical_memory_init(). After free the memory, +* we must not use it. So restore the origin memory region to make sure +* the tests can run as normal and not affected by the double array. +*/ + memblock.reserved.regions = orig_region; + memblock.reserved.cnt =INIT_MEMBLOCK_RESERVED_REGIONS ; + }test_pass_pop ();
Signature · 1 line
2.34.1
Mike Rapoport
maintainerRe: [Patch v2 1/8] memblock tests: reserve the 129th memory block at all possible position
Quoted 2 lines · Wei Yang
In stead of add 129th memory block at the last position, let's try all possible position.
Why do you insist on changing the existing test rather than adding a new one?
Quoted 4 lines
Signed-off-by Wei Yang <[email protected]>--- tools/testing/memblock/tests/basic_api.c | 121 ++++++++++++----------- 1 file changed, 65 insertions(+), 56 deletions(-)
Signature · 2 lines
Sincerely yours, Mike.
Re: [Patch v2 1/8] memblock tests: reserve the 129th memory block at all possible position
Quoted 7 lines · Mike Rapoport
Quoted 2 lines · Wei Yang
In stead of add 129th memory block at the last position, let's try all possible position.
Why do you insist on changing the existing test rather than adding a new one?
Sounds there is some misunderstanding between us.
I am not sure about your idea at first, so I sent a draft to confirm with you. Then I came up with another version which could trigger the overlap bug.
You mentioned to keep both and not objection to the first draft, which is the same as this one, I thought this is what you expect.
Well, I will add a new one next round. Do you have some suggestion on the function name? memblock_reserve_many_all_position_check ?
Quoted 9 lines
Quoted 4 lines
Signed-off-by Wei Yang <[email protected]>--- tools/testing/memblock/tests/basic_api.c | 121 ++++++++++++----------- 1 file changed, 65 insertions(+), 56 deletions(-)Signature · 2 lines
Sincerely yours, Mike.
Signature · 2 lines
Wei Yang Help you, Help me
Mike Rapoport
maintainerRe: [Patch v2 1/8] memblock tests: reserve the 129th memory block at all possible position
Quoted 9 lines · Wei Yang
Quoted 7 lines · Mike Rapoport
Quoted 2 lines · Wei Yang
In stead of add 129th memory block at the last position, let's try all possible position.
Why do you insist on changing the existing test rather than adding a new one?
Quoted 7 lines
Sounds there is some misunderstanding between us.
I am not sure about your idea at first, so I sent a draft to confirm with you. Then I came up with another version which could trigger the overlap bug.
You mentioned to keep both and not objection to the first draft, which is the same as this one, I thought this is what you expect.
Sorry if I wasn't clear. My intention was to keep the existing test and add a new one rather than update the old test.
Quoted 2 lines
Well, I will add a new one next round. Do you have some suggestion on the function name?
memblock_reserve_many_all_position_check?
How about memblock_reserve_all_locations_check?
Quoted 13 lines
Quoted 9 lines
Quoted 4 lines
Signed-off-by Wei Yang <[email protected]>--- tools/testing/memblock/tests/basic_api.c | 121 ++++++++++++----------- 1 file changed, 65 insertions(+), 56 deletions(-)Signature · 2 lines
Sincerely yours, Mike.Signature · 2 lines
Wei Yang Help you, Help me
Signature · 2 lines
Sincerely yours, Mike.
Re: [Patch v2 1/8] memblock tests: reserve the 129th memory block at all possible position
Quoted 12 lines · Mike Rapoport
Quoted 11 lines · Wei Yang
Quoted 7 lines · Mike Rapoport
Quoted 2 lines · Wei Yang
In stead of add 129th memory block at the last position, let's try all possible position.
Why do you insist on changing the existing test rather than adding a new one?
Sounds there is some misunderstanding between us.
Quoted 9 lines
Quoted 5 lines
I am not sure about your idea at first, so I sent a draft to confirm with you. Then I came up with another version which could trigger the overlap bug.
You mentioned to keep both and not objection to the first draft, which is the same as this one, I thought this is what you expect.
Sorry if I wasn't clear. My intention was to keep the existing test and add a new one rather than update the old test.
You are so kind. Nice to work with you.
Quoted 5 lines
Quoted 2 lines
Well, I will add a new one next round. Do you have some suggestion on the function name?
memblock_reserve_many_all_position_check?How about
memblock_reserve_all_locations_check?
Yep, sounds good :-)
Quoted 17 lines
Quoted 13 lines
Quoted 9 lines
Quoted 4 lines
Signed-off-by Wei Yang <[email protected]>--- tools/testing/memblock/tests/basic_api.c | 121 ++++++++++++----------- 1 file changed, 65 insertions(+), 56 deletions(-)Signature · 2 lines
Sincerely yours, Mike.Signature · 2 lines
Wei Yang Help you, Help meSignature · 2 lines
Sincerely yours, Mike.
Signature · 2 lines
Wei Yang Help you, Help me
PATCH v2 4/8 mm/memblock: remove consecutive regions at once
In reply to Wei YangIn memblock_remove_range, we would remove consecutive regions after isolation.
Instead of removing each region from end to start one by one, we could remove those consecutive regions at once.
--- mm/memblock.c | 24 ++++++++++++++++-------- 1 file changed, 16 insertions(+), 8 deletions(-)
mm/memblock.c +16 -8
@@ -343,12 +343,15 @@ static phys_addr_t __init_memblock memblock_find_in_range(phys_addr_t start,return ret; } -static void __init_memblock memblock_remove_region(struct memblock_type *type, unsigned long r) +static void __init_memblock memblock_remove_regions(struct memblock_type *type, +unsigned long r, +unsigned int num) { - type->total_size -= type->regions[r].size; -memmove (&type->regions[r], &type->regions[r +1 ], - (type->cnt - (r +1 )) *sizeof (type->regions[r])); - type->cnt--; +for (int i =0 ; i < num; i++) + type->total_size -= type->regions[r + i].size; +memmove (&type->regions[r], &type->regions[r + num], + (type->cnt - (r + num)) *sizeof (type->regions[r])); + type->cnt -= num;/* Special case for empty arrays */ if (type->cnt ==0 ) {
@@ -360,6 +363,11 @@ static void __init_memblock memblock_remove_region(struct memblock_type *type, u } } +static void __init_memblock memblock_remove_region(struct memblock_type *type, unsigned long r) +{ +memblock_remove_regions (type, r,1 ); +} +#ifndef CONFIG_ARCH_KEEP_MEMBLOCK /** * memblock_discard - discard memory and reserved arrays if they were allocated
@@ -846,14 +854,14 @@ static int __init_memblock memblock_remove_range(struct memblock_type *type, phys_addr_t base, phys_addr_t size) { int start_rgn, end_rgn; -int i, ret; +int ret; ret =memblock_isolate_range (type, base, size, &start_rgn, &end_rgn);if (ret)return ret; -for (i = end_rgn -1 ; i >= start_rgn; i--) -memblock_remove_region (type, i); +if (end_rgn - start_rgn) +memblock_remove_regions (type, start_rgn, end_rgn - start_rgn);return 0 ; }
Signature · 1 line
2.34.1
Mike Rapoport
maintainerRe: [Patch v2 4/8] mm/memblock: remove consecutive regions at once
Quoted 5 lines · Wei Yang
In
memblock_remove_range, we would remove consecutive regions after isolation.Instead of removing each region from end to start one by one, we could remove those consecutive regions at once.
I don't think this is worth the churn.
Quoted 63 lines
Signed-off-by Wei Yang <[email protected]>--- mm/memblock.c | 24 ++++++++++++++++-------- 1 file changed, 16 insertions(+), 8 deletions(-)mm/memblock.c +16
-8@@ -343,12 +343,15 @@ static phys_addr_t __init_memblock memblock_find_in_range(phys_addr_t start,return ret; } -static void __init_memblock memblock_remove_region(struct memblock_type *type, unsigned long r) +static void __init_memblock memblock_remove_regions(struct memblock_type *type, +unsigned long r, +unsigned int num) { - type->total_size -= type->regions[r].size; -memmove (&type->regions[r], &type->regions[r +1 ], - (type->cnt - (r +1 )) *sizeof (type->regions[r])); - type->cnt--; +for (int i =0 ; i < num; i++) + type->total_size -= type->regions[r + i].size; +memmove (&type->regions[r], &type->regions[r + num], + (type->cnt - (r + num)) *sizeof (type->regions[r])); + type->cnt -= num;/* Special case for empty arrays */ if (type->cnt ==0 ) {@@ -360,6 +363,11 @@ static void __init_memblock memblock_remove_region(struct memblock_type *type, u } } +static void __init_memblock memblock_remove_region(struct memblock_type *type, unsigned long r) +{ +memblock_remove_regions (type, r,1 ); +} +#ifndef CONFIG_ARCH_KEEP_MEMBLOCK /** * memblock_discard - discard memory and reserved arrays if they were allocated @@ -846,14 +854,14 @@ static int __init_memblock memblock_remove_range(struct memblock_type *type, phys_addr_t base, phys_addr_t size) { int start_rgn, end_rgn; -int i, ret; +int ret; ret =memblock_isolate_range (type, base, size, &start_rgn, &end_rgn);if (ret)return ret; -for (i = end_rgn -1 ; i >= start_rgn; i--) -memblock_remove_region (type, i); +if (end_rgn - start_rgn) +memblock_remove_regions (type, start_rgn, end_rgn - start_rgn);return 0 ; }Signature · 1 line
2.34.1
Signature · 2 lines
Sincerely yours, Mike.
Re: [Patch v2 4/8] mm/memblock: remove consecutive regions at once
Quoted 9 lines · Mike Rapoport
Quoted 5 lines · Wei Yang
In
memblock_remove_range, we would remove consecutive regions after isolation.Instead of removing each region from end to start one by one, we could remove those consecutive regions at once.
I don't think this is worth the churn.
Will drop it.
PATCH v2 5/8 memblock tests: add memblock_overlaps_region_checks
In reply to Wei YangAdd a test case for memblock_overlaps_region().
--- tools/testing/memblock/tests/basic_api.c | 48 ++++++++++++++++++++++++ tools/testing/memblock/tests/common.h | 3 ++ 2 files changed, 51 insertions(+)
tools/testing/memblock/tests/basic_api.c +48 -0
@@ -2261,6 +2261,53 @@ static int memblock_trim_memory_checks(void)return 0 ; } +static int memblock_overlaps_region_check (void ) +{ +struct region r = { + .base =SZ_1G , + .size =SZ_4M + }; + +PREFIX_PUSH (); + +reset_memblock_regions (); +memblock_add (r.base, r.size); + +/* Far Away */ +ASSERT_FALSE (memblock_overlaps_region (&memblock.memory,SZ_1M ,SZ_1M )); +ASSERT_FALSE (memblock_overlaps_region (&memblock.memory,SZ_2G ,SZ_1M )); + +/* Neighbor */ +ASSERT_FALSE (memblock_overlaps_region (&memblock.memory,SZ_1G -SZ_1M ,SZ_1M )); +ASSERT_FALSE (memblock_overlaps_region (&memblock.memory,SZ_1G +SZ_4M ,SZ_1M )); + +/* Partial Overlap */ +ASSERT_TRUE (memblock_overlaps_region (&memblock.memory,SZ_1G -SZ_1M ,SZ_2M )); +ASSERT_TRUE (memblock_overlaps_region (&memblock.memory,SZ_1G +SZ_2M ,SZ_2M )); + +/* Totally Overlap */ +ASSERT_TRUE (memblock_overlaps_region (&memblock.memory,SZ_1G ,SZ_4M )); +ASSERT_TRUE (memblock_overlaps_region (&memblock.memory,SZ_1G -SZ_2M ,SZ_8M )); +ASSERT_TRUE (memblock_overlaps_region (&memblock.memory,SZ_1G +SZ_1M ,SZ_1M )); + +test_pass_pop (); + +return 0 ; +} + +static int memblock_overlaps_region_checks (void ) +{ +prefix_reset (); +prefix_push ("memblock_overlaps_region" ); +test_print ("Running memblock_overlaps_region tests...\n" ); + +memblock_overlaps_region_check (); + +prefix_pop (); + +return 0 ; +} +int memblock_basic_checks (void ) {memblock_initialization_check ();
@@ -2270,6 +2317,7 @@ int memblock_basic_checks(void)memblock_free_checks ();memblock_bottom_up_checks ();memblock_trim_memory_checks (); +memblock_overlaps_region_checks ();return 0 ; }
tools/testing/memblock/tests/common.h +3 -0
@@ -40,6 +40,9 @@ enum test_flags {assert ((_expected) == (_seen)); \ } while (0) +#define ASSERT_TRUE (_seen) ASSERT_EQ(true, _seen) +#define ASSERT_FALSE (_seen) ASSERT_EQ(false, _seen) +/** * ASSERT_NE(): * Check the condition
Signature · 1 line
2.34.1
PATCH v2 6/8 mm/memblock: return true directly on finding overlap region
In reply to Wei YangNot necessary to break and check i against type->cnt again.
--- mm/memblock.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
mm/memblock.c +2 -2
@@ -194,8 +194,8 @@ bool __init_memblock memblock_overlaps_region(struct memblock_type *type,for (i =0 ; i < type->cnt; i++)if (memblock_addrs_overlap (base, size, type->regions[i].base, type->regions[i].size)) -break ; -return i < type->cnt; +return true; +return false; }/**
Signature · 1 line
2.34.1
PATCH v2 7/8 mm/memblock: use PAGE_ALIGN_DOWN to get pgend in free_memmap
In reply to Wei YangLeverage the macro PAGE_ALIGN_DOWN to get pgend.
--- mm/memblock.c | 2 +- tools/include/linux/mm.h | 1 + 2 files changed, 2 insertions(+), 1 deletion(-)
mm/memblock.c +1 -1
@@ -2046,7 +2046,7 @@ static void __init free_memmap(unsigned long start_pfn, unsigned long end_pfn)* downwards. */ pg =PAGE_ALIGN (__pa (start_pg)); - pgend =__pa (end_pg) &PAGE_MASK ; + pgend =PAGE_ALIGN_DOWN (__pa (end_pg));/* * If there are free pages between these, free the section of the
tools/include/linux/mm.h +1 -0
@@ -15,6 +15,7 @@#define ALIGN_DOWN (x, a) __ALIGN_KERNEL((x) - ((a) - 1), (a))#define PAGE_ALIGN (addr) ALIGN(addr, PAGE_SIZE) +#define PAGE_ALIGN_DOWN (addr) ALIGN_DOWN(addr, PAGE_SIZE)#define __va (x) ((void *)((unsigned long)(x)))#define __pa (x) ((unsigned long)(x))
Signature · 1 line
2.34.1
PATCH v2 8/8 mm/memblock: default region's nid may be MAX_NUMNODES
In reply to Wei YangOn x86, the call flow looks like this:
numa_init()
memblock_set_node(..., MAX_NUMNODES)
numa_register_memblks()
memblock_validate_numa_coverage()
If there is a hole, the nid for this region would stay to be MAX_NUMNODES. Then memblock_validate_numa_coverage() will miss to report it.
--- mm/memblock.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
mm/memblock.c +1 -1
@@ -759,7 +759,7 @@ bool __init_memblock memblock_validate_numa_coverage(unsigned long threshold_byt/* calculate lose page */ for_each_mem_pfn_range (i,MAX_NUMNODES , &start_pfn, &end_pfn, &nid) { -if (nid ==NUMA_NO_NODE ) +if (nid ==MAX_NUMNODES || nid ==NUMA_NO_NODE ) nr_pages += end_pfn - start_pfn; }
Signature · 1 line
2.34.1
A message that is not in the archive (sent off-list or not yet received): [email protected]
Mike Rapoport
maintainerRe: [Patch v2 2/8] memblock tests: add memblock_reserve_many_may_conflict_check()
Quoted 18 lines · Wei Yang
This may trigger the case fixed by commit
48c3b583bbdd("mm/memblock:
fix overlapping allocation when doubling reserved array").This is done by adding the 129th reserve region into memblock.memory. If
memblock_double_array()use this reserve region as new array, it fails.Signed-off-by Wei Yang <[email protected]>--- tools/testing/memblock/tests/basic_api.c | 123 +++++++++++++++++++++++ tools/testing/memblock/tests/common.c | 4 +- tools/testing/memblock/tests/common.h | 1 + 3 files changed, 126 insertions(+), 2 deletions(-)tools/testing/memblock/tests/basic_api.c +0
-0
tools/testing/memblock/tests/basic_api.c static int memblock_reserve_many_check(void) · line 994
Quoted 4 lines
tools/testing/memblock/tests/basic_api.c +1
-0@@ -991,6 +991,128 @@ static int memblock_reserve_many_check(void)return 0 ; } +/* Keep the gap so these memory region will not be merged. */
The gap where? What regions should not be merged? Also please add a comment with the test description
Quoted 11 lines
tools/testing/memblock/tests/basic_api.c +11
-0+#define MEMORY_BASE_OFFSET (idx, offset) ((offset) + (MEM_SIZE * 2) * (idx)) +static int memblock_reserve_many_may_conflict_check (void ) +{ +int i, skip; +void *orig_region; +struct region r = { + .base =SZ_16K , + .size =SZ_16K , + }; +phys_addr_t new_reserved_regions_size; +
tools/testing/memblock/tests/basic_api.c static int memblock_reserve_many_check(void) · line 1015
Quoted 10 lines
tools/testing/memblock/tests/basic_api.c +10
-0+/* +* 0 1 129 +* +---+ +---+ +---+ +* |32K| |32K| .. |32K| +* +---+ +---+ +---+ +* +* Pre-allocate the range for 129 memory block + one range for double +* memblock.reserved.regions at idx 0. +* See commit 48c3b583bbdd ("mm/memblock: fix overlapping allocation +* when doubling reserved array")
Sorry, but I'm failing to parse it
Quoted 116 lines
tools/testing/memblock/tests/basic_api.c +101
-0+*/ +dummy_physical_memory_init (); +phys_addr_t memory_base =dummy_physical_memory_base (); +phys_addr_t offset =PAGE_ALIGN (memory_base); + +PREFIX_PUSH (); + +/* Reserve the 129th memory block for all possible positions*/ +for (skip =1 ; skip <=INIT_MEMBLOCK_REGIONS +1 ; skip++) { +reset_memblock_regions (); +memblock_allow_resize (); + +reset_memblock_attributes (); +/* Add a valid memory region used by double_array(). */ +memblock_add (MEMORY_BASE_OFFSET (0 , offset),MEM_SIZE ); +/* +* Add a memory region which will be reserved as 129th memory +* region. This is not expected to be used by double_array(). +*/ +memblock_add (MEMORY_BASE_OFFSET (skip, offset),MEM_SIZE ); + +for (i =1 ; i <=INIT_MEMBLOCK_REGIONS +1 ; i++) { +if (i == skip) +continue ; + +/* Reserve some fakes memory region to fulfill the memblock. */ +memblock_reserve (MEMORY_BASE_OFFSET (i, offset),MEM_SIZE ); + +if (i < skip) { +ASSERT_EQ (memblock.reserved.cnt, i); +ASSERT_EQ (memblock.reserved.total_size, i *MEM_SIZE ); + }else { +ASSERT_EQ (memblock.reserved.cnt, i -1 ); +ASSERT_EQ (memblock.reserved.total_size, (i -1 ) *MEM_SIZE ); + } + } + + orig_region = memblock.reserved.regions; + +/* This reserve the 129 memory_region, and makes it double array. */ +memblock_reserve (MEMORY_BASE_OFFSET (skip, offset),MEM_SIZE ); + +/* +* This is the memory region size used by the doubled reserved.regions, +* and it has been reserved due to it has been used. The size is used to +* calculate the total_size that the memblock.reserved have now. +*/ + new_reserved_regions_size =PAGE_ALIGN ((INIT_MEMBLOCK_REGIONS *2 ) * +sizeof (struct memblock_region )); +/* +* The double_array() will find a free memory region as the new +* reserved.regions, and the used memory region will be reserved, so +* there will be one more region exist in the reserved memblock. And the +* one more reserved region's size is new_reserved_regions_size. +*/ +ASSERT_EQ (memblock.reserved.cnt,INIT_MEMBLOCK_REGIONS +2 ); +ASSERT_EQ (memblock.reserved.total_size, (INIT_MEMBLOCK_REGIONS +1 ) *MEM_SIZE + + new_reserved_regions_size); +ASSERT_EQ (memblock.reserved.max,INIT_MEMBLOCK_REGIONS *2 ); + +/* +* The first reserved region is allocated for double array +* with the size of new_reserved_regions_size and the base to be +* MEMORY_BASE_OFFSET(0, offset) + SZ_32K - new_reserved_regions_size +*/ +ASSERT_EQ (memblock.reserved.regions[0 ].base + memblock.reserved.regions[0 ].size, +MEMORY_BASE_OFFSET (0 , offset) +SZ_32K ); +ASSERT_EQ (memblock.reserved.regions[0 ].size, new_reserved_regions_size); + +/* +* Now memblock_double_array() works fine. Let's check after the +* double_array(), the memblock_reserve() still works as normal. +*/ +memblock_reserve (r.base, r.size); +ASSERT_EQ (memblock.reserved.regions[0 ].base, r.base); +ASSERT_EQ (memblock.reserved.regions[0 ].size, r.size); + +ASSERT_EQ (memblock.reserved.cnt,INIT_MEMBLOCK_REGIONS +3 ); +ASSERT_EQ (memblock.reserved.total_size, (INIT_MEMBLOCK_REGIONS +1 ) *MEM_SIZE + + new_reserved_regions_size + + r.size); +ASSERT_EQ (memblock.reserved.max,INIT_MEMBLOCK_REGIONS *2 ); + +/* +* The current reserved.regions is occupying a range of memory that +* allocated from dummy_physical_memory_init(). After free the memory, +* we must not use it. So restore the origin memory region to make sure +* the tests can run as normal and not affected by the double array. +*/ + memblock.reserved.regions = orig_region; + memblock.reserved.cnt =INIT_MEMBLOCK_RESERVED_REGIONS ; + } + +dummy_physical_memory_cleanup (); + +test_pass_pop (); + +return 0 ; +} +static int memblock_reserve_checks (void ) {prefix_reset ();@@ -1006,6 +1128,7 @@ static int memblock_reserve_checks(void)memblock_reserve_between_check ();memblock_reserve_near_max_check ();memblock_reserve_many_check (); +memblock_reserve_many_may_conflict_check ();prefix_pop ();tools/testing/memblock/tests/common.c +0
-0
tools/testing/memblock/tests/common.c void reset_memblock_attributes(void) · line 64
Quoted 5 lines
tools/testing/memblock/tests/common.c +1
-1@@ -61,7 +61,7 @@ void reset_memblock_attributes(void)static inline void fill_memblock (void ) { -memset (memory_block.base,1 ,MEM_SIZE ); +memset (memory_block.base,1 ,MEM_ALLOC_SIZE );
I believe PHYS_MEM_SIZE is a better name.
Quoted 17 lines
tools/testing/memblock/tests/common.c +1
-1}void setup_memblock (void )@@ -103,7 +103,7 @@ void setup_numa_memblock(const unsigned int node_fracs[])void dummy_physical_memory_init (void ) { - memory_block.base =malloc (MEM_SIZE ); + memory_block.base =malloc (MEM_ALLOC_SIZE );assert (memory_block.base);fill_memblock (); }tools/testing/memblock/tests/common.h +0
-0
tools/testing/memblock/tests/common.h line 15
Quoted 4 lines
tools/testing/memblock/tests/common.h +1
-0@@ -12,6 +12,7 @@#include <../selftests/kselftest.h> #define MEM_SIZE SZ_32K +#define MEM_ALLOC_SIZE SZ_16M
Do we really need 16M? Wouldn't one or two suffice?
Quoted 6 lines
tools/testing/memblock/tests/common.h +0
-0#define NUMA_NODES 8#define INIT_MEMBLOCK_REGIONS 128Signature · 1 line
2.34.1
Signature · 2 lines
Sincerely yours, Mike.
Re: [Patch v2 2/8] memblock tests: add memblock_reserve_many_may_conflict_check()
Quoted 22 lines · Mike Rapoport
Quoted 21 lines · Wei Yang
This may trigger the case fixed by commit
48c3b583bbdd("mm/memblock:
fix overlapping allocation when doubling reserved array").This is done by adding the 129th reserve region into memblock.memory. If
memblock_double_array()use this reserve region as new array, it fails.Signed-off-by Wei Yang <[email protected]>--- tools/testing/memblock/tests/basic_api.c | 123 +++++++++++++++++++++++ tools/testing/memblock/tests/common.c | 4 +- tools/testing/memblock/tests/common.h | 1 + 3 files changed, 126 insertions(+), 2 deletions(-)tools/testing/memblock/tests/basic_api.c +0
-0@@ -991,6 +991,128 @@ static int memblock_reserve_many_check(void)return 0 ; }
Quoted 3 lines
tools/testing/memblock/tests/basic_api.c static int memblock_reserve_many_check(void) · line 994
Quoted 1 line
tools/testing/memblock/tests/basic_api.c +1
-0+/* Keep the gap so these memory region will not be merged. */ The gap where? What regions should not be merged?
We would reserve range [1-129] like below. The gap is between each range, which is 32K, to prevent merge during memblock_reserve().
0 1 129
+---+ +---+ +---+
|32K| |32K| .. |32K|
+---+ +---+ +---+
|gap |
|<-->|Quoted 1 line
Also please add a comment with the test description
Sure.
Quoted 142 lines
Quoted 12 lines
tools/testing/memblock/tests/basic_api.c +11
-0+#define MEMORY_BASE_OFFSET (idx, offset) ((offset) + (MEM_SIZE * 2) * (idx)) +static int memblock_reserve_many_may_conflict_check (void ) +{ +int i, skip; +void *orig_region; +struct region r = { + .base =SZ_16K , + .size =SZ_16K , + }; +phys_addr_t new_reserved_regions_size; +tools/testing/memblock/tests/basic_api.c static int memblock_reserve_many_check(void) · line 1015
Quoted 10 lines
tools/testing/memblock/tests/basic_api.c +10
-0+/* +* 0 1 129 +* +---+ +---+ +---+ +* |32K| |32K| .. |32K| +* +---+ +---+ +---+ +* +* Pre-allocate the range for 129 memory block + one range for double +* memblock.reserved.regions at idx 0. +* See commit 48c3b583bbdd ("mm/memblock: fix overlapping allocation +* when doubling reserved array") Sorry, but I'm failing to parse it
Quoted 117 lines
tools/testing/memblock/tests/basic_api.c +101
-0+*/ +dummy_physical_memory_init (); +phys_addr_t memory_base =dummy_physical_memory_base (); +phys_addr_t offset =PAGE_ALIGN (memory_base); + +PREFIX_PUSH (); + +/* Reserve the 129th memory block for all possible positions*/ +for (skip =1 ; skip <=INIT_MEMBLOCK_REGIONS +1 ; skip++) { +reset_memblock_regions (); +memblock_allow_resize (); + +reset_memblock_attributes (); +/* Add a valid memory region used by double_array(). */ +memblock_add (MEMORY_BASE_OFFSET (0 , offset),MEM_SIZE ); +/* +* Add a memory region which will be reserved as 129th memory +* region. This is not expected to be used by double_array(). +*/ +memblock_add (MEMORY_BASE_OFFSET (skip, offset),MEM_SIZE ); + +for (i =1 ; i <=INIT_MEMBLOCK_REGIONS +1 ; i++) { +if (i == skip) +continue ; + +/* Reserve some fakes memory region to fulfill the memblock. */ +memblock_reserve (MEMORY_BASE_OFFSET (i, offset),MEM_SIZE ); + +if (i < skip) { +ASSERT_EQ (memblock.reserved.cnt, i); +ASSERT_EQ (memblock.reserved.total_size, i *MEM_SIZE ); + }else { +ASSERT_EQ (memblock.reserved.cnt, i -1 ); +ASSERT_EQ (memblock.reserved.total_size, (i -1 ) *MEM_SIZE ); + } + } + + orig_region = memblock.reserved.regions; + +/* This reserve the 129 memory_region, and makes it double array. */ +memblock_reserve (MEMORY_BASE_OFFSET (skip, offset),MEM_SIZE ); + +/* +* This is the memory region size used by the doubled reserved.regions, +* and it has been reserved due to it has been used. The size is used to +* calculate the total_size that the memblock.reserved have now. +*/ + new_reserved_regions_size =PAGE_ALIGN ((INIT_MEMBLOCK_REGIONS *2 ) * +sizeof (struct memblock_region )); +/* +* The double_array() will find a free memory region as the new +* reserved.regions, and the used memory region will be reserved, so +* there will be one more region exist in the reserved memblock. And the +* one more reserved region's size is new_reserved_regions_size. +*/ +ASSERT_EQ (memblock.reserved.cnt,INIT_MEMBLOCK_REGIONS +2 ); +ASSERT_EQ (memblock.reserved.total_size, (INIT_MEMBLOCK_REGIONS +1 ) *MEM_SIZE + + new_reserved_regions_size); +ASSERT_EQ (memblock.reserved.max,INIT_MEMBLOCK_REGIONS *2 ); + +/* +* The first reserved region is allocated for double array +* with the size of new_reserved_regions_size and the base to be +* MEMORY_BASE_OFFSET(0, offset) + SZ_32K - new_reserved_regions_size +*/ +ASSERT_EQ (memblock.reserved.regions[0 ].base + memblock.reserved.regions[0 ].size, +MEMORY_BASE_OFFSET (0 , offset) +SZ_32K ); +ASSERT_EQ (memblock.reserved.regions[0 ].size, new_reserved_regions_size); + +/* +* Now memblock_double_array() works fine. Let's check after the +* double_array(), the memblock_reserve() still works as normal. +*/ +memblock_reserve (r.base, r.size); +ASSERT_EQ (memblock.reserved.regions[0 ].base, r.base); +ASSERT_EQ (memblock.reserved.regions[0 ].size, r.size); + +ASSERT_EQ (memblock.reserved.cnt,INIT_MEMBLOCK_REGIONS +3 ); +ASSERT_EQ (memblock.reserved.total_size, (INIT_MEMBLOCK_REGIONS +1 ) *MEM_SIZE + + new_reserved_regions_size + + r.size); +ASSERT_EQ (memblock.reserved.max,INIT_MEMBLOCK_REGIONS *2 ); + +/* +* The current reserved.regions is occupying a range of memory that +* allocated from dummy_physical_memory_init(). After free the memory, +* we must not use it. So restore the origin memory region to make sure +* the tests can run as normal and not affected by the double array. +*/ + memblock.reserved.regions = orig_region; + memblock.reserved.cnt =INIT_MEMBLOCK_RESERVED_REGIONS ; + } + +dummy_physical_memory_cleanup (); + +test_pass_pop (); + +return 0 ; +} +static int memblock_reserve_checks (void ) {prefix_reset ();@@ -1006,6 +1128,7 @@ static int memblock_reserve_checks(void)memblock_reserve_between_check ();memblock_reserve_near_max_check ();memblock_reserve_many_check (); +memblock_reserve_many_may_conflict_check ();prefix_pop ();tools/testing/memblock/tests/common.c +0
-0@@ -61,7 +61,7 @@ void reset_memblock_attributes(void)
Quoted 7 lines
tools/testing/memblock/tests/common.c void reset_memblock_attributes(void) · line 64
Quoted 4 lines
tools/testing/memblock/tests/common.c +1
-1static inline void fill_memblock (void ) { -memset (memory_block.base,1 ,MEM_SIZE ); +memset (memory_block.base,1 ,MEM_ALLOC_SIZE );I believe
PHYS_MEM_SIZEis a better name.
Sounds better.
Quoted 19 lines
Quoted 19 lines
tools/testing/memblock/tests/common.c +1
-1}void setup_memblock (void )@@ -103,7 +103,7 @@ void setup_numa_memblock(const unsigned int node_fracs[])void dummy_physical_memory_init (void ) { - memory_block.base =malloc (MEM_SIZE ); + memory_block.base =malloc (MEM_ALLOC_SIZE );assert (memory_block.base);fill_memblock (); }tools/testing/memblock/tests/common.h +0
-0@@ -12,6 +12,7 @@#include <../selftests/kselftest.h>
Quoted 5 lines
tools/testing/memblock/tests/common.h line 15
Quoted 2 lines
tools/testing/memblock/tests/common.h +1
-0#define MEM_SIZE SZ_32K +#define MEM_ALLOC_SIZE SZ_16MDo we really need 16M? Wouldn't one or two suffice?
0 1 129 +---+ +---+ +---+ |32K| |32K| .. |32K| +---+ +---+ +---+
This is the range we are manipulating, which is roughly
130 * 64K = 128 * 64K + 128K = 8M + 128K
So I choose the next power of 2, 16M.
Since we insert the 129th range at all possible position, we may allocate the double array at any place when the bug is triggered. So we need to allocate the whole range instead of just allocate range 0 as we expect.
Quoted 10 lines
Quoted 6 lines
tools/testing/memblock/tests/common.h +0
-0#define NUMA_NODES 8#define INIT_MEMBLOCK_REGIONS 128Signature · 1 line
2.34.1Signature · 2 lines
Sincerely yours, Mike.
Signature · 2 lines
Wei Yang Help you, Help me
Re: [Patch v2 2/8] memblock tests: add memblock_reserve_many_may_conflict_check()
In reply to Mike RapoportQuoted 42 lines · Mike Rapoport
Quoted 18 lines · Wei Yang
This may trigger the case fixed by commit
48c3b583bbdd("mm/memblock:
fix overlapping allocation when doubling reserved array").This is done by adding the 129th reserve region into memblock.memory. If
memblock_double_array()use this reserve region as new array, it fails.Signed-off-by Wei Yang <[email protected]>--- tools/testing/memblock/tests/basic_api.c | 123 +++++++++++++++++++++++ tools/testing/memblock/tests/common.c | 4 +- tools/testing/memblock/tests/common.h | 1 + 3 files changed, 126 insertions(+), 2 deletions(-)tools/testing/memblock/tests/basic_api.c +0
-0tools/testing/memblock/tests/basic_api.c static int memblock_reserve_many_check(void) · line 994
Quoted 4 lines
tools/testing/memblock/tests/basic_api.c +1
-0@@ -991,6 +991,128 @@ static int memblock_reserve_many_check(void)return 0 ; } +/* Keep the gap so these memory region will not be merged. */ The gap where? What regions should not be merged? Also please add a comment with the test description
Quoted 15 lines
tools/testing/memblock/tests/basic_api.c +15
-0+#define MEMORY_BASE_OFFSET (idx, offset) ((offset) + (MEM_SIZE * 2) * (idx)) +static int memblock_reserve_many_may_conflict_check (void ) +{ +int i, skip; +void *orig_region; +struct region r = { + .base =SZ_16K , + .size =SZ_16K , + }; +phys_addr_t new_reserved_regions_size; + +/* +* 0 1 129 +* +---+ +---+ +---+ +* |32K| |32K| .. |32K|
Quoted 9 lines
tools/testing/memblock/tests/basic_api.c static int memblock_reserve_many_check(void) · line 1015
Quoted 6 lines
tools/testing/memblock/tests/basic_api.c +6
-0+* +---+ +---+ +---+ +* +* Pre-allocate the range for 129 memory block + one range for double +* memblock.reserved.regions at idx 0. +* See commit 48c3b583bbdd ("mm/memblock: fix overlapping allocation +* when doubling reserved array") Sorry, but I'm failing to parse it
Sorry for missed this one. You mean how this case is triggered?
Suppose current memblock looks like this:
0 1
memblock.memory +---+ +---+
|32K| |32K|
+---+ +---+ 0 2 128
memblock.reserved+---+ +---+ +---+
|32K| |32K| .. |32K|
+---+ +---+ +---+
^
|Now we want to reserve range 1 here. Since there are 128 blocks in memblock.reserved, it will double array. Now memblock_double_array() will "find" available range in memblock.memory.
Since we set top down, it will find memblock.memory.regions[1]. This conflict the range we want to reserve.
Signature · 2 lines
Wei Yang Help you, Help me
Mike Rapoport
maintainerRe: [Patch v2 2/8] memblock tests: add memblock_reserve_many_may_conflict_check()
Quoted 68 lines · Wei Yang
Quoted 42 lines · Mike Rapoport
Quoted 18 lines · Wei Yang
This may trigger the case fixed by commit
48c3b583bbdd("mm/memblock:
fix overlapping allocation when doubling reserved array").This is done by adding the 129th reserve region into memblock.memory. If
memblock_double_array()use this reserve region as new array, it fails.Signed-off-by Wei Yang <[email protected]>--- tools/testing/memblock/tests/basic_api.c | 123 +++++++++++++++++++++++ tools/testing/memblock/tests/common.c | 4 +- tools/testing/memblock/tests/common.h | 1 + 3 files changed, 126 insertions(+), 2 deletions(-)tools/testing/memblock/tests/basic_api.c +0
-0tools/testing/memblock/tests/basic_api.c static int memblock_reserve_many_check(void) · line 994
Quoted 4 lines
tools/testing/memblock/tests/basic_api.c +1
-0@@ -991,6 +991,128 @@ static int memblock_reserve_many_check(void)return 0 ; } +/* Keep the gap so these memory region will not be merged. */ The gap where? What regions should not be merged? Also please add a comment with the test description
Quoted 15 lines
tools/testing/memblock/tests/basic_api.c +15
-0+#define MEMORY_BASE_OFFSET (idx, offset) ((offset) + (MEM_SIZE * 2) * (idx)) +static int memblock_reserve_many_may_conflict_check (void ) +{ +int i, skip; +void *orig_region; +struct region r = { + .base =SZ_16K , + .size =SZ_16K , + }; +phys_addr_t new_reserved_regions_size; + +/* +* 0 1 129 +* +---+ +---+ +---+ +* |32K| |32K| .. |32K| Quoted 9 lines
tools/testing/memblock/tests/basic_api.c static int memblock_reserve_many_check(void) · line 1015
Quoted 6 lines
tools/testing/memblock/tests/basic_api.c +6
-0+* +---+ +---+ +---+ +* +* Pre-allocate the range for 129 memory block + one range for double +* memblock.reserved.regions at idx 0. +* See commit 48c3b583bbdd ("mm/memblock: fix overlapping allocation +* when doubling reserved array") Sorry, but I'm failing to parse it
Sorry for missed this one. You mean how this case is triggered?
Suppose current memblock looks like this:
0 1 memblock.memory +---+ +---+ |32K| |32K| +---+ +---+0 2 128 memblock.reserved+---+ +---+ +---+ |32K| |32K| .. |32K| +---+ +---+ +---+ ^ |
Quoted 6 lines
Now we want to reserve range 1 here. Since there are 128 blocks in memblock.reserved, it will double array. Now
memblock_double_array()will "find" available range in memblock.memory.Since we set top down, it will find memblock.memory.regions[1]. This conflict the range we want to reserve.
Please include something like this explanation in the test description
Quoted 3 lines
Signature · 2 lines
Wei Yang Help you, Help me
Signature · 2 lines
Sincerely yours, Mike.
A message that is not in the archive (sent off-list or not yet received): [email protected]
Mike Rapoport
maintainerRe: [Patch v2 3/8] mm/memblock: fix comment for memblock_isolate_range()
Quoted 15 lines · Wei Yang
The isolated range is [*@start_rgn, *@end_rgn - 1], while the comment says "the end region inside the range" is *@end_rgn.
Let's correct it.
Signed-off-by Wei Yang <[email protected]>--- mm/memblock.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)mm/memblock.c +0
-0
mm/memblock.c bool __init_memblock memblock_validate_numa_coverage(unsigned long threshold_byt · line 775
Quoted 5 lines
mm/memblock.c +1
-1@@ -772,12 +772,12 @@ bool __init_memblock memblock_validate_numa_coverage(unsigned long threshold_byt* @base: base of range to isolate * @size: size of range to isolate * @start_rgn: out parameter for the start of isolated region -* @end_rgn: out parameter for the end of isolated region +* @end_rgn: out parameter for the (end + 1) of isolated region
end can be inclusive or exclusive, please let's keep this line ...
mm/memblock.c bool __init_memblock memblock_validate_numa_coverage(unsigned long threshold_byt · line 780
Quoted 6 lines
mm/memblock.c +1
-1* * Walk @type and ensure that regions don't cross the boundaries defined by * [@base, @base + @size). Crossing regions are split at the boundaries, * which may create at most two more regions. The index of the first -* region inside the range is returned in *@start_rgn and end in *@end_rgn. +* region inside the range is returned in *@start_rgn and (end + 1) in *@end_rgn.
... and emphasise here that end is exclusive, e.g
The index of the first region inside the range is returned in *@start_rgn and the index of the first region after the range is returned in *@end_rgn.
Quoted 6 lines
mm/memblock.c +0
-0* * Return: * 0 on success, -errno on failure. Signature · 1 line
2.34.1
Signature · 2 lines
Sincerely yours, Mike.
Re: [Patch v2 3/8] mm/memblock: fix comment for memblock_isolate_range()
Quoted 31 lines · Mike Rapoport
Quoted 15 lines · Wei Yang
The isolated range is [*@start_rgn, *@end_rgn - 1], while the comment says "the end region inside the range" is *@end_rgn.
Let's correct it.
Signed-off-by Wei Yang <[email protected]>--- mm/memblock.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)mm/memblock.c +0
-0mm/memblock.c bool __init_memblock memblock_validate_numa_coverage(unsigned long threshold_byt · line 775
Quoted 5 lines
mm/memblock.c +1
-1@@ -772,12 +772,12 @@ bool __init_memblock memblock_validate_numa_coverage(unsigned long threshold_byt* @base: base of range to isolate * @size: size of range to isolate * @start_rgn: out parameter for the start of isolated region -* @end_rgn: out parameter for the end of isolated region +* @end_rgn: out parameter for the (end + 1) of isolated region end can be inclusive or exclusive, please let's keep this line ...
mm/memblock.c bool __init_memblock memblock_validate_numa_coverage(unsigned long threshold_byt · line 780
Quoted 6 lines
mm/memblock.c +1
-1* * Walk @type and ensure that regions don't cross the boundaries defined by * [@base, @base + @size). Crossing regions are split at the boundaries, * which may create at most two more regions. The index of the first -* region inside the range is returned in *@start_rgn and end in *@end_rgn. +* region inside the range is returned in *@start_rgn and (end + 1) in *@end_rgn.
Quoted 5 lines
... and emphasise here that end is exclusive, e.g
The index of the first region inside the range is returned in *@start_rgn and the index of the first region after the range is returned in *@end_rgn.
Looks better, thanks.
Quoted 10 lines
Quoted 6 lines
mm/memblock.c +0
-0* * Return: * 0 on success, -errno on failure. Signature · 1 line
2.34.1Signature · 2 lines
Sincerely yours, Mike.
Signature · 2 lines
Wei Yang Help you, Help me