arm64: drop pfn_valid_within() and simplify pfn_valid()
Mike Rapoport
maintainercover arm64: drop pfn_valid_within() and simplify pfn_valid()
From: Mike Rapoport <[email protected]>
Hi,
These patches aim to remove CONFIG_HOLES_IN_ZONE and essentially hardwire pfn_valid_within() to 1.
The idea is to mark NOMAP pages as reserved in the memory map and restore the intended semantics of pfn_valid() to designate availability of struct page for a pfn.
With this the core mm will be able to cope with the fact that it cannot use NOMAP pages and the holes created by NOMAP ranges within MAX_ORDER blocks will be treated correctly even without the need for pfn_valid_within.
The patches are only boot tested on qemu-system-aarch64 so I'd really appreciate memory stress tests on real hardware.
If this actually works we'll be one step closer to drop custom pfn_valid() on arm64 altogether.
Mike Rapoport (3):
memblock: update initialization of reserved pages
arm64: decouple check whether pfn is normal memory from pfn_valid()
arm64: drop pfn_valid_within() and simplify pfn_valid()
arch/arm64/Kconfig | 3 --- arch/arm64/include/asm/memory.h | 2 +- arch/arm64/include/asm/page.h | 1 + arch/arm64/kvm/mmu.c | 2 +- arch/arm64/mm/init.c | 10 ++++++++-- arch/arm64/mm/ioremap.c | 4 ++-- arch/arm64/mm/mmu.c | 2 +- mm/memblock.c | 23 +++++++++++++++++++++-- 8 files changed, 35 insertions(+), 12 deletions(-)
base-commit: e49d033bddf5b565044e2abe4241353959bc9120
Signature · 1 line
2.28.0
Mike Rapoport
maintainerPATCH 1/3 memblock: update initialization of reserved pages
From: Mike Rapoport <[email protected]>
The struct pages representing a reserved memory region are initialized using reserve_bootmem_range() function. This function is called for each reserved region just before the memory is freed from memblock to the buddy page allocator.
The struct pages for MEMBLOCK_NOMAP regions are kept with the default values set by the memory map initialization which makes it necessary to have a special treatment for such pages in pfn_valid() and pfn_valid_within().
Split out initialization of the reserved pages to a function with a meaningful name and treat the MEMBLOCK_NOMAP regions the same way as the reserved regions and mark struct pages for the NOMAP regions as PageReserved.
--- mm/memblock.c | 23 +++++++++++++++++++++-- 1 file changed, 21 insertions(+), 2 deletions(-)
mm/memblock.c +21 -2
@@ -2002,6 +2002,26 @@ static unsigned long __init __free_memory_core(phys_addr_t start,return end_pfn - start_pfn; } +static void __init memmap_init_reserved_pages (void ) +{ +struct memblock_region *region; +phys_addr_t start, end; +u64 i; + +/* initialize struct pages for the reserved regions */ +for_each_reserved_mem_range (i, &start, &end) +reserve_bootmem_region (start, end); + +/* and also treat struct pages for the NOMAP regions as PageReserved */ +for_each_mem_region (region) { +if (memblock_is_nomap (region)) { + start = region->base; + end = start + region->size; +reserve_bootmem_region (start, end); + } + } +} +static unsigned long __init free_low_memory_core_early (void ) {unsigned long count =0 ;
@@ -2010,8 +2030,7 @@ static unsigned long __init free_low_memory_core_early(void)memblock_clear_hotplug (0 ,-1 ); -for_each_reserved_mem_range (i, &start, &end) -reserve_bootmem_region (start, end); +memmap_init_reserved_pages ();/* * We need to use NUMA_NO_NODE instead of NODE_DATA(0)->node_id
Signature · 1 line
2.28.0
Re: [RFC/RFT PATCH 1/3] memblock: update initialization of reserved pages
Quoted 12 lines · Mike Rapoport
From: Mike Rapoport <[email protected]>
The struct pages representing a reserved memory region are initialized using
reserve_bootmem_range()function. This function is called for each reserved region just before the memory is freed from memblock to the buddy page allocator.The struct pages for
MEMBLOCK_NOMAPregions are kept with the default values set by the memory map initialization which makes it necessary to have a special treatment for such pages inpfn_valid()andpfn_valid_within().
Quoted 4 lines
Split out initialization of the reserved pages to a function with a meaningful name and treat the
MEMBLOCK_NOMAPregions the same way as the reserved regions and mark struct pages for the NOMAP regions as PageReserved.
This would definitely need updating the comment for MEMBLOCK_NOMAP definition in include/linux/memblock.h just to make the semantics is clear, though arm64 is currently the only user for MEMBLOCK_NOMAP.
Quoted 48 lines
Signed-off-by Mike Rapoport <[email protected]>--- mm/memblock.c | 23 +++++++++++++++++++++-- 1 file changed, 21 insertions(+), 2 deletions(-)mm/memblock.c +21
-2@@ -2002,6 +2002,26 @@ static unsigned long __init __free_memory_core(phys_addr_t start,return end_pfn - start_pfn; } +static void __init memmap_init_reserved_pages (void ) +{ +struct memblock_region *region; +phys_addr_t start, end; +u64 i; + +/* initialize struct pages for the reserved regions */ +for_each_reserved_mem_range (i, &start, &end) +reserve_bootmem_region (start, end); + +/* and also treat struct pages for the NOMAP regions as PageReserved */ +for_each_mem_region (region) { +if (memblock_is_nomap (region)) { + start = region->base; + end = start + region->size; +reserve_bootmem_region (start, end); + } + } +} +static unsigned long __init free_low_memory_core_early (void ) {unsigned long count =0 ;@@ -2010,8 +2030,7 @@ static unsigned long __init free_low_memory_core_early(void)memblock_clear_hotplug (0 ,-1 ); -for_each_reserved_mem_range (i, &start, &end) -reserve_bootmem_region (start, end); +memmap_init_reserved_pages ();/* * We need to use NUMA_NO_NODE instead of NODE_DATA(0)->node_id
Mike Rapoport
maintainerRe: [RFC/RFT PATCH 1/3] memblock: update initialization of reserved pages
Quoted 15 lines · Anshuman Khandual
Quoted 12 lines · Mike Rapoport
From: Mike Rapoport <[email protected]>
The struct pages representing a reserved memory region are initialized using
reserve_bootmem_range()function. This function is called for each reserved region just before the memory is freed from memblock to the buddy page allocator.The struct pages for
MEMBLOCK_NOMAPregions are kept with the default values set by the memory map initialization which makes it necessary to have a special treatment for such pages inpfn_valid()andpfn_valid_within().
Quoted 7 lines
Quoted 4 lines
Split out initialization of the reserved pages to a function with a meaningful name and treat the
MEMBLOCK_NOMAPregions the same way as the reserved regions and mark struct pages for the NOMAP regions as PageReserved.This would definitely need updating the comment for
MEMBLOCK_NOMAPdefinition ininclude/linux/memblock.hjust to make the semantics is clear,
Sure
Quoted 49 lines
though arm64 is currently the only user for
MEMBLOCK_NOMAP.Quoted 47 lines
Signed-off-by Mike Rapoport <[email protected]>--- mm/memblock.c | 23 +++++++++++++++++++++-- 1 file changed, 21 insertions(+), 2 deletions(-)mm/memblock.c +21
-2@@ -2002,6 +2002,26 @@ static unsigned long __init __free_memory_core(phys_addr_t start,return end_pfn - start_pfn; } +static void __init memmap_init_reserved_pages (void ) +{ +struct memblock_region *region; +phys_addr_t start, end; +u64 i; + +/* initialize struct pages for the reserved regions */ +for_each_reserved_mem_range (i, &start, &end) +reserve_bootmem_region (start, end); + +/* and also treat struct pages for the NOMAP regions as PageReserved */ +for_each_mem_region (region) { +if (memblock_is_nomap (region)) { + start = region->base; + end = start + region->size; +reserve_bootmem_region (start, end); + } + } +} +static unsigned long __init free_low_memory_core_early (void ) {unsigned long count =0 ;@@ -2010,8 +2030,7 @@ static unsigned long __init free_low_memory_core_early(void)memblock_clear_hotplug (0 ,-1 ); -for_each_reserved_mem_range (i, &start, &end) -reserve_bootmem_region (start, end); +memmap_init_reserved_pages ();/* * We need to use NUMA_NO_NODE instead of NODE_DATA(0)->node_id
Signature · 2 lines
Sincerely yours, Mike.
Re: [RFC/RFT PATCH 1/3] memblock: update initialization of reserved pages
In reply to Mike RapoportQuoted 11 lines · Mike Rapoport
From: Mike Rapoport <[email protected]>
The struct pages representing a reserved memory region are initialized using
reserve_bootmem_range()function. This function is called for each reserved region just before the memory is freed from memblock to the buddy page allocator.The struct pages for
MEMBLOCK_NOMAPregions are kept with the default values set by the memory map initialization which makes it necessary to have a special treatment for such pages inpfn_valid()andpfn_valid_within().
I assume these pages are never given to the buddy, because we don't have a direct mapping. So to the kernel, it's essentially just like a memory hole with benefits.
I can spot that we want to export such memory like any special memory thingy/hole in /proc/iomem -- "reserved", which makes sense.
I would assume that MEMBLOCK_NOMAP is a special type of *reserved* memory. IOW, that for_each_reserved_mem_range() should already succeed on these as well -- we should mark anything that is MEMBLOCK_NOMAP implicitly as reserved. Or are there valid reasons not to do so? What can anyone do with that memory?
I assume they are pretty much useless for the kernel, right? Like other reserved memory ranges.
Quoted 53 lines
Split out initialization of the reserved pages to a function with a meaningful name and treat the
MEMBLOCK_NOMAPregions the same way as the reserved regions and mark struct pages for the NOMAP regions as PageReserved.Signed-off-by Mike Rapoport <[email protected]>--- mm/memblock.c | 23 +++++++++++++++++++++-- 1 file changed, 21 insertions(+), 2 deletions(-)mm/memblock.c +21
-2@@ -2002,6 +2002,26 @@ static unsigned long __init __free_memory_core(phys_addr_t start,return end_pfn - start_pfn; } +static void __init memmap_init_reserved_pages (void ) +{ +struct memblock_region *region; +phys_addr_t start, end; +u64 i; + +/* initialize struct pages for the reserved regions */ +for_each_reserved_mem_range (i, &start, &end) +reserve_bootmem_region (start, end); + +/* and also treat struct pages for the NOMAP regions as PageReserved */ +for_each_mem_region (region) { +if (memblock_is_nomap (region)) { + start = region->base; + end = start + region->size; +reserve_bootmem_region (start, end); + } + } +} +static unsigned long __init free_low_memory_core_early (void ) {unsigned long count =0 ;@@ -2010,8 +2030,7 @@ static unsigned long __init free_low_memory_core_early(void)memblock_clear_hotplug (0 ,-1 ); -for_each_reserved_mem_range (i, &start, &end) -reserve_bootmem_region (start, end); +memmap_init_reserved_pages ();/* * We need to use NUMA_NO_NODE instead of NODE_DATA(0)->node_id
Signature · 3 lines
Thanks, David / dhildenb
Re: [RFC/RFT PATCH 1/3] memblock: update initialization of reserved pages
Quoted 21 lines · David Hildenbrand
Quoted 11 lines · Mike Rapoport
From: Mike Rapoport <[email protected]>
The struct pages representing a reserved memory region are initialized using
reserve_bootmem_range()function. This function is called for each reserved region just before the memory is freed from memblock to the buddy page allocator.The struct pages for
MEMBLOCK_NOMAPregions are kept with the default values set by the memory map initialization which makes it necessary to have a special treatment for such pages inpfn_valid()andpfn_valid_within().I assume these pages are never given to the buddy, because we don't have a direct mapping. So to the kernel, it's essentially just like a memory hole with benefits.
I can spot that we want to export such memory like any special memory thingy/hole in
/proc/iomem-- "reserved", which makes sense.
Quoted 9 lines
I would assume that
MEMBLOCK_NOMAPis a special type of *reserved* memory. IOW, thatfor_each_reserved_mem_range()should already succeed on these as well -- we should mark anything that isMEMBLOCK_NOMAPimplicitly as reserved. Or are there valid reasons not to do so? What can anyone do with that memory?I assume they are pretty much useless for the kernel, right? Like other reserved memory ranges.
On ARM, we need to know whether any physical regions that do not contain system memory contain something with device semantics or not. One of the examples is ACPI tables: these are in reserved memory, and so they are not covered by the linear region. However, when the ACPI core ioremap()s an arbitrary memory region, we don't know whether it is mapping a memory region or a device region unless we keep track of this in some way. (Device mappings require device attributes, but firmware tables require memory attributes, as they might be accessed using misaligned reads)
Quoted 66 lines
Quoted 45 lines
Split out initialization of the reserved pages to a function with a meaningful name and treat the
MEMBLOCK_NOMAPregions the same way as the reserved regions and mark struct pages for the NOMAP regions as PageReserved.Signed-off-by Mike Rapoport <[email protected]>--- mm/memblock.c | 23 +++++++++++++++++++++-- 1 file changed, 21 insertions(+), 2 deletions(-)mm/memblock.c +20
-0@@ -2002,6 +2002,26 @@ static unsigned long __init __free_memory_core(phys_addr_t start,return end_pfn - start_pfn; } +static void __init memmap_init_reserved_pages (void ) +{ +struct memblock_region *region; +phys_addr_t start, end; +u64 i; + +/* initialize struct pages for the reserved regions */ +for_each_reserved_mem_range (i, &start, &end) +reserve_bootmem_region (start, end); + +/* and also treat struct pages for the NOMAP regions as PageReserved */ +for_each_mem_region (region) { +if (memblock_is_nomap (region)) { + start = region->base; + end = start + region->size; +reserve_bootmem_region (start, end); + } + } +} +static unsigned long __init free_low_memory_core_early (void ) {unsigned long count =0 ;mm/memblock.c static unsigned long __init free_low_memory_core_early(void) · line 2036
Quoted 9 lines
mm/memblock.c +1
-2@@ -2010,8 +2030,7 @@ static unsigned long __init free_low_memory_core_early(void)memblock_clear_hotplug (0 ,-1 ); -for_each_reserved_mem_range (i, &start, &end) -reserve_bootmem_region (start, end); +memmap_init_reserved_pages ();/* * We need to use NUMA_NO_NODE instead of NODE_DATA(0)->node_id --
Thanks,David / dhildenb
_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Re: [RFC/RFT PATCH 1/3] memblock: update initialization of reserved pages
Quoted 32 lines · Ard Biesheuvel
Quoted 21 lines · David Hildenbrand
Quoted 11 lines · Mike Rapoport
From: Mike Rapoport <[email protected]>
The struct pages representing a reserved memory region are initialized using
reserve_bootmem_range()function. This function is called for each reserved region just before the memory is freed from memblock to the buddy page allocator.The struct pages for
MEMBLOCK_NOMAPregions are kept with the default values set by the memory map initialization which makes it necessary to have a special treatment for such pages inpfn_valid()andpfn_valid_within().I assume these pages are never given to the buddy, because we don't have a direct mapping. So to the kernel, it's essentially just like a memory hole with benefits.
I can spot that we want to export such memory like any special memory thingy/hole in
/proc/iomem-- "reserved", which makes sense.Quoted 9 lines
I would assume that
MEMBLOCK_NOMAPis a special type of *reserved* memory. IOW, thatfor_each_reserved_mem_range()should already succeed on these as well -- we should mark anything that isMEMBLOCK_NOMAPimplicitly as reserved. Or are there valid reasons not to do so? What can anyone do with that memory?I assume they are pretty much useless for the kernel, right? Like other reserved memory ranges.
Quoted 9 lines
On ARM, we need to know whether any physical regions that do not contain system memory contain something with device semantics or not. One of the examples is ACPI tables: these are in reserved memory, and so they are not covered by the linear region. However, when the ACPI core ioremap()s an arbitrary memory region, we don't know whether it is mapping a memory region or a device region unless we keep track of this in some way. (Device mappings require device attributes, but firmware tables require memory attributes, as they might be accessed using misaligned reads)
Using generically sounding NOMAP ("don't create direct mapping") to identify device regions feels like a hack. I know, it was introduced just for that purpose.
Looking at memblock_mark_nomap(), we consider "device regions"
1) ACPI tables
2) VIDEO_TYPE_EFI memory
3) some device-tree regions in of/fdt.c
IIUC, right now we end up creating a memmap for this NOMAP memory, but hide it away in pfn_valid(). This patch set at least fixes that.
Assuming these pages are never mapped to user space via the struct page (which better be the case), we could further use a new pagetype to mark these pages in a special way, such that we can identify them directly via pfn_to_page().
Then, we could mostly avoid having to query memblock at runtime to figure out that this is special memory. This would obviously be an extension to this series. Just a thought.
Signature · 3 lines
Thanks, David / dhildenb
Mike Rapoport
maintainerRe: [RFC/RFT PATCH 1/3] memblock: update initialization of reserved pages
Quoted 51 lines · David Hildenbrand
Quoted 32 lines · Ard Biesheuvel
Quoted 21 lines · David Hildenbrand
Quoted 11 lines · Mike Rapoport
From: Mike Rapoport <[email protected]>
The struct pages representing a reserved memory region are initialized using
reserve_bootmem_range()function. This function is called for each reserved region just before the memory is freed from memblock to the buddy page allocator.The struct pages for
MEMBLOCK_NOMAPregions are kept with the default values set by the memory map initialization which makes it necessary to have a special treatment for such pages inpfn_valid()andpfn_valid_within().I assume these pages are never given to the buddy, because we don't have a direct mapping. So to the kernel, it's essentially just like a memory hole with benefits.
I can spot that we want to export such memory like any special memory thingy/hole in
/proc/iomem-- "reserved", which makes sense.Quoted 9 lines
I would assume that
MEMBLOCK_NOMAPis a special type of *reserved* memory. IOW, thatfor_each_reserved_mem_range()should already succeed on these as well -- we should mark anything that isMEMBLOCK_NOMAPimplicitly as reserved. Or are there valid reasons not to do so? What can anyone do with that memory?I assume they are pretty much useless for the kernel, right? Like other reserved memory ranges.
Quoted 9 lines
On ARM, we need to know whether any physical regions that do not contain system memory contain something with device semantics or not. One of the examples is ACPI tables: these are in reserved memory, and so they are not covered by the linear region. However, when the ACPI core ioremap()s an arbitrary memory region, we don't know whether it is mapping a memory region or a device region unless we keep track of this in some way. (Device mappings require device attributes, but firmware tables require memory attributes, as they might be accessed using misaligned reads)
Using generically sounding NOMAP ("don't create direct mapping") to identify device regions feels like a hack. I know, it was introduced just for that purpose.
Looking at
memblock_mark_nomap(), we consider "device regions"1) ACPI tables
Quoted 7 lines
2)
VIDEO_TYPE_EFImemory3) some device-tree regions in
of/fdt.cIIUC, right now we end up creating a memmap for this NOMAP memory, but hide it away in
pfn_valid(). This patch set at least fixes that.
Currently we have memmap entries with struct page set to defaults for the NOMAP memory. AFAIU hiding them in pfn_valid()/pfn_valid_within() was a solution to failures in pfn walkers that presumed that for a pfn_valid() there will be a struct page that really reflects the state of that page.
Quoted 4 lines
Assuming these pages are never mapped to user space via the struct page (which better be the case), we could further use a new pagetype to mark these pages in a special way, such that we can identify them directly via
pfn_to_page().
Not sure we really need a new pagetype here, PG_Reserved seems to be quite enough to say "don't touch this". I generally agree that we could make PG_Reserved a PageType and then have several sub-types for reserved memory. This definitely will add clarity but I'm not sure that this justifies amount of churn and effort required to audit uses of PageResrved().
Quoted 3 lines
Then, we could mostly avoid having to query memblock at runtime to figure out that this is special memory. This would obviously be an extension to this series. Just a thought.
Stop pushing memblock out of kernel! ;-)
Now, seriously, we can minimize memblock involvement in run-time and this series in yet another step in that direction.
Signature · 2 lines
Sincerely yours, Mike.
Re: [RFC/RFT PATCH 1/3] memblock: update initialization of reserved pages
Quoted 11 lines
Not sure we really need a new pagetype here,
PG_Reservedseems to be quite enough to say "don't touch this". I generally agree that we could makePG_Reserveda PageType and then have several sub-types for reserved memory. This definitely will add clarity but I'm not sure that this justifies amount of churn and effort required to audit uses ofPageResrved().Quoted 3 lines
Then, we could mostly avoid having to query memblock at runtime to figure out that this is special memory. This would obviously be an extension to this series. Just a thought.
Stop pushing memblock out of kernel! ;-)
Can't stop. Won't stop. :D
It's lovely for booting up a kernel until we have other data-structures in place ;)
Signature · 3 lines
Thanks, David / dhildenb
Mike Rapoport
maintainerRe: [RFC/RFT PATCH 1/3] memblock: update initialization of reserved pages
Quoted 9 lines · David Hildenbrand
Quoted 9 lines
Not sure we really need a new pagetype here,
PG_Reservedseems to be quite enough to say "don't touch this". I generally agree that we could makePG_Reserveda PageType and then have several sub-types for reserved memory. This definitely will add clarity but I'm not sure that this justifies amount of churn and effort required to audit uses ofPageResrved().Quoted 3 lines
Then, we could mostly avoid having to query memblock at runtime to figure out that this is special memory. This would obviously be an extension to this series. Just a thought.
Quoted 6 lines
Quoted 1 line
Stop pushing memblock out of kernel! ;-)
Can't stop. Won't stop. :D
It's lovely for booting up a kernel until we have other data-structures in place ;)
A bit more seriously, we don't have any data structure that reliably represents physical memory layout and arch-independent fashion. memblock is probably the best starting point for eventually having one.
Signature · 2 lines
Sincerely yours, Mike.
Re: [RFC/RFT PATCH 1/3] memblock: update initialization of reserved pages
Quoted 12 lines · Mike Rapoport
Quoted 11 lines · David Hildenbrand
Quoted 10 lines
Not sure we really need a new pagetype here,
PG_Reservedseems to be quite enough to say "don't touch this". I generally agree that we could makePG_Reserveda PageType and then have several sub-types for reserved memory. This definitely will add clarity but I'm not sure that this justifies amount of churn and effort required to audit uses ofPageResrved().Quoted 3 lines
Then, we could mostly avoid having to query memblock at runtime to figure out that this is special memory. This would obviously be an extension to this series. Just a thought.
Stop pushing memblock out of kernel! ;-)
Quoted 8 lines
Quoted 4 lines
Can't stop. Won't stop. :D
It's lovely for booting up a kernel until we have other data-structures in place ;)
A bit more seriously, we don't have any data structure that reliably represents physical memory layout and arch-independent fashion. memblock is probably the best starting point for eventually having one.
We have the (slowish) kernel resource tree after boot and the (faster) memmap. I really don't see why we really need another slowish variant.
We might be better off to just extend and speed up the kernel resource tree.
Memblock as is is not a reasonable datastructure to keep around after boot: for example, how we handle boottime allocations and reserve regions both as reserved.
Signature · 3 lines
Thanks, David / dhildenb
Mike Rapoport
maintainerRe: [RFC/RFT PATCH 1/3] memblock: update initialization of reserved pages
In reply to Ard BiesheuvelQuoted 32 lines · Ard Biesheuvel
Quoted 21 lines · David Hildenbrand
Quoted 11 lines · Mike Rapoport
From: Mike Rapoport <[email protected]>
The struct pages representing a reserved memory region are initialized using
reserve_bootmem_range()function. This function is called for each reserved region just before the memory is freed from memblock to the buddy page allocator.The struct pages for
MEMBLOCK_NOMAPregions are kept with the default values set by the memory map initialization which makes it necessary to have a special treatment for such pages inpfn_valid()andpfn_valid_within().I assume these pages are never given to the buddy, because we don't have a direct mapping. So to the kernel, it's essentially just like a memory hole with benefits.
I can spot that we want to export such memory like any special memory thingy/hole in
/proc/iomem-- "reserved", which makes sense.Quoted 9 lines
I would assume that
MEMBLOCK_NOMAPis a special type of *reserved* memory. IOW, thatfor_each_reserved_mem_range()should already succeed on these as well -- we should mark anything that isMEMBLOCK_NOMAPimplicitly as reserved. Or are there valid reasons not to do so? What can anyone do with that memory?I assume they are pretty much useless for the kernel, right? Like other reserved memory ranges.
Quoted 9 lines
On ARM, we need to know whether any physical regions that do not contain system memory contain something with device semantics or not. One of the examples is ACPI tables: these are in reserved memory, and so they are not covered by the linear region. However, when the ACPI core ioremap()s an arbitrary memory region, we don't know whether it is mapping a memory region or a device region unless we keep track of this in some way. (Device mappings require device attributes, but firmware tables require memory attributes, as they might be accessed using misaligned reads)
I mostly agree, but my understanding is that regions of *physical* memory that are occupied by various pieces of EFI/ACPI information require special treatment because it was defined this way in the APCI spec.
And since ARM cannot tolerate aliased mappings with different caching mode the whole bunch of firmware memory should be ioremap()ed to access it.
Quoted 64 lines
Quoted 64 lines
Quoted 43 lines
Split out initialization of the reserved pages to a function with a meaningful name and treat the
MEMBLOCK_NOMAPregions the same way as the reserved regions and mark struct pages for the NOMAP regions as PageReserved.Signed-off-by Mike Rapoport <[email protected]>--- mm/memblock.c | 23 +++++++++++++++++++++-- 1 file changed, 21 insertions(+), 2 deletions(-)mm/memblock.c +20
-0@@ -2002,6 +2002,26 @@ static unsigned long __init __free_memory_core(phys_addr_t start,return end_pfn - start_pfn; } +static void __init memmap_init_reserved_pages (void ) +{ +struct memblock_region *region; +phys_addr_t start, end; +u64 i; + +/* initialize struct pages for the reserved regions */ +for_each_reserved_mem_range (i, &start, &end) +reserve_bootmem_region (start, end); + +/* and also treat struct pages for the NOMAP regions as PageReserved */ +for_each_mem_region (region) { +if (memblock_is_nomap (region)) { + start = region->base; + end = start + region->size; +reserve_bootmem_region (start, end); + } + } +} +static unsigned long __init free_low_memory_core_early (void ) {unsigned long count =0 ;mm/memblock.c static unsigned long __init free_low_memory_core_early(void) · line 2036
Quoted 9 lines
mm/memblock.c +1
-2@@ -2010,8 +2030,7 @@ static unsigned long __init free_low_memory_core_early(void)memblock_clear_hotplug (0 ,-1 ); -for_each_reserved_mem_range (i, &start, &end) -reserve_bootmem_region (start, end); +memmap_init_reserved_pages ();/* * We need to use NUMA_NO_NODE instead of NODE_DATA(0)->node_id --
Thanks,David / dhildenb
_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Signature · 2 lines
Sincerely yours, Mike.
Mike Rapoport
maintainerRe: [RFC/RFT PATCH 1/3] memblock: update initialization of reserved pages
In reply to David HildenbrandQuoted 8 lines · David Hildenbrand
Quoted 7 lines · Mike Rapoport
From: Mike Rapoport <[email protected]>
The struct pages representing a reserved memory region are initialized using
reserve_bootmem_range()function. This function is called for each reserved region just before the memory is freed from memblock to the buddy page allocator.
Quoted 8 lines
Quoted 4 lines
The struct pages for
MEMBLOCK_NOMAPregions are kept with the default values set by the memory map initialization which makes it necessary to have a special treatment for such pages inpfn_valid()andpfn_valid_within().I assume these pages are never given to the buddy, because we don't have a direct mapping. So to the kernel, it's essentially just like a memory hole with benefits.
The pages should not be accessed as normal memory so they do not have a direct (or in ARMish linear) mapping and are never given to buddy. After looking at ACPI standard I don't see a fundamental reason for this but they've already made this mess and we need to cope with it.
Quoted 2 lines
I can spot that we want to export such memory like any special memory thingy/hole in
/proc/iomem-- "reserved", which makes sense.
It does, but let's wait with /proc/iomem changes. We don't really have a 100% consistent view of it on different architectures, so adding yet another type there does not seem, well, urgent.
Quoted 8 lines
I would assume that
MEMBLOCK_NOMAPis a special type of *reserved* memory. IOW, thatfor_each_reserved_mem_range()should already succeed on these as well -- we should mark anything that isMEMBLOCK_NOMAPimplicitly as reserved. Or are there valid reasons not to do so? What can anyone do with that memory?I assume they are pretty much useless for the kernel, right? Like other reserved memory ranges.
I agree that there is a lot of commonality between NOMAP and reserved. The problem is that even semantics for reserved is different between architectures. Moreover, on the same architecture there could be E820_TYPE_RESERVED and memblock.reserved with different properties.
I'd really prefer moving in baby steps here because any change in the boot mm can bear several month of early hangs debugging ;-)
Quoted 47 lines
Quoted 47 lines
Split out initialization of the reserved pages to a function with a meaningful name and treat the
MEMBLOCK_NOMAPregions the same way as the reserved regions and mark struct pages for the NOMAP regions as PageReserved.Signed-off-by Mike Rapoport <[email protected]>--- mm/memblock.c | 23 +++++++++++++++++++++-- 1 file changed, 21 insertions(+), 2 deletions(-)mm/memblock.c +21
-2@@ -2002,6 +2002,26 @@ static unsigned long __init __free_memory_core(phys_addr_t start,return end_pfn - start_pfn; } +static void __init memmap_init_reserved_pages (void ) +{ +struct memblock_region *region; +phys_addr_t start, end; +u64 i; + +/* initialize struct pages for the reserved regions */ +for_each_reserved_mem_range (i, &start, &end) +reserve_bootmem_region (start, end); + +/* and also treat struct pages for the NOMAP regions as PageReserved */ +for_each_mem_region (region) { +if (memblock_is_nomap (region)) { + start = region->base; + end = start + region->size; +reserve_bootmem_region (start, end); + } + } +} +static unsigned long __init free_low_memory_core_early (void ) {unsigned long count =0 ;@@ -2010,8 +2030,7 @@ static unsigned long __init free_low_memory_core_early(void)memblock_clear_hotplug (0 ,-1 ); -for_each_reserved_mem_range (i, &start, &end) -reserve_bootmem_region (start, end); +memmap_init_reserved_pages ();/* * We need to use NUMA_NO_NODE instead of NODE_DATA(0)->node_id
Signature · 2 lines
Sincerely yours, Mike.
Re: [RFC/RFT PATCH 1/3] memblock: update initialization of reserved pages
Mike Rapoport <[email protected]> schrieb am Mi. 14. Apr. 2021 um 22:06:
Quoted 27 lines · David Hildenbrand
Quoted 5 lines · Mike Rapoport
Quoted 4 lines
From: Mike Rapoport <[email protected]>
The struct pages representing a reserved memory region are initialized using
reserve_bootmem_range()function. This function is called foreach
Quoted 1 line
Quoted 1 line
reserved region just before the memory is freed from memblock to the
buddy
Quoted 8 lines
Quoted 6 lines
page allocator.
The struct pages for
MEMBLOCK_NOMAPregions are kept with the default values set by the memory map initialization which makes it necessary to have a special treatment for such pages inpfn_valid()andpfn_valid_within().I assume these pages are never given to the buddy, because we don't have
a
Quoted 1 line
direct mapping. So to the kernel, it's essentially just like a memory
hole
Quoted 1 line
with benefits.
The pages should not be accessed as normal memory so they do not have a direct (or in ARMish linear) mapping and are never given to buddy. After looking at ACPI standard I don't see a fundamental reason for this but they've already made this mess and we need to cope with it.
Quoted 7 lines
Quoted 2 lines
I can spot that we want to export such memory like any special memory thingy/hole in
/proc/iomem-- "reserved", which makes sense.It does, but let's wait with
/proc/iomemchanges. We don't really have a 100% consistent view of it on different architectures, so adding yet another type there does not seem, well, urgent.
To clarify: this is already done on arm64.
Quoted 12 lines
Quoted 1 line
I would assume that
MEMBLOCK_NOMAPis a special type of *reserved*memory.
Quoted 1 line
IOW, that
for_each_reserved_mem_range()should already succeed on theseas
Quoted 2 lines
well -- we should mark anything that is
MEMBLOCK_NOMAPimplicitly as reserved. Or are there valid reasons not to do so? What can anyone dowith
Quoted 4 lines
that memory?
I assume they are pretty much useless for the kernel, right? Like other reserved memory ranges.
Quoted 7 lines
I agree that there is a lot of commonality between NOMAP and reserved. The problem is that even semantics for reserved is different between architectures. Moreover, on the same architecture there could be
E820_TYPE_RESERVEDand memblock.reserved with different properties.I'd really prefer moving in baby steps here because any change in the boot mm can bear several month of early hangs debugging ;-)
Yeah I know. We just should have the desired target state figured out :)
Quoted 53 lines
Quoted 3 lines
Quoted 3 lines
Split out initialization of the reserved pages to a function with a meaningful name and treat the
MEMBLOCK_NOMAPregions the same way asthe
Quoted 8 lines
Quoted 8 lines
reserved regions and mark struct pages for the NOMAP regions as PageReserved.
Signed-off-by Mike Rapoport <[email protected]>--- mm/memblock.c | 23 +++++++++++++++++++++-- 1 file changed, 21 insertions(+), 2 deletions(-)Quoted 5 lines
mm/memblock.c static unsigned long __init
Quoted 5 lines
mm/memblock.c +0
-0@@ -2002,6 +2002,26 @@ static unsigned long __init__free_memory_core(phys_addr_t start,
Quoted 5 lines
Quoted 5 lines
mm/memblock.c +3
-0return end_pfn - start_pfn; } +static void __init memmap_init_reserved_pages (void ) +{ +struct memblock_region *region;Quoted 8 lines
mm/memblock.c static unsigned long __init __free_memory_core(phys_addr_t start,
Quoted 8 lines
mm/memblock.c +8
-0+phys_addr_t start, end; +u64 i; + +/* initialize struct pages for the reserved regions */ +for_each_reserved_mem_range (i, &start, &end) +reserve_bootmem_region (start, end); + +/* and also treat struct pages for the NOMAP regions as PageReserved */
Quoted 5 lines
Quoted 5 lines
mm/memblock.c +5
-0+for_each_mem_region (region) { +if (memblock_is_nomap (region)) { + start = region->base; + end = start + region->size; +reserve_bootmem_region (start, end);Quoted 8 lines
mm/memblock.c static unsigned long __init
Quoted 8 lines
mm/memblock.c +4
-0+ } + } +} +static unsigned long __init free_low_memory_core_early (void ) {unsigned long count =0 ;@@ -2010,8 +2030,7 @@ static unsigned long __init
free_low_memory_core_early(void)Quoted 6 lines
mm/memblock.c static unsigned long __init free_low_memory_core_early(void) · line 2036
Quoted 6 lines
mm/memblock.c +1
-2memblock_clear_hotplug (0 ,-1 ); -for_each_reserved_mem_range (i, &start, &end) -reserve_bootmem_region (start, end); +memmap_init_reserved_pages ();/* * We need to use NUMA_NO_NODE instead of NODE_DATA(0)->node_id
Quoted 5 lines
--
Sincerely yours,
Mike.--
Thanks,
David / dhildenb
Mike Rapoport
maintainerPATCH 2/3 arm64: decouple check whether pfn is normal memory from pfn_valid()
In reply to Mike RapoportFrom: Mike Rapoport <[email protected]>
The intended semantics of pfn_valid() is to verify whether there is a struct page for the pfn in question and nothing else.
Yet, on arm64 it is used to distinguish memory areas that are mapped in the linear map vs those that require ioremap() to access them.
Introduce a dedicated pfn_is_memory() to perform such check and use it where appropriate.
--- arch/arm64/include/asm/memory.h | 2 +- arch/arm64/include/asm/page.h | 1 + arch/arm64/kvm/mmu.c | 2 +- arch/arm64/mm/init.c | 6 ++++++ arch/arm64/mm/ioremap.c | 4 ++-- arch/arm64/mm/mmu.c | 2 +- 6 files changed, 12 insertions(+), 5 deletions(-)
arch/arm64/include/asm/memory.h +1 -1
@@ -351,7 +351,7 @@ static inline void *phys_to_virt(phys_addr_t x)#define virt_addr_valid (addr) ({ \ __typeof__(addr) __addr = __tag_reset(addr); \ - __is_lm_address(__addr) && pfn_valid(virt_to_pfn(__addr)); \ + __is_lm_address(__addr) && pfn_is_memory(virt_to_pfn(__addr)); \ })void dump_mem_limit (void );
arch/arm64/include/asm/page.h +1 -0
@@ -38,6 +38,7 @@ void copy_highpage(struct page *to, struct page *from);typedef struct page *pgtable_t ;extern int pfn_valid (unsigned long ); +extern int pfn_is_memory (unsigned long );#include <asm/memory.h>
arch/arm64/kvm/mmu.c +1 -1
@@ -85,7 +85,7 @@ void kvm_flush_remote_tlbs(struct kvm *kvm)static bool kvm_is_device_pfn (unsigned long pfn) { -return !pfn_valid (pfn); +return !pfn_is_memory (pfn); }/*
arch/arm64/mm/init.c +6 -0
@@ -258,6 +258,12 @@ int pfn_valid(unsigned long pfn) }EXPORT_SYMBOL (pfn_valid); +int pfn_is_memory (unsigned long pfn) +{ +return memblock_is_map_memory (PFN_PHYS (pfn)); +} +EXPORT_SYMBOL (pfn_is_memory); +static phys_addr_t memory_limit =PHYS_ADDR_MAX ;/*
arch/arm64/mm/ioremap.c +2 -2
@@ -43,7 +43,7 @@ static void __iomem *__ioremap_caller(phys_addr_t phys_addr, size_t size,/* * Don't allow RAM to be mapped. */ -if (WARN_ON (pfn_valid (__phys_to_pfn (phys_addr)))) +if (WARN_ON (pfn_is_memory (__phys_to_pfn (phys_addr))))return NULL ; area =get_vm_area_caller (size,VM_IOREMAP , caller);
@@ -84,7 +84,7 @@ EXPORT_SYMBOL(iounmap);void __iomem *ioremap_cache (phys_addr_t phys_addr,size_t size) {/* For normal memory we already have a cacheable mapping. */ -if (pfn_valid (__phys_to_pfn (phys_addr))) +if (pfn_is_memory (__phys_to_pfn (phys_addr)))return (void __iomem *)__phys_to_virt (phys_addr); return __ioremap_caller(phys_addr, size, __pgprot(PROT_NORMAL),
arch/arm64/mm/mmu.c +1 -1
@@ -81,7 +81,7 @@ void set_swapper_pgd(pgd_t *pgdp, pgd_t pgd)pgprot_t phys_mem_access_prot (struct file *file,unsigned long pfn,unsigned long size,pgprot_t vma_prot) { -if (!pfn_valid (pfn)) +if (!pfn_is_memory (pfn))return pgprot_noncached (vma_prot);else if (file->f_flags &O_SYNC )return pgprot_writecombine (vma_prot);
Signature · 1 line
2.28.0
Re: [RFC/RFT PATCH 2/3] arm64: decouple check whether pfn is normal memory from pfn_valid()
Quoted 4 lines · Mike Rapoport
From: Mike Rapoport <[email protected]>
The intended semantics of
pfn_valid()is to verify whether there is a struct page for the pfn in question and nothing else.
Should there be a comment affirming this semantics interpretation, above the generic pfn_valid() in include/linux/mmzone.h ?
Quoted 60 lines
Yet, on arm64 it is used to distinguish memory areas that are mapped in the linear map vs those that require
ioremap()to access them.Introduce a dedicated
pfn_is_memory()to perform such check and use it where appropriate.Signed-off-by Mike Rapoport <[email protected]>--- arch/arm64/include/asm/memory.h | 2 +- arch/arm64/include/asm/page.h | 1 + arch/arm64/kvm/mmu.c | 2 +- arch/arm64/mm/init.c | 6 ++++++ arch/arm64/mm/ioremap.c | 4 ++-- arch/arm64/mm/mmu.c | 2 +- 6 files changed, 12 insertions(+), 5 deletions(-)arch/arm64/include/asm/memory.h +1
-1@@ -351,7 +351,7 @@ static inline void *phys_to_virt(phys_addr_t x)#define virt_addr_valid (addr) ({ \ __typeof__(addr) __addr = __tag_reset(addr); \ - __is_lm_address(__addr) && pfn_valid(virt_to_pfn(__addr)); \ + __is_lm_address(__addr) && pfn_is_memory(virt_to_pfn(__addr)); \ })void dump_mem_limit (void );arch/arm64/include/asm/page.h +1
-0@@ -38,6 +38,7 @@ void copy_highpage(struct page *to, struct page *from);typedef struct page *pgtable_t ;extern int pfn_valid (unsigned long ); +extern int pfn_is_memory (unsigned long );#include <asm/memory.h> arch/arm64/kvm/mmu.c +1
-1@@ -85,7 +85,7 @@ void kvm_flush_remote_tlbs(struct kvm *kvm)static bool kvm_is_device_pfn (unsigned long pfn) { -return !pfn_valid (pfn); +return !pfn_is_memory (pfn); }/* arch/arm64/mm/init.c +0
-0
arch/arm64/mm/init.c int pfn_valid(unsigned long pfn) · line 265
Quoted 8 lines
arch/arm64/mm/init.c +5
-0@@ -258,6 +258,12 @@ int pfn_valid(unsigned long pfn) }EXPORT_SYMBOL (pfn_valid); +int pfn_is_memory (unsigned long pfn) +{ +return memblock_is_map_memory (PFN_PHYS (pfn)); +} +EXPORT_SYMBOL(pfn_is_memory);> +
Should not this be generic though ? There is nothing platform or arm64 specific in here. Wondering as pfn_is_memory() just indicates that the pfn is linear mapped, should not it be renamed as pfn_is_linear_memory() instead ? Regardless, it's fine either way.
Quoted 39 lines
arch/arm64/mm/init.c +0
-0static phys_addr_t memory_limit =PHYS_ADDR_MAX ;/* arch/arm64/mm/ioremap.c +2
-2@@ -43,7 +43,7 @@ static void __iomem *__ioremap_caller(phys_addr_t phys_addr, size_t size,/* * Don't allow RAM to be mapped. */ -if (WARN_ON (pfn_valid (__phys_to_pfn (phys_addr)))) +if (WARN_ON (pfn_is_memory (__phys_to_pfn (phys_addr))))return NULL ; area =get_vm_area_caller (size,VM_IOREMAP , caller);@@ -84,7 +84,7 @@ EXPORT_SYMBOL(iounmap);void __iomem *ioremap_cache (phys_addr_t phys_addr,size_t size) {/* For normal memory we already have a cacheable mapping. */ -if (pfn_valid (__phys_to_pfn (phys_addr))) +if (pfn_is_memory (__phys_to_pfn (phys_addr)))return (void __iomem *)__phys_to_virt (phys_addr); return __ioremap_caller(phys_addr, size, __pgprot(PROT_NORMAL),arch/arm64/mm/mmu.c +1
-1@@ -81,7 +81,7 @@ void set_swapper_pgd(pgd_t *pgdp, pgd_t pgd)pgprot_t phys_mem_access_prot (struct file *file,unsigned long pfn,unsigned long size,pgprot_t vma_prot) { -if (!pfn_valid (pfn)) +if (!pfn_is_memory (pfn))return pgprot_noncached (vma_prot);else if (file->f_flags &O_SYNC )return pgprot_writecombine (vma_prot);
Mike Rapoport
maintainerRe: [RFC/RFT PATCH 2/3] arm64: decouple check whether pfn is normal memory from pfn_valid()
Quoted 9 lines · Anshuman Khandual
Quoted 4 lines · Mike Rapoport
From: Mike Rapoport <[email protected]>
The intended semantics of
pfn_valid()is to verify whether there is a struct page for the pfn in question and nothing else.Should there be a comment affirming this semantics interpretation, above the generic
pfn_valid()ininclude/linux/mmzone.h?
Yeah, that would have been helpful :)
Quoted 63 lines
Quoted 63 lines
Yet, on arm64 it is used to distinguish memory areas that are mapped in the linear map vs those that require
ioremap()to access them.Introduce a dedicated
pfn_is_memory()to perform such check and use it where appropriate.Signed-off-by Mike Rapoport <[email protected]>--- arch/arm64/include/asm/memory.h | 2 +- arch/arm64/include/asm/page.h | 1 + arch/arm64/kvm/mmu.c | 2 +- arch/arm64/mm/init.c | 6 ++++++ arch/arm64/mm/ioremap.c | 4 ++-- arch/arm64/mm/mmu.c | 2 +- 6 files changed, 12 insertions(+), 5 deletions(-)arch/arm64/include/asm/memory.h +1
-1@@ -351,7 +351,7 @@ static inline void *phys_to_virt(phys_addr_t x)#define virt_addr_valid (addr) ({ \ __typeof__(addr) __addr = __tag_reset(addr); \ - __is_lm_address(__addr) && pfn_valid(virt_to_pfn(__addr)); \ + __is_lm_address(__addr) && pfn_is_memory(virt_to_pfn(__addr)); \ })void dump_mem_limit (void );arch/arm64/include/asm/page.h +1
-0@@ -38,6 +38,7 @@ void copy_highpage(struct page *to, struct page *from);typedef struct page *pgtable_t ;extern int pfn_valid (unsigned long ); +extern int pfn_is_memory (unsigned long );#include <asm/memory.h> arch/arm64/kvm/mmu.c +1
-1@@ -85,7 +85,7 @@ void kvm_flush_remote_tlbs(struct kvm *kvm)static bool kvm_is_device_pfn (unsigned long pfn) { -return !pfn_valid (pfn); +return !pfn_is_memory (pfn); }/* arch/arm64/mm/init.c +0
-0@@ -258,6 +258,12 @@ int pfn_valid(unsigned long pfn) }EXPORT_SYMBOL (pfn_valid);
Quoted 8 lines
arch/arm64/mm/init.c int pfn_valid(unsigned long pfn)
Quoted 5 lines
arch/arm64/mm/init.c +5
-0+int pfn_is_memory (unsigned long pfn) +{ +return memblock_is_map_memory (PFN_PHYS (pfn)); +} +EXPORT_SYMBOL(pfn_is_memory);> +Should not this be generic though ? There is nothing platform or arm64 specific in here.
As NOMAP itself is quite ARM specific, this check is currently only relevant for arm64 and maybe arm32.
But probably having an EXPORT_SYMBOL wrapper for memblock_is_map_memory(), say in memblock does make sense for all architectures that have KEEP_MEMBLOCK.
Quoted 3 lines
Wondering as
pfn_is_memory()just indicates that the
pfn is linear mapped, should not it be renamed aspfn_is_linear_memory()instead ? Regardless, it's fine either way.
Yeah, I agree that naming could be better here. I think that for a generic name we'd need pfn_is_directly_mapped() so that it can be used on x86 ;-)
Quoted 39 lines
Quoted 39 lines
arch/arm64/mm/init.c +0
-0static phys_addr_t memory_limit =PHYS_ADDR_MAX ;/* arch/arm64/mm/ioremap.c +2
-2@@ -43,7 +43,7 @@ static void __iomem *__ioremap_caller(phys_addr_t phys_addr, size_t size,/* * Don't allow RAM to be mapped. */ -if (WARN_ON (pfn_valid (__phys_to_pfn (phys_addr)))) +if (WARN_ON (pfn_is_memory (__phys_to_pfn (phys_addr))))return NULL ; area =get_vm_area_caller (size,VM_IOREMAP , caller);@@ -84,7 +84,7 @@ EXPORT_SYMBOL(iounmap);void __iomem *ioremap_cache (phys_addr_t phys_addr,size_t size) {/* For normal memory we already have a cacheable mapping. */ -if (pfn_valid (__phys_to_pfn (phys_addr))) +if (pfn_is_memory (__phys_to_pfn (phys_addr)))return (void __iomem *)__phys_to_virt (phys_addr); return __ioremap_caller(phys_addr, size, __pgprot(PROT_NORMAL),arch/arm64/mm/mmu.c +1
-1@@ -81,7 +81,7 @@ void set_swapper_pgd(pgd_t *pgdp, pgd_t pgd)pgprot_t phys_mem_access_prot (struct file *file,unsigned long pfn,unsigned long size,pgprot_t vma_prot) { -if (!pfn_valid (pfn)) +if (!pfn_is_memory (pfn))return pgprot_noncached (vma_prot);else if (file->f_flags &O_SYNC )return pgprot_writecombine (vma_prot);
Signature · 2 lines
Sincerely yours, Mike.
Re: [RFC/RFT PATCH 2/3] arm64: decouple check whether pfn is normal memory from pfn_valid()
In reply to Anshuman KhandualQuoted 73 lines · Anshuman Khandual
Quoted 4 lines · Mike Rapoport
From: Mike Rapoport <[email protected]>
The intended semantics of
pfn_valid()is to verify whether there is a struct page for the pfn in question and nothing else.Should there be a comment affirming this semantics interpretation, above the generic
pfn_valid()ininclude/linux/mmzone.h?Quoted 63 lines
Yet, on arm64 it is used to distinguish memory areas that are mapped in the linear map vs those that require
ioremap()to access them.Introduce a dedicated
pfn_is_memory()to perform such check and use it where appropriate.Signed-off-by Mike Rapoport <[email protected]>--- arch/arm64/include/asm/memory.h | 2 +- arch/arm64/include/asm/page.h | 1 + arch/arm64/kvm/mmu.c | 2 +- arch/arm64/mm/init.c | 6 ++++++ arch/arm64/mm/ioremap.c | 4 ++-- arch/arm64/mm/mmu.c | 2 +- 6 files changed, 12 insertions(+), 5 deletions(-)arch/arm64/include/asm/memory.h +1
-1@@ -351,7 +351,7 @@ static inline void *phys_to_virt(phys_addr_t x)#define virt_addr_valid (addr) ({ \ __typeof__(addr) __addr = __tag_reset(addr); \ - __is_lm_address(__addr) && pfn_valid(virt_to_pfn(__addr)); \ + __is_lm_address(__addr) && pfn_is_memory(virt_to_pfn(__addr)); \ })void dump_mem_limit (void );arch/arm64/include/asm/page.h +1
-0@@ -38,6 +38,7 @@ void copy_highpage(struct page *to, struct page *from);typedef struct page *pgtable_t ;extern int pfn_valid (unsigned long ); +extern int pfn_is_memory (unsigned long );#include <asm/memory.h> arch/arm64/kvm/mmu.c +1
-1@@ -85,7 +85,7 @@ void kvm_flush_remote_tlbs(struct kvm *kvm)static bool kvm_is_device_pfn (unsigned long pfn) { -return !pfn_valid (pfn); +return !pfn_is_memory (pfn); }/* arch/arm64/mm/init.c +0
-0@@ -258,6 +258,12 @@ int pfn_valid(unsigned long pfn) }EXPORT_SYMBOL (pfn_valid);
Quoted 10 lines
arch/arm64/mm/init.c int pfn_valid(unsigned long pfn)
Quoted 5 lines
arch/arm64/mm/init.c +5
-0+int pfn_is_memory (unsigned long pfn) +{ +return memblock_is_map_memory (PFN_PHYS (pfn)); +} +EXPORT_SYMBOL(pfn_is_memory);> +Should not this be generic though ? There is nothing platform or arm64 specific in here. Wondering as
pfn_is_memory()just indicates that the pfn is linear mapped, should not it be renamed aspfn_is_linear_memory()instead ? Regardless, it's fine either way.
TBH, I dislike (generic) pfn_is_memory(). It feels like we're mixing concepts. NOMAP memory vs !NOMAP memory; even NOMAP is some kind of memory after all. pfn_is_map_memory() would be more expressive, although still sub-optimal.
We'd actually want some kind of arm64-specific pfn_is_system_memory() or the inverse pfn_is_device_memory() -- to be improved.
Signature · 3 lines
Thanks, David / dhildenb
Mike Rapoport
maintainerRe: [RFC/RFT PATCH 2/3] arm64: decouple check whether pfn is normal memory from pfn_valid()
Quoted 77 lines · David Hildenbrand
Quoted 67 lines · Anshuman Khandual
Quoted 4 lines · Mike Rapoport
From: Mike Rapoport <[email protected]>
The intended semantics of
pfn_valid()is to verify whether there is a struct page for the pfn in question and nothing else.Should there be a comment affirming this semantics interpretation, above the generic
pfn_valid()ininclude/linux/mmzone.h?Quoted 57 lines
Yet, on arm64 it is used to distinguish memory areas that are mapped in the linear map vs those that require
ioremap()to access them.Introduce a dedicated
pfn_is_memory()to perform such check and use it where appropriate.Signed-off-by Mike Rapoport <[email protected]>--- arch/arm64/include/asm/memory.h | 2 +- arch/arm64/include/asm/page.h | 1 + arch/arm64/kvm/mmu.c | 2 +- arch/arm64/mm/init.c | 6 ++++++ arch/arm64/mm/ioremap.c | 4 ++-- arch/arm64/mm/mmu.c | 2 +- 6 files changed, 12 insertions(+), 5 deletions(-)arch/arm64/include/asm/memory.h +1
-1@@ -351,7 +351,7 @@ static inline void *phys_to_virt(phys_addr_t x)#define virt_addr_valid (addr) ({ \ __typeof__(addr) __addr = __tag_reset(addr); \ - __is_lm_address(__addr) && pfn_valid(virt_to_pfn(__addr)); \ + __is_lm_address(__addr) && pfn_is_memory(virt_to_pfn(__addr)); \ })void dump_mem_limit (void );arch/arm64/include/asm/page.h +1
-0@@ -38,6 +38,7 @@ void copy_highpage(struct page *to, struct page *from);typedef struct page *pgtable_t ;extern int pfn_valid (unsigned long ); +extern int pfn_is_memory (unsigned long );#include <asm/memory.h> arch/arm64/kvm/mmu.c +1
-1@@ -85,7 +85,7 @@ void kvm_flush_remote_tlbs(struct kvm *kvm)static bool kvm_is_device_pfn (unsigned long pfn) { -return !pfn_valid (pfn); +return !pfn_is_memory (pfn); }/* arch/arm64/mm/init.c +2
-0@@ -258,6 +258,12 @@ int pfn_valid(unsigned long pfn) }EXPORT_SYMBOL (pfn_valid); +int pfn_is_memory (unsigned long pfn) +{Quoted 8 lines
arch/arm64/mm/init.c int pfn_valid(unsigned long pfn)
Quoted 3 lines
arch/arm64/mm/init.c +3
-0+return memblock_is_map_memory (PFN_PHYS (pfn)); +} +EXPORT_SYMBOL(pfn_is_memory);> +Should not this be generic though ? There is nothing platform or arm64 specific in here. Wondering as
pfn_is_memory()just indicates that the pfn is linear mapped, should not it be renamed aspfn_is_linear_memory()instead ? Regardless, it's fine either way.
Quoted 2 lines
TBH, I dislike (generic)
pfn_is_memory(). It feels like we're mixing concepts.
Yeah, at the moment NOMAP is very much arm specific so I'd keep it this way for now.
Quoted 6 lines
NOMAP memory vs !NOMAP memory; even NOMAP is some kind of memory
after all.pfn_is_map_memory()would be more expressive, although still sub-optimal.We'd actually want some kind of arm64-specific
pfn_is_system_memory()or the inversepfn_is_device_memory()-- to be improved.
In my current version (to be posted soon) I've started with pfn_lineary_mapped() but then ended up with pfn_mapped() to make it "upward" compatible with architectures that use direct rather than linear map :)
Signature · 2 lines
Sincerely yours, Mike.
Re: [RFC/RFT PATCH 2/3] arm64: decouple check whether pfn is normal memory from pfn_valid()
Quoted 88 lines · Mike Rapoport
Quoted 77 lines · David Hildenbrand
Quoted 67 lines · Anshuman Khandual
Quoted 4 lines · Mike Rapoport
From: Mike Rapoport <[email protected]>
The intended semantics of
pfn_valid()is to verify whether there is a struct page for the pfn in question and nothing else.Should there be a comment affirming this semantics interpretation, above the generic
pfn_valid()ininclude/linux/mmzone.h?Quoted 57 lines
Yet, on arm64 it is used to distinguish memory areas that are mapped in the linear map vs those that require
ioremap()to access them.Introduce a dedicated
pfn_is_memory()to perform such check and use it where appropriate.Signed-off-by Mike Rapoport <[email protected]>--- arch/arm64/include/asm/memory.h | 2 +- arch/arm64/include/asm/page.h | 1 + arch/arm64/kvm/mmu.c | 2 +- arch/arm64/mm/init.c | 6 ++++++ arch/arm64/mm/ioremap.c | 4 ++-- arch/arm64/mm/mmu.c | 2 +- 6 files changed, 12 insertions(+), 5 deletions(-)arch/arm64/include/asm/memory.h +1
-1@@ -351,7 +351,7 @@ static inline void *phys_to_virt(phys_addr_t x)#define virt_addr_valid (addr) ({ \ __typeof__(addr) __addr = __tag_reset(addr); \ - __is_lm_address(__addr) && pfn_valid(virt_to_pfn(__addr)); \ + __is_lm_address(__addr) && pfn_is_memory(virt_to_pfn(__addr)); \ })void dump_mem_limit (void );arch/arm64/include/asm/page.h +1
-0@@ -38,6 +38,7 @@ void copy_highpage(struct page *to, struct page *from);typedef struct page *pgtable_t ;extern int pfn_valid (unsigned long ); +extern int pfn_is_memory (unsigned long );#include <asm/memory.h> arch/arm64/kvm/mmu.c +1
-1@@ -85,7 +85,7 @@ void kvm_flush_remote_tlbs(struct kvm *kvm)static bool kvm_is_device_pfn (unsigned long pfn) { -return !pfn_valid (pfn); +return !pfn_is_memory (pfn); }/* arch/arm64/mm/init.c +2
-0@@ -258,6 +258,12 @@ int pfn_valid(unsigned long pfn) }EXPORT_SYMBOL (pfn_valid); +int pfn_is_memory (unsigned long pfn) +{Quoted 8 lines
arch/arm64/mm/init.c int pfn_valid(unsigned long pfn)
Quoted 3 lines
arch/arm64/mm/init.c +3
-0+return memblock_is_map_memory (PFN_PHYS (pfn)); +} +EXPORT_SYMBOL(pfn_is_memory);> +Should not this be generic though ? There is nothing platform or arm64 specific in here. Wondering as
pfn_is_memory()just indicates that the pfn is linear mapped, should not it be renamed aspfn_is_linear_memory()instead ? Regardless, it's fine either way.Quoted 2 lines
TBH, I dislike (generic)
pfn_is_memory(). It feels like we're mixing concepts.Yeah, at the moment NOMAP is very much arm specific so I'd keep it this way for now.
Quoted 4 lines
NOMAP memory vs !NOMAP memory; even NOMAP is some kind of memory
after all.pfn_is_map_memory()would be more expressive, although still sub-optimal.
Quoted 7 lines
Quoted 2 lines
We'd actually want some kind of arm64-specific
pfn_is_system_memory()or the inversepfn_is_device_memory()-- to be improved.In my current version (to be posted soon) I've started with
pfn_lineary_mapped()but then ended up withpfn_mapped()to make it "upward" compatible with architectures that use direct rather than linear map :)
And even that is moot. It doesn't tell you if a PFN is *actually* mapped (hello secretmem).
I'd suggest to just use memblock_is_map_memory() in arch specific code. Then it's clear what we are querying exactly and what the semantics might be.
Signature · 3 lines
Thanks, David / dhildenb
Mike Rapoport
maintainerRe: [RFC/RFT PATCH 2/3] arm64: decouple check whether pfn is normal memory from pfn_valid()
Quoted 97 lines · David Hildenbrand
Quoted 88 lines · Mike Rapoport
Quoted 77 lines · David Hildenbrand
Quoted 67 lines · Anshuman Khandual
Quoted 4 lines · Mike Rapoport
From: Mike Rapoport <[email protected]>
The intended semantics of
pfn_valid()is to verify whether there is a struct page for the pfn in question and nothing else.Should there be a comment affirming this semantics interpretation, above the generic
pfn_valid()ininclude/linux/mmzone.h?Quoted 57 lines
Yet, on arm64 it is used to distinguish memory areas that are mapped in the linear map vs those that require
ioremap()to access them.Introduce a dedicated
pfn_is_memory()to perform such check and use it where appropriate.Signed-off-by Mike Rapoport <[email protected]>--- arch/arm64/include/asm/memory.h | 2 +- arch/arm64/include/asm/page.h | 1 + arch/arm64/kvm/mmu.c | 2 +- arch/arm64/mm/init.c | 6 ++++++ arch/arm64/mm/ioremap.c | 4 ++-- arch/arm64/mm/mmu.c | 2 +- 6 files changed, 12 insertions(+), 5 deletions(-)arch/arm64/include/asm/memory.h +1
-1@@ -351,7 +351,7 @@ static inline void *phys_to_virt(phys_addr_t x)#define virt_addr_valid (addr) ({ \ __typeof__(addr) __addr = __tag_reset(addr); \ - __is_lm_address(__addr) && pfn_valid(virt_to_pfn(__addr)); \ + __is_lm_address(__addr) && pfn_is_memory(virt_to_pfn(__addr)); \ })void dump_mem_limit (void );arch/arm64/include/asm/page.h +1
-0@@ -38,6 +38,7 @@ void copy_highpage(struct page *to, struct page *from);typedef struct page *pgtable_t ;extern int pfn_valid (unsigned long ); +extern int pfn_is_memory (unsigned long );#include <asm/memory.h> arch/arm64/kvm/mmu.c +1
-1@@ -85,7 +85,7 @@ void kvm_flush_remote_tlbs(struct kvm *kvm)static bool kvm_is_device_pfn (unsigned long pfn) { -return !pfn_valid (pfn); +return !pfn_is_memory (pfn); }/* arch/arm64/mm/init.c +2
-0@@ -258,6 +258,12 @@ int pfn_valid(unsigned long pfn) }EXPORT_SYMBOL (pfn_valid); +int pfn_is_memory (unsigned long pfn) +{Quoted 8 lines
arch/arm64/mm/init.c int pfn_valid(unsigned long pfn)
Quoted 3 lines
arch/arm64/mm/init.c +3
-0+return memblock_is_map_memory (PFN_PHYS (pfn)); +} +EXPORT_SYMBOL(pfn_is_memory);> +Should not this be generic though ? There is nothing platform or arm64 specific in here. Wondering as
pfn_is_memory()just indicates that the pfn is linear mapped, should not it be renamed aspfn_is_linear_memory()instead ? Regardless, it's fine either way.Quoted 2 lines
TBH, I dislike (generic)
pfn_is_memory(). It feels like we're mixing concepts.Yeah, at the moment NOMAP is very much arm specific so I'd keep it this way for now.
Quoted 4 lines
NOMAP memory vs !NOMAP memory; even NOMAP is some kind of memory
after all.pfn_is_map_memory()would be more expressive, although still sub-optimal.Quoted 7 lines
Quoted 2 lines
We'd actually want some kind of arm64-specific
pfn_is_system_memory()or the inversepfn_is_device_memory()-- to be improved.In my current version (to be posted soon) I've started with
pfn_lineary_mapped()but then ended up withpfn_mapped()to make it "upward" compatible with architectures that use direct rather than linear map :)
Quoted 5 lines
And even that is moot. It doesn't tell you if a PFN is *actually* mapped (hello secretmem).
I'd suggest to just use
memblock_is_map_memory()in arch specific code. Then it's clear what we are querying exactly and what the semantics might be.
Ok, let's export memblock_is_map_memory() for the KEEP_MEMBLOCK case.
Signature · 2 lines
Sincerely yours, Mike.
Mike Rapoport
maintainerPATCH 3/3 arm64: drop pfn_valid_within() and simplify pfn_valid()
In reply to Mike RapoportFrom: Mike Rapoport <[email protected]>
The arm64's version of pfn_valid() differs from the generic because of two reasons:
* Parts of the memory map are freed during boot. This makes it necessary to verify that there is actual physical memory that corresponds to a pfn which is done by querying memblock.
* There are NOMAP memory regions. These regions are not mapped in the linear map and until the previous commit the struct pages representing these areas had default values.
As the consequence of absence of the special treatment of NOMAP regions in the memory map it was necessary to use memblock_is_map_memory() in pfn_valid() and to have pfn_valid_within() aliased to pfn_valid() so that generic mm functionality would not treat a NOMAP page as a normal page.
Since the NOMAP regions are now marked as PageReserved(), pfn walkers and the rest of core mm will treat them as unusable memory and thus pfn_valid_within() is no longer required at all and can be disabled by removing CONFIG_HOLES_IN_ZONE on arm64.
pfn_valid() can be slightly simplified by replacing memblock_is_map_memory() with memblock_is_memory().
--- arch/arm64/Kconfig | 3 --- arch/arm64/mm/init.c | 4 ++-- 2 files changed, 2 insertions(+), 5 deletions(-)
arch/arm64/Kconfig +0 -3
@@ -1040,9 +1040,6 @@ config NEED_PER_CPU_EMBED_FIRST_CHUNK def_bool y depends on NUMA -config HOLES_IN_ZONE - def_bool y - source "kernel/Kconfig.hz" config ARCH_SPARSEMEM_ENABLE
arch/arm64/mm/init.c +2 -2
@@ -243,7 +243,7 @@ int pfn_valid(unsigned long pfn)/* * ZONE_DEVICE memory does not have the memblock entries. -* memblock_is_map_memory() check for ZONE_DEVICE based +* memblock_is_memory() check for ZONE_DEVICE based * addresses will always fail. Even the normal hotplugged * memory will never have MEMBLOCK_NOMAP flag set in their * memblock entries. Skip memblock search for all non early
@@ -254,7 +254,7 @@ int pfn_valid(unsigned long pfn) return pfn_section_valid(ms, pfn); }#endif -return memblock_is_map_memory (addr); +return memblock_is_memory (addr); }EXPORT_SYMBOL (pfn_valid);
Signature · 1 line
2.28.0
Quoted 18 lines · Mike Rapoport
From: Mike Rapoport <[email protected]>
The arm64's version of
pfn_valid()differs from the generic because of two reasons:* Parts of the memory map are freed during boot. This makes it necessary to verify that there is actual physical memory that corresponds to a pfn which is done by querying memblock.
* There are NOMAP memory regions. These regions are not mapped in the linear map and until the previous commit the struct pages representing these areas had default values.
As the consequence of absence of the special treatment of NOMAP regions in the memory map it was necessary to use
memblock_is_map_memory()inpfn_valid()and to havepfn_valid_within()aliased topfn_valid()so that generic mm functionality would not treat a NOMAP page as a normal page.
Quoted 4 lines
Since the NOMAP regions are now marked as
PageReserved(), pfn walkers and the rest of core mm will treat them as unusable memory and thuspfn_valid_within()is no longer required at all and can be disabled by removingCONFIG_HOLES_IN_ZONEon arm64.
But what about the memory map that are freed during boot (mentioned above). Would not they still cause CONFIG_HOLES_IN_ZONE to be applicable and hence pfn_valid_within() ?
Quoted 3 lines
pfn_valid()can be slightly simplified by replacingmemblock_is_map_memory()withmemblock_is_memory().
Just to understand this better, pfn_valid() will now return true for all MEMBLOCK_NOMAP based memory but that is okay as core MM would still ignore them as unusable memory for being PageReserved().
Quoted 44 lines
Signed-off-by Mike Rapoport <[email protected]>--- arch/arm64/Kconfig | 3 --- arch/arm64/mm/init.c | 4 ++-- 2 files changed, 2 insertions(+), 5 deletions(-)arch/arm64/Kconfig +0
-3@@ -1040,9 +1040,6 @@ config NEED_PER_CPU_EMBED_FIRST_CHUNK def_bool y depends on NUMA -config HOLES_IN_ZONE - def_bool y - source "kernel/Kconfig.hz" config ARCH_SPARSEMEM_ENABLEarch/arm64/mm/init.c +2
-2@@ -243,7 +243,7 @@ int pfn_valid(unsigned long pfn)/* * ZONE_DEVICE memory does not have the memblock entries. -* memblock_is_map_memory() check for ZONE_DEVICE based +* memblock_is_memory() check for ZONE_DEVICE based * addresses will always fail. Even the normal hotplugged * memory will never have MEMBLOCK_NOMAP flag set in their * memblock entries. Skip memblock search for all non early @@ -254,7 +254,7 @@ int pfn_valid(unsigned long pfn) return pfn_section_valid(ms, pfn); }#endif -return memblock_is_map_memory (addr); +return memblock_is_memory (addr); }EXPORT_SYMBOL (pfn_valid);
Mike Rapoport
maintainerQuoted 20 lines · Anshuman Khandual
Quoted 18 lines · Mike Rapoport
From: Mike Rapoport <[email protected]>
The arm64's version of
pfn_valid()differs from the generic because of two reasons:* Parts of the memory map are freed during boot. This makes it necessary to verify that there is actual physical memory that corresponds to a pfn which is done by querying memblock.
* There are NOMAP memory regions. These regions are not mapped in the linear map and until the previous commit the struct pages representing these areas had default values.
As the consequence of absence of the special treatment of NOMAP regions in the memory map it was necessary to use
memblock_is_map_memory()inpfn_valid()and to havepfn_valid_within()aliased topfn_valid()so that generic mm functionality would not treat a NOMAP page as a normal page.
Quoted 8 lines
Quoted 4 lines
Since the NOMAP regions are now marked as
PageReserved(), pfn walkers and the rest of core mm will treat them as unusable memory and thuspfn_valid_within()is no longer required at all and can be disabled by removingCONFIG_HOLES_IN_ZONEon arm64.But what about the memory map that are freed during boot (mentioned above). Would not they still cause
CONFIG_HOLES_IN_ZONEto be applicable and hencepfn_valid_within()?
The CONFIG_HOLES_IN_ZONE name is misleading as actually pfn_valid_within() is only required for holes within a MAX_ORDER_NR_PAGES blocks (see comment near pfn_valid_within() definition in mmzone.h). The freeing of the memory map during boot avoids breaking MAX_ORDER blocks and the holes for which memory map is freed are always aligned at MAX_ORDER.
AFAIU, the only case when there could be a hole in a MAX_ORDER block is when EFI/ACPI reserves memory for its use and this memory becomes NOMAP in the kernel. We still create struct pages for this memory, but they never get values other than defaults, so core mm has no idea that this memory should be touched, hence the need for pfn_valid_within() aliased to pfn_valid() on arm64.
Quoted 6 lines
Quoted 2 lines
pfn_valid()can be slightly simplified by replacingmemblock_is_map_memory()withmemblock_is_memory().Just to understand this better,
pfn_valid()will now return true for allMEMBLOCK_NOMAPbased memory but that is okay as core MM would still ignore them as unusable memory for beingPageReserved().
Right, pfn_valid() will return true for all memory, including MEMBLOCK_NOMAP. Since core mm deals with PageResrved() for memory used by the firmware, e.g. on x86, I don't see why it won't work on arm64.
Quoted 44 lines
Quoted 44 lines
Signed-off-by Mike Rapoport <[email protected]>--- arch/arm64/Kconfig | 3 --- arch/arm64/mm/init.c | 4 ++-- 2 files changed, 2 insertions(+), 5 deletions(-)arch/arm64/Kconfig +0
-3@@ -1040,9 +1040,6 @@ config NEED_PER_CPU_EMBED_FIRST_CHUNK def_bool y depends on NUMA -config HOLES_IN_ZONE - def_bool y - source "kernel/Kconfig.hz" config ARCH_SPARSEMEM_ENABLEarch/arm64/mm/init.c +2
-2@@ -243,7 +243,7 @@ int pfn_valid(unsigned long pfn)/* * ZONE_DEVICE memory does not have the memblock entries. -* memblock_is_map_memory() check for ZONE_DEVICE based +* memblock_is_memory() check for ZONE_DEVICE based * addresses will always fail. Even the normal hotplugged * memory will never have MEMBLOCK_NOMAP flag set in their * memblock entries. Skip memblock search for all non early @@ -254,7 +254,7 @@ int pfn_valid(unsigned long pfn) return pfn_section_valid(ms, pfn); }#endif -return memblock_is_map_memory (addr); +return memblock_is_memory (addr); }EXPORT_SYMBOL (pfn_valid);
Signature · 2 lines
Sincerely yours, Mike.
Adding James here.
+ James Morse <[email protected]>
Quoted 6 lines · Mike Rapoport
From: Mike Rapoport <[email protected]>
Hi,
These patches aim to remove
CONFIG_HOLES_IN_ZONEand essentially hardwirepfn_valid_within()to 1.
That would be really great for arm64 platform as it will save CPU cycles on many generic MM paths, given that our pfn_valid() has been expensive.
Quoted 2 lines
The idea is to mark NOMAP pages as reserved in the memory map and restore
Though I am not really sure, would that possibly be problematic for UEFI/EFI use cases as it might have just treated them as normal struct pages till now.
Quoted 2 lines
the intended semantics of
pfn_valid()to designate availability of struct page for a pfn.
Right, that would be better as the current semantics is not ideal.
Quoted 7 lines
With this the core mm will be able to cope with the fact that it cannot use NOMAP pages and the holes created by NOMAP ranges within
MAX_ORDERblocks will be treated correctly even without the need forpfn_valid_within.The patches are only boot tested on qemu-system-aarch64 so I'd really appreciate memory stress tests on real hardware.
Did some preliminary memory stress tests on a guest with portions of memory marked as MEMBLOCK_NOMAP and did not find any obvious problem. But this might require some testing on real UEFI environment with firmware using MEMBLOCK_NOMAP memory to make sure that changing these struct pages to PageReserved() is safe.
Quoted 3 lines
If this actually works we'll be one step closer to drop custom
pfn_valid()on arm64 altogether.
Right, planning to rework and respin the RFC originally sent last month.
https://patchwork.kernel.org/project/linux-mm/patch/[email protected]/
Mike Rapoport
maintainerQuoted 12 lines · Anshuman Khandual
Adding James here.
+ James Morse <[email protected]>
Quoted 6 lines · Mike Rapoport
From: Mike Rapoport <[email protected]>
Hi,
These patches aim to remove
CONFIG_HOLES_IN_ZONEand essentially hardwirepfn_valid_within()to 1.
Quoted 8 lines
That would be really great for arm64 platform as it will save CPU cycles on many generic MM paths, given that our
pfn_valid()has been expensive.Quoted 2 lines
The idea is to mark NOMAP pages as reserved in the memory map and restore
Though I am not really sure, would that possibly be problematic for UEFI/EFI use cases as it might have just treated them as normal struct pages till now.
I don't think there should be a problem because now the struct pages for UEFI/ACPI never got to be used by the core mm. They were (rightfully) skipped by memblock_free_all() from one side and pfn_valid() and pfn_valid_within() return false for them in various pfn walkers from the other side.
Quoted 10 lines
Quoted 2 lines
the intended semantics of
pfn_valid()to designate availability of struct page for a pfn.Right, that would be better as the current semantics is not ideal.
Quoted 5 lines
With this the core mm will be able to cope with the fact that it cannot use NOMAP pages and the holes created by NOMAP ranges within
MAX_ORDERblocks will be treated correctly even without the need forpfn_valid_within.
Quoted 7 lines
Quoted 2 lines
The patches are only boot tested on qemu-system-aarch64 so I'd really appreciate memory stress tests on real hardware.
Did some preliminary memory stress tests on a guest with portions of memory marked as
MEMBLOCK_NOMAPand did not find any obvious problem. But this might require some testing on real UEFI environment with firmware usingMEMBLOCK_NOMAPmemory to make sure that changing these struct pages toPageReserved()is safe.
I surely have no access for such machines :)
Quoted 6 lines
Quoted 2 lines
If this actually works we'll be one step closer to drop custom
pfn_valid()on arm64 altogether.Right, planning to rework and respin the RFC originally sent last month.
https://patchwork.kernel.org/project/linux-mm/patch/[email protected]/
Signature · 2 lines
Sincerely yours, Mike.