xen: Replace arch_mfn_in_directmap() with arch_mfns_in_directmap()
PATCH v2 1/1 xen: Replace arch_mfn_in_directmap() with arch_mfns_in_directmap()
From: Julien Grall <[email protected]>
The name of arch_mfn_in_directmap() suggests that it will check against that the passed MFN should be in the directmap.
However, the current callers are passing the next MFN and the implementation will return true for up to one MFN past the directmap.
It would be more meaningful to test the exact MFN rather than the next one.
That said, the current expectation is the memory will be direct-mapped from 0 up to a given MFN. This may not be a valid assumption on all the architectures.
For instance, on Arm32 only the xenheap that will be direct-mapped. This may not be allocated a the beginning of the RAM.
So take the opportunity to rework the parameters and pass the number of pages we want to check. This also requires to rename the helper to better match the implementation.
Note that the implementation of the helper on arm32 is left as-is for now.
----
Changes in v2:
- Pass a region instead the last MFN.
- Rename the helper to arch_mfns_in_directmap()
- This was https://lore.kernel.org/xen-devel/[email protected]/
Looking at the history, it looks like the check in init_node_heap() was <= and it was simply moved to a new helper without any adjustment as part of c6fdc9696a "boot allocator: use arch helper for virt_to_mfn on DIRECTMAP_VIRT region".
--- xen/arch/arm/include/asm/arm32/mm.h | 2 +- xen/arch/arm/include/asm/arm64/mm.h | 2 +- xen/arch/x86/include/asm/mm.h | 6 +++--- xen/common/page_alloc.c | 4 ++-- 4 files changed, 7 insertions(+), 7 deletions(-)
xen/arch/arm/include/asm/arm32/mm.h +1 -1
@@ -5,7 +5,7 @@* Only a limited amount of RAM, called xenheap, is always mapped on ARM32. * For convenience always return false. */ -static inline bool arch_mfn_in_directmap (unsigned long mfn) +static inline bool arch_mfns_in_directmap (unsigned long mfn,unsigned long nr) {return false; }
xen/arch/arm/include/asm/arm64/mm.h +1 -1
@@ -5,7 +5,7 @@* On ARM64, all the RAM is currently direct mapped in Xen. * Hence return always true. */ -static inline bool arch_mfn_in_directmap (unsigned long mfn) +static inline bool arch_mfns_in_directmap (unsigned long mfn,unsigned long nr) {return true; }
xen/arch/x86/include/asm/mm.h +3 -3
@@ -635,13 +635,13 @@ void write_32bit_pse_identmap(uint32_t *l2);/* * x86 maps part of physical memory via the directmap region. -* Return whether the input MFN falls in that range. +* Return whether the range of MFN falls in the directmap region. */ -static inline bool arch_mfn_in_directmap (unsigned long mfn) +static inline bool arch_mfns_in_directmap (unsigned long mfn,unsigned long nr) {unsigned long eva =min (DIRECTMAP_VIRT_END ,HYPERVISOR_VIRT_END ); -return mfn <= (virt_to_mfn (eva -1 ) +1 ); +return (mfn + nr) <= (virt_to_mfn (eva -1 ) +1 ); }#endif /* __ASM_X86_MM_H__ */
xen/common/page_alloc.c +2 -2
@@ -588,7 +588,7 @@ static unsigned long init_node_heap(int node, unsigned long mfn, needed =0 ; }else if ( *use_tail && nr >= needed && -arch_mfn_in_directmap (mfn + nr) && +arch_mfns_in_directmap (mfn + nr - needed, needed) && (!xenheap_bits || !((mfn + nr -1 ) >> (xenheap_bits -PAGE_SHIFT ))) ) {
@@ -597,7 +597,7 @@ static unsigned long init_node_heap(int node, unsigned long mfn,PAGE_SIZE -sizeof (**avail) *NR_ZONES ; }else if ( nr >= needed && -arch_mfn_in_directmap (mfn + needed) && +arch_mfns_in_directmap (mfn, needed) && (!xenheap_bits || !((mfn + needed -1 ) >> (xenheap_bits -PAGE_SHIFT ))) ) {
Signature · 1 line
2.32.0
Quoted 18 lines · Julien Grall
From: Julien Grall <[email protected]>
The name of
arch_mfn_in_directmap()suggests that it will check against that the passed MFN should be in the directmap.However, the current callers are passing the next MFN and the implementation will return true for up to one MFN past the directmap.
It would be more meaningful to test the exact MFN rather than the next one.
That said, the current expectation is the memory will be direct-mapped from 0 up to a given MFN. This may not be a valid assumption on all the architectures.
For instance, on Arm32 only the xenheap that will be direct-mapped. This may not be allocated a the beginning of the RAM.
Quoted 8 lines
So take the opportunity to rework the parameters and pass the number of pages we want to check. This also requires to rename the helper to better match the implementation.
Note that the implementation of the helper on arm32 is left as-is for now.
Signed-off-by Julien Grall <[email protected]>
Quoted 18 lines · Julien Grall
From: Julien Grall <[email protected]>
The name of
arch_mfn_in_directmap()suggests that it will check against that the passed MFN should be in the directmap.However, the current callers are passing the next MFN and the implementation will return true for up to one MFN past the directmap.
It would be more meaningful to test the exact MFN rather than the next one.
That said, the current expectation is the memory will be direct-mapped from 0 up to a given MFN. This may not be a valid assumption on all the architectures.
For instance, on Arm32 only the xenheap that will be direct-mapped. This may not be allocated a the beginning of the RAM.
Quoted 8 lines
So take the opportunity to rework the parameters and pass the number of pages we want to check. This also requires to rename the helper to better match the implementation.
Note that the implementation of the helper on arm32 is left as-is for now.
Signed-off-by Julien Grall <[email protected]>
Quoted 90 lines
----
Changes in v2:
- Pass a region instead the last MFN.
- Rename the helper toarch_mfns_in_directmap()
- This was https://lore.kernel.org/xen-devel/[email protected]/Looking at the history, it looks like the check in
init_node_heap()was <= and it was simply moved to a new helper without any adjustment as part ofc6fdc9696a"boot allocator: use arch helper forvirt_to_mfnonDIRECTMAP_VIRTregion".--- xen/arch/arm/include/asm/arm32/mm.h | 2 +- xen/arch/arm/include/asm/arm64/mm.h | 2 +- xen/arch/x86/include/asm/mm.h | 6 +++--- xen/common/page_alloc.c | 4 ++-- 4 files changed, 7 insertions(+), 7 deletions(-)xen/arch/arm/include/asm/arm32/mm.h +1
-1@@ -5,7 +5,7 @@* Only a limited amount of RAM, called xenheap, is always mapped on ARM32. * For convenience always return false. */ -static inline bool arch_mfn_in_directmap (unsigned long mfn) +static inline bool arch_mfns_in_directmap (unsigned long mfn,unsigned long nr) {return false; }xen/arch/arm/include/asm/arm64/mm.h +1
-1@@ -5,7 +5,7 @@* On ARM64, all the RAM is currently direct mapped in Xen. * Hence return always true. */ -static inline bool arch_mfn_in_directmap (unsigned long mfn) +static inline bool arch_mfns_in_directmap (unsigned long mfn,unsigned long nr) {return true; }xen/arch/x86/include/asm/mm.h +3
-3@@ -635,13 +635,13 @@ void write_32bit_pse_identmap(uint32_t *l2);/* * x86 maps part of physical memory via the directmap region. -* Return whether the input MFN falls in that range. +* Return whether the range of MFN falls in the directmap region. */ -static inline bool arch_mfn_in_directmap (unsigned long mfn) +static inline bool arch_mfns_in_directmap (unsigned long mfn,unsigned long nr) {unsigned long eva =min (DIRECTMAP_VIRT_END ,HYPERVISOR_VIRT_END ); -return mfn <= (virt_to_mfn (eva -1 ) +1 ); +return (mfn + nr) <= (virt_to_mfn (eva -1 ) +1 ); }#endif /* __ASM_X86_MM_H__ */ xen/common/page_alloc.c +2
-2@@ -588,7 +588,7 @@ static unsigned long init_node_heap(int node, unsigned long mfn, needed =0 ; }else if ( *use_tail && nr >= needed && -arch_mfn_in_directmap (mfn + nr) && +arch_mfns_in_directmap (mfn + nr - needed, needed) && (!xenheap_bits || !((mfn + nr -1 ) >> (xenheap_bits -PAGE_SHIFT ))) ) {@@ -597,7 +597,7 @@ static unsigned long init_node_heap(int node, unsigned long mfn,PAGE_SIZE -sizeof (**avail) *NR_ZONES ; }else if ( nr >= needed && -arch_mfn_in_directmap (mfn + needed) && +arch_mfns_in_directmap (mfn, needed) && (!xenheap_bits || !((mfn + needed -1 ) >> (xenheap_bits -PAGE_SHIFT ))) ) {Signature · 1 line
2.32.0