From d4234d131b0a3f9e65973f1cdc71bb3560f5d14b Mon Sep 17 00:00:00 2001
From: Zhenhua Huang <quic_zhenhuah@quicinc.com>
Date: Tue, 4 Mar 2025 15:27:00 +0800
Subject: arm64: mm: Populate vmemmap at the page level if not section aligned
On the arm64 platform with 4K base page config, SECTION_SIZE_BITS is set
to 27, making one section 128M. The related page struct which vmemmap
points to is 2M then.
Commit c1cc1552616d ("arm64: MMU initialisation") optimizes the
vmemmap to populate at the PMD section level which was suitable
initially since hot plug granule is always one section(128M). However,
commit ba72b4c8cf60 ("mm/sparsemem: support sub-section hotplug")
introduced a 2M(SUBSECTION_SIZE) hot plug granule, which disrupted the
existing arm64 assumptions.
The first problem is that if start or end is not aligned to a section
boundary, such as when a subsection is hot added, populating the entire
section is wasteful.
The next problem is if we hotplug something that spans part of 128 MiB
section (subsections, let's call it memblock1), and then hotplug something
that spans another part of a 128 MiB section(subsections, let's call it
memblock2), and subsequently unplug memblock1, vmemmap_free() will clear
the entire PMD entry which also supports memblock2 even though memblock2
is still active.
Assuming hotplug/unplug sizes are guaranteed to be symmetric. Do the
fix similar to x86-64: populate to pages levels if start/end is not aligned
with section boundary.
Cc: stable@vger.kernel.org # v5.4+
Fixes: ba72b4c8cf60 ("mm/sparsemem: support sub-section hotplug")
Acked-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Zhenhua Huang <quic_zhenhuah@quicinc.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Link: https://lore.kernel.org/r/20250304072700.3405036-1-quic_zhenhuah@quicinc.com
Signed-off-by: Will Deacon <will@kernel.org>
---
arch/arm64/mm/mmu.c | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/arch/arm64/mm/mmu.c b/arch/arm64/mm/mmu.c
index b4df5bc5b1b8..1dfe1a8efdbe 100644
--- a/arch/arm64/mm/mmu.c
+++ b/arch/arm64/mm/mmu.c
@@ -1177,8 +1177,11 @@ int __meminit vmemmap_populate(unsigned long start, unsigned long end, int node,
struct vmem_altmap *altmap)
{
WARN_ON((start < VMEMMAP_START) || (end > VMEMMAP_END));
+ /* [start, end] should be within one section */
+ WARN_ON_ONCE(end - start > PAGES_PER_SECTION * sizeof(struct page));
- if (!IS_ENABLED(CONFIG_ARM64_4K_PAGES))
+ if (!IS_ENABLED(CONFIG_ARM64_4K_PAGES) ||
+ (end - start < PAGES_PER_SECTION * sizeof(struct page)))
return vmemmap_populate_basepages(start, end, node, altmap);
else
return vmemmap_populate_hugepages(start, end, node, altmap);
--
cgit v1.2.3
From f7edb07ad7c66eab3dce57384f33b9799d579133 Mon Sep 17 00:00:00 2001
From: Piotr Jaroszynski <pjaroszynski@nvidia.com>
Date: Tue, 4 Mar 2025 00:51:27 -0800
Subject: Fix mmu notifiers for range-based invalidates
Update the __flush_tlb_range_op macro not to modify its parameters as
these are unexepcted semantics. In practice, this fixes the call to
mmu_notifier_arch_invalidate_secondary_tlbs() in
__flush_tlb_range_nosync() to use the correct range instead of an empty
range with start=end. The empty range was (un)lucky as it results in
taking the invalidate-all path that doesn't cause correctness issues,
but can certainly result in suboptimal perf.
This has been broken since commit 6bbd42e2df8f ("mmu_notifiers: call
invalidate_range() when invalidating TLBs") when the call to the
notifiers was added to __flush_tlb_range(). It predates the addition of
the __flush_tlb_range_op() macro from commit 360839027a6e ("arm64: tlb:
Refactor the core flush algorithm of __flush_tlb_range") that made the
bug hard to spot.
Fixes: 6bbd42e2df8f ("mmu_notifiers: call invalidate_range() when invalidating TLBs")
Signed-off-by: Piotr Jaroszynski <pjaroszynski@nvidia.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will@kernel.org>
Cc: Robin Murphy <robin.murphy@arm.com>
Cc: Alistair Popple <apopple@nvidia.com>
Cc: Raghavendra Rao Ananta <rananta@google.com>
Cc: SeongJae Park <sj@kernel.org>
Cc: Jason Gunthorpe <jgg@nvidia.com>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: Nicolin Chen <nicolinc@nvidia.com>
Cc: linux-arm-kernel@lists.infradead.org
Cc: iommu@lists.linux.dev
Cc: linux-mm@kvack.org
Cc: linux-kernel@vger.kernel.org
Cc: stable@vger.kernel.org
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Reviewed-by: Alistair Popple <apopple@nvidia.com>
Link: https://lore.kernel.org/r/20250304085127.2238030-1-pjaroszynski@nvidia.com
Signed-off-by: Will Deacon <will@kernel.org>
---
arch/arm64/include/asm/tlbflush.h | 22 ++++++++++++----------
1 file changed, 12 insertions(+), 10 deletions(-)
diff --git a/arch/arm64/include/asm/tlbflush.h b/arch/arm64/include/asm/tlbflush.h
index bc94e036a26b..8104aee4f9a0 100644
--- a/arch/arm64/include/asm/tlbflush.h
+++ b/arch/arm64/include/asm/tlbflush.h
@@ -396,33 +396,35 @@ static inline void arch_tlbbatch_flush(struct arch_tlbflush_unmap_batch *batch)
#define __flush_tlb_range_op(op, start, pages, stride, \
asid, tlb_level, tlbi_user, lpa2) \
do { \
+ typeof(start) __flush_start = start; \
+ typeof(pages) __flush_pages = pages; \
int num = 0; \
int scale = 3; \
int shift = lpa2 ? 16 : PAGE_SHIFT; \
unsigned long addr; \
\
- while (pages > 0) { \
+ while (__flush_pages > 0) { \
if (!system_supports_tlb_range() || \
- pages == 1 || \
- (lpa2 && start != ALIGN(start, SZ_64K))) { \
- addr = __TLBI_VADDR(start, asid); \
+ __flush_pages == 1 || \
+ (lpa2 && __flush_start != ALIGN(__flush_start, SZ_64K))) { \
+ addr = __TLBI_VADDR(__flush_start, asid); \
__tlbi_level(op, addr, tlb_level); \
if (tlbi_user) \
__tlbi_user_level(op, addr, tlb_level); \
- start += stride; \
- pages -= stride >> PAGE_SHIFT; \
+ __flush_start += stride; \
+ __flush_pages -= stride >> PAGE_SHIFT; \
continue; \
} \
\
- num = __TLBI_RANGE_NUM(pages, scale); \
+ num = __TLBI_RANGE_NUM(__flush_pages, scale); \
if (num >= 0) { \
- addr = __TLBI_VADDR_RANGE(start >> shift, asid, \
+ addr = __TLBI_VADDR_RANGE(__flush_start >> shift, asid, \
scale, num, tlb_level); \
__tlbi(r##op, addr); \
if (tlbi_user) \
__tlbi_user(r##op, addr); \
- start += __TLBI_RANGE_PAGES(num, scale) << PAGE_SHIFT; \
- pages -= __TLBI_RANGE_PAGES(num, scale); \
+ __flush_start += __TLBI_RANGE_PAGES(num, scale) << PAGE_SHIFT; \
+ __flush_pages -= __TLBI_RANGE_PAGES(num, scale);\
} \
scale--; \
} \
--
cgit v1.2.3