vmemmap_populate() does not validate the requested vmemmap address range to be inside the platform assigned space i.e [VMEMMAP_START..VMEMMAP_END] for vmemmap. Instead it would just go ahead and create the mapping which might then overlap with other sections in the kernel virtual address space.
Just adding an warning here for range overrun which would help detect the problem earlier on, before a potential struct page corruption. This also makes vmemmap_populate() symmetrical with vmemmap_free() which already has a similar warning. Cc: Catalin Marinas <[email protected]> Cc: Will Deacon <[email protected]> Cc: [email protected] Cc: [email protected] Signed-off-by: Anshuman Khandual <[email protected]> --- This applies on v5.11-rc2 arch/arm64/mm/mmu.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/arch/arm64/mm/mmu.c b/arch/arm64/mm/mmu.c index ae0c3d023824..f938e8020523 100644 --- a/arch/arm64/mm/mmu.c +++ b/arch/arm64/mm/mmu.c @@ -1094,6 +1094,7 @@ static void free_empty_tables(unsigned long addr, unsigned long end, int __meminit vmemmap_populate(unsigned long start, unsigned long end, int node, struct vmem_altmap *altmap) { + WARN_ON((start < VMEMMAP_START) || (end > VMEMMAP_END)); return vmemmap_populate_basepages(start, end, node, altmap); } #else /* !ARM64_SWAPPER_USES_SECTION_MAPS */ @@ -1107,6 +1108,7 @@ int __meminit vmemmap_populate(unsigned long start, unsigned long end, int node, pud_t *pudp; pmd_t *pmdp; + WARN_ON((start < VMEMMAP_START) || (end > VMEMMAP_END)); do { next = pmd_addr_end(addr, end); -- 2.20.1
