On 5/30/25 14:10, Ryan Roberts wrote:

On 30/05/2025 09:20, Dev Jain wrote:

quoted

arm64 disables vmalloc-huge when kernel page table dumping is enabled,
because an intermediate table may be removed, potentially causing the
ptdump code to dereference an invalid address. We want to be able to
analyze block vs page mappings for kernel mappings with ptdump, so to
enable vmalloc-huge with ptdump, synchronize between page table removal in
pmd_free_pte_page()/pud_free_pmd_page() and ptdump pagetable walking. We
use mmap_read_lock and not write lock because we don't need to synchronize
between two different vm_structs; two vmalloc objects running this same
code path will point to different page tables, hence there is no race. 

Signed-off-by: Dev Jain <dev.jain@arm.com>
---
 arch/arm64/include/asm/vmalloc.h | 6 ++----
 arch/arm64/mm/mmu.c              | 7 +++++++
 2 files changed, 9 insertions(+), 4 deletions(-)

diff --git a/arch/arm64/include/asm/vmalloc.h b/arch/arm64/include/asm/vmalloc.h
index 38fafffe699f..28b7173d8693 100644
--- a/arch/arm64/include/asm/vmalloc.h
+++ b/arch/arm64/include/asm/vmalloc.h

@@ -12,15 +12,13 @@ static inline bool arch_vmap_pud_supported(pgprot_t prot)
 	/*
 	 * SW table walks can't handle removal of intermediate entries.
 	 */
-	return pud_sect_supported() &&
-	       !IS_ENABLED(CONFIG_PTDUMP_DEBUGFS);
+	return pud_sect_supported();
 }
 
 #define arch_vmap_pmd_supported arch_vmap_pmd_supported
 static inline bool arch_vmap_pmd_supported(pgprot_t prot)
 {
-	/* See arch_vmap_pud_supported() */
-	return !IS_ENABLED(CONFIG_PTDUMP_DEBUGFS);
+	return true;
 }
 
 #endif

diff --git a/arch/arm64/mm/mmu.c b/arch/arm64/mm/mmu.c
index ea6695d53fb9..798cebd9e147 100644
--- a/arch/arm64/mm/mmu.c
+++ b/arch/arm64/mm/mmu.c

@@ -1261,7 +1261,11 @@ int pmd_free_pte_page(pmd_t *pmdp, unsigned long addr)
 	}
 
 	table = pte_offset_kernel(pmdp, addr);
+
+	/* Synchronize against ptdump_walk_pgd() */
+	mmap_read_lock(&init_mm);
 	pmd_clear(pmdp);
+	mmap_read_unlock(&init_mm);

So this works because ptdump_walk_pgd() takes the write_lock (which is mutually
exclusive with any read_lock holders) for the duration of the table walk, so it
will either consistently see the pgtables before or after this removal. It will
never disappear during the walk, correct?

Agreed.

I guess there is a risk of this showing up as contention with other init_mm
write_lock holders. But I expect that pmd_free_pte_page()/pud_free_pmd_page()
are called sufficiently rarely that the risk is very small. Let's fix any perf
problem if/when we see it.

Checking against CONFIG_PTDUMP_DEBUGFS being enabled is simple enough without much
cost. So why not make this conditional only for scenarios, where this read lock is
really required. Something like

--- a/arch/arm64/mm/mmu.c
+++ b/arch/arm64/mm/mmu.c

@@ -1293,11 +1293,15 @@ int pud_free_pmd_page(pud_t *pudp, unsigned long addr)
                pmd_free_pte_page(pmdp, next);
        } while (pmdp++, next += PMD_SIZE, next != end);
 
-       /* Synchronize against ptdump_walk_pgd() */
-       mmap_read_lock(&init_mm);
-       pud_clear(pudp);
-       mmap_read_unlock(&init_mm);
        __flush_tlb_kernel_pgtable(addr);
+       if (IS_ENABLED(CONFIG_PTDUMP_DEBUGFS)) {
+               /* Synchronize against ptdump_walk_pgd() */
+               mmap_read_lock(&init_mm);
+               pud_clear(pudp);
+               mmap_read_unlock(&init_mm);
+       } else {
+               pud_clear(pudp);
+       }
        pmd_free(NULL, table);
        return 1;
 }

quoted

 	__flush_tlb_kernel_pgtable(addr);

And the tlbi doesn't need to be serialized because there is no security issue.
The walker can be trusted to only dereference memory that it sees as it walks
the pgtable (obviously).

Agreed.

quoted

 	pte_free_kernel(NULL, table);
 	return 1;

@@ -1289,7 +1293,10 @@ int pud_free_pmd_page(pud_t *pudp, unsigned long addr)
 		pmd_free_pte_page(pmdp, next);
 	} while (pmdp++, next += PMD_SIZE, next != end);
 
+	/* Synchronize against ptdump_walk_pgd() */
+	mmap_read_lock(&init_mm);
 	pud_clear(pudp);
+	mmap_read_unlock(&init_mm);

Hmm, so pud_free_pmd_page() is now going to cause us to acquire and release the
(upto) lock 513 times (for a 4K kernel). I wonder if there is an argument for
clearing the pud first (under the lock), then the pmds can all be cleared
without a lock, since the walker won't be able to see the pmds once the pud is
cleared.

Makes sense if pud_free_pmd_page() would have been the only caller but seems like
vmap_try_huge_pmd() calls pmd_free_pte_page() directly as well.

Thanks,
Ryan

quoted

 	__flush_tlb_kernel_pgtable(addr);
 	pmd_free(NULL, table);
 	return 1;