[PATCH] riscv: mm: Fix out-of-bounds page-table walk during memory hot-remove

[PATCH] riscv: mm: Fix out-of-bounds page-table walk during memory hot-remove

[Karl Mehltretter]

remove_pud_mapping() and remove_p4d_mapping() obtain a child table base
with pud_offset(p4dp, 0) and p4d_offset(pgd, 0), then add the index for
addr.

RISC-V folds page-table levels at runtime. When a level is folded, its
offset helper returns the parent entry itself, but the index can still be
nonzero. Adding it walks past the parent table. Sv48 folds P4D, while Sv39
folds both P4D and PUD, so memory hot-remove can descend into unrelated
memory and pass an invalid page to __free_pages(). This can trigger:

  kernel BUG at include/linux/mm.h:1810!
  VM_BUG_ON_PAGE(page_ref_count(page) == 0)
  arch_remove_memory+0x1e/0x5c
  try_remove_memory+0x15e/0x200
  remove_memory+0x24/0x3c

Only add the index when the corresponding page-table level is enabled,
matching p4d_offset() and pud_offset().

Fixes: c75a74f4ba19 ("riscv: mm: Add memory hotplugging support")
Assisted-by: Claude:claude-fable-5
Signed-off-by: Karl Mehltretter <kmehltretter@gmail.com>
---
 arch/riscv/mm/init.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/riscv/mm/init.c b/arch/riscv/mm/init.c
index 3e450890be07..338a8cbf4b80 100644
--- a/arch/riscv/mm/init.c
+++ b/arch/riscv/mm/init.c
@@ -1618,7 +1618,7 @@ static void __meminit remove_pud_mapping(pud_t *pud_base, unsigned long addr, un
 
 	for (; addr < end; addr = next) {
 		next = pud_addr_end(addr, end);
-		pudp = pud_base + pud_index(addr);
+		pudp = pgtable_l4_enabled ? pud_base + pud_index(addr) : pud_base;
 		pud = pudp_get(pudp);
 		if (!pud_present(pud))
 			continue;
@@ -1649,7 +1649,7 @@ static void __meminit remove_p4d_mapping(p4d_t *p4d_base, unsigned long addr, un
 
 	for (; addr < end; addr = next) {
 		next = p4d_addr_end(addr, end);
-		p4dp = p4d_base + p4d_index(addr);
+		p4dp = pgtable_l5_enabled ? p4d_base + p4d_index(addr) : p4d_base;
 		p4d = p4dp_get(p4dp);
 		if (!p4d_present(p4d))
 			continue;
-- 
2.53.0

[PATCH] riscv: mm: Fix out-of-bounds page-table walk during memory hot-remove

[Karl Mehltretter]

remove_pud_mapping() and remove_p4d_mapping() obtain a child table base
with pud_offset(p4dp, 0) and p4d_offset(pgd, 0), then add the index for
addr.

RISC-V folds page-table levels at runtime. When a level is folded, its
offset helper returns the parent entry itself, but the index can still be
nonzero. Adding it walks past the parent table. Sv48 folds P4D, while Sv39
folds both P4D and PUD, so memory hot-remove can descend into unrelated
memory and pass an invalid page to __free_pages(). This can trigger:

  kernel BUG at include/linux/mm.h:1810!
  VM_BUG_ON_PAGE(page_ref_count(page) == 0)
  arch_remove_memory+0x1e/0x5c
  try_remove_memory+0x15e/0x200
  remove_memory+0x24/0x3c

Only add the index when the corresponding page-table level is enabled,
matching p4d_offset() and pud_offset().

Fixes: c75a74f4ba19 ("riscv: mm: Add memory hotplugging support")
Assisted-by: Claude:claude-fable-5
Signed-off-by: Karl Mehltretter <kmehltretter@gmail.com>
---
 arch/riscv/mm/init.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/riscv/mm/init.c b/arch/riscv/mm/init.c
index 3e450890be07..338a8cbf4b80 100644
--- a/arch/riscv/mm/init.c
+++ b/arch/riscv/mm/init.c
@@ -1618,7 +1618,7 @@ static void __meminit remove_pud_mapping(pud_t *pud_base, unsigned long addr, un
 
 	for (; addr < end; addr = next) {
 		next = pud_addr_end(addr, end);
-		pudp = pud_base + pud_index(addr);
+		pudp = pgtable_l4_enabled ? pud_base + pud_index(addr) : pud_base;
 		pud = pudp_get(pudp);
 		if (!pud_present(pud))
 			continue;
@@ -1649,7 +1649,7 @@ static void __meminit remove_p4d_mapping(p4d_t *p4d_base, unsigned long addr, un
 
 	for (; addr < end; addr = next) {
 		next = p4d_addr_end(addr, end);
-		p4dp = p4d_base + p4d_index(addr);
+		p4dp = pgtable_l5_enabled ? p4d_base + p4d_index(addr) : p4d_base;
 		p4d = p4dp_get(p4dp);
 		if (!p4d_present(p4d))
 			continue;
-- 
2.53.0


_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv

Re: [PATCH] riscv: mm: Fix out-of-bounds page-table walk during memory hot-remove

[Paul Walmsley]

On Wed, 29 Jul 2026, Karl Mehltretter wrote:

> remove_pud_mapping() and remove_p4d_mapping() obtain a child table base
> with pud_offset(p4dp, 0) and p4d_offset(pgd, 0), then add the index for
> addr.
> 
> RISC-V folds page-table levels at runtime. When a level is folded, its
> offset helper returns the parent entry itself, but the index can still be
> nonzero. Adding it walks past the parent table. Sv48 folds P4D, while Sv39
> folds both P4D and PUD, so memory hot-remove can descend into unrelated
> memory and pass an invalid page to __free_pages(). This can trigger:
> 
>   kernel BUG at include/linux/mm.h:1810!
>   VM_BUG_ON_PAGE(page_ref_count(page) == 0)
>   arch_remove_memory+0x1e/0x5c
>   try_remove_memory+0x15e/0x200
>   remove_memory+0x24/0x3c
> 
> Only add the index when the corresponding page-table level is enabled,
> matching p4d_offset() and pud_offset().
> 
> Fixes: c75a74f4ba19 ("riscv: mm: Add memory hotplugging support")
> Assisted-by: Claude:claude-fable-5
> Signed-off-by: Karl Mehltretter <kmehltretter@gmail.com>

Thanks, queued for v7.2-rc.


- Paul

_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv

Re: [PATCH] riscv: mm: Fix out-of-bounds page-table walk during memory hot-remove

[Paul Walmsley]

On Wed, 29 Jul 2026, Karl Mehltretter wrote:

> remove_pud_mapping() and remove_p4d_mapping() obtain a child table base
> with pud_offset(p4dp, 0) and p4d_offset(pgd, 0), then add the index for
> addr.
> 
> RISC-V folds page-table levels at runtime. When a level is folded, its
> offset helper returns the parent entry itself, but the index can still be
> nonzero. Adding it walks past the parent table. Sv48 folds P4D, while Sv39
> folds both P4D and PUD, so memory hot-remove can descend into unrelated
> memory and pass an invalid page to __free_pages(). This can trigger:
> 
>   kernel BUG at include/linux/mm.h:1810!
>   VM_BUG_ON_PAGE(page_ref_count(page) == 0)
>   arch_remove_memory+0x1e/0x5c
>   try_remove_memory+0x15e/0x200
>   remove_memory+0x24/0x3c
> 
> Only add the index when the corresponding page-table level is enabled,
> matching p4d_offset() and pud_offset().
> 
> Fixes: c75a74f4ba19 ("riscv: mm: Add memory hotplugging support")
> Assisted-by: Claude:claude-fable-5
> Signed-off-by: Karl Mehltretter <kmehltretter@gmail.com>

Thanks, queued for v7.2-rc.


- Paul