mm: Fix possible NULL pointer dereference in __swap_duplicate
Subsystems (2)
PATCH v3 1/1 mm: Fix possible NULL pointer dereference in __swap_duplicate
Add a NULL check on the return value of swp_swap_info in __swap_duplicate to prevent crashes caused by NULL pointer dereference.
The reason why swp_swap_info() returns NULL is unclear; it may be due to CPU cache issues or DDR bit flips. The probability of this issue is very small, and the stack info we encountered is as follows:
Unable to handle kernel NULL pointer dereference at virtual address
0000000000000058
[RB/E]rb_sreason_str_set: sreason_str set null_pointer Mem abort info:
ESR = 0x0000000096000005
EC = 0x25: DABT (current EL), IL = 32 bits
SET = 0, FnV = 0
EA = 0, S1PTW = 0
FSC = 0x05: level 1 translation fault
Data abort info:
ISV = 0, ISS = 0x00000005, ISS2 = 0x00000000
CM = 0, WnR = 0, TnD = 0, TagAccess = 0
GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0
user pgtable: 4k pages, 39-bit VAs, pgdp=00000008a80e5000 [0000000000000058] pgd=0000000000000000, p4d=0000000000000000, pud=0000000000000000
Internal error: Oops: 0000000096000005 [#1] PREEMPT SMP Skip md ftrace buffer dump for: 0x1609e0 ... pc : swap_duplicate+0x44/0x164 lr : copy_page_range+0x508/0x1e78
sp : ffffffc0f2a699e0
x29: ffffffc0f2a699e0 x28: ffffff8a5b28d388 x27: ffffff8b06603388
x26: ffffffdf7291fe70 x25: 0000000000000006 x24: 0000000000100073
x23: 00000000002d2d2f x22: 0000000000000008 x21: 0000000000000000
x20: 00000000002d2d2f x19: 18000000002d2d2f x18: ffffffdf726faec0
x17: 0000000000000000 x16: 0010000000000001 x15: 0040000000000001
x14: 0400000000000001 x13: ff7ffffffffffb7f x12: ffeffffffffffbff
x11: ffffff8a5c7e1898 x10: 0000000000000018 x9 : 0000000000000006 x8 : 1800000000000000 x7 : 0000000000000000 x6 : ffffff8057c01f10 x5 : 000000000000a318 x4 : 0000000000000000 x3 : 0000000000000000 x2 : 0000006daf200000 x1 : 0000000000000001 x0 : 18000000002d2d2f
Call trace: swap_duplicate+0x44/0x164 copy_page_range+0x508/0x1e78 copy_process+0x1278/0x21cc kernel_clone+0x90/0x438 __arm64_sys_clone+0x5c/0x8c invoke_syscall+0x58/0x110 do_el0_svc+0x8c/0xe0 el0_svc+0x38/0x9c el0t_64_sync_handler+0x44/0xec el0t_64_sync+0x1a8/0x1ac Code: 9139c35a 71006f3f 54000568 f8797b55 (f9402ea8) ---[ end trace 0000000000000000 ]--- Kernel panic - not syncing: Oops: Fatal exception
SMP: stopping secondary CPUs
The patch seems to only provide a workaround, but there are no more effective software solutions to handle the bit flips problem. This path will change the issue from a system crash to a process exception, thereby reducing the impact on the entire machine.
---
v1 -> v2:
- Add WARN_ON_ONCE.
- update the commit info.
v2 -> v3: Delete the review tags (This is my issue, and I apologize).
---
mm/swapfile.c | 2 ++
1 file changed, 2 insertions(+)
mm/swapfile.c +2 -0
@@ -3521,6 +3521,8 @@ static int __swap_duplicate(swp_entry_t entry, unsigned char usage, int nr)int err, i; si =swp_swap_info (entry); +if (WARN_ON_ONCE (!si)) +return -EINVAL ; offset =swp_offset (entry);VM_WARN_ON (nr >SWAPFILE_CLUSTER - offset %SWAPFILE_CLUSTER );
Signature · 1 line
2.17.1
Quoted 39 lines · gaoxu
Add a NULL check on the return value of
swp_swap_infoin__swap_duplicateto prevent crashes caused by NULL pointer dereference.The reason why
swp_swap_info()returns NULL is unclear; it may be due to CPU cache issues or DDR bit flips. The probability of this issue is very small, and the stack info we encountered is as follows:Unable to handle kernel NULL pointer dereference at virtual address0000000000000058
[RB/E]rb_sreason_str_set:sreason_strsetnull_pointerMem abort info:
ESR =0x0000000096000005
EC =0x25: DABT (current EL), IL = 32 bits
SET = 0, FnV = 0
EA = 0, S1PTW = 0
FSC =0x05: level 1 translation fault
Data abort info:
ISV = 0, ISS =0x00000005, ISS2 =0x00000000
CM = 0, WnR = 0, TnD = 0, TagAccess = 0
GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0
user pgtable: 4k pages, 39-bit VAs, pgdp=00000008a80e5000 [0000000000000058] pgd=0000000000000000, p4d=0000000000000000, pud=0000000000000000Internal error: Oops: 0000000096000005 [#1] PREEMPT SMP Skip md ftrace buffer dump for: 0x1609e0 ... pc : swap_duplicate+0x44/0x164 lr : copy_page_range+0x508/0x1e78sp :
ffffffc0f2a699e0
x29:ffffffc0f2a699e0x28:ffffff8a5b28d388x27:ffffff8b06603388
x26:ffffffdf7291fe70x25: 0000000000000006 x24: 0000000000100073
x23:00000000002d2d2fx22: 0000000000000008 x21: 0000000000000000
x20:00000000002d2d2fx19:18000000002d2d2fx18:ffffffdf726faec0
x17: 0000000000000000 x16: 0010000000000001 x15: 0040000000000001
x14: 0400000000000001 x13:ff7ffffffffffb7fx12: ffeffffffffffbff
x11:ffffff8a5c7e1898x10: 0000000000000018 x9 : 0000000000000006 x8 : 1800000000000000 x7 : 0000000000000000 x6 :ffffff8057c01f10x5 :000000000000a318x4 : 0000000000000000 x3 : 0000000000000000 x2 :0000006daf200000x1 : 0000000000000001 x0 :18000000002d2d2f
Quoted 3 lines
Call trace: swap_duplicate+0x44/0x164 copy_page_range+0x508/0x1e78
This is really strange since we already have a swap entry check before calling swap_duplicate().
copy_nonpresent_pte(struct mm_struct *dst_mm, struct mm_struct *src_mm,
pte_t *dst_pte, pte_t *src_pte, struct vm_area_struct *dst_vma, struct vm_area_struct *src_vma, unsigned long addr, int *rss)
{
unsigned long vm_flags = dst_vma->vm_flags;
pte_t orig_pte = ptep_get(src_pte);
pte_t pte = orig_pte;
struct folio *folio;
struct page *page;
swp_entry_t entry = pte_to_swp_entry(orig_pte);
if (likely(!non_swap_entry(entry))) {
if (swap_duplicate(entry) < 0)
return -EIO;...
}
likely the swap_type is larger than MAX_SWAPFILES so we get a NULL?
static struct swap_info_struct *swap_type_to_swap_info(int type) {
if (type >= MAX_SWAPFILES)
return NULL;
return READ_ONCE(swap_info[type]); /* rcu_dereference() */
}
But non_swap_entry() guarantees that swp_type is smaller than MAX_SWAPFILES.
static inline int non_swap_entry(swp_entry_t entry)
{
return swp_type(entry) >= MAX_SWAPFILES;
}So another possibility is that we have an overflow of swap_info[] where type is < MAX_SWAPFILES but is not a valid existing swapfile?
I don't see how the current patch contributes to debugging or fixing anything related to
this dumped stack. Can we dump swp_type() as well?
Quoted 34 lines
copy_process+0x1278/0x21cc kernel_clone+0x90/0x438 __arm64_sys_clone+0x5c/0x8c invoke_syscall+0x58/0x110 do_el0_svc+0x8c/0xe0 el0_svc+0x38/0x9c el0t_64_sync_handler+0x44/0xec el0t_64_sync+0x1a8/0x1ac Code: 9139c35a 71006f3f 54000568 f8797b55 (f9402ea8) ---[ end trace 0000000000000000 ]--- Kernel panic - not syncing: Oops: Fatal exceptionSMP: stopping secondary CPUs
The patch seems to only provide a workaround, but there are no more effective software solutions to handle the bit flips problem. This path will change the issue from a system crash to a process exception, thereby reducing the impact on the entire machine.
Signed-off-by gao xu <[email protected]>---
v1 -> v2:
- AddWARN_ON_ONCE.
- update the commit info.
v2 -> v3: Delete the review tags (This is my issue, and I apologize).
---
mm/swapfile.c| 2 ++1 file changed, 2 insertions(+)mm/swapfile.c +0
-0
mm/swapfile.c static int __swap_duplicate(swp_entry_t entry, unsigned char usage, int nr) · line 3524
Quoted 4 lines
mm/swapfile.c +1
-0@@ -3521,6 +3521,8 @@ static int __swap_duplicate(swp_entry_t entry, unsigned char usage, int nr)int err, i; si =swp_swap_info (entry); +if (WARN_ON_ONCE (!si))
I mean, printk something related to swp_type(). This is really strange, but the current
stack won't help with debugging.
Quoted 6 lines
mm/swapfile.c +1
-1+return -EINVAL ; offset =swp_offset (entry);VM_WARN_ON (nr >SWAPFILE_CLUSTER - offset %SWAPFILE_CLUSTER ); --2.17.1
Thanks
Barry
Quoted 71 lines · gaoxu
Quoted 37 lines
Add a NULL check on the return value of
swp_swap_infoin__swap_duplicateto prevent crashes caused by NULL pointer dereference.The reason why
swp_swap_info()returns NULL is unclear; it may be due to CPU cache issues or DDR bit flips. The probability of this issue is very small, and the stack info we encountered is as follows:Unable to handle kernel NULL pointer dereference at virtual address0000000000000058
[RB/E]rb_sreason_str_set:sreason_strsetnull_pointerMem abort info:
ESR =0x0000000096000005
EC =0x25: DABT (current EL), IL = 32 bits
SET = 0, FnV = 0
EA = 0, S1PTW = 0
FSC =0x05: level 1 translation fault Data abort info:
ISV = 0, ISS =0x00000005, ISS2 =0x00000000
CM = 0, WnR = 0, TnD = 0, TagAccess = 0
GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0 user pgtable: 4k pages,
39-bit VAs, pgdp=00000008a80e5000 [0000000000000058] pgd=0000000000000000, p4d=0000000000000000, pud=0000000000000000Internal error: Oops: 0000000096000005 [#1] PREEMPT SMP Skip md ftrace buffer dump for: 0x1609e0 ... pc : swap_duplicate+0x44/0x164 lr : copy_page_range+0x508/0x1e78sp :
ffffffc0f2a699e0
x29:ffffffc0f2a699e0x28:ffffff8a5b28d388x27:ffffff8b06603388
x26:ffffffdf7291fe70x25: 0000000000000006 x24: 0000000000100073
x23:00000000002d2d2fx22: 0000000000000008 x21: 0000000000000000
x20:00000000002d2d2fx19:18000000002d2d2fx18:ffffffdf726faec0
x17: 0000000000000000 x16: 0010000000000001 x15: 0040000000000001
x14: 0400000000000001 x13:ff7ffffffffffb7fx12: ffeffffffffffbff
x11:ffffff8a5c7e1898x10: 0000000000000018 x9 : 0000000000000006 x8 : 1800000000000000 x7 : 0000000000000000 x6 :ffffff8057c01f10x5 :000000000000a318x4 : 0000000000000000 x3 : 0000000000000000 x2 :0000006daf200000x1 : 0000000000000001 x0 :18000000002d2d2fCall trace:Quoted 2 lines
swap_duplicate+0x44/0x164 copy_page_range+0x508/0x1e78This is really strange since we already have a swap entry check before calling
swap_duplicate().copy_nonpresent_pte(struct
mm_struct*dst_mm, structmm_struct*src_mm,
pte_t*dst_pte,pte_t*src_pte, structvm_area_struct*dst_vma,
structvm_area_struct*src_vma, unsigned long addr, int*rss) { unsigned long vm_flags = dst_vma->vm_flags; pte_t orig_pte = ptep_get(src_pte); pte_t pte = orig_pte; struct folio *folio; struct page *page; swp_entry_t entry = pte_to_swp_entry(orig_pte); if (likely(!non_swap_entry(entry))) { if (swap_duplicate(entry) < 0) return -EIO;...
}likely the
swap_typeis larger thanMAX_SWAPFILESso we get a NULL?static struct swap_info_struct *swap_type_to_swap_info(int type) { if (type >= MAX_SWAPFILES) return NULL; return READ_ONCE(swap_info[type]); /* rcu_dereference() */ }
Quoted 8 lines
But
non_swap_entry()guarantees thatswp_typeis smaller thanMAX_SWAPFILES.static inline int non_swap_entry(swp_entry_t entry) { return swp_type(entry) >= MAX_SWAPFILES; }So another possibility is that we have an overflow of swap_info[] where type is <
MAX_SWAPFILESbut is not a valid existing swapfile?
In the log of this issue, there is a printed entry: get_swap_device:
Bad swap file entry 18000000002d2d2f.
It can be calculated that swp_type(18000000002d2d2f) = 6.
In the Android 15-linux6.6:
system: MAX_SWAPFILES = 28, nr_swapfiles = 1.
Since swp_type(18000000002d2d2f)=6 is less than MAX_SWAPFILES but greater than nr_swapfiles, the value of this entry is abnormal.
static unsigned int nr_swapfiles;
static struct swap_info_struct *swap_info[MAX_SWAPFILES];swap_info is a static array, with its values initialized to 0.
The size of the array is MAX_SWAPFILES, and the size of valid values in the array is nr_swapfiles. Therefore, when we validate the validity of swp_type(entry), we should compare it with nr_swapfiles, not MAX_SWAPFILES.
The code for validating swp_type may need to be modified as follows:
static inline int non_swap_entry(swp_entry_t entry)
{
- return swp_type(entry) >= MAX_SWAPFILES;
+ return swp_type(entry) >= nr_swapfiles;
}
static struct swap_info_struct *swap_type_to_swap_info(int type) {
- if (type >= MAX_SWAPFILES)
+ if (type >= nr_swapfiles)
return NULL; return READ_ONCE(swap_info[type]); /* rcu_dereference() */ }
Quoted 39 lines
I don't see how the current patch contributes to debugging or fixing anything related to this dumped stack. Can we dump
swp_type()as well?mm/swapfile.c static int __swap_duplicate(swp_entry_t entry,
Quoted 34 lines
copy_process+0x1278/0x21cc kernel_clone+0x90/0x438 __arm64_sys_clone+0x5c/0x8c invoke_syscall+0x58/0x110 do_el0_svc+0x8c/0xe0 el0_svc+0x38/0x9c el0t_64_sync_handler+0x44/0xec el0t_64_sync+0x1a8/0x1ac Code: 9139c35a 71006f3f 54000568 f8797b55 (f9402ea8) ---[ end trace0000000000000000 ]--- Kernel panic - not syncing: Oops: Fatal exception
SMP: stopping secondary CPUsThe patch seems to only provide a workaround, but there are no more effective software solutions to handle the bit flips problem. This path will change the issue from a system crash to a process exception, thereby reducing the impact on the entire machine.
Signed-off-by gao xu <[email protected]>---
v1 -> v2:
- AddWARN_ON_ONCE.
- update the commit info.
v2 -> v3: Delete the review tags (This is my issue, and I apologize).
---
mm/swapfile.c| 2 ++1 file changed, 2 insertions(+)mm/swapfile.c index 7448a3876..a0bfdba94 renamed from mm/swapfile.c +0
-0100644
mm/swapfile.c +0
-0@@ -3521,6 +3521,8 @@ static int __swap_duplicate(swp_entry_t entry,unsigned char usage, int nr)
Quoted 7 lines
mm/swapfile.c static int __swap_duplicate(swp_entry_t entry, unsigned char usage, int nr) · line 3524
Quoted 4 lines
mm/swapfile.c +1
-0int err, i; si =swp_swap_info (entry); +if (WARN_ON_ONCE (!si))I mean, printk something related to
swp_type(). This is really strange, but the current stack won't help with debugging.
The log can find info related to "get_swap_device: Bad swap file entry xxx" when an entry encounters an exception.
Add a print info log like the following:pr_err("%s%08d\n", Bad swap type, swp_type(entry));
Quoted 11 lines
mm/swapfile.c static int __swap_duplicate(swp_entry_t entry, unsigned char usage, int nr)
Quoted 5 lines
mm/swapfile.c +1
-0+return -EINVAL ; offset =swp_offset (entry); VM_WARN_ON(nr > SWAPFILE_CLUSTER - offset %
SWAPFILE_CLUSTER);Quoted 2 lines
--
2.17.1Thanks
Barry
Thank you!
Quoted 88 lines · gaoxu
Quoted 72 lines · gaoxu
Quoted 37 lines
Add a NULL check on the return value of
swp_swap_infoin__swap_duplicateto prevent crashes caused by NULL pointer dereference.The reason why
swp_swap_info()returns NULL is unclear; it may be due to CPU cache issues or DDR bit flips. The probability of this issue is very small, and the stack info we encountered is as follows:Unable to handle kernel NULL pointer dereference at virtual address0000000000000058
[RB/E]rb_sreason_str_set:sreason_strsetnull_pointerMem abort info:
ESR =0x0000000096000005
EC =0x25: DABT (current EL), IL = 32 bits
SET = 0, FnV = 0
EA = 0, S1PTW = 0
FSC =0x05: level 1 translation fault Data abort info:
ISV = 0, ISS =0x00000005, ISS2 =0x00000000
CM = 0, WnR = 0, TnD = 0, TagAccess = 0
GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0 user pgtable: 4k pages,
39-bit VAs, pgdp=00000008a80e5000 [0000000000000058] pgd=0000000000000000, p4d=0000000000000000, pud=0000000000000000Internal error: Oops: 0000000096000005 [#1] PREEMPT SMP Skip md ftrace buffer dump for: 0x1609e0 ... pc : swap_duplicate+0x44/0x164 lr : copy_page_range+0x508/0x1e78sp :
ffffffc0f2a699e0
x29:ffffffc0f2a699e0x28:ffffff8a5b28d388x27:ffffff8b06603388
x26:ffffffdf7291fe70x25: 0000000000000006 x24: 0000000000100073
x23:00000000002d2d2fx22: 0000000000000008 x21: 0000000000000000
x20:00000000002d2d2fx19:18000000002d2d2fx18:ffffffdf726faec0
x17: 0000000000000000 x16: 0010000000000001 x15: 0040000000000001
x14: 0400000000000001 x13:ff7ffffffffffb7fx12: ffeffffffffffbff
x11:ffffff8a5c7e1898x10: 0000000000000018 x9 : 0000000000000006 x8 : 1800000000000000 x7 : 0000000000000000 x6 :ffffff8057c01f10x5 :000000000000a318x4 : 0000000000000000 x3 : 0000000000000000 x2 :0000006daf200000x1 : 0000000000000001 x0 :18000000002d2d2fCall trace:Quoted 2 lines
swap_duplicate+0x44/0x164 copy_page_range+0x508/0x1e78This is really strange since we already have a swap entry check before calling
swap_duplicate().copy_nonpresent_pte(struct
mm_struct*dst_mm, structmm_struct*src_mm,
pte_t*dst_pte,pte_t*src_pte, structvm_area_struct*dst_vma,
structvm_area_struct*src_vma, unsigned long addr, int*rss) { unsigned long vm_flags = dst_vma->vm_flags; pte_t orig_pte = ptep_get(src_pte); pte_t pte = orig_pte; struct folio *folio; struct page *page; swp_entry_t entry = pte_to_swp_entry(orig_pte); if (likely(!non_swap_entry(entry))) { if (swap_duplicate(entry) < 0) return -EIO;...
}likely the
swap_typeis larger thanMAX_SWAPFILESso we get a NULL?static struct swap_info_struct *swap_type_to_swap_info(int type) { if (type >= MAX_SWAPFILES) return NULL; return READ_ONCE(swap_info[type]); /* rcu_dereference() */ }Quoted 8 lines
But
non_swap_entry()guarantees thatswp_typeis smaller thanMAX_SWAPFILES.static inline int non_swap_entry(swp_entry_t entry) { return swp_type(entry) >= MAX_SWAPFILES; }So another possibility is that we have an overflow of swap_info[] where type is <
MAX_SWAPFILESbut is not a valid existing swapfile?In the log of this issue, there is a printed entry:
get_swap_device:
Bad swap file entry18000000002d2d2f.
It can be calculated thatswp_type(18000000002d2d2f)= 6.
In the Android 15-linux6.6:
system:MAX_SWAPFILES= 28,nr_swapfiles= 1.
Since swp_type(18000000002d2d2f)=6 is less thanMAX_SWAPFILESbut greater thannr_swapfiles, the value of this entry is abnormal.
Quoted 7 lines
static unsigned int
nr_swapfiles;
static structswap_info_struct*swap_info[MAX_SWAPFILES];swap_infois a static array, with its values initialized to 0.
The size of the array isMAX_SWAPFILES, and the size of valid values in the array isnr_swapfiles. Therefore, when we validate the validity ofswp_type(entry), we should compare it withnr_swapfiles, notMAX_SWAPFILES.
The code for validatingswp_typemay need to be modified as follows:
That might be true, but on a normal system, we only need to distinguish between a swap entry and a migrate entry. Therefore, comparing with MAX_SWAPFILES is sufficient.
Quoted 58 lines
static inline int
non_swap_entry(swp_entry_t entry){(quoted diff) +1
-1- return swp_type(entry) >= MAX_SWAPFILES; + return swp_type(entry) >= nr_swapfiles;}
static struct
swap_info_struct*swap_type_to_swap_info(int type) {(quoted diff) +1
-1- if (type >= MAX_SWAPFILES) + if (type >= nr_swapfiles) return NULL; return READ_ONCE(swap_info[type]); /* rcu_dereference() */}
Quoted 44 lines
I don't see how the current patch contributes to debugging or fixing anything related to this dumped stack. Can we dump
swp_type()as well?mm/swapfile.c static int __swap_duplicate(swp_entry_t entry,
Quoted 34 lines
copy_process+0x1278/0x21cc kernel_clone+0x90/0x438 __arm64_sys_clone+0x5c/0x8c invoke_syscall+0x58/0x110 do_el0_svc+0x8c/0xe0 el0_svc+0x38/0x9c el0t_64_sync_handler+0x44/0xec el0t_64_sync+0x1a8/0x1ac Code: 9139c35a 71006f3f 54000568 f8797b55 (f9402ea8) ---[ end trace0000000000000000 ]--- Kernel panic - not syncing: Oops: Fatal exception
SMP: stopping secondary CPUsThe patch seems to only provide a workaround, but there are no more effective software solutions to handle the bit flips problem. This path will change the issue from a system crash to a process exception, thereby reducing the impact on the entire machine.
Signed-off-by gao xu <[email protected]>---
v1 -> v2:
- AddWARN_ON_ONCE.
- update the commit info.
v2 -> v3: Delete the review tags (This is my issue, and I apologize).
---
mm/swapfile.c| 2 ++1 file changed, 2 insertions(+)mm/swapfile.c index 7448a3876..a0bfdba94 renamed from mm/swapfile.c +0
-0100644
mm/swapfile.c +0
-0@@ -3521,6 +3521,8 @@ static int __swap_duplicate(swp_entry_t entry,unsigned char usage, int nr)
Quoted 4 lines
mm/swapfile.c +1
-0int err, i; si =swp_swap_info (entry); +if (WARN_ON_ONCE (!si))
Quoted 6 lines
Quoted 2 lines
I mean, printk something related to
swp_type(). This is really strange, but the current stack won't help with debugging.The log can find info related to "
get_swap_device: Bad swap file entry xxx" when an entry encounters an exception.
Add a print info log like the following:pr_err("%s%08d\n", Bad swap type, swp_type(entry));
This is really strange. It would be better to have the entire PTE value dumped so we can determine if a bit-flip occurred on critical bits like PTE_PRESENT.
In that case, a present PTE could be misinterpreted as a swap entry.
On arm64,
/*
* Encode and decode a swap entry:
* bits 0-1: present (must be zero) * bits 2: remember PG_anon_exclusive * bits 3-7: swap type * bits 8-57: swap offset * bit 58: PTE_PROT_NONE (must be zero)
*/
#define __SWP_TYPE_SHIFT 3 #define __SWP_TYPE_BITS 5 #define __SWP_OFFSET_BITS 50 #define __SWP_TYPE_MASK ((1 << __SWP_TYPE_BITS) - 1) #define __SWP_OFFSET_SHIFT (__SWP_TYPE_BITS + __SWP_TYPE_SHIFT) #define __SWP_OFFSET_MASK ((1UL << __SWP_OFFSET_BITS) - 1)
_swp_type is bits3-7.
For a present pte, bits 3-7 are:
AP[7-6], NS[5], AttributeIndex[4-2].
Quoted 8 lines
Quoted 8 lines
mm/swapfile.c static int __swap_duplicate(swp_entry_t entry, unsigned char usage, int nr)
Quoted 5 lines
mm/swapfile.c +1
-0+return -EINVAL ; offset =swp_offset (entry); VM_WARN_ON(nr > SWAPFILE_CLUSTER - offset %
SWAPFILE_CLUSTER);Quoted 2 lines
--
2.17.1
Thanks
Barry
Quoted 172 lines
Thank you!
Quoted 6 lines · gaoxu
Quoted 6 lines · gaoxu
Quoted 3 lines
Add a NULL check on the return value of
swp_swap_infoin__swap_duplicateto prevent crashes caused by NULL pointerdereference.
Quoted 20 lines
Quoted 20 lines
Quoted 20 lines
The reason why
swp_swap_info()returns NULL is unclear; it may be due to CPU cache issues or DDR bit flips. The probability of this issue is very small, and the stack info we encountered is as follows:Unable to handle kernel NULL pointer dereference at virtualaddress
0000000000000058
[RB/E]rb_sreason_str_set:sreason_strsetnull_pointerMem abort info:
ESR =0x0000000096000005
EC =0x25: DABT (current EL), IL = 32 bits
SET = 0, FnV = 0
EA = 0, S1PTW = 0
FSC =0x05: level 1 translation fault Data abort info:
ISV = 0, ISS =0x00000005, ISS2 =0x00000000
CM = 0, WnR = 0, TnD = 0, TagAccess = 0
GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0 user pgtable: 4k
pages, 39-bit VAs, pgdp=00000008a80e5000 [0000000000000058] pgd=0000000000000000, p4d=0000000000000000, pud=0000000000000000Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Internal error: Oops: 0000000096000005 [#1] PREEMPT SMP Skip md ftrace buffer dump for: 0x1609e0 ... pc : swap_duplicate+0x44/0x164 lr : copy_page_range+0x508/0x1e78sp :
ffffffc0f2a699e0
x29:ffffffc0f2a699e0x28:ffffff8a5b28d388x27:ffffff8b06603388
x26:ffffffdf7291fe70x25: 0000000000000006 x24: 0000000000100073
x23:00000000002d2d2fx22: 0000000000000008 x21:0000000000000000
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
x20:
00000000002d2d2fx19:18000000002d2d2fx18:ffffffdf726faec0
x17: 0000000000000000 x16: 0010000000000001 x15:0040000000000001
Quoted 10 lines
Quoted 10 lines
Quoted 9 lines
x14: 0400000000000001 x13:
ff7ffffffffffb7fx12: ffeffffffffffbff
x11:ffffff8a5c7e1898x10: 0000000000000018 x9 : 0000000000000006 x8 : 1800000000000000 x7 : 0000000000000000 x6 :ffffff8057c01f10x5 :000000000000a318x4 : 0000000000000000 x3 : 0000000000000000 x2 :0000006daf200000x1 : 0000000000000001 x0 :18000000002d2d2fCall
trace:swap_duplicate+0x44/0x164 copy_page_range+0x508/0x1e78Quoted 4 lines
Quoted 4 lines
This is really strange since we already have a swap entry check before calling
swap_duplicate().copy_nonpresent_pte(struct
mm_struct*dst_mm, structmm_struct
*src_mm,Quoted 41 lines
Quoted 25 lines
pte_t*dst_pte,pte_t*src_pte, structvm_area_struct*dst_vma,
structvm_area_struct*src_vma, unsigned long addr,
int*rss) { unsigned long vm_flags = dst_vma->vm_flags; pte_t orig_pte = ptep_get(src_pte); pte_t pte = orig_pte; struct folio *folio; struct page *page; swp_entry_t entry = pte_to_swp_entry(orig_pte); if (likely(!non_swap_entry(entry))) { if (swap_duplicate(entry) < 0) return -EIO; ... }likely the
swap_typeis larger thanMAX_SWAPFILESso we get a NULL?static struct swap_info_struct *swap_type_to_swap_info(int type) { if (type >= MAX_SWAPFILES) return NULL; return READ_ONCE(swap_info[type]); /* rcu_dereference() */ }Quoted 8 lines
But
non_swap_entry()guarantees thatswp_typeis smaller thanMAX_SWAPFILES.static inline int non_swap_entry(swp_entry_t entry) { return swp_type(entry) >= MAX_SWAPFILES; }So another possibility is that we have an overflow of swap_info[] where type is <
MAX_SWAPFILESbut is not a valid existing swapfile?In the log of this issue, there is a printed entry:
get_swap_device:
Bad swap file entry18000000002d2d2f.
It can be calculated thatswp_type(18000000002d2d2f)= 6. In the Android 15-linux6.6:
system:MAX_SWAPFILES= 28,nr_swapfiles= 1.
Since swp_type(18000000002d2d2f)=6 is less thanMAX_SWAPFILESbut greater thannr_swapfiles, the value of this entry is abnormal.Quoted 6 lines
static unsigned int
nr_swapfiles;
static structswap_info_struct*swap_info[MAX_SWAPFILES];swap_infois a static array, with its values initialized to 0.
The size of the array isMAX_SWAPFILES, and the size of valid values in the array isnr_swapfiles. Therefore, when we validate the validity ofswp_type(entry), we should compare it withnr_swapfiles, not
MAX_SWAPFILES.Quoted 1 line
The code for validating
swp_typemay need to be modified as follows:That might be true, but on a normal system, we only need to distinguish between a swap entry and a migrate entry. Therefore, comparing with
MAX_SWAPFILESis sufficient.Quoted 60 lines
static inline int
non_swap_entry(swp_entry_t entry){(quoted diff) +1
-1- return swp_type(entry) >= MAX_SWAPFILES; + return swp_type(entry) >= nr_swapfiles;}
static struct
swap_info_struct*swap_type_to_swap_info(int type) {(quoted diff) +1
-1- if (type >= MAX_SWAPFILES) + if (type >= nr_swapfiles) return NULL; return READ_ONCE(swap_info[type]); /* rcu_dereference() */ }Quoted 49 lines
I don't see how the current patch contributes to debugging or fixing anything related to this dumped stack. Can we dump
swp_type()as well?Quoted 36 lines
copy_process+0x1278/0x21cc kernel_clone+0x90/0x438 __arm64_sys_clone+0x5c/0x8c invoke_syscall+0x58/0x110 do_el0_svc+0x8c/0xe0 el0_svc+0x38/0x9c el0t_64_sync_handler+0x44/0xec el0t_64_sync+0x1a8/0x1ac Code: 9139c35a 71006f3f 54000568 f8797b55 (f9402ea8) ---[ endtrace
0000000000000000 ]--- Kernel panic - not syncing: Oops: Fatal exception
SMP: stopping secondary CPUsThe patch seems to only provide a workaround, but there are no more effective software solutions to handle the bit flips problem. This path will change the issue from a system crash to a process exception, thereby reducing the impact on the entire machine.
Signed-off-by gao xu <[email protected]>---
v1 -> v2:
- AddWARN_ON_ONCE.
- update the commit info.
v2 -> v3: Delete the review tags (This is my issue, and I apologize).
---
mm/swapfile.c| 2 ++1 file changed, 2 insertions(+)mm/swapfile.c index renamed from mm/swapfile.c +0
-0
7448a3876..a0bfdba94
100644mm/swapfile.c static int __swap_duplicate(swp_entry_t · line 3521
Quoted 1 line
mm/swapfile.c +0
-0@@ -3521,6 +3521,8 @@ static int __swap_duplicate(swp_entry_t entry,unsigned char usage, int nr)
mm/swapfile.c static int __swap_duplicate(swp_entry_t entry, unsigned char usage, int nr) · line 3524
Quoted 4 lines
mm/swapfile.c +1
-0int err, i; si =swp_swap_info (entry); +if (WARN_ON_ONCE (!si))I mean, printk something related to
swp_type(). This is really strange, but the current stack won't help with debugging.
Quoted 7 lines
Quoted 4 lines
The log can find info related to "
get_swap_device: Bad swap file entry xxx" when an entry encounters an exception.
Add a print info log like the following:pr_err("%s%08d\n", Bad swap type, swp_type(entry));This is really strange. It would be better to have the entire PTE value dumped so we can determine if a bit-flip occurred on critical bits like
PTE_PRESENT.
Do you mean to convert the SWP entry to PTE and then print it out? pr_err("%s%08lx\n", Bad pte, pte_val(swp_entry_to_pte(entry)));
Or is it sufficient to print the SWP entry directly? pr_err("%s%08lx\n", Bad swap entry, entry.val);
Quoted 37 lines
In that case, a present PTE could be misinterpreted as a swap entry.
On arm64,
/*
* Encode and decode a swap entry:* bits 0-1: present (must be zero) * bits 2: remember PG_anon_exclusive * bits 3-7: swap type * bits 8-57: swap offset * bit 58: PTE_PROT_NONE (must be zero)*/
#define __SWP_TYPE_SHIFT 3 #define __SWP_TYPE_BITS 5 #define __SWP_OFFSET_BITS 50 #define __SWP_TYPE_MASK ((1 << __SWP_TYPE_BITS) - 1) #define __SWP_OFFSET_SHIFT (__SWP_TYPE_BITS +
__SWP_TYPE_SHIFT)
#define__SWP_OFFSET_MASK((1UL <<__SWP_OFFSET_BITS) - 1)
_swp_typeis bits3-7.For a present pte, bits 3-7 are:
AP[7-6], NS[5], AttributeIndex[4-2].Quoted 8 lines
Quoted 8 lines
Quoted 5 lines
mm/swapfile.c +1
-0+return -EINVAL ; offset =swp_offset (entry);VM_WARN_ON(nr >
SWAPFILE_CLUSTER- offset %
SWAPFILE_CLUSTER);Quoted 2 lines
--
2.17.1Thanks
Barry
Quoted 178 lines · gaoxu
Quoted 178 lines
Thank you!
Quoted 6 lines · gaoxu
Quoted 6 lines · gaoxu
Quoted 3 lines
Add a NULL check on the return value of
swp_swap_infoin__swap_duplicateto prevent crashes caused by NULL pointerdereference.
Quoted 20 lines
Quoted 20 lines
Quoted 20 lines
The reason why
swp_swap_info()returns NULL is unclear; it may be due to CPU cache issues or DDR bit flips. The probability of this issue is very small, and the stack info we encountered is as follows:Unable to handle kernel NULL pointer dereference at virtualaddress
0000000000000058
[RB/E]rb_sreason_str_set:sreason_strsetnull_pointerMem abort info:
ESR =0x0000000096000005
EC =0x25: DABT (current EL), IL = 32 bits
SET = 0, FnV = 0
EA = 0, S1PTW = 0
FSC =0x05: level 1 translation fault Data abort info:
ISV = 0, ISS =0x00000005, ISS2 =0x00000000
CM = 0, WnR = 0, TnD = 0, TagAccess = 0
GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0 user pgtable: 4k
pages, 39-bit VAs, pgdp=00000008a80e5000 [0000000000000058] pgd=0000000000000000, p4d=0000000000000000, pud=0000000000000000Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Internal error: Oops: 0000000096000005 [#1] PREEMPT SMP Skip md ftrace buffer dump for: 0x1609e0 ... pc : swap_duplicate+0x44/0x164 lr : copy_page_range+0x508/0x1e78sp :
ffffffc0f2a699e0
x29:ffffffc0f2a699e0x28:ffffff8a5b28d388x27:ffffff8b06603388
x26:ffffffdf7291fe70x25: 0000000000000006 x24: 0000000000100073
x23:00000000002d2d2fx22: 0000000000000008 x21:0000000000000000
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
x20:
00000000002d2d2fx19:18000000002d2d2fx18:ffffffdf726faec0
x17: 0000000000000000 x16: 0010000000000001 x15:0040000000000001
Quoted 10 lines
Quoted 10 lines
Quoted 9 lines
x14: 0400000000000001 x13:
ff7ffffffffffb7fx12: ffeffffffffffbff
x11:ffffff8a5c7e1898x10: 0000000000000018 x9 : 0000000000000006 x8 : 1800000000000000 x7 : 0000000000000000 x6 :ffffff8057c01f10x5 :000000000000a318x4 : 0000000000000000 x3 : 0000000000000000 x2 :0000006daf200000x1 : 0000000000000001 x0 :18000000002d2d2fCall
trace:swap_duplicate+0x44/0x164 copy_page_range+0x508/0x1e78Quoted 4 lines
Quoted 4 lines
This is really strange since we already have a swap entry check before calling
swap_duplicate().copy_nonpresent_pte(struct
mm_struct*dst_mm, structmm_struct
*src_mm,Quoted 41 lines
Quoted 25 lines
pte_t*dst_pte,pte_t*src_pte, structvm_area_struct*dst_vma,
structvm_area_struct*src_vma, unsigned long addr,
int*rss) { unsigned long vm_flags = dst_vma->vm_flags; pte_t orig_pte = ptep_get(src_pte); pte_t pte = orig_pte; struct folio *folio; struct page *page; swp_entry_t entry = pte_to_swp_entry(orig_pte); if (likely(!non_swap_entry(entry))) { if (swap_duplicate(entry) < 0) return -EIO; ... }likely the
swap_typeis larger thanMAX_SWAPFILESso we get a NULL?static struct swap_info_struct *swap_type_to_swap_info(int type) { if (type >= MAX_SWAPFILES) return NULL; return READ_ONCE(swap_info[type]); /* rcu_dereference() */ }Quoted 8 lines
But
non_swap_entry()guarantees thatswp_typeis smaller thanMAX_SWAPFILES.static inline int non_swap_entry(swp_entry_t entry) { return swp_type(entry) >= MAX_SWAPFILES; }So another possibility is that we have an overflow of swap_info[] where type is <
MAX_SWAPFILESbut is not a valid existing swapfile?In the log of this issue, there is a printed entry:
get_swap_device:
Bad swap file entry18000000002d2d2f.
It can be calculated thatswp_type(18000000002d2d2f)= 6. In the Android 15-linux6.6:
system:MAX_SWAPFILES= 28,nr_swapfiles= 1.
Since swp_type(18000000002d2d2f)=6 is less thanMAX_SWAPFILESbut greater thannr_swapfiles, the value of this entry is abnormal.Quoted 6 lines
static unsigned int
nr_swapfiles;
static structswap_info_struct*swap_info[MAX_SWAPFILES];swap_infois a static array, with its values initialized to 0.
The size of the array isMAX_SWAPFILES, and the size of valid values in the array isnr_swapfiles. Therefore, when we validate the validity ofswp_type(entry), we should compare it withnr_swapfiles, not
MAX_SWAPFILES.Quoted 1 line
The code for validating
swp_typemay need to be modified as follows:That might be true, but on a normal system, we only need to distinguish between a swap entry and a migrate entry. Therefore, comparing with
MAX_SWAPFILESis sufficient.Quoted 64 lines
static inline int
non_swap_entry(swp_entry_t entry){(quoted diff) +1
-1- return swp_type(entry) >= MAX_SWAPFILES; + return swp_type(entry) >= nr_swapfiles;}
static struct
swap_info_struct*swap_type_to_swap_info(int type) {(quoted diff) +1
-1- if (type >= MAX_SWAPFILES) + if (type >= nr_swapfiles) return NULL; return READ_ONCE(swap_info[type]); /* rcu_dereference() */ }Quoted 42 lines
I don't see how the current patch contributes to debugging or fixing anything related to this dumped stack. Can we dump
swp_type()as well?Quoted 36 lines
copy_process+0x1278/0x21cc kernel_clone+0x90/0x438 __arm64_sys_clone+0x5c/0x8c invoke_syscall+0x58/0x110 do_el0_svc+0x8c/0xe0 el0_svc+0x38/0x9c el0t_64_sync_handler+0x44/0xec el0t_64_sync+0x1a8/0x1ac Code: 9139c35a 71006f3f 54000568 f8797b55 (f9402ea8) ---[ endtrace
0000000000000000 ]--- Kernel panic - not syncing: Oops: Fatal exception
SMP: stopping secondary CPUsThe patch seems to only provide a workaround, but there are no more effective software solutions to handle the bit flips problem. This path will change the issue from a system crash to a process exception, thereby reducing the impact on the entire machine.
Signed-off-by gao xu <[email protected]>---
v1 -> v2:
- AddWARN_ON_ONCE.
- update the commit info.
v2 -> v3: Delete the review tags (This is my issue, and I apologize).
---
mm/swapfile.c| 2 ++1 file changed, 2 insertions(+)mm/swapfile.c index renamed from mm/swapfile.c +0
-0
7448a3876..a0bfdba94
100644mm/swapfile.c static int __swap_duplicate(swp_entry_t · line 3521
Quoted 1 line
mm/swapfile.c +0
-0@@ -3521,6 +3521,8 @@ static int __swap_duplicate(swp_entry_t entry,unsigned char usage, int nr)
Quoted 7 lines
mm/swapfile.c static int __swap_duplicate(swp_entry_t entry, unsigned char usage, int nr) · line 3524
Quoted 4 lines
mm/swapfile.c +1
-0int err, i; si =swp_swap_info (entry); +if (WARN_ON_ONCE (!si))I mean, printk something related to
swp_type(). This is really strange, but the current stack won't help with debugging.The log can find info related to "
get_swap_device: Bad swap file entry xxx" when an entry encounters an exception.
Add a print info log like the following:pr_err("%s%08d\n", Bad swap type, swp_type(entry));
Quoted 7 lines
Quoted 2 lines
This is really strange. It would be better to have the entire PTE value dumped so we can determine if a bit-flip occurred on critical bits like
PTE_PRESENT.Do you mean to convert the SWP entry to PTE and then print it out?
pr_err("%s%08lx\n", Bad pte, pte_val(swp_entry_to_pte(entry)));Or is it sufficient to print the SWP entry directly?
pr_err("%s%08lx\n", Bad swap entry, entry.val);
Yes, I think so. With that, we can convert it to PTE offline and debug using that value.
By the way, I don’t have a strong opinion on whether this patch gets merged or not, but it’s still nice to have. :)
I’m more interested in the bug itself and curious whether other Android products using the same kernel will encounter the same issue.
Quoted 27 lines
Quoted 27 lines
In that case, a present PTE could be misinterpreted as a swap entry.
On arm64,
/*
* Encode and decode a swap entry:* bits 0-1: present (must be zero) * bits 2: remember PG_anon_exclusive * bits 3-7: swap type * bits 8-57: swap offset * bit 58: PTE_PROT_NONE (must be zero)*/
#define __SWP_TYPE_SHIFT 3 #define __SWP_TYPE_BITS 5 #define __SWP_OFFSET_BITS 50 #define __SWP_TYPE_MASK ((1 << __SWP_TYPE_BITS) - 1) #define __SWP_OFFSET_SHIFT (__SWP_TYPE_BITS +
__SWP_TYPE_SHIFT)
#define__SWP_OFFSET_MASK((1UL <<__SWP_OFFSET_BITS) - 1)
_swp_typeis bits3-7.For a present pte, bits 3-7 are:
AP[7-6], NS[5], AttributeIndex[4-2].Quoted 1 line
Quoted 1 line
Quoted 8 lines
Quoted 8 lines
Quoted 7 lines
Quoted 7 lines
Quoted 4 lines
mm/swapfile.c +1
-0+return -EINVAL ; offset =swp_offset (entry);VM_WARN_ON(nr >
SWAPFILE_CLUSTER- offset %
SWAPFILE_CLUSTER);Quoted 2 lines
--
2.17.1
Thanks
Barry