md/raid5: validate payload size before calculating payload_len
Subsystems (2)
PATCH v3 1/1 md/raid5: validate payload size before calculating payload_len
Commit b0cc3ae97e89 ("md/raid5: validate payload size before accessing journal metadata") introduced a bounds check to verify that a given payload fits into the current metadata block. In order to calculate the payload size, it has to access the members payload->header.type and payload->size, the offset of which may already be past the end of the metadata block, because the loop condition only checks that the first byte of the payload is inside, and the payload entries have variable sizes. Later on, payload->checksum[0] and payload->checksum[1] are accessed without verifying that payload->size is large enough.
Add another check to make sure that payload->size can be safely accessed, and make sure that the checksum fields are accessible.
This issue has been found by AI (gemma4, Gemini) during a backport review.
---
Changes v2 -> v3: Fixed another issue reported by Sashiko
* Modify loop condition to make sure the payload->header.type can be accessed
Changes v1 -> v2: Fixed issues reported by Sashiko
* fix use of wrong variable name payload_len instead of mb_offset
* add checks to ensure payload->size is large enough to hold the checksum values
--- drivers/md/raid5-cache.c | 34 ++++++++++++++++++++++++++-------- 1 file changed, 26 insertions(+), 8 deletions(-)
drivers/md/raid5-cache.c +26 -8
@@ -2001,27 +2001,36 @@ r5l_recovery_verify_data_checksum_for_mb(struct r5l_log *log,if (!page)return -ENOMEM ; -while (mb_offset <le32_to_cpu (mb->meta_size)) { +while (mb_offset +sizeof (struct r5l_payload_header ) <le32_to_cpu (mb->meta_size)) { +uint32_t payload_size;sector_t payload_len; payload = (void *)mb + mb_offset; payload_flush = (void *)mb + mb_offset;if (le16_to_cpu (payload->header.type) ==R5LOG_PAYLOAD_DATA ) { +if (mb_offset +sizeof (struct r5l_payload_data_parity ) + >le32_to_cpu (mb->meta_size)) + goto mismatch; + payload_size =le32_to_cpu (payload->size) >> (PAGE_SHIFT -9 ); payload_len =sizeof (struct r5l_payload_data_parity ) + - (sector_t )sizeof (__le32) * - (le32_to_cpu (payload->size) >> (PAGE_SHIFT -9 )); -if (mb_offset + payload_len >le32_to_cpu (mb->meta_size)) + (sector_t )sizeof (__le32) * payload_size; +if (mb_offset + payload_len >le32_to_cpu (mb->meta_size) || + payload_size <1 ) goto mismatch;if (r5l_recovery_verify_data_checksum ( log, ctx, page, log_offset, payload->checksum[0 ]) <0 ) goto mismatch; }else if (le16_to_cpu (payload->header.type) ==R5LOG_PAYLOAD_PARITY ) { +if (mb_offset +sizeof (struct r5l_payload_data_parity ) + >le32_to_cpu (mb->meta_size)) + goto mismatch; + payload_size =le32_to_cpu (payload->size) >> (PAGE_SHIFT -9 ); payload_len =sizeof (struct r5l_payload_data_parity ) + - (sector_t )sizeof (__le32) * - (le32_to_cpu (payload->size) >> (PAGE_SHIFT -9 )); -if (mb_offset + payload_len >le32_to_cpu (mb->meta_size)) + (sector_t )sizeof (__le32) * payload_size; +if (mb_offset + payload_len >le32_to_cpu (mb->meta_size) || + payload_size < conf->max_degraded) goto mismatch; if (r5l_recovery_verify_data_checksum( log, ctx, page, log_offset,
@@ -2035,6 +2044,9 @@ r5l_recovery_verify_data_checksum_for_mb(struct r5l_log *log, payload->checksum[1]) < 0) goto mismatch; }else if (le16_to_cpu (payload->header.type) ==R5LOG_PAYLOAD_FLUSH ) { +if (mb_offset +sizeof (struct r5l_payload_flush ) + >le32_to_cpu (mb->meta_size)) + goto mismatch; payload_len =sizeof (struct r5l_payload_flush ) + (sector_t )le32_to_cpu (payload_flush->size);if (mb_offset + payload_len >le32_to_cpu (mb->meta_size))
@@ -2096,7 +2108,7 @@ r5c_recovery_analyze_meta_block(struct r5l_log *log, mb_offset =sizeof (struct r5l_meta_block ); log_offset =r5l_ring_add (log, ctx->pos,BLOCK_SECTORS ); -while (mb_offset <le32_to_cpu (mb->meta_size)) { +while (mb_offset +sizeof (struct r5l_payload_header ) <le32_to_cpu (mb->meta_size)) {sector_t payload_len;int dd;
@@ -2106,6 +2118,9 @@ r5c_recovery_analyze_meta_block(struct r5l_log *log,if (le16_to_cpu (payload->header.type) ==R5LOG_PAYLOAD_FLUSH ) {int i, count; +if (mb_offset +sizeof (struct r5l_payload_flush ) > +le32_to_cpu (mb->meta_size)) +return -EINVAL ; payload_len =sizeof (struct r5l_payload_flush ) + (sector_t )le32_to_cpu (payload_flush->size); if (mb_offset + payload_len >
@@ -2130,6 +2145,9 @@ r5c_recovery_analyze_meta_block(struct r5l_log *log, }/* DATA or PARITY payload */ +if (mb_offset +sizeof (struct r5l_payload_data_parity ) > +le32_to_cpu (mb->meta_size)) +return -EINVAL ; payload_len =sizeof (struct r5l_payload_data_parity ) + (sector_t )sizeof (__le32) * (le32_to_cpu (payload->size) >> (PAGE_SHIFT -9 ));
Signature · 1 line
2.51.0
Hi Martin,
Thanks for fixing the follow-up out-of-bounds accesses. The diff looks good to me.
Thanks,
Junrui
Quoted 126 lines · Martin Wilck
Commit
b0cc3ae97e89("md/raid5: validate payload size before accessing journal metadata") introduced a bounds check to verify that a given payload fits into the current metadata block. In order to calculate the payload size, it has to access the memberspayload->header.typeandpayload->size, the offset of which may already be past the end of the metadata block, because the loop condition only checks that the first byte of the payload is inside, and the payload entries have variable sizes. Later on, payload->checksum[0] and payload->checksum[1] are accessed without verifying thatpayload->sizeis large enough.Add another check to make sure that
payload->sizecan be safely accessed, and make sure that the checksum fields are accessible.This issue has been found by AI (gemma4, Gemini) during a backport review.
Cc Junrui Luo <[email protected]>Fixes b0cc3ae97e89 ("md/raid5: validate payload size before accessing journal metadata")Signed-off-by Martin Wilck <[email protected]>---
Changes v2 -> v3: Fixed another issue reported by Sashiko* Modify loop condition to make sure the
payload->header.typecan be accessedChanges v1 -> v2: Fixed issues reported by Sashiko
* fix use of wrong variable name
payload_leninstead ofmb_offset
* add checks to ensurepayload->sizeis large enough to hold the checksum values---
drivers/md/raid5-cache.c| 34 ++++++++++++++++++++++++++--------1 file changed, 26 insertions(+), 8 deletions(-)drivers/md/raid5-cache.c +6
-1@@ -2001,27 +2001,36 @@ r5l_recovery_verify_data_checksum_for_mb(struct r5l_log *log,if (!page)return -ENOMEM ; -while (mb_offset <le32_to_cpu (mb->meta_size)) { +while (mb_offset +sizeof (struct r5l_payload_header ) <le32_to_cpu (mb->meta_size)) { +uint32_t payload_size;sector_t payload_len; payload = (void *)mb + mb_offset; payload_flush = (void *)mb + mb_offset;if (le16_to_cpu (payload->header.type) ==R5LOG_PAYLOAD_DATA ) { +if (mb_offset +sizeof (struct r5l_payload_data_parity ) + >le32_to_cpu (mb->meta_size)) + goto mismatch; + payload_size =le32_to_cpu (payload->size) >> (PAGE_SHIFT -9 );
payload_len=sizeof(struct r5l_payload_data_parity)+
- (sector_t)sizeof(__le32) *
- (le32_to_cpu(payload->size)>> (PAGE_SHIFT- 9));
- if (mb_offset+payload_len>le32_to_cpu(mb->meta_size))
+ (sector_t)sizeof(__le32) *payload_size;
+ if (mb_offset+payload_len>le32_to_cpu(mb->meta_size)||
+payload_size< 1)
goto mismatch;
if (r5l_recovery_verify_data_checksum(
log, ctx, page,log_offset,
payload->checksum[0]) < 0)
goto mismatch;
} else if (le16_to_cpu(payload->header.type)==R5LOG_PAYLOAD_PARITY) {
+ if (mb_offset+sizeof(struct r5l_payload_data_parity)
+ >le32_to_cpu(mb->meta_size))
+ goto mismatch;
+payload_size=le32_to_cpu(payload->size)>> (PAGE_SHIFT- 9);payload_len=sizeof(struct r5l_payload_data_parity)+
- (sector_t)sizeof(__le32) *
- (le32_to_cpu(payload->size)>> (PAGE_SHIFT- 9));
- if (mb_offset+payload_len>le32_to_cpu(mb->meta_size))
+ (sector_t)sizeof(__le32) *payload_size;
+ if (mb_offset+payload_len>le32_to_cpu(mb->meta_size)||
+payload_size<conf->max_degraded)
goto mismatch;
if (r5l_recovery_verify_data_checksum(
log, ctx, page,log_offset,drivers/md/raid5-cache.c +3
-0@@ -2035,6 +2044,9 @@ r5l_recovery_verify_data_checksum_for_mb(struct r5l_log *log, payload->checksum[1]) < 0) goto mismatch; }else if (le16_to_cpu (payload->header.type) ==R5LOG_PAYLOAD_FLUSH ) { +if (mb_offset +sizeof (struct r5l_payload_flush ) + >le32_to_cpu (mb->meta_size)) + goto mismatch;
payload_len=sizeof(struct r5l_payload_flush)+ (sector_t)le32_to_cpu(payload_flush->size);
if (mb_offset+payload_len>le32_to_cpu(mb->meta_size))drivers/md/raid5-cache.c +4
-1@@ -2096,7 +2108,7 @@ r5c_recovery_analyze_meta_block(struct r5l_log *log, mb_offset =sizeof (struct r5l_meta_block ); log_offset =r5l_ring_add (log, ctx->pos,BLOCK_SECTORS ); -while (mb_offset <le32_to_cpu (mb->meta_size)) { +while (mb_offset +sizeof (struct r5l_payload_header ) <le32_to_cpu (mb->meta_size)) {sector_t payload_len;int dd;@@ -2106,6 +2118,9 @@ r5c_recovery_analyze_meta_block(struct r5l_log *log,if (le16_to_cpu (payload->header.type) ==R5LOG_PAYLOAD_FLUSH ) {int i, count; +if (mb_offset +sizeof (struct r5l_payload_flush ) > +le32_to_cpu (mb->meta_size)) +return -EINVAL ;
payload_len=sizeof(struct r5l_payload_flush)+
(sector_t)le32_to_cpu(payload_flush->size);
if (mb_offset+payload_len>drivers/md/raid5-cache.c +3
-0@@ -2130,6 +2145,9 @@ r5c_recovery_analyze_meta_block(struct r5l_log *log, }/* DATA or PARITY payload */ +if (mb_offset +sizeof (struct r5l_payload_data_parity ) > +le32_to_cpu (mb->meta_size)) +return -EINVAL ;
payload_len=sizeof(struct r5l_payload_data_parity)+ (sector_t)sizeof(__le32) * (le32_to_cpu(payload->size)>> (PAGE_SHIFT- 9));Signature · 1 line
2.51.0
yu kuai
maintainerHi,
Quoted 39 lines · Martin Wilck
Commit
b0cc3ae97e89("md/raid5: validate payload size before accessing journal metadata") introduced a bounds check to verify that a given payload fits into the current metadata block. In order to calculate the payload size, it has to access the memberspayload->header.typeandpayload->size, the offset of which may already be past the end of the metadata block, because the loop condition only checks that the first byte of the payload is inside, and the payload entries have variable sizes. Later on, payload->checksum[0] and payload->checksum[1] are accessed without verifying thatpayload->sizeis large enough.Add another check to make sure that
payload->sizecan be safely accessed, and make sure that the checksum fields are accessible.This issue has been found by AI (gemma4, Gemini) during a backport review.
Cc Junrui Luo <[email protected]>Fixes b0cc3ae97e89 ("md/raid5: validate payload size before accessing journal metadata")Signed-off-by Martin Wilck <[email protected]>---
Changes v2 -> v3: Fixed another issue reported by Sashiko* Modify loop condition to make sure the
payload->header.typecan be accessedChanges v1 -> v2: Fixed issues reported by Sashiko
* fix use of wrong variable name
payload_leninstead ofmb_offset
* add checks to ensurepayload->sizeis large enough to hold the checksum values--- drivers/md/raid5-cache.c | 34 ++++++++++++++++++++++++++-------- 1 file changed, 26 insertions(+), 8 deletions(-)drivers/md/raid5-cache.c +0
-0
drivers/md/raid5-cache.c r5l_recovery_verify_data_checksum_for_mb(struct r5l_log *log, · line 2004
Quoted 5 lines
drivers/md/raid5-cache.c +1
-1@@ -2001,27 +2001,36 @@ r5l_recovery_verify_data_checksum_for_mb(struct r5l_log *log,if (!page)return -ENOMEM ; -while (mb_offset <le32_to_cpu (mb->meta_size)) { +while (mb_offset +sizeof (struct r5l_payload_header ) <le32_to_cpu (mb->meta_size)) {
Offset can't be meta_size, however, is it possible that the remaining space is exactly one payload size? If so I think <= should be used here.
drivers/md/raid5-cache.c r5l_recovery_verify_data_checksum_for_mb(struct r5l_log *log, · line 2005
Quoted 1 line
drivers/md/raid5-cache.c +1
-0+uint32_t payload_size;
checkpatch report u32 is prefered.
Quoted 76 lines
drivers/md/raid5-cache.c +24
-7sector_t payload_len; payload = (void *)mb + mb_offset; payload_flush = (void *)mb + mb_offset;if (le16_to_cpu (payload->header.type) ==R5LOG_PAYLOAD_DATA ) { +if (mb_offset +sizeof (struct r5l_payload_data_parity ) + >le32_to_cpu (mb->meta_size)) + goto mismatch; + payload_size =le32_to_cpu (payload->size) >> (PAGE_SHIFT -9 ); payload_len =sizeof (struct r5l_payload_data_parity ) + - (sector_t )sizeof (__le32) * - (le32_to_cpu (payload->size) >> (PAGE_SHIFT -9 )); -if (mb_offset + payload_len >le32_to_cpu (mb->meta_size)) + (sector_t )sizeof (__le32) * payload_size; +if (mb_offset + payload_len >le32_to_cpu (mb->meta_size) || + payload_size <1 ) goto mismatch;if (r5l_recovery_verify_data_checksum ( log, ctx, page, log_offset, payload->checksum[0 ]) <0 ) goto mismatch; }else if (le16_to_cpu (payload->header.type) ==R5LOG_PAYLOAD_PARITY ) { +if (mb_offset +sizeof (struct r5l_payload_data_parity ) + >le32_to_cpu (mb->meta_size)) + goto mismatch; + payload_size =le32_to_cpu (payload->size) >> (PAGE_SHIFT -9 ); payload_len =sizeof (struct r5l_payload_data_parity ) + - (sector_t )sizeof (__le32) * - (le32_to_cpu (payload->size) >> (PAGE_SHIFT -9 )); -if (mb_offset + payload_len >le32_to_cpu (mb->meta_size)) + (sector_t )sizeof (__le32) * payload_size; +if (mb_offset + payload_len >le32_to_cpu (mb->meta_size) || + payload_size < conf->max_degraded) goto mismatch; if (r5l_recovery_verify_data_checksum( log, ctx, page, log_offset,@@ -2035,6 +2044,9 @@ r5l_recovery_verify_data_checksum_for_mb(struct r5l_log *log, payload->checksum[1]) < 0) goto mismatch; }else if (le16_to_cpu (payload->header.type) ==R5LOG_PAYLOAD_FLUSH ) { +if (mb_offset +sizeof (struct r5l_payload_flush ) + >le32_to_cpu (mb->meta_size)) + goto mismatch; payload_len =sizeof (struct r5l_payload_flush ) + (sector_t )le32_to_cpu (payload_flush->size);if (mb_offset + payload_len >le32_to_cpu (mb->meta_size))@@ -2096,7 +2108,7 @@ r5c_recovery_analyze_meta_block(struct r5l_log *log, mb_offset =sizeof (struct r5l_meta_block ); log_offset =r5l_ring_add (log, ctx->pos,BLOCK_SECTORS ); -while (mb_offset <le32_to_cpu (mb->meta_size)) { +while (mb_offset +sizeof (struct r5l_payload_header ) <le32_to_cpu (mb->meta_size)) {sector_t payload_len;int dd;@@ -2106,6 +2118,9 @@ r5c_recovery_analyze_meta_block(struct r5l_log *log,if (le16_to_cpu (payload->header.type) ==R5LOG_PAYLOAD_FLUSH ) {int i, count; +if (mb_offset +sizeof (struct r5l_payload_flush ) > +le32_to_cpu (mb->meta_size)) +return -EINVAL ; payload_len =sizeof (struct r5l_payload_flush ) + (sector_t )le32_to_cpu (payload_flush->size); if (mb_offset + payload_len >@@ -2130,6 +2145,9 @@ r5c_recovery_analyze_meta_block(struct r5l_log *log, }/* DATA or PARITY payload */ +if (mb_offset +sizeof (struct r5l_payload_data_parity ) > +le32_to_cpu (mb->meta_size)) +return -EINVAL ; payload_len =sizeof (struct r5l_payload_data_parity ) + (sector_t )sizeof (__le32) * (le32_to_cpu (payload->size) >> (PAGE_SHIFT -9 ));
Signature · 2 lines
Thanks, Kuai
Quoted 46 lines · yu kuai
Hi,
Quoted 43 lines · Martin Wilck
Commit
b0cc3ae97e89("md/raid5: validate payload size before accessing
journal metadata") introduced a bounds check to verify that a given payload
fits into the current metadata block. In order to calculate the payload
size, it has to access the memberspayload->header.typeandpayload->size,
the offset of which may already be past the end of the metadata block,
because the loop condition only checks that the first byte of the payload
is inside, and the payload entries have variable sizes. Later on, payload->checksum[0] and payload->checksum[1] are accessed without verifying thatpayload->sizeis large enough.Add another check to make sure that
payload->sizecan be safely accessed, and make sure that the checksum fields are accessible.This issue has been found by AI (gemma4, Gemini) during a backport review.
Cc Junrui Luo <[email protected]>Fixes b0cc3ae97e89 ("md/raid5: validate payload size beforeaccessing journal metadata")
Signed-off-by Martin Wilck <[email protected]>---
Changes v2 -> v3: Fixed another issue reported by Sashiko* Modify loop condition to make sure the
payload->header.typecan be accessedChanges v1 -> v2: Fixed issues reported by Sashiko
* fix use of wrong variable name
payload_leninstead ofmb_offset
* add checks to ensurepayload->sizeis large enough to hold the checksum values---
drivers/md/raid5-cache.c| 34 ++++++++++++++++++++++++++--------1 file changed, 26 insertions(+), 8 deletions(-)
Quoted 16 lines
Quoted 11 lines
drivers/md/raid5-cache.c +0
-0@@ -2001,27 +2001,36 @@ r5l_recovery_verify_data_checksum_for_mb(struct r5l_log *log, if (!page) return -ENOMEM;drivers/md/raid5-cache.c +1
-1-while (mb_offset <le32_to_cpu (mb->meta_size)) { + while (mb_offset + sizeof(struct r5l_payload_header) <Quoted 1 line
le32_to_cpu(mb->meta_size)) {Offset can't be
meta_size, however, is it possible that the remaining space is exactly one
payload size? If so I think <= should be used here.
Right. But practically speaking, as this must be either a struct r5l_payload_data_parity or a struct r5l_payload_flush, and both contain additional elements besides the header, if "==" applies here, we will run into one of the following error conditions later, because the size element won't be accessible.
Anyway, for the sake of correctness, I'll change it to "<=".
Quoted 4 lines
drivers/md/raid5-cache.c r5l_recovery_verify_data_checksum_for_mb(struct r5l_log *log, · line 2005
Quoted 2 lines
drivers/md/raid5-cache.c +1
-0+uint32_t payload_size;checkpatch report u32 is prefered.
Strange, it did not for me. I'll change it.
Thanks,
Martin
Signature · 4 lines
Dr. Martin Wilck <[email protected]> SUSE Software Solutions Germany GmbH, Frankenstr. 146, 90461 Nürnberg, Germany Geschäftsführer: Jochen Jaser, Andrew McDonald (HRB 36809,AG Nürnberg)