Bluetooth: bnep: reject short frames before parsing
Subsystems (2)
PATCH v4 1/1 Bluetooth: bnep: reject short frames before parsing
A BNEP peer can send a short BNEP SDU. bnep_rx_frame() reads the packet type byte immediately and, for control packets, reads the control opcode and setup UUID-size byte before proving that those bytes are present. bnep_rx_control() also dereferences the control opcode without rejecting an empty control payload.
Use skb_pull_data() for the fixed fields in bnep_rx_frame() so a NULL return gates each dereference. Split the control handler so the frame path can pass an opcode that has already been pulled, and keep the byte-buffer wrapper for extension control payloads.
Validation reproduced this kernel report:
KASAN slab-out-of-bounds in bnep_rx_frame.isra.0+0x130c/0x1790
The buggy address belongs to the object at ffff88800c0f7908 which belongs to the cache kmalloc-8 of size 8 The buggy address is located 0 bytes to the right of allocated 1-byte region [ffff88800c0f7908, ffff88800c0f7909) Read of size 1 Call trace: dump_stack_lvl+0xb3/0x140 (?:?) print_address_description+0x57/0x3a0 (?:?) bnep_rx_frame+0x130c/0x1790 (net/bluetooth/bnep/core.c:306) print_report+0xb9/0x2b0 (?:?) __virt_addr_valid+0x1ba/0x3a0 (?:?) srso_alias_return_thunk+0x5/0xfbef5 (?:?) kasan_addr_to_slab+0x21/0x60 (?:?) kasan_report+0xe0/0x110 (?:?) process_one_work+0xfce/0x17e0 (kernel/workqueue.c:3200) worker_thread+0x65c/0xe40 (?:?) __kthread_parkme+0x184/0x230 (?:?) kthread+0x35e/0x470 (?:?) _raw_spin_unlock_irq+0x28/0x50 (?:?) ret_from_fork+0x586/0x870 (?:?) __switch_to+0x74f/0xdc0 (?:?) ret_from_fork_asm+0x1a/0x30 (?:?)
--- net/bluetooth/bnep/core.c | 51 ++++++++++++++++++++++++--------------- 1 file changed, 31 insertions(+), 20 deletions(-)
net/bluetooth/bnep/core.c +31 -20
@@ -206,14 +206,11 @@ static int bnep_ctrl_set_mcfilter(struct bnep_session *s, u8 *data, int len)return 0 ; } -static int bnep_rx_control (struct bnep_session *s,void *data,int len) +static int bnep_rx_control_cmd (struct bnep_session *s,u8 cmd,void *data, +int len) { -u8 cmd = *(u8 *)data;int err =0 ; - data++; - len--; -switch (cmd) {case BNEP_CMD_NOT_UNDERSTOOD :case BNEP_SETUP_CONN_RSP :
@@ -254,6 +251,14 @@ static int bnep_rx_control(struct bnep_session *s, void *data, int len)return err; } +static int bnep_rx_control (struct bnep_session *s,void *data,int len) +{ +if (len <1 ) +return -EILSEQ ; + +return bnep_rx_control_cmd (s, *(u8 *)data, data +1 , len -1 ); +} +static int bnep_rx_extension (struct bnep_session *s,struct sk_buff *skb) {struct bnep_ext_hdr *h;
@@ -299,19 +304,26 @@ static int bnep_rx_frame(struct bnep_session *s, struct sk_buff *skb) {struct net_device *dev = s->dev;struct sk_buff *nskb; +u8 *data;u8 type, ctrl_type; dev->stats.rx_bytes += skb->len; - type = *(u8 *) skb->data; -skb_pull (skb,1 ); - ctrl_type = *(u8 *)skb->data; + data =skb_pull_data (skb,sizeof (type)); +if (!data) + goto badframe; + type = *data;if ((type &BNEP_TYPE_MASK ) >=sizeof (__bnep_rx_hlen)) goto badframe;if ((type &BNEP_TYPE_MASK ) ==BNEP_CONTROL ) { -if (bnep_rx_control (s, skb->data, skb->len) <0 ) { + data =skb_pull_data (skb,sizeof (ctrl_type)); +if (!data) + goto badframe; + ctrl_type = *data; + +if (bnep_rx_control_cmd (s, ctrl_type, skb->data, skb->len) <0 ) { dev->stats.tx_errors++;kfree_skb (skb);return 0 ;
@@ -325,23 +337,22 @@ static int bnep_rx_frame(struct bnep_session *s, struct sk_buff *skb)/* Verify and pull ctrl message since it's already processed */ switch (ctrl_type) {case BNEP_SETUP_CONN_REQ : -/* Pull: ctrl type (1 b), len (1 b), data (len bytes) */ -if (!skb_pull (skb,2 + *(u8 *)(skb->data +1 ) *2 )) +/* Pull: len (1 b), data (len * 2 bytes) */ + data =skb_pull_data (skb,1 ); +if (!data) + goto badframe; +if (!skb_pull (skb, *data *2 )) goto badframe;break ;case BNEP_FILTER_MULTI_ADDR_SET : -case BNEP_FILTER_NET_TYPE_SET : { -u8 *hdr; - -/* Pull ctrl type (1 b) + len (2 b) */ - hdr =skb_pull_data (skb,3 ); -if (!hdr) +case BNEP_FILTER_NET_TYPE_SET : +/* Pull: len (2 b), data (len bytes) */ + data =skb_pull_data (skb,sizeof (u16)); +if (!data) goto badframe; -/* Pull data (len bytes); length is big-endian */ -if (!skb_pull (skb,get_unaligned_be16 (&hdr[1 ]))) +if (!skb_pull (skb,get_unaligned_be16 (data))) goto badframe;break ; - }default :kfree_skb (skb);return 0 ;
Signature · 1 line
2.43.0
This is automated email and please do not reply to this email!
Dear submitter,
Thank you for submitting the patches to the linux bluetooth mailing list. This is a CI test results with your patch series:
PW Link:https://patchwork.kernel.org/project/bluetooth/list/?series=1098453
---Test result---
Test Summary:
CheckPatch PASS 0.49 seconds GitLint PASS 0.19 seconds SubjectPrefix PASS 0.06 seconds BuildKernel PASS 19.75 seconds CheckAllWarning PASS 25.20 seconds CheckSparse PASS 22.00 seconds BuildKernel32 PASS 19.79 seconds TestRunnerSetup PASS 412.06 seconds TestRunner_bnep-tester PASS 14.59 seconds IncrementalBuild PASS 22.14 seconds
https://github.com/bluez/bluetooth-next/pull/228
---
Regards,
Linux Bluetooth
Hi Luiz,
Gentle ping on this patch. The current version follows the previous review feedback and the bluetooth test bot reported no failures.
Please let me know if you would prefer any further changes.
Thanks,
Zhang Cen
Quoted 31 lines · <[email protected]> 于2026年5月21日周四 16:07写道:
This is automated email and please do not reply to this email!
Dear submitter,
Thank you for submitting the patches to the linux bluetooth mailing list. This is a CI test results with your patch series:
PW Link:https://patchwork.kernel.org/project/bluetooth/list/?series=1098453---Test result---
Test Summary:
CheckPatch PASS 0.49 seconds GitLint PASS 0.19 seconds SubjectPrefix PASS 0.06 seconds BuildKernel PASS 19.75 seconds CheckAllWarning PASS 25.20 seconds CheckSparse PASS 22.00 seconds BuildKernel32 PASS 19.79 seconds TestRunnerSetup PASS 412.06 seconds TestRunner_bnep-tester PASS 14.59 seconds IncrementalBuild PASS 22.14 secondshttps://github.com/bluez/bluetooth-next/pull/228
---
Regards,
Linux Bluetooth
Hi Luiz,
Gentle ping on this patch. The current version follows the previous review feedback and the bluetooth test bot reported no failures.
Please let me know if you would prefer any further changes.
Thanks,
Zhang Cen
Quoted 147 lines · Zhang Cen <[email protected]> 于2026年5月21日周四 13:26写道:
A BNEP peer can send a short BNEP SDU.
bnep_rx_frame()reads the packet type byte immediately and, for control packets, reads the control opcode and setup UUID-size byte before proving that those bytes are present.bnep_rx_control()also dereferences the control opcode without rejecting an empty control payload.Use
skb_pull_data()for the fixed fields inbnep_rx_frame()so a NULL return gates each dereference. Split the control handler so the frame path can pass an opcode that has already been pulled, and keep the byte-buffer wrapper for extension control payloads.Validation reproduced this kernel report:
KASAN slab-out-of-bounds in bnep_rx_frame.isra.0+0x130c/0x1790The buggy address belongs to the object at ffff88800c0f7908 which belongs to the cache kmalloc-8 of size 8 The buggy address is located 0 bytes to the right of allocated 1-byte region [ffff88800c0f7908, ffff88800c0f7909) Read of size 1 Call trace: dump_stack_lvl+0xb3/0x140 (?:?) print_address_description+0x57/0x3a0 (?:?) bnep_rx_frame+0x130c/0x1790 (net/bluetooth/bnep/core.c:306) print_report+0xb9/0x2b0 (?:?) __virt_addr_valid+0x1ba/0x3a0 (?:?) srso_alias_return_thunk+0x5/0xfbef5 (?:?) kasan_addr_to_slab+0x21/0x60 (?:?) kasan_report+0xe0/0x110 (?:?) process_one_work+0xfce/0x17e0 (kernel/workqueue.c:3200) worker_thread+0x65c/0xe40 (?:?) __kthread_parkme+0x184/0x230 (?:?) kthread+0x35e/0x470 (?:?) _raw_spin_unlock_irq+0x28/0x50 (?:?) ret_from_fork+0x586/0x870 (?:?) __switch_to+0x74f/0xdc0 (?:?) ret_from_fork_asm+0x1a/0x30 (?:?)Fixes 1da177e4c3f4 ("Linux-2.6.12-rc2")Signed-off-by Zhang Cen <[email protected]>--- net/bluetooth/bnep/core.c | 51 ++++++++++++++++++++++++--------------- 1 file changed, 31 insertions(+), 20 deletions(-)net/bluetooth/bnep/core.c +31
-21@@ -206,14 +206,11 @@ static int bnep_ctrl_set_mcfilter(struct bnep_session *s, u8 *data, int len)return 0 ; } -static int bnep_rx_control (struct bnep_session *s,void *data,int len) +static int bnep_rx_control_cmd (struct bnep_session *s,u8 cmd,void *data, +int len) { -u8 cmd = *(u8 *)data;int err =0 ; - data++; - len--; -switch (cmd) {case BNEP_CMD_NOT_UNDERSTOOD :case BNEP_SETUP_CONN_RSP :@@ -254,6 +251,14 @@ static int bnep_rx_control(struct bnep_session *s, void *data, int len)return err; } +static int bnep_rx_control (struct bnep_session *s,void *data,int len) +{ +if (len <1 ) +return -EILSEQ ; + +return bnep_rx_control_cmd (s, *(u8 *)data, data +1 , len -1 ); +} +static int bnep_rx_extension (struct bnep_session *s,struct sk_buff *skb) {struct bnep_ext_hdr *h;@@ -299,19 +304,26 @@ static int bnep_rx_frame(struct bnep_session *s, struct sk_buff *skb) {struct net_device *dev = s->dev;struct sk_buff *nskb; +u8 *data;u8 type, ctrl_type; dev->stats.rx_bytes += skb->len; - type = *(u8 *) skb->data; -skb_pull (skb,1 ); - ctrl_type = *(u8 *)skb->data; + data =skb_pull_data (skb,sizeof (type)); +if (!data) + goto badframe; + type = *data;if ((type &BNEP_TYPE_MASK ) >=sizeof (__bnep_rx_hlen)) goto badframe;if ((type &BNEP_TYPE_MASK ) ==BNEP_CONTROL ) { -if (bnep_rx_control (s, skb->data, skb->len) <0 ) { + data =skb_pull_data (skb,sizeof (ctrl_type)); +if (!data) + goto badframe; + ctrl_type = *data; + +if (bnep_rx_control_cmd (s, ctrl_type, skb->data, skb->len) <0 ) { dev->stats.tx_errors++;kfree_skb (skb);return 0 ;@@ -325,23 +337,22 @@ static int bnep_rx_frame(struct bnep_session *s, struct sk_buff *skb)/* Verify and pull ctrl message since it's already processed */ switch (ctrl_type) {case BNEP_SETUP_CONN_REQ : -/* Pull: ctrl type (1 b), len (1 b), data (len bytes) */ -if (!skb_pull (skb,2 + *(u8 *)(skb->data +1 ) *2 )) +/* Pull: len (1 b), data (len * 2 bytes) */ + data =skb_pull_data (skb,1 ); +if (!data) + goto badframe; +if (!skb_pull (skb, *data *2 )) goto badframe;break ;case BNEP_FILTER_MULTI_ADDR_SET : -case BNEP_FILTER_NET_TYPE_SET : { -u8 *hdr; - -/* Pull ctrl type (1 b) + len (2 b) */ - hdr =skb_pull_data (skb,3 ); -if (!hdr) +case BNEP_FILTER_NET_TYPE_SET : +/* Pull: len (2 b), data (len bytes) */ + data =skb_pull_data (skb,sizeof (u16)); +if (!data) goto badframe; -/* Pull data (len bytes); length is big-endian */ -if (!skb_pull (skb,get_unaligned_be16 (&hdr[1 ]))) +if (!skb_pull (skb,get_unaligned_be16 (data))) goto badframe;break ; - }default :kfree_skb (skb);return 0 ; --2.43.0
Luiz Augusto von Dentz
maintainerHi Zhang,
Quoted 111 lines · Zhang Cen
A BNEP peer can send a short BNEP SDU.
bnep_rx_frame()reads the packet type byte immediately and, for control packets, reads the control opcode and setup UUID-size byte before proving that those bytes are present.bnep_rx_control()also dereferences the control opcode without rejecting an empty control payload.Use
skb_pull_data()for the fixed fields inbnep_rx_frame()so a NULL return gates each dereference. Split the control handler so the frame path can pass an opcode that has already been pulled, and keep the byte-buffer wrapper for extension control payloads.Validation reproduced this kernel report:
KASAN slab-out-of-bounds in bnep_rx_frame.isra.0+0x130c/0x1790The buggy address belongs to the object at ffff88800c0f7908 which belongs to the cache kmalloc-8 of size 8 The buggy address is located 0 bytes to the right of allocated 1-byte region [ffff88800c0f7908, ffff88800c0f7909) Read of size 1 Call trace: dump_stack_lvl+0xb3/0x140 (?:?) print_address_description+0x57/0x3a0 (?:?) bnep_rx_frame+0x130c/0x1790 (net/bluetooth/bnep/core.c:306) print_report+0xb9/0x2b0 (?:?) __virt_addr_valid+0x1ba/0x3a0 (?:?) srso_alias_return_thunk+0x5/0xfbef5 (?:?) kasan_addr_to_slab+0x21/0x60 (?:?) kasan_report+0xe0/0x110 (?:?) process_one_work+0xfce/0x17e0 (kernel/workqueue.c:3200) worker_thread+0x65c/0xe40 (?:?) __kthread_parkme+0x184/0x230 (?:?) kthread+0x35e/0x470 (?:?) _raw_spin_unlock_irq+0x28/0x50 (?:?) ret_from_fork+0x586/0x870 (?:?) __switch_to+0x74f/0xdc0 (?:?) ret_from_fork_asm+0x1a/0x30 (?:?)Fixes 1da177e4c3f4 ("Linux-2.6.12-rc2")Signed-off-by Zhang Cen <[email protected]>--- net/bluetooth/bnep/core.c | 51 ++++++++++++++++++++++++--------------- 1 file changed, 31 insertions(+), 20 deletions(-)net/bluetooth/bnep/core.c +21
-9@@ -206,14 +206,11 @@ static int bnep_ctrl_set_mcfilter(struct bnep_session *s, u8 *data, int len)return 0 ; } -static int bnep_rx_control (struct bnep_session *s,void *data,int len) +static int bnep_rx_control_cmd (struct bnep_session *s,u8 cmd,void *data, +int len) { -u8 cmd = *(u8 *)data;int err =0 ; - data++; - len--; -switch (cmd) {case BNEP_CMD_NOT_UNDERSTOOD :case BNEP_SETUP_CONN_RSP :@@ -254,6 +251,14 @@ static int bnep_rx_control(struct bnep_session *s, void *data, int len)return err; } +static int bnep_rx_control (struct bnep_session *s,void *data,int len) +{ +if (len <1 ) +return -EILSEQ ; + +return bnep_rx_control_cmd (s, *(u8 *)data, data +1 , len -1 ); +} +static int bnep_rx_extension (struct bnep_session *s,struct sk_buff *skb) {struct bnep_ext_hdr *h;@@ -299,19 +304,26 @@ static int bnep_rx_frame(struct bnep_session *s, struct sk_buff *skb) {struct net_device *dev = s->dev;struct sk_buff *nskb; +u8 *data;u8 type, ctrl_type; dev->stats.rx_bytes += skb->len; - type = *(u8 *) skb->data; -skb_pull (skb,1 ); - ctrl_type = *(u8 *)skb->data; + data =skb_pull_data (skb,sizeof (type)); +if (!data) + goto badframe; + type = *data;if ((type &BNEP_TYPE_MASK ) >=sizeof (__bnep_rx_hlen)) goto badframe;if ((type &BNEP_TYPE_MASK ) ==BNEP_CONTROL ) { -if (bnep_rx_control (s, skb->data, skb->len) <0 ) { + data =skb_pull_data (skb,sizeof (ctrl_type)); +if (!data) + goto badframe; + ctrl_type = *data; + +if (bnep_rx_control_cmd (s, ctrl_type, skb->data, skb->len) <0 ) { dev->stats.tx_errors++;kfree_skb (skb);return 0 ;
net/bluetooth/bnep/core.c static int bnep_rx_frame(struct bnep_session *s, struct sk_buff *skb) · line 344
Quoted 10 lines
net/bluetooth/bnep/core.c +5
-2@@ -325,23 +337,22 @@ static int bnep_rx_frame(struct bnep_session *s, struct sk_buff *skb)/* Verify and pull ctrl message since it's already processed */ switch (ctrl_type) {case BNEP_SETUP_CONN_REQ : -/* Pull: ctrl type (1 b), len (1 b), data (len bytes) */ -if (!skb_pull (skb,2 + *(u8 *)(skb->data +1 ) *2 )) +/* Pull: len (1 b), data (len * 2 bytes) */ + data =skb_pull_data (skb,1 ); +if (!data) + goto badframe; +if (!skb_pull (skb, *data *2 ))
This still looks incorrect to me, why should we do * 2? If look at BNEP_FILTER_NET_TYPE_SET it just use the header to pull using get_unaligned, I know this is an existing problem but except if the BNEP_SETUP_CONN_REQ header length works on a multiple of 2 this makes no sense to me.
Quoted 26 lines
net/bluetooth/bnep/core.c +5
-10goto badframe;break ;case BNEP_FILTER_MULTI_ADDR_SET : -case BNEP_FILTER_NET_TYPE_SET : { -u8 *hdr; - -/* Pull ctrl type (1 b) + len (2 b) */ - hdr =skb_pull_data (skb,3 ); -if (!hdr) +case BNEP_FILTER_NET_TYPE_SET : +/* Pull: len (2 b), data (len bytes) */ + data =skb_pull_data (skb,sizeof (u16)); +if (!data) goto badframe; -/* Pull data (len bytes); length is big-endian */ -if (!skb_pull (skb,get_unaligned_be16 (&hdr[1 ]))) +if (!skb_pull (skb,get_unaligned_be16 (data))) goto badframe;break ; - }default :kfree_skb (skb);return 0 ; --2.43.0
Signature · 1 line
Luiz Augusto von Dentz
Hi Luiz,
Thanks for taking another look.
Quoted 5 lines
This still looks incorrect to me, why should we do * 2? If look at
BNEP_FILTER_NET_TYPE_SETit just use the header to pull usingget_unaligned, I know this is an existing problem but except if theBNEP_SETUP_CONN_REQheader length works on a multiple of 2 this makes no sense to me.
The reason I used * 2 there is that I read that byte as the UUID size, not as a total payload length.
For BNEP_SETUP_CONN_REQ, the byte after the control type is the UUID Size field. The setup request then carries two UUIDs of that size: the destination service UUID and the source service UUID. So the intended skip was:
UUID Size field itself
destination service UUID, uuid_size bytes
source service UUID, uuid_size bytes
That is different from BNEP_FILTER_NET_TYPE_SET, where the header has a 16-bit length field and the value from get_unaligned_be16() is already the total filter-list payload length.
That said, I agree the patch makes this much less clear than it should. The comment says "data (len * 2 bytes)", which makes it look like a generic length field, and the raw * 2 is too opaque.
I can send a v5 that names the byte uuid_size and uses uuid_size + uuid_size with a comment explaining that the two parts are the destination and source service UUIDs.If I have misunderstood anything here, sorry for the confusion. I would be happy to help with any further changes
or testing.
Thanks,
Zhang Cen
Luiz Augusto von Dentz
maintainerHi Cen,
Quoted 14 lines · Cen Zhang
Hi Luiz,
Thanks for taking another look.
Quoted 5 lines
This still looks incorrect to me, why should we do * 2? If look at
BNEP_FILTER_NET_TYPE_SETit just use the header to pull usingget_unaligned, I know this is an existing problem but except if theBNEP_SETUP_CONN_REQheader length works on a multiple of 2 this makes no sense to me.The reason I used
* 2there is that I read that byte as the UUID size, not as a total payload length.
Quoted 8 lines
For
BNEP_SETUP_CONN_REQ, the byte after the control type is the UUID Size field. The setup request then carries two UUIDs of that size: the destination service UUID and the source service UUID. So the intended skip was:UUID Size field itself
destination service UUID,uuid_sizebytes
source service UUID,uuid_sizebytes
This is what you get if you don't know where to look at:
struct bnep_setup_conn_req {
uint8_t type;
uint8_t ctrl;
uint8_t uuid_size;
uint8_t service[0];
} __attribute__((packed));https://github.com/bluez/bluez/blob/master/lib/bluetooth/bnep.h#L79
It is in fact a tuple of src + dst service as parsed by btmon:
https://github.com/bluez/bluez/blob/master/monitor/bnep.c#L104
Quoted 8 lines
That is different from
BNEP_FILTER_NET_TYPE_SET, where the header has a 16-bit length field and the value fromget_unaligned_be16()is already the total filter-list payload length.That said, I agree the patch makes this much less clear than it should. The comment says "data (len * 2 bytes)", which makes it look like a generic length field, and the raw
* 2is too opaque.
Quoted 6 lines
I can send a v5 that names the byte
uuid_sizeand usesuuid_size + uuid_sizewith a comment explaining that the two parts are the destination and source service UUIDs.If I have misunderstood anything here, sorry for the confusion. I would be happy to help with any further changes
or testing.
Probably just reference the above, but the fact that you couldn't mention any of the BlueZ references really bothers me. Although the result was correct, the lack of references makes me believe I'm just arguing with an AI prompt which doesn't have BlueZ userspace loaded as context.
Signature · 1 line
Luiz Augusto von Dentz
Hi Luiz,
Quoted 9 lines · Luiz Augusto von Dentz
This is what you get if you don't know where to look at:
struct bnep_setup_conn_req { uint8_t type; uint8_t ctrl; uint8_t uuid_size; uint8_t service[0]; } __attribute__((packed));
Quoted 4 lines
https://github.com/bluez/bluez/blob/master/lib/bluetooth/bnep.h#L79
It is in fact a tuple of src + dst service as parsed by btmon:
https://github.com/bluez/bluez/blob/master/monitor/bnep.c#L104
Right, service[] is the dst + src service tuple, each uuid_size, so the skip after the control type is 1 + 2 * uuid_size. For v5, I'll name the byte uuid_size and make the setup-request layout explicit in the comment so it is not an opaque length.
Quoted 5 lines
Probably just reference the above, but the fact that you couldn't mention any of the BlueZ references really bothers me. Although the result was correct, the lack of references makes me believe I'm just arguing with an AI prompt which doesn't have BlueZ userspace loaded as context.
You are right to call this out, and I'm sorry for the noise.
The initial lead for this issue did come from AI assistance. I then reproduced the bug under KASAN and checked the kernel-side code, but I should have done a better job before replying.
I've also been reading the recent kernel discussions around AI-assisted reports and patches, and I understand the concern better now. A patch is not just a bug report saying what can go wrong; it also has to explain why the change is correct in the subsystem context, so reviewers are not left doing that work from scratch. I was thinking too much like a bug reporter there and not carefully enough as a patch author. I will not rely on LLM output as a substitute for reading and understanding the code.
I'll send v5 with the above and will be more careful with Bluetooth patches and follow-ups going forward.
Thanks for taking the time to point this out.
Zhang Cen