bpf: Fail verification for sign-extension of packet data/data_end/data_meta
Subsystems (6)
Yonghong Song
reviewerPATCH v3 1/2 bpf: Fail verification for sign-extension of packet data/data_end/data_meta
syzbot reported a kernel crash due to
commit 1f1e864b6555 ("bpf: Handle sign-extenstin ctx member accesses").
The reason is due to sign-extension of 32-bit load for
packet data/data_end/data_meta uapi field.
The original code looks like:
r2 = *(s32 *)(r1 + 76) /* load __sk_buff->data */
r3 = *(u32 *)(r1 + 80) /* load __sk_buff->data_end */
r0 = r2
r0 += 8
if r3 > r0 goto +1
...Note that __sk_buff->data load has 32-bit sign extension.
After verification and convert_ctx_accesses(), the final asm code looks like:
r2 = *(u64 *)(r1 +208)
r2 = (s32)r2
r3 = *(u64 *)(r1 +80)
r0 = r2
r0 += 8
if r3 > r0 goto pc+1
...Note that 'r2 = (s32)r2' may make the kernel __sk_buff->data address invalid which may cause runtime failure.
Currently, in C code, typically we have
void *data = (void *)(long)skb->data;
void *data_end = (void *)(long)skb->data_end;
...
and it will generate
r2 = *(u64 *)(r1 +208)
r3 = *(u64 *)(r1 +80)
r0 = r2
r0 += 8
if r3 > r0 goto pc+1If we allow sign-extension,
void *data = (void *)(long)(int)skb->data;
void *data_end = (void *)(long)skb->data_end;
...
the generated code looks like
r2 = *(u64 *)(r1 +208)
r2 <<= 32
r2 s>>= 32
r3 = *(u64 *)(r1 +80)
r0 = r2
r0 += 8
if r3 > r0 goto pc+1and this will cause verification failure since "r2 <<= 32" is not allowed as "r2" is a packet pointer.
To fix this issue for case
r2 = *(s32 *)(r1 + 76) /* load __sk_buff->data */
this patch added additional checking in is_valid_access() callback function for packet data/data_end/data_meta access. If those accesses are with sign-extenstion, the verification will fail.
[1] https://lore.kernel.org/bpf/[email protected]/
--- include/linux/bpf.h | 1 + kernel/bpf/verifier.c | 5 +++-- net/core/filter.c | 21 ++++++++++++++++----- 3 files changed, 20 insertions(+), 7 deletions(-)
Changelog from v2:
- Rebase due to conflict
include/linux/bpf.h +1 -0
@@ -920,6 +920,7 @@ static_assert(__BPF_REG_TYPE_MAX <= BPF_BASE_TYPE_LIMIT);*/ struct bpf_insn_access_aux { enum bpf_reg_type reg_type; + bool is_ldsx; union { int ctx_field_size; struct {
kernel/bpf/verifier.c +3 -2
@@ -5626,12 +5626,13 @@ static int check_packet_access(struct bpf_verifier_env *env, u32 regno, int off,/* check access to 'struct bpf_context' fields. Supports fixed offsets only */ static int check_ctx_access (struct bpf_verifier_env *env,int insn_idx,int off,int size,enum bpf_access_type t,enum bpf_reg_type *reg_type, -struct btf **btf,u32 *btf_id,bool *is_retval) +struct btf **btf,u32 *btf_id,bool *is_retval,bool is_ldsx) {struct bpf_insn_access_aux info = { .reg_type = *reg_type, .log = &env->log, .is_retval = false, + .is_ldsx = is_ldsx, }; if (env->ops->is_valid_access &&
@@ -6945,7 +6946,7 @@ static int check_mem_access(struct bpf_verifier_env *env, int insn_idx, u32 regnreturn err; err =check_ctx_access (env, insn_idx, off, size, t, ®_type, &btf, - &btf_id, &is_retval); + &btf_id, &is_retval, is_ldsx);if (err)verbose_linfo (env, insn_idx,"; " );if (!err && t ==BPF_READ && value_regno >=0 ) {
net/core/filter.c +16 -5
@@ -8572,13 +8572,16 @@ static bool bpf_skb_is_valid_access(int off, int size, enum bpf_access_type type if (off + size > offsetofend(struct __sk_buff, cb[4]))return false;break ; + case bpf_ctx_range(struct __sk_buff, data): + case bpf_ctx_range(struct __sk_buff, data_meta): + case bpf_ctx_range(struct __sk_buff, data_end): +if (info->is_ldsx || size != size_default) +return false; +break ; case bpf_ctx_range_till(struct __sk_buff, remote_ip6[0], remote_ip6[3]): case bpf_ctx_range_till(struct __sk_buff, local_ip6[0], local_ip6[3]): case bpf_ctx_range_till(struct __sk_buff, remote_ip4, remote_ip4): case bpf_ctx_range_till(struct __sk_buff, local_ip4, local_ip4): - case bpf_ctx_range(struct __sk_buff, data): - case bpf_ctx_range(struct __sk_buff, data_meta): - case bpf_ctx_range(struct __sk_buff, data_end):if (size != size_default)return false;break ;
@@ -9022,6 +9025,14 @@ static bool xdp_is_valid_access(int off, int size, } }return false; + }else { +switch (off) { +case offsetof(struct xdp_md , data_meta): +case offsetof(struct xdp_md , data): +case offsetof(struct xdp_md , data_end): +if (info->is_ldsx) +return false; + } }switch (off) {
@@ -9347,12 +9358,12 @@ static bool flow_dissector_is_valid_access(int off, int size,switch (off) { case bpf_ctx_range(struct __sk_buff, data): -if (size != size_default) +if (info->is_ldsx || size != size_default)return false; info->reg_type =PTR_TO_PACKET ;return true; case bpf_ctx_range(struct __sk_buff, data_end): -if (size != size_default) +if (info->is_ldsx || size != size_default)return false; info->reg_type =PTR_TO_PACKET_END ;return true;
Signature · 1 line
2.43.0
Yonghong Song
reviewerPATCH v3 2/2 selftests/bpf: Add tests for ldsx of pkt data/data_end/data_meta accesses
The following tests are added to verifier_ldsx.c:
- sign extension of data/data_end/data_meta for tcx programs. The actual checking is in bpf_skb_is_valid_access() which is called by sk_filter, cg_skb, lwt, tc(tcx) and sk_skb.
- sign extension of data/data_end/data_meta for xdp programs.
- sign extension of data/data_end for flow_dissector programs.
All newly-added tests have verification failure with message "invalid bpf_context access". Without previous patch, all these tests succeeded verification.
--- .../selftests/bpf/progs/verifier_ldsx.c | 112 ++++++++++++++++++ 1 file changed, 112 insertions(+)
Changelog from v1:
- Simplify test case and still trigger the verification failure.
tools/testing/selftests/bpf/progs/verifier_ldsx.c +112 -0
@@ -144,6 +144,118 @@ __naked void ldsx_s32_range(void) : __clobber_all); } +SEC("xdp") +__description("LDSX, xdp s32 xdp_md->data") +__failure __msg("invalid bpf_context access") +__naked void ldsx_ctx_1 (void ) +{ + asmvolatile ( +"r2 = *(s32 *)(r1 + %[xdp_md_data]);" +"r0 = 0;" +"exit;" + : + : __imm_const(xdp_md_data, offsetof(struct xdp_md, data)) + : __clobber_all); +} + +SEC("xdp") +__description("LDSX, xdp s32 xdp_md->data_end") +__failure __msg("invalid bpf_context access") +__naked void ldsx_ctx_2(void) +{ + asmvolatile ( +"r2 = *(s32 *)(r1 + %[xdp_md_data_end]);" +"r0 = 0;" +"exit;" + : + : __imm_const(xdp_md_data_end, offsetof(struct xdp_md, data_end)) + : __clobber_all); +} + +SEC("xdp") +__description("LDSX, xdp s32 xdp_md->data_meta") +__failure __msg("invalid bpf_context access") +__naked void ldsx_ctx_3(void) +{ + asmvolatile ( +"r2 = *(s32 *)(r1 + %[xdp_md_data_meta]);" +"r0 = 0;" +"exit;" + : + : __imm_const(xdp_md_data_meta, offsetof(struct xdp_md, data_meta)) + : __clobber_all); +} + +SEC("tcx/ingress") +__description("LDSX, tcx s32 __sk_buff->data") +__failure __msg("invalid bpf_context access") +__naked void ldsx_ctx_4(void) +{ + asmvolatile ( +"r2 = *(s32 *)(r1 + %[sk_buff_data]);" +"r0 = 0;" +"exit;" + : + : __imm_const(sk_buff_data, offsetof(struct __sk_buff, data)) + : __clobber_all); +} + +SEC("tcx/ingress") +__description("LDSX, tcx s32 __sk_buff->data_end") +__failure __msg("invalid bpf_context access") +__naked void ldsx_ctx_5(void) +{ + asmvolatile ( +"r2 = *(s32 *)(r1 + %[sk_buff_data_end]);" +"r0 = 0;" +"exit;" + : + : __imm_const(sk_buff_data_end, offsetof(struct __sk_buff, data_end)) + : __clobber_all); +} + +SEC("tcx/ingress") +__description("LDSX, tcx s32 __sk_buff->data_meta") +__failure __msg("invalid bpf_context access") +__naked void ldsx_ctx_6(void) +{ + asmvolatile ( +"r2 = *(s32 *)(r1 + %[sk_buff_data_meta]);" +"r0 = 0;" +"exit;" + : + : __imm_const(sk_buff_data_meta, offsetof(struct __sk_buff, data_meta)) + : __clobber_all); +} + +SEC("flow_dissector") +__description("LDSX, flow_dissector s32 __sk_buff->data") +__failure __msg("invalid bpf_context access") +__naked void ldsx_ctx_7(void) +{ + asmvolatile ( +"r2 = *(s32 *)(r1 + %[sk_buff_data]);" +"r0 = 0;" +"exit;" + : + : __imm_const(sk_buff_data, offsetof(struct __sk_buff, data)) + : __clobber_all); +} + +SEC("flow_dissector") +__description("LDSX, flow_dissector s32 __sk_buff->data_end") +__failure __msg("invalid bpf_context access") +__naked void ldsx_ctx_8(void) +{ + asmvolatile ( +"r2 = *(s32 *)(r1 + %[sk_buff_data_end]);" +"r0 = 0;" +"exit;" + : + : __imm_const(sk_buff_data_end, offsetof(struct __sk_buff, data_end)) + : __clobber_all); +} +#else SEC ("socket" )
Signature · 1 line
2.43.0
Re: [PATCH bpf-next v3 2/2] selftests/bpf: Add tests for ldsx of pkt data/data_end/data_meta accesses
CI has tested the following submission:
Dear patch submitter,
CI has tested the following submission:
Status: SUCCESS Name: [bpf-next,v3,2/2] selftests/bpf: Add tests for ldsx of pkt data/data_end/data_meta accesses Patchwork: https://patchwork.kernel.org/project/netdevbpf/list/?series=873296&state=* Matrix: https://github.com/kernel-patches/bpf/actions/runs/10062073488
No further action is necessary on your part.
Please note: this email is coming from an unmonitored mailbox. If you have questions or feedback, please reach out to the Meta Kernel CI team at [email protected].
This series was applied to bpf/bpf-next.git (master) by Alexei Starovoitov <[email protected]>:
Hello:
This series was applied to bpf/bpf-next.git (master) by Alexei Starovoitov <[email protected]>:
Quoted 12 lines · On Tue, 23 Jul 2024 08:34:39 -0700 you wrote:
syzbot reported a kernel crash due to
commit1f1e864b6555("bpf: Handle sign-extenstin ctx member accesses").
The reason is due to sign-extension of 32-bit load for
packet data/data_end/data_meta uapi field.The original code looks like:
r2 = *(s32 *)(r1 + 76) /* load __sk_buff->data */ r3 = *(u32 *)(r1 + 80) /* load __sk_buff->data_end */ r0 = r2 r0 += 8 if r3 > r0 goto +1 ...
Quoted 3 lines
Note that
__sk_buff->dataload has 32-bit sign extension.[...]
Here is the summary with links:
- [bpf-next,v3,1/2] bpf: Fail verification for sign-extension of packet data/data_end/data_meta
https://git.kernel.org/bpf/bpf-next/c/8924c0a1d51b - [bpf-next,v3,2/2] selftests/bpf: Add tests for ldsx of pkt data/data_end/data_meta accesses https://git.kernel.org/bpf/bpf-next/c/eb1b55c4875a
You are awesome, thank you!
Signature · 2 lines
Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/patchwork/pwbot.html