Message ID | 20200904194900.3031377-1-yhs@fb.com |
---|---|
State | Changes Requested |
Delegated to: | BPF Maintainers |
Headers | show |
Series | bpf: permit map_ptr arithmetic with opcode add and offset 0 | expand |
On Fri, Sep 4, 2020 at 12:49 PM Yonghong Song <yhs@fb.com> wrote: > > Commit 41c48f3a98231 ("bpf: Support access > to bpf map fields") added support to access map fields > with CORE support. For example, > > struct bpf_map { > __u32 max_entries; > } __attribute__((preserve_access_index)); > > struct bpf_array { > struct bpf_map map; > __u32 elem_size; > } __attribute__((preserve_access_index)); > > struct { > __uint(type, BPF_MAP_TYPE_ARRAY); > __uint(max_entries, 4); > __type(key, __u32); > __type(value, __u32); > } m_array SEC(".maps"); > > SEC("cgroup_skb/egress") > int cg_skb(void *ctx) > { > struct bpf_array *array = (struct bpf_array *)&m_array; > > /* .. array->map.max_entries .. */ > } > > In kernel, bpf_htab has similar structure, > > struct bpf_htab { > struct bpf_map map; > ... > } > > In the above cg_skb(), to access array->map.max_entries, with CORE, the clang will > generate two builtin's. > base = &m_array; > /* access array.map */ > map_addr = __builtin_preserve_struct_access_info(base, 0, 0); > /* access array.map.max_entries */ > max_entries_addr = __builtin_preserve_struct_access_info(map_addr, 0, 0); > max_entries = *max_entries_addr; > > In the current llvm, if two builtin's are in the same function or > in the same function after inlining, the compiler is smart enough to chain > them together and generates like below: > base = &m_array; > max_entries = *(base + reloc_offset); /* reloc_offset = 0 in this case */ > and we are fine. > > But if we force no inlining for one of functions in test_map_ptr() selftest, e.g., > check_default(), the above two __builtin_preserve_* will be in two different > functions. In this case, we will have code like: > func check_hash(): > reloc_offset_map = 0; > base = &m_array; > map_base = base + reloc_offset_map; > check_default(map_base, ...) > func check_default(map_base, ...): > max_entries = *(map_base + reloc_offset_max_entries); > > In kernel, map_ptr (CONST_PTR_TO_MAP) does not allow any arithmetic. > The above "map_base = base + reloc_offset_map" will trigger a verifier failure. > ; VERIFY(check_default(&hash->map, map)); > 0: (18) r7 = 0xffffb4fe8018a004 > 2: (b4) w1 = 110 > 3: (63) *(u32 *)(r7 +0) = r1 > R1_w=invP110 R7_w=map_value(id=0,off=4,ks=4,vs=8,imm=0) R10=fp0 > ; VERIFY_TYPE(BPF_MAP_TYPE_HASH, check_hash); > 4: (18) r1 = 0xffffb4fe8018a000 > 6: (b4) w2 = 1 > 7: (63) *(u32 *)(r1 +0) = r2 > R1_w=map_value(id=0,off=0,ks=4,vs=8,imm=0) R2_w=invP1 R7_w=map_value(id=0,off=4,ks=4,vs=8,imm=0) R10=fp0 > 8: (b7) r2 = 0 > 9: (18) r8 = 0xffff90bcb500c000 > 11: (18) r1 = 0xffff90bcb500c000 > 13: (0f) r1 += r2 > R1 pointer arithmetic on map_ptr prohibited > > To fix the issue, let us permit map_ptr + 0 arithmetic which will > result in exactly the same map_ptr. > > Signed-off-by: Yonghong Song <yhs@fb.com> > --- > kernel/bpf/verifier.c | 3 +++ > 1 file changed, 3 insertions(+) > > diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c > index b4e9c56b8b32..92aa985e99df 100644 > --- a/kernel/bpf/verifier.c > +++ b/kernel/bpf/verifier.c > @@ -5317,6 +5317,9 @@ static int adjust_ptr_min_max_vals(struct bpf_verifier_env *env, > dst, reg_type_str[ptr_reg->type]); > return -EACCES; > case CONST_PTR_TO_MAP: > + if (known && smin_val == 0 && opcode == BPF_ADD) does smin_val imply that var_off is strictly zero? if that's the case, can you please leave a comment stating this clearly, it's hard to tell if that's enough of a check. > + break; > + /* fall-through */ > case PTR_TO_PACKET_END: > case PTR_TO_SOCKET: > case PTR_TO_SOCKET_OR_NULL: > -- > 2.24.1 >
On 9/4/20 1:30 PM, Andrii Nakryiko wrote: > On Fri, Sep 4, 2020 at 12:49 PM Yonghong Song <yhs@fb.com> wrote: >> >> Commit 41c48f3a98231 ("bpf: Support access >> to bpf map fields") added support to access map fields >> with CORE support. For example, >> >> struct bpf_map { >> __u32 max_entries; >> } __attribute__((preserve_access_index)); >> >> struct bpf_array { >> struct bpf_map map; >> __u32 elem_size; >> } __attribute__((preserve_access_index)); >> >> struct { >> __uint(type, BPF_MAP_TYPE_ARRAY); >> __uint(max_entries, 4); >> __type(key, __u32); >> __type(value, __u32); >> } m_array SEC(".maps"); >> >> SEC("cgroup_skb/egress") >> int cg_skb(void *ctx) >> { >> struct bpf_array *array = (struct bpf_array *)&m_array; >> >> /* .. array->map.max_entries .. */ >> } >> >> In kernel, bpf_htab has similar structure, >> >> struct bpf_htab { >> struct bpf_map map; >> ... >> } >> >> In the above cg_skb(), to access array->map.max_entries, with CORE, the clang will >> generate two builtin's. >> base = &m_array; >> /* access array.map */ >> map_addr = __builtin_preserve_struct_access_info(base, 0, 0); >> /* access array.map.max_entries */ >> max_entries_addr = __builtin_preserve_struct_access_info(map_addr, 0, 0); >> max_entries = *max_entries_addr; >> >> In the current llvm, if two builtin's are in the same function or >> in the same function after inlining, the compiler is smart enough to chain >> them together and generates like below: >> base = &m_array; >> max_entries = *(base + reloc_offset); /* reloc_offset = 0 in this case */ >> and we are fine. >> >> But if we force no inlining for one of functions in test_map_ptr() selftest, e.g., >> check_default(), the above two __builtin_preserve_* will be in two different >> functions. In this case, we will have code like: >> func check_hash(): >> reloc_offset_map = 0; >> base = &m_array; >> map_base = base + reloc_offset_map; >> check_default(map_base, ...) >> func check_default(map_base, ...): >> max_entries = *(map_base + reloc_offset_max_entries); >> >> In kernel, map_ptr (CONST_PTR_TO_MAP) does not allow any arithmetic. >> The above "map_base = base + reloc_offset_map" will trigger a verifier failure. >> ; VERIFY(check_default(&hash->map, map)); >> 0: (18) r7 = 0xffffb4fe8018a004 >> 2: (b4) w1 = 110 >> 3: (63) *(u32 *)(r7 +0) = r1 >> R1_w=invP110 R7_w=map_value(id=0,off=4,ks=4,vs=8,imm=0) R10=fp0 >> ; VERIFY_TYPE(BPF_MAP_TYPE_HASH, check_hash); >> 4: (18) r1 = 0xffffb4fe8018a000 >> 6: (b4) w2 = 1 >> 7: (63) *(u32 *)(r1 +0) = r2 >> R1_w=map_value(id=0,off=0,ks=4,vs=8,imm=0) R2_w=invP1 R7_w=map_value(id=0,off=4,ks=4,vs=8,imm=0) R10=fp0 >> 8: (b7) r2 = 0 >> 9: (18) r8 = 0xffff90bcb500c000 >> 11: (18) r1 = 0xffff90bcb500c000 >> 13: (0f) r1 += r2 >> R1 pointer arithmetic on map_ptr prohibited >> >> To fix the issue, let us permit map_ptr + 0 arithmetic which will >> result in exactly the same map_ptr. >> >> Signed-off-by: Yonghong Song <yhs@fb.com> >> --- >> kernel/bpf/verifier.c | 3 +++ >> 1 file changed, 3 insertions(+) >> >> diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c >> index b4e9c56b8b32..92aa985e99df 100644 >> --- a/kernel/bpf/verifier.c >> +++ b/kernel/bpf/verifier.c >> @@ -5317,6 +5317,9 @@ static int adjust_ptr_min_max_vals(struct bpf_verifier_env *env, >> dst, reg_type_str[ptr_reg->type]); >> return -EACCES; >> case CONST_PTR_TO_MAP: >> + if (known && smin_val == 0 && opcode == BPF_ADD) > > does smin_val imply that var_off is strictly zero? if that's the case, > can you please leave a comment stating this clearly, it's hard to tell > if that's enough of a check. It should be, if register state is maintained properly, the following function (or its functionality) should have been called. static void __update_reg64_bounds(struct bpf_reg_state *reg) { /* min signed is max(sign bit) | min(other bits) */ reg->smin_value = max_t(s64, reg->smin_value, reg->var_off.value | (reg->var_off.mask & S64_MIN)); /* max signed is min(sign bit) | max(other bits) */ reg->smax_value = min_t(s64, reg->smax_value, reg->var_off.value | (reg->var_off.mask & S64_MAX)); reg->umin_value = max(reg->umin_value, reg->var_off.value); reg->umax_value = min(reg->umax_value, reg->var_off.value | reg->var_off.mask); } for scalar constant, reg->var_off.mask should be 0. so we will have reg->smin_value = reg->smax_value = (s64)reg->var_off.value. The smin_val is also used below, e.g., BPF_ADD, for a known value. That is why I am using smin_val here. Will add a comment and submit v2. > >> + break; >> + /* fall-through */ >> case PTR_TO_PACKET_END: >> case PTR_TO_SOCKET: >> case PTR_TO_SOCKET_OR_NULL: >> -- >> 2.24.1 >>
On Fri, Sep 4, 2020 at 4:20 PM Yonghong Song <yhs@fb.com> wrote: > > > > On 9/4/20 1:30 PM, Andrii Nakryiko wrote: > > On Fri, Sep 4, 2020 at 12:49 PM Yonghong Song <yhs@fb.com> wrote: > >> > >> Commit 41c48f3a98231 ("bpf: Support access > >> to bpf map fields") added support to access map fields > >> with CORE support. For example, > >> > >> struct bpf_map { > >> __u32 max_entries; > >> } __attribute__((preserve_access_index)); > >> > >> struct bpf_array { > >> struct bpf_map map; > >> __u32 elem_size; > >> } __attribute__((preserve_access_index)); > >> > >> struct { > >> __uint(type, BPF_MAP_TYPE_ARRAY); > >> __uint(max_entries, 4); > >> __type(key, __u32); > >> __type(value, __u32); > >> } m_array SEC(".maps"); > >> > >> SEC("cgroup_skb/egress") > >> int cg_skb(void *ctx) > >> { > >> struct bpf_array *array = (struct bpf_array *)&m_array; > >> > >> /* .. array->map.max_entries .. */ > >> } > >> > >> In kernel, bpf_htab has similar structure, > >> > >> struct bpf_htab { > >> struct bpf_map map; > >> ... > >> } > >> > >> In the above cg_skb(), to access array->map.max_entries, with CORE, the clang will > >> generate two builtin's. > >> base = &m_array; > >> /* access array.map */ > >> map_addr = __builtin_preserve_struct_access_info(base, 0, 0); > >> /* access array.map.max_entries */ > >> max_entries_addr = __builtin_preserve_struct_access_info(map_addr, 0, 0); > >> max_entries = *max_entries_addr; > >> > >> In the current llvm, if two builtin's are in the same function or > >> in the same function after inlining, the compiler is smart enough to chain > >> them together and generates like below: > >> base = &m_array; > >> max_entries = *(base + reloc_offset); /* reloc_offset = 0 in this case */ > >> and we are fine. > >> > >> But if we force no inlining for one of functions in test_map_ptr() selftest, e.g., > >> check_default(), the above two __builtin_preserve_* will be in two different > >> functions. In this case, we will have code like: > >> func check_hash(): > >> reloc_offset_map = 0; > >> base = &m_array; > >> map_base = base + reloc_offset_map; > >> check_default(map_base, ...) > >> func check_default(map_base, ...): > >> max_entries = *(map_base + reloc_offset_max_entries); > >> > >> In kernel, map_ptr (CONST_PTR_TO_MAP) does not allow any arithmetic. > >> The above "map_base = base + reloc_offset_map" will trigger a verifier failure. > >> ; VERIFY(check_default(&hash->map, map)); > >> 0: (18) r7 = 0xffffb4fe8018a004 > >> 2: (b4) w1 = 110 > >> 3: (63) *(u32 *)(r7 +0) = r1 > >> R1_w=invP110 R7_w=map_value(id=0,off=4,ks=4,vs=8,imm=0) R10=fp0 > >> ; VERIFY_TYPE(BPF_MAP_TYPE_HASH, check_hash); > >> 4: (18) r1 = 0xffffb4fe8018a000 > >> 6: (b4) w2 = 1 > >> 7: (63) *(u32 *)(r1 +0) = r2 > >> R1_w=map_value(id=0,off=0,ks=4,vs=8,imm=0) R2_w=invP1 R7_w=map_value(id=0,off=4,ks=4,vs=8,imm=0) R10=fp0 > >> 8: (b7) r2 = 0 > >> 9: (18) r8 = 0xffff90bcb500c000 > >> 11: (18) r1 = 0xffff90bcb500c000 > >> 13: (0f) r1 += r2 > >> R1 pointer arithmetic on map_ptr prohibited > >> > >> To fix the issue, let us permit map_ptr + 0 arithmetic which will > >> result in exactly the same map_ptr. > >> > >> Signed-off-by: Yonghong Song <yhs@fb.com> > >> --- > >> kernel/bpf/verifier.c | 3 +++ > >> 1 file changed, 3 insertions(+) > >> > >> diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c > >> index b4e9c56b8b32..92aa985e99df 100644 > >> --- a/kernel/bpf/verifier.c > >> +++ b/kernel/bpf/verifier.c > >> @@ -5317,6 +5317,9 @@ static int adjust_ptr_min_max_vals(struct bpf_verifier_env *env, > >> dst, reg_type_str[ptr_reg->type]); > >> return -EACCES; > >> case CONST_PTR_TO_MAP: > >> + if (known && smin_val == 0 && opcode == BPF_ADD) > > > > does smin_val imply that var_off is strictly zero? if that's the case, > > can you please leave a comment stating this clearly, it's hard to tell > > if that's enough of a check. > > It should be, if register state is maintained properly, the following > function (or its functionality) should have been called. > > static void __update_reg64_bounds(struct bpf_reg_state *reg) > { > /* min signed is max(sign bit) | min(other bits) */ > reg->smin_value = max_t(s64, reg->smin_value, > reg->var_off.value | (reg->var_off.mask > & S64_MIN)); > /* max signed is min(sign bit) | max(other bits) */ > reg->smax_value = min_t(s64, reg->smax_value, > reg->var_off.value | (reg->var_off.mask > & S64_MAX)); > reg->umin_value = max(reg->umin_value, reg->var_off.value); > reg->umax_value = min(reg->umax_value, > reg->var_off.value | reg->var_off.mask); > } > > for scalar constant, reg->var_off.mask should be 0. so we will have > reg->smin_value = reg->smax_value = (s64)reg->var_off.value. > > The smin_val is also used below, e.g., BPF_ADD, for a known value. > That is why I am using smin_val here. > > Will add a comment and submit v2. it would be way-way more obvious (and reliable in the long run, probably) if you just used (known && reg->var_off.value == 0). or just tnum_equals_const(reg->var_off, 0)? > > > > >> + break; > >> + /* fall-through */ > >> case PTR_TO_PACKET_END: > >> case PTR_TO_SOCKET: > >> case PTR_TO_SOCKET_OR_NULL: > >> -- > >> 2.24.1 > >>
On Fri, Sep 4, 2020 at 5:08 PM Andrii Nakryiko <andrii.nakryiko@gmail.com> wrote: > > On Fri, Sep 4, 2020 at 4:20 PM Yonghong Song <yhs@fb.com> wrote: > > > > > > > > On 9/4/20 1:30 PM, Andrii Nakryiko wrote: > > > On Fri, Sep 4, 2020 at 12:49 PM Yonghong Song <yhs@fb.com> wrote: > > >> > > >> Commit 41c48f3a98231 ("bpf: Support access > > >> to bpf map fields") added support to access map fields > > >> with CORE support. For example, > > >> > > >> struct bpf_map { > > >> __u32 max_entries; > > >> } __attribute__((preserve_access_index)); > > >> > > >> struct bpf_array { > > >> struct bpf_map map; > > >> __u32 elem_size; > > >> } __attribute__((preserve_access_index)); > > >> > > >> struct { > > >> __uint(type, BPF_MAP_TYPE_ARRAY); > > >> __uint(max_entries, 4); > > >> __type(key, __u32); > > >> __type(value, __u32); > > >> } m_array SEC(".maps"); > > >> > > >> SEC("cgroup_skb/egress") > > >> int cg_skb(void *ctx) > > >> { > > >> struct bpf_array *array = (struct bpf_array *)&m_array; > > >> > > >> /* .. array->map.max_entries .. */ > > >> } > > >> > > >> In kernel, bpf_htab has similar structure, > > >> > > >> struct bpf_htab { > > >> struct bpf_map map; > > >> ... > > >> } > > >> > > >> In the above cg_skb(), to access array->map.max_entries, with CORE, the clang will > > >> generate two builtin's. > > >> base = &m_array; > > >> /* access array.map */ > > >> map_addr = __builtin_preserve_struct_access_info(base, 0, 0); > > >> /* access array.map.max_entries */ > > >> max_entries_addr = __builtin_preserve_struct_access_info(map_addr, 0, 0); > > >> max_entries = *max_entries_addr; > > >> > > >> In the current llvm, if two builtin's are in the same function or > > >> in the same function after inlining, the compiler is smart enough to chain > > >> them together and generates like below: > > >> base = &m_array; > > >> max_entries = *(base + reloc_offset); /* reloc_offset = 0 in this case */ > > >> and we are fine. > > >> > > >> But if we force no inlining for one of functions in test_map_ptr() selftest, e.g., > > >> check_default(), the above two __builtin_preserve_* will be in two different > > >> functions. In this case, we will have code like: > > >> func check_hash(): > > >> reloc_offset_map = 0; > > >> base = &m_array; > > >> map_base = base + reloc_offset_map; > > >> check_default(map_base, ...) > > >> func check_default(map_base, ...): > > >> max_entries = *(map_base + reloc_offset_max_entries); > > >> > > >> In kernel, map_ptr (CONST_PTR_TO_MAP) does not allow any arithmetic. > > >> The above "map_base = base + reloc_offset_map" will trigger a verifier failure. > > >> ; VERIFY(check_default(&hash->map, map)); > > >> 0: (18) r7 = 0xffffb4fe8018a004 > > >> 2: (b4) w1 = 110 > > >> 3: (63) *(u32 *)(r7 +0) = r1 > > >> R1_w=invP110 R7_w=map_value(id=0,off=4,ks=4,vs=8,imm=0) R10=fp0 > > >> ; VERIFY_TYPE(BPF_MAP_TYPE_HASH, check_hash); > > >> 4: (18) r1 = 0xffffb4fe8018a000 > > >> 6: (b4) w2 = 1 > > >> 7: (63) *(u32 *)(r1 +0) = r2 > > >> R1_w=map_value(id=0,off=0,ks=4,vs=8,imm=0) R2_w=invP1 R7_w=map_value(id=0,off=4,ks=4,vs=8,imm=0) R10=fp0 > > >> 8: (b7) r2 = 0 > > >> 9: (18) r8 = 0xffff90bcb500c000 > > >> 11: (18) r1 = 0xffff90bcb500c000 > > >> 13: (0f) r1 += r2 > > >> R1 pointer arithmetic on map_ptr prohibited > > >> > > >> To fix the issue, let us permit map_ptr + 0 arithmetic which will > > >> result in exactly the same map_ptr. > > >> > > >> Signed-off-by: Yonghong Song <yhs@fb.com> > > >> --- > > >> kernel/bpf/verifier.c | 3 +++ > > >> 1 file changed, 3 insertions(+) > > >> > > >> diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c > > >> index b4e9c56b8b32..92aa985e99df 100644 > > >> --- a/kernel/bpf/verifier.c > > >> +++ b/kernel/bpf/verifier.c > > >> @@ -5317,6 +5317,9 @@ static int adjust_ptr_min_max_vals(struct bpf_verifier_env *env, > > >> dst, reg_type_str[ptr_reg->type]); > > >> return -EACCES; > > >> case CONST_PTR_TO_MAP: > > >> + if (known && smin_val == 0 && opcode == BPF_ADD) > > > > > > does smin_val imply that var_off is strictly zero? if that's the case, > > > can you please leave a comment stating this clearly, it's hard to tell > > > if that's enough of a check. > > > > It should be, if register state is maintained properly, the following > > function (or its functionality) should have been called. > > > > static void __update_reg64_bounds(struct bpf_reg_state *reg) > > { > > /* min signed is max(sign bit) | min(other bits) */ > > reg->smin_value = max_t(s64, reg->smin_value, > > reg->var_off.value | (reg->var_off.mask > > & S64_MIN)); > > /* max signed is min(sign bit) | max(other bits) */ > > reg->smax_value = min_t(s64, reg->smax_value, > > reg->var_off.value | (reg->var_off.mask > > & S64_MAX)); > > reg->umin_value = max(reg->umin_value, reg->var_off.value); > > reg->umax_value = min(reg->umax_value, > > reg->var_off.value | reg->var_off.mask); > > } > > > > for scalar constant, reg->var_off.mask should be 0. so we will have > > reg->smin_value = reg->smax_value = (s64)reg->var_off.value. > > > > The smin_val is also used below, e.g., BPF_ADD, for a known value. > > That is why I am using smin_val here. > > > > Will add a comment and submit v2. > > it would be way-way more obvious (and reliable in the long run, > probably) if you just used (known && reg->var_off.value == 0). or just > tnum_equals_const(reg->var_off, 0)? Pls dont. smin_val == 0 is a standard way to do this. Just check all other places in this function and everywhere else.
On 9/5/20 2:10 AM, Alexei Starovoitov wrote: > On Fri, Sep 4, 2020 at 5:08 PM Andrii Nakryiko > <andrii.nakryiko@gmail.com> wrote: >> On Fri, Sep 4, 2020 at 4:20 PM Yonghong Song <yhs@fb.com> wrote: [...] >>> for scalar constant, reg->var_off.mask should be 0. so we will have >>> reg->smin_value = reg->smax_value = (s64)reg->var_off.value. >>> >>> The smin_val is also used below, e.g., BPF_ADD, for a known value. >>> That is why I am using smin_val here. >>> >>> Will add a comment and submit v2. >> >> it would be way-way more obvious (and reliable in the long run, >> probably) if you just used (known && reg->var_off.value == 0). or just >> tnum_equals_const(reg->var_off, 0)? > > Pls dont. smin_val == 0 is a standard way to do this. > Just check all other places in this function and everywhere else. Also, we taint the reg earlier in that function if its known and min != max: if ((known && (smin_val != smax_val || umin_val != umax_val)) || smin_val > smax_val || umin_val > umax_val) { /* Taint dst register if offset had invalid bounds derived from * e.g. dead branches. */ __mark_reg_unknown(env, dst_reg); return 0; }
diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c index b4e9c56b8b32..92aa985e99df 100644 --- a/kernel/bpf/verifier.c +++ b/kernel/bpf/verifier.c @@ -5317,6 +5317,9 @@ static int adjust_ptr_min_max_vals(struct bpf_verifier_env *env, dst, reg_type_str[ptr_reg->type]); return -EACCES; case CONST_PTR_TO_MAP: + if (known && smin_val == 0 && opcode == BPF_ADD) + break; + /* fall-through */ case PTR_TO_PACKET_END: case PTR_TO_SOCKET: case PTR_TO_SOCKET_OR_NULL:
Commit 41c48f3a98231 ("bpf: Support access to bpf map fields") added support to access map fields with CORE support. For example, struct bpf_map { __u32 max_entries; } __attribute__((preserve_access_index)); struct bpf_array { struct bpf_map map; __u32 elem_size; } __attribute__((preserve_access_index)); struct { __uint(type, BPF_MAP_TYPE_ARRAY); __uint(max_entries, 4); __type(key, __u32); __type(value, __u32); } m_array SEC(".maps"); SEC("cgroup_skb/egress") int cg_skb(void *ctx) { struct bpf_array *array = (struct bpf_array *)&m_array; /* .. array->map.max_entries .. */ } In kernel, bpf_htab has similar structure, struct bpf_htab { struct bpf_map map; ... } In the above cg_skb(), to access array->map.max_entries, with CORE, the clang will generate two builtin's. base = &m_array; /* access array.map */ map_addr = __builtin_preserve_struct_access_info(base, 0, 0); /* access array.map.max_entries */ max_entries_addr = __builtin_preserve_struct_access_info(map_addr, 0, 0); max_entries = *max_entries_addr; In the current llvm, if two builtin's are in the same function or in the same function after inlining, the compiler is smart enough to chain them together and generates like below: base = &m_array; max_entries = *(base + reloc_offset); /* reloc_offset = 0 in this case */ and we are fine. But if we force no inlining for one of functions in test_map_ptr() selftest, e.g., check_default(), the above two __builtin_preserve_* will be in two different functions. In this case, we will have code like: func check_hash(): reloc_offset_map = 0; base = &m_array; map_base = base + reloc_offset_map; check_default(map_base, ...) func check_default(map_base, ...): max_entries = *(map_base + reloc_offset_max_entries); In kernel, map_ptr (CONST_PTR_TO_MAP) does not allow any arithmetic. The above "map_base = base + reloc_offset_map" will trigger a verifier failure. ; VERIFY(check_default(&hash->map, map)); 0: (18) r7 = 0xffffb4fe8018a004 2: (b4) w1 = 110 3: (63) *(u32 *)(r7 +0) = r1 R1_w=invP110 R7_w=map_value(id=0,off=4,ks=4,vs=8,imm=0) R10=fp0 ; VERIFY_TYPE(BPF_MAP_TYPE_HASH, check_hash); 4: (18) r1 = 0xffffb4fe8018a000 6: (b4) w2 = 1 7: (63) *(u32 *)(r1 +0) = r2 R1_w=map_value(id=0,off=0,ks=4,vs=8,imm=0) R2_w=invP1 R7_w=map_value(id=0,off=4,ks=4,vs=8,imm=0) R10=fp0 8: (b7) r2 = 0 9: (18) r8 = 0xffff90bcb500c000 11: (18) r1 = 0xffff90bcb500c000 13: (0f) r1 += r2 R1 pointer arithmetic on map_ptr prohibited To fix the issue, let us permit map_ptr + 0 arithmetic which will result in exactly the same map_ptr. Signed-off-by: Yonghong Song <yhs@fb.com> --- kernel/bpf/verifier.c | 3 +++ 1 file changed, 3 insertions(+)