From: Kui-Feng Lee <thinker.li@gmail.com>
To: bpf@vger.kernel.org, ast@kernel.org, martin.lau@linux.dev,
andrii@kernel.org
Cc: sinquersw@gmail.com, kuifeng@meta.com,
Kui-Feng Lee <thinker.li@gmail.com>
Subject: [RFC bpf-next v4 3/6] bpf: Handle BPF_UPTR in verifier.
Date: Fri, 16 Aug 2024 12:12:10 -0700 [thread overview]
Message-ID: <20240816191213.35573-4-thinker.li@gmail.com> (raw)
In-Reply-To: <20240816191213.35573-1-thinker.li@gmail.com>
Give PTR_TO_MEM | PTR_MAYBE_NULL to the memory pointed by an uptr with the
size of the pointed type to make them readable and writable.
Signed-off-by: Kui-Feng Lee <thinker.li@gmail.com>
---
kernel/bpf/verifier.c | 37 ++++++++++++++++++++++++++++++++++++-
1 file changed, 36 insertions(+), 1 deletion(-)
diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index e3932f8ce10a..5bc5b37b63cc 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -5340,6 +5340,10 @@ static int map_kptr_match_type(struct bpf_verifier_env *env,
int perm_flags;
const char *reg_name = "";
+ if (kptr_field->type == BPF_UPTR)
+ /* BPF programs should not change any user kptr */
+ return -EACCES;
+
if (btf_is_kernel(reg->btf)) {
perm_flags = PTR_MAYBE_NULL | PTR_TRUSTED | MEM_RCU;
@@ -5488,6 +5492,29 @@ static u32 btf_ld_kptr_type(struct bpf_verifier_env *env, struct btf_field *kptr
return ret;
}
+static int mark_uptr_ld_reg(struct bpf_verifier_env *env, u32 regno,
+ struct btf_field *field)
+{
+ struct bpf_reg_state *val_reg;
+ const struct btf_type *t;
+ u32 type_id, tsz;
+
+ val_reg = reg_state(env, regno);
+ type_id = field->kptr.btf_id;
+ t = btf_type_id_size(field->kptr.btf, &type_id, &tsz);
+ if (!t) {
+ verbose(env, "The type of uptr is invalid");
+ return -EACCES;
+ }
+
+ mark_reg_known_zero(env, cur_regs(env), regno);
+ val_reg->type = PTR_TO_MEM | PTR_MAYBE_NULL;
+ val_reg->mem_size = tsz;
+ val_reg->id = ++env->id_gen;
+
+ return 0;
+}
+
static int check_map_kptr_access(struct bpf_verifier_env *env, u32 regno,
int value_regno, int insn_idx,
struct btf_field *kptr_field)
@@ -5516,9 +5543,16 @@ static int check_map_kptr_access(struct bpf_verifier_env *env, u32 regno,
verbose(env, "store to referenced kptr disallowed\n");
return -EACCES;
}
+ if (class != BPF_LDX && kptr_field->type == BPF_UPTR) {
+ verbose(env, "store to uptr disallowed\n");
+ return -EACCES;
+ }
if (class == BPF_LDX) {
val_reg = reg_state(env, value_regno);
+ if (kptr_field->type == BPF_UPTR)
+ return mark_uptr_ld_reg(env, value_regno, kptr_field);
+
/* We can simply mark the value_regno receiving the pointer
* value from map as PTR_TO_BTF_ID, with the correct type.
*/
@@ -5576,6 +5610,7 @@ static int check_map_access(struct bpf_verifier_env *env, u32 regno,
case BPF_KPTR_UNREF:
case BPF_KPTR_REF:
case BPF_KPTR_PERCPU:
+ case BPF_UPTR:
if (src != ACCESS_DIRECT) {
verbose(env, "kptr cannot be accessed indirectly by helper\n");
return -EACCES;
@@ -6956,7 +6991,7 @@ static int check_mem_access(struct bpf_verifier_env *env, int insn_idx, u32 regn
return err;
if (tnum_is_const(reg->var_off))
kptr_field = btf_record_find(reg->map_ptr->record,
- off + reg->var_off.value, BPF_KPTR);
+ off + reg->var_off.value, BPF_KPTR | BPF_UPTR);
if (kptr_field) {
err = check_map_kptr_access(env, regno, value_regno, insn_idx, kptr_field);
} else if (t == BPF_READ && value_regno >= 0) {
--
2.34.1
next prev parent reply other threads:[~2024-08-16 19:12 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-08-16 19:12 [RFC bpf-next v4 0/6] Share user memory to BPF program through task storage map Kui-Feng Lee
2024-08-16 19:12 ` [RFC bpf-next v4 1/6] bpf: define BPF_UPTR a new enumerator of btf_field_type Kui-Feng Lee
2024-08-16 19:12 ` [RFC bpf-next v4 2/6] bpf: Parse and support "uptr" tag Kui-Feng Lee
2024-08-16 19:12 ` Kui-Feng Lee [this message]
2024-08-16 19:12 ` [RFC bpf-next v4 4/6] bpf: pin, translate, and unpin __uptr from syscalls Kui-Feng Lee
2024-08-28 23:24 ` Alexei Starovoitov
2024-09-04 22:21 ` Martin KaFai Lau
2024-09-06 20:11 ` Martin KaFai Lau
2024-09-06 23:44 ` Alexei Starovoitov
2024-09-07 1:32 ` Martin KaFai Lau
2024-09-07 4:03 ` Martin KaFai Lau
2024-08-16 19:12 ` [RFC bpf-next v4 5/6] libbpf: define __uptr Kui-Feng Lee
2024-08-27 23:13 ` Andrii Nakryiko
2024-08-28 17:53 ` Kui-Feng Lee
2024-08-16 19:12 ` [RFC bpf-next v4 6/6] selftests/bpf: test __uptr on the value of a task storage map Kui-Feng Lee
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20240816191213.35573-4-thinker.li@gmail.com \
--to=thinker.li@gmail.com \
--cc=andrii@kernel.org \
--cc=ast@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=kuifeng@meta.com \
--cc=martin.lau@linux.dev \
--cc=sinquersw@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.