On Tue, 22 Sep 2020 at 19:38, Martin KaFai Lau [off-list ref] wrote:

On Tue, Sep 22, 2020 at 10:56:55AM +0100, Lorenz Bauer wrote:

quoted

On Tue, 22 Sep 2020 at 08:04, Martin KaFai Lau [off-list ref] wrote:

quoted

check_reg_type() checks whether a reg can be used as an arg of a
func_proto.  For PTR_TO_BTF_ID, the check is actually not
completely done until the reg->btf_id is pointing to a
kernel struct that is acceptable by the func_proto.

Thus, this patch moves the btf_id check into check_reg_type().
The compatible_reg_types[] usage is localized in check_reg_type() now.

The "if (!btf_id) verbose(...); " is also removed since it won't happen.

Signed-off-by: Martin KaFai Lau <redacted>
---
 kernel/bpf/verifier.c | 65 +++++++++++++++++++++++--------------------
 1 file changed, 35 insertions(+), 30 deletions(-)

diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index 15ab889b0a3f..3ce61c412ea0 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c

@@ -4028,20 +4028,29 @@ static const struct bpf_reg_types *compatible_reg_types[] = {
        [__BPF_ARG_TYPE_MAX]            = NULL,
 };

-static int check_reg_type(struct bpf_verifier_env *env, u32 regno,
-                         const struct bpf_reg_types *compatible)
+static int check_reg_type(struct bpf_verifier_env *env, u32 arg,
+                         enum bpf_arg_type arg_type,
+                         const struct bpf_func_proto *fn)

Yes. I think that works as good.

An idea for the mid term, I think this map helper's arg override logic
should belong to a new map_ops and this new map_ops can return the whole
"fn" instead of overriding on an arg-by-arg base.

Yeah, agreed. I've had a similar idea, but no time to implement it yet.

-- 
Lorenz Bauer  |  Systems Engineer
6th Floor, County Hall/The Riverside Building, SE1 7PB, UK

www.cloudflare.com