| Message ID | 20220707211931.3415440-1-deso@posteo.net (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | 06cd4e9d5d969d713e6b710f0e5ca0bc8476ae41 |
| Delegated to: | BPF |
| Headers | show |
| Series | [bpf-next] bpf: Correctly propagate errors up from bpf_core_composites_match | expand |
Hello: This patch was applied to bpf/bpf-next.git (master) by Andrii Nakryiko <andrii@kernel.org>: On Thu, 7 Jul 2022 21:19:31 +0000 you wrote: > This change addresses a comment made earlier [0] about a missing return > of an error when __bpf_core_types_match is invoked from > bpf_core_composites_match, which could have let to us erroneously > ignoring errors. > > Regarding the typedef name check pointed out in the same context, it is > not actually an issue, because callers of the function perform a name > check for the root type anyway. To make that more obvious, let's add > comments to the function (similar to what we have for > bpf_core_types_are_compat, which is called in pretty much the same > context). > > [...] Here is the summary with links: - [bpf-next] bpf: Correctly propagate errors up from bpf_core_composites_match https://git.kernel.org/bpf/bpf-next/c/06cd4e9d5d96 You are awesome, thank you!
diff --git a/tools/lib/bpf/relo_core.c b/tools/lib/bpf/relo_core.c index fe25330..c4b0e8 100644 --- a/tools/lib/bpf/relo_core.c +++ b/tools/lib/bpf/relo_core.c @@ -1500,6 +1500,8 @@ static int bpf_core_composites_match(const struct btf *local_btf, const struct b err = __bpf_core_types_match(local_btf, local_m->type, targ_btf, targ_m->type, behind_ptr, level - 1); + if (err < 0) + return err; if (err > 0) { matched = true; break; @@ -1512,7 +1514,8 @@ static int bpf_core_composites_match(const struct btf *local_btf, const struct b return 1; } -/* Check that two types "match". +/* Check that two types "match". This function assumes that root types were + * already checked for name match. * * The matching relation is defined as follows: * - modifiers and typedefs are stripped (and, hence, effectively ignored) @@ -1561,6 +1564,10 @@ int __bpf_core_types_match(const struct btf *local_btf, __u32 local_id, const st if (!local_t || !targ_t) return -EINVAL; + /* While the name check happens after typedefs are skipped, root-level + * typedefs would still be name-matched as that's the contract with + * callers. + */ if (!bpf_core_names_match(local_btf, local_t->name_off, targ_btf, targ_t->name_off)) return 0;
This change addresses a comment made earlier [0] about a missing return of an error when __bpf_core_types_match is invoked from bpf_core_composites_match, which could have let to us erroneously ignoring errors. Regarding the typedef name check pointed out in the same context, it is not actually an issue, because callers of the function perform a name check for the root type anyway. To make that more obvious, let's add comments to the function (similar to what we have for bpf_core_types_are_compat, which is called in pretty much the same context). [0]: https://lore.kernel.org/bpf/165708121449.4919.13204634393477172905.git-patchwork-notify@kernel.org/T/#m55141e8f8cfd2e8d97e65328fa04852870d01af6 Suggested-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Daniel Müller <deso@posteo.net> --- tools/lib/bpf/relo_core.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-)