From: Andrii Nakryiko <andriin@fb.com>
To: <bpf@vger.kernel.org>, <netdev@vger.kernel.org>, <ast@fb.com>,
<daniel@iogearbox.net>, <yhs@fb.com>
Cc: <andrii.nakryiko@gmail.com>, <kernel-team@fb.com>,
Andrii Nakryiko <andriin@fb.com>
Subject: [PATCH v2 bpf-next 3/5] libbpf: add support for field size relocations
Date: Fri, 1 Nov 2019 15:28:08 -0700 [thread overview]
Message-ID: <20191101222810.1246166-4-andriin@fb.com> (raw)
In-Reply-To: <20191101222810.1246166-1-andriin@fb.com>
Add bpf_core_field_size() macro, capturing a relocation against field size.
Adjust bits of internal libbpf relocation logic to allow capturing size
relocations of various field types: arrays, structs/unions, enums, etc.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
---
tools/lib/bpf/bpf_core_read.h | 7 ++++++
tools/lib/bpf/libbpf.c | 40 ++++++++++++++++++++++++++++-------
2 files changed, 39 insertions(+), 8 deletions(-)
diff --git a/tools/lib/bpf/bpf_core_read.h b/tools/lib/bpf/bpf_core_read.h
index 0935cd68e7de..11461b2623b0 100644
--- a/tools/lib/bpf/bpf_core_read.h
+++ b/tools/lib/bpf/bpf_core_read.h
@@ -96,6 +96,13 @@ enum bpf_field_info_kind {
#define bpf_core_field_exists(field) \
__builtin_preserve_field_info(field, BPF_FIELD_EXISTS)
+/*
+ * Convenience macro to get byte size of a field. Works for integers,
+ * struct/unions, pointers, arrays, and enums.
+ */
+#define bpf_core_field_size(field) \
+ __builtin_preserve_field_info(field, BPF_FIELD_BYTE_SIZE)
+
/*
* bpf_core_read() abstracts away bpf_probe_read() call and captures offset
* relocation for source address using __builtin_preserve_access_index()
diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
index f1a01ed93c4a..f93d17e7cccb 100644
--- a/tools/lib/bpf/libbpf.c
+++ b/tools/lib/bpf/libbpf.c
@@ -2569,8 +2569,10 @@ static struct ids_vec *bpf_core_find_cands(const struct btf *local_btf,
/* Check two types for compatibility, skipping const/volatile/restrict and
* typedefs, to ensure we are relocating compatible entities:
* - any two STRUCTs/UNIONs are compatible and can be mixed;
- * - any two FWDs are compatible;
+ * - any two FWDs are compatible, if their names match (modulo flavor suffix);
* - any two PTRs are always compatible;
+ * - for ENUMs, names should be the same (ignoring flavor suffix) or at
+ * least one of enums should be anonymous;
* - for ENUMs, check sizes, names are ignored;
* - for INT, size and signedness are ignored;
* - for ARRAY, dimensionality is ignored, element types are checked for
@@ -2598,11 +2600,23 @@ static int bpf_core_fields_are_compat(const struct btf *local_btf,
return 0;
switch (btf_kind(local_type)) {
- case BTF_KIND_FWD:
case BTF_KIND_PTR:
return 1;
- case BTF_KIND_ENUM:
- return local_type->size == targ_type->size;
+ case BTF_KIND_FWD:
+ case BTF_KIND_ENUM: {
+ const char *local_name, *targ_name;
+ size_t local_len, targ_len;
+
+ local_name = btf__name_by_offset(local_btf,
+ local_type->name_off);
+ targ_name = btf__name_by_offset(targ_btf, targ_type->name_off);
+ local_len = bpf_core_essential_name_len(local_name);
+ targ_len = bpf_core_essential_name_len(targ_name);
+ /* one of them is anonymous or both w/ same flavor-less names */
+ return local_len == 0 || targ_len == 0 ||
+ (local_len == targ_len &&
+ strncmp(local_name, targ_name, local_len) == 0);
+ }
case BTF_KIND_INT:
/* just reject deprecated bitfield-like integers; all other
* integers are by default compatible between each other
@@ -2791,16 +2805,23 @@ static int bpf_core_calc_field_relo(const struct bpf_program *prog,
const struct btf_member *m;
const struct btf_type *mt;
bool bitfield;
+ __s64 sz;
/* a[n] accessor needs special handling */
if (!acc->name) {
- if (relo->kind != BPF_FIELD_BYTE_OFFSET) {
- pr_warn("prog '%s': relo %d at insn #%d can't be applied to array access'\n",
+ if (relo->kind == BPF_FIELD_BYTE_OFFSET) {
+ *val = spec->bit_offset / 8;
+ } else if (relo->kind == BPF_FIELD_BYTE_SIZE) {
+ sz = btf__resolve_size(spec->btf, acc->type_id);
+ if (sz < 0)
+ return -EINVAL;
+ *val = sz;
+ } else {
+ pr_warn("prog '%s': relo %d at insn #%d can't be applied to array access\n",
bpf_program__title(prog, false),
relo->kind, relo->insn_off / 8);
return -EINVAL;
}
- *val = spec->bit_offset / 8;
if (validate)
*validate = true;
return 0;
@@ -2828,7 +2849,10 @@ static int bpf_core_calc_field_relo(const struct bpf_program *prog,
byte_off = bit_off / 8 / byte_sz * byte_sz;
}
} else {
- byte_sz = mt->size;
+ sz = btf__resolve_size(spec->btf, m->type);
+ if (sz < 0)
+ return -EINVAL;
+ byte_sz = sz;
byte_off = spec->bit_offset / 8;
bit_sz = byte_sz * 8;
}
--
2.17.1
next prev parent reply other threads:[~2019-11-01 22:28 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-11-01 22:28 [PATCH v2 bpf-next 0/5] Bitfield and size relocations support in libbpf Andrii Nakryiko
2019-11-01 22:28 ` [PATCH v2 bpf-next 1/5] selftests/bpf: remove too strict field offset relo test cases Andrii Nakryiko
2019-11-01 22:28 ` [PATCH v2 bpf-next 2/5] libbpf: add support for relocatable bitfields Andrii Nakryiko
2019-11-01 22:28 ` Andrii Nakryiko [this message]
2019-11-01 22:28 ` [PATCH v2 bpf-next 4/5] selftest/bpf: add relocatable bitfield reading tests Andrii Nakryiko
2019-11-01 22:28 ` [PATCH v2 bpf-next 5/5] selftests/bpf: add field size relocation tests Andrii Nakryiko
2019-11-04 15:38 ` [PATCH v2 bpf-next 0/5] Bitfield and size relocations support in libbpf Daniel Borkmann
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=20191101222810.1246166-4-andriin@fb.com \
--to=andriin@fb.com \
--cc=andrii.nakryiko@gmail.com \
--cc=ast@fb.com \
--cc=bpf@vger.kernel.org \
--cc=daniel@iogearbox.net \
--cc=kernel-team@fb.com \
--cc=netdev@vger.kernel.org \
--cc=yhs@fb.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).