From: Andrii Nakryiko <andrii@kernel.org>
To: <dwarves@vger.kernel.org>
Cc: <bpf@vger.kernel.org>, <kernel-team@fb.com>, <ast@kernel.org>,
Andrii Nakryiko <andrii@kernel.org>,
Arnaldo Carvalho de Melo <acme@kernel.org>,
Andrii Nakryiko <andriin@fb.com>
Subject: [PATCH v2 dwarves 3/8] btf_encoder: fix emitting __ARRAY_SIZE_TYPE__ as index range type
Date: Thu, 8 Oct 2020 16:39:55 -0700 [thread overview]
Message-ID: <20201008234000.740660-4-andrii@kernel.org> (raw)
In-Reply-To: <20201008234000.740660-1-andrii@kernel.org>
From: Andrii Nakryiko <andriin@fb.com>
Fix the logic of determining if __ARRAY_SIZE_TYPE__ needs to be emitted.
Previously, such type could be emitted unnecessarily due to some particular CU
not having an int type in it. That would happen even if there was no array
type in that CU. Fix it by keeping track of 'int' type across CUs and only
emitting __ARRAY_SIZE_TYPE__ if a given CU has array type, but we still
haven't found 'int' type.
Testing against vmlinux shows that now there are no __ARRAY_SIZE_TYPE__
integers emitted.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
---
btf_encoder.c | 30 +++++++++++++++++++++---------
1 file changed, 21 insertions(+), 9 deletions(-)
diff --git a/btf_encoder.c b/btf_encoder.c
index 6e6bce202438..2e5df03e040f 100644
--- a/btf_encoder.c
+++ b/btf_encoder.c
@@ -147,6 +147,8 @@ static int32_t enumeration_type__encode(struct btf_elf *btfe, struct cu *cu, str
return type_id;
}
+static bool need_index_type;
+
static int tag__encode_btf(struct cu *cu, struct tag *tag, uint32_t core_id, struct btf_elf *btfe,
uint32_t array_index_id, uint32_t type_id_off)
{
@@ -179,6 +181,7 @@ static int tag__encode_btf(struct cu *cu, struct tag *tag, uint32_t core_id, str
return structure_type__encode(btfe, cu, tag, type_id_off);
case DW_TAG_array_type:
/* TODO: Encode one dimension at a time. */
+ need_index_type = true;
return btf_elf__add_array(btfe, ref_type_id, array_index_id, array_type__nelems(tag));
case DW_TAG_enumeration_type:
return enumeration_type__encode(btfe, cu, tag);
@@ -193,6 +196,7 @@ static int tag__encode_btf(struct cu *cu, struct tag *tag, uint32_t core_id, str
static struct btf_elf *btfe;
static uint32_t array_index_id;
+static bool has_index_type;
int btf_encoder__encode()
{
@@ -228,7 +232,6 @@ static struct variable *hashaddr__find_variable(const struct hlist_head hashtabl
int cu__encode_btf(struct cu *cu, int verbose, bool force,
bool skip_encoding_vars)
{
- bool add_index_type = false;
uint32_t type_id_off;
uint32_t core_id;
struct function *fn;
@@ -254,18 +257,26 @@ int cu__encode_btf(struct cu *cu, int verbose, bool force,
if (!btfe)
return -1;
- /* cu__find_base_type_by_name() takes "type_id_t *id" */
- type_id_t id;
- if (!cu__find_base_type_by_name(cu, "int", &id)) {
- add_index_type = true;
- id = cu->types_table.nr_entries;
- }
- array_index_id = id;
+ has_index_type = false;
+ need_index_type = false;
+ array_index_id = 0;
if (verbose)
printf("File %s:\n", btfe->filename);
}
+ if (!has_index_type) {
+ /* cu__find_base_type_by_name() takes "type_id_t *id" */
+ type_id_t id;
+ if (cu__find_base_type_by_name(cu, "int", &id)) {
+ has_index_type = true;
+ array_index_id = id;
+ } else {
+ has_index_type = false;
+ array_index_id = cu->types_table.nr_entries;
+ }
+ }
+
btf_elf__verbose = verbose;
type_id_off = btf__get_nr_types(btfe->btf);
@@ -279,12 +290,13 @@ int cu__encode_btf(struct cu *cu, int verbose, bool force,
}
}
- if (add_index_type) {
+ if (need_index_type && !has_index_type) {
struct base_type bt = {};
bt.name = 0;
bt.bit_size = 32;
btf_elf__add_base_type(btfe, &bt, "__ARRAY_SIZE_TYPE__");
+ has_index_type = true;
}
cu__for_each_function(cu, core_id, fn) {
--
2.24.1
next prev parent reply other threads:[~2020-10-08 23:40 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-10-08 23:39 [PATCH v2 dwarves 0/8] Switch BTF loading and encoding to libbpf APIs Andrii Nakryiko
2020-10-08 23:39 ` [PATCH v2 dwarves 1/8] btf_loader: use libbpf to load BTF Andrii Nakryiko
2020-10-09 14:53 ` Arnaldo Carvalho de Melo
2020-10-09 17:49 ` Andrii Nakryiko
2020-10-08 23:39 ` [PATCH v2 dwarves 2/8] btf_encoder: use libbpf APIs to encode BTF type info Andrii Nakryiko
2020-10-08 23:39 ` Andrii Nakryiko [this message]
2020-10-08 23:39 ` [PATCH v2 dwarves 4/8] btf_encoder: discard CUs after BTF encoding Andrii Nakryiko
2020-10-09 15:47 ` Arnaldo Carvalho de Melo
2020-10-08 23:39 ` [PATCH v2 dwarves 5/8] btf_encoder: revamp how per-CPU variables are encoded Andrii Nakryiko
2020-10-09 15:53 ` Arnaldo Carvalho de Melo
2020-10-08 23:39 ` [PATCH v2 dwarves 6/8] dwarf_loader: increase the size of lookup hash map Andrii Nakryiko
2020-10-08 23:39 ` [PATCH v2 dwarves 7/8] strings: use BTF's string APIs for strings management Andrii Nakryiko
2020-10-08 23:40 ` [PATCH v2 dwarves 8/8] btf_encoder: support cross-compiled ELF binaries with different endianness Andrii Nakryiko
2020-10-09 16:22 ` [PATCH v2 dwarves 0/8] Switch BTF loading and encoding to libbpf APIs Arnaldo Carvalho de Melo
2020-10-09 17:59 ` Andrii Nakryiko
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=20201008234000.740660-4-andrii@kernel.org \
--to=andrii@kernel.org \
--cc=acme@kernel.org \
--cc=andriin@fb.com \
--cc=ast@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=dwarves@vger.kernel.org \
--cc=kernel-team@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).