dwarves.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Douglas RAILLARD <douglas.raillard@arm.com>
To: acme@redhat.com
Cc: dwarves@vger.kernel.org, douglas.raillard@arm.com
Subject: [PATCH v3 3/6] btf_loader.c: Infer alignment info
Date: Thu, 28 Oct 2021 13:27:07 +0100	[thread overview]
Message-ID: <20211028122710.881181-4-douglas.raillard@arm.com> (raw)
In-Reply-To: <20211028122710.881181-1-douglas.raillard@arm.com>

From: Douglas Raillard <douglas.raillard@arm.com>

BTF does not carry alignment information, but it carries the offset in
structs. This allows inferring the original alignment, yielding a C
header dump that is not identical to the original C code, but is
guaranteed to lead to the same memory layout.

This allows using the output of pahole in another program to poke at
memory, with the assurance that we will not read garbage.

Note: Since the alignment is inferred from the offset, it sometimes
happens that the offset was already correctly aligned, which means the
inferred alignment will be smaller than in the original source. This
does not impact the ability to read existing structs, but it could
impact creating such struct if other client code expects higher
alignment than the one exposed in the generated header.

Signed-off-by: Douglas Raillard <douglas.raillard@arm.com>
---
 btf_loader.c | 36 ++++++++++++++++++++++++++++++++++++
 dwarves.c    |  2 +-
 dwarves.h    |  2 ++
 3 files changed, 39 insertions(+), 1 deletion(-)

diff --git a/btf_loader.c b/btf_loader.c
index 9c2daee..2885252 100644
--- a/btf_loader.c
+++ b/btf_loader.c
@@ -471,10 +471,37 @@ static int btf__load_sections(struct btf *btf, struct cu *cu)
 	return btf__load_types(btf, cu);
 }
 
+static uint32_t class__infer_alignment(uint32_t byte_offset,
+				       uint32_t natural_alignment,
+				       uint32_t smallest_offset)
+{
+	uint32_t alignment = 0;
+	uint32_t offset_delta = byte_offset - smallest_offset;
+
+	if (offset_delta) {
+		if (byte_offset % 2 == 0) {
+			/* Find the power of 2 immediately higher than
+			 * offset_delta
+			 */
+			alignment = 1 << (8 * sizeof(offset_delta) -
+					      __builtin_clz(offset_delta));
+		} else {
+			alignment = 0;
+		}
+	}
+
+	/* Natural alignment, nothing to do */
+	if (alignment <= natural_alignment || alignment == 1)
+		alignment = 0;
+
+	return alignment;
+}
+
 static int class__fixup_btf_bitfields(struct tag *tag, struct cu *cu)
 {
 	struct class_member *pos;
 	struct type *tag_type = tag__type(tag);
+	uint32_t smallest_offset = 0;
 
 	type__for_each_data_member(tag_type, pos) {
 		struct tag *type = tag__strip_typedefs_and_modifiers(&pos->tag, cu);
@@ -508,8 +535,17 @@ static int class__fixup_btf_bitfields(struct tag *tag, struct cu *cu)
 				pos->byte_offset = pos->bit_offset / 8;
 			}
 		}
+
+		pos->alignment = class__infer_alignment(pos->byte_offset,
+							tag__natural_alignment(type, cu),
+							smallest_offset);
+		smallest_offset = pos->byte_offset + pos->byte_size;
 	}
 
+	tag_type->alignment = class__infer_alignment(tag_type->size,
+						     tag__natural_alignment(tag, cu),
+						     smallest_offset);
+
 	return 0;
 }
 
diff --git a/dwarves.c b/dwarves.c
index b6f2489..bb8af5b 100644
--- a/dwarves.c
+++ b/dwarves.c
@@ -1515,7 +1515,7 @@ void class__find_holes(struct class *class)
 
 static size_t type__natural_alignment(struct type *type, const struct cu *cu);
 
-static size_t tag__natural_alignment(struct tag *tag, const struct cu *cu)
+size_t tag__natural_alignment(struct tag *tag, const struct cu *cu)
 {
 	size_t natural_alignment = 1;
 
diff --git a/dwarves.h b/dwarves.h
index 30d33fa..c2fea0a 100644
--- a/dwarves.h
+++ b/dwarves.h
@@ -1002,6 +1002,8 @@ struct type {
 
 void __type__init(struct type *type);
 
+size_t tag__natural_alignment(struct tag *tag, const struct cu *cu);
+
 static inline struct class *type__class(const struct type *type)
 {
 	return (struct class *)type;
-- 
2.25.1


  parent reply	other threads:[~2021-10-28 12:27 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-10-28 12:27 [PATCH v3 0/6] Infer BTF alignment Douglas RAILLARD
2021-10-28 12:27 ` [PATCH v3 1/6] fprintf: Fix nested struct printing Douglas RAILLARD
2021-10-28 12:40   ` Arnaldo Carvalho de Melo
2021-10-28 12:27 ` [PATCH v3 2/6] btf_loader.c: Refactor class__fixup_btf_bitfields Douglas RAILLARD
2021-10-28 13:15   ` Arnaldo Carvalho de Melo
2021-10-28 12:27 ` Douglas RAILLARD [this message]
2021-10-28 13:15   ` [PATCH v3 3/6] btf_loader.c: Infer alignment info Arnaldo Carvalho de Melo
2021-10-28 12:27 ` [PATCH v3 4/6] dwarves_fprintf: Move cacheline_size into struct conf_fprintf Douglas RAILLARD
2021-10-28 13:18   ` Arnaldo Carvalho de Melo
2021-10-28 12:27 ` [PATCH v3 5/6] btf_loader.c: Propagate struct conf_load Douglas RAILLARD
2021-10-28 13:19   ` Arnaldo Carvalho de Melo
2021-10-28 12:27 ` [PATCH v3 6/6] btf_loader.c: Use cacheline size to infer alignment Douglas RAILLARD
2021-10-28 13:24   ` Arnaldo Carvalho de Melo
2021-10-28 14:12     ` Douglas Raillard

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=20211028122710.881181-4-douglas.raillard@arm.com \
    --to=douglas.raillard@arm.com \
    --cc=acme@redhat.com \
    --cc=dwarves@vger.kernel.org \
    /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).