From: Jeff King <peff@peff.net>
To: Taylor Blau <me@ttaylorr.com>
Cc: git@vger.kernel.org, Junio C Hamano <gitster@pobox.com>,
Derrick Stolee <derrickstolee@github.com>,
Abhradeep Chakraborty <chakrabortyabhradeep79@gmail.com>
Subject: Re: [PATCH 3/6] pack-bitmap.c: drop unnecessary 'inline's
Date: Tue, 21 Mar 2023 13:40:33 -0400 [thread overview]
Message-ID: <20230321174033.GE3119834@coredump.intra.peff.net> (raw)
In-Reply-To: <2e3e3a7145a5851fcf5c485b38d14344c9b824d7.1679342296.git.me@ttaylorr.com>
On Mon, Mar 20, 2023 at 04:02:46PM -0400, Taylor Blau wrote:
> Both `read_be32()` and `read_u8()` are defined as inline dating back
> to b5007211b6 (pack-bitmap: do not use gcc packed attribute,
> 2014-11-27), though that commit does not hint at why the functions were
> defined with that attribute.
I think any non-header inline like this can implicitly be assumed to be
"I thought it would make things faster". ;)
> However (at least with GCC 12.2.0, at the time of writing), the
> resulting pack-bitmap.o contains the same instructions with or without
> the inline attribute applied to these functions:
>
> $ make O=3 pack-bitmap.o && mv pack-bitmap.o{,.before}
> [ apply this patch ]
> $ make O=3 pack-bitmap.o && mv pack-bitmap.o{,.after}
> $ objdump -d pack-bitmap.o.before >before
> $ objdump -d pack-bitmap.o.after >after
> $ diff -u pack-bitmap.o.{before,after}
> --- before 2023-03-15 18:54:17.021580095 -0400
> +++ after 2023-03-15 18:54:21.853552218 -0400
> @@ -1,5 +1,5 @@
>
> -pack-bitmap.o.before: file format elf64-x86-64
> +pack-bitmap.o.after: file format elf64-x86-64
>
> Disassembly of section .text:
>
> So defining these functions as inline is at best a noop, and at worst
> confuses the reader into thinking that there is some trickier reason
> that they are defined as inline when there isn't.
Nice digging. The "inline" is really just a hint to the compiler here,
and obviously it does not need that hint. I do wonder if that is still
true after you make them more complicated in a later patch in the
series.
On the other hand, I doubt that these need to be very optimized at all.
If there were a tight loop of single-byte reads, the function overhead
might be noticeable. But generally we're reading only a few items from
the beginning of each entry, and then reading the bulk of the data via
ewah_read_mmap().
So I think the overall argument is "let the compiler decide what is good
to inline and what is not".
-Peff
next prev parent reply other threads:[~2023-03-21 17:41 UTC|newest]
Thread overview: 29+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-03-20 20:02 [PATCH 0/6] pack-bitmap: miscellaneous mmap read hardening Taylor Blau
2023-03-20 20:02 ` [PATCH 1/6] pack-bitmap.c: hide bitmap internals in `read_u8()` Taylor Blau
2023-03-21 17:35 ` Jeff King
2023-03-24 17:52 ` Derrick Stolee
2023-03-20 20:02 ` [PATCH 2/6] pack-bitmap.c: hide bitmap internals in `read_be32()` Taylor Blau
2023-03-20 20:02 ` [PATCH 3/6] pack-bitmap.c: drop unnecessary 'inline's Taylor Blau
2023-03-21 17:40 ` Jeff King [this message]
2023-03-20 20:02 ` [PATCH 4/6] pack-bitmap.c: factor out manual `map_pos` manipulation Taylor Blau
2023-03-21 17:56 ` Jeff King
2023-03-24 18:04 ` Derrick Stolee
2023-03-24 18:29 ` Jeff King
2023-03-24 23:23 ` Taylor Blau
2023-03-25 4:57 ` Jeff King
2023-03-24 23:13 ` Taylor Blau
2023-03-24 23:24 ` Taylor Blau
2023-03-24 23:08 ` Taylor Blau
2023-03-20 20:02 ` [PATCH 5/6] pack-bitmap.c: use `bitmap_index_seek()` where possible Taylor Blau
2023-03-21 18:05 ` Jeff King
2023-03-24 18:06 ` Derrick Stolee
2023-03-24 18:35 ` Jeff King
2023-03-24 19:43 ` Junio C Hamano
2023-03-24 20:37 ` Jeff King
2023-03-24 21:38 ` Junio C Hamano
2023-03-24 22:57 ` Taylor Blau
2023-03-20 20:02 ` [PATCH 6/6] pack-bitmap.c: factor out `bitmap_index_seek_commit()` Taylor Blau
2023-03-21 18:13 ` Jeff King
2023-03-21 18:16 ` Taylor Blau
2023-03-21 18:27 ` Jeff King
2023-03-24 18:09 ` Derrick Stolee
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
List information: http://vger.kernel.org/majordomo-info.html
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20230321174033.GE3119834@coredump.intra.peff.net \
--to=peff@peff.net \
--cc=chakrabortyabhradeep79@gmail.com \
--cc=derrickstolee@github.com \
--cc=git@vger.kernel.org \
--cc=gitster@pobox.com \
--cc=me@ttaylorr.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.
Code repositories for project(s) associated with this public inbox
https://80x24.org/mirrors/git.git
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 read-only IMAP folder(s) and NNTP newsgroup(s).