public inbox for libc-alpha@sourceware.org
 help / color / mirror / Atom feed
From: k4lizen <k4lizen@proton.me>
To: "libc-alpha@sourceware.org" <libc-alpha@sourceware.org>
Subject: [PATCH v2] malloc: send freed small chunks to smallbin
Date: Mon, 08 Jul 2024 05:39:40 +0000	[thread overview]
Message-ID: <UVNjyu5K3y4Pxi6Rwr1-pohsKhwE3nlbYOnMjmOuSvqGnpUGISJu00igzozfDrHYX26FyZQ1t0PRMc_6XBViXLGIL6YhNmr9gHtz5_g7eXk=@proton.me> (raw)

[-- Attachment #1: Type: text/plain, Size: 2705 bytes --]

Large chunks get added to the unsorted bin since
sorting them takes time, for small chunks the
benefit of adding them to the unsorted bin is
non-existant, actually hurting performance.

Splitting and malloc_consolidate still add small
chunks to unsorted, but we can hint the compiler
that that is a relatively rare occurance.
Benchmarking shows this to be consistently good.
---
malloc/malloc.c | 59 +++++++++++++++++++++++++++++++++----------------
1 file changed, 40 insertions(+), 19 deletions(-)

diff --git a/malloc/malloc.c b/malloc/malloc.c
index bcb6e5b83c..ad77cd083e 100644
--- a/malloc/malloc.c
+++ b/malloc/malloc.c
@@ -4156,9 +4156,9 @@ _int_malloc (mstate av, size_t bytes)
#endif
}

- /* place chunk in bin */
-
- if (in_smallbin_range (size))
+ /* Place chunk in bin. Only malloc_consolidate() and splitting can put
+ small chunks into the unsorted bin. */
+ if (__glibc_unlikely (in_smallbin_range (size)))
{
victim_index = smallbin_index (size);
bck = bin_at (av, victim_index);
@@ -4723,23 +4723,45 @@ _int_free_create_chunk (mstate av, mchunkptr p, INTERNAL_SIZE_T size,
} else
clear_inuse_bit_at_offset(nextchunk, 0);

- /*
- Place the chunk in unsorted chunk list. Chunks are
- not placed into regular bins until after they have
- been given one chance to be used in malloc.
- */
+ mchunkptr bck, fwd;

- mchunkptr bck = unsorted_chunks (av);
- mchunkptr fwd = bck->fd;
- if (__glibc_unlikely (fwd->bk != bck))
- malloc_printerr ("free(): corrupted unsorted chunks");
- p->fd = fwd;
+ if(!in_smallbin_range (size))
+ {
+ /*
+ Place large chunks in unsorted chunk list. Large chunks are
+ not placed into regular bins until after they have
+ been given one chance to be used in malloc.
+
+ This branch is first in the if-statement to help branch
+ prediction on consecutive adjacent frees.
+ */
+
+ bck = unsorted_chunks (av);
+ fwd = bck->fd;
+ if (__glibc_unlikely (fwd->bk != bck))
+ malloc_printerr ("free(): corrupted unsorted chunks");
+ p->fd_nextsize = NULL;
+ p->bk_nextsize = NULL;
+ }
+ else
+ {
+ /*
+ Place small chunks directly in their smallbin, so they
+ don't pollute the unsorted bin.
+ */
+
+ int chunk_index = smallbin_index (size);
+ bck = bin_at (av, chunk_index);
+ fwd = bck->fd;
+
+ if (__glibc_unlikely (fwd->bk != bck))
+ malloc_printerr ("free(): chunks in smallbin corrupted");
+
+ mark_bin (av, chunk_index);
+ }
+
p->bk = bck;
- if (!in_smallbin_range(size))
- {
- p->fd_nextsize = NULL;
- p->bk_nextsize = NULL;
- }
+ p->fd = fwd;
bck->fd = p;
fwd->bk = p;

@@ -4748,7 +4770,6 @@ _int_free_create_chunk (mstate av, mchunkptr p, INTERNAL_SIZE_T size,

check_free_chunk(av, p);
}
-
else
{
/* If the chunk borders the current high end of memory,
--
2.45.2

             reply	other threads:[~2024-07-08  5:39 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-07-08  5:39 k4lizen [this message]
2024-07-08  5:51 ` k4lizen
2024-10-21 17:19 Wilco Dijkstra

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='UVNjyu5K3y4Pxi6Rwr1-pohsKhwE3nlbYOnMjmOuSvqGnpUGISJu00igzozfDrHYX26FyZQ1t0PRMc_6XBViXLGIL6YhNmr9gHtz5_g7eXk=@proton.me' \
    --to=k4lizen@proton.me \
    --cc=libc-alpha@sourceware.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 read-only IMAP folder(s) and NNTP newsgroup(s).