unofficial mirror of libc-alpha@sourceware.org
 help / color / Atom feed
From: Carlos O'Donell <carlos@redhat.com>
To: Wilco Dijkstra <Wilco.Dijkstra@arm.com>,
	Ondřej Bílka <neleai@seznam.cz>,
	Florian Weimer <fweimer@redhat.com>
Cc: Joseph Myers <joseph@codesourcery.com>,
	"libc-alpha@sourceware.org" <libc-alpha@sourceware.org>,
	nd <nd@arm.com>
Subject: Re: [PATCH v2] Add malloc micro benchmark
Date: Wed, 28 Feb 2018 10:21:20 -0800
Message-ID: <a9367dd0-f130-a23c-df7b-14d50ed10cfa@redhat.com> (raw)
In-Reply-To: <DB6PR0801MB20537F726B53D8373CBD6BD083C70@DB6PR0801MB2053.eurprd08.prod.outlook.com>

On 02/28/2018 09:01 AM, Wilco Dijkstra wrote:
> Ondøej Bílka wrote:
>   
>>> I think a heap-style allocator which does not segregate allocations
>>> of different sizes still has its place, and why not provide one in
>>> glibc?
>>>
>> That isn't case for any allocator and it is asking for trouble. You want
>> to avoid sitation where two big chunks couldn't be merged because of
>> tiny chunk between them.
> 
> Agreed, you always want to special case small blocks. I don't believe there is
> any advantage in using a single big heap.
> 
>> For larger size this representation is still problematic and you could
>> improve performance with another representation that also avoids
>> alignment problem by placing metadata elsewhere(for mine only 4 bytes are needed).
> 
> Larger sizes would be helped a lot once small blocks are dealt with separately.
> So I don't think we need complicated balanced binary trees when dealing with a
> small number of large blocks. You won't need an unsorted list either, large blocks
> can be merged in O(1) time.
> 
> There may be an advantage to place meta data elsewhere, for example it could make
> adding/removing/walking free lists much faster (spatial locality) or to make heap
> overflow attacks almost impossible,

I agree with many of the things you and Ondrej are proposing.

The outcome of our current discussions should be an incremental plan to go from
where we are today, to where we want to be tomorrow.

However, I do *not* believe it is a good plan to simply throw away the present
allocator and claim it should be replaced from scratch. We do not have that luxury
as a core project, we must remain answerable to our users.

A high-level concrete problem today with glibc's malloc, and the only problem being
reported by our users is that it consumes too much RSS. Solving that problem in the
abstract is what we should be looking at.

If we think that having multiple heaps for different sized objects is the way to
do this, then we should think about how to go down that path with an experiment.

Any cleanup we do before that is a win.

-- 
Cheers,
Carlos.


  reply index

Thread overview: 32+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-12-01 13:51 [PATCH] " Wilco Dijkstra
2017-12-01 16:13 ` Carlos O'Donell
2017-12-18 15:18   ` Wilco Dijkstra
2017-12-18 16:32     ` Carlos O'Donell
2018-01-02 18:20       ` [PATCH v2] " Wilco Dijkstra
2018-01-02 18:45         ` DJ Delorie
2018-01-03 12:12           ` Wilco Dijkstra
2018-01-03 15:07             ` Carlos O'Donell
2018-01-04 13:48               ` Wilco Dijkstra
2018-01-04 16:37                 ` Adhemerval Zanella
2018-01-05 14:32                 ` Carlos O'Donell
2018-01-05 15:50                   ` Adhemerval Zanella
2018-01-05 16:17                     ` Carlos O'Donell
2018-01-05 16:46                       ` Adhemerval Zanella
2018-01-05 17:27                         ` Carlos O'Donell
2018-01-05 14:33         ` Carlos O'Donell
2018-01-05 16:28           ` Joseph Myers
2018-01-05 17:26             ` Carlos O'Donell
2018-02-28 12:40               ` Florian Weimer
2018-02-28 14:11                 ` Ondřej Bílka
2018-02-28 14:16                   ` Florian Weimer
2018-02-28 16:16                     ` Carlos O'Donell
2018-02-28 20:17                       ` Ondřej Bílka
2018-02-28 16:46                     ` Ondřej Bílka
2018-02-28 17:01                       ` Wilco Dijkstra
2018-02-28 18:21                         ` Carlos O'Donell [this message]
2018-02-28 19:56                         ` Ondřej Bílka
2018-02-28 21:56                           ` DJ Delorie
2018-03-01 11:24                             ` Ondřej Bílka
2017-12-18 23:02     ` [PATCH] " DJ Delorie
2017-12-28 14:09       ` Wilco Dijkstra
2017-12-28 19:01         ` DJ Delorie

Reply instructions:

You may reply publically 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: https://www.gnu.org/software/libc/involved.html

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=a9367dd0-f130-a23c-df7b-14d50ed10cfa@redhat.com \
    --to=carlos@redhat.com \
    --cc=Wilco.Dijkstra@arm.com \
    --cc=fweimer@redhat.com \
    --cc=joseph@codesourcery.com \
    --cc=libc-alpha@sourceware.org \
    --cc=nd@arm.com \
    --cc=neleai@seznam.cz \
    /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

unofficial mirror of libc-alpha@sourceware.org

Archives are clonable:
	git clone --mirror https://public-inbox.org/libc-alpha/0 libc-alpha/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 libc-alpha libc-alpha/ https://public-inbox.org/libc-alpha \
		libc-alpha@sourceware.org
	public-inbox-index libc-alpha


Newsgroups are available over NNTP:
	nntp://news.public-inbox.org/inbox.comp.lib.glibc.alpha
	nntp://ou63pmih66umazou.onion/inbox.comp.lib.glibc.alpha
	nntp://news.gmane.org/gmane.comp.lib.glibc.alpha

 note: .onion URLs require Tor: https://www.torproject.org/

AGPL code for this site: git clone https://public-inbox.org/ public-inbox