From: Sebastian Huber <sebastian.huber@embedded-brains.de>
To: Patrick Bader <mail@patrickbader.eu>, newlib@sourceware.org
Subject: Re: Increased RAM usage after upgrade to newlib 4.3.0
Date: Thu, 23 Feb 2023 07:02:51 +0100 [thread overview]
Message-ID: <cee20c8a-7881-7cec-07da-ca5b41719b00@embedded-brains.de> (raw)
In-Reply-To: <9c0c090d-87fc-9f71-40f2-dd92c10cf6b8@patrickbader.eu>
Hello Patrick,
On 17.02.23 20:10, Patrick Bader wrote:
> Dear newlib group members,
>
> I have just updated my firmware from newlib 4.2.0 to newlib 4.3.0 from
> the Arch Linux packages and found out that static memory consumption
> rose by about 300 bytes due to the standard input and output streams.
> However, I am not using these streams, but am using dynamic memory,
> which is using the _reent struct internally.
>
> As I found out, this struct is always populated with the global __sf
> file descriptors in the 4.3.0 version of newlib, which are about 300
> bytes in total.
> The library is built with --enable-newlib-reent-small with the nano.spec
> and in the 4.2.0 version, _reent is initialized with some fake file
> descriptors: "__sf_fake_stdin,..." which are way smaller.
> Those fake versions are missing in newlib 4.3.0 and the __sf descriptors
> are used always.
>
> Is this intended behaviour or a regression? If it is intended, is there
> a way or some flag to get rid of the extra 300 bytes?
> I also do not need all those functions to be reentrant, but found no
> compilation parameter to turn this off.
this regression is probably a side-effect of the support for splitting
up the struct _reent into individual thread-local storage objects. See
the --enable-newlib-reent-thread-local configuration option.
--
embedded brains GmbH
Herr Sebastian HUBER
Dornierstr. 4
82178 Puchheim
Germany
email: sebastian.huber@embedded-brains.de
phone: +49-89-18 94 741 - 16
fax: +49-89-18 94 741 - 08
Registergericht: Amtsgericht München
Registernummer: HRB 157899
Vertretungsberechtigte Geschäftsführer: Peter Rasmussen, Thomas Dörfler
Unsere Datenschutzerklärung finden Sie hier:
https://embedded-brains.de/datenschutzerklaerung/
prev parent reply other threads:[~2023-02-23 6:02 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-02-17 19:10 Patrick Bader
2023-02-23 6:02 ` Sebastian Huber [this message]
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=cee20c8a-7881-7cec-07da-ca5b41719b00@embedded-brains.de \
--to=sebastian.huber@embedded-brains.de \
--cc=mail@patrickbader.eu \
--cc=newlib@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).