From: Torbjorn SVENSSON <torbjorn.svensson@st.com>
To: Matthew Joyce <matthew.joyce@embedded-brains.de>,
"newlib@sourceware.org" <newlib@sourceware.org>
Subject: RE: [PATCH v2 05/11] Remove __sinit_locks / __sinit_recursive_mutex
Date: Thu, 12 May 2022 13:25:54 +0000 [thread overview]
Message-ID: <AM6PR10MB21976F9C85346B01A8F0265B81CB9@AM6PR10MB2197.EURPRD10.PROD.OUTLOOK.COM> (raw)
In-Reply-To: <20220512121143.21473-6-matthew.joyce@embedded-brains.de>
Hello,
I'm not against this change, but I have some questions related to the change...
What about the retargetable locking scheme?
Should the mutex __sinit_recursive_mutex be removed from newlib/libc/misc/lock.c too?
Without removing the mutex, I suppose there would be a mutex that is never used in the system.
Removing the mutex would pose an API break as I assume that the functions defined in newlib/libc/include/sys/lock.h (and implemented in newlib/libc/misc/lock.c) are to be considered API as it's the functions that the application needs to provide in order to ensure newlib is thread safe.
Does this change force newlib to do a major release?
Kind regards,
Torbjörn
-----Original Message-----
From: Newlib <newlib-bounces+torbjorn.svensson=st.com@sourceware.org> On Behalf Of Matthew Joyce
Sent: den 12 maj 2022 14:12
To: newlib@sourceware.org
Subject: [PATCH v2 05/11] Remove __sinit_locks / __sinit_recursive_mutex
From: Matt Joyce <matthew.joyce@embedded-brains.de>
Removed __sinit_lock_acquire() and __sinit_lock_release(). Replace these
with __sfp_lock_acquire() and __sfp_lock_release(), respectively. This
eliminates a potential deadlock issue between __sinit() and __sfp().
Removed now unused __sinit_recursive_mutex.
---
newlib/libc/stdio/findfp.c | 19 +++----------------
newlib/libc/stdio/local.h | 2 --
2 files changed, 3 insertions(+), 18 deletions(-)
diff --git a/newlib/libc/stdio/findfp.c b/newlib/libc/stdio/findfp.c
index 66867e664..afbdad9b1 100644
--- a/newlib/libc/stdio/findfp.c
+++ b/newlib/libc/stdio/findfp.c
@@ -235,11 +235,11 @@ cleanup_stdio (struct _reent *ptr)
void
__sinit (struct _reent *s)
{
- __sinit_lock_acquire ();
+ __sfp_lock_acquire ();
if (s->__cleanup)
{
- __sinit_lock_release ();
+ __sfp_lock_release ();
return;
}
@@ -268,13 +268,12 @@ __sinit (struct _reent *s)
stderr_init (s->_stderr);
#endif /* _REENT_GLOBAL_STDIO_STREAMS */
- __sinit_lock_release ();
+ __sfp_lock_release ();
}
#ifndef __SINGLE_THREAD__
__LOCK_INIT_RECURSIVE(static, __sfp_recursive_mutex);
-__LOCK_INIT_RECURSIVE(static, __sinit_recursive_mutex);
void
__sfp_lock_acquire (void)
@@ -288,18 +287,6 @@ __sfp_lock_release (void)
__lock_release_recursive (__sfp_recursive_mutex);
}
-void
-__sinit_lock_acquire (void)
-{
- __lock_acquire_recursive (__sinit_recursive_mutex);
-}
-
-void
-__sinit_lock_release (void)
-{
- __lock_release_recursive (__sinit_recursive_mutex);
-}
-
/* Walkable file locking routine. */
static int
__fp_lock (struct _reent * ptr __unused, FILE * fp)
diff --git a/newlib/libc/stdio/local.h b/newlib/libc/stdio/local.h
index 30c534dcd..9c6f63fdb 100644
--- a/newlib/libc/stdio/local.h
+++ b/newlib/libc/stdio/local.h
@@ -287,8 +287,6 @@ char *_llicvt (char *, long long, char);
#else
void __sfp_lock_acquire (void);
void __sfp_lock_release (void);
-void __sinit_lock_acquire (void);
-void __sinit_lock_release (void);
#endif
/* Types used in positional argument support in vfprinf/vfwprintf.
--
2.31.1
next prev parent reply other threads:[~2022-05-12 13:26 UTC|newest]
Thread overview: 27+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-05-12 12:11 [PATCH v2 00/11] Decouple global file object list from _GLOBAL_REENT Matthew Joyce
2022-05-12 12:11 ` [PATCH v2 01/11] Remove duplicate stdio initializations Matthew Joyce
2022-05-12 12:11 ` [PATCH v2 02/11] Remove duplicate sglue initializations Matthew Joyce
2022-05-12 12:11 ` [PATCH v2 03/11] Declare global __sf[] only once Matthew Joyce
2022-05-12 12:11 ` [PATCH v2 04/11] Add two __sglue initialization macros Matthew Joyce
2022-05-12 12:11 ` [PATCH v2 05/11] Remove __sinit_locks / __sinit_recursive_mutex Matthew Joyce
2022-05-12 13:25 ` Torbjorn SVENSSON [this message]
2022-05-12 13:51 ` Sebastian Huber
2022-05-12 12:11 ` [PATCH v2 06/11] Move __sglue initializations to __sfp() Matthew Joyce
2022-05-12 12:11 ` [PATCH v2 07/11] Add CLEANUP_FILE define Matthew Joyce
2022-05-12 12:11 ` [PATCH v2 08/11] Add stdio_exit_handler() Matthew Joyce
2022-05-12 19:11 ` Corinna Vinschen
2022-05-17 8:36 ` Takashi Yano
2022-05-17 10:13 ` Sebastian Huber
2022-05-12 12:11 ` [PATCH v2 09/11] stdio: Replace _fwalk_reent() with _fwalk_sglue() Matthew Joyce
2022-05-12 12:11 ` [PATCH v2 10/11] Add global __sglue object for all configurations Matthew Joyce
2022-05-12 12:11 ` [PATCH v2 11/11] Remove __sglue member for one configuration Matthew Joyce
2022-05-12 19:09 ` Corinna Vinschen
2022-05-12 19:14 ` [PATCH v2 00/11] Decouple global file object list from _GLOBAL_REENT Corinna Vinschen
2022-05-12 19:44 ` Corinna Vinschen
2022-05-13 5:44 ` Sebastian Huber
2022-05-13 7:53 ` Corinna Vinschen
2022-05-13 10:54 ` Sebastian Huber
2022-05-13 11:20 ` Corinna Vinschen
2022-05-13 8:03 ` Corinna Vinschen
2022-05-13 11:20 ` Sebastian Huber
2022-05-13 11:30 ` Corinna Vinschen
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=AM6PR10MB21976F9C85346B01A8F0265B81CB9@AM6PR10MB2197.EURPRD10.PROD.OUTLOOK.COM \
--to=torbjorn.svensson@st.com \
--cc=matthew.joyce@embedded-brains.de \
--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).