ist or are both NULL. ppc gets that one wrong too.
--
Alan Modra
Australia Development Lab, IBM
y to come up with a patch, but I'd appreciate some guidance what
> direction it should take.
--
Alan Modra
Australia Development Lab, IBM
Ping?
https://lists.gnu.org/archive/html/bug-gnulib/2013-12/msg00107.html
--
Alan Modra
Australia Development Lab, IBM
a) Delete nonsense about "not polluting the namespace with stddef.h
symbols" since string.h includes stddef.h a little later anyway.
b) Don't roll our own slow memcpy in _obstack_newchunk.
c) Rename obstack_free to _obstack_free. This makes the naming consistent
with other obstack functions and
ready takes signigicant time in
libiberty. I don't want to double that time.
Alan Modra (5):
obstack tidy
64-bit obstack support, part 1
64-bit obstack support, part 2
64-bit obstack support, part 3
obstack usability
lib/obstack.
And >2G on 32-bit.
* include/gnu-versions.h (_GNU_OBSTACK_INTERFACE_VERSION): Bump.
* include/obstack.h: Add _obstack2_newchunk hidden proto.
* malloc/obstack.h: Import from gnulib.
* malloc/obstack.c: Likewise.
* malloc/obstackv1.c: New file.
* mall
a) Correct calls to alloc function, to use a size_t arg. "long" is
just wrong on targets like x86_64-mingw64 where "long" is 32 bits
and "size_t" 64 bits.
b) Consolidate _obstack_begin and _obstack_begin1 code.
* lib/obstack.h (struct obstack ): Correct prototype to
use "siz
This finally enables full 64-bit obstack support. Renaming the
_obstack_* functions to _obstack2_* is necessary to prevent a shared
library using version 1 obstack functions from having these functions
overridden by version 2 functions in the main executable. Not all
shared libraries are created
This gets us 4G obstack support, without changing ABI compatibility.
a) Replace "int" size parameters, return values, and macro local vars
with __OBSTACK_SIZE_T, an "unsigned int" for now.
b) Make obstack.chunk_size a __CHUNK_SIZE_T, an "unsigned long" for now.
c) Make all obstack macros checki
When using obstack.h/c in other projects it isn't very nice that gnulib
headers like exitfail.h and gettext.h need to be present. Far worse
is a dependency on gnulib's version of stdlib.h for __attribute_pure__
and _Noreturn. This only works if a project is willing to import
rather a lot of gnuli
On Sat, Jul 26, 2014 at 09:20:36AM +0200, Andreas Schwab wrote:
> Alan Modra writes:
>
> > +# elif defined _MSC_VER && 1200 <= _MSC_VER
>
> Please write that as "_MSC_VER >= 1200".
Thanks for looking over the patch. This line was taken out of
gnul
On Sat, Jul 26, 2014 at 07:48:21AM -0400, Paul Eggert wrote:
> On 07/26/2014 01:57 AM, Alan Modra wrote:
> >When using obstack.h/c in other projects it isn't very nice that gnulib
> >headers like exitfail.h and gettext.h need to be present. Far worse
> >is a depende
f you also want
> ptrdiff_t for use in obstack.h, with GCC's stddef.h you can do:
> #define __need_ptrdiff_t
> #include
>
> That said, I don't think anybody actually cares about a GNU extension
> header like obstack.h implicitly defining common stan
o be
negotiated before a patch can be committed.. It's not just gnulib
and glibc, but eg. to update the ancient obstack.h in libiberty, gcc,
gdb and gas all needed fixes to remove improper use of obstack
internals.
https://sourceware.org/bugzilla/show_bug.cgi?id=14483
https://bugzilla
On Sat, Jul 26, 2014 at 03:27:07PM +0930, Alan Modra wrote:
> This patch series gives obstacks some much needed TLC. The first patch
Ping. It's been over a week, and I haven't seen any comments on the
bulk of the patch series, except for Roland's comment about including
stddef
On Mon, Aug 04, 2014 at 02:16:06PM +0930, Alan Modra wrote:
> On Sat, Jul 26, 2014 at 03:27:07PM +0930, Alan Modra wrote:
> > This patch series gives obstacks some much needed TLC. The first patch
http://lists.gnu.org/archive/html/bug-gnulib/2014-07/msg00106.html
> Ping. It's
On Fri, Aug 29, 2014 at 09:57:49AM +0200, Thien-Thi Nguyen wrote:
> () Alan Modra
> () Fri, 29 Aug 2014 11:15:57 +0930
>
>On Mon, Aug 04, 2014 at 02:16:06PM +0930, Alan Modra wrote:
>> On Sat, Jul 26, 2014 at 03:27:07PM +0930, Alan Modra wrote:
>> > This
har *) (h)->chunk))\
? ((h)->next_free = (h)->chunk_limit) : 0), \
(h)->object_base = (h)->next_free,\
- (h)->temp.tempptr)
+ (h)->temp.p)
# define obstack_free(h, obj)\
- ((h)->temp.tempint = (char *) (obj) - (char *) (h)->chunk, \
- h)->temp.tempint > 0 \
- && (h)->temp.tempint < (h)->chunk_limit - (char *) (h)->chunk))\
-? (void) ((h)->next_free = (h)->object_base
\
- = (h)->temp.tempint + (char *) (h)->chunk) \
-: (__obstack_free) (h, (h)->temp.tempint + (char *) (h)->chunk)))
+ ((h)->temp.p = (void *) (obj), \
+ (((h)->temp.p > (void *) (h)->chunk \
+ && (h)->temp.p < (void *) (h)->chunk_limit) \
+? (void) ((h)->next_free = (h)->object_base = (char *) (h)->temp.p) \
+: _obstack_free ((h), (h)->temp.p)))
#endif /* not __GNUC__ */
@@ -512,4 +513,4 @@ extern int obstack_exit_failure;
} /* C++ */
#endif
-#endif /* obstack.h */
+#endif /* _OBSTACK_H */
--
Alan Modra
Australia Development Lab, IBM
nt/misc/../sysdeps/unix/syscall-template.S:81: multiple
definition of `vm86@GLIBC_2.21'
/home/alan/build/glibc32/libc_pic.a(oldgetrlimit.os): In function
`__old_getrlimit':
/src/glibc-current/misc/../sysdeps/unix/syscall-template.S:81: multiple
definition of `getrlimit@GLIBC_2.21'
ecially for 64-bit hosts. */
+# define _OBSTACK_SIZE_T size_t
+# define _CHUNK_SIZE_T size_t
+#endif
/* If B is the base of an object addressed by P, return the result of
aligning P to the next multiple of A + 1. B and P must be of type
@@ -126,8 +139,6 @@
__BPTR_ALIGN (sizeof (ptrdiff_t) < sizeof (void *) ? (B) : (char *) 0, \
P, A)
-#include
-
#ifndef __attribute_pure__
# define __attribute_pure__ _GL_ATTRIBUTE_PURE
#endif
--
Alan Modra
Australia Development Lab, IBM
mp.i < (h)->chunk_limit - (char *) (h)->chunk))
\
? (void) ((h)->next_free = (h)->object_base
\
= (h)->temp.i + (char *) (h)->chunk) \
- : (__obstack_free) (h, (h)->temp.i + (char *) (h)->chunk)))
+: _obstack_free (h, (h)->temp.i + (char *) (h)->chunk)))
#endif /* not __GNUC__ */
@@ -512,4 +497,4 @@ extern int obstack_exit_failure;
} /* C++ */
#endif
-#endif /* obstack.h */
+#endif /* _OBSTACK_H */
--
Alan Modra
Australia Development Lab, IBM
uct _obstack_chunk *(*)(void *, long))(newchunkfun))
+ ((h)->chunkfun = (struct _obstack_chunk *(*)(void *, size_t))(newchunkfun))
#define obstack_freefun(h, newfreefun) \
((h)->freefun = (void (*)(void *, struct _obstack_chunk *))(newfreefun))
--
Alan Modra
Australia Development Lab, IBM
\
- && (h)->temp.i < (h)->chunk_limit - (char *) (h)->chunk))
\
-? (void) ((h)->next_free = (h)->object_base
\
- = (h)->temp.i + (char *) (h)->chunk) \
-: _obstack_free (h, (h)->temp.i + (char *) (h)->chunk)))
+ ((h)->temp.p = (void *) (obj), \
+ (((h)->temp.p > (void *) (h)->chunk \
+ && (h)->temp.p < (void *) (h)->chunk_limit) \
+? (void) ((h)->next_free = (h)->object_base = (char *) (h)->temp.p) \
+: _obstack_free ((h), (h)->temp.p)))
#endif /* not __GNUC__ */
--
Alan Modra
Australia Development Lab, IBM
gt;chunk))\
+ ((h)->temp.i = (char *) (obj) - (char *) (h)->chunk, \
+ h)->temp.i > 0\
+ && (h)->temp.i < (h)->chunk_limit - (char *) (h)->chunk))
\
? (void) ((h)->next_free = (h)->object_base
\
- = (h)->temp.tempint + (char *) (h)->chunk) \
-: (__obstack_free) (h, (h)->temp.tempint + (char *) (h)->chunk)))
+ = (h)->temp.i + (char *) (h)->chunk) \
+: (__obstack_free) (h, (h)->temp.i + (char *) (h)->chunk)))
#endif /* not __GNUC__ */
--
Alan Modra
Australia Development Lab, IBM
e argument to make
the current object smaller. Just don't try to shrink it beyond zero
length---there's no telling what will happen if you do that.
It is no longer possible to shrink an obstack with obstack_blank (but
you can still do that with obstack_blank_fast).
--
Alan Modra
Australia Development Lab, IBM
On Wed, Oct 29, 2014 at 06:34:02PM +, Joseph S. Myers wrote:
> On Wed, 29 Oct 2014, Alan Modra wrote:
>
> > And >2G on 32-bit.
> >
>
> [BZ #14483]
>
> in the ChangeLog entry.
Thanks, added. Since the stdalign.h change won't do for glibc, I&#
On Wed, Oct 29, 2014 at 08:35:18PM -0700, Paul Eggert wrote:
> Alan Modra wrote:
> >It is no longer possible to shrink an obstack with obstack_blank (but
> >you can still do that with obstack_blank_fast).
>
> Ouch, I hadn't noticed that. That's an incompatible chan
attached.
--
Alan Modra
Australia Development Lab, IBM
>From fb249df2535a4fcd913589e27cd3c30a60cd5308 Mon Sep 17 00:00:00 2001
From: Alan Modra
Date: Fri, 31 Oct 2014 10:54:55 +1030
Subject: [PATCH 1/2] 64-bit obstack support, part 1a
a) Correct calls to alloc function, to use a size_t arg
ed F
+ _obstack_newchunk F
I also looked at exported obstack symbols for various other
non-standard builds, and checked that __GI__obstack_newchunk matched
the proper _obstack_newchunk (I had it wrong at one stage).
--
Alan Modra
Australia Development Lab, IBM
to me the easiest solution is to always include obstack.o in
libgnu.a. If the system libc is up to date, then obstack.o will be
empty.
And another small tweak: Don't use alignof.h if __IBM__ALIGNOF__ is
defined (alignof.h just uses __alignof__ in that case), or when
__alignof__ is define
id) ((h)->chunkfun.extra = _OBSTACK_CAST (void *(*) (void *, size_t), \
newchunkfun)))
and in obstack_freefun. I think you need to keep the casts, otherwise
you can only specify new extra arg alloc and free functions.
--
Alan Modra
Australia Development Lab, IBM
ould
- kill >2G obstacks on 32-bit targets,
- lose the nice symmetry with other obstack functions,
and obstack_blank_fast is the right interface to use for shrinking.
--
Alan Modra
Australia Development Lab, IBM
!new_chunk)
(*obstack_alloc_failed_handler)();
h->chunk = new_chunk;
--
Alan Modra
Australia Development Lab, IBM
eption?
Not really. binutils, gdb, and gcc are all built with -Wc++compat
nowadays so I have applied this patch to the obstack source in
libiberty, to pacify the C++ fanatics.
I'm not concerned that the sources diverge, given the low frequency of
updates to libiberty's obstack code in the pa
34 matches
Mail list logo
