Hi,
Roland McGrath, le Mon 13 Nov 2006 16:02:00 -0800, a écrit :
> I meant some existing header if there is an appropriate one.
I can't find any that would suit well.
The functions at stake are initialization functions, idle functions,
and halt functions. Would you prefer kernel/init.h, kernel/
OK, for kicks I tried to build glibc with --with-tls and --without-__thread
using my gnumach with the gdt stuff I did a couple weeks ago and Jeroens
patch from here:
http://www.dekkers.cx/hurd/glibc-tls.patch
Here is what I got:
CPP='gcc-4.0 -E -x c-header'
/devel2/bdefreese/glibc_11142006/
> I'm not inventing anything. Those functions are defined in model_dep.c...
You are indeed inventing model_dep.h, a header file name that did not
exist before. Where a global function is defined is a private matter.
Where it is declared for its users to #include is a matter of reasonable
interfac
Roland McGrath, le Mon 13 Nov 2006 16:02:00 -0800, a écrit :
> > Roland McGrath, le Mon 13 Nov 2006 13:44:30 -0800, a écrit :
> > > > I don't disagree. Have a suggestion for a better place to declare it?
> > >
> > > Some machine-independent header included by the relevant code.
> >
> > kern/mode
> In other *nixes it appears to be declared in sys/ddi.h. Does that make more
> sense?
That's Solarisism, not an appropriate model.
___
Bug-hurd mailing list
Bug-hurd@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-hurd
Neill Miller, le Tue 14 Nov 2006 11:29:32 -0600, a écrit :
> One more time, this time adding the -ffreestanding option after the
> -fstack-protector option:
>
> movl%gs:20, %eax
So gcc is still buggy.
Samuel
___
Bug-hurd mailing list
Bug
On Tue, 14 Nov 2006 11:08:02 -0600
Neill Miller <[EMAIL PROTECTED]> wrote:
One more time, this time adding the -ffreestanding option after the
-fstack-protector option:
> > $ echo 'void f (void) { volatile char a[8]; a[3]; }' | gcc -S -x c -O2
> > -fstack-protector - -o -; uname -a; /lib/libc-*
On Tue, 14 Nov 2006 17:57:50 +0100
Thomas Schwinge <[EMAIL PROTECTED]> wrote:
Oops, that was on PPC. Let me try this again ;-)
Note that while it's an AMD 64 chip, it's running in full 32-bit mode with no
64 bit anything. Hope that helps.
> #v+
> $ echo 'void f (void) { volatile char a[8]; a[
Neill Miller, le Tue 14 Nov 2006 11:08:02 -0600, a écrit :
> [EMAIL PROTECTED] ~ $ echo 'void f (void) { volatile char a[8]; a[3]; }' |
> gcc -S -x c -O2 -fstack-protector - -o -; uname -a; /lib/libc-*.so
> lwz 0,-28680(2)
That's the TLS STACK_CHK_GUARD instruction. Could you retry with
On Tue, 14 Nov 2006 17:57:50 +0100
Thomas Schwinge <[EMAIL PROTECTED]> wrote:
Hello,
Hope this helps!
> ... because this is a Debian testing system and they're still using a
> 2.3-based glibc. That's probably it. If someone has easy access to a
> system with a really recent GCC 4.1 or GCC 4.2
Update of bug #17136 (project hurd):
Status:None => Invalid
Assigned to:None => tschwinge
Open/Closed:Open => Closed
Hello!
On Tue, Nov 14, 2006 at 04:26:26PM +0100, Samuel Thibault wrote:
> Thomas Schwinge, le Tue 14 Nov 2006 16:15:08 +0100, a ?crit :
> > Using `-fstack-protector' with GCC 4.1 made it include assembler code
> > using ``%gs:0x14'' even with `-ffreestanding'. However, this isn't the
> > correct
Hello!
On Tue, Nov 14, 2006 at 05:38:21PM +0100, I wrote:
> On Tue, Nov 14, 2006 at 04:26:26PM +0100, Samuel Thibault wrote:
> > Thomas Schwinge, le Tue 14 Nov 2006 16:15:08 +0100, a ?crit :
> > > Using `-fstack-protector' with GCC 4.1 made it include assembler code
> > > using ``%gs:0x14'' even w
Thomas Schwinge, le Tue 14 Nov 2006 16:15:08 +0100, a écrit :
> Using `-fstack-protector' with GCC 4.1 made it include assembler code
> using ``%gs:0x14'' even with `-ffreestanding'. However, this isn't the
> correct thing to do in kernel
> space (with `-ffreestanding'). I think I've now tracked
[Also sent to Roland and to bug-hurd.]
Hello!
Using `-fstack-protector' with GCC 4.1 made it include assembler code
using ``%gs:0x14'' even with `-ffreestanding'. However, this isn't the
correct thing to do on a) GNU/Hurd user space and neither b) in kernel
space (with `-ffreestanding'). I thi
- Original Message -
From: "Roland McGrath" <[EMAIL PROTECTED]>
To: "Samuel Thibault" <[EMAIL PROTECTED]>
Cc: "Barry deFreese" <[EMAIL PROTECTED]>;
Sent: Monday, November 13, 2006 7:02 PM
Subject: Re: Gnumach cleanup Round 4
Roland McGrath, le Mon 13 Nov 2006 13:44:30 -0800, a écrit
Hi,
Barry deFreese, le Tue 14 Nov 2006 03:59:38 -0500, a écrit :
> I realize that this reverts a couple of the things I sent on my last
> patch but several of the lpr functions really return nothing so should
> have been void in the first place.
Not necessarily.
> Index: i386/i386/trap.c
> =
17 matches
Mail list logo
