[parisc-linux] glibc-2.3.3: a warning with gcc-3.3 become an error with gcc-3.5

All of lore.kernel.org
 help / color / mirror / Atom feed

* [parisc-linux] glibc-2.3.3: a warning with gcc-3.3 become an error with gcc-3.5
@ 2004-03-01 18:43 Joel Soete
  2004-03-01 19:38 ` Carlos O'Donell
  0 siblings, 1 reply; 28+ messages in thread
From: Joel Soete @ 2004-03-01 18:43 UTC (permalink / raw)
  To: parisc-linux; +Cc: Carlos O'Donell

Hi all,

As the title mail said, the following warning with gcc-3.3:
../sysdeps/hppa/fpu/feupdateenv.c: In function `feupdateenv':
../sysdeps/hppa/fpu/feupdateenv.c:31: warning: assignment of read-only member
`__status_word'

become an error with gcc-3.5:
../sysdeps/hppa/fpu/feupdateenv.c: In function `feupdateenv':
../sysdeps/hppa/fpu/feupdateenv.c:31: error: assignment of read-only member
`__status_word'
make[2]: *** [/Develop/parisc-linux/build/glibc/math/feupdateenv.o] Error
1
make[2]: Leaving directory `/Develop/parisc-linux/sources/glibc-2.3.3-20040301/math'
make[1]: *** [math/others] Error 2
make[1]: Leaving directory `/Develop/parisc-linux/sources/glibc-2.3.3-20040301'
make: *** [all] Error 2

And regarding the definition of feupdateenv():
int
feupdateenv (const fenv_t *envp)
[snip]

I couldn't blame gcc.

So I would suggest to use sysdeps/mips/fpu/feupdateenv.c as model:
int
feupdateenv (const fenv_t *envp)
{
  int temp;

  /* Save current exceptions.  */
  _FPU_GETCW (temp);
  temp &= FE_ALL_EXCEPT;

  /* Install new environment.  */
  fesetenv (envp);

  /* Raise the safed exception.  Incidently for us the implementation
     defined format of the values in objects of type fexcept_t is the
     same as the ones specified using the FE_* constants.  */
  feraiseexcept (temp);

  /* Success.  */
  return 0;
}

but replacing:
 * 'int temp;' by 'fenv_t *temp;'
 * '_FPU_GETCW (temp);' by 'fegetenv(temp);' (as we don't have _FPU_GETCW()
)

But not sure it's the best solution.

Thanks in advance for all advise,
    Joel


----------------------------------------------------------------------------------------
Tiscali ADSL: 19,50 €/mois, pendant 3 mois! L'Internet rapide, c'est pour
tout le monde.
http://reg.tiscali.be/default.asp?lg=fr

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [parisc-linux] glibc-2.3.3: a warning with gcc-3.3 become an error with gcc-3.5
  2004-03-01 18:43 [parisc-linux] glibc-2.3.3: a warning with gcc-3.3 become an error with gcc-3.5 Joel Soete
@ 2004-03-01 19:38 ` Carlos O'Donell
  2004-03-02 14:59   ` Joel Soete
  0 siblings, 1 reply; 28+ messages in thread
From: Carlos O'Donell @ 2004-03-01 19:38 UTC (permalink / raw)
  To: Joel Soete; +Cc: parisc-linux

On Mon, Mar 01, 2004 at 07:43:01PM +0100, Joel Soete wrote:
> But not sure it's the best solution.
> Thanks in advance for all advise,

Show me a patch.

c.

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [parisc-linux] glibc-2.3.3: a warning with gcc-3.3 become an error with gcc-3.5
  2004-03-01 19:38 ` Carlos O'Donell
@ 2004-03-02 14:59   ` Joel Soete
  2004-03-02 15:27     ` Matthew Wilcox
  2004-03-02 19:50     ` [parisc-linux] glibc-2.3.3: a warning with gcc-3.3 become an error with gcc-3.5 Carlos O'Donell
  0 siblings, 2 replies; 28+ messages in thread
From: Joel Soete @ 2004-03-02 14:59 UTC (permalink / raw)
  To: Carlos O'Donell; +Cc: parisc-linux

>
>Show me a patch.
>
something like this perhaps?

--- ./feupdateenv.c.orig	2004-03-02 15:21:25.959085856 +0100
+++ ./feupdateenv.c	2004-03-02 07:53:27.938166120 +0100
@@ -28,11 +28,16 @@
   /* Get the current exception status. */
   __asm__ ("fstd %%fr0,0(%1)" : "=m" (*sw) : "r" (sw));
   sw[0] &= FE_ALL_EXCEPT;
-  envp->__status_word = envp->__status_word | sw[0];
+/*  envp->__status_word = envp->__status_word | sw[0]; */
   
   /* Install new environment.  */
   fesetenv (envp);
 
+  /* Raise the safed exception.  Incidently for us the implementation
+     defined format of the values in objects of type fexcept_t is the
+     same as the ones specified using the FE_* constants.  */
+  feraiseexcept (sw[0]);
+
   /* Success.  */
   return 0;
 }

Thanks in advance for attention,
    Joel


----------------------------------------------------------------------------------------
Tiscali ADSL: 19,50 €/mois, pendant 3 mois! L'Internet rapide, c'est pour
tout le monde.
http://reg.tiscali.be/default.asp?lg=fr

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [parisc-linux] glibc-2.3.3: a warning with gcc-3.3 become an error with gcc-3.5
  2004-03-02 14:59   ` Joel Soete
@ 2004-03-02 15:27     ` Matthew Wilcox
  2004-04-02 18:03       ` [parisc-linux] [PATCH] glibc-2.3.3 error with debian unstable compiler Carlos O'Donell
  2004-03-02 19:50     ` [parisc-linux] glibc-2.3.3: a warning with gcc-3.3 become an error with gcc-3.5 Carlos O'Donell
  1 sibling, 1 reply; 28+ messages in thread
From: Matthew Wilcox @ 2004-03-02 15:27 UTC (permalink / raw)
  To: Joel Soete; +Cc: Carlos O'Donell, parisc-linux

On Tue, Mar 02, 2004 at 03:59:34PM +0100, Joel Soete wrote:
> +  /* Raise the safed exception.  Incidently for us the implementation

"saved exception" or "safe exception"?

-- 
"Next the statesmen will invent cheap lies, putting the blame upon 
the nation that is attacked, and every man will be glad of those
conscience-soothing falsities, and will diligently study them, and refuse
to examine any refutations of them; and thus he will by and by convince 
himself that the war is just, and will thank God for the better sleep 
he enjoys after this process of grotesque self-deception." -- Mark Twain

^ permalink raw reply	[flat|nested] 28+ messages in thread

* [parisc-linux] [PATCH] glibc-2.3.3 error with debian unstable compiler...
  2004-03-02 15:27     ` Matthew Wilcox
@ 2004-04-02 18:03       ` Carlos O'Donell
  2004-04-06 10:46         ` [parisc-linux] " Joel Soete
  0 siblings, 1 reply; 28+ messages in thread
From: Carlos O'Donell @ 2004-04-02 18:03 UTC (permalink / raw)
  To: Matthew Wilcox; +Cc: parisc-linux

On Tue, Mar 02, 2004 at 03:27:44PM +0000, Matthew Wilcox wrote:
> On Tue, Mar 02, 2004 at 03:59:34PM +0100, Joel Soete wrote:
> > +  /* Raise the safed exception.  Incidently for us the implementation
> 
> "saved exception" or "safe exception"?

That's a typo that *every* other arch has because they copied the code
from i386, I think rth was the only one anal enough to fix the spelling
mistake ;)

I snatched the C99 spec and fixed our implementation based on my reading
of the current code, other arch code, and the spec. The correct
feupdateenv is provided below.

I'm sending this upstream to debian, so we can build again.

c.

Index: feupdateenv.c
===================================================================
RCS file: /cvs/glibc/libc/sysdeps/hppa/fpu/feupdateenv.c,v
retrieving revision 1.3
diff -u -p -r1.3 feupdateenv.c
--- feupdateenv.c	18 Dec 2003 03:57:50 -0000	1.3
+++ feupdateenv.c	2 Apr 2004 06:56:56 -0000
@@ -27,11 +27,10 @@ feupdateenv (const fenv_t *envp)
 
   /* Get the current exception status. */
   __asm__ ("fstd %%fr0,0(%1)" : "=m" (*sw) : "r" (sw));
-  sw[0] &= FE_ALL_EXCEPT;
-  envp->__status_word = envp->__status_word | sw[0];
-  
   /* Install new environment.  */
   fesetenv (envp);
+  /* Raise the saved exceptions */
+  feraiseexcept(sw[0] & FE_ALL_EXCEPT);
 
   /* Success.  */
   return 0;

^ permalink raw reply	[flat|nested] 28+ messages in thread

* [parisc-linux] RE: [PATCH] glibc-2.3.3 error with debian unstable compiler...
  2004-04-02 18:03       ` [parisc-linux] [PATCH] glibc-2.3.3 error with debian unstable compiler Carlos O'Donell
@ 2004-04-06 10:46         ` Joel Soete
  2004-04-06 13:14           ` Carlos O'Donell
  2004-04-20 12:22           ` Joel Soete
  0 siblings, 2 replies; 28+ messages in thread
From: Joel Soete @ 2004-04-06 10:46 UTC (permalink / raw)
  To: Carlos O'Donell, Matthew Wilcox; +Cc: parisc-linux

Carlos,

I just apply your patch against the glibc cvs dated 20040405 (yesterday for
me).
It compiles well with last updated gcc (3.3.3-6 dpkg iirc) :)

I just complete the 'make -k check' (on chroot disk) which finished with
nothing more then well known bug :)

... And the compile (always in the chroot disk) of the last k-2.6.5-pa3 just
finishes with success :)

hth,
    Joel

----------------------------------------------------------------------------------------
Tiscali ADSL: 35 €/mois, la meilleure offre du marché!
http://reg.tiscali.be/default.asp?lg=fr

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [parisc-linux] RE: [PATCH] glibc-2.3.3 error with debian unstable compiler...
  2004-04-06 10:46         ` [parisc-linux] " Joel Soete
@ 2004-04-06 13:14           ` Carlos O'Donell
  2004-04-20 12:22           ` Joel Soete
  1 sibling, 0 replies; 28+ messages in thread
From: Carlos O'Donell @ 2004-04-06 13:14 UTC (permalink / raw)
  To: Joel Soete; +Cc: Matthew Wilcox, parisc-linux

On Tue, Apr 06, 2004 at 12:46:54PM +0200, Joel Soete wrote:
> Carlos,
> 
> I just apply your patch against the glibc cvs dated 20040405
> (yesterday for me).  It compiles well with last updated gcc (3.3.3-6
> dpkg iirc) :)
> 
> I just complete the 'make -k check' (on chroot disk) which finished
> with nothing more then well known bug :)
> 
> ... And the compile (always in the chroot disk) of the last
> k-2.6.5-pa3 just finishes with success :)


Thanks for the check Joel!

c.

^ permalink raw reply	[flat|nested] 28+ messages in thread

* RE: [parisc-linux] RE: [PATCH] glibc-2.3.3 error with debian unstable compiler...
  2004-04-06 10:46         ` [parisc-linux] " Joel Soete
  2004-04-06 13:14           ` Carlos O'Donell
@ 2004-04-20 12:22           ` Joel Soete
  2004-04-20 19:35             ` Carlos O'Donell
  1 sibling, 1 reply; 28+ messages in thread
From: Joel Soete @ 2004-04-20 12:22 UTC (permalink / raw)
  To: Carlos O'Donell; +Cc: parisc-linux

Carlos,

Last week I encounter some pb to co last cvs, sorry.

This week and can apply your patch against the glibc cvs dated 20040419 (just
some hunk failled because in the uptream: great).

I just complete the 'make -k check' (on chroot disk) which finished with
just one more error then well known ones:
make[2]: *** [/Develop/parisc-linux/build/glibc/stdio-common/tst-rndseek.out]
Error 1
make[1]: *** [stdio-common/tests] Error 2

... And the compile (always in the chroot disk) of the last k-2.6.6-rc1-pa0
works still fine and even boot well on a b180.

hth,
    Joel



----------------------------------------------------------------------------------------
Tiscali ADSL: 35 €/mois, la meilleure offre du marché!
http://reg.tiscali.be/default.asp?lg=fr

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [parisc-linux] RE: [PATCH] glibc-2.3.3 error with debian unstable compiler...
  2004-04-20 12:22           ` Joel Soete
@ 2004-04-20 19:35             ` Carlos O'Donell
  2004-04-21  6:15               ` Joel Soete
  0 siblings, 1 reply; 28+ messages in thread
From: Carlos O'Donell @ 2004-04-20 19:35 UTC (permalink / raw)
  To: Joel Soete; +Cc: parisc-linux

On Tue, Apr 20, 2004 at 02:22:30PM +0200, Joel Soete wrote:
> Carlos,
> 
> Last week I encounter some pb to co last cvs, sorry.
> 
> This week and can apply your patch against the glibc cvs dated 20040419 (just
> some hunk failled because in the uptream: great).
> 
> I just complete the 'make -k check' (on chroot disk) which finished with
> just one more error then well known ones:
> make[2]: *** [/Develop/parisc-linux/build/glibc/stdio-common/tst-rndseek.out]
> Error 1
> make[1]: *** [stdio-common/tests] Error 2
> 
> ... And the compile (always in the chroot disk) of the last k-2.6.6-rc1-pa0
> works still fine and even boot well on a b180.

I'm seeing an extra error with realtime signal delivery tst-timer4, but
not this error. I'm working on fixing tst-timer4, can you find out why
tst-rndseek fails?

What does tst-rndseek.out say? I don't see this error?

Cheers,
Carlos.

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [parisc-linux] RE: [PATCH] glibc-2.3.3 error with debian unstable compiler...
  2004-04-20 19:35             ` Carlos O'Donell
@ 2004-04-21  6:15               ` Joel Soete
  2004-04-21 11:55                 ` Carlos O'Donell
  0 siblings, 1 reply; 28+ messages in thread
From: Joel Soete @ 2004-04-21  6:15 UTC (permalink / raw)
  To: Carlos O'Donell; +Cc: parisc-linux

 
>> I just complete the 'make -k check' (on chroot disk) which finished with
>> just one more error then well known ones:
>> make[2]: *** [/Develop/parisc-linux/build/glibc/stdio-common/tst-rndseek.out]
>> Error 1
>> make[1]: *** [stdio-common/tests] Error 2
>> 
>> ... And the compile (always in the chroot disk) of the last k-2.6.6-rc1-pa0
>> works still fine and even boot well on a b180.
>
>I'm seeing an extra error with realtime signal delivery tst-timer4, but
>not this error. I'm working on fixing tst-timer4,
:) and I don't see your's (not exactly the same co may be)

# grep Error glibc-2.3.3-20040419_Make-check 
make[2]: *** [/Develop/parisc-linux/build/glibc/math/test-float.out] Error
1
make[2]: *** [/Develop/parisc-linux/build/glibc/math/test-double.out] Error
1
make[2]: *** [/Develop/parisc-linux/build/glibc/math/test-ifloat.out] Error
1
make[1]: *** [math/tests] Error 2
make[2]: *** [/Develop/parisc-linux/build/glibc/stdlib/tst-strtod.out] Error
1
make[2]: *** [/Develop/parisc-linux/build/glibc/stdlib/bug-strtod.out] Error
1
make[1]: *** [stdlib/tests] Error 2
make[2]: *** [/Develop/parisc-linux/build/glibc/stdio-common/tst-rndseek.out]
Error 1
make[1]: *** [stdio-common/tests] Error 2
make[2]: [/Develop/parisc-linux/build/glibc/posix/annexc.out] Error 1 (ignored)
make[2]: *** [/Develop/parisc-linux/build/glibc/elf/tst-tls13.out] Error
1
make[1]: *** [elf/tests] Error 2
make: *** [check] Error 2

> can you find out why tst-rndseek fails?
>
>What does tst-rndseek.out say? I don't see this error?
grrr just a time out (I would have to put the 'export TIMEOUTFACTOR=4' in
the root profile on my chroot disk :? )

GCONV_PATH=/Develop/parisc-linux/build/glibc/iconvdata LC_ALL=C   /Develop/parisc-linux/build/glibc/elf/ld.so.1
--library-path /Develop/parisc-linux/build/glibc:/Develop/parisc-linux/build/glibc/math:/Develop/parisc-linux/build/glibc/elf:/Develop/parisc-linux/build/glibc/dlfcn:/Develop/parisc-linux/build/glibc/nss:/Develop/parisc-linux/build/glibc/nis:/Develop/p
arisc-linux/build/glibc/rt:/Develop/parisc-linux/build/glibc/resolv:/Develop/parisc-linux/build/glibc/crypt:/Develop/parisc-linux/build/glibc/linuxthreads
/Develop/parisc-linux/build/glibc/stdio-common/tst-rndseek  > /Develop/parisc-linux/build/glibc/stdio-common/tst-rndseek.out
Timed out: killed the child process

hth,
    Joel


----------------------------------------------------------------------------------------
Tiscali ADSL: 35 €/mois, la meilleure offre du marché!
http://reg.tiscali.be/default.asp?lg=fr

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [parisc-linux] RE: [PATCH] glibc-2.3.3 error with debian unstable compiler...
  2004-04-21  6:15               ` Joel Soete
@ 2004-04-21 11:55                 ` Carlos O'Donell
  2004-04-27  7:04                   ` Joel Soete
  0 siblings, 1 reply; 28+ messages in thread
From: Carlos O'Donell @ 2004-04-21 11:55 UTC (permalink / raw)
  To: Joel Soete; +Cc: parisc-linux

> >What does tst-rndseek.out say? I don't see this error?
> grrr just a time out (I would have to put the 'export TIMEOUTFACTOR=4' in
> the root profile on my chroot disk :? )
> 
> GCONV_PATH=/Develop/parisc-linux/build/glibc/iconvdata LC_ALL=C   /Develop/parisc-linux/build/glibc/elf/ld.so.1
> --library-path /Develop/parisc-linux/build/glibc:/Develop/parisc-linux/build/glibc/math:/Develop/parisc-linux/build/glibc/elf:/Develop/parisc-linux/build/glibc/dlfcn:/Develop/parisc-linux/build/glibc/nss:/Develop/parisc-linux/build/glibc/nis:/Develop/p
> arisc-linux/build/glibc/rt:/Develop/parisc-linux/build/glibc/resolv:/Develop/parisc-linux/build/glibc/crypt:/Develop/parisc-linux/build/glibc/linuxthreads
> /Develop/parisc-linux/build/glibc/stdio-common/tst-rndseek  > /Develop/parisc-linux/build/glibc/stdio-common/tst-rndseek.out
> Timed out: killed the child process

Yes, the default timeout is too short, add a TIMEOUTFACTOR of 5 or more.

c.

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [parisc-linux] RE: [PATCH] glibc-2.3.3 error with debian unstable compiler...
  2004-04-21 11:55                 ` Carlos O'Donell
@ 2004-04-27  7:04                   ` Joel Soete
  2004-04-27 14:57                     ` Carlos O'Donell
  0 siblings, 1 reply; 28+ messages in thread
From: Joel Soete @ 2004-04-27  7:04 UTC (permalink / raw)
  To: Carlos O'Donell; +Cc: parisc-linux


Hello Carlos,

Here is my weekly test: cvs 20040426.
Nothing new to mention: always well known errors collected after make -k
check. I just build a 2.6.6-_rc2_-pa2 (for the fun) without pb (which boot
fine on my b180)

Cheers,
    Joel

PS: 
> Yes, the default timeout is too short, add a TIMEOUTFACTOR of 5 or more.

This time I put TIMEOUTFACTOR=4 in the chroot root profile and seems to be
enough to bypass timeout errors :)


----------------------------------------------------------------------------------------
Tiscali ADSL: 35 €/mois, la meilleure offre du marché!
http://reg.tiscali.be/default.asp?lg=fr

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [parisc-linux] RE: [PATCH] glibc-2.3.3 error with debian unstable compiler...
  2004-04-27  7:04                   ` Joel Soete
@ 2004-04-27 14:57                     ` Carlos O'Donell
  2004-04-27 17:06                       ` Joel Soete
  0 siblings, 1 reply; 28+ messages in thread
From: Carlos O'Donell @ 2004-04-27 14:57 UTC (permalink / raw)
  To: Joel Soete; +Cc: parisc-linux

> Here is my weekly test: cvs 20040426.
> Nothing new to mention: always well known errors collected after make -k
> check. I just build a 2.6.6-_rc2_-pa2 (for the fun) without pb (which boot
> fine on my b180)

You aren't seeing a failure in tst-timer4?
Can you verify this by running tst-timer4 by hand and sending the output
ot the list?

c.

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [parisc-linux] RE: [PATCH] glibc-2.3.3 error with debian unstable compiler...
  2004-04-27 14:57                     ` Carlos O'Donell
@ 2004-04-27 17:06                       ` Joel Soete
  2004-04-27 17:15                         ` Carlos O'Donell
  0 siblings, 1 reply; 28+ messages in thread
From: Joel Soete @ 2004-04-27 17:06 UTC (permalink / raw)
  To: Carlos O'Donell; +Cc: parisc-linux


> You aren't seeing a failure in tst-timer4?

uhm ... No (definitely :^)

> Can you verify this by running tst-timer4 by hand and sending the output
> ot the list?
Yes naturally, here there are:

# ./tst-timer4 
clock_gettime returned timespec = { 1083087449, 554351000 }
clock_getres returned timespec = { 0, 10000000 }
# echo $?
0

Well duno if results are right (first line different on each test but the
second is constant with or without TIMEOUTFACTOR)?

Don't hesitate with more questions ;)

Joel

----------------------------------------------------------------------------------------
Tiscali ADSL: 35 €/mois, la meilleure offre du marché!
http://reg.tiscali.be/default.asp?lg=fr

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [parisc-linux] RE: [PATCH] glibc-2.3.3 error with debian unstable compiler...
  2004-04-27 17:06                       ` Joel Soete
@ 2004-04-27 17:15                         ` Carlos O'Donell
  2004-04-28  5:39                           ` Joel Soete
  0 siblings, 1 reply; 28+ messages in thread
From: Carlos O'Donell @ 2004-04-27 17:15 UTC (permalink / raw)
  To: Joel Soete; +Cc: parisc-linux

On Tue, Apr 27, 2004 at 07:06:06PM +0200, Joel Soete wrote:
> > Can you verify this by running tst-timer4 by hand and sending the output
> > ot the list?
> Yes naturally, here there are:
> 
> # ./tst-timer4 
> clock_gettime returned timespec = { 1083087449, 554351000 }
> clock_getres returned timespec = { 0, 10000000 }
> # echo $?
> 0
> 
> Well duno if results are right (first line different on each test but the
> second is constant with or without TIMEOUTFACTOR)?
> 
> Don't hesitate with more questions ;)

Okay, it's a 64-bit kernel problem them. What kernel are you running
right now?

c.

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [parisc-linux] RE: [PATCH] glibc-2.3.3 error with debian unstable compiler...
  2004-04-27 17:15                         ` Carlos O'Donell
@ 2004-04-28  5:39                           ` Joel Soete
  2004-04-28  9:31                             ` Joel Soete
  0 siblings, 1 reply; 28+ messages in thread
From: Joel Soete @ 2004-04-28  5:39 UTC (permalink / raw)
  To: Carlos O'Donell; +Cc: parisc-linux


> Okay, it's a 64-bit kernel problem them. What kernel are you running
> right now?

a merged 2.6.6-rc1-pa2 and upstream 2.6.6-rc2 :) 32bits build with lastest
gcc-3.3, binutils (debian unstable) and libc6 (cvs 20040426 + your patches
:)

I also have a b2k on which I can try to build 64bits kernel with gcc-3.0?

Joel


----------------------------------------------------------------------------------------
Tiscali ADSL: 35 €/mois, la meilleure offre du marché!
http://reg.tiscali.be/default.asp?lg=fr

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [parisc-linux] RE: [PATCH] glibc-2.3.3 error with debian unstable compiler...
  2004-04-28  5:39                           ` Joel Soete
@ 2004-04-28  9:31                             ` Joel Soete
  2004-04-28 15:54                               ` Carlos O'Donell
  0 siblings, 1 reply; 28+ messages in thread
From: Joel Soete @ 2004-04-28  9:31 UTC (permalink / raw)
  To: Carlos O'Donell; +Cc: parisc-linux

Carlos,
> > 
> > Okay, it's a 64-bit kernel problem them. What kernel are you running
> > right now?
> 
> a merged 2.6.6-rc1-pa2 and upstream 2.6.6-rc2 :) 32bits build with lastest
> gcc-3.3, binutils (debian unstable) and libc6 (cvs 20040426 + your patches
> :)
>
> I also have a b2k on which I can try to build 64bits kernel with gcc-3.0?

Certainly: I just rebuild my own merge of 2.6.6-rc1-pa6 and 2.6.6-rc3 with
gcc-hppa64 (from ftp.p-l.org) and reboot my b2k with it and effectively the
result of tst-timer4 is a bit different:
# ./tst-timer4
clock_gettime returned timespec = { 1083142973, 718981000 }
clock_getres returned timespec = { 0, 10000000 }
*** sig1_sigval.sival_ptr (nil) != 0xbff00690
*** sig2_sigval.sival_ptr 0 != 163
# echo $?
0

hth,
    Joel


----------------------------------------------------------------------------------------
Tiscali ADSL: 35 €/mois, la meilleure offre du marché!
http://reg.tiscali.be/default.asp?lg=fr

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [parisc-linux] RE: [PATCH] glibc-2.3.3 error with debian unstable compiler...
  2004-04-28  9:31                             ` Joel Soete
@ 2004-04-28 15:54                               ` Carlos O'Donell
  2004-04-28 16:29                                 ` Joel Soete
  0 siblings, 1 reply; 28+ messages in thread
From: Carlos O'Donell @ 2004-04-28 15:54 UTC (permalink / raw)
  To: Joel Soete; +Cc: parisc-linux

On Wed, Apr 28, 2004 at 11:31:29AM +0200, Joel Soete wrote:
> Certainly: I just rebuild my own merge of 2.6.6-rc1-pa6 and 2.6.6-rc3 with
> gcc-hppa64 (from ftp.p-l.org) and reboot my b2k with it and effectively the
> result of tst-timer4 is a bit different:
> # ./tst-timer4
> clock_gettime returned timespec = { 1083142973, 718981000 }
> clock_getres returned timespec = { 0, 10000000 }
> *** sig1_sigval.sival_ptr (nil) != 0xbff00690
> *** sig2_sigval.sival_ptr 0 != 163
> # echo $?
> 0

It does help, we have a padding problem in siginfo, I never had enough
test cases to prove the __SI_RT parts were in the right place. This goes
to show that what you can't test is probably wrong.

I have to work on this today...

c.

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [parisc-linux] RE: [PATCH] glibc-2.3.3 error with debian unstable compiler...
  2004-04-28 15:54                               ` Carlos O'Donell
@ 2004-04-28 16:29                                 ` Joel Soete
  2004-04-28 17:32                                   ` Joel Soete
  0 siblings, 1 reply; 28+ messages in thread
From: Joel Soete @ 2004-04-28 16:29 UTC (permalink / raw)
  To: Carlos O'Donell; +Cc: parisc-linux

>
> It does help, we have a padding problem in siginfo, I never had enough
> test cases to prove the __SI_RT parts were in the right place. This goes
> to show that what you can't test is probably wrong.
> 
> I have to work on this today...

I trust it is not related but btw I encounter another strange pb with this
kernel:
I launch a find on a big disk and get severall messages:
find: /a/dir/of/mine: Input/Output error
...

the same when ls -l /a/dir/of/mine (I/O error)

well this is a merge of our 2.6.6-rc1-pa6 and upsteam 2.6.6-rc3 64bits.
if i reboot with 2.6.6-rc1-pa0 32bits no pb.

So i will try to rebuild 2.6.6-rc1-pa6 alone in 64bit and see if the pb is
there or in my merge?

Cheers,
    Joel



----------------------------------------------------------------------------------------
Tiscali ADSL: 35 €/mois, la meilleure offre du marché!
http://reg.tiscali.be/default.asp?lg=fr

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [parisc-linux] RE: [PATCH] glibc-2.3.3 error with debian unstable compiler...
  2004-04-28 16:29                                 ` Joel Soete
@ 2004-04-28 17:32                                   ` Joel Soete
  0 siblings, 0 replies; 28+ messages in thread
From: Joel Soete @ 2004-04-28 17:32 UTC (permalink / raw)
  To: Carlos O'Donell; +Cc: parisc-linux

Well,

That must be something wrong in my merge 2.6.6-rc1-pa6 works fine. I will
wait the p-l merge.

Joel

> -- Original Message --
> Date: Wed, 28 Apr 2004 18:29:55 +0200
> From: "Joel Soete" <soete.joel@tiscali.be>
> To: "Carlos O'Donell" <carlos@baldric.uwo.ca>
> Cc: parisc-linux@lists.parisc-linux.org
> Subject: Re: [parisc-linux] RE: [PATCH] glibc-2.3.3 error with debian unstable
compiler...
> 
> 
> >
> It does help, we have a padding problem in siginfo, I never had enough
> test cases to prove the __SI_RT parts were in the right place. This goes
> to show that what you can't test is probably wrong.
> 
> I have to work on this today...

I t
> ust it is not related but btw I encounter another strange pb with this
kernel:
I launch a find on a big disk and get severall messages:
find: /a/dir/of/mine: Input/Output error
...

the same when ls -l /a/dir/of/mine (I/O error)

well this is a
> merge of our 2.6.6-rc1-pa6 and upsteam 2.6.6-rc3 64bits.
if i reboot with 2.6.6-rc1-pa0 32bits no pb.

So i will try to rebuild 2.6.6-rc1-pa6 alone in 64bit and see if the pb is
there or in my merge?

Cheers,
    Joel



--------------------
> -------------------------------------------------------------------
Tiscali ADSL: 35 ?/mois, la meilleure offre du marché!
http://reg.tiscali.be/default.asp?lg=fr



_______________________________________________
parisc-linux mailing list
pari
> c-linux@lists.parisc-linux.org
http://lists.parisc-linux.org/mailman/listinfo/parisc-linux



----------------------------------------------------------------------------------------
Tiscali ADSL: 35 €/mois, la meilleure offre du marché!
http://reg.tiscali.be/default.asp?lg=fr

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [parisc-linux] glibc-2.3.3: a warning with gcc-3.3 become an error with gcc-3.5
  2004-03-02 14:59   ` Joel Soete
  2004-03-02 15:27     ` Matthew Wilcox
@ 2004-03-02 19:50     ` Carlos O'Donell
  2004-03-06 22:14       ` Joel Soete
  1 sibling, 1 reply; 28+ messages in thread
From: Carlos O'Donell @ 2004-03-02 19:50 UTC (permalink / raw)
  To: Joel Soete; +Cc: parisc-linux

Joel,

Could you lookup the C99 specification please?

> --- ./feupdateenv.c.orig	2004-03-02 15:21:25.959085856 +0100
> +++ ./feupdateenv.c	2004-03-02 07:53:27.938166120 +0100
> @@ -28,11 +28,16 @@
>    /* Get the current exception status. */
>    __asm__ ("fstd %%fr0,0(%1)" : "=m" (*sw) : "r" (sw));
>    sw[0] &= FE_ALL_EXCEPT;
> -  envp->__status_word = envp->__status_word | sw[0];
> +/*  envp->__status_word = envp->__status_word | sw[0]; */

This is required to set the flags. Please review the specfications.

>    /* Install new environment.  */
>    fesetenv (envp);
>  
> +  /* Raise the safed exception.  Incidently for us the implementation
> +     defined format of the values in objects of type fexcept_t is the
> +     same as the ones specified using the FE_* constants.  */
> +  feraiseexcept (sw[0]);
> +

AFAIK, this is not correct, the spec says that it will be an OR of the
raised exceptions, e.g. an OR of the raised flags. 

Calling feraiseexcept could fault the program.

>    /* Success.  */
>    return 0;
>  }

c.

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [parisc-linux] glibc-2.3.3: a warning with gcc-3.3 become an error with gcc-3.5
  2004-03-02 19:50     ` [parisc-linux] glibc-2.3.3: a warning with gcc-3.3 become an error with gcc-3.5 Carlos O'Donell
@ 2004-03-06 22:14       ` Joel Soete
  2004-03-06 22:57         ` Carlos O'Donell
  0 siblings, 1 reply; 28+ messages in thread
From: Joel Soete @ 2004-03-06 22:14 UTC (permalink / raw)
  To: Carlos O'Donell; +Cc: parisc-linux



Carlos O'Donell wrote:
> Joel,
> 
> Could you lookup the C99 specification please?
> 
> 
>>--- ./feupdateenv.c.orig	2004-03-02 15:21:25.959085856 +0100
>>+++ ./feupdateenv.c	2004-03-02 07:53:27.938166120 +0100
>>@@ -28,11 +28,16 @@
>>   /* Get the current exception status. */
>>   __asm__ ("fstd %%fr0,0(%1)" : "=m" (*sw) : "r" (sw));
>>   sw[0] &= FE_ALL_EXCEPT;
>>-  envp->__status_word = envp->__status_word | sw[0];
>>+/*  envp->__status_word = envp->__status_word | sw[0]; */
> 
> 
> This is required to set the flags. Please review the specfications.
> 
> 
>>   /* Install new environment.  */
>>   fesetenv (envp);
>> 
>>+  /* Raise the safed exception.  Incidently for us the implementation
>>+     defined format of the values in objects of type fexcept_t is the
>>+     same as the ones specified using the FE_* constants.  */
>>+  feraiseexcept (sw[0]);
>>+
> 
> 
> AFAIK, this is not correct, the spec says that it will be an OR of the
> raised exceptions, e.g. an OR of the raised flags. 
> 
> Calling feraiseexcept could fault the program.
> 
> 
>>   /* Success.  */
>>   return 0;
>> }
> 
Taken into account of your relevant remarks here is my new proposal:
=========><=========
--- sysdeps/hppa/fpu/feupdateenv.c.orig 2004-03-06 21:18:40.789605000 +0100
+++ sysdeps/hppa/fpu/feupdateenv.c      2004-03-06 21:22:21.839605000 +0100
@@ -23,15 +23,16 @@
  int
  feupdateenv (const fenv_t *envp)
  {
-  unsigned int sw[2];
+  struct fenv_t saved_env, new_env;

    /* Get the current exception status. */
-  __asm__ ("fstd %%fr0,0(%1)" : "=m" (*sw) : "r" (sw));
-  sw[0] &= FE_ALL_EXCEPT;
-  envp->__status_word = envp->__status_word | sw[0];
+  fegetenv(&saved_env);
+  saved_env.__status_word &= FE_ALL_EXCEPT;
+  new_env.__status_word = envp->__status_word | saved_env.__status_word;
+  for (i = 0; i < 7; i++) new_env.__exception[i] = envp->__exception[i];

    /* Install new environment.  */
-  fesetenv (envp);
+  fesetenv (&new_env);

    /* Success.  */
    return 0;
=========><=========

Joel

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [parisc-linux] glibc-2.3.3: a warning with gcc-3.3 become an error with gcc-3.5
  2004-03-06 22:14       ` Joel Soete
@ 2004-03-06 22:57         ` Carlos O'Donell
  2004-03-07 11:07           ` Joel Soete
  2004-03-14 16:21           ` Joel Soete
  0 siblings, 2 replies; 28+ messages in thread
From: Carlos O'Donell @ 2004-03-06 22:57 UTC (permalink / raw)
  To: Joel Soete; +Cc: parisc-linux

On Sat, Mar 06, 2004 at 10:14:12PM +0000, Joel Soete wrote:
> Taken into account of your relevant remarks here is my new proposal:
> =========><=========
> --- sysdeps/hppa/fpu/feupdateenv.c.orig 2004-03-06 21:18:40.789605000 +0100
> +++ sysdeps/hppa/fpu/feupdateenv.c      2004-03-06 21:22:21.839605000 +0100
> @@ -23,15 +23,16 @@
>  int
>  feupdateenv (const fenv_t *envp)
>  {
> -  unsigned int sw[2];
> +  struct fenv_t saved_env, new_env;
> 
>    /* Get the current exception status. */
> -  __asm__ ("fstd %%fr0,0(%1)" : "=m" (*sw) : "r" (sw));
> -  sw[0] &= FE_ALL_EXCEPT;
> -  envp->__status_word = envp->__status_word | sw[0];

Yeah, this is bad, because 'envp' is const, and we shouldn't be changing
its value.

> +  fegetenv(&saved_env);
> +  saved_env.__status_word &= FE_ALL_EXCEPT;
> +  new_env.__status_word = envp->__status_word | saved_env.__status_word;
> +  for (i = 0; i < 7; i++) new_env.__exception[i] = envp->__exception[i];

Why do you use "new_env" at all? Just use saved_env. Then you don't have
to have a silly loop to copy the exceptions, they'll already be there.
Then you can call fesetenv(&saved_env);
 
>    /* Install new environment.  */
> -  fesetenv (envp);
> +  fesetenv (&new_env);
> 
>    /* Success.  */
>    return 0;
> =========><=========

c.

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [parisc-linux] glibc-2.3.3: a warning with gcc-3.3 become an error with gcc-3.5
  2004-03-06 22:57         ` Carlos O'Donell
@ 2004-03-07 11:07           ` Joel Soete
  2004-03-07 20:59             ` Carlos O'Donell
  2004-03-14 16:21           ` Joel Soete
  1 sibling, 1 reply; 28+ messages in thread
From: Joel Soete @ 2004-03-07 11:07 UTC (permalink / raw)
  To: Carlos O'Donell; +Cc: parisc-linux



Carlos O'Donell wrote:
> On Sat, Mar 06, 2004 at 10:14:12PM +0000, Joel Soete wrote:
> 
>>Taken into account of your relevant remarks here is my new proposal:
>>=========><=========
>>--- sysdeps/hppa/fpu/feupdateenv.c.orig 2004-03-06 21:18:40.789605000 +0100
>>+++ sysdeps/hppa/fpu/feupdateenv.c      2004-03-06 21:22:21.839605000 +0100
>>@@ -23,15 +23,16 @@
>> int
>> feupdateenv (const fenv_t *envp)
>> {
>>-  unsigned int sw[2];
>>+  struct fenv_t saved_env, new_env;
>>
>>   /* Get the current exception status. */
>>-  __asm__ ("fstd %%fr0,0(%1)" : "=m" (*sw) : "r" (sw));
>>-  sw[0] &= FE_ALL_EXCEPT;
>>-  envp->__status_word = envp->__status_word | sw[0];
> 
> 
> Yeah, this is bad, because 'envp' is const, and we shouldn't be changing
> its value.
> 
> 
>>+  fegetenv(&saved_env);
>>+  saved_env.__status_word &= FE_ALL_EXCEPT;
>>+  new_env.__status_word = envp->__status_word | saved_env.__status_word;
>>+  for (i = 0; i < 7; i++) new_env.__exception[i] = envp->__exception[i];
> 
> 
> Why do you use "new_env" at all?

The reason is very simple: I just begin with glibc and my knowledge is still at stone age in this field.
This is an evidence that sun appears red in the early morning or at the end of the evening but white for the rest of a suny day, why?
The same here, hppa don't need feraiseexcept() but the alpha, i386, ia64, m68k, mips, (ppc), s390 and sparc need and I don't yet 
understand why ;)

> Just use saved_env. Then you don't have
> to have a silly loop to copy the exceptions, they'll already be there.
> Then you can call fesetenv(&saved_env);
>  
> 
>>   /* Install new environment.  */
>>-  fesetenv (envp);
>>+  fesetenv (&new_env);
>>
>>   /* Success.  */
>>   return 0;
>>=========><=========
> 
I will send you another patch which take into account your relevant remarks.

In the mean time I encounter this new pb:
/Sources/parisc-linux/xc/bin/hppa-linux-gcc dl-sym.c -c -std=gnu99 -O2 -Wall -Winline -Wstrict-prototypes -Wwrite-strings -g 
-I../include -I. -I/Sources/parisc-linux/build/glibc/elf -I.. -I../libio  -I/Sources/parisc-linux/build/glibc 
-I../sysdeps/hppa/elf -I../linuxthreads/sysdeps/unix/sysv/linux/hppa -I../linuxthreads/sysdeps/unix/sysv/linux 
-I../linuxthreads/sysdeps/pthread -I../sysdeps/pthread -I../linuxthreads/sysdeps/unix/sysv -I../linuxthreads/sysdeps/unix 
-I../linuxthreads/sysdeps/hppa -I../sysdeps/unix/sysv/linux/hppa -I../sysdeps/unix/sysv/linux -I../sysdeps/gnu 
-I../sysdeps/unix/common -I../sysdeps/unix/mman -I../sysdeps/unix/inet -I../sysdeps/unix/sysv -I../sysdeps/unix -I../sysdeps/posix 
-I../sysdeps/hppa/hppa1.1 -I../sysdeps/wordsize-32 -I../sysdeps/ieee754/flt-32 -I../sysdeps/ieee754/dbl-64 
-I../sysdeps/ieee754/ldbl-128 -I../sysdeps/hppa/fpu -I../sysdeps/hppa -I../sysdeps/ieee754 -I../sysdeps/generic/elf 
-I../sysdeps/generic -nostdinc -isystem /Sources/parisc-linux/xc/lib/gcc/hppa-linux/3.5.0/include -isystem 
/Sources/parisc-linux/sources/linux/include -D_LIBC_REENTRANT -include ../include/libc-symbols.h       -o 
/Sources/parisc-linux/build/glibc/elf/dl-sym.o -MD -MP -MF
/Sources/parisc-linux/build/glibc/elf/dl-sym.o.dt
In file included from dl-sym.c:27:
../sysdeps/generic/dl-tls.h:2:2: #error "architecture-specific version of <dl-tls.h> missing"
make[2]: *** [/Sources/parisc-linux/build/glibc/elf/dl-sym.o] Error 1
make[2]: Leaving directory `/Sources/parisc-linux/sources/glibc-2.3.3-20040306/elf'
make[1]: *** [elf/subdir_lib] Error 2
make[1]: Leaving directory `/Sources/parisc-linux/sources/glibc-2.3.3-20040306'
make: *** [all] Error 2

and would suggest this patch:
--- dl-sym.c.orig       2004-03-07 11:44:11.219605000 +0100
+++ dl-sym.c    2004-03-07 11:41:21.219605000 +0100
@@ -24,10 +24,10 @@
  #include <dlfcn.h>
  #include <ldsodefs.h>
  #include <dl-hash.h>
-#include <dl-tls.h>
-

  #if defined USE_TLS && defined SHARED
+#include <dl-tls.h>
+
  /* Systems which do not have tls_index also probably have to define
     DONT_USE_TLS_INDEX.  */
=========><=========

Thanks,
	Joel

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [parisc-linux] glibc-2.3.3: a warning with gcc-3.3 become an error with gcc-3.5
  2004-03-07 11:07           ` Joel Soete
@ 2004-03-07 20:59             ` Carlos O'Donell
  0 siblings, 0 replies; 28+ messages in thread
From: Carlos O'Donell @ 2004-03-07 20:59 UTC (permalink / raw)
  To: Joel Soete; +Cc: parisc-linux

On Sun, Mar 07, 2004 at 11:07:19AM +0000, Joel Soete wrote:
> >Why do you use "new_env" at all?
> 
> The reason is very simple: I just begin with glibc and my knowledge is 
> still at stone age in this field.
> This is an evidence that sun appears red in the early morning or at the end 
> of the evening but white for the rest of a suny day, why?
> The same here, hppa don't need feraiseexcept() but the alpha, i386, ia64, 
> m68k, mips, (ppc), s390 and sparc need and I don't yet understand why ;)

Because our feraiseexcept() is different from theirs?
 
> In the mean time I encounter this new pb:
> /Sources/parisc-linux/xc/bin/hppa-linux-gcc dl-sym.c -c -std=gnu99 -O2 
> -Wall -Winline -Wstrict-prototypes -Wwrite-strings -g -I../include -I. 
> -I/Sources/parisc-linux/build/glibc/elf -I.. -I../libio  
> -I/Sources/parisc-linux/build/glibc -I../sysdeps/hppa/elf 
> -I../linuxthreads/sysdeps/unix/sysv/linux/hppa 
> -I../linuxthreads/sysdeps/unix/sysv/linux -I../linuxthreads/sysdeps/pthread 
> -I../sysdeps/pthread -I../linuxthreads/sysdeps/unix/sysv 
> -I../linuxthreads/sysdeps/unix -I../linuxthreads/sysdeps/hppa 
> -I../sysdeps/unix/sysv/linux/hppa -I../sysdeps/unix/sysv/linux 
> -I../sysdeps/gnu -I../sysdeps/unix/common -I../sysdeps/unix/mman 
> -I../sysdeps/unix/inet -I../sysdeps/unix/sysv -I../sysdeps/unix 
> -I../sysdeps/posix -I../sysdeps/hppa/hppa1.1 -I../sysdeps/wordsize-32 
> -I../sysdeps/ieee754/flt-32 -I../sysdeps/ieee754/dbl-64 
> -I../sysdeps/ieee754/ldbl-128 -I../sysdeps/hppa/fpu -I../sysdeps/hppa 
> -I../sysdeps/ieee754 -I../sysdeps/generic/elf -I../sysdeps/generic 
> -nostdinc -isystem 
> /Sources/parisc-linux/xc/lib/gcc/hppa-linux/3.5.0/include -isystem 
> /Sources/parisc-linux/sources/linux/include -D_LIBC_REENTRANT -include 
> ../include/libc-symbols.h       -o 
> /Sources/parisc-linux/build/glibc/elf/dl-sym.o -MD -MP -MF
> /Sources/parisc-linux/build/glibc/elf/dl-sym.o.dt
> In file included from dl-sym.c:27:
> ../sysdeps/generic/dl-tls.h:2:2: #error "architecture-specific version of 
> <dl-tls.h> missing"
> make[2]: *** [/Sources/parisc-linux/build/glibc/elf/dl-sym.o] Error 1
> make[2]: Leaving directory 
> `/Sources/parisc-linux/sources/glibc-2.3.3-20040306/elf'
> make[1]: *** [elf/subdir_lib] Error 2
> make[1]: Leaving directory 
> `/Sources/parisc-linux/sources/glibc-2.3.3-20040306'
> make: *** [all] Error 2
> 
> and would suggest this patch:
> --- dl-sym.c.orig       2004-03-07 11:44:11.219605000 +0100
> +++ dl-sym.c    2004-03-07 11:41:21.219605000 +0100
> @@ -24,10 +24,10 @@
>  #include <dlfcn.h>
>  #include <ldsodefs.h>
>  #include <dl-hash.h>
> -#include <dl-tls.h>
> -
> 
>  #if defined USE_TLS && defined SHARED
> +#include <dl-tls.h>
> +
>  /* Systems which do not have tls_index also probably have to define
>     DONT_USE_TLS_INDEX.  */

I'm doing a build right now, if I see this issue I'll report it upstream
with this fix.

c.

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [parisc-linux] glibc-2.3.3: a warning with gcc-3.3 become an error with gcc-3.5
  2004-03-06 22:57         ` Carlos O'Donell
  2004-03-07 11:07           ` Joel Soete
@ 2004-03-14 16:21           ` Joel Soete
  2004-03-14 19:26             ` Carlos O'Donell
  1 sibling, 1 reply; 28+ messages in thread
From: Joel Soete @ 2004-03-14 16:21 UTC (permalink / raw)
  To: Carlos O'Donell; +Cc: parisc-linux

Carlos,

Carlos O'Donell wrote:
> On Sat, Mar 06, 2004 at 10:14:12PM +0000, Joel Soete wrote:
> 
>>Taken into account of your relevant remarks here is my new proposal:
>>=========><=========
>>--- sysdeps/hppa/fpu/feupdateenv.c.orig 2004-03-06 21:18:40.789605000 +0100
>>+++ sysdeps/hppa/fpu/feupdateenv.c      2004-03-06 21:22:21.839605000 +0100
>>@@ -23,15 +23,16 @@
>> int
>> feupdateenv (const fenv_t *envp)
>> {
>>-  unsigned int sw[2];
>>+  struct fenv_t saved_env, new_env;
>>
>>   /* Get the current exception status. */
>>-  __asm__ ("fstd %%fr0,0(%1)" : "=m" (*sw) : "r" (sw));
>>-  sw[0] &= FE_ALL_EXCEPT;
>>-  envp->__status_word = envp->__status_word | sw[0];
> 
> 
> Yeah, this is bad, because 'envp' is const, and we shouldn't be changing
> its value.
> 
> 
>>+  fegetenv(&saved_env);
>>+  saved_env.__status_word &= FE_ALL_EXCEPT;
>>+  new_env.__status_word = envp->__status_word | saved_env.__status_word;
>>+  for (i = 0; i < 7; i++) new_env.__exception[i] = envp->__exception[i];
> 
> 
> Why do you use "new_env" at all? Just use saved_env. Then you don't have
> to have a silly loop to copy the exceptions, they'll already be there.
> Then you can call fesetenv(&saved_env);
>  
> 
>>   /* Install new environment.  */
>>-  fesetenv (envp);
>>+  fesetenv (&new_env);
>>
>>   /* Success.  */
>>   return 0;
>>=========><=========
> 
Taken into account of your remarks here is the final patch:
--- glibc-2.3.3-20040314/sysdeps/hppa/fpu/feupdateenv.c.orig    2003-12-18 04:57:50.000000000 +0100
+++ glibc-2.3.3-20040314/sysdeps/hppa/fpu/feupdateenv.c 2004-03-14 12:06:13.286349000 +0100
@@ -23,15 +23,15 @@
  int
  feupdateenv (const fenv_t *envp)
  {
-  unsigned int sw[2];
+  fenv_t tmp_env;

    /* Get the current exception status. */
-  __asm__ ("fstd %%fr0,0(%1)" : "=m" (*sw) : "r" (sw));
-  sw[0] &= FE_ALL_EXCEPT;
-  envp->__status_word = envp->__status_word | sw[0];
+  fegetenv(&tmp_env);
+  tmp_env.__status_word &= FE_ALL_EXCEPT;
+  tmp_env.__status_word |= envp->__status_word;

    /* Install new environment.  */
-  fesetenv (envp);
+  fesetenv (&tmp_env);

    /* Success.  */
    return 0;
=========><=========

it compiles well with gcc-3.3.

I hope it will not break anything :)

Joel

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [parisc-linux] glibc-2.3.3: a warning with gcc-3.3 become an error with gcc-3.5
  2004-03-14 16:21           ` Joel Soete
@ 2004-03-14 19:26             ` Carlos O'Donell
  2004-03-14 19:44               ` Joel Soete
  0 siblings, 1 reply; 28+ messages in thread
From: Carlos O'Donell @ 2004-03-14 19:26 UTC (permalink / raw)
  To: Joel Soete; +Cc: parisc-linux

On Sun, Mar 14, 2004 at 04:21:37PM +0000, Joel Soete wrote:
> it compiles well with gcc-3.3.
> I hope it will not break anything :)

Hope?? What does the testsuite for glibc say?

c.

^ permalink raw reply	[flat|nested] 28+ messages in thread

* Re: [parisc-linux] glibc-2.3.3: a warning with gcc-3.3 become an error with gcc-3.5
  2004-03-14 19:26             ` Carlos O'Donell
@ 2004-03-14 19:44               ` Joel Soete
  0 siblings, 0 replies; 28+ messages in thread
From: Joel Soete @ 2004-03-14 19:44 UTC (permalink / raw)
  To: Carlos O'Donell; +Cc: parisc-linux



Carlos O'Donell wrote:
> On Sun, Mar 14, 2004 at 04:21:37PM +0000, Joel Soete wrote:
> 
>>it compiles well with gcc-3.3.
>>I hope it will not break anything :)
> 
> 
> Hope?? What does the testsuite for glibc say?
> 
Right now I reach to build glibc completely (with just a warning less ;) )
This c110 is unfortunately too slow to let me complete the make -k check now, I will so continue tomorrow ;)

Regards,
	Joel

^ permalink raw reply	[flat|nested] 28+ messages in thread

end of thread, other threads:[~2004-04-28 17:32 UTC | newest]

Thread overview: 28+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2004-03-01 18:43 [parisc-linux] glibc-2.3.3: a warning with gcc-3.3 become an error with gcc-3.5 Joel Soete
2004-03-01 19:38 ` Carlos O'Donell
2004-03-02 14:59   ` Joel Soete
2004-03-02 15:27     ` Matthew Wilcox
2004-04-02 18:03       ` [parisc-linux] [PATCH] glibc-2.3.3 error with debian unstable compiler Carlos O'Donell
2004-04-06 10:46         ` [parisc-linux] " Joel Soete
2004-04-06 13:14           ` Carlos O'Donell
2004-04-20 12:22           ` Joel Soete
2004-04-20 19:35             ` Carlos O'Donell
2004-04-21  6:15               ` Joel Soete
2004-04-21 11:55                 ` Carlos O'Donell
2004-04-27  7:04                   ` Joel Soete
2004-04-27 14:57                     ` Carlos O'Donell
2004-04-27 17:06                       ` Joel Soete
2004-04-27 17:15                         ` Carlos O'Donell
2004-04-28  5:39                           ` Joel Soete
2004-04-28  9:31                             ` Joel Soete
2004-04-28 15:54                               ` Carlos O'Donell
2004-04-28 16:29                                 ` Joel Soete
2004-04-28 17:32                                   ` Joel Soete
2004-03-02 19:50     ` [parisc-linux] glibc-2.3.3: a warning with gcc-3.3 become an error with gcc-3.5 Carlos O'Donell
2004-03-06 22:14       ` Joel Soete
2004-03-06 22:57         ` Carlos O'Donell
2004-03-07 11:07           ` Joel Soete
2004-03-07 20:59             ` Carlos O'Donell
2004-03-14 16:21           ` Joel Soete
2004-03-14 19:26             ` Carlos O'Donell
2004-03-14 19:44               ` Joel Soete

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.