[Linux-ia64] [hinkelbein@ti.uni-mannheim.de: Re: Strange problems building bigloo on ia64]

[Linux-ia64] [hinkelbein@ti.uni-mannheim.de: Re: Strange problems building bigloo on ia64]

[Matthew Wilcox]

I don't grok ia64 asm enough to answer his question...

----- Forwarded message from Christian Hinkelbein <hinkelbein@ti.uni-mannheim.de> -----

Date: Mon, 19 May 2003 10:21:04 +0200
From: Christian Hinkelbein <hinkelbein@ti.uni-mannheim.de>
To: Yann Dirson <ydirson@altern.org>
Cc: debian-ia64@lists.debian.org,
	Manuel Serrano <Manuel.Serrano@sophia.inria.fr>
Subject: Re: Strange problems building bigloo on ia64

Hi,

Yann Dirson schrieb:
> 
> Hi ia64 porters,
> 
> #193608 already reports that there are 64bit issues on this package,
> but I see in the logs a lot of messages which make me suspect some gcc
> problem:
> 
> /tmp/ccnfO17g.s: Assembler messages:
> /tmp/ccnfO17g.s:958: Warning: Use of 'mov' may violate WAW dependency 'GR%, % in 1 - 127' (impliedf), specific resource number is 14

i see some similary warnings when compiling following
with gcc3.2 and binutils 2.12

//-----------
// waw.c
int sh;

int f(int s)
{
  int test = sh;

  if(s)
    return (test = 1) ? 0:(13);
  else
    return (test = 2) ? 0:(13);
}
//-----------

~/src/ia64 > gcc -c -O3 waw.c
/tmp/cchXUlJF.s: Assembler messages:
/tmp/cchXUlJF.s:27: Warning: Use of 'addl' may violate WAW dependency 'GR%, % in
1 - 127' (impliedf), specific resource number is 8
/tmp/cchXUlJF.s:27: Warning: Only the first path encountering the conflict is
reported
/tmp/cchXUlJF.s:25: Warning: This is the location of the conflicting usage
~/src/ia64 >

This only occurs when using option -O3 to gcc.
This looks like beeing an assembler problem, not one from gcc.
Also, when looking into the assembler output, i could not find an error:

        .file   "waw.c"
        .pred.safe_across_calls p1-p5,p16-p63
        .text
        .align 16
        .global f#
        .proc f#
f:
        .prologue
        .body
        .mii
        addl r3 = @ltoff(sh#), gp
        cmp4.eq p8, p9 = 0, r32
        ;;
        nop.i 0
        .mmi
        ld8 r2 = [r3]
        ;;
        ld4 r14 = [r2]
        nop.i 0
        ;;
        .mii
        (p9) cmp4.eq p6, p7 = 1, r14     ! it is this stuff that triggers the
asm warning.
        (p8) cmp4.eq p6, p7 = 2, r14     !
        ;;                               !
        (p6) mov r8 = r0                 !
        .mib                             !
        (p7) addl r8 = 13, r0            !
        nop.i 0                          !
        br.ret.sptk.many rp              !
        .endp f#
        .common sh#,4,4
        .ident  "GCC: (GNU) 3.2"

until now i could savely ignore the warning. 
if a ia64 asm guru could please comment ?

i dont see the following anything to do with the asm issue frome above.
 
> 
> And the build terminates in a strange way as well:
> 
> /build/buildd/bigloo-2.5c+really2.5d-alpha-2003-05-16/bin/bigloo -no-hello -O2 -fsharing -unsafe -static-bigloo -c jas/lib.scm -o jas/lib.o
> /build/buildd/bigloo-2.5c+really2.5d-alpha-2003-05-16/bin/bigloo -no-hello -O2 -fsharing -unsafe -static-bigloo -c jas/classfile.scm -o jas/classfile.o
> Assembler messages:
> FATAL: can't create classfile.o: No such file or directory
> make[2]: *** [jas/classfile.o] Error 1
> make[2]: Leaving directory `/build/buildd/bigloo-2.5c+really2.5d-alpha-2003-05-16/bde'
> make[1]: *** [boot] Error 2
> make[1]: Leaving directory `/build/buildd/bigloo-2.5c+really2.5d-alpha-2003-05-16'
> 
> Ideas anyone ?
> --
> Yann Dirson    <ydirson@altern.org> |    Why make M$-Bill richer & richer ?
> Debian-related: <dirson@debian.org> |   Support Debian GNU/Linux:
> Pro:    <yann.dirson@fr.alcove.com> |  Freedom, Power, Stability, Gratuity
>      http://ydirson.free.fr/        | Check <http://www.debian.org/>
> 
> --
> To UNSUBSCRIBE, email to debian-ia64-request@lists.debian.org
> with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org


 cheers
  christian


-- 
To UNSUBSCRIBE, email to debian-ia64-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org


----- End forwarded message -----

-- 
"It's not Hollywood.  War is real, war is primarily not about defeat or
victory, it is about death.  I've seen thousands and thousands of dead bodies.
Do you think I want to have an academic debate on this subject?" -- Robert Fisk

Re: [Linux-ia64] [hinkelbein@ti.uni-mannheim.de: Re: Strange problems

[CH Gowri Kumar]

Comments inline.

>         .file   "waw.c"
>         .pred.safe_across_calls p1-p5,p16-p63
>         .text
>         .align 16
>         .global f#
>         .proc f#
> f:
>         .prologue
>         .body
>         .mii
>         addl r3 = @ltoff(sh#), gp
>         cmp4.eq p8, p9 = 0, r32
>         ;;
>         nop.i 0
>         .mmi
>         ld8 r2 = [r3]
>         ;;
>         ld4 r14 = [r2]
>         nop.i 0
>         ;;
>         .mii
>         (p9) cmp4.eq p6, p7 = 1, r14     ! it is this stuff that triggers the
> asm warning.
>         (p8) cmp4.eq p6, p7 = 2, r14     !
>         ;;                               !
>         (p6) mov r8 = r0                 !
>         .mib                             !
>         (p7) addl r8 = 13, r0            !

It is at this place the assembler is cribbing.
There are two statements which are writing to the same register ofcourse
with appropriate predicate registers.

(p6) mov r8 = r0
(p7) addl r8 = r13,r0

But the predicate register p6 and p7 are set by the just above statements
>         (p9) cmp4.eq p6, p7 = 1, r14     
>         (p8) cmp4.eq p6, p7 = 2, r14     
which are again dependent on the predicates p8 and p9 which are set by the
statment:
>         cmp4.eq p8, p9 = 0, r32

I feel that the assembler is not able to recognize that the two statements 
which are writing to the same register(r8) are mutually exclusive ( only 
one of them can be executed), as it is two levels dependant on predicate 
registers.

Anyway, it's just a warning, not an error.

Regards,
Gowri Kumar

Re: [Linux-ia64] [hinkelbein@ti.uni-mannheim.de: Re: Strange problems building bigloo on ia64]

[Keith Owens]

On Mon, 19 May 2003 14:36:51 +0100, 
Matthew Wilcox <willy@debian.org> wrote:
>
>I don't grok ia64 asm enough to answer his question...
>Yann Dirson schrieb:
>~/src/ia64 > gcc -c -O3 waw.c
>/tmp/cchXUlJF.s: Assembler messages:
>/tmp/cchXUlJF.s:27: Warning: Use of 'addl' may violate WAW dependency 'GR%, % in
>1 - 127' (impliedf), specific resource number is 8
>/tmp/cchXUlJF.s:27: Warning: Only the first path encountering the conflict is
>reported
>/tmp/cchXUlJF.s:25: Warning: This is the location of the conflicting usage
>~/src/ia64 >
>
>This only occurs when using option -O3 to gcc.
>This looks like beeing an assembler problem, not one from gcc.
>Also, when looking into the assembler output, i could not find an error:
>
>        .file   "waw.c"
>        .pred.safe_across_calls p1-p5,p16-p63
>        .text
>        .align 16
>        .global f#
>        .proc f#
>f:
>        .prologue
>        .body
>        .mii
>        addl r3 = @ltoff(sh#), gp
>        cmp4.eq p8, p9 = 0, r32
>        ;;
>        nop.i 0
>        .mmi
>        ld8 r2 = [r3]
>        ;;
>        ld4 r14 = [r2]
>        nop.i 0
>        ;;
>        .mii
>        (p9) cmp4.eq p6, p7 = 1, r14     ! it is this stuff that triggers the
>asm warning.
>        (p8) cmp4.eq p6, p7 = 2, r14     !
>        ;;                               !
>        (p6) mov r8 = r0                 !
>        .mib                             !
>        (p7) addl r8 = 13, r0            !
>        nop.i 0                          !
>        br.ret.sptk.many rp              !
>        .endp f#
>        .common sh#,4,4
>        .ident  "GCC: (GNU) 3.2"

The assembler normally recognises that two predicates are exclusive.
In this case the setting of p6 and p7 is itself predicated (by p8, p9)
so the assembler cannot analyse the logic flow well enough to see that
p6 and p7 are exclusive.  IMHO gcc should issue .pred.rel directives
for complex predicates like this, e.g.

	.pred.rel "mutex", p6, p7
        (p6) mov r8 = r0
        (p7) addl r8 = 13, r0
	.pred.rel "clear", p6, p7

gcc knows that the predicates are exclusive and it is far easier for
gcc to issue the .pred.rel directives than for the assembler to do
complex logic path analysis to deduce what gcc already knows.