Re: gcc 4.0.0 and xen-devel

Re: gcc 4.0.0 and xen-devel

[David F Barrera]

On Tue, 2005-06-07 at 17:08 -0300, Flavio Leitner wrote:
> 
> Hello!
> 
> I tried to compile xen 2.0.6 with gcc 4.0.0 and got lots of warnings
> like:
> 
> 'warning: pointer targets in passing argument X of 'function' differ in
> signedness' 
> 
I am seeing the same thing on xen-unstable (xen devel 3.0) and GCC 4.0.0
on a FC4 platform. 
I submitted a defect on the Xensource Bugzilla, "Bugzilla Bug 73 
gcc4 build break ...", to track this issue.

> Is there any intention of compile xen stable/unstable with gcc 4.0.0? 
> 
> Thanks!
-- 
Regards,

David F Barrera
Linux Technology Center
Systems and Technology Group, IBM

"The wisest men follow their own direction. "
                                                        Euripides

Re: [Xen-users] gcc 4.0.0 and xen-devel

[David F Barrera]

This problem has been corrected. Fix is in the latest BK source tree.

On Wed, 2005-06-08 at 08:20 -0500, David F Barrera wrote:
> On Tue, 2005-06-07 at 17:08 -0300, Flavio Leitner wrote:
> > 
> > Hello!
> > 
> > I tried to compile xen 2.0.6 with gcc 4.0.0 and got lots of warnings
> > like:
> > 
> > 'warning: pointer targets in passing argument X of 'function' differ in
> > signedness' 
> > 
> I am seeing the same thing on xen-unstable (xen devel 3.0) and GCC 4.0.0
> on a FC4 platform. 
> I submitted a defect on the Xensource Bugzilla, "Bugzilla Bug 73 
> gcc4 build break ...", to track this issue.
> 
> > Is there any intention of compile xen stable/unstable with gcc 4.0.0? 
> > 
> > Thanks!
-- 
Regards,

David F Barrera
Linux Technology Center
Systems and Technology Group, IBM

"The wisest men follow their own direction. "
                                                        Euripides

Re: gcc 4.0.0 and xen-devel

[Flavio Leitner]

I checkout bktree from bk://xen.bkbits.net/xeno-unstable.bk 
and now I'm getting these messages:

make world or make dist

<snip>
gcc -nostdinc -fno-builtin -fno-common -fno-strict-aliasing -iwithprefix
include -Wall -Werror -Wno-pointer-arith -pipe
-I/home/fbl/xeno-unstable.bk/xen/include
-I/home/fbl/xeno-unstable.bk/xen/include/asm-x86/mach-generic
-I/home/fbl/xeno-unstable.bk/xen/include/asm-x86/mach-default -O3
-fomit-frame-pointer -msoft-float -m32 -march=i686 -DNDEBUG -c i8259.c
-o i8259.o
{standard input}: Assembler messages:
{standard input}:6: Error: suffix or operands invalid for `mov'
{standard input}:6: Error: suffix or operands invalid for `mov'
{standard input}:6: Error: suffix or operands invalid for `mov'
{standard input}:6: Error: suffix or operands invalid for `mov'
{standard input}:1296: Error: suffix or operands invalid for `mov'
{standard input}:1296: Error: suffix or operands invalid for `mov'
<snip>


On Wed, Jun 08, 2005 at 11:02:21AM -0500, David F Barrera wrote:
> This problem has been corrected. Fix is in the latest BK source tree.
> 
> On Wed, 2005-06-08 at 08:20 -0500, David F Barrera wrote:
> > On Tue, 2005-06-07 at 17:08 -0300, Flavio Leitner wrote:
> > > 
> > > Hello!
> > > 
> > > I tried to compile xen 2.0.6 with gcc 4.0.0 and got lots of warnings
> > > like:
> > > 
> > > 'warning: pointer targets in passing argument X of 'function' differ in
> > > signedness' 
> > > 
> > I am seeing the same thing on xen-unstable (xen devel 3.0) and GCC 4.0.0
> > on a FC4 platform. 
> > I submitted a defect on the Xensource Bugzilla, "Bugzilla Bug 73 
> > gcc4 build break ...", to track this issue.
> > 
> > > Is there any intention of compile xen stable/unstable with gcc 4.0.0? 
> > > 
> > > Thanks!
> -- 
> Regards,
> 
> David F Barrera
> Linux Technology Center
> Systems and Technology Group, IBM
> 
> "The wisest men follow their own direction. "
>                                                         Euripides
> 
> 
> _______________________________________________
> Xen-users mailing list
> Xen-users@lists.xensource.com
> http://lists.xensource.com/xen-users

-- 
Flávio Bruno Leitner <fbl@conectiva.com.br>
[0EA2 7F40 4CF4 1E63 4AF6  33C0 3E10 E205 F251 EDDA]

Re: [Xen-devel] Re: gcc 4.0.0 and xen-devel

[Flavio Leitner]

More info below where is failing.

On Thu, Jun 09, 2005 at 10:13:37AM -0300, Flavio Leitner wrote:
> 
> I checkout bktree from bk://xen.bkbits.net/xeno-unstable.bk 
> and now I'm getting these messages:
> 
> make world or make dist
> 
> <snip>
> gcc -nostdinc -fno-builtin -fno-common -fno-strict-aliasing -iwithprefix
> include -Wall -Werror -Wno-pointer-arith -pipe
> -I/home/fbl/xeno-unstable.bk/xen/include
> -I/home/fbl/xeno-unstable.bk/xen/include/asm-x86/mach-generic
> -I/home/fbl/xeno-unstable.bk/xen/include/asm-x86/mach-default -O3
> -fomit-frame-pointer -msoft-float -m32 -march=i686 -DNDEBUG -c i8259.c
> -o i8259.o
> {standard input}: Assembler messages:
> {standard input}:6: Error: suffix or operands invalid for `mov'
> {standard input}:6: Error: suffix or operands invalid for `mov'
> {standard input}:6: Error: suffix or operands invalid for `mov'
> {standard input}:6: Error: suffix or operands invalid for `mov'
> {standard input}:1296: Error: suffix or operands invalid for `mov'
> {standard input}:1296: Error: suffix or operands invalid for `mov'
> <snip>

Gcc 4.0.0
Binutils 2.16.90.0.2

The assembler output:
       .file   "i8259.c"
#APP

.align 16,0x90
common_interrupt:
        cld;pushl %eax;pushl %ebp;pushl %edi;pushl %esi;pushl %edx;pushl %ecx;pushl %ebx;testl $0x00020000,40(%esp);jz 2f;call setup_vm86_frame;jmp 3f;2:testb $3,36(%esp);jz 1f;movl %ds,56(%esp);movl %es,52(%esp);movl %fs,60(%esp);movl %gs,64(%esp);3:movl $(0xe010),%eax;movl %eax,%ds;movl %eax,%es;1:movl %esp,%eax
        pushl %eax
        call do_IRQ
        addl $4,%esp
        jmp ret_from_intr
<snip>

spurious_interrupt:
        pushl $0xff<<16
        cld;pushl %eax;pushl %ebp;pushl %edi;pushl %esi;pushl %edx;pushl %ecx;pushl %ebx;testl $0x00020000,40(%esp);jz 2f;call setup_vm86_frame;jmp 3f;2:testb $3,36(%esp);jz 1f;movl %ds,56(%esp);movl %es,52(%esp);movl %fs,60(%esp);movl %gs,64(%esp);3:movl $(0xe010),%eax;movl %eax,%ds;movl %eax,%es;1:call smp_spurious_interrupt
        jmp ret_from_intr

#NO_APP


-- 
Flávio Bruno Leitner <fbl@conectiva.com.br>
[0EA2 7F40 4CF4 1E63 4AF6  33C0 3E10 E205 F251 EDDA]

Re: Re: [Xen-users] gcc 4.0.0 and xen-devel

[Vincent Hanquez]

On Thu, Jun 09, 2005 at 10:55:07AM -0300, Flavio Leitner wrote:
> 
> More info below where is failing.
> 
> On Thu, Jun 09, 2005 at 10:13:37AM -0300, Flavio Leitner wrote:
> > 
> > I checkout bktree from bk://xen.bkbits.net/xeno-unstable.bk 
> > and now I'm getting these messages:
> > 
> > make world or make dist
> > 
> > <snip>
> > gcc -nostdinc -fno-builtin -fno-common -fno-strict-aliasing -iwithprefix
> > include -Wall -Werror -Wno-pointer-arith -pipe
> > -I/home/fbl/xeno-unstable.bk/xen/include
> > -I/home/fbl/xeno-unstable.bk/xen/include/asm-x86/mach-generic
> > -I/home/fbl/xeno-unstable.bk/xen/include/asm-x86/mach-default -O3
> > -fomit-frame-pointer -msoft-float -m32 -march=i686 -DNDEBUG -c i8259.c
> > -o i8259.o
> > {standard input}: Assembler messages:
> > {standard input}:6: Error: suffix or operands invalid for `mov'
> > {standard input}:6: Error: suffix or operands invalid for `mov'
> > {standard input}:6: Error: suffix or operands invalid for `mov'
> > {standard input}:6: Error: suffix or operands invalid for `mov'
> > {standard input}:1296: Error: suffix or operands invalid for `mov'
> > {standard input}:1296: Error: suffix or operands invalid for `mov'
> > <snip>
> 
> Gcc 4.0.0
> Binutils 2.16.90.0.2

these binutils version are unstable CVS (not a good idea to have that IMHO) 

and as far as I remember they remove support for "movl %ds, ...".
now is must use: "mov %ds, ..."

does the following patch fix your problems ?

diff -Nru a/xen/include/asm-x86/x86_32/asm_defns.h b/xen/include/asm-x86/x86_32/asm_defns.h
--- a/xen/include/asm-x86/x86_32/asm_defns.h	2005-06-09 17:52:47 +01:00
+++ b/xen/include/asm-x86/x86_32/asm_defns.h	2005-06-09 17:52:47 +01:00
@@ -19,10 +19,10 @@
         "jmp 3f;"                                                       \
         "2:testb $3,"STR(UREGS_cs)"(%esp);"                             \
         "jz 1f;"                                                        \
-        "movl %ds,"STR(UREGS_ds)"(%esp);"                               \
-        "movl %es,"STR(UREGS_es)"(%esp);"                               \
-        "movl %fs,"STR(UREGS_fs)"(%esp);"                               \
-        "movl %gs,"STR(UREGS_gs)"(%esp);"                               \
+        "mov %ds,"STR(UREGS_ds)"(%esp);"                                \
+        "mov %es,"STR(UREGS_es)"(%esp);"                                \
+        "mov %fs,"STR(UREGS_fs)"(%esp);"                                \
+        "mov %gs,"STR(UREGS_gs)"(%esp);"                                \
         "3:"
 
 #define SAVE_ALL_NOSEGREGS(_reg)                \



-- 
Vincent Hanquez

Re: [Xen-devel] Re: gcc 4.0.0 and xen-devel

[Flavio Leitner]

> and as far as I remember they remove support for "movl %ds, ...".
> now is must use: "mov %ds, ..."

Yes, I found this on google.
http://www.kernel.org/pub/linux/devel/binutils/release.binutils-2.16.90.0.2


> does the following patch fix your problems ?

Part of it because there are more places. Thanks!

Well, now it breaks because of warnings.

===== xen/arch/x86/x86_32/entry.S 1.112 vs edited =====
--- 1.112/xen/arch/x86/x86_32/entry.S	2005-06-02 17:05:31 -04:00
+++ edited/xen/arch/x86/x86_32/entry.S	2005-06-09 10:32:50 -04:00
@@ -176,10 +176,10 @@
 restore_all_guest:
         testl $X86_EFLAGS_VM,UREGS_eflags(%esp)
         jnz  restore_all_vm86
-FLT1:   movl UREGS_ds(%esp),%ds
-FLT2:   movl UREGS_es(%esp),%es
-FLT3:   movl UREGS_fs(%esp),%fs
-FLT4:   movl UREGS_gs(%esp),%gs
+FLT1:   mov UREGS_ds(%esp),%ds
+FLT2:   mov UREGS_es(%esp),%es
+FLT3:   mov UREGS_fs(%esp),%fs
+FLT4:   mov UREGS_gs(%esp),%gs
 restore_all_vm86:
         popl %ebx
         popl %ecx
@@ -306,7 +306,7 @@
         testl $(2|X86_EFLAGS_VM),%ecx
         jz   ring1 /* jump if returning to an existing ring-1 activation */
         movl VCPU_kernel_sp(%ebx),%esi
-FLT6:   movl VCPU_kernel_ss(%ebx),%gs
+FLT6:   mov VCPU_kernel_ss(%ebx),%gs
         testl $X86_EFLAGS_VM,UREGS_eflags+4(%esp)
         jz   nvm86_1
         subl $16,%esi       /* push ES/DS/FS/GS (VM86 stack frame) */
@@ -326,7 +326,7 @@
         jmp 1f
 ring1:  /* obtain ss/esp from oldss/oldesp -- a ring-1 activation exists */
         movl UREGS_esp+4(%esp),%esi
-FLT13:  movl UREGS_ss+4(%esp),%gs 
+FLT13:  mov UREGS_ss+4(%esp),%gs 
 1:      /* Construct a stack frame: EFLAGS, CS/EIP */
         movb TRAPBOUNCE_flags(%edx),%cl
         subl $12,%esi
@@ -383,7 +383,7 @@
 nvm86_3:/* Rewrite our stack frame and return to ring 1. */
         /* IA32 Ref. Vol. 3: TF, VM, RF and NT flags are cleared on trap. */
         andl $0xfffcbeff,UREGS_eflags+4(%esp)
-        movl %gs,UREGS_ss+4(%esp)
+        mov %gs,UREGS_ss+4(%esp)
         movl %esi,UREGS_esp+4(%esp)
         movzwl TRAPBOUNCE_cs(%edx),%eax
         movl %eax,UREGS_cs+4(%esp)
@@ -670,7 +670,7 @@
 
         # GS:ESI == Ring-1 stack activation
         movl UREGS_esp(%esp),%esi
-VFLT1:  movl UREGS_ss(%esp),%gs
+VFLT1:  mov UREGS_ss(%esp),%gs
 
         # ES:EDI == Ring-0 stack activation
         leal UREGS_eip(%esp),%edi
===== xen/include/asm-x86/x86_32/asm_defns.h 1.10 vs edited =====
--- 1.10/xen/include/asm-x86/x86_32/asm_defns.h	2005-04-28 17:35:48 -04:00
+++ edited/xen/include/asm-x86/x86_32/asm_defns.h	2005-06-09 10:30:45 -04:00
@@ -19,10 +19,10 @@
         "jmp 3f;"                                                       \
         "2:testb $3,"STR(UREGS_cs)"(%esp);"                             \
         "jz 1f;"                                                        \
-        "movl %ds,"STR(UREGS_ds)"(%esp);"                               \
-        "movl %es,"STR(UREGS_es)"(%esp);"                               \
-        "movl %fs,"STR(UREGS_fs)"(%esp);"                               \
-        "movl %gs,"STR(UREGS_gs)"(%esp);"                               \
+        "mov %ds,"STR(UREGS_ds)"(%esp);"                               \
+        "mov %es,"STR(UREGS_es)"(%esp);"                               \
+        "mov %fs,"STR(UREGS_fs)"(%esp);"                               \
+        "mov %gs,"STR(UREGS_gs)"(%esp);"                               \
         "3:"
 
 #define SAVE_ALL_NOSEGREGS(_reg)                \
@@ -31,8 +31,8 @@
 
 #define SET_XEN_SEGMENTS(_reg)                                  \
         "movl $("STR(__HYPERVISOR_DS)"),%e"STR(_reg)"x;"        \
-        "movl %e"STR(_reg)"x,%ds;"                              \
-        "movl %e"STR(_reg)"x,%es;"
+        "mov %e"STR(_reg)"x,%ds;"                              \
+        "mov %e"STR(_reg)"x,%es;"
 
 #define SAVE_ALL(_reg)                          \
         __SAVE_ALL_PRE                          \
@@ -56,10 +56,10 @@
         jmp 3f;                                         \
         2:testb $3,UREGS_cs(%esp);                      \
         jz 1f;                                          \
-        movl %ds,UREGS_ds(%esp);                        \
-        movl %es,UREGS_es(%esp);                        \
-        movl %fs,UREGS_fs(%esp);                        \
-        movl %gs,UREGS_gs(%esp);                        \
+        mov %ds,UREGS_ds(%esp);                        \
+        mov %es,UREGS_es(%esp);                        \
+        mov %fs,UREGS_fs(%esp);                        \
+        mov %gs,UREGS_gs(%esp);                        \
         3:
 
 #define SAVE_ALL_NOSEGREGS(_reg)                \
@@ -68,8 +68,8 @@
 
 #define SET_XEN_SEGMENTS(_reg)                          \
         movl $(__HYPERVISOR_DS),%e ## _reg ## x;        \
-        movl %e ## _reg ## x,%ds;                       \
-        movl %e ## _reg ## x,%es;
+        mov %e ## _reg ## x,%ds;                       \
+        mov %e ## _reg ## x,%es;
 
 #define SAVE_ALL(_reg)                          \
         __SAVE_ALL_PRE                          \

-- 
Flávio Bruno Leitner <fbl@conectiva.com.br>
[0EA2 7F40 4CF4 1E63 4AF6  33C0 3E10 E205 F251 EDDA]

Re: Re: [Xen-users] gcc 4.0.0 and xen-devel

[Vincent Hanquez]

On Thu, Jun 09, 2005 at 02:45:11PM -0300, Flavio Leitner wrote:

> > does the following patch fix your problems ?
> 
> Part of it because there are more places. Thanks!
> 
> Well, now it breaks because of warnings.

Applied thanks.
please don't forget a Signed-off-by line next time.

-- 
Vincent Hanquez

Re: Re: [Xen-users] gcc 4.0.0 and xen-devel

[David F Barrera]

I am getting gcc 4.0 errors on xen-unstable, x86_64:

(The last changeset on the source that I am using was
ChangeSet 2005/06/10 10:34:30+01:00 kaf24@firebug.cl.cam.ac.uk)

gcc -pthread -fno-strict-aliasing -DNDEBUG -O2 -Wall -g -pipe -Wp,-
D_FORTIFY_SOURCE=2 -m64 -D_GNU_SOURCE -fPIC -m64 -fPIC -
I../../tools/python/xen/lowlevel/xu -I../../tools/libxc -
I../../tools/xenstore -I../../tools/xcs -Ixen/lowlevel/xs -
I/usr/include/python2.4 -c xen/lowlevel/xs/xs.c -o build/temp.linux-
x86_64-2.4/xen/lowlevel/xs/xs.o -fno-strict-aliasing -Wall -Werror
cc1: warnings being treated as errors
xen/lowlevel/xs/xs.c: In function ‘xspy_read’:
xen/lowlevel/xs/xs.c:96: warning: pointer targets in passing argument 3
of ‘xs_read’ differ in signedness
xen/lowlevel/xs/xs.c: In function ‘xspy_ls’:
xen/lowlevel/xs/xs.c:144: warning: pointer targets in passing argument 3
of ‘xs_directory’ differ in signedness
xen/lowlevel/xs/xs.c: In function ‘xspy_get_permissions’:
xen/lowlevel/xs/xs.c:193: warning: pointer targets in passing argument 3
of ‘xs_get_permissions’ differ in signedness
error: command 'gcc' failed with exit status 1
make[3]: *** [build] Error 1
make[3]: Leaving directory `/tmp/xen-unstable/tools/python'
make[2]: *** [install] Error 2
make[2]: Leaving directory `/tmp/xen-unstable/tools'
make[1]: *** [tools] Error 2
make[1]: Leaving directory `/tmp/xen-unstable'
make: *** [world] Error 2


On Fri, 2005-06-10 at 13:43 +0200, Vincent Hanquez wrote:
> On Thu, Jun 09, 2005 at 02:45:11PM -0300, Flavio Leitner wrote:
> 
> > > does the following patch fix your problems ?
> > 
> > Part of it because there are more places. Thanks!
> > 
> > Well, now it breaks because of warnings.
> 
> Applied thanks.
> please don't forget a Signed-off-by line next time.
> 
-- 
Regards,

David F Barrera
Linux Technology Center
Systems and Technology Group, IBM

"The wisest men follow their own direction. "
                                                        Euripides

Re: Re: [Xen-users] gcc 4.0.0 and xen-devel

[Vincent Hanquez]

On Fri, Jun 10, 2005 at 08:02:20AM -0500, David F Barrera wrote:
> I am getting gcc 4.0 errors on xen-unstable, x86_64:
> 
> (The last changeset on the source that I am using was
> ChangeSet 2005/06/10 10:34:30+01:00 kaf24@firebug.cl.cam.ac.uk)

thoses warning has been corrected already.
try to pull a new version.

-- 
Vincent Hanquez