The situation did not change during the last two months, even with
the latest GCC imports.  beast.FreeBSD.org still blows up attempting
to build world and kernel for i386:

: --------------------------------------------------------------
: >>> stage 4: building everything..
: --------------------------------------------------------------
: cd /j/ru/src-5; MAKEOBJDIRPREFIX=/j/ru/obj/i386  MACHINE_ARCH=i386  MACHINE=i386  
:CPUTYPE=  GROFF_BIN_PATH=/j/ru/obj/i386/j/ru/src-5/alpha/usr/bin  
:GROFF_FONT_PATH=/j/ru/obj/i386/j/ru/src-5/alpha/usr/share/groff_font  
:GROFF_TMAC_PATH=/j/ru/obj/i386/j/ru/src-5/alpha/usr/share/tmac  
:DESTDIR=/j/ru/obj/i386/j/ru/src-5/alpha  INSTALL="sh /j/ru/src-5/tools/install.sh"  
:PATH=/j/ru/obj/i386/j/ru/src-5/alpha/usr/sbin:/j/ru/obj/i386/j/ru/src-5/alpha/usr/bin:/j/ru/obj/i386/j/ru/src-5/alpha/usr/games:/sbin:/bin:/usr/sbin:/usr/bin
: make -f Makefile.inc1 all
[...]
: ===> usr.bin/getconf
: cc -O -pipe -mcpu=pentiumpro -I/j/ru/src-5/usr.bin/getconf    -c limits.c
: {standard input}: Assembler messages:
: {standard input}:334: Error: bignum invalid
: *** Error code 1
: 
: Stop in /j/ru/src-5/usr.bin/getconf.
: *** Error code 1

The difference between the native i386 assembler output and the output
from the i386 cross-compiler for the limits.c is as follows:

--- limits.s (alpha->i386)      Sun Oct 13 15:57:20 2002
+++ limits.s (i386 native)      Sun Oct 13 16:02:15 2002
@@ -331,10 +331,10 @@
        .quad   -2147483648
        .long   1
        .long   .LC55
-       .quad   -9223372036854775808
+       .quad   0x8000000000000000
        .long   1
        .long   .LC56
-       .quad   9223372036854775807
+       .quad   0x7fffffffffffffff
        .long   1
        .long   .LC57
        .quad   0
@@ -367,13 +367,13 @@
        .quad   255
        .long   1
        .long   .LC67
-       .quad   4294967295
+       .quad   0xffffffff
        .long   1
        .long   .LC68
        .quad   -1
        .long   1
        .long   .LC69
-       .quad   4294967295
+       .quad   0xffffffff
        .long   1
        .long   .LC70
        .quad   65535

: --------------------------------------------------------------
: >>> Kernel build for GENERIC started on Sun Oct 13 05:25:16 PDT 2002
: --------------------------------------------------------------
: ===> GENERIC
[...]
: cd /j/ru/obj/i386/j/ru/src-5/sys/GENERIC;  MAKEOBJDIRPREFIX=/j/ru/obj/i386  
:MACHINE_ARCH=i386  MACHINE=i386  CPUTYPE=  
:GROFF_BIN_PATH=/j/ru/obj/i386/j/ru/src-5/alpha/usr/bin  
:GROFF_FONT_PATH=/j/ru/obj/i386/j/ru/src-5/alpha/usr/share/groff_font  
:GROFF_TMAC_PATH=/j/ru/obj/i386/j/ru/src-5/alpha/usr/share/tmac  
:DESTDIR=/j/ru/obj/i386/j/ru/src-5/alpha  INSTALL="sh /j/ru/src-5/tools/install.sh"  
:PATH=/j/ru/obj/i386/j/ru/src-5/alpha/usr/sbin:/j/ru/obj/i386/j/ru/src-5/alpha/usr/bin:/j/ru/obj/i386/j/ru/src-5/alpha/usr/games:/sbin:/bin:/usr/sbin:/usr/bin
: make KERNEL=kernel all
[...]
: cc -c -O -pipe -mcpu=pentiumpro -Wall -Wredundant-decls -Wnested-externs 
:-Wstrict-prototypes  -Wmissing-prototypes -Wpointer-arith -Winline -Wcast-qual  
:-fformat-extensions -ansi -g -nostdinc -I-  -I. -I/j/ru/src-5/sys 
:-I/j/ru/src-5/sys/dev -I/j/ru/src-5/sys/contrib/dev/acpica 
:-I/j/ru/src-5/sys/contrib/ipfilter -D_KERNEL -include opt_global.h -fno-common  
:-mno-align-long-strings -mpreferred-stack-boundary=2 -ffreestanding -Werror  
:/j/ru/src-5/sys/dev/advansys/adw_pci.c
: {standard input}: Assembler messages:
: {standard input}:22: Error: bignum invalid
: {standard input}:28: Error: bignum invalid
: *** Error code 1
: 
: Stop in /j/ru/obj/i386/j/ru/src-5/sys/GENERIC.
: *** Error code 1


On Wed, Aug 07, 2002 at 06:04:11PM +0300, Ruslan Ermilov wrote:
> On Wed, Aug 07, 2002 at 01:51:44PM +0300, Ruslan Ermilov wrote:
> > On Wed, Aug 07, 2002 at 12:16:02PM +0300, Ruslan Ermilov wrote:
> > [...]
> > > Wilko,
> > > 
> > > I am unable to cross-build i386 on alpha due to this bug in gcc.
> > > Native compiler has no problem building the generated cgram.c,
> > > and I have verified that native and cross compilers produce the
> > > same cgram.c output.  Perhaps, David could shed some light on
> > > this, and even probably fix it:
> > > 
> > > %%%
> > > --------------------------------------------------------------
> > > >>> stage 4: building everything..
> > > --------------------------------------------------------------
> > > cd /usr/src; MAKEOBJDIRPREFIX=/usr/obj/i386  MACHINE_ARCH=i386  MACHINE=i386  
>CPUTYPE=  OBJFORMAT_PATH=/usr/obj/i386/usr/src/alpha/usr/libexec  
>GROFF_BIN_PATH=/usr/obj/i386/usr/src/alpha/usr/bin  
>GROFF_FONT_PATH=/usr/obj/i386/usr/src/alpha/usr/share/groff_font  
>GROFF_TMAC_PATH=/usr/obj/i386/usr/src/alpha/usr/share/tmac  
>DESTDIR=/usr/obj/i386/usr/src/alpha  INSTALL="sh /usr/src/tools/install.sh"  
>PATH=/usr/obj/i386/usr/src/alpha/usr/sbin:/usr/obj/i386/usr/src/alpha/usr/bin:/usr/obj/i386/usr/src/alpha/usr/games:/sbin:/bin:/usr/sbin:/usr/bin
> make -f Makefile.inc1 all
> > > ===> share/info
> > > [...]
> > > ===> usr.bin/xlint/lint1
> > > cc -O -pipe -mcpu=pentiumpro -I. -I/usr/src/usr.bin/xlint/lint1 
>-I/usr/src/usr.bin/xlint/lint1/../arch/i386 -I/usr/src/usr.bin/xlint/lint1/../common  
>  -c cgram.c
> > > cc -O -pipe -mcpu=pentiumpro -I. -I/usr/src/usr.bin/xlint/lint1 
>-I/usr/src/usr.bin/xlint/lint1/../arch/i386 -I/usr/src/usr.bin/xlint/lint1/../common  
>  -c scan.c
> > > gzip -cn lint.7 > lint.7.gz
> > > groff -Tascii -mtty-char -man -t lint.7 | gzip -cn > lint.7.cat.gz
> > > /usr/src/usr.bin/xlint/lint1/cgram.y: In function `toicon':
> > > /usr/src/usr.bin/xlint/lint1/cgram.y:1711: unrecognizable insn:
> > > (insn 247 219 248 (set (reg:CC 17 flags)
> > >         (compare:CC (subreg:SI (reg/v:DI 75) 0)
> > >             (const_int 2147483648 [0x80000000]))) -1 (nil)
> > >     (expr_list:REG_DEAD (reg/v:DI 75)
> > >         (nil)))
> > > /usr/src/usr.bin/xlint/lint1/cgram.y:1711: Internal compiler error in 
>extract_insn, at recog.c:2150
> > > Please submit a full bug report,
> > > with preprocessed source if appropriate.
> > > See <URL:http://www.gnu.org/software/gcc/bugs.html> for instructions.
> > > *** Error code 1
> > > 1 error
> > > %%%
> > > 
> > > Meanwhile, I will temporary remove usr.bin/xlint from the build, and
> > > resume the i386 release building on ds10.
> > > 
> > While attempting to cross-compile kernel, it bombs out with, attempting
> > to build `random' module like this:
> > 
> > ===> random
> > cc -O -pipe -mcpu=pentiumpro -I/usr/src/sys/modules/random/../..  -D_KERNEL -Wall 
>-Wredundant-decls -Wnested-externs -Wstrict-prototypes  -Wmissing-prototypes 
>-Wpointer-arith -Winline -Wcast-qual  -Wno-format -ansi -DKLD_MODULE -nostdinc -I-  
>-I/usr/src/sys/modules/random/../.. -I. -I@ -I@/dev -I@/../include -fno-common -g 
>-mpreferred-stack-boundary=2 -ffreestanding -Wall -Wredundant-decls -Wnested-externs 
>-Wstrict-prototypes  -Wmissing-prototypes -Wpointer-arith -Winline -Wcast-qual  
>-Wno-format -ansi -c /usr/src/sys/crypto/sha2/sha2.c
> > {standard input}: Assembler messages:
> > {standard input}:97: Error: bignum invalid
> > {standard input}:98: Error: bignum invalid
> > {standard input}:101: Error: bignum invalid
> > 
> > Excluding it as well...  :(
> > 
> I see the same symptoms on beast.FreeBSD.org running fresh 5.0-CURRENT,
> so the problem is not unique to your DS10 host.
> 
> Excluding "random" module did not help; buildkernel died similarly
> later:
> 
> : cc -c -O -pipe -mcpu=pentiumpro -Wall -Wredundant-decls -Wnested-externs 
>-Wstrict-prototypes  -Wmissing-prototypes -Wpointer-arith -Winline -Wcast-qual  
>-Wno-format -ansi -g -nostdinc -I-  -I. -I/usr/src/sys -I/usr/src/sys/dev 
>-I/usr/src/sys/contrib/dev/acpica -I/usr/src/sys/contrib/ipfilter 
>-I/usr/src/sys/../include -D_KERNEL -include opt_global.h -fno-common   
>-mpreferred-stack-boundary=2 -ffreestanding -Werror  
>/usr/src/sys/dev/advansys/adw_pci.c
> : {standard input}: Assembler messages:
> : {standard input}:24: Error: bignum invalid
> : {standard input}:30: Error: bignum invalid
> : *** Error code 1
> : Stop in /usr/obj/i386/usr/src/sys/GENERIC.
> : *** Error code 1
> 
> David, can you please look into this?



-- 
Ruslan Ermilov          Sysadmin and DBA,
[EMAIL PROTECTED]           Sunbay Software AG,
[EMAIL PROTECTED]          FreeBSD committer,
+380.652.512.251        Simferopol, Ukraine

http://www.FreeBSD.org  The Power To Serve
http://www.oracle.com   Enabling The Information Age

Attachment: msg44558/pgp00000.pgp
Description: PGP signature

Reply via email to