* AIX cross-compiler woes
@ 1999-04-27 4:28 Jo Dillon
1999-04-30 16:07 ` David Edelsohn
` (2 more replies)
0 siblings, 3 replies; 6+ messages in thread
From: Jo Dillon @ 1999-04-27 4:28 UTC (permalink / raw)
To: egcs
Hi - I have a cross-compiler built to create AIX/PowerPC/Coff binaries on
Intel/Linux. It's egcs 1.1.1 with the gpc patches applied, and it fails
on some (but not all) Pascal binaries, with results like this:
/usr/local/powerpc/bin/powerpc-ibm-aix3.2.5-gpc -c -w realreg.pas -o realreg.coff
realreg.pas: In function `Proxysearch':
realreg.pas:463: internal error--unrecognizable insn:
(insn 8383 8380 1439 (set (mem:SI (plus:SI (reg:SI 14 r14)
(const_int 65536)))
(reg:SI 17 r17)) -1 (nil)
(nil))
../../../egcs-1.1.1/gcc/p/../toplev.c:1378: Internal compiler error in function fatal_insn
The line in question is the end of the file. I had a few problems compiling
the cross-compiler; it seemed not to build a crt0.o, which meant that
make cross-install didn't work properly. I don't actually need a crt0.o
or a runtime library since I simply want to create XCoff .o's to include
into a Macintosh program, so I got gcc to install anyway; I was wondering
if this could be the cause of the above problem - perhaps the cross-gpc
is finding the wrong gcc backend or something? - but this seems unlikely
since most files do compile without problems and provide me with usable
XCoff binaries that I've tested on my Mac. Are there any known issues with
the AIX backend?
--
Jo
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: AIX cross-compiler woes
1999-04-27 4:28 AIX cross-compiler woes Jo Dillon
@ 1999-04-30 16:07 ` David Edelsohn
1999-04-30 23:15 ` David Edelsohn
1999-04-30 23:15 ` Jo Dillon
1999-05-23 23:44 ` Jeffrey A Law
2 siblings, 1 reply; 6+ messages in thread
From: David Edelsohn @ 1999-04-30 16:07 UTC (permalink / raw)
To: Jo Dillon; +Cc: egcs
>>>>> Jo Dillon writes:
Jo> Hi - I have a cross-compiler built to create AIX/PowerPC/Coff binaries on
Jo> Intel/Linux. It's egcs 1.1.1 with the gpc patches applied, and it fails
Jo> on some (but not all) Pascal binaries, with results like this:
Jo> /usr/local/powerpc/bin/powerpc-ibm-aix3.2.5-gpc -c -w realreg.pas -o realreg.coff
Jo> realreg.pas: In function `Proxysearch':
Jo> realreg.pas:463: internal error--unrecognizable insn:
Jo> (insn 8383 8380 1439 (set (mem:SI (plus:SI (reg:SI 14 r14)
Jo> (const_int 65536)))
Jo> (reg:SI 17 r17)) -1 (nil)
Jo> (nil))
Jo> ../../../egcs-1.1.1/gcc/p/../toplev.c:1378: Internal compiler error in function fatal_insn
The source line listed is the end of the file because this is a
problem occurring at a later compilation stage processing the entire
function.
This was a known problem which may be fixed in egcs-1.1.2.
David
^ permalink raw reply [flat|nested] 6+ messages in thread
* AIX cross-compiler woes
1999-04-27 4:28 AIX cross-compiler woes Jo Dillon
1999-04-30 16:07 ` David Edelsohn
@ 1999-04-30 23:15 ` Jo Dillon
1999-05-23 23:44 ` Jeffrey A Law
2 siblings, 0 replies; 6+ messages in thread
From: Jo Dillon @ 1999-04-30 23:15 UTC (permalink / raw)
To: egcs
Hi - I have a cross-compiler built to create AIX/PowerPC/Coff binaries on
Intel/Linux. It's egcs 1.1.1 with the gpc patches applied, and it fails
on some (but not all) Pascal binaries, with results like this:
/usr/local/powerpc/bin/powerpc-ibm-aix3.2.5-gpc -c -w realreg.pas -o realreg.coff
realreg.pas: In function `Proxysearch':
realreg.pas:463: internal error--unrecognizable insn:
(insn 8383 8380 1439 (set (mem:SI (plus:SI (reg:SI 14 r14)
(const_int 65536)))
(reg:SI 17 r17)) -1 (nil)
(nil))
../../../egcs-1.1.1/gcc/p/../toplev.c:1378: Internal compiler error in function fatal_insn
The line in question is the end of the file. I had a few problems compiling
the cross-compiler; it seemed not to build a crt0.o, which meant that
make cross-install didn't work properly. I don't actually need a crt0.o
or a runtime library since I simply want to create XCoff .o's to include
into a Macintosh program, so I got gcc to install anyway; I was wondering
if this could be the cause of the above problem - perhaps the cross-gpc
is finding the wrong gcc backend or something? - but this seems unlikely
since most files do compile without problems and provide me with usable
XCoff binaries that I've tested on my Mac. Are there any known issues with
the AIX backend?
--
Jo
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: AIX cross-compiler woes
1999-04-30 16:07 ` David Edelsohn
@ 1999-04-30 23:15 ` David Edelsohn
0 siblings, 0 replies; 6+ messages in thread
From: David Edelsohn @ 1999-04-30 23:15 UTC (permalink / raw)
To: Jo Dillon; +Cc: egcs
>>>>> Jo Dillon writes:
Jo> Hi - I have a cross-compiler built to create AIX/PowerPC/Coff binaries on
Jo> Intel/Linux. It's egcs 1.1.1 with the gpc patches applied, and it fails
Jo> on some (but not all) Pascal binaries, with results like this:
Jo> /usr/local/powerpc/bin/powerpc-ibm-aix3.2.5-gpc -c -w realreg.pas -o realreg.coff
Jo> realreg.pas: In function `Proxysearch':
Jo> realreg.pas:463: internal error--unrecognizable insn:
Jo> (insn 8383 8380 1439 (set (mem:SI (plus:SI (reg:SI 14 r14)
Jo> (const_int 65536)))
Jo> (reg:SI 17 r17)) -1 (nil)
Jo> (nil))
Jo> ../../../egcs-1.1.1/gcc/p/../toplev.c:1378: Internal compiler error in function fatal_insn
The source line listed is the end of the file because this is a
problem occurring at a later compilation stage processing the entire
function.
This was a known problem which may be fixed in egcs-1.1.2.
David
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: AIX cross-compiler woes
1999-04-27 4:28 AIX cross-compiler woes Jo Dillon
1999-04-30 16:07 ` David Edelsohn
1999-04-30 23:15 ` Jo Dillon
@ 1999-05-23 23:44 ` Jeffrey A Law
1999-05-31 21:36 ` Jeffrey A Law
2 siblings, 1 reply; 6+ messages in thread
From: Jeffrey A Law @ 1999-05-23 23:44 UTC (permalink / raw)
To: Jo Dillon; +Cc: egcs
In message <19990427122827.A14570@thelonious.new.ox.ac.uk>you write:
> Hi - I have a cross-compiler built to create AIX/PowerPC/Coff binaries on
> Intel/Linux. It's egcs 1.1.1 with the gpc patches applied, and it fails
> on some (but not all) Pascal binaries, with results like this:
>
> /usr/local/powerpc/bin/powerpc-ibm-aix3.2.5-gpc -c -w realreg.pas -o realre
> g.coff
> realreg.pas: In function `Proxysearch':
> realreg.pas:463: internal error--unrecognizable insn:
> (insn 8383 8380 1439 (set (mem:SI (plus:SI (reg:SI 14 r14)
> (const_int 65536)))
> (reg:SI 17 r17)) -1 (nil)
> (nil))
> ../../../egcs-1.1.1/gcc/p/../toplev.c:1378: Internal compiler error in func
> tion fatal_insn
I believe this bug was fixed in egcs-1.1.2.
jeff
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: AIX cross-compiler woes
1999-05-23 23:44 ` Jeffrey A Law
@ 1999-05-31 21:36 ` Jeffrey A Law
0 siblings, 0 replies; 6+ messages in thread
From: Jeffrey A Law @ 1999-05-31 21:36 UTC (permalink / raw)
To: Jo Dillon; +Cc: egcs
In message <19990427122827.A14570@thelonious.new.ox.ac.uk>you write:
> Hi - I have a cross-compiler built to create AIX/PowerPC/Coff binaries on
> Intel/Linux. It's egcs 1.1.1 with the gpc patches applied, and it fails
> on some (but not all) Pascal binaries, with results like this:
>
> /usr/local/powerpc/bin/powerpc-ibm-aix3.2.5-gpc -c -w realreg.pas -o realre
> g.coff
> realreg.pas: In function `Proxysearch':
> realreg.pas:463: internal error--unrecognizable insn:
> (insn 8383 8380 1439 (set (mem:SI (plus:SI (reg:SI 14 r14)
> (const_int 65536)))
> (reg:SI 17 r17)) -1 (nil)
> (nil))
> ../../../egcs-1.1.1/gcc/p/../toplev.c:1378: Internal compiler error in func
> tion fatal_insn
I believe this bug was fixed in egcs-1.1.2.
jeff
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~1999-05-31 21:36 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
1999-04-27 4:28 AIX cross-compiler woes Jo Dillon
1999-04-30 16:07 ` David Edelsohn
1999-04-30 23:15 ` David Edelsohn
1999-04-30 23:15 ` Jo Dillon
1999-05-23 23:44 ` Jeffrey A Law
1999-05-31 21:36 ` Jeffrey A Law
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).