Weird... 2.6.9 kills FC2 gcc

Weird... 2.6.9 kills FC2 gcc

[Jeff Garzik]

The following appears in 2.6.9 release kernel, building with stock FC2 
gcc on x86, but does not appear in 2.6.9-final:

>   AS      arch/i386/kernel/vsyscall.o
> cc1: internal compiler error: Segmentation fault
> Please submit a full bug report,
> with preprocessed source if appropriate.
> See <URL:http://bugzilla.redhat.com/bugzilla> for instructions.
> make[1]: *** [arch/i386/kernel/vsyscall.o] Error 1
> make: *** [arch/i386/kernel] Error 2



This is 100% reproducible, at the same location (vsyscall), which is 
strange because vsyscall didn't change AFAICS.

I'll build a gcc 3.4.2 without Fedora Core patches and see if the 
behavior persists.

But in the meantime, if anybody else knows what line of code causes this 
segfault, please speak up :)

	Jeff

Re: Weird... 2.6.9 kills FC2 gcc

[Mark Haverkamp]

On Mon, 2004-10-18 at 18:10, Jeff Garzik wrote:
> The following appears in 2.6.9 release kernel, building with stock FC2 
> gcc on x86, but does not appear in 2.6.9-final:
> 
> >   AS      arch/i386/kernel/vsyscall.o
> > cc1: internal compiler error: Segmentation fault
> > Please submit a full bug report,
> > with preprocessed source if appropriate.
> > See <URL:http://bugzilla.redhat.com/bugzilla> for instructions.
> > make[1]: *** [arch/i386/kernel/vsyscall.o] Error 1
> > make: *** [arch/i386/kernel] Error 2
> 
> 
> 
> This is 100% reproducible, at the same location (vsyscall), which is 
> strange because vsyscall didn't change AFAICS.
> 
> I'll build a gcc 3.4.2 without Fedora Core patches and see if the 
> behavior persists.
> 
> But in the meantime, if anybody else knows what line of code causes this 
> segfault, please speak up :)

As an experiment, I commented out the include of init.h and replaced the
__INITDATA and __FINIT with the .section and .previous.  It then
compiled OK.

Mark.


> 
> 	Jeff
> 
> 
> 
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
-- 
Mark Haverkamp <markh@osdl.org>

[PATCH] Re: Weird... 2.6.9 kills FC2 gcc

[Jeff Garzik]

[-- Attachment #1: Type: text/plain, Size: 1095 bytes --]

Mark Haverkamp wrote:
> On Mon, 2004-10-18 at 18:10, Jeff Garzik wrote:
> 
>>The following appears in 2.6.9 release kernel, building with stock FC2 
>>gcc on x86, but does not appear in 2.6.9-final:
>>
>>
>>>  AS      arch/i386/kernel/vsyscall.o
>>>cc1: internal compiler error: Segmentation fault
>>>Please submit a full bug report,
>>>with preprocessed source if appropriate.
>>>See <URL:http://bugzilla.redhat.com/bugzilla> for instructions.
>>>make[1]: *** [arch/i386/kernel/vsyscall.o] Error 1
>>>make: *** [arch/i386/kernel] Error 2
>>
>>
>>
>>This is 100% reproducible, at the same location (vsyscall), which is 
>>strange because vsyscall didn't change AFAICS.
>>
>>I'll build a gcc 3.4.2 without Fedora Core patches and see if the 
>>behavior persists.
>>
>>But in the meantime, if anybody else knows what line of code causes this 
>>segfault, please speak up :)
> 
> 
> As an experiment, I commented out the include of init.h and replaced the
> __INITDATA and __FINIT with the .section and .previous.  It then
> compiled OK.


Yep, the same fix (patch attached) works for me.

	Jeff



[-- Attachment #2: patch --]
[-- Type: text/plain, Size: 455 bytes --]

===== arch/i386/kernel/vsyscall.S 1.1 vs edited =====
--- 1.1/arch/i386/kernel/vsyscall.S	2003-04-22 22:14:57 -04:00
+++ edited/arch/i386/kernel/vsyscall.S	2004-10-18 22:20:32 -04:00
@@ -1,6 +1,5 @@
-#include <linux/init.h>
 
-__INITDATA
+.section	".init.data","aw"
 
 	.globl vsyscall_int80_start, vsyscall_int80_end
 vsyscall_int80_start:
@@ -12,4 +11,4 @@
 	.incbin "arch/i386/kernel/vsyscall-sysenter.so"
 vsyscall_sysenter_end:
 
-__FINIT
+.previous

Re: [PATCH] Re: Weird... 2.6.9 kills FC2 gcc

[Jeff Garzik]

More data points:

No problems at all on x86-64.

No ICE on 32-bit x86 gcc 3.4.2, with 2.6.9 release kernel.

So this ICE appears to be a bug specific to 3.3.x or perhaps Fedora.

	Jeff

Re: [PATCH] Re: Weird... 2.6.9 kills FC2 gcc

[Dmitry Torokhov]

On Monday 18 October 2004 10:31 pm, Jeff Garzik wrote:
> 
> More data points:
> 
> No problems at all on x86-64.
> 
> No ICE on 32-bit x86 gcc 3.4.2, with 2.6.9 release kernel.
> 
> So this ICE appears to be a bug specific to 3.3.x or perhaps Fedora.
> 
> 	Jeff
> 

For what it worth this is on mutated RH 8.0:

[dtor@core dtor]$ make
  CHK     include/linux/version.h
make[1]: `arch/i386/kernel/asm-offsets.s' is up to date.
  CHK     include/asm-i386/asm_offsets.h
  CHK     include/linux/compile.h
  AS      arch/i386/kernel/vsyscall.o
include/linux/compiler.h:20: warning: parameter name starts with a digit in #define
include/linux/compiler.h:20: badly punctuated parameter list in #define
make[1]: *** [arch/i386/kernel/vsyscall.o] Error 1
make: *** [arch/i386/kernel] Error 2
[dtor@core dtor]$ gcc -v
Reading specs from /usr/lib/gcc-lib/i386-redhat-linux/3.2.2/specs
Configured with: ../configure --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --enable-shared --enable-threads=posix --disable-checking --with-system-zlib --enable-__cxa_atexit --host=i386-redhat-linux
Thread model: posix
gcc version 3.2.2 20030222 (Red Hat Linux 3.2.2-5)


-- 
Dmitry

Re: [PATCH] Re: Weird... 2.6.9 kills FC2 gcc

[Meelis Roos]

JG> So this ICE appears to be a bug specific to 3.3.x or perhaps Fedora.

Debian gcc-3.3 3.3.5-1 is also affected.

-- 
Meelis Roos

Re: Weird... 2.6.9 kills FC2 gcc

[Gene Heskett]

On Monday 18 October 2004 21:10, Jeff Garzik wrote:
>The following appears in 2.6.9 release kernel, building with stock
> FC2
>
>gcc on x86, but does not appear in 2.6.9-final:
>>   AS      arch/i386/kernel/vsyscall.o
>> cc1: internal compiler error: Segmentation fault
>> Please submit a full bug report,
>> with preprocessed source if appropriate.
>> See <URL:http://bugzilla.redhat.com/bugzilla> for instructions.
>> make[1]: *** [arch/i386/kernel/vsyscall.o] Error 1
>> make: *** [arch/i386/kernel] Error 2
>
>This is 100% reproducible, at the same location (vsyscall), which is
>strange because vsyscall didn't change AFAICS.
>
>I'll build a gcc 3.4.2 without Fedora Core patches and see if the
>behavior persists.
>
>But in the meantime, if anybody else knows what line of code causes
> this segfault, please speak up :)
>
> Jeff

I'm an FC2, gcc-3.3.3 user, and it works here without that error, 
Jeff.

-- 
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.27% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com attorneys please note, additions to this message
by Gene Heskett are:
Copyright 2004 by Maurice Eugene Heskett, all rights reserved.

Re: [PATCH] Re: Weird... 2.6.9 kills FC2 gcc

[Mark Haverkamp]

On Mon, 2004-10-18 at 23:31 -0400, Jeff Garzik wrote:
> More data points:
> 
> No problems at all on x86-64.
> 
> No ICE on 32-bit x86 gcc 3.4.2, with 2.6.9 release kernel.
> 
> So this ICE appears to be a bug specific to 3.3.x or perhaps Fedora.
> 
> 	Jeff
> 

I tried building this on FC3 with a 3.4.2 gcc and it compiles OK.

Mark.


-- 
Mark Haverkamp <markh@osdl.org>

Re: [PATCH] Re: Weird... 2.6.9 kills FC2 gcc

[Jeff Garzik]

Mark Haverkamp wrote:
> On Mon, 2004-10-18 at 23:31 -0400, Jeff Garzik wrote:
> 
>>More data points:
>>
>>No problems at all on x86-64.
>>
>>No ICE on 32-bit x86 gcc 3.4.2, with 2.6.9 release kernel.
>>
>>So this ICE appears to be a bug specific to 3.3.x or perhaps Fedora.
>>
>>	Jeff
>>
> 
> 
> I tried building this on FC3 with a 3.4.2 gcc and it compiles OK.


Yeah, it looks like 3.3.x from FC2 and Debian both ICE, but 3.4.x (from 
any sources) is OK.

	Jeff