Fix gdb 7.0 compilation on IRIX 5.3

Fix gdb 7.0 compilation on IRIX 5.3

[Rainer Orth]

I just tried to compile gdb 7.0 on IRIX 5.3, but failed initially:

/vol/src/gnu/gdb/gdb-7.0/gdb/dwarf2read.c: In function 'dwarf2_read_section':
/vol/src/gnu/gdb/gdb-7.0/gdb/dwarf2read.c:1402: error: 'MAP_FAILED' undeclared (first use in this function)
/vol/src/gnu/gdb/gdb-7.0/gdb/dwarf2read.c:1402: error: (Each undeclared identifier is reported only once
/vol/src/gnu/gdb/gdb-7.0/gdb/dwarf2read.c:1402: error: for each function it appears in.)
make[2]: *** [dwarf2read.o] Error 1

The following trivial patch fixes this and allowed the build to finish.
Unfortunately, the resulting gdb segfaults when run against itself or
some recent xgcc binary.  I'll investigate and report separately.

	Rainer

-- 
-----------------------------------------------------------------------------
Rainer Orth, Center for Biotechnology, Bielefeld University


2009-11-23  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>

	* dwarf2read.c [HAVE_MMAP] (MAP_FAILED): Define if missing.

===================================================================
RCS file: gdb/RCS/dwarf2read.c,v
retrieving revision 1.1
diff -up -r1.1 gdb/dwarf2read.c
--- gdb/dwarf2read.c	2009/09/15 16:09:32	1.1
+++ gdb/dwarf2read.c	2009/11/23 18:13:08
@@ -58,6 +58,9 @@
 #endif
 #ifdef HAVE_MMAP
 #include <sys/mman.h>
+#ifndef MAP_FAILED
+#define MAP_FAILED -1
+#endif
 #endif
 
 #if 0

Re: Fix gdb 7.0 compilation on IRIX 5.3

[Tom Tromey]

>>>>> "Rainer" == Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE> writes:

Rainer> 2009-11-23  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
Rainer> 	* dwarf2read.c [HAVE_MMAP] (MAP_FAILED): Define if missing.

Rainer> +#ifndef MAP_FAILED
Rainer> +#define MAP_FAILED -1
Rainer> +#endif

I would have expected `(void *) -1' here.  This is ok if it was
warning-free.  Or the other form is also ok.

thanks,
Tom

Re: Fix gdb 7.0 compilation on IRIX 5.3

[Rainer Orth]

Tom Tromey <tromey@redhat.com> writes:

>>>>>> "Rainer" == Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE> writes:
>
> Rainer> 2009-11-23  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
> Rainer> 	* dwarf2read.c [HAVE_MMAP] (MAP_FAILED): Define if missing.
>
> Rainer> +#ifndef MAP_FAILED
> Rainer> +#define MAP_FAILED -1
> Rainer> +#endif
>
> I would have expected `(void *) -1' here.  This is ok if it was
> warning-free.  Or the other form is also ok.

I didn't see the warning, and used the cast-free form since gcc had
both.  I notice now that they use casts to char * when MAP_FAILED is
used. 

Could you check this in for me?  I couldn't check out an rw cvs tree from
sourceware.org, although I should have rw access to src.

ro@manam 443 > cvs -d :ext:ro@sourceware.org:/cvs/src co gdb
Permission denied (publickey,gssapi-with-mic).
cvs [checkout aborted]: end of file from server (consult above messages if any)

Btw, the crashes I've observed were already known: PR gdb/10819.

     Rainer

-- 
-----------------------------------------------------------------------------
Rainer Orth, Center for Biotechnology, Bielefeld University


===================================================================
RCS file: gdb/RCS/dwarf2read.c,v
retrieving revision 1.1
diff -up -r1.1 gdb/dwarf2read.c
--- gdb/dwarf2read.c	2009/09/15 16:09:32	1.1
+++ gdb/dwarf2read.c	2009/11/23 20:36:20
@@ -58,6 +58,9 @@
 #endif
 #ifdef HAVE_MMAP
 #include <sys/mman.h>
+#ifndef MAP_FAILED
+#define MAP_FAILED ((void *) -1)
+#endif
 #endif
 
 #if 0

Re: Fix gdb 7.0 compilation on IRIX 5.3

[Tom Tromey]

>>>>> "Rainer" == Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE> writes:

Rainer> I didn't see the warning, and used the cast-free form since gcc had
Rainer> both.  I notice now that they use casts to char * when MAP_FAILED is
Rainer> used. 

Rainer> Could you check this in for me?

Sure, I'll do it soon.

Rainer> I couldn't check out an rw cvs tree from sourceware.org,
Rainer> although I should have rw access to src.
Rainer> ro@manam 443 > cvs -d :ext:ro@sourceware.org:/cvs/src co gdb
Rainer> Permission denied (publickey,gssapi-with-mic).
Rainer> cvs [checkout aborted]: end of file from server (consult above messages if any)

Peculiar.  I checked and you are in the right group.
Maybe you are using the wrong key or something.

Tom

Re: Fix gdb 7.0 compilation on IRIX 5.3

[Rainer Orth]

Tom Tromey <tromey@redhat.com> writes:

>>>>>> "Rainer" == Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE> writes:
>
> Rainer> I didn't see the warning, and used the cast-free form since gcc had
> Rainer> both.  I notice now that they use casts to char * when MAP_FAILED is
> Rainer> used. 
>
> Rainer> Could you check this in for me?
>
> Sure, I'll do it soon.

Excellent, thanks.

> Rainer> I couldn't check out an rw cvs tree from sourceware.org,
> Rainer> although I should have rw access to src.
> Rainer> ro@manam 443 > cvs -d :ext:ro@sourceware.org:/cvs/src co gdb
> Rainer> Permission denied (publickey,gssapi-with-mic).
> Rainer> cvs [checkout aborted]: end of file from server (consult above messages if any)
>
> Peculiar.  I checked and you are in the right group.
> Maybe you are using the wrong key or something.

Indeed: I'm still using a v1 key for gcc.gnu.org and sourceware.org, but
also have a v2 one.  When I configure ssh accordingly, a r/w checkout
works.

Thanks.
        Rainer

-- 
-----------------------------------------------------------------------------
Rainer Orth, Center for Biotechnology, Bielefeld University

Re: Fix gdb 7.0 compilation on IRIX 5.3

[Joel Brobecker]

> Could you check this in for me?  I couldn't check out an rw cvs tree from
> sourceware.org, although I should have rw access to src.

Someone beat me to it, but I confirm that this change has been checked
in HEAD.

How about the 7.0 branch?

-- 
Joel

Re: Fix gdb 7.0 compilation on IRIX 5.3

[Tom Tromey]

>>>>> "Joel" == Joel Brobecker <brobecker@adacore.com> writes:

>> Could you check this in for me?  I couldn't check out an rw cvs tree from
>> sourceware.org, although I should have rw access to src.

Joel> Someone beat me to it, but I confirm that this change has been checked
Joel> in HEAD.

Joel> How about the 7.0 branch?

I just committed it there as well.

Tom