* ia64 target?
@ 2001-03-21 15:59 Andrew Cagney
2001-03-21 15:59 ` Kevin Buettner
0 siblings, 1 reply; 13+ messages in thread
From: Andrew Cagney @ 2001-03-21 15:59 UTC (permalink / raw)
To: GDB Discussion
Hello,
What is a good ia64 target? The maintainer file currently notes:
ia64 (--target=ia64-elf broken)
if I try to configure that, nothing interesting happens (the GDB
directory didn't appear to even be created).
Could someone suggest a better generic ia64 target that can be used to
build a cross debug ia64 GDB.
enjoy,
Andrew
^ permalink raw reply [flat|nested] 13+ messages in thread* Re: ia64 target?
2001-03-21 15:59 ia64 target? Andrew Cagney
@ 2001-03-21 15:59 ` Kevin Buettner
2001-03-21 15:59 ` J.T. Conklin
2001-03-21 15:59 ` Andrew Cagney
0 siblings, 2 replies; 13+ messages in thread
From: Kevin Buettner @ 2001-03-21 15:59 UTC (permalink / raw)
To: Andrew Cagney, GDB Discussion
On Mar 13, 5:51pm, Andrew Cagney wrote:
> What is a good ia64 target? The maintainer file currently notes:
>
> ia64 (--target=ia64-elf broken)
>
> if I try to configure that, nothing interesting happens (the GDB
> directory didn't appear to even be created).
>
> Could someone suggest a better generic ia64 target that can be used to
> build a cross debug ia64 GDB.
How about...
--target=ia64-unknown-linux
I've successfully done a configure and make with
--target=ia64-unknown-linux on the following hosts:
i386-unknown-freebsd4.2
i386-unknown-netbsdelf1.5
i386-pc-solaris2.8
i686-pc-linux-gnu
ia64-unknown-linux
Of course, this only takes care of the build portion. If you want to
(somehow) test the resulting gdb, you need to have access to an IA-64
machine running Linux or a simulator (running Linux).
See http://www.software.hp.com/products/LIA64/overview4a.htm for info
on an IA-64 simulator. (I did much of the initial development with an
i686-pc-linux-gnu cross ia64-unknown-linux GDB using HP's ski
simulator. It is able to boot Linux and runs amazingly well.)
It may also be possible to use HP's NUE environment to do the testing
I have not tried this.
Kevin
^ permalink raw reply [flat|nested] 13+ messages in thread* Re: ia64 target?
2001-03-21 15:59 ` Kevin Buettner
@ 2001-03-21 15:59 ` J.T. Conklin
2001-03-21 15:59 ` Andrew Cagney
2001-03-21 15:59 ` Kevin Buettner
2001-03-21 15:59 ` Andrew Cagney
1 sibling, 2 replies; 13+ messages in thread
From: J.T. Conklin @ 2001-03-21 15:59 UTC (permalink / raw)
To: Kevin Buettner; +Cc: Andrew Cagney, GDB Discussion
>>>>> "Kevin" == Kevin Buettner <kevinb@cygnus.com> writes:
>> What is a good ia64 target? The maintainer file currently notes:
>>
>> ia64 (--target=ia64-elf broken)
>>
>> if I try to configure that, nothing interesting happens (the GDB
>> directory didn't appear to even be created).
>>
>> Could someone suggest a better generic ia64 target that can be used to
>> build a cross debug ia64 GDB.
Kevin> How about...
Kevin>
Kevin> --target=ia64-unknown-linux
Kevin>
Kevin> I've successfully done a configure and make with
Kevin> --target=ia64-unknown-linux on the following hosts:
Kevin>
Kevin> i386-unknown-freebsd4.2
Kevin> i386-unknown-netbsdelf1.5
Kevin> i386-pc-solaris2.8
Kevin> i686-pc-linux-gnu
Kevin> ia64-unknown-linux
Note that all of these hosts use ELF/svr4 shared libraries. On
others, it fails to build. I believe all of the Linux targets suffer
from this bug.
That might be considered an OK limitation for Linux targets. But if
we're going to have a config listed in the MAINTAINERS list so we can
use it to verify that global changes don't break builds, we have to
ensure that those configs build on all hosts.
--jtc
--
J.T. Conklin
RedBack Networks
^ permalink raw reply [flat|nested] 13+ messages in thread* Re: ia64 target?
2001-03-21 15:59 ` J.T. Conklin
@ 2001-03-21 15:59 ` Andrew Cagney
2001-03-21 15:59 ` J.T. Conklin
2001-03-21 15:59 ` Kevin Buettner
1 sibling, 1 reply; 13+ messages in thread
From: Andrew Cagney @ 2001-03-21 15:59 UTC (permalink / raw)
To: jtc; +Cc: Kevin Buettner, GDB Discussion
"J.T. Conklin" wrote:
> That might be considered an OK limitation for Linux targets. But if
> we're going to have a config listed in the MAINTAINERS list so we can
> use it to verify that global changes don't break builds, we have to
> ensure that those configs build on all hosts.
Keven recently fixed this - you really can cross compile shared library
support for linux targets (I'm pretty sure I double checked a !linux X
ia64-linux before updating this).
Andrew
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: ia64 target?
2001-03-21 15:59 ` Andrew Cagney
@ 2001-03-21 15:59 ` J.T. Conklin
2001-03-21 15:59 ` Andrew Cagney
0 siblings, 1 reply; 13+ messages in thread
From: J.T. Conklin @ 2001-03-21 15:59 UTC (permalink / raw)
To: Andrew Cagney; +Cc: Kevin Buettner, GDB Discussion
>>>>> "Andrew" == Andrew Cagney <ac131313@cygnus.com> writes:
>> That might be considered an OK limitation for Linux targets. But if
>> we're going to have a config listed in the MAINTAINERS list so we can
>> use it to verify that global changes don't break builds, we have to
>> ensure that those configs build on all hosts.
Andrew> Keven recently fixed this - you really can cross compile shared library
Andrew> support for linux targets (I'm pretty sure I double checked a !linux X
Andrew> ia64-linux before updating this).
Ummm, I wouldn't have sent a message if the ia64-linux target wasn't
failing on my nightly builds (i386-netbsd1.3 host).
--jtc
gcc -c -g -O2 -I. -I../../gdb -I../../gdb/config -DHAVE_CONFIG_H -I../../gdb/../include/opcode -I../../gdb/../readline/.. -I../bfd -I../../gdb/../bfd -I../../gdb/../include -I../intl -I../../gdb/../intl -DUI_OUT=1 -Wimplicit -Wreturn-type -Wcomment -Wtrigraphs -Wformat -Wparentheses -Wpointer-arith ../../gdb/solib-legacy.c
In file included from ../../gdb/solib-legacy.c:28:
/usr/include/link.h:85: field `nlist' has incomplete type
../../gdb/solib-legacy.c: In function `legacy_svr4_fetch_link_map_offsets':
../../gdb/solib-legacy.c:53: sizeof applied to an incomplete type
../../gdb/solib-legacy.c:55: dereferencing pointer to incomplete type
../../gdb/solib-legacy.c:56: dereferencing pointer to incomplete type
../../gdb/solib-legacy.c:58: sizeof applied to an incomplete type
../../gdb/solib-legacy.c:60: dereferencing pointer to incomplete type
../../gdb/solib-legacy.c:61: dereferencing pointer to incomplete type
../../gdb/solib-legacy.c:63: dereferencing pointer to incomplete type
../../gdb/solib-legacy.c:64: dereferencing pointer to incomplete type
--
J.T. Conklin
RedBack Networks
^ permalink raw reply [flat|nested] 13+ messages in thread* Re: ia64 target?
2001-03-21 15:59 ` J.T. Conklin
@ 2001-03-21 15:59 ` Andrew Cagney
2001-03-21 15:59 ` Kevin Buettner
0 siblings, 1 reply; 13+ messages in thread
From: Andrew Cagney @ 2001-03-21 15:59 UTC (permalink / raw)
To: jtc; +Cc: Kevin Buettner, GDB Discussion
"J.T. Conklin" wrote:
>
> >>>>> "Andrew" == Andrew Cagney <ac131313@cygnus.com> writes:
> >> That might be considered an OK limitation for Linux targets. But if
> >> we're going to have a config listed in the MAINTAINERS list so we can
> >> use it to verify that global changes don't break builds, we have to
> >> ensure that those configs build on all hosts.
>
> Andrew> Keven recently fixed this - you really can cross compile shared library
> Andrew> support for linux targets (I'm pretty sure I double checked a !linux X
> Andrew> ia64-linux before updating this).
>
> Ummm, I wouldn't have sent a message if the ia64-linux target wasn't
> failing on my nightly builds (i386-netbsd1.3 host).
Puzzled expression :-)
Just put it back to broken for now and then Kevin and/or I will figure
out what happened.
Andrew
> gcc -c -g -O2 -I. -I../../gdb -I../../gdb/config -DHAVE_CONFIG_H -I../../gdb/../include/opcode -I../../gdb/../readline/.. -I../bfd -I../../gdb/../bfd -I../../gdb/../include -I../intl -I../../gdb/../intl -DUI_OUT=1 -Wimplicit -Wreturn-type -Wcomment -Wtrigraphs -Wformat -Wparentheses -Wpointer-arith ../../gdb/solib-legacy.c
> In file included from ../../gdb/solib-legacy.c:28:
> /usr/include/link.h:85: field `nlist' has incomplete type
> ../../gdb/solib-legacy.c: In function `legacy_svr4_fetch_link_map_offsets':
> ../../gdb/solib-legacy.c:53: sizeof applied to an incomplete type
> ../../gdb/solib-legacy.c:55: dereferencing pointer to incomplete type
> ../../gdb/solib-legacy.c:56: dereferencing pointer to incomplete type
> ../../gdb/solib-legacy.c:58: sizeof applied to an incomplete type
> ../../gdb/solib-legacy.c:60: dereferencing pointer to incomplete type
> ../../gdb/solib-legacy.c:61: dereferencing pointer to incomplete type
> ../../gdb/solib-legacy.c:63: dereferencing pointer to incomplete type
> ../../gdb/solib-legacy.c:64: dereferencing pointer to incomplete type
>
> --
> J.T. Conklin
> RedBack Networks
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: ia64 target?
2001-03-21 15:59 ` Andrew Cagney
@ 2001-03-21 15:59 ` Kevin Buettner
2001-03-21 15:59 ` J.T. Conklin
0 siblings, 1 reply; 13+ messages in thread
From: Kevin Buettner @ 2001-03-21 15:59 UTC (permalink / raw)
To: Andrew Cagney, jtc; +Cc: Kevin Buettner, GDB Discussion
On Mar 16, 8:09pm, Andrew Cagney wrote:
> > Ummm, I wouldn't have sent a message if the ia64-linux target wasn't
> > failing on my nightly builds (i386-netbsd1.3 host).
>
> Puzzled expression :-)
>
> Just put it back to broken for now and then Kevin and/or I will figure
> out what happened.
I was able to reproduce J.T.'s problem on NetBSD 1.4 and have fixed it.
See http://sources.redhat.com/ml/gdb-patches/2001-03/msg00341.html
for the details. (I would appreciate it though if J.T. would retest
on his NetBSD 1.3 box.)
Kevin
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: ia64 target?
2001-03-21 15:59 ` Kevin Buettner
@ 2001-03-21 15:59 ` J.T. Conklin
2001-03-21 15:59 ` Kevin Buettner
0 siblings, 1 reply; 13+ messages in thread
From: J.T. Conklin @ 2001-03-21 15:59 UTC (permalink / raw)
To: Kevin Buettner; +Cc: GDB Discussion
>>>>> "Kevin" == Kevin Buettner <kevinb@cygnus.com> writes:
>> > Ummm, I wouldn't have sent a message if the ia64-linux target wasn't
>> > failing on my nightly builds (i386-netbsd1.3 host).
>>
>> Puzzled expression :-)
>>
>> Just put it back to broken for now and then Kevin and/or I will figure
>> out what happened.
Kevin> I was able to reproduce J.T.'s problem on NetBSD 1.4 and have
Kevin> fixed it. See
Kevin> http://sources.redhat.com/ml/gdb-patches/2001-03/msg00341.html
Kevin> for the details. (I would appreciate it though if J.T. would
Kevin> retest on his NetBSD 1.3 box.)
It's closer.
gcc -c -g -O2 -I. -I../../gdb -I../../gdb/config -DHAVE_CONFIG_H -I../../gdb/../include/opcode -I../../gdb/../readline/.. -I../bfd -I../../gdb/../bfd -I../../gdb/../include -I../intl -I../../gdb/../intl -DUI_OUT=1 -Wimplicit -Wreturn-type -Wcomment -Wtrigraphs -Wformat -Wparentheses -Wpointer-arith ../../gdb/solib-legacy.c
In file included from ../../gdb/solib-legacy.c:28:
/usr/include/link.h:85: field `nlist' has incomplete type
It appears that 1.3 requires nlist.h to be included before link.h.
--jtc
--
J.T. Conklin
RedBack Networks
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: ia64 target?
2001-03-21 15:59 ` J.T. Conklin
2001-03-21 15:59 ` Andrew Cagney
@ 2001-03-21 15:59 ` Kevin Buettner
2001-03-21 15:59 ` J.T. Conklin
1 sibling, 1 reply; 13+ messages in thread
From: Kevin Buettner @ 2001-03-21 15:59 UTC (permalink / raw)
To: jtc; +Cc: Kevin Buettner, Andrew Cagney, GDB Discussion
On Mar 16, 2:32pm, J.T. Conklin wrote:
> Kevin> How about...
> Kevin>
> Kevin> --target=ia64-unknown-linux
> Kevin>
> Kevin> I've successfully done a configure and make with
> Kevin> --target=ia64-unknown-linux on the following hosts:
> Kevin>
> Kevin> i386-unknown-freebsd4.2
> Kevin> i386-unknown-netbsdelf1.5
> Kevin> i386-pc-solaris2.8
> Kevin> i686-pc-linux-gnu
> Kevin> ia64-unknown-linux
>
> Note that all of these hosts use ELF/svr4 shared libraries. On
> others, it fails to build. I believe all of the Linux targets suffer
> from this bug.
Are you sure?
I've just checked and solib-svr4.o (and solib-legacy.o) appear in
the TDEPFILES list in each of the linux.mt files. This means that
shared library support will be compiled in for the target.
The thing that's missing to make this actual usable on IA-64
is the definition of a suitable link_map_offsets fetcher. But
even without this, I think it should still build.
Which host OS are you trying this with? (If it's something that runs
on i386, I'd like to fetch a copy and try it out for myself.)
Kevin
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: ia64 target?
2001-03-21 15:59 ` Kevin Buettner
@ 2001-03-21 15:59 ` J.T. Conklin
0 siblings, 0 replies; 13+ messages in thread
From: J.T. Conklin @ 2001-03-21 15:59 UTC (permalink / raw)
To: Kevin Buettner; +Cc: Andrew Cagney, GDB Discussion
>>>>> "Kevin" == Kevin Buettner <kevinb@cygnus.com> writes:
Kevin> On Mar 16, 2:32pm, J.T. Conklin wrote:
Kevin> How about...
Kevin>
Kevin> --target=ia64-unknown-linux
Kevin>
Kevin> I've successfully done a configure and make with
Kevin> --target=ia64-unknown-linux on the following hosts:
Kevin>
Kevin> i386-unknown-freebsd4.2
Kevin> i386-unknown-netbsdelf1.5
Kevin> i386-pc-solaris2.8
Kevin> i686-pc-linux-gnu
Kevin> ia64-unknown-linux
>>
>> Note that all of these hosts use ELF/svr4 shared libraries. On
>> others, it fails to build. I believe all of the Linux targets suffer
>> from this bug.
Kevin> Are you sure?
Yes.
Both ia64-linux and i386-linux fail to build on a i386-netbsd1.3 host.
I suspect these targets would also fail to build on on sparc-sunos, or
any other host without ELF/svr4 shared libraries because solib-legacy.c
includes the host's "link.h" header.
--jtc
--
J.T. Conklin
RedBack Networks
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: ia64 target?
2001-03-21 15:59 ` Kevin Buettner
2001-03-21 15:59 ` J.T. Conklin
@ 2001-03-21 15:59 ` Andrew Cagney
1 sibling, 0 replies; 13+ messages in thread
From: Andrew Cagney @ 2001-03-21 15:59 UTC (permalink / raw)
To: Kevin Buettner; +Cc: GDB Discussion
Kevin Buettner wrote:
>
> On Mar 13, 5:51pm, Andrew Cagney wrote:
>
> > What is a good ia64 target? The maintainer file currently notes:
> >
> > ia64 (--target=ia64-elf broken)
> >
> > if I try to configure that, nothing interesting happens (the GDB
> > directory didn't appear to even be created).
> >
> > Could someone suggest a better generic ia64 target that can be used to
> > build a cross debug ia64 GDB.
>
> How about...
>
> --target=ia64-unknown-linux
I'll take this one, thanks.
> Of course, this only takes care of the build portion. If you want to
> (somehow) test the resulting gdb, you need to have access to an IA-64
> machine running Linux or a simulator (running Linux).
I don't think this is necessary. It is for someone making purely
mechanical changes (eg moving a function declaration, deleting a global
variable).
Andrew
^ permalink raw reply [flat|nested] 13+ messages in thread
end of thread, other threads:[~2001-03-21 15:59 UTC | newest]
Thread overview: 13+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2001-03-21 15:59 ia64 target? Andrew Cagney
2001-03-21 15:59 ` Kevin Buettner
2001-03-21 15:59 ` J.T. Conklin
2001-03-21 15:59 ` Andrew Cagney
2001-03-21 15:59 ` J.T. Conklin
2001-03-21 15:59 ` Andrew Cagney
2001-03-21 15:59 ` Kevin Buettner
2001-03-21 15:59 ` J.T. Conklin
2001-03-21 15:59 ` Kevin Buettner
2001-03-21 15:59 ` J.T. Conklin
2001-03-21 15:59 ` Kevin Buettner
2001-03-21 15:59 ` J.T. Conklin
2001-03-21 15:59 ` Andrew Cagney
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).