Mirror of the gdb mailing list
 help / color / mirror / Atom feed
* Re: Problems building gdbserver in binutils 2.46
       [not found] <CAG_eJLe_vuR93fE6H4hENPh3=R2XE72T2tL1ZaTnY93zW2J5-Q@mail.gmail.com>
@ 2026-02-09  7:45 ` Jan Beulich via Gdb
  2026-02-09 12:20   ` Tom Kacvinsky via Gdb
  2026-02-09 14:52   ` Andreas Schwab via Gdb
  0 siblings, 2 replies; 6+ messages in thread
From: Jan Beulich via Gdb @ 2026-02-09  7:45 UTC (permalink / raw)
  To: Tom Kacvinsky, Nick Clifton; +Cc: Binutils, GDB

On 08.02.2026 16:34, Tom Kacvinsky wrote:
> Using CentOS 70 - which has glibc 2.17 and kernel 3.10.0 - and using the
> tool chains
> 
>   GCC 12.1.0
>   binutils 2.39
> libtool 2.47
> 
> and then building binutil 2.46 from source without specifying
> --disable-gdbserver, I got this build error

Doesn't this point at a (long-standing) regression in how branching is done?
Up to 2.24 gdb*/ weren't part of the branches, hence such an issue couldn't
arise.

I assume you would observe the same when building on the master branch; there
the issue wants dealing with on the gdb side then; Cc-ing the correct list.

Jan

> CXX    nat/x86-linux.o
> 
> In file included from
> ../../binutils-gdb/gdbserver/../gdb/nat/x86-linux.c:22:
> 
> ../../binutils-gdb/gdbserver/../gdb/nat/x86-linux.c: In function ‘bool
> i386_ptrace_get_tls_data(int, gdb::array_view<user_desc>)’:
> 
> ../../binutils-gdb/gdbserver/../gdb/nat/x86-linux.c:207:19: error:
> ‘PTRACE_GET_THREAD_AREA’ was not declared in this scope
> 
>   207 |       if (ptrace (PTRACE_GET_THREAD_AREA, pid, addr, data) < 0)
> 
>       |                   ^~~~~~~~~~~~~~~~~~~~~~
> 
> ../../binutils-gdb/gdbserver/../gdb/nat/gdb_ptrace.h:142:37: note: in
> definition of macro ‘ptrace’
> 
>   142 |          ptrace ((PTRACE_TYPE_ARG1) request, pid, addr, data)
> 
>       |                                     ^~~~~~~
> 
> ../../binutils-gdb/gdbserver/../gdb/nat/x86-linux.c: In function ‘bool
> i386_ptrace_set_tls_data(int, gdb::array_view<user_desc>)’:
> 
> ../../binutils-gdb/gdbserver/../gdb/nat/x86-linux.c:226:19: error:
> ‘PTRACE_SET_THREAD_AREA’ was not declared in this scope
> 
>   226 |       if (ptrace (PTRACE_SET_THREAD_AREA, pid, addr, data) < 0)
> 
>       |                   ^~~~~~~~~~~~~~~~~~~~~~
> 
> ../../binutils-gdb/gdbserver/../gdb/nat/gdb_ptrace.h:142:37: note: in
> definition of macro ‘ptrace’
> 
>   142 |          ptrace ((PTRACE_TYPE_ARG1) request, pid, addr, data)
> 
> 
> I know this is an EOL Linux distribution, but for the moment that is
> what I need to build our product.  From what research I have done,
> PTRACE_GET_THREAD_AREA and PTRACE_TYPE_ARG1 are only available in newer
> versions of glibc.
> 
> Any ideas how to fix this so one doesn't have to specify
> --disable-gdbserver? I am worried that future builds of gdb from the
> binutils-gdb repo will no longer build on CentOS 7.
> 
> Thanks,
> 
> Tom
> 


^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: Problems building gdbserver in binutils 2.46
  2026-02-09  7:45 ` Problems building gdbserver in binutils 2.46 Jan Beulich via Gdb
@ 2026-02-09 12:20   ` Tom Kacvinsky via Gdb
  2026-02-09 14:43     ` Tom Kacvinsky via Gdb
  2026-02-09 14:52   ` Andreas Schwab via Gdb
  1 sibling, 1 reply; 6+ messages in thread
From: Tom Kacvinsky via Gdb @ 2026-02-09 12:20 UTC (permalink / raw)
  To: Jan Beulich, Nick Clifton, GDB, Binutils

Hi,

On Mon, Feb 9, 2026 at 2:45 AM Jan Beulich <jbeulich@suse.com> wrote:

> On 08.02.2026 16:34, Tom Kacvinsky wrote:
> > Using CentOS 70 - which has glibc 2.17 and kernel 3.10.0 - and using the
> > tool chains
> >
> >   GCC 12.1.0
> >   binutils 2.39
> > libtool 2.47
> >
> > and then building binutil 2.46 from source without specifying
> > --disable-gdbserver, I got this build error
>
> Doesn't this point at a (long-standing) regression in how branching is
> done?
> Up to 2.24 gdb*/ weren't part of the branches, hence such an issue couldn't
> arise.
>
> I assume you would observe the same when building on the master branch;
> there
> the issue wants dealing with on the gdb side then; Cc-ing the correct list.
>

Yes, I switched to the master branch (from binutils-2_46) and the problems
exists
there, too.

So what I do being that I only want gas and ld, I configure with

../binutils-gdb/configure --prefix=${BINUTILS_PREFIX} \

    --disable-gdb \

    --disable-gdbserver \

    --disable-gprof \

    --disable-gprofng \

    --enable-multilib


and then the build completes.


Tom

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: Problems building gdbserver in binutils 2.46
  2026-02-09 12:20   ` Tom Kacvinsky via Gdb
@ 2026-02-09 14:43     ` Tom Kacvinsky via Gdb
  0 siblings, 0 replies; 6+ messages in thread
From: Tom Kacvinsky via Gdb @ 2026-02-09 14:43 UTC (permalink / raw)
  To: Jan Beulich, Nick Clifton, GDB, Binutils

Hi,

On Mon, Feb 9, 2026 at 7:20 AM Tom Kacvinsky <tkacvins@gmail.com> wrote:

> Hi,
>
> On Mon, Feb 9, 2026 at 2:45 AM Jan Beulich <jbeulich@suse.com> wrote:
>
>> On 08.02.2026 16:34, Tom Kacvinsky wrote:
>> > Using CentOS 70 - which has glibc 2.17 and kernel 3.10.0 - and using the
>> > tool chains
>> >
>> >   GCC 12.1.0
>> >   binutils 2.39
>> > libtool 2.47
>> >
>> > and then building binutil 2.46 from source without specifying
>> > --disable-gdbserver, I got this build error
>>
>> Doesn't this point at a (long-standing) regression in how branching is
>> done?
>> Up to 2.24 gdb*/ weren't part of the branches, hence such an issue
>> couldn't
>> arise.
>>
>> I assume you would observe the same when building on the master branch;
>> there
>> the issue wants dealing with on the gdb side then; Cc-ing the correct
>> list.
>>
>
> Yes, I switched to the master branch (from binutils-2_46) and the problems
> exists
> there, too.
>
> So what I do being that I only want gas and ld, I configure with
>
> ../binutils-gdb/configure --prefix=${BINUTILS_PREFIX} \
>
>     --disable-gdb \
>
>     --disable-gdbserver \
>
>     --disable-gprof \
>
>     --disable-gprofng \
>
>     --enable-multilib
>
>
> and then the build completes.
>
>
I did note that PTRACE_GET_THREAD_AREA is available in glibc 2.31 (on my
openSUSE Leap 15.5 VM).  Not sure which glibc version introduced that. If
Google's AI overview can be trusted, it says that PTRACE_TYPE_ARG1 is
supposed to be auto-detected/set by configure.  I looked at the configure
scripts and did not see a "#define PTRACE_TYPE_ARG1 ..."
Tom

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: Problems building gdbserver in binutils 2.46
  2026-02-09  7:45 ` Problems building gdbserver in binutils 2.46 Jan Beulich via Gdb
  2026-02-09 12:20   ` Tom Kacvinsky via Gdb
@ 2026-02-09 14:52   ` Andreas Schwab via Gdb
  2026-02-09 14:54     ` Jan Beulich via Gdb
  1 sibling, 1 reply; 6+ messages in thread
From: Andreas Schwab via Gdb @ 2026-02-09 14:52 UTC (permalink / raw)
  To: Jan Beulich via Gdb; +Cc: Tom Kacvinsky, Nick Clifton, Jan Beulich, Binutils

On Feb 09 2026, Jan Beulich via Gdb wrote:

> Doesn't this point at a (long-standing) regression in how branching is done?
> Up to 2.24 gdb*/ weren't part of the branches, hence such an issue couldn't
> arise.

It is only present in the repository, but never in the tarball.

-- 
Andreas Schwab, SUSE Labs, schwab@suse.de
GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE  1748 E4D4 88E3 0EEA B9D7
"And now for something completely different."

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: Problems building gdbserver in binutils 2.46
  2026-02-09 14:52   ` Andreas Schwab via Gdb
@ 2026-02-09 14:54     ` Jan Beulich via Gdb
  2026-02-09 15:24       ` Andreas Schwab via Gdb
  0 siblings, 1 reply; 6+ messages in thread
From: Jan Beulich via Gdb @ 2026-02-09 14:54 UTC (permalink / raw)
  To: Andreas Schwab; +Cc: Tom Kacvinsky, Nick Clifton, Binutils, GDB

On 09.02.2026 15:52, Andreas Schwab wrote:
> On Feb 09 2026, Jan Beulich via Gdb wrote:
> 
>> Doesn't this point at a (long-standing) regression in how branching is done?
>> Up to 2.24 gdb*/ weren't part of the branches, hence such an issue couldn't
>> arise.
> 
> It is only present in the repository, but never in the tarball.

Well, yes, hence I said "on the branches". I'm puzzled by that, as I'd expect
the tarballs to represent the branches at the respective tags.

Jan

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: Problems building gdbserver in binutils 2.46
  2026-02-09 14:54     ` Jan Beulich via Gdb
@ 2026-02-09 15:24       ` Andreas Schwab via Gdb
  0 siblings, 0 replies; 6+ messages in thread
From: Andreas Schwab via Gdb @ 2026-02-09 15:24 UTC (permalink / raw)
  To: Jan Beulich; +Cc: Tom Kacvinsky, Nick Clifton, Binutils, GDB

On Feb 09 2026, Jan Beulich wrote:

> Well, yes, hence I said "on the branches". I'm puzzled by that, as I'd expect
> the tarballs to represent the branches at the respective tags.

Never has been, see src-release.sh how they are created.

-- 
Andreas Schwab, SUSE Labs, schwab@suse.de
GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE  1748 E4D4 88E3 0EEA B9D7
"And now for something completely different."

^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2026-02-09 15:26 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
     [not found] <CAG_eJLe_vuR93fE6H4hENPh3=R2XE72T2tL1ZaTnY93zW2J5-Q@mail.gmail.com>
2026-02-09  7:45 ` Problems building gdbserver in binutils 2.46 Jan Beulich via Gdb
2026-02-09 12:20   ` Tom Kacvinsky via Gdb
2026-02-09 14:43     ` Tom Kacvinsky via Gdb
2026-02-09 14:52   ` Andreas Schwab via Gdb
2026-02-09 14:54     ` Jan Beulich via Gdb
2026-02-09 15:24       ` Andreas Schwab via Gdb

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox