* S/390 Linux doesn't link on trunk @ 2002-04-07 14:39 Daniel Jacobowitz 2002-04-08 20:50 ` Andrew Cagney 2002-04-15 10:06 ` Dr. Jochen Röhrig 0 siblings, 2 replies; 7+ messages in thread From: Daniel Jacobowitz @ 2002-04-07 14:39 UTC (permalink / raw) To: gdb Just FYI; I didn't try the branch. The error is: gcc -O2 -g -Wl,-rpath,/usr/lib -o gdb \ main.o libgdb.a cli-dump.o cli-decode.o cli-script.o cli-cmds.o cli-setshow.o cli-utils.o mi-out.o mi-console.o mi-cmds.o mi-cmd-var.o mi-cmd-break.o mi-cmd-stack.o mi-cmd-disas.o mi-main.o mi-parse.o mi-getopt.o tui-file.o tui.o tuiData.o tuiSource.o tuiStack.o tuiIO.o tuiGeneralWin.o tuiLayout.o tuiWin.o tuiCommand.o tuiDisassem.o tuiSourceWin.o tuiRegs.o tuiDataWin.o tui-out.o tui-hooks.o ../bfd/libbfd.a -lreadline ../opcodes/libopcodes.a ../libiberty/libiberty.a -lncurses -lm ../libiberty/libiberty.a \ -ldl -rdynamic libgdb.a(inftarg.o): In function `init_child_ops': /home/buildd/build/gdb-5.2.cvs20020401/objdir/gdb/../../gdb/inftarg.c:754: undefined reference to `child_pid_to_exec_file' collect2: ld returned 1 exit status make[2]: *** [gdb] Error 1 make[2]: Leaving directory `/home/buildd/build/gdb-5.2.cvs20020401/objdir/gdb' make[1]: *** [all-gdb] Error 2 make[1]: Leaving directory `/home/buildd/build/gdb-5.2.cvs20020401/objdir' make: *** [build-stamp] Error 2 -- Daniel Jacobowitz Carnegie Mellon University MontaVista Software Debian GNU/Linux Developer ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: S/390 Linux doesn't link on trunk 2002-04-07 14:39 S/390 Linux doesn't link on trunk Daniel Jacobowitz @ 2002-04-08 20:50 ` Andrew Cagney 2002-04-08 21:36 ` Daniel Jacobowitz 2002-04-15 10:06 ` Dr. Jochen Röhrig 1 sibling, 1 reply; 7+ messages in thread From: Andrew Cagney @ 2002-04-08 20:50 UTC (permalink / raw) To: Daniel Jacobowitz; +Cc: gdb > Just FYI; I didn't try the branch. The error is: > > gcc -O2 -g -Wl,-rpath,/usr/lib -o gdb \ > main.o libgdb.a cli-dump.o cli-decode.o cli-script.o cli-cmds.o cli-setshow.o cli-utils.o mi-out.o mi-console.o mi-cmds.o mi-cmd-var.o mi-cmd-break.o mi-cmd-stack.o mi-cmd-disas.o mi-main.o mi-parse.o mi-getopt.o tui-file.o tui.o tuiData.o tuiSource.o tuiStack.o tuiIO.o tuiGeneralWin.o tuiLayout.o tuiWin.o tuiCommand.o tuiDisassem.o tuiSourceWin.o tuiRegs.o tuiDataWin.o tui-out.o tui-hooks.o ../bfd/libbfd.a -lreadline ../opcodes/libopcodes.a ../libiberty/libiberty.a -lncurses -lm ../libiberty/libiberty.a \ > -ldl -rdynamic > libgdb.a(inftarg.o): In function `init_child_ops': > /home/buildd/build/gdb-5.2.cvs20020401/objdir/gdb/../../gdb/inftarg.c:754: undefined reference to `child_pid_to_exec_file' > collect2: ld returned 1 exit status > make[2]: *** [gdb] Error 1 > make[2]: Leaving directory `/home/buildd/build/gdb-5.2.cvs20020401/objdir/gdb' > make[1]: *** [all-gdb] Error 2 > make[1]: Leaving directory `/home/buildd/build/gdb-5.2.cvs20020401/objdir' > make: *** [build-stamp] Error 2 I take it this is native? The cross appears to build fine. Andrew ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: S/390 Linux doesn't link on trunk 2002-04-08 20:50 ` Andrew Cagney @ 2002-04-08 21:36 ` Daniel Jacobowitz 2002-04-10 10:36 ` Jim Blandy 0 siblings, 1 reply; 7+ messages in thread From: Daniel Jacobowitz @ 2002-04-08 21:36 UTC (permalink / raw) To: Andrew Cagney; +Cc: gdb On Mon, Apr 08, 2002 at 11:50:19PM -0400, Andrew Cagney wrote: > >Just FYI; I didn't try the branch. The error is: > > > >gcc -O2 -g -Wl,-rpath,/usr/lib -o gdb \ > >main.o libgdb.a cli-dump.o cli-decode.o cli-script.o cli-cmds.o > >cli-setshow.o cli-utils.o mi-out.o mi-console.o mi-cmds.o mi-cmd-var.o > >mi-cmd-break.o mi-cmd-stack.o mi-cmd-disas.o mi-main.o mi-parse.o > >mi-getopt.o tui-file.o tui.o tuiData.o tuiSource.o tuiStack.o tuiIO.o > >tuiGeneralWin.o tuiLayout.o tuiWin.o tuiCommand.o tuiDisassem.o > >tuiSourceWin.o tuiRegs.o tuiDataWin.o tui-out.o tui-hooks.o > >../bfd/libbfd.a -lreadline ../opcodes/libopcodes.a > >../libiberty/libiberty.a -lncurses -lm ../libiberty/libiberty.a \ > >-ldl -rdynamic > >libgdb.a(inftarg.o): In function `init_child_ops': > >/home/buildd/build/gdb-5.2.cvs20020401/objdir/gdb/../../gdb/inftarg.c:754: > >undefined reference to `child_pid_to_exec_file' > >collect2: ld returned 1 exit status > >make[2]: *** [gdb] Error 1 > >make[2]: Leaving directory > >`/home/buildd/build/gdb-5.2.cvs20020401/objdir/gdb' > >make[1]: *** [all-gdb] Error 2 > >make[1]: Leaving directory `/home/buildd/build/gdb-5.2.cvs20020401/objdir' > >make: *** [build-stamp] Error 2 > > I take it this is native? The cross appears to build fine. Yes, it is. I don't have access to the build machine (as far as I know, at least...). -- Daniel Jacobowitz Carnegie Mellon University MontaVista Software Debian GNU/Linux Developer ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: S/390 Linux doesn't link on trunk 2002-04-08 21:36 ` Daniel Jacobowitz @ 2002-04-10 10:36 ` Jim Blandy 0 siblings, 0 replies; 7+ messages in thread From: Jim Blandy @ 2002-04-10 10:36 UTC (permalink / raw) To: Daniel Jacobowitz; +Cc: Andrew Cagney, gdb Daniel Jacobowitz <drow@mvista.com> writes: > Yes, it is. I don't have access to the build machine (as far as I > know, at least...). I believe that S/390 linux has had some recent changes in its debugging interface somewhere. (Core file? /proc? ptrace? I can't remember.) Something that needs to be a struct became an array, or something idiotic like that. Anyway, sourceware GDB did build on the S/390 as of the end of last December. It's probably bit rot and version skew. What a helpful message this is. ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: S/390 Linux doesn't link on trunk 2002-04-07 14:39 S/390 Linux doesn't link on trunk Daniel Jacobowitz 2002-04-08 20:50 ` Andrew Cagney @ 2002-04-15 10:06 ` Dr. Jochen Röhrig 2002-04-15 10:11 ` Daniel Jacobowitz 1 sibling, 1 reply; 7+ messages in thread From: Dr. Jochen Röhrig @ 2002-04-15 10:06 UTC (permalink / raw) To: gdb; +Cc: Daniel Jacobowitz, msnyder On Sunday, 7. April 2002 23:38, you wrote: > Just FYI; I didn't try the branch. The error is: > > gcc -O2 -g -Wl,-rpath,/usr/lib -o gdb \ > main.o libgdb.a cli-dump.o cli-decode.o cli-script.o cli-cmds.o > cli-setshow.o cli-utils.o mi-out.o mi-console.o mi-cmds.o mi-cmd-var.o > mi-cmd-break.o mi-cmd-stack.o mi-cmd-disas.o mi-main.o mi-parse.o > mi-getopt.o tui-file.o tui.o tuiData.o tuiSource.o tuiStack.o tuiIO.o > tuiGeneralWin.o tuiLayout.o tuiWin.o tuiCommand.o tuiDisassem.o > tuiSourceWin.o tuiRegs.o tuiDataWin.o tui-out.o tui-hooks.o > ../bfd/libbfd.a -lreadline ../opcodes/libopcodes.a > ../libiberty/libiberty.a -lncurses -lm ../libiberty/libiberty.a \ -ldl > -rdynamic > libgdb.a(inftarg.o): In function `init_child_ops': > /home/buildd/build/gdb-5.2.cvs20020401/objdir/gdb/../../gdb/inftarg.c:754: > undefined reference to `child_pid_to_exec_file' collect2: ld returned 1 The problem can be circumvented by adding a #undef CHILD_PID_TO_EXEC_FILE after the "#include "config/nm-linux.h"-line in config/s390/nm-linux.h (at least the current Debian gdb-source package version 5.2.cvs20020401-3 builds after this modification and gbd seems to work). The problem arises from the unconditional definition of CHILD_PID_TO_EXEC_FILE at the end of config/nm-linux.h. According to the CVS-log this was added in Revision 1.11 of config/nm-linux.h on January, 8th. There were also changes made in config/<arch>/linux.mh for some architectures (linking with linux-proc.o), which, as far as I can judge it, avoid the above described problem on theses architectures. However s390 (and, as it seems, some other architectures), doesn't have a config/<arch>/linux.mh so no modifications were made for s390. My question now: are we missing something for s390 because we don't have the config/s390/linux.mh-file or are the changes to config/s390/nm-linux.h that I described above a correct solution for the problem? Or would it instead be a better solution to add a "NATDEPFILES += linux-proc.o" to config/s390/s390.mh? Jochen -- Dr. Jochen Röhrig Linux for eServer Development IBM Deutschland Entwicklung GmbH, Schönaicher Str. 220, 71032 Böblingen roehrig@de.ibm.com ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: S/390 Linux doesn't link on trunk 2002-04-15 10:06 ` Dr. Jochen Röhrig @ 2002-04-15 10:11 ` Daniel Jacobowitz 2002-04-15 11:28 ` Michael Snyder 0 siblings, 1 reply; 7+ messages in thread From: Daniel Jacobowitz @ 2002-04-15 10:11 UTC (permalink / raw) To: Dr. Jochen Röhrig; +Cc: gdb, msnyder On Mon, Apr 15, 2002 at 07:02:25PM +0200, Dr. Jochen Röhrig wrote: > The problem arises from the unconditional definition of > CHILD_PID_TO_EXEC_FILE at the end of config/nm-linux.h. According to the > CVS-log this was added in Revision 1.11 of config/nm-linux.h on January, 8th. > There were also changes made in config/<arch>/linux.mh for some architectures > (linking with linux-proc.o), which, as far as I can judge it, avoid the above > described problem on theses architectures. However s390 (and, as it seems, > some other architectures), doesn't have a config/<arch>/linux.mh so no > modifications were made for s390. > > My question now: are we missing something for s390 because we don't have the > config/s390/linux.mh-file or are the changes to config/s390/nm-linux.h that I > described above a correct solution for the problem? Or would it instead be a > better solution to add a "NATDEPFILES += linux-proc.o" to config/s390/s390.mh? Aha! That's a good solution. A better one is to kill config/s390/s390.mh, create config/s390/linux.mh, and use that instead. That's much more accurate and will help prevent this sort of error recurring. Michael, sound right to you? I can't think of any reason not to use linux-proc.o on S/390. -- Daniel Jacobowitz Carnegie Mellon University MontaVista Software Debian GNU/Linux Developer ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: S/390 Linux doesn't link on trunk 2002-04-15 10:11 ` Daniel Jacobowitz @ 2002-04-15 11:28 ` Michael Snyder 0 siblings, 0 replies; 7+ messages in thread From: Michael Snyder @ 2002-04-15 11:28 UTC (permalink / raw) To: Daniel Jacobowitz; +Cc: Dr. Jochen Röhrig, gdb Daniel Jacobowitz wrote: > > On Mon, Apr 15, 2002 at 07:02:25PM +0200, Dr. Jochen Röhrig wrote: > > The problem arises from the unconditional definition of > > CHILD_PID_TO_EXEC_FILE at the end of config/nm-linux.h. According to the > > CVS-log this was added in Revision 1.11 of config/nm-linux.h on January, 8th. > > There were also changes made in config/<arch>/linux.mh for some architectures > > (linking with linux-proc.o), which, as far as I can judge it, avoid the above > > described problem on theses architectures. However s390 (and, as it seems, > > some other architectures), doesn't have a config/<arch>/linux.mh so no > > modifications were made for s390. > > > > My question now: are we missing something for s390 because we don't have the > > config/s390/linux.mh-file or are the changes to config/s390/nm-linux.h that I > > described above a correct solution for the problem? Or would it instead be a > > better solution to add a "NATDEPFILES += linux-proc.o" to config/s390/s390.mh? > > Aha! That's a good solution. A better one is to kill > config/s390/s390.mh, create config/s390/linux.mh, and use that instead. > That's much more accurate and will help prevent this sort of error > recurring. > > Michael, sound right to you? I can't think of any reason not to use > linux-proc.o on S/390. So long as /proc works as advertised on s390/Linux, sure, go for it. ^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2002-04-15 18:28 UTC | newest] Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2002-04-07 14:39 S/390 Linux doesn't link on trunk Daniel Jacobowitz 2002-04-08 20:50 ` Andrew Cagney 2002-04-08 21:36 ` Daniel Jacobowitz 2002-04-10 10:36 ` Jim Blandy 2002-04-15 10:06 ` Dr. Jochen Röhrig 2002-04-15 10:11 ` Daniel Jacobowitz 2002-04-15 11:28 ` Michael Snyder
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox