Why HP-UX must use its own linker for gcc?

[liuyl_it]

Anyone can explain why GNU do not provide or support itself's linker within the binutils for HP-UX？

The related post reference link:
https://community.hpe.com/t5/Languages-and-Scripting/HPUX-linker-with-gnu-compilers/m-p/2797432#M3941
https://community.hpe.com/t5/System-Administration/How2-install-gnu-linker-on-B-11-23-ia64/td-p/3389167

[Dennis Handly]

For what architecture?  For PA32, it uses SOM format, not ELF.

For PA64 and Integrity, it uses ELF.

What fancy features do you need from GNU ld?

[liuyl_it]

Because the linker for HP-UX itself must use the following shared library file！
And this file libu2comp.so is just bundled with the HP-UX aCC commercial software！

# gcc scsi_test.c -o scsi_test
ld: Unable to load shared library "/opt/langtools/lib/hpux64/libu2comp.so"
Fatal error.
collect2: ld returned 1 exit status

Notes: There is also no GNU linker within the binutils for either IA/PA64 or PA32！

[Dennis Handly]

> Because the linker for HP-UX itself must use the following shared library file

This is a user error and it's not true.  Newer lds were suppose to have an error message saying mixing PA32 with ELF.

> # gcc scsi_test.c

Compile with -c and use "file scsi_test.o" to see what the architecture is.  Are you trying to build PA32 on Integrity?

>There is also no GNU linker within the binutils for either IA/PA64 or PA32

Why would you need one?

[liuyl_it]

1) It is directly built on IA64:
#
# file scsi_test.o
scsi_test.o: ELF-32 relocatable object file - IA64
#

And referring to the following posts:
https://community.hpe.com/t5/General/from-PA-RISC-to-Itanium-Unable-to-load-shared-library/td-p/4866432   <<<< no this cause
https://community.hpe.com/t5/Languages-and-Scripting/Unable-to-load-quot-libu2comp-so-quot/td-p/6971384  <<<< this might be

2) I think that the GNU linker should not use this shared library file that bundled with HP-UX aCC！
Notes: It seems that the GNU linker integrates with the binutils for the most UNIXes except HP-UX！

[Dennis Handly]

> 1) It is directly built on IA64:  scsi_test.o: ELF-32 relocatable object file - IA64

Ok but there may be other objects that confuse ld(1).

Note: Your URLs contain junk chars on the end.  You can fix that by editing your post with Edit Reply and using the hyperlink menu to edit them.
https://community.hpe.com/t5/General/from-PA-RISC-to-Itanium-Unable-to-load-shared-library/td-p/4866432 <<<< no this cause

This could be a case where that shlib was actually needed to optimize?

https://community.hpe.com/t5/Languages-and-Scripting/Unable-to-load-quot-libu2comp-so-quot/td-p/6971384  <<<< this might be

Yes, here is where I explain how to find the bad object:

https://community.hpe.com/t5/Languages-and-Scripting/Unable-to-load-quot-libu2comp-so-quot/td-p/6971384#U6972297

 > 2) I think that the GNU linker should not use this shared library file that bundled with HP-UX aCC

That's not the problem.  The problem is that ld(1) is not giving you an error when you're passing in the wrong architecture objects, archives or shlibs.

[liuyl_it]

1) Bad object ?

#
#
# export LDOPTS="+vtype files"
#
#
# gcc scsi_test.c -o scsi_test_sp
Loading /usr/lib/hpux32/unix98.o:
ld: Unable to load shared library "/opt/langtools/lib/hpux64/libu2comp.so"
Fatal error.
collect2: ld returned 1 exit status
#
#
#
# ls -l /usr/lib/hpux32/unix98.o
-r--r--r-- 1 bin bin 1560 Jan 21 2014 /usr/lib/hpux32/unix98.o
#
#
#
# export LDOPTS="-V -T"
#
#
# echo $LDOPTS
-V -T
#
#
# gcc scsi_test.c -o scsi_test_sp
ld: 92453-07 linker ld HP Itanium(R) B.12.61 IPF/IPF
ld: Unable to load shared library "/opt/langtools/lib/hpux64/libu2comp.so"
Fatal error.
collect2: ld returned 1 exit status
#
#

[Dennis Handly]

> Loading /usr/lib/hpux32/unix98.o:
> ld: Unable to load shared library "/opt/langtools/lib/hpux64/libu2comp.so"

This indicates that unix98.o is bad but I'm not sure how.

Trying using "-Wl,-v" with gcc, to have linker verbose output.  I only need to see first 10 or so lines.

And next step is to use gcc to compile your object but link with cc.

[liuyl_it]

Now this time I feel that we might find the hidden problem soon ......！

#
#
# gcc -Wl,-v scsi_test.c -o scsi_test_exec
collect2 version 4.3.1 (IA-64) HP-UX
/usr/ccs/bin/ld -z +Accept TypeMismatch -u main -o scsi_test_exec /usr/lib/hpux32/unix98.o -L/opt/hp-gcc-4.3.1/lib/gcc/ia64-hp-hpux11.23/4.3.1 -L/usr/ccs/lib -L/opt/hp-gcc-4.3.1/lib/gcc/ia64-hp-hpux11.23/4.3.1/../../.. -v scsi_test.c -lgcc -lgcc_eh -lunwind -lc -lgcc -lgcc_eh -lunwind
/usr/ccs/bin/ld -z +Accept TypeMismatch -u main -o scsi_test_exec /usr/lib/hpux32/unix98.o -L/opt/hp-gcc-4.3.1/lib/gcc/ia64-hp-hpux11.23/4.3.1 -L/usr/ccs/lib -L/opt/hp-gcc-4.3.1/lib/gcc/ia64-hp-hpux11.23/4.3.1/../../.. -v scsi_test.c -lgcc -lgcc_eh -lunwind -lc -lgcc -lgcc_eh -lunwind
LPATH is :
Loading /usr/lib/hpux32/unix98.o:
/usr/lib/hpux32/unix98.o:
__xpg4_extended_mask is DEFINED GLOBAL OBJECT
/usr/lib/hpux32/unix98.o:
section .sdata PROGBITS AWS 4 8 added to data segment
section .debug_line PROGBITS 26 8 added to nonsegment segment
section .debug_actual PROGBITS 24 8 added to nonsegment segment
section .note NOTE 848 4 added to note segment
ld: Unable to load shared library "/opt/langtools/lib/hpux64/libu2comp.so"
Fatal error.
collect2: ld returned 1 exit status
#
#
# gcc -c scsi_test.c -o scsi_test_relocate.o
gcc: scsi_test.c: linker input file unused because linking not done
#
#

[liuyl_it]

Eventually I found that I had made an embarrassing  low-level mistake！
It seems that the original soruce filename(scsi_test.c) was wrong with some invisible charactors！
Now I just delete this original soruce file and then recreate it, so at last al l right！

But I still do not understand why gcc would report such ld libu2comp.so error under this situation！

[ranganath ramachandra]

gcc got misled by the unexpected content in the file and considered it an object file, so passed it to the linker; the linker in turn took it for an instrumented object file to be (re-)processed by the compiler backend - that is the case when it tried to load libucomp.so .

how did you locate the problem?

[Dennis Handly · Accepted solution]

> It seems that the original source filename (scsi_test.c) was wrong with some invisible characters.

Yes, that would confused the driver if it didn't end in ".c".  cc(1)/c99(1) work the same way.

> I still do not understand why gcc would report such ld libu2comp.so error under this situation

Because ld(1) is broken and doesn't make sure it is an ELF file.

> the linker in turn took it for an instrumented object file to be (re-)processed by the compiler backend - that is the case when it tried to load libucomp.so .

Wasn't this fixed years ago?

> how did you locate the problem?

Well, the source file was listed on the verbose ld output.  :-)

[ranganath ramachandra]

---

@Dennis Handly wrote:

> It seems that the original source filename (scsi_test.c) was wrong with some invisible characters.

Yes, that would confused the driver if it didn't end in ".c".  cc(1)/c99(1) work the same way.

> I still do not understand why gcc would report such ld libu2comp.so error under this situation

Because ld(1) is broken and doesn't make sure it is an ELF file.

> the linker in turn took it for an instrumented object file to be (re-)processed by the compiler backend - that is the case when it tried to load libucomp.so .

Wasn't this fixed years ago?

At least now.

[ cathh02 tmp ] $ gcc -Wl,-V test.cx
ld: 92453-07 linker ld HP Itanium(R) B.12.63 IPF/IPF
ld: Mismatched ABI (not an ELF file) for test.cx
Fatal error.
collect2: error: ld returned 1 exit status

I got the same results with 12.58, couldn't find 12.61 though.

[liuyl_it]

I noticed the following error message that returned via gcc -c :
gcc: scsi_test.c: linker input file unused because linking not done

Then the below related post at stackoverflow should implied what cause my problem might be due to!
https://stackoverflow.com/questions/10739072/gcclinker-input-file-unused-because-linking-not-done

But I still really cannot explain why the source file was listed on the verbose ld output.

By the way,  it seems that the earlier version of binutils for HP-UX provided the GNU linker long ago！

[Dennis Handly]

> But I still really cannot explain why the source file was listed on the verbose ld output.

Because it's not a source file because it doesn't end in ".c".  You said there was junk on the end?

ll -b will show the unprintable chars.

[liuyl_it]

Now I wonder what the other error messages would be returned via ld with the same problematic source file  if the system has installed with the shared library of libu2comp.so ？

[ranganath ramachandra]

---

@liuyl_it wrote:

Now I wonder what the other error messages would be returned via ld with the same problematic source file  if the system has installed with the shared library of libu2comp.so ？

---

As long as the file is properly named, any errors that can be detected with the installed compilter would be reported already.

If the file is not named correctly, error reports may not be reliable. So don't bother about the errors when the file is incorrectly named.

From the compiler manual page cc_bundled(1):

 Other Suffixes
             All other arguments, such as those names ending with .o,
             .a, or .so are taken to be relocatable object files and
             are passed to ld (see ld(1)) to be included in the link
             operation.

 The behaviour of the linker for unrecognized input formats seems to be undocumented.

[liuyl_it]

These days I have searched the knowledges for the instrumented object/code/execution！
But I just only found few information about them including the following reference link:

https://www.sharcnet.ca/Software/Intel/IFC/compiler_f/main_for/copts/common_options/option_prof_gen.htm

So could you tell me the significance of the "instrumented object file"  concept in my this issue？

[ranganath ramachandra]

---

So could you tell me the significance of the "instrumented object file"  concept in my this issue？

---

In your case there is no instrumented object file. Because of the bad filename, the compiler passed the source file to the linker. The linker mistakenly assumed it to be an instrumented object file. If your original problem goes away after renaming the source file, you don't have to bother about it any more.