[Bug gold/26023] New: Null pointer dereferencing in gc_process_relocs

[zeccav at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=26023

Bug ID: 26023
   Summary: Null pointer dereferencing in gc_process_relocs
   Product: binutils
   Version: 2.34
Status: UNCONFIRMED
  Severity: normal
  Priority: P2
 Component: gold
  Assignee: ccoutant at gmail dot com
  Reporter: zeccav at gmail dot com
CC: ian at airs dot com
  Target Milestone: ---

While running the make check procedure a null pointer is referenced
in gold/gc.h line 218

check_section_for_function_pointers =
symtab->icf()->check_section_for_function_pointers(src_section_name,
   target);

You may reproduce it by prepending the following

gold_assert(symtab);
gold_assert(symtab->icf()); /* internal error here */

This is running plugin_test_5, gc_comdat_test, gc_tls_test,
gc_orphan_section_test, pr14265, pr20717, gc_dynamic_list_test,
file_in_many_sections.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug ld/26018] Make --dynamic-list* override -Bsymbolic and -Bsymbolic-functions

[hjl.tools at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=26018

H.J. Lu  changed:

   What|Removed |Added

 CC||hjl.tools at gmail dot com

--- Comment #1 from H.J. Lu  ---
--dynamic-list=DYNAMIC-LIST-FILE'
 Specify the name of a dynamic list file to the linker.  This is
 typically used when creating shared libraries to specify a list of
 global symbols whose references shouldn't be bound to the
 definition within the shared library, or creating dynamically
 linked executables to specify a list of symbols which should be
 added to the symbol table in the executable.  This option is only
 meaningful on ELF platforms which support shared libraries.

The --dynamic-list* options are intended for shared libraries.  The goal
IS NOT put them in dynamic symbol table since all global symbols are in
dynamic symbol table already in a shared library.  The goal is to make
them PREEMPTIBLE.  So making the --dynamic-list* options override -Bsymbolic
and -Bsymbolic-functions is incorrect since there is NOTHING to override.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug ld/25910] Support --export-dynamic-symbol $symbol

[hjl.tools at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=25910

--- Comment #3 from H.J. Lu  ---
You should

1. Extend  --dynamic-list* to executables.  Symbol binding is unchanged.
Only add symbols to dynamic symbol table.
2. Make --export-dynamic-symbol an alias of  --dynamic-list* for executables.
and ignore --export-dynamic-symbol for -shared.  It has to be done at the
end of command-line parsing.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug gas/26025] New: riscv gas inserts relocation for BFD_RELOC_RISCV_CFA in last section of file, instead of relevant section

[je...@j-software.dk]

https://sourceware.org/bugzilla/show_bug.cgi?id=26025

Bug ID: 26025
   Summary: riscv gas inserts relocation for BFD_RELOC_RISCV_CFA
in last section of file, instead of relevant section
   Product: binutils
   Version: unspecified
Status: UNCONFIRMED
  Severity: normal
  Priority: P2
 Component: gas
  Assignee: unassigned at sourceware dot org
  Reporter: je...@j-software.dk
  Target Milestone: ---

Created attachment 12562
  --> https://sourceware.org/bugzilla/attachment.cgi?id=12562&action=edit
Assembly triggering the bug

Somehow a specific sequence of Dwarf debug info triggers this bug.

It causes R_RISCV_SET32 and R_RISCV_SUB32 relocations to be added to the last
section of the file, in the attached example that would be
.text.z_DEBUGEND_$P$TEST

When trying to link this it would generate a segfault because it's not the
right section, and an empty one

I'm guessing it's because relocations can't be added in riscv_pre_output_hook
the way it's currently done

Steps to reproduce:
riscv32-linux-as test.s -o test.o
riscv32-linux-ld test.o
riscv32-linux-objdump -r test.o

objdump would show two relocations in the last section that should be in
.debug_frame instead

Same thing happens on 64bit toolchain as well

-- 
You are receiving this mail because:
You are on the CC list for the bug.

Binutils mailing list subscription does not work

[Paul Kaletta]

Hi,

https://www.gnu.org/software/binutils/  
contains an HTML form for subscribing to the binutils mailing list.

Unfortunately, the „Send in the request“ button results in a 404 Not Found 
error (The requested URL /cgi-bin/ml-request was not found on this server).

Best regards,
Paul Kaletta

Re: Binutils mailing list subscription does not work

[Stephen Casner]

On Fri, 22 May 2020, Paul Kaletta wrote:

> https://www.gnu.org/software/binutils/
>  contains an HTML form for
> subscribing to the binutils mailing list.
>
> Unfortunately, the „Send in the request“ button results in a 404 Not
> Found error (The requested URL /cgi-bin/ml-request was not found on
> this server).

I'll answer since I was the person to most recently ask the same
question about a month ago.  The reply I received was:

" Some external links are still broken after the recent move of the
mailing list to a new system.

Use:  to subscribe
via a web interface or send an e-mail to:
 with "subscribe" in the subject to
get it handled the classic way."

-- Steve

[Bug gas/26025] riscv gas inserts relocation for BFD_RELOC_RISCV_CFA in last section of file, instead of relevant section

[je...@j-software.dk]

https://sourceware.org/bugzilla/show_bug.cgi?id=26025

--- Comment #1 from je...@j-software.dk ---
Created attachment 12563
  --> https://sourceware.org/bugzilla/attachment.cgi?id=12563&action=edit
Potential fix

This patch removes the crash and seems to put the relocations in the right
place

-- 
You are receiving this mail because:
You are on the CC list for the bug.