public inbox for git-commits@fedoraproject.org
help / color / mirror / Atom feed
* [rpms/gdb] gdb-17.2-rebase-f44: - Fix failed gdb_assert due to the PIE patchset (BZ 559414).
@ 2026-06-27 23:54 Jan Kratochvil
  0 siblings, 0 replies; only message in thread
From: Jan Kratochvil @ 2026-06-27 23:54 UTC (permalink / raw)
  To: git-commits

A new commit has been pushed.

Repo   : rpms/gdb
Branch : gdb-17.2-rebase-f44
Commit : a04bc95fa3524f77c3d24050ac4cd404d74810f4
Author : Jan Kratochvil <jkratoch@fedoraproject.org>
Date   : 2010-01-31T01:57:45+00:00
Stats  : +178/-23 in 3 file(s)
URL    : https://src.fedoraproject.org/rpms/gdb/c/a04bc95fa3524f77c3d24050ac4cd404d74810f4?branch=gdb-17.2-rebase-f44

Log:
- Fix failed gdb_assert due to the PIE patchset (BZ 559414).

---
diff --git a/gdb-archer-pie-assert-temp-workaround.patch b/gdb-archer-pie-assert-temp-workaround.patch
deleted file mode 100644
index 42a9275..0000000
--- a/gdb-archer-pie-assert-temp-workaround.patch
+++ /dev/null
@@ -1,17 +0,0 @@
---- gdb-7.0.50.20100121/gdb/solib-svr4.c-orig	2010-01-21 21:09:54.000000000 +0100
-+++ gdb-7.0.50.20100121/gdb/solib-svr4.c	2010-01-21 21:10:01.000000000 +0100
-@@ -596,7 +596,14 @@ scan_dyntag (int dyntag, bfd *abfd, CORE
-        target_section++)
-     if (sect == target_section->the_bfd_section)
-       break;
-+  /* FIXME: https://bugzilla.redhat.com/show_bug.cgi?id=556310
-+     It may crash here but it is not understood why.  */
-+#if 0
-   gdb_assert (target_section < current_target_sections->sections_end);
-+#else
-+  if (target_section == current_target_sections->sections_end)
-+    return 0;
-+#endif
- 
-   /* Read in .dynamic from the BFD.  We will get the actual value
-      from memory later.  */

diff --git a/gdb-bz559414-pie-assert-fix.patch b/gdb-bz559414-pie-assert-fix.patch
new file mode 100644
index 0000000..31c531f
--- /dev/null
+++ b/gdb-bz559414-pie-assert-fix.patch
@@ -0,0 +1,172 @@
+http://sourceware.org/ml/gdb-patches/2010-01/msg00661.html
+Subject: [patch] Fix failed gdb_assert due to the PIE patchset
+
+Hi,
+
+the PIE patchset brought in an occasionally failing gdb_assert (according to
+Fedora ABRT bugreports).
+solib-svr4.c:600: internal-error: scan_dyntag: Assertion `target_section < current_target_sections->sections_end' failed.
+
+Now I understand it as generally current_target_sections are about target-side
+data while symbols are the GDB side and there is only loose relation of those.
+
+ABFD can be from multiple places acting for multiple purposes as parsed by
+elf_lookup_lib_symbol:
+
+exec_bfd:
+	It cannot be from exec_one - this BFD has no associated objfile to
+	pass to symfile_objfile.  OK.
+symfile_objfile:
+	It gets translated by elf_lookup_lib_symbol into exec_bfd which
+	represents the target data.  OK.
+so_list's objfile's separate_debug_objfile file:
+	gdb_assert-forbidden as input to elf_lookup_lib_symbol.  OK.
+so_list's objfile's main file:
+	update_solib_list must have called add_target_sections.  OK.
+objfile's separate_debug_objfile file:
+	gdb_assert-forbidden as input to elf_lookup_lib_symbol.  OK.
+objfile's main file (if it has no associated so_list):
+	Such ABFD from add_symbol_file_command has no add_target_sections
+	called.  Still it can have current context BLOCK which crashes
+	scan_dyntag.  BUG.
+
+
+No regressions on {x86_64,x86_64-m32,i686}-fedora12-linux-gnu.
+
+The patch mostly reverts the PIE part there.
+
+OK to check in?
+
+
+Sorry,
+Jan
+
+
+gdb/
+2010-01-31  Jan Kratochvil  <jan.kratochvil@redhat.com>
+
+	* solib-svr4.c (scan_dyntag): New variable dyn_addr.  Replace gdb_assert
+	by a conditional setting DYN_ADDR.  Use DYN_ADDR.
+	* config/djgpp/fnchange.lst: Add translations for
+	symbol-without-target_section.exp and symbol-without-target_section.c.
+
+gdb/testsuite/
+2010-01-31  Jan Kratochvil  <jan.kratochvil@redhat.com>
+
+	* gdb.base/symbol-without-target_section.exp,
+	gdb.base/symbol-without-target_section.c: New.
+
+--- a/gdb/config/djgpp/fnchange.lst
++++ b/gdb/config/djgpp/fnchange.lst
+@@ -394,6 +394,8 @@
+ @V@/gdb/testsuite/gdb.base/solib-symbol-main.c @V@/gdb/testsuite/gdb.base/so-symmain.c
+ @V@/gdb/testsuite/gdb.base/solib-overlap-lib.c @V@/gdb/testsuite/gdb.base/so-ovrlib.c
+ @V@/gdb/testsuite/gdb.base/solib-overlap-main.c @V@/gdb/testsuite/gdb.base/so-ovrmain.c
++@V@/gdb/testsuite/gdb.base/gdb.base/symbol-without-target_section.exp @V@/gdb/testsuite/gdb.base/symnosec.exp
++@V@/gdb/testsuite/gdb.base/gdb.base/symbol-without-target_section.c @V@/gdb/testsuite/gdb.base/symnosec.c
+ @V@/gdb/testsuite/gdb.base/type-opaque-lib.c @V@/gdb/testsuite/gdb.base/ty-opqlib.c
+ @V@/gdb/testsuite/gdb.base/type-opaque-main.c @V@/gdb/testsuite/gdb.base/ty-opqmain.c
+ @V@/gdb/testsuite/gdb.base/watchpoint-hw.c @V@/gdb/testsuite/gdb.base/wp-hw.c
+--- a/gdb/solib-svr4.c
++++ b/gdb/solib-svr4.c
+@@ -570,7 +570,7 @@ scan_dyntag (int dyntag, bfd *abfd, CORE_ADDR *ptr)
+ {
+   int arch_size, step, sect_size;
+   long dyn_tag;
+-  CORE_ADDR dyn_ptr;
++  CORE_ADDR dyn_ptr, dyn_addr;
+   gdb_byte *bufend, *bufstart, *buf;
+   Elf32_External_Dyn *x_dynp_32;
+   Elf64_External_Dyn *x_dynp_64;
+@@ -597,7 +597,17 @@ scan_dyntag (int dyntag, bfd *abfd, CORE_ADDR *ptr)
+        target_section++)
+     if (sect == target_section->the_bfd_section)
+       break;
+-  gdb_assert (target_section < current_target_sections->sections_end);
++  if (target_section < current_target_sections->sections_end)
++    dyn_addr = target_section->addr;
++  else
++    {
++      /* ABFD may come from OBJFILE acting only as a symbol file without being
++	 loaded into the target (see add_symbol_file_command).  This case is
++	 such fallback to the file VMA address without the possibility of
++	 having the section relocated to its actual in-memory address.  */
++
++      dyn_addr = bfd_section_vma (abfd, sect);
++    }
+ 
+   /* Read in .dynamic from the BFD.  We will get the actual value
+      from memory later.  */
+@@ -639,7 +649,7 @@ scan_dyntag (int dyntag, bfd *abfd, CORE_ADDR *ptr)
+ 	     CORE_ADDR ptr_addr;
+ 
+ 	     ptr_type = builtin_type (target_gdbarch)->builtin_data_ptr;
+-	     ptr_addr = target_section->addr + (buf - bufstart) + arch_size / 8;
++	     ptr_addr = dyn_addr + (buf - bufstart) + arch_size / 8;
+ 	     if (target_read_memory (ptr_addr, ptr_buf, arch_size / 8) == 0)
+ 	       dyn_ptr = extract_typed_address (ptr_buf, ptr_type);
+ 	     *ptr = dyn_ptr;
+--- /dev/null
++++ b/gdb/testsuite/gdb.base/symbol-without-target_section.c
+@@ -0,0 +1,18 @@
++/* This testcase is part of GDB, the GNU debugger.
++
++   Copyright 2010 Free Software Foundation, Inc.
++
++   This program is free software; you can redistribute it and/or modify
++   it under the terms of the GNU General Public License as published by
++   the Free Software Foundation; either version 3 of the License, or
++   (at your option) any later version.
++
++   This program is distributed in the hope that it will be useful,
++   but WITHOUT ANY WARRANTY; without even the implied warranty of
++   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
++   GNU General Public License for more details.
++
++   You should have received a copy of the GNU General Public License
++   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
++
++int symbol_without_target_section;
+--- /dev/null
++++ b/gdb/testsuite/gdb.base/symbol-without-target_section.exp
+@@ -0,0 +1,39 @@
++# Copyright 2010 Free Software Foundation, Inc.
++
++# This program is free software; you can redistribute it and/or modify
++# it under the terms of the GNU General Public License as published by
++# the Free Software Foundation; either version 3 of the License, or
++# (at your option) any later version.
++#
++# This program is distributed in the hope that it will be useful,
++# but WITHOUT ANY WARRANTY; without even the implied warranty of
++# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
++# GNU General Public License for more details.
++#
++# You should have received a copy of the GNU General Public License
++# along with this program.  If not, see <http://www.gnu.org/licenses/>.
++
++# Exploit formerly failed assertion in scan_dyntag when it got called for
++# objfile having ".dynamic" section but without having data loaded in target.
++# Such file is ${binmainfile} through add-symbol-file here.  Set context first
++# by "list main" to have some local BLOCK set in lookup_symbol_global.
++
++set testfile symbol-without-target_section
++set srclibfile ${testfile}.c
++set binlibfile ${testfile}.x
++set srcmainfile start.c
++set binmainfile ${testfile}
++if { [gdb_compile "${srcdir}/${subdir}/${srclibfile}" \
++		  "${objdir}/${subdir}/${binlibfile}" object {debug}] != "" } {
++    untested ${testfile}.exp
++    return -1
++}
++if {[build_executable ${testfile}.exp ${binmainfile} ${srcmainfile} {debug}] == -1} {
++    return -1
++}
++clean_restart ${binlibfile}
++
++gdb_test "add-symbol-file ${objdir}/${subdir}/${binmainfile} 0" "" "add-symbol-file" \
++	 "add symbol table from file \".*\" at.*\\(y or n\\) " "y"
++gdb_test "list main"
++gdb_test "print symbol_without_target_section"
+

diff --git a/gdb.spec b/gdb.spec
index 3fd11bf..131fee9 100644
--- a/gdb.spec
+++ b/gdb.spec
@@ -36,7 +36,7 @@ Version: 7.0.50.20100128
 
 # The release always contains a leading reserved number, start it at 1.
 # `upstream' is not a part of `name' to stay fully rpm dependencies compatible for the testing.
-Release: 10%{?_with_upstream:.upstream}%{dist}
+Release: 11%{?_with_upstream:.upstream}%{dist}
 
 License: GPLv3+
 Group: Development/Debuggers
@@ -427,8 +427,8 @@ Patch408: gdb-ppc-power7-test.patch
 # Revert: Add -Wunused-function to compile flags.
 Patch412: gdb-unused-revert.patch
 
-# It may crash here but it is not understood why.
-Patch414: gdb-archer-pie-assert-temp-workaround.patch
+# Fix failed gdb_assert due to the PIE patchset (BZ 559414).
+Patch414: gdb-bz559414-pie-assert-fix.patch
 
 BuildRequires: ncurses-devel%{?_isa} texinfo gettext flex bison expat-devel%{?_isa}
 Requires: readline%{?_isa}
@@ -682,13 +682,10 @@ rm -f gdb/jv-exp.c gdb/m2-exp.c gdb/objc-exp.c gdb/p-exp.c
 %patch415 -p1
 %patch393 -p1
 %patch335 -p1
-# Patch414: gdb-archer-pie-assert-temp-workaround.patch
-# It may crash here but it is not understood why.
 # Patch415: gdb-6.6-buildid-locate-core-as-arg.patch
 # Currently disabled for RHEL as it is a new experimental feature not present
 # in FSF GDB and possibly affecting new user scripts.
 %if 0%{!?rhel:1}
-%patch414 -p1 -R
 %patch415 -p1 -R
 %endif
 %if 0%{!?el5:1}
@@ -1013,6 +1010,9 @@ fi
 %endif
 
 %changelog
+* Sun Jan 31 2010 Jan Kratochvil <jan.kratochvil@redhat.com> - 7.0.50.20100128-11.fc13
+- Fix failed gdb_assert due to the PIE patchset (BZ 559414).
+
 * Thu Jan 28 2010 Jan Kratochvil <jan.kratochvil@redhat.com> - 7.0.50.20100128-10.fc13
 - Upgrade to the FSF GDB snapshot: 7.0.50.20100128
 - archer-jankratochvil-fedora13 commit: 39c5a8b75fad3acd7204903db5dee025055a4594

^ permalink raw reply related	[flat|nested] only message in thread

only message in thread, other threads:[~2026-06-27 23:54 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-06-27 23:54 [rpms/gdb] gdb-17.2-rebase-f44: - Fix failed gdb_assert due to the PIE patchset (BZ 559414) Jan Kratochvil

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