[PATCH 0/2] Internal error for shared library with no .text section

[PATCH 0/2] Internal error for shared library with no .text section

[Andrew Burgess]

A Fedora GDB user ran into an issue with a shared library that had no
.text section triggering an internal error from GDB.

Patch #1 is a small refactor.

Patch #2 addresses the issue.

Thanks,
Andrew


---

Andrew Burgess (2):
  gdb: int to bool in struct entry_info
  gdb: don't use .text as default entry point section

 gdb/objfiles.h                                |  8 +-
 gdb/symfile.c                                 | 19 ++--
 .../gdb.base/bad-solib-entry-addr-lib.s       | 33 +++++++
 gdb/testsuite/gdb.base/bad-solib-entry-addr.c | 22 +++++
 .../gdb.base/bad-solib-entry-addr.exp         | 95 +++++++++++++++++++
 5 files changed, 165 insertions(+), 12 deletions(-)
 create mode 100644 gdb/testsuite/gdb.base/bad-solib-entry-addr-lib.s
 create mode 100644 gdb/testsuite/gdb.base/bad-solib-entry-addr.c
 create mode 100644 gdb/testsuite/gdb.base/bad-solib-entry-addr.exp


base-commit: cc4600477f621c764011b8e2f5bf1a2ebcffb61a
-- 
2.25.4

[PATCH 1/2] gdb: int to bool in struct entry_info

[Andrew Burgess]

Convert 'struct entry_info' to use bool for flag fields.  The places
where these flags are set are updated too.

There should be no user visible changes after this commit.
---
 gdb/objfiles.h | 8 ++++----
 gdb/symfile.c  | 8 ++++----
 2 files changed, 8 insertions(+), 8 deletions(-)

diff --git a/gdb/objfiles.h b/gdb/objfiles.h
index 89ac559ce81..429d8c82681 100644
--- a/gdb/objfiles.h
+++ b/gdb/objfiles.h
@@ -115,11 +115,11 @@ struct entry_info
   /* The index of the section in which the entry point appears.  */
   int the_bfd_section_index;
 
-  /* Set to 1 iff ENTRY_POINT contains a valid value.  */
-  unsigned entry_point_p : 1;
+  /* Set to true iff ENTRY_POINT contains a valid value.  */
+  bool entry_point_p : 1;
 
-  /* Set to 1 iff this object was initialized.  */
-  unsigned initialized : 1;
+  /* Set to true iff this object was initialized.  */
+  bool initialized : 1;
 };
 
 #define SECT_OFF_DATA(objfile) \
diff --git a/gdb/symfile.c b/gdb/symfile.c
index d0ea9506c63..d583960d430 100644
--- a/gdb/symfile.c
+++ b/gdb/symfile.c
@@ -776,7 +776,7 @@ init_entry_point_info (struct objfile *objfile)
 
   if (ei->initialized)
     return;
-  ei->initialized = 1;
+  ei->initialized = true;
 
   /* Save startup file's range of PC addresses to help blockframe.c
      decide where the bottom of the stack is.  */
@@ -786,7 +786,7 @@ init_entry_point_info (struct objfile *objfile)
       /* Executable file -- record its entry point so we'll recognize
 	 the startup file because it contains the entry point.  */
       ei->entry_point = bfd_get_start_address (objfile->obfd.get ());
-      ei->entry_point_p = 1;
+      ei->entry_point_p = true;
     }
   else if (bfd_get_file_flags (objfile->obfd.get ()) & DYNAMIC
 	   && bfd_get_start_address (objfile->obfd.get ()) != 0)
@@ -795,12 +795,12 @@ init_entry_point_info (struct objfile *objfile)
 	 runnable.  There's no clear way to indicate this, so just check
 	 for values other than zero.  */
       ei->entry_point = bfd_get_start_address (objfile->obfd.get ());
-      ei->entry_point_p = 1;
+      ei->entry_point_p = true;
     }
   else
     {
       /* Examination of non-executable.o files.  Short-circuit this stuff.  */
-      ei->entry_point_p = 0;
+      ei->entry_point_p = false;
     }
 
   if (ei->entry_point_p)
-- 
2.25.4

[PATCH 2/2] gdb: don't use .text as default entry point section

[Andrew Burgess]

We got a Fedora GDB bug report that a user tried to debug an Appimage,
and GDB would reliably crash like this:

  (gdb) run
  Starting program: /tmp/build/gdb/testsuite/outputs/gdb.base/bad-solib-entry-addr/bad-solib-entry-addr
  ../../src/gdb/symfile.c:843: internal-error: sect_index_text not initialized
  A problem internal to GDB has been detected,
  further debugging may prove unreliable.
  ----- Backtrace -----
  ... etc ...

The specific AppImage being debugged can be found here, I've modified
the URL with a warning marker.  I make no claims about whether it is
safe to download the image, and running it is definitely at your own
risk.  If you wish to, delete the warning marker to download:

  https://github.com/Murmele/Gittyup/<RUN AT YOUR OWN RISK>/releases/download/gittyup_v1.4.0/Gittyup-1.4.0-x86_64.AppImage

At the point of the above crash GDB's stack is:

  #9  0x000000000190c6ed in internal_error_loc (file=0x1e4a94e "../../src/gdb/symfile.c", line=838, fmt=0x1e4aa50 "sect_index_text not initialized") at ../../src/gdbsupport/errors.cc:57
  #10 0x0000000000f5f5ac in init_entry_point_info (objfile=0x5a98e80) at ../../src/gdb/symfile.c:838
  #11 0x0000000000f5f943 in syms_from_objfile (objfile=0x5a98e80, addrs=0x7ffd78728490, add_flags=...) at ../../src/gdb/symfile.c:962
  #12 0x0000000000f5fe6d in symbol_file_add_with_addrs (abfd=..., name=0x3e76e50 "/tmp/.mount_GittyujmIBkD/usr/bin/../../home/runner/work/Gittyup/Qt/5.15.2/gcc_64/lib/./libicudata.so.56", add_flags=..., addrs=0x7ffd78728490, flags=..., parent=0x0) at ../../src/gdb/symfile.c:1071
  #13 0x0000000000f601aa in symbol_file_add_from_bfd (abfd=..., name=0x3e76e50 "/tmp/.mount_GittyujmIBkD/usr/bin/../../home/runner/work/Gittyup/Qt/5.15.2/gcc_64/lib/./libicudata.so.56", add_flags=..., addrs=0x7ffd78728490, flags=..., parent=0x0) at ../../src/gdb/symfile.c:1145
  #14 0x0000000000f0f2ad in solib_read_symbols (so=..., flags=...) at ../../src/gdb/solib.c:627
  #15 0x0000000000f10263 in solib_add (pattern=0x0, from_tty=0, readsyms=1) at ../../src/gdb/solib.c:960

From this we can see GDB is trying to add the shared library:

  /tmp/.mount_GittyujmIBkD/usr/bin/../../home/runner/work/Gittyup/Qt/5.15.2/gcc_64/lib/./libicudata.so.56

The internal error is triggered from these lines in
init_entry_point_info:

  if (!found)
    ei->the_bfd_section_index = SECT_OFF_TEXT (objfile);

Where SECT_OFF_TEXT is:

  #define SECT_OFF_TEXT(objfile) \
     ((objfile->sect_index_text == -1) \
      ? (internal_error (_("sect_index_text not initialized")), -1)	\
      : objfile->sect_index_text)

So we can see that objfile::sect_index_text is -1, which leads to the
internal error.

Looking at the 'readelf -Wa ...' output for the shared library in
question we see this:

  ELF Header:
    Magic:   7f 45 4c 46 02 01 01 00 00 00 00 00 00 00 00 00
    Class:                             ELF64
    Data:                              2's complement, little endian
    Version:                           1 (current)
    OS/ABI:                            UNIX - System V
    ABI Version:                       0
    Type:                              DYN (Shared object file)
    Machine:                           Advanced Micro Devices X86-64
    Version:                           0x1
    Entry point address:               0x2d7
    Start of program headers:          25051136 (bytes into file)
    Start of section headers:          25047552 (bytes into file)
    Flags:                             0x0
    Size of this header:               64 (bytes)
    Size of program headers:           56 (bytes)
    Number of program headers:         7
    Size of section headers:           64 (bytes)
    Number of section headers:         11
    Section header string table index: 7

  Section Headers:
    [Nr] Name              Type            Address          Off    Size   ES Flg Lk Inf Al
    [ 0]                   NULL            0000000000000000 000000 000000 00      0   0  0
    [ 1] .note.gnu.build-id NOTE            0000000000000190 000190 000024 00   A  0   0  4
    [ 2] .gnu.hash         GNU_HASH        00000000000001b8 0001b8 000034 00   A  3   0  8
    [ 3] .dynsym           DYNSYM          00000000000001f0 0001f0 000090 18   A 10   2  8
    [ 4] .rodata           PROGBITS        00000000000002e0 0002e0 17e27d0 00   A  0   0 16
    [ 5] .eh_frame         PROGBITS        00000000017e2ab0 17e2ab0 000000 00   A  0   0  8
    [ 6] .dynamic          DYNAMIC         00000000019e2f10 17e2f10 0000f0 10  WA 10   0  8
    [ 7] .shstrtab         STRTAB          0000000000000000 17e3000 000063 00      0   0  1
    [ 8] .symtab           SYMTAB          0000000000000000 17e3068 000150 18      9  10  8
    [ 9] .strtab           STRTAB          0000000000000000 17e31b8 000044 00      0   0  1
    [10] .dynstr           STRTAB          00000000019e3188 17e4188 000420 00   A  0   0  8
  Key to Flags:
    W (write), A (alloc), X (execute), M (merge), S (strings), I (info),
    L (link order), O (extra OS processing required), G (group), T (TLS),
    C (compressed), x (unknown), o (OS specific), E (exclude),
    D (mbind), l (large), p (processor specific)

  There are no section groups in this file.

  Program Headers:
    Type           Offset   VirtAddr           PhysAddr           FileSiz  MemSiz   Flg Align
    LOAD           0x000000 0x0000000000000000 0x0000000000000000 0x17e2ab0 0x17e2ab0 R   0x200000
    GNU_STACK      0x000000 0x0000000000000000 0x0000000000000000 0x000000 0x000000 RW  0x10
    NOTE           0x000190 0x0000000000000190 0x0000000000000190 0x000024 0x000024 R   0x4
    LOAD           0x17e2f10 0x00000000019e2f10 0x00000000019e2f10 0x0000f0 0x0000f0 RW  0x200000
    DYNAMIC        0x17e2f10 0x00000000019e2f10 0x00000000019e2f10 0x0000f0 0x0000f0 RW  0x8
    GNU_RELRO      0x17e2f10 0x00000000019e2f10 0x00000000019e2f10 0x0000f0 0x0000f0 R   0x1
    LOAD           0x17e4000 0x00000000019e3000 0x00000000019e3000 0x0005a8 0x0005a8 RW  0x1000

Things to note here are:

  1. There really is no .text section, or any executable sections,

  2. there are 3 LOAD segments.  This will be important later, and

  3. the "Entry point address" is outside all sections, and is
     non-zero.

Next we can investigate where objfile::sect_index_text is set to
something other than -1.  Starting in init_objfile_sect_indices, if
the objfile has a ".text" section then sect_index_text can be set.
This case clearly doesn't apply.

Next symfile_find_segment_sections is called.  This tries to match a
common case where we have either 1 or 2 LOAD segments, and assumes a
default distribution of sections to segments.  However, we have 3 LOAD
segments, so these lines:

  if (data->segments.size () != 1 && data->segments.size () != 2)
    return;

result in an early return from symfile_find_segment_sections without
sect_index_text being set.

Back in init_objfile_sect_indices, if no sections have an offset then
we set any currently unset sect_index_* values, including
sect_index_text, to point at section 0.  However, in our case the
objfile is a relocatable shared library, so the sections will have an
offset, and so this final fallback case doesn't apply.

The result is that init_objfile_sect_indices never sets
sect_index_text.  This worries me a little as
init_objfile_sect_indices contains this comment:

  /* This is where things get really weird...  We MUST have valid
     indices for the various sect_index_* members or gdb will abort.
     So if for example, there is no ".text" section, we have to
     accommodate that.  First, check for a file with the standard
     one or two segments.  */

Notice the emphasis on MUST in that comment, and indeed, we exit this
function without setting sect_index_text, and GDB does indeed abort.
The comment seems to imply that the following code is going to try to
figure out a suitable stand-in sect_index_text for when there is no
".text" section, but clearly I've run into a case that isn't covered.

All of this code relating to setting sect_index_text was introduced in
commit:

  commit 31d99776c73d6fca13163da59c852b0fa99f89b8
  Date:   Mon Jun 18 15:46:38 2007 +0000

Which unfortunately is from a time where we didn't write useful commit
messages, so to understand the commit you need to go read the mailing
list archive, but they don't offer much more insight:

  https://sourceware.org/pipermail/gdb-patches/2007-May/050527.html

Clearly the comment in init_objfile_sect_indices would suggest that
the fix here is to figure out some "fake" value for sect_index_text,
and that would certainly avoid the problem here.  But, at least for
this problem, I think there's maybe a better solution.

The original internal error is triggered by a use of SECT_OFF_TEXT in
init_entry_point_info.  We have an entry point address, we try to find
the section index for the section containing the entry point, and
failing that, we assume the entry point is in the text section.  This
fall-back assumption means that, if the text section has an offset
applied, then the entry point will also have that same offset
applied.  But it's not clear to me why picking the text section is
going to be any more valid than any other section, especially in a
case like this where we don't even have a text section, so the
sect_index_text might itself point to some other arbitrary section.

Earlier in init_entry_point_info we already have a fall-back case
where we set entry_info::entry_point_p to false to indicate that the
objfile has no entry point, so this is always a possibility.  So I
wondered about writing something like:

  if (!found)
    {
      if (objfile->sect_index_text != -1)
	ei->the_bfd_section_index = SECT_OFF_TEXT (objfile);
      else
        ei->entry_point_p = false;
    }

If we have no text section index then we just claim the objfile has no
entry point.  But I didn't like this for two reasons, first, the
comment back in init_objfile_sect_indices saying that the index should
be set, this seems to indicate that we should not be making decisions
later within GDB based on whether the index is set or not.

And second, using the text section as a fall back, when the entry
address is outside every section, just seems off.  So I wondered, why
not just reject the entry point completely in this case?  Which is how
I ended up with:

  if (!found)
    ei->entry_point_p = false;

With this patch in place I was able to start debugging the AppImage
linked above.

I created a simple test case which reproduces this issue.  It's a
little contrived because it has to hit all the points required to
trigger this bug:

  1. No .text section,

  2. more than 2 LOAD segments, and

  3. entry address outside every section.

I have no idea what caused the original shared library to take on
these characteristics, it might even be a tool issue building the
original shared library.  I haven't investigated this, as I don't
think it really matters, GDB shouldn't be crashing just because the
incoming objects are a little weird.

I've attached a link to the Fedora bug in the 'Bug:' tag, but it's a
little confusing.  An automated system has merged together two bug
reports.  As such the overall bug report linked too is for a
completely different issue, only comments 21, 22, 23, and 24 relate to
the bug being fixed here.

Bug: https://bugzilla.redhat.com/show_bug.cgi?id=2366461#c21
---
 gdb/symfile.c                                 | 11 ++-
 .../gdb.base/bad-solib-entry-addr-lib.s       | 33 +++++++
 gdb/testsuite/gdb.base/bad-solib-entry-addr.c | 22 +++++
 .../gdb.base/bad-solib-entry-addr.exp         | 95 +++++++++++++++++++
 4 files changed, 157 insertions(+), 4 deletions(-)
 create mode 100644 gdb/testsuite/gdb.base/bad-solib-entry-addr-lib.s
 create mode 100644 gdb/testsuite/gdb.base/bad-solib-entry-addr.c
 create mode 100644 gdb/testsuite/gdb.base/bad-solib-entry-addr.exp

diff --git a/gdb/symfile.c b/gdb/symfile.c
index d583960d430..0413f7f7ce1 100644
--- a/gdb/symfile.c
+++ b/gdb/symfile.c
@@ -806,7 +806,6 @@ init_entry_point_info (struct objfile *objfile)
   if (ei->entry_point_p)
     {
       CORE_ADDR entry_point =  ei->entry_point;
-      int found;
 
       /* Make certain that the address points at real code, and not a
 	 function descriptor.  */
@@ -818,7 +817,7 @@ init_entry_point_info (struct objfile *objfile)
       ei->entry_point
 	= gdbarch_addr_bits_remove (objfile->arch (), entry_point);
 
-      found = 0;
+      bool found = false;
       for (obj_section &osect : objfile->sections ())
 	{
 	  struct bfd_section *sect = osect.the_bfd_section;
@@ -829,13 +828,17 @@ init_entry_point_info (struct objfile *objfile)
 	    {
 	      ei->the_bfd_section_index
 		= gdb_bfd_section_index (objfile->obfd.get (), sect);
-	      found = 1;
+	      found = true;
 	      break;
 	    }
 	}
 
+      /* We store the section index so that the entry address can be
+	 relocated when used.  If the entry address is outside of any
+	 section then we cannot relocate it.  Just claim that there is no
+	 entry address in this case.  */
       if (!found)
-	ei->the_bfd_section_index = SECT_OFF_TEXT (objfile);
+	ei->entry_point_p = false;
     }
 }
 
diff --git a/gdb/testsuite/gdb.base/bad-solib-entry-addr-lib.s b/gdb/testsuite/gdb.base/bad-solib-entry-addr-lib.s
new file mode 100644
index 00000000000..595a5c5e132
--- /dev/null
+++ b/gdb/testsuite/gdb.base/bad-solib-entry-addr-lib.s
@@ -0,0 +1,33 @@
+/* Copyright 2026 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/>.  */
+
+	.section .rodata
+	.globl lib_var
+	.type lib_var, @object
+lib_var:
+	.long 42
+	.size lib_var, .-lib_var
+
+	/* This executable section exists solely to force the linker
+	   to create an additional LOAD segment (with R+X permissions),
+	   giving the library 3+ LOAD segments instead of the default 2.
+
+	   This matters because GDB's symfile_find_segment_sections only
+	   runs for objects with exactly 1 or 2 segments.  When it runs,
+	   it sets sect_index_text from the first section in segment 1,
+	   which masks the bug we are testing for, see the .exp file for
+	   details.  */
+	.section .not_text, "ax", @progbits
+	.byte 0
diff --git a/gdb/testsuite/gdb.base/bad-solib-entry-addr.c b/gdb/testsuite/gdb.base/bad-solib-entry-addr.c
new file mode 100644
index 00000000000..9c5c7b491e0
--- /dev/null
+++ b/gdb/testsuite/gdb.base/bad-solib-entry-addr.c
@@ -0,0 +1,22 @@
+/* Copyright 2026 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/>.  */
+
+extern int lib_var;
+
+int
+main ()
+{
+  return lib_var;
+}
diff --git a/gdb/testsuite/gdb.base/bad-solib-entry-addr.exp b/gdb/testsuite/gdb.base/bad-solib-entry-addr.exp
new file mode 100644
index 00000000000..da82066ee03
--- /dev/null
+++ b/gdb/testsuite/gdb.base/bad-solib-entry-addr.exp
@@ -0,0 +1,95 @@
+# Copyright 2026 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/>.
+
+# This test aims to exercise a specific situation which was seen
+# causing GDB to crash.  An application has a shared library that
+# meets the following conditions:
+#
+# 1. No ".text" section,
+# 2. at least 3 LOAD segments, and
+# 3. a non-zero entry address that is outside of every section.
+#
+# When these 3 conditions are met then GDB would run into a problem in
+# init_entry_point_info for the shared library.  The non-zero entry
+# address means GDB would try to find the section corresponding to the
+# address.  As the address is outside of all sections then GDB would
+# try to use the .text section as a fall-back.  But there is no text
+# section, so an internal error would be triggered.
+#
+# The 3 LOAD segments is important because, if there are only 1 or 2
+# segments GDB has a default in symfile_find_segment_sections where
+# it assumes the text section is the first section in the first
+# segment, which means init_entry_point_info will have a .text section
+# to use.
+#
+# This test has a very simple assembler file (which hopefully contains
+# no architecture specific content), which will compile to a couple of
+# sections, but no .text section.  These sections force the creation
+# of more than 2 LOAD segments.  The compiler flags then set the entry
+# address to 0x1, which (we hope) is outside all sections.  This
+# assembler file is compiled into a shared library.
+
+require allow_shlib_tests
+require !use_gdb_stub
+require {istarget *-linux*}
+
+standard_testfile .c -lib.s
+
+# Build shared library.
+set lib_testfile ${testfile}-lib.so
+set lib_srcfile ${srcfile2}
+set lib_binfile [standard_output_file ${lib_testfile}]
+if { [build_executable "build solib" $lib_testfile $lib_srcfile \
+	  {shlib additional_flags=-nostartfiles additional_flags=-Wl,-e,0x1}] == -1 } {
+    return
+}
+
+# Build the test executable.
+if { [build_executable "build exec" $testfile $srcfile \
+	  [list debug pie shlib=${lib_binfile}]] == -1 } {
+    return
+}
+
+# Confirm we have more than 2 LOAD segments in the shared library.
+set readelf_program [gdb_find_readelf]
+set command "exec $readelf_program -Wl $lib_binfile"
+verbose -log "command is $command"
+set result [catch {{*}$command} output]
+verbose -log "result is $result"
+verbose -log "output is $output"
+if {$result != 0} {
+    fail "read program headers from $lib_testfile"
+    return
+}
+if {![regexp {\nProgram Headers:\n *Type [^\n]* Align\n(.*?)\n\n} $output trash phdr]} {
+    fail "no Program Headers found"
+    return
+}
+set load_segment_count 0
+foreach line [regexp -line -all -inline {^ *LOAD .* 0x[0-9]+$} $phdr] {
+    incr load_segment_count
+}
+if { $load_segment_count <= 2 } {
+    fail "not enough LOAD segments"
+    return
+}
+
+# Start GDB and run to main.  If the shared library is causing issues
+# then we will see an internal error once the inferior starts running.
+clean_restart $testfile
+
+if {![runto_main message]} {
+  return 0
+}
-- 
2.25.4