Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Sandra Loosemore <sandra@codesourcery.com>
To: <gdb-patches@sourceware.org>
Cc: Yao Qi <qiyaoltc@gmail.com>
Subject: [patch, dwarf2] avoid segfault on missing directory table
Date: Fri, 08 May 2015 17:10:00 -0000	[thread overview]
Message-ID: <554CEDD3.80407@codesourcery.com> (raw)

[-- Attachment #1: Type: text/plain, Size: 731 bytes --]

A while back we were given a nios2-elf executable that caused GDB to 
segfault while reading its debug information.  The binary turned out to 
have invalid DWARF-2 information in the .debug_line section: although 
the file name table had references to entries in the directory table, 
the directory table itself was empty.  The executable was produced by 
some very old version of GCC (4.1.2?), and we verified that more current 
toolchain versions don't produce such bad debug information any more. 
But, since it's generally a bad thing for GDB to segfault, here is a 
patch that makes the DWARF-2 reader more robust by making sure the 
directory table is non-NULL before trying to access entries in it.

OK to commit?

-Sandra


[-- Attachment #2: gdb-dwarf.log --]
[-- Type: text/x-log, Size: 392 bytes --]

2015-05-07  Yao Qi  <yao@codesourcery.com>
	    Sandra Loosemore  <sandra@codesourcery.com>

	Avoid segfault on missing directory table.

	gdb/
	* dwarf2read.c (setup_type_unit_groups): Do NULL pointer check
	to 'lh->include_dirs' before accessing to it.
	(psymtab_include_file_name): Likewise.
	(dwarf_decode_lines_1): Likewise.
	(dwarf_decode_lines): Likewise.
	(file_file_name): Likewise.

[-- Attachment #3: gdb-dwarf.patch --]
[-- Type: text/x-patch, Size: 2328 bytes --]

diff --git a/gdb/dwarf2read.c b/gdb/dwarf2read.c
index 4982922..e2ea7e2 100644
--- a/gdb/dwarf2read.c
+++ b/gdb/dwarf2read.c
@@ -9320,7 +9320,7 @@ setup_type_unit_groups (struct die_info *die, struct dwarf2_cu *cu)
 	  const char *dir = NULL;
 	  struct file_entry *fe = &lh->file_names[i];
 
-	  if (fe->dir_index)
+	  if (fe->dir_index && lh->include_dirs != NULL)
 	    dir = lh->include_dirs[fe->dir_index - 1];
 	  dwarf2_start_subfile (fe->name, dir);
 
@@ -17396,7 +17396,7 @@ psymtab_include_file_name (const struct line_header *lh, int file_index,
   char *copied_name = NULL;
   int file_is_pst;
 
-  if (fe.dir_index)
+  if (fe.dir_index && lh->include_dirs != NULL)
     dir_name = lh->include_dirs[fe.dir_index - 1];
 
   if (!IS_ABSOLUTE_PATH (include_name)
@@ -17595,7 +17595,7 @@ dwarf_decode_lines_1 (struct line_header *lh, struct dwarf2_cu *cu,
           struct file_entry *fe = &lh->file_names[file - 1];
           const char *dir = NULL;
 
-          if (fe->dir_index)
+          if (fe->dir_index && lh->include_dirs != NULL)
             dir = lh->include_dirs[fe->dir_index - 1];
 
 	  dwarf2_start_subfile (fe->name, dir);
@@ -17815,7 +17815,7 @@ dwarf_decode_lines_1 (struct line_header *lh, struct dwarf2_cu *cu,
                 else
                   {
                     fe = &lh->file_names[file - 1];
-                    if (fe->dir_index)
+                    if (fe->dir_index && lh->include_dirs != NULL)
                       dir = lh->include_dirs[fe->dir_index - 1];
                     if (!decode_for_pst_p)
                       {
@@ -17958,7 +17958,7 @@ dwarf_decode_lines (struct line_header *lh, const char *comp_dir,
 	  struct file_entry *fe;
 
 	  fe = &lh->file_names[i];
-	  if (fe->dir_index)
+	  if (fe->dir_index && lh->include_dirs != NULL)
 	    dir = lh->include_dirs[fe->dir_index - 1];
 	  dwarf2_start_subfile (fe->name, dir);
 
@@ -20640,7 +20640,8 @@ file_file_name (int file, struct line_header *lh)
     {
       struct file_entry *fe = &lh->file_names[file - 1];
 
-      if (IS_ABSOLUTE_PATH (fe->name) || fe->dir_index == 0)
+      if (IS_ABSOLUTE_PATH (fe->name) || fe->dir_index == 0
+	  || lh->include_dirs == NULL)
         return xstrdup (fe->name);
       return concat (lh->include_dirs[fe->dir_index - 1], SLASH_STRING,
 		     fe->name, NULL);

             reply	other threads:[~2015-05-08 17:10 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-05-08 17:10 Sandra Loosemore [this message]
2015-05-08 17:23 ` Joel Brobecker

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=554CEDD3.80407@codesourcery.com \
    --to=sandra@codesourcery.com \
    --cc=gdb-patches@sourceware.org \
    --cc=qiyaoltc@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox