bad patch - also not on list?

bad patch - also not on list?

[Tom Tromey]

I noticed that this commit:

commit b05efa39b47995db08c5537e4504271c8727702a
Author: Vsevolod Alekseyev <sevaa@sprynet.com>
Date:   Fri Nov 10 15:26:48 2023 +0000

    readelf..debug-dump=loc displays bogus base addresses
    
      PR 30880
      * dwarf.c (read_and_display_attr_value): Fix loclist handling. (display_loclists_list): Likewise.

... removes several hunks from display_gdb_index, causing readelf to be
incapable of reading v9 indices.

I think this was probably done in error.  However, I also could not find
this patch on the mailing list.  So, I wonder what's going in there.

I recommend reverting it.  However at the very least the regression
ought to be fixed.

thanks,
Tom

Re: bad patch - also not on list?

[Nick Clifton]

Hi Tom,

> I noticed that this commit:
> 
> commit b05efa39b47995db08c5537e4504271c8727702a

> ... removes several hunks from display_gdb_index, causing readelf to be
> incapable of reading v9 indices.

Ah, I did notice that and wonder.

> I think this was probably done in error.

I think so.  It looks like it might be just one delta that is causing
the problem:


@@ -10934,33 +10788,6 @@ display_gdb_index (struct dwarf_section *section,
  	}
      }

-  if (version >= 9)
-    {
-      printf (_("\nShortcut table:\n"));
-
-      if (shortcut_table_offset + 8 > constant_pool_offset)
-	{
-	  warn (_("Corrupt shortcut table in the %s section.\n"), section->name);
-	  return 0;
-	}
-
-      uint32_t lang = byte_get_little_endian (shortcut_table, 4);
-      printf (_("Language of main: "));
-      display_lang (lang);
-      printf ("\n");
-
-      uint32_t name_offset = byte_get_little_endian (shortcut_table + 4, 4);
-      printf (_("Name of main: "));
-      if (name_offset >= section->size - constant_pool_offset)
-	{
-	  printf (_("<corrupt offset: %x>\n"), name_offset);
-	  warn (_("Corrupt name offset of 0x%x found for name of main\n"),
-		name_offset);
-	}
-      else
-	printf ("%s\n", constant_pool + name_offset);
-    }
-
    return 1;
  }

If that code is restored, does v9 functionality then return ?


> However, I also could not find
> this patch on the mailing list.   

It was posted to the PR, and I thought that I had seen it on the
mailing list too.  But obviously I was wrong.  Sorry about that.

Cheers
   Nick

RE: bad patch - also not on list?

[Vsevolod Alekseyev]

I don't recall removing that fragment on purpose. Could've been an artifact of working with a less than latest version. Sorry for that.

To the best of my recollection, the shortcut table had nothing to do with the issue I was trying to address.

Please feel free to revert.



-----Original Message-----
From: Nick Clifton <nickc@redhat.com> 
Sent: Monday, November 20, 2023 9:33 AM
To: Tom Tromey <tom@tromey.com>; binutils@sourceware.org
Cc: Vsevolod Alekseyev <sevaa@sprynet.com>
Subject: Re: bad patch - also not on list?

Hi Tom,

> I noticed that this commit:
> 
> commit b05efa39b47995db08c5537e4504271c8727702a

> ... removes several hunks from display_gdb_index, causing readelf to 
> be incapable of reading v9 indices.

Ah, I did notice that and wonder.

> I think this was probably done in error.

I think so.  It looks like it might be just one delta that is causing the problem:


@@ -10934,33 +10788,6 @@ display_gdb_index (struct dwarf_section *section,
  	}
      }

-  if (version >= 9)
-    {
-      printf (_("\nShortcut table:\n"));
-
-      if (shortcut_table_offset + 8 > constant_pool_offset)
-	{
-	  warn (_("Corrupt shortcut table in the %s section.\n"), section->name);
-	  return 0;
-	}
-
-      uint32_t lang = byte_get_little_endian (shortcut_table, 4);
-      printf (_("Language of main: "));
-      display_lang (lang);
-      printf ("\n");
-
-      uint32_t name_offset = byte_get_little_endian (shortcut_table + 4, 4);
-      printf (_("Name of main: "));
-      if (name_offset >= section->size - constant_pool_offset)
-	{
-	  printf (_("<corrupt offset: %x>\n"), name_offset);
-	  warn (_("Corrupt name offset of 0x%x found for name of main\n"),
-		name_offset);
-	}
-      else
-	printf ("%s\n", constant_pool + name_offset);
-    }
-
    return 1;
  }

If that code is restored, does v9 functionality then return ?


> However, I also could not find
> this patch on the mailing list.   

It was posted to the PR, and I thought that I had seen it on the mailing list too.  But obviously I was wrong.  Sorry about that.

Cheers
   Nick

Re: bad patch - also not on list?

[Tom Tromey]

>>>>> "Nick" == Nick Clifton <nickc@redhat.com> writes:

Nick> I think so.  It looks like it might be just one delta that is causing
Nick> the problem:
...
Nick> If that code is restored, does v9 functionality then return ?

There were several hunks affecting that function, all of them have to be
fixed.  I can do it if you want.

Tom

Re: bad patch - also not on list?

[Nick Clifton]

Hi Tom,

> Nick> I think so.  It looks like it might be just one delta that is causing
> Nick> the problem:
> ...
> Nick> If that code is restored, does v9 functionality then return ?
> 
> There were several hunks affecting that function, all of them have to be
> fixed.  I can do it if you want.

Yes - please do.

Cheers
   Nick