[Bug gas/13684] New: Labels generated by .macro are assembled with 0x00000000 as its value in the body of the assembly.

2012-02-12 Thread Roland.Koluvek at gmail dot com
http://sourceware.org/bugzilla/show_bug.cgi?id=13684

 Bug #: 13684
   Summary: Labels generated by .macro are assembled with
0x as its value in the body of the assembly.
   Product: binutils
   Version: 2.21
Status: NEW
  Severity: critical
  Priority: P2
 Component: gas
AssignedTo: unassig...@sourceware.org
ReportedBy: roland.kolu...@gmail.com
Classification: Unclassified


Created attachment 6211
  -- http://sourceware.org/bugzilla/attachment.cgi?id=6211
A sample program to demonstrate the label problem.

The attachment will give you a clear example of the problem.

In my assembly I precede all labels with _.
The symbol table has the correct value, All .word _ are assembled to
0x. All branch addresses to any .macro generated label are the same
with incorrect offset.

I think it would be clear to you by looking at my attachment. But look at:
line 40 -- b _word3_long and a valid line 39 -- b _home.
line 46 -- b _word2, code generated is the same as line 40.  Again check
line 54 --. b _word1.  Now look at lines 56 to 59.
.word _home is ok, but .word _word2, .word _word1, and .word _word3_long
assembled to 0x.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
--- You are receiving this mail because: ---
You are on the CC list for the bug.

___
bug-binutils mailing list
bug-binutils@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-binutils


[Bug gas/13684] Labels generated by .macro are assembled with 0x00000000 as its value in the body of the assembly.

2012-02-12 Thread amodra at gmail dot com
http://sourceware.org/bugzilla/show_bug.cgi?id=13684

Alan Modra amodra at gmail dot com changed:

   What|Removed |Added

 Status|NEW |RESOLVED
 CC||amodra at gmail dot com
 Resolution||INVALID

--- Comment #1 from Alan Modra amodra at gmail dot com 2012-02-12 10:44:06 
UTC ---
Not a bug.  Please take a look at the object file output with objdump -dr.
You'll see at the end
 260:0200 .word0x0200
260: R_ARM_ABS32.text
...
264: R_ARM_ABS32_word2
268: R_ARM_ABS32_word1
26c: R_ARM_ABS32_word3_long

These relocations will be processed by the linker to insert the final value of
_word2, _word1 etc.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
--- You are receiving this mail because: ---
You are on the CC list for the bug.

___
bug-binutils mailing list
bug-binutils@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-binutils


[Bug ld/12742] ld error unrecognized option '-plugin' on mingw and how to build with plugins enabled on mingw

2012-02-12 Thread ktietz at redhat dot com
http://sourceware.org/bugzilla/show_bug.cgi?id=12742

--- Comment #4 from Kai Tietz ktietz at redhat dot com 2012-02-12 10:51:28 
UTC ---
Yes, with a dummy libdl.a (or with that version of dlfcn-win32 mentioned in
comment #1), bootstrap works.
I tried to find the cause for the -ldl adding on link, but couldn't find the
explicit point.  It might be reasoned by the fact that tcl.m4 code checks here
for uname, which is of course for candian-cross the wrong choice.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
--- You are receiving this mail because: ---
You are on the CC list for the bug.

___
bug-binutils mailing list
bug-binutils@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-binutils


[Bug binutils/13686] New: objdump -d does not recognise alternative encodings of lfence/mfence/sfence

2012-02-12 Thread m...@mythic-beasts.com
http://sourceware.org/bugzilla/show_bug.cgi?id=13686

 Bug #: 13686
   Summary: objdump -d does not recognise alternative encodings
of lfence/mfence/sfence
   Product: binutils
   Version: 2.23 (HEAD)
Status: NEW
  Severity: normal
  Priority: P2
 Component: binutils
AssignedTo: unassig...@sourceware.org
ReportedBy: m...@mythic-beasts.com
Classification: Unclassified


The canonical encoding of sfence is 0f ae f8, but the AMD manual says there
are 8 possible encodings of it, allowing any value in the 3-bit ModRM r/m field
(from Table A-7. Opcode 0F 01 and 0F AE ModRM Extensions):

0f ae f8
0f ae f9
0f ae fa
0f ae fb
0f ae fc
0f ae fd
0f ae fe
0f ae ff

objdump only recognises the canonical encoding:

   0:0f ae f8 sfence 
   3:0f ae(bad)  
   5:f9   stc
   6:0f ae(bad)  
   8:fa   cli
   9:0f ae(bad)  
   b:fb   sti
   c:0f ae(bad)  
   e:fc   cld
   f:0f ae(bad)  
  11:fd   std
  12:0f ae(bad)  
  14:fe 0fdecb   (%edi)
  16:ae   scas   %es:(%edi),%al
  17:ff   .byte 0xff

The same applies for lfence and mfence.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
--- You are receiving this mail because: ---
You are on the CC list for the bug.

___
bug-binutils mailing list
bug-binutils@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-binutils