binutils 2.20.1

binutils 2.20.1

[Kenny Simpson]

When is 2.20.1 expected out?  Is there a list of issues resolved in this version?

thanks,
-Kenny



      

Re: binutils 2.20.1

[Tristan Gingold]

On Feb 3, 2010, at 8:15 AM, Kenny Simpson wrote:

> When is 2.20.1 expected out?

Soon.  There is still one issue to be solved (ARM related).

>  Is there a list of issues resolved in this version?

No.  You can refer to bugzilla or ChangeLogs.

Tristan.

Re: binutils 2.20.1

[Christophe LYON]

On 03.02.2010 14:08, Tristan Gingold wrote:
>
> On Feb 3, 2010, at 8:15 AM, Kenny Simpson wrote:
>
>> When is 2.20.1 expected out?
>
> Soon.  There is still one issue to be solved (ARM related).
>

Are you talking about
http://sourceware.org/ml/binutils/2010-01/msg00342.html ?

If so, sorry for the delay I am still working on it, but during this 
process I have discovered other issues.

In short, I think that
http://sourceware.org/ml/binutils/2009-03/msg00511.html
has broken thumb->arm intersection branches (and probably vice-versa). I 
have starting working on a patch I intend to propose for this one before 
finishing the other patch.

Christophe.

Re: binutils 2.20.1

[Tristan Gingold]

On Feb 3, 2010, at 2:22 PM, Christophe LYON wrote:

> On 03.02.2010 14:08, Tristan Gingold wrote:
>> 
>> On Feb 3, 2010, at 8:15 AM, Kenny Simpson wrote:
>> 
>>> When is 2.20.1 expected out?
>> 
>> Soon.  There is still one issue to be solved (ARM related).
>> 
> 
> Are you talking about
> http://sourceware.org/ml/binutils/2010-01/msg00342.html ?

No, I was thinking about:

http://sourceware.org/ml/binutils/2010-02/msg00013.html

As far as I know, the 'ARM stub sizing fix' hasn't been merged on the branch.  If this was your attempt,
please tell us.

> If so, sorry for the delay I am still working on it, but during this process I have discovered other issues.
> 
> In short, I think that
> http://sourceware.org/ml/binutils/2009-03/msg00511.html
> has broken thumb->arm intersection branches (and probably vice-versa). I have starting working on a patch I intend to propose for this one before finishing the other patch.

Ok.  Do you plan to propose the fix for the 2.20 branch ?
(This is really up to you.  I think there is no hurry but a rough schedule would be nice)

Tristan.

Re: binutils 2.20.1

[Richard Earnshaw]

On Wed, 2010-02-03 at 14:33 +0100, Tristan Gingold wrote:
> On Feb 3, 2010, at 2:22 PM, Christophe LYON wrote:
> 
> > On 03.02.2010 14:08, Tristan Gingold wrote:
> >> 
> >> On Feb 3, 2010, at 8:15 AM, Kenny Simpson wrote:
> >> 
> >>> When is 2.20.1 expected out?
> >> 
> >> Soon.  There is still one issue to be solved (ARM related).
> >> 
> > 
> > Are you talking about
> > http://sourceware.org/ml/binutils/2010-01/msg00342.html ?
> 
> No, I was thinking about:
> 
> http://sourceware.org/ml/binutils/2010-02/msg00013.html
> 
> As far as I know, the 'ARM stub sizing fix' hasn't been merged on the branch.  If this was your attempt,
> please tell us.
> 

Matt posted a fix for that, but it hasn't been approved yet.  It looks
fine to me, but Matt can't commit it and I'm snowed under with other
stuff at the moment.

R.

Re: binutils 2.20.1

[Tristan Gingold]

On Feb 3, 2010, at 2:40 PM, Richard Earnshaw wrote:

> 
> On Wed, 2010-02-03 at 14:33 +0100, Tristan Gingold wrote:
>> On Feb 3, 2010, at 2:22 PM, Christophe LYON wrote:
>> 
>>> On 03.02.2010 14:08, Tristan Gingold wrote:
>>>> 
>>>> On Feb 3, 2010, at 8:15 AM, Kenny Simpson wrote:
>>>> 
>>>>> When is 2.20.1 expected out?
>>>> 
>>>> Soon.  There is still one issue to be solved (ARM related).
>>>> 
>>> 
>>> Are you talking about
>>> http://sourceware.org/ml/binutils/2010-01/msg00342.html ?
>> 
>> No, I was thinking about:
>> 
>> http://sourceware.org/ml/binutils/2010-02/msg00013.html
>> 
>> As far as I know, the 'ARM stub sizing fix' hasn't been merged on the branch.  If this was your attempt,
>> please tell us.
>> 
> 
> Matt posted a fix for that, but it hasn't been approved yet.  It looks
> fine to me, but Matt can't commit it and I'm snowed under with other
> stuff at the moment.

Thanks.  I will do it.

Tristan.

Re: binutils 2.20.1

[Ramana Radhakrishnan]

On Wed, 2010-02-03 at 14:22 +0100, Christophe LYON wrote:
> On 03.02.2010 14:08, Tristan Gingold wrote:
> >
> > On Feb 3, 2010, at 8:15 AM, Kenny Simpson wrote:
> >
> >> When is 2.20.1 expected out?
> >
> > Soon.  There is still one issue to be solved (ARM related).
> >
> 
> Are you talking about
> http://sourceware.org/ml/binutils/2010-01/msg00342.html ?
> 
> If so, sorry for the delay I am still working on it, but during this 
> process I have discovered other issues.
> 
> In short, I think that
> http://sourceware.org/ml/binutils/2009-03/msg00511.html

Is there a testcase that shows a problem caused by this patch ? I'd be
interested in hearing about it as the original author of that patch and
helping in any way possible in fixing the breakage if any. 

cheers
Ramana

Re: binutils 2.20.1

[Christophe LYON]

>> In short, I think that
>> http://sourceware.org/ml/binutils/2009-03/msg00511.html
>
> Is there a testcase that shows a problem caused by this patch ? I'd be
> interested in hearing about it as the original author of that patch and
> helping in any way possible in fixing the breakage if any.
>

I start a new thread for that, even though I was not quite ready to post 
it ;-)

Christophe.

Re: binutils 2.20.1

[Alan Modra]

http://sourceware.org/ml/binutils/2010-01/msg00360.html needs an ARM
maintainer to review and adjust the ARM testsuite.

-- 
Alan Modra
Australia Development Lab, IBM

Re: Binutils 2.20.1

[Tristan Gingold]

On Jan 6, 2010, at 2:21 PM, Chris Sutcliffe wrote:

>> I (my fault) haven't actively tracked all the patches that are candidate for the release branch but
>> many have been committed and this is time to ask if you want a particular patch to be put in the next
>> minor release.
> 
> I'd appreciate it if the fix for the bug described here:
> 
> https://bugs.launchpad.net/ubuntu/+source/binutils/+bug/461303
> 
> And subsequently linked to by:
> 
> https://sourceforge.net/tracker/?func=detail&aid=2913876&group_id=2435&atid=102435
> 
> Could be included.

Hi,

this one has already been backported.

Tristan.

Re: Binutils 2.20.1

[Tristan Gingold]

On Jan 5, 2010, at 8:34 PM, Bob Plantz wrote:

> On Tue, 2010-01-05 at 14:12 +0100, Tristan Gingold wrote:
>> Hi,
>> 
>> there was a number of glitches in the binutils release 2.20 so it makes sense to make a new minor release.
>> 
>> I (my fault) haven't actively tracked all the patches that are candidate for the release branch but
>> many have been committed and this is time to ask if you want a particular patch to be put in the next
>> minor release.
>> 
>> I'd like to release 2.20.1 before February.
>> 
>> Tristan. 
> 
> I apologize for not knowing how to get glitches fixed in gas, but I
> would like to see a minor problem fixed. I filed bug report #11122 with
> bugzilla about this.
> 
> The problem arises when compiling a C source file and using -Wa,-adhl to
> get a C source listing with the assembly language. The C source lines
> are numbered from 0. They should be numbered from 1. The .loc gas
> directives correctly number from 1.

Hi,

this bug has been fixed (thanks to Alan Modra) and backported to the branch.

Tristan.

Re: Binutils 2.20.1

[Tristan Gingold]

On Jan 5, 2010, at 4:20 PM, Matthias Klose wrote:

> On 05.01.2010 14:12, Tristan Gingold wrote:
> 
> could the following two arm patches applied to the branch?
> 
> opcodes/
> 
> 2009-11-17  Edward Nevill <edward.nevill@arm.com>
> 
>        * arm-dis.c (print_insn_thumb32): Handle undefined instruction.

Hi,

I applied this patch.

> gas/
> 2009-12-21  Ramana Radhakrishnan  <ramana.radhakrishnan@arm.com>
>            Richard Earnshaw  <richard.earnshaw@arm.com>
> 
>        * config/tc-arm.c (encode_thumb2_b_bl_offset): New. Refactored
>        from md_apply_fix.
>        (md_apply_fix): Fixup range checks for Thumb2 version
>        of unconditional calls. Call encode_thumb2_b_bl_offset for
>        unconditional branches / function calls.

This one was already backported.

Thanks,
Tristan.

Re: Binutils 2.20.1

[Tristan Gingold]

On Jan 5, 2010, at 3:39 PM, Matthias Klose wrote:

> On 05.01.2010 14:12, Tristan Gingold wrote:
>> I (my fault) haven't actively tracked all the patches that are candidate for the release branch but
>> many have been committed and this is time to ask if you want a particular patch to be put in the next
>> minor release.
> 
> a fix for 10144, if available.

Hi,

according to bugzilla, 10144 is still open.

> are any gold updates planned for 2.20.1? For a rebuild test with gold as the default linker I would like to see a fix for 11072 on the branch (objcopy --add-gnu-debuglink is used a lot for package building). Other missing features/options are summarized by Peter Fritzsche in [1].

In addition to Ian reply, I would say that gold would be slightly improved because 2.20 was builded before
a batch of commit.

> is there a chance that the --enable-gold= patch is approved for the trunk and backported to the branch?

The patch is not yet in mainline.

Tristan.

Re: Binutils 2.20.1

[Ian Lance Taylor]

[-- Attachment #1: Type: text/plain, Size: 1201 bytes --]

Matthias Klose <doko@ubuntu.com> writes:

> are any gold updates planned for 2.20.1? For a rebuild test with gold
> as the default linker I would like to see a fix for 11072 on the
> branch (objcopy --add-gnu-debuglink is used a lot for package
> building). Other missing features/options are summarized by Peter
> Fritzsche in [1].

Since you asked, I have committed the patch for PR 11072 to the
binutils 2.20 branch.

I believe that all the options mentioned in Peter Fritzsche's Debian
bug report are now supported in the mainline version of gold.  But I'm
a little reluctant to simply copy mainline into the 2.20 branch as
they have had very little testing.  I was willing to risk total
breakage for 2.20.0 but I don't want 2.20.1 to be worse than 2.20.0.


> is there a chance that the --enable-gold= patch is approved for the
> trunk and backported to the branch?

I think this patch is still pending approval by a binutils maintainer.

Ian


11072 patch committed to 2.20 branch:


2010-01-13  Ian Lance Taylor  <iant@google.com>

	Bring over from mainline:
	2010-01-08  Ian Lance Taylor  <iant@google.com>

	PR 11072
	* layout.cc (Layout::include_section): Remove .gnu_debuglink
	sections.



[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: debuglink --]
[-- Type: text/x-diff, Size: 989 bytes --]

Index: layout.cc
===================================================================
RCS file: /cvs/src/src/gold/layout.cc,v
retrieving revision 1.134.2.3
diff -p -u -r1.134.2.3 layout.cc
--- layout.cc	5 Nov 2009 06:25:01 -0000	1.134.2.3
+++ layout.cc	13 Jan 2010 15:36:13 -0000
@@ -1,6 +1,6 @@
 // layout.cc -- lay out output file sections for gold
 
-// Copyright 2006, 2007, 2008, 2009 Free Software Foundation, Inc.
+// Copyright 2006, 2007, 2008, 2009, 2010 Free Software Foundation, Inc.
 // Written by Ian Lance Taylor <iant@google.com>.
 
 // This file is part of gold.
@@ -349,6 +349,11 @@ Layout::include_section(Sized_relobj<siz
           if (is_prefix_of(".gnu.lto_", name))
             return false;
         }
+      // The GNU linker strips .gnu_debuglink sections, so we do too.
+      // This is a feature used to keep debugging information in
+      // separate files.
+      if (strcmp(name, ".gnu_debuglink") == 0)
+	return false;
       return true;
 
     default:

Re: Binutils 2.20.1

[Chris Sutcliffe]

> I (my fault) haven't actively tracked all the patches that are candidate for the release branch but
> many have been committed and this is time to ask if you want a particular patch to be put in the next
> minor release.

I'd appreciate it if the fix for the bug described here:

https://bugs.launchpad.net/ubuntu/+source/binutils/+bug/461303

And subsequently linked to by:

https://sourceforge.net/tracker/?func=detail&aid=2913876&group_id=2435&atid=102435

Could be included.

Thank you,

Chris

-- 
Chris Sutcliffe
http://emergedesktop.org

Re: Binutils 2.20.1

[Bob Plantz]

On Tue, 2010-01-05 at 14:12 +0100, Tristan Gingold wrote:
> Hi,
> 
> there was a number of glitches in the binutils release 2.20 so it makes sense to make a new minor release.
> 
> I (my fault) haven't actively tracked all the patches that are candidate for the release branch but
> many have been committed and this is time to ask if you want a particular patch to be put in the next
> minor release.
> 
> I'd like to release 2.20.1 before February.
> 
> Tristan. 

I apologize for not knowing how to get glitches fixed in gas, but I
would like to see a minor problem fixed. I filed bug report #11122 with
bugzilla about this.

The problem arises when compiling a C source file and using -Wa,-adhl to
get a C source listing with the assembly language. The C source lines
are numbered from 0. They should be numbered from 1. The .loc gas
directives correctly number from 1.

Example:
Using
   gcc -O0 -g -Wa,-adhl nada.c > nada.lst
to get gas listing embedded in C source code, C source code lines are numbered
from 0, .loc in gas starts with 1. Previous gas (4.19) started both at 1.

Example:
   1              		.file	"nada.c"
   9              	.Ltext0:
  10              	.globl foo
  12              	foo:
  13              	.LFB0:
  14              		.file 1 "nada.c"
   0:nada.c        **** int foo()
   1:nada.c        **** {
  15              		.loc 1 2 0
  16              		.cfi_startproc
  17 0000 55       		pushq	%rbp
  18              	.LCFI0:
  19              		.cfi_def_cfa_offset 16
  20 0001 4889E5   		movq	%rsp, %rbp
  21              		.cfi_offset 6, -16
  22              	.LCFI1:
  23              		.cfi_def_cfa_register 6
   2:nada.c        ****    return 0;
  24              		.loc 1 3 0
  25 0004 B8000000 		movl	$0, %eax
  25      00
   3:nada.c        **** }
  26              		.loc 1 4 0
  27 0009 C9       		leave
  28 000a C3       		ret
  29              		.cfi_endproc
  30              	.LFE0:
  32              	.Letext0:

Again, I apologize for not having a good understanding of the
change system, so perhaps doing something inappropriate here.

--Bob

Re: Binutils 2.20.1

[Matthias Klose]

On 05.01.2010 14:12, Tristan Gingold wrote:

could the following two arm patches applied to the branch?

opcodes/

2009-11-17  Edward Nevill <edward.nevill@arm.com>

         * arm-dis.c (print_insn_thumb32): Handle undefined instruction.

gas/
2009-12-21  Ramana Radhakrishnan  <ramana.radhakrishnan@arm.com>
             Richard Earnshaw  <richard.earnshaw@arm.com>

         * config/tc-arm.c (encode_thumb2_b_bl_offset): New. Refactored
         from md_apply_fix.
         (md_apply_fix): Fixup range checks for Thumb2 version
         of unconditional calls. Call encode_thumb2_b_bl_offset for
         unconditional branches / function calls.

Re: Binutils 2.20.1

[Matthias Klose]

On 05.01.2010 14:12, Tristan Gingold wrote:
> I (my fault) haven't actively tracked all the patches that are candidate for the release branch but
> many have been committed and this is time to ask if you want a particular patch to be put in the next
> minor release.

a fix for 10144, if available.

are any gold updates planned for 2.20.1? For a rebuild test with gold as the 
default linker I would like to see a fix for 11072 on the branch (objcopy 
--add-gnu-debuglink is used a lot for package building). Other missing 
features/options are summarized by Peter Fritzsche in [1].

is there a chance that the --enable-gold= patch is approved for the trunk and 
backported to the branch?

   Matthias

[1] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=555886#54

Binutils 2.20.1

[Tristan Gingold]

Hi,

there was a number of glitches in the binutils release 2.20 so it makes sense to make a new minor release.

I (my fault) haven't actively tracked all the patches that are candidate for the release branch but
many have been committed and this is time to ask if you want a particular patch to be put in the next
minor release.

I'd like to release 2.20.1 before February.

Tristan.