headers
David Edelsohn | 1 Jun 2006 01:18
Picon
Favicon

Re: [PATCH] PR 27786

>>>>> Gerald Pfeifer writes:

> In fact, from what is on file at the FSF, we must not accept any
> non-trivial patches because there is an employer disclaimer, but
> no copyright assignment on file.
> 
> Unless I'm mistaken, we should not have taken Asher's previous
> patches into our tree.

	Gerald, you are mistaken on all counts.  This already has been
discussed with Ted Teah at the FSF.  The terminology in the FSF copyrights
file is not clear, but we have complete paperwork for Asher if he is
working for LLNL.

David
Permalink | Reply | 0 comments |
headers
Mark Mitchell | 1 Jun 2006 01:20

Re: [lto] c-decl.c, stor-layout.c: Use VEC for pending_sizes.

Kazu Hirata wrote:
> Hi,
> 
> Attached is a patch to use VEC for pending_sizes.
> 
> One of the things we'd like to do on the LTO branch is to eliminate
> TREE_LIST.  Without this patch, pending_sizes uses TREE_LIST.
> 
> This patch converts pending_sizes to use the VEC API.  Note that the
> order of elements have not been changed.  A bunch of tree_cons
> followed by nreverse is the same as VEC_safe_push followed by
> forward VEC_iterate.
> 
> Tested on x86_64-pc-linux-gnu.  OK to apply to the LTO branch?

OK.  This is also OK for 4.3, once in Stage 1.

--

-- 
Mark Mitchell
CodeSourcery
mark <at> codesourcery.com
(650) 331-3385 x713
Permalink | Reply | 0 comments |
headers
Diego Novillo | 1 Jun 2006 01:30
Picon
Favicon

Re: [lto] tree-cfg.c: Move reinstall_phi_args to tree-ssa.c.

Kazu Hirata wrote on 05/31/06 09:26:

> 2006-05-31  Kazu Hirata  <kazu <at> codesourcery.com>
> 
> 	* tree-cfg.c (reinstall_phi_args): Move to ...
> 	* tree-ssa.c: ... here.
> 	(flush_pending_stmts): Use reinstall_phi_args.
> 	* tree-flow.h: Add a prototype for reinstall_phi_args.
> 
OK.
Permalink | Reply | 0 comments |
headers
David Edelsohn | 1 Jun 2006 01:32
Picon
Favicon

Re: [PATCH] Enable -mcpu=power6 for PowerPC

>>>>> Andrew Pinski writes: 

> Semi unrelated to this but doesn't the Power4 and Power5/5+ both
> have MASK_PPC_GPOPT?
>
> Also I would hope that the Power6 have MASK_PPC_GPOPT (sqrt) also.

	As has been discussed many times, PowerPC hardware sqrt does not
always improve performance.  Library implementations (on platforms other
than Darwin) seem to do quite well, so GPOPT is not always enabled.

	Maybe you should take your own advice and look at the mailinglist
archives before asking :-).

David
Permalink | Reply | 0 comments |
headers
Kazu Hirata | 1 Jun 2006 01:57

[lto] coverage.c: Call build_constructor instead of build_constructor_from_list. (Part 1)

Hi,

Attached is a patch to call build_constructor instead of
build_constructor_from_list.

build_constructor_from_list takes a linked list constructed with
tree_cons and internally converts that to an instance of VEC.

This patches skips the construction of the linked list and calls
build_constructor, which takes VEC directly.

Note that this patch does not change the order of elements.  A bunch
of tree_cons followed by nreverse is the same as a bunch of
VEC_safe_push.

Tested on x86_64-pc-linux-gnu.  OK to apply to the LTO branch (and 4.3
once in stage 1)?

Kazu Hirata

2006-05-31  Kazu Hirata  <kazu <at> codesourcery.com>

	* coverage.c (build_ctr_info_value): Call build_constructor
	instead of build_constructor_from_list.

Index: coverage.c
===================================================================
--- coverage.c	(revision 114242)
+++ coverage.c	(working copy)
 <at>  <at>  -706,15 +706,15  <at>  <at>  build_ctr_info_type (void)
(Continue reading)

Permalink | Reply | 1 comment |
headers
Mark Mitchell | 1 Jun 2006 02:03

Re: [lto] coverage.c: Call build_constructor instead of build_constructor_from_list. (Part 1)

Kazu Hirata wrote:

> Tested on x86_64-pc-linux-gnu.  OK to apply to the LTO branch (and 4.3
> once in stage 1)?

OK.

--

-- 
Mark Mitchell
CodeSourcery
mark <at> codesourcery.com
(650) 331-3385 x713
Permalink | Reply | 0 comments |
headers
Daniel Berlin | 1 Jun 2006 02:07

Re: [4.2 projects] Omega data dependence test

Diego Novillo wrote:
> I guess your idea is to use Omega as one of the data dependency
> analyzers in GRAPHITE? 

The plan was to use it as one of the data dependence testers for, well,
data dependence :)

> Omega is generally a fairly expensive
> analyzer, is your goal replacing the data dependency routines that we
> already have in tree-data-ref.c?

My goal with Omega (which may not be sebastians) was to use it when the
cheaper tests fail.

This is fairly typical usage of data dependence testers, testing with
progressively more expensive ones up to one that can you an exact result.

OTOH, Omega was relatively cheap (compared to other non-simplifying
exact techniques) when it was written (1992), and it's probably even
cheaper now.

IIRC, it was like 89ms per test back in 1992.

>> Index: omega.c
>> > ===================================================================
>> > --- omega.c	(revision 0)
>> > +++ omega.c	(revision 0)
>> >  <at>  <at>  -0,0 +1,5575  <at>  <at> 
>> > +/* Source code for an implementation of the Omega test, an integer
>> > +   programming algorithm for dependence analysis, by William Pugh,
(Continue reading)

Permalink | Reply | 1 comment |
headers
David Edelsohn | 1 Jun 2006 02:14
Picon
Favicon

Re: [4.2 projects] Omega data dependence test

>>>>> Daniel Berlin writes:

Daniel> I asked various people before we started down this road, and was told it
Daniel> was fine.
Daniel> Maybe this is different due to recent events.
Daniel> The author of the code is happy to verify it really is public domain,
Daniel> and has absolutley no license restrictions, if that is what is wanted.

	Public domain is fine.  The FSF website explicitly states "Public
domain status is compatible with the GNU GPL."  The FSF explicitly accepts
a public domain disclaimer in lieu of an assignment and has done so in the
past for GCC.

David
Permalink | Reply | 0 comments |
headers
Gerald Pfeifer | 1 Jun 2006 02:37

Re: PR/25453: document --disable-bootstrap

Hi Paolo,

On Thu, 25 May 2006, Paolo Bonzini wrote:
> Here it is, long overdue.
>
> 2006-05-25  Paolo Bonzini  <bonzini <at> gnu.org>
>
>       * doc/install.texi: Document --disable-bootsrtap and
>       --enable-bootstrap.

I had hoped for one of the build machinery maintainers to approve
this, but since you To:ed me in addition to the list, the new
documentation in the patch reads fine to my eyes, and I assume
you well know what your code is doing ;-), let's go ahead with
this patch.

Ok.

Thanks,
Gerald
Permalink | Reply | 0 comments |
headers
Asher Langton | 1 Jun 2006 03:24

Add myself to Write After Approval list

Committed.

2006-05-31  Asher Langton  <langton2 <at> llnl.gov>

	* MAINTAINERS (Write After Approval): Add myself.
Attachment (commit.diff): application/octet-stream, 853 bytes
Permalink | Reply | 0 comments |

Gmane