Author: rafael
Date: Fri Sep 28 07:53:01 2007
New Revision: 42433
URL: http://llvm.org/viewvc/llvm-project?rev=42433&view=rev
Log:
Refactor the memcpy lowering for the x86 target.
The only generated code difference is that now we call memcpy when
the size of the array is unknown. This matches GCC
Author: djg
Date: Fri Sep 28 10:43:33 2007
New Revision: 42434
URL: http://llvm.org/viewvc/llvm-project?rev=42434&view=rev
Log:
Use 32-bit data directives for DW_FORM_data4 format data, even on
targets with 64-bit addresses.
Modified:
llvm/trunk/lib/CodeGen/DwarfWriter.cpp
Modified: llvm/tru
Hi Rafael,
Are you sure this is better? Did you do any measurement? Our goal
isn't to match gcc output. :)
Perhaps you can add some unit tests to llvm-test?
Thanks,
Evan
On Sep 28, 2007, at 5:53 AM, Rafael Espindola <[EMAIL PROTECTED]
> wrote:
> Author: rafael
> Date: Fri Sep 28 07:53:01
On Sep 28, 2007, at 10:00 AM, Chris Lattner wrote:
> Hi Dale,
>>
>>def int_powi_f32 : Intrinsic<[llvm_float_ty, llvm_float_ty,
>> llvm_i32_ty]>;
>>def int_powi_f64 : Intrinsic<[llvm_double_ty, llvm_double_ty,
>> llvm_i32_ty]>;
>> + def int_powi_f80 : Intrinsic<[llvm_f80_ty, llvm_f80_ty,
Changes in directory llvm-www/releases:
index.html updated: 1.40 -> 1.41
---
Log message:
fix link
---
Diffs of the changes: (+1 -1)
index.html |2 +-
1 files changed, 1 insertion(+), 1 deletion(-)
Index: llvm-www/releases/index.html
diff -u llvm-www/releases/index.html:1.40 llvm-ww
>>> +def X86cmp_new : SDNode<"X86ISD::CMP_NEW" , SDTX86CmpTest>;
>>
>> X86ISD::CMP_NEW SelectionDAG nodes return an i32 flags value,
>> however the
>> corresponding X86cmp_new pattern-matching nodes use SDTX86CmpTest,
>> which has
>> no result values. Is this intentional? If so, how is the
On Sep 28, 2007, at 8:36 AM, Evan Cheng wrote:
> Are you sure this is better? Did you do any measurement? Our goal
> isn't to match gcc output. :)
I think this behavior makes sense. If the size is variable, it could
be arbitrarily large. We should assume that memcpy (the library
implementati
Hi Dale,
On Sep 27, 2007, at 6:08 PM, Dale Johannesen wrote:
> ==
>
> --- llvm/trunk/include/llvm/Intrinsics.td (original)
> +++ llvm/trunk/include/llvm/Intrinsics.td Thu Sep 27 20:08:20 2007
> @@ -76,6 +76,9 @@
> def
Author: djg
Date: Fri Sep 28 11:50:28 2007
New Revision: 42435
URL: http://llvm.org/viewvc/llvm-project?rev=42435&view=rev
Log:
Make the checks for DW_FORM_data4 consistent with the others, and
add more such code for DIEDwarfLabel::SizeOf and DIEObjectLabel::SizeOf.
Modified:
llvm/trunk/lib/C
On Sep 28, 2007, at 10:19 AM, Dale Johannesen wrote:
>
> On Sep 28, 2007, at 10:00 AM, Chris Lattner wrote:
>
>> Hi Dale,
>>>
>>>def int_powi_f32 : Intrinsic<[llvm_float_ty, llvm_float_ty,
>>> llvm_i32_ty]>;
>>>def int_powi_f64 : Intrinsic<[llvm_double_ty, llvm_double_ty,
>>> llvm_i32_ty]
Changes in directory llvm-www/releases:
index.html updated: 1.39 -> 1.40
---
Log message:
add 2.1 doc link and fix link to svn head dox
---
Diffs of the changes: (+2 -1)
index.html |3 ++-
1 files changed, 2 insertions(+), 1 deletion(-)
Index: llvm-www/releases/index.html
diff -u llv
BTW, thank you for splitting up this patch! It makes it much easier
to review.
> //===-- gc-3-collector.patch (+531) ---===//
>
> include/llvm/CodeGen/Collector.h (+134)
> lib/CodeGen/Collector.cpp (+359)
> lib/CodeGen/README.txt (+38)
>
> Collector is the base cla
> I think this behavior makes sense. If the size is variable, it could
> be arbitrarily large. We should assume that memcpy (the library
> implementation) is tuned as best as possible for handling the unknown
> size case.
Yes, we know that libc memcpy is better for big values and that inline
is
>>> The llvm.gcroot intrinsic can be declared to take any types you
>>> care to give it in the module, so long as its eventual type is
>>> like void(**, *).
>>
>> Actually no, I need to update the langref.html document to make
>> this explicit, thanks for reminding me. The issue is that if y
Author: johannes
Date: Fri Sep 28 13:44:17 2007
New Revision: 42440
URL: http://llvm.org/viewvc/llvm-project?rev=42440&view=rev
Log:
Fix long double -> uint64 conversion.
Modified:
llvm/trunk/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp
Modified: llvm/trunk/lib/CodeGen/SelectionDAG/LegalizeDAG.
>> Why not use anyfp? This would allow us to have one sqrt and one powi
>> definition.
>
> Hard to explain; it just feels like the sort of thing where all the
> values ought to be spelled out.
FWIW, sqrt was one of the main reasons we created anyfloat. The
codegen portion of the vector-intrins
Author: johannes
Date: Fri Sep 28 13:46:28 2007
New Revision: 42441
URL: http://llvm.org/viewvc/llvm-project?rev=42441&view=rev
Log:
Record latest work.
Modified:
llvm/trunk/CREDITS.TXT
Modified: llvm/trunk/CREDITS.TXT
URL:
http://llvm.org/viewvc/llvm-project/llvm/trunk/CREDITS.TXT?rev=424
Author: evancheng
Date: Fri Sep 28 14:24:24 2007
New Revision: 42442
URL: http://llvm.org/viewvc/llvm-project?rev=42442&view=rev
Log:
Trim some unneeded fields.
Modified:
llvm/trunk/include/llvm/CodeGen/ScheduleDAG.h
llvm/trunk/lib/CodeGen/SelectionDAG/ScheduleDAG.cpp
llvm/trunk/lib/C
On Jul 20, 2007, at 12:34 PM, Chandler Carruth wrote:
> Author: chandlerc
> Date: Fri Jul 20 14:34:37 2007
> New Revision: 40115
>
> URL: http://llvm.org/viewvc/llvm-project?rev=40115&view=rev
> Log:
> This introduces the atomic operation intrinsics into the
> documentation. This is
> a preview
On Sep 17, 2007, at 3:45 AM, Duncan Sands wrote:
>>> My hope was that some of these trampoline calls could be eliminated
>>> using this optimization. But for...
>>
>> It would be pretty straight-forward to implement an IPO pass that
>> turns trampoline pointers into fat pointers in cases like this
On Sep 28, 2007, at 11:58 AM, Dan Gohman wrote:
>>> Why not use anyfp? This would allow us to have one sqrt and one
>>> powi
>>> definition.
>>
>> Hard to explain; it just feels like the sort of thing where all the
>> values ought to be spelled out.
>
> FWIW, sqrt was one of the main reasons w
Evan, ping?
-Chris
On Aug 1, 2007, at 5:47 PM, Chris Lattner wrote:
> On Jul 31, 2007, at 1:04 AM, Evan Cheng wrote:
>> Author: evancheng
>> Date: Tue Jul 31 03:04:03 2007
>> New Revision: 40628
>>
>> URL: http://llvm.org/viewvc/llvm-project?rev=40628&view=rev
>> Log:
>> Redo and generalize prev
>> FWIW, sqrt was one of the main reasons we created anyfloat. The
>> codegen portion of the vector-intrinsics patch depends on sqrt
>> using anyfloat.
>
> ? It wasn't before my change, either.
The codegen portion of that patch isn't committed yet. But if there's any
reason why anyfloat isn't sui
Author: evancheng
Date: Fri Sep 28 14:37:35 2007
New Revision: 42443
URL: http://llvm.org/viewvc/llvm-project?rev=42443&view=rev
Log:
Remove a poor scheduling heuristic.
Modified:
llvm/trunk/lib/CodeGen/SelectionDAG/ScheduleDAGRRList.cpp
Modified: llvm/trunk/lib/CodeGen/SelectionDAG/Schedule
On Sep 28, 2007, at 12:35 PM, Dan Gohman wrote:
>>> FWIW, sqrt was one of the main reasons we created anyfloat. The
>>> codegen portion of the vector-intrinsics patch depends on sqrt
>>> using anyfloat.
>>
>> ? It wasn't before my change, either.
>
> The codegen portion of that patch isn't commit
Author: lattner
Date: Fri Sep 28 15:09:40 2007
New Revision: 42444
URL: http://llvm.org/viewvc/llvm-project?rev=42444&view=rev
Log:
Make Value::getNameLen not crash on an empty name, fixing PR1712.
Modified:
llvm/trunk/lib/VMCore/Value.cpp
Modified: llvm/trunk/lib/VMCore/Value.cpp
URL:
http
Hi Chris,
> >>> My hope was that some of these trampoline calls could be eliminated
> >>> using this optimization. But for...
> >>
> >> It would be pretty straight-forward to implement an IPO pass that
> >> turns trampoline pointers into fat pointers in cases like this.
> >
> > that's a very inte
> URL: http://llvm.org/viewvc/llvm-project?rev=42367&view=rev
> Log:
> Added C and Ocaml bindings for functions, basic blocks, and
> instruction creation. No support yet for instruction introspection.
>
> Also eliminated allocas from the Ocaml bindings for portability,
> and avoided unnecessary cas
Hi Dale, on my machine (x86-32, linux) LLVM considers long double to
have a size of 10 bytes and an alignment of 4 bytes. GCC gives it a
size of 12 bytes and an alignment of 4 bytes (of course only the first
80 bits (10 bytes) are actually used - the rest is padding).
I can see several problems w
On Sep 24, 2007, at 1:02 PM, Devang Patel wrote:
> Author: dpatel
> Date: Mon Sep 24 15:02:42 2007
> New Revision: 42270
>
> URL: http://llvm.org/viewvc/llvm-project?rev=42270&view=rev
> Log:
> Do not promote null values because it may be unsafe to do so.
Hi Devang,
Unfortunately, this approac
Author: lattner
Date: Fri Sep 28 15:50:50 2007
New Revision: 42448
URL: http://llvm.org/viewvc/llvm-project?rev=42448&view=rev
Log:
really fix PR1581, thanks to Daniel Dunbar for pointing
this out.
Modified:
llvm/trunk/lib/System/DynamicLibrary.cpp
Modified: llvm/trunk/lib/System/DynamicLibr
I tried having the padding by part of the type but concluded this way is
better (although there's a bug, see below.)
The idea is that the size represents how many bytes actually have data,
and the padding in structs and arrays is handled by the alignment (btw,
size and alignment are both 16 on Dar
>> It is very feasible and would be pretty straight-forward. Consider
>> this code in C:
>>
>> static void foo(funcptr_t P) {
>>P();
>> }
>>
>> void a() {
>>tmp = maketrampoline(somefunc1, somedata);
>>foo(tmp);
>> }
>> void b() {
>>tmp = maketrampoline(somefunc2, somedata);
>>
> URL: http://llvm.org/viewvc/llvm-project?rev=42285&view=rev
> Log:
> Added support for new condition code modeling scheme (i.e. physical
> register dependency). These are a bunch of instructions that are
> duplicated so the x86 backend can support both the old and new
> schemes at the same
Hi Dale,
> I tried having the padding by part of the type but concluded this way is
> better (although there's a bug, see below.)
>
> The idea is that the size represents how many bytes actually have data,
> and the padding in structs and arrays is handled by the alignment (btw,
> size and alignm
On Sep 28, 2007, at 2:38 PM, Duncan Sands wrote:
> Hi Chris,
>
>>> Also
>>> 5. Within foo, you need to check that nothing too complicated is
>>> done to P1, for example it isn't stored in a global or passed to
>>> some other function.
>>
>> Actually, worse case, you could just make a trampoline i
On Sep 28, 2007, at 2:04 PM, Chris Lattner <[EMAIL PROTECTED]> wrote:
> BTW, thank you for splitting up this patch! It makes it much easier
> to review.
No problem!
>> //===-- gc-3-collector.patch (+531) ---===//
>>
>> include/llvm/CodeGen/Collector.h (+134)
>> lib/Cod
Author: johannes
Date: Fri Sep 28 13:06:58 2007
New Revision: 42439
URL: http://llvm.org/viewvc/llvm-project?rev=42439&view=rev
Log:
minor long double related changes
Modified:
llvm/trunk/lib/CodeGen/AsmPrinter.cpp
llvm/trunk/lib/CodeGen/IntrinsicLowering.cpp
llvm/trunk/lib/Transform
>> 2. Adding a -gc option to llc isn't what we really want.
>
> Right. I think it should be attached to the module (perhaps via the
> [sub-]target) so that the module is self-contained and magic command-
> line options are not needed.
>
> I toyed with the idea of attaching it to the function to all
Hi Chris,
> > Also
> > 5. Within foo, you need to check that nothing too complicated is
> > done to P1, for example it isn't stored in a global or passed to
> > some other function.
>
> Actually, worse case, you could just make a trampoline in foo.
> However, you're right that this is only wor
Author: evancheng
Date: Fri Sep 28 17:32:30 2007
New Revision: 42453
URL: http://llvm.org/viewvc/llvm-project?rev=42453&view=rev
Log:
If two instructions are both two-address code, favors (schedule closer to
terminator) the one that has a CopyToReg use. This fixes
2006-05-11-InstrSched.ll with -ne
Author: tbrethou
Date: Fri Sep 28 17:50:54 2007
New Revision: 42454
URL: http://llvm.org/viewvc/llvm-project?rev=42454&view=rev
Log:
Update how to release document.
Add release version to getting started guide.
Modified:
llvm/trunk/docs/GettingStarted.html
llvm/trunk/docs/HowToReleaseLLVM
Changes in directory llvm-www:
www-index.html updated: 1.147 -> 1.148
---
Log message:
2.2 schedule
---
Diffs of the changes: (+9 -1)
www-index.html | 10 +-
1 files changed, 9 insertions(+), 1 deletion(-)
Index: llvm-www/www-index.html
diff -u llvm-www/www-index.html:1.147 llv
Changes in directory llvm-www:
www-index.html updated: 1.148 -> 1.149
---
Log message:
Fix typo.
---
Diffs of the changes: (+2 -2)
www-index.html |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
Index: llvm-www/www-index.html
diff -u llvm-www/www-index.html:1.148 llvm-www/ww
Author: djg
Date: Fri Sep 28 18:53:40 2007
New Revision: 42457
URL: http://llvm.org/viewvc/llvm-project?rev=42457&view=rev
Log:
Teach SplitVectorOp how to split INSERT_VECTOR_ELT.
Added:
llvm/trunk/test/CodeGen/X86/illegal-insert.ll
Modified:
llvm/trunk/lib/CodeGen/SelectionDAG/LegalizeDA
Author: dpatel
Date: Fri Sep 28 18:54:26 2007
New Revision: 42458
URL: http://llvm.org/viewvc/llvm-project?rev=42458&view=rev
Log:
Do not ignore packed member size while selecting union type.
Modified:
llvm-gcc-4.0/trunk/gcc/llvm-types.cpp
Modified: llvm-gcc-4.0/trunk/gcc/llvm-types.cpp
URL:
Author: dpatel
Date: Fri Sep 28 18:54:26 2007
New Revision: 42458
URL: http://llvm.org/viewvc/llvm-project?rev=42458&view=rev
Log:
Do not ignore packed member size while selecting union type.
Added:
llvm/trunk/test/CFrontend/2007-09-28-PackedUnionMember.c
Added: llvm/trunk/test/CFrontend/200
Author: gordon
Date: Fri Sep 28 19:42:21 2007
New Revision: 42460
URL: http://llvm.org/viewvc/llvm-project?rev=42460&view=rev
Log:
PR1601: etags not configured correctly
Resolving this by deleting vestigal etags remnants.
Modified:
llvm/trunk/Makefile.config.in
llvm/trunk/autoconf/configu
Author: gordon
Date: Fri Sep 28 19:42:56 2007
New Revision: 42461
URL: http://llvm.org/viewvc/llvm-project?rev=42461&view=rev
Log:
Regenerate.
Modified:
llvm/trunk/configure
Modified: llvm/trunk/configure
URL:
http://llvm.org/viewvc/llvm-project/llvm/trunk/configure?rev=42461&r1=42460&r2=42
On 9/28/07, Devang Patel <[EMAIL PROTECTED]> wrote:
> URL: http://llvm.org/viewvc/llvm-project?rev=42458&view=rev
> Log:
> Do not ignore packed member size while selecting union type.
>
Thanks! I think this should be applied to 4.2 as well.
-bw
___
llvm-
Author: dannyb
Date: Fri Sep 28 19:50:40 2007
New Revision: 42462
URL: http://llvm.org/viewvc/llvm-project?rev=42462&view=rev
Log:
Switch to densemap rather than std::set
Modified:
llvm/trunk/lib/Analysis/IPA/Andersens.cpp
Modified: llvm/trunk/lib/Analysis/IPA/Andersens.cpp
URL:
http://llvm
On Sep 28, 2007, at 16:18, Chris Lattner wrote:
URL: http://llvm.org/viewvc/llvm-project?rev=42367&view=rev
Log:
Added C and Ocaml bindings for functions, basic blocks, and
instruction creation. No support yet for instruction introspection.
Also eliminated allocas from the Ocaml bindings for po
Author: dpatel
Date: Fri Sep 28 20:20:11 2007
New Revision: 42464
URL: http://llvm.org/viewvc/llvm-project?rev=42464&view=rev
Log:
Do not ignore packed member size while selecting union type.
Modified:
llvm-gcc-4.2/trunk/gcc/llvm-types.cpp
Modified: llvm-gcc-4.2/trunk/gcc/llvm-types.cpp
URL
Author: gordon
Date: Fri Sep 28 20:38:42 2007
New Revision: 42465
URL: http://llvm.org/viewvc/llvm-project?rev=42465&view=rev
Log:
Demoting CHelpers.h to include/llvm/Support.
Added:
llvm/trunk/include/llvm/Support/CHelpers.h
- copied unchanged from r42459, llvm/trunk/include/llvm/CHelp
On 2007-09-28, at 21:02, Gordon Henriksen wrote:
On Sep 28, 2007, at 16:18, Chris Lattner wrote:
URL: http://llvm.org/viewvc/llvm-project?rev=42367&view=rev
Log:
Added C and Ocaml bindings for functions, basic blocks, and
instruction creation. No support yet for instruction introspection.
Als
Author: gordon
Date: Fri Sep 28 21:13:43 2007
New Revision: 42466
URL: http://llvm.org/viewvc/llvm-project?rev=42466&view=rev
Log:
Collector is the base class for garbage collection code generators.
This version enhances the previous patch to add root initialization
as discussed here:
http://list
Author: gordon
Date: Fri Sep 28 21:23:08 2007
New Revision: 42467
URL: http://llvm.org/viewvc/llvm-project?rev=42467&view=rev
Log:
This is done already.
Modified:
llvm/trunk/lib/CodeGen/README.txt
Modified: llvm/trunk/lib/CodeGen/README.txt
URL:
http://llvm.org/viewvc/llvm-project/llvm/trun
Author: evancheng
Date: Fri Sep 28 22:06:21 2007
New Revision: 42468
URL: http://llvm.org/viewvc/llvm-project?rev=42468&view=rev
Log:
x86 llcbeta back to -regalloc=local -fast
Modified:
test-suite/trunk/Makefile.programs
Modified: test-suite/trunk/Makefile.programs
URL:
http://llvm.org/view
58 matches
Mail list logo
