On Jan 4, 2008, at 11:21 PM, Bill Wendling wrote:
> On Jan 4, 2008, at 11:02 PM, Chris Lattner wrote:
>> You aren't looking for loads *from the global*, you're looking for
>> loads from the *stub for the global*, which are always invariant.
>> You
>> just need to know whether the load is from
Hrm. ReMat of load from argument area is being handled as a special
case in LiveIntervals::isReMaterializable(). Having
isReallyTrivallyReMaterializable() returns true for it shouldn't make
any difference. I'll look at this.
Evan
On Jan 4, 2008, at 10:10 PM, Chris Lattner wrote:
> Author:
On Jan 5, 2008, at 12:28 AM, Chris Lattner wrote:
>
> On Jan 4, 2008, at 11:21 PM, Bill Wendling wrote:
>
>> On Jan 4, 2008, at 11:02 PM, Chris Lattner wrote:
>>> You aren't looking for loads *from the global*, you're looking for
>>> loads from the *stub for the global*, which are always invarian
Cool!
Evan
On Jan 4, 2008, at 5:04 PM, Chris Lattner wrote:
> On Jan 4, 2008, at 4:41 PM, Evan Cheng wrote:
>> Author: evancheng
>> Date: Fri Jan 4 18:41:47 2008
>> New Revision: 45605
>>
>> URL: http://llvm.org/viewvc/llvm-project?rev=45605&view=rev
>> Log:
>> Combine MovePCtoStack + POP32r in
On Jan 5, 2008, at 12:53 AM, Evan Cheng wrote:
>
> On Jan 5, 2008, at 12:28 AM, Chris Lattner wrote:
>
>>
>> On Jan 4, 2008, at 11:21 PM, Bill Wendling wrote:
>>
>>> On Jan 4, 2008, at 11:02 PM, Chris Lattner wrote:
You aren't looking for loads *from the global*, you're looking for
load
Author: void
Date: Sat Jan 5 03:18:04 2008
New Revision: 45626
URL: http://llvm.org/viewvc/llvm-project?rev=45626&view=rev
Log:
Chris and Evan noticed that this check was compleatly fubared. I was
checking that there was a from a global instead of a load from the stub
for a global, which is the o
On Jan 5, 2008, at 12:53 AM, Evan Cheng wrote:
> Right. If it's a load from GV and GVRequiresExtraLoad() returns true
> then it's a load from a stub.
>
Okay. Once I did the check, it now seems to treat the example program:
volatile char G[100];
int B(char *F, int N) {
for (; N > 0; --N)
F[
On Jan 5, 2008, at 1:18 AM, Bill Wendling <[EMAIL PROTECTED]> wrote:
> Author: void
> Date: Sat Jan 5 03:18:04 2008
> New Revision: 45626
>
> URL: http://llvm.org/viewvc/llvm-project?rev=45626&view=rev
> Log:
> Chris and Evan noticed that this check was compleatly fubared. I was
> checking that
> Nice evan! With this, sinking is now able to compile code like this:
Nifty! I would like to test Mozilla & Qt with this change: they heavily
use shared libraries thus almost everything is PIC.
--
WBR, Anton Korobeynikov
___
llvm-commits mailing list
Author: baldrick
Date: Sat Jan 5 11:48:09 2008
New Revision: 45628
URL: http://llvm.org/viewvc/llvm-project?rev=45628&view=rev
Log:
The Ada front-end only sets flag_exceptions after
processing the file (but before gimplification).
At system init time flag_exceptions is thus never
set, which means
>> I am not 100% sure, but I think that
>> X86Subtarget::GVRequiresExtraLoad
>> is the place to start looking.
>
> Right. If it's a load from GV and GVRequiresExtraLoad() returns true
> then it's a load from a stub.
>
> Here is something to consider for further enhancement. There are
> potentiall
On Jan 5, 2008, at 12:36 AM, Evan Cheng wrote:
> Hrm. ReMat of load from argument area is being handled as a special
> case in LiveIntervals::isReMaterializable(). Having
> isReallyTrivallyReMaterializable() returns true for it shouldn't make
> any difference. I'll look at this.
Thanks for inves
On Jan 5, 2008, at 8:58 AM, Anton Korobeynikov wrote:
>> Nice evan! With this, sinking is now able to compile code like this:
> Nifty! I would like to test Mozilla & Qt with this change: they
> heavily
> use shared libraries thus almost everything is PIC.
Nice, I hope to get this enabled by d
Author: lattner
Date: Sat Jan 5 14:15:42 2008
New Revision: 45629
URL: http://llvm.org/viewvc/llvm-project?rev=45629&view=rev
Log:
Fix build issue on certain compilers.
Modified:
llvm/trunk/lib/CodeGen/MachineDominators.cpp
Modified: llvm/trunk/lib/CodeGen/MachineDominators.cpp
URL:
http:/
Author: sampo
Date: Sat Jan 5 14:47:37 2008
New Revision: 45630
URL: http://llvm.org/viewvc/llvm-project?rev=45630&view=rev
Log:
If custom lowering of insert element fails, the result Val will be 0.
Don't overwrite a variable used by the fallthrough code path in this
case.
Modified:
llvm/tr
Author: sampo
Date: Sat Jan 5 14:51:30 2008
New Revision: 45631
URL: http://llvm.org/viewvc/llvm-project?rev=45631&view=rev
Log:
Remove an incorrect optimization that is performed correctly by
the target independent legalizer.
Modified:
llvm/trunk/lib/Target/X86/X86ISelLowering.cpp
Modified
Author: lattner
Date: Sat Jan 5 16:54:53 2008
New Revision: 45635
URL: http://llvm.org/viewvc/llvm-project?rev=45635&view=rev
Log:
move Node Transformation printing from CodeGenDAGPatterns -> DAGISelEmitter.
The only difference in output is that we now print them in alphabetical
order instead of
Author: lattner
Date: Sat Jan 5 16:43:57 2008
New Revision: 45634
URL: http://llvm.org/viewvc/llvm-project?rev=45634&view=rev
Log:
move predicate printing code from CodeGenDAGPatterns -> DAGISelEmitter.
Modified:
llvm/trunk/utils/TableGen/CodeGenDAGPatterns.cpp
llvm/trunk/utils/TableGen/
Author: lattner
Date: Sat Jan 5 16:30:17 2008
New Revision: 45633
URL: http://llvm.org/viewvc/llvm-project?rev=45633&view=rev
Log:
fix a fixme by improving const correctness.
Modified:
llvm/trunk/utils/TableGen/CodeGenDAGPatterns.h
llvm/trunk/utils/TableGen/DAGISelEmitter.cpp
llvm/tr
Author: lattner
Date: Sat Jan 5 16:58:54 2008
New Revision: 45636
URL: http://llvm.org/viewvc/llvm-project?rev=45636&view=rev
Log:
now that computing CodegenDAGPatterns doesn't implicitly print stuff
out, DAGISelEmitter can compute it in its ctor, which simplifies some code.
Now we can use Code
Author: lattner
Date: Sat Jan 5 17:37:52 2008
New Revision: 45639
URL: http://llvm.org/viewvc/llvm-project?rev=45639&view=rev
Log:
fix build on case sensitive file systems.
Modified:
llvm/trunk/utils/TableGen/CodeGenDAGPatterns.cpp
Modified: llvm/trunk/utils/TableGen/CodeGenDAGPatterns.cpp
Author: void
Date: Sat Jan 5 17:30:51 2008
New Revision: 45638
URL: http://llvm.org/viewvc/llvm-project?rev=45638&view=rev
Log:
Fix comment.
Modified:
llvm/trunk/lib/Target/X86/X86InstrInfo.cpp
Modified: llvm/trunk/lib/Target/X86/X86InstrInfo.cpp
URL:
http://llvm.org/viewvc/llvm-project/ll
Author: lattner
Date: Sat Jan 5 17:29:51 2008
New Revision: 45637
URL: http://llvm.org/viewvc/llvm-project?rev=45637&view=rev
Log:
make this build with newer gcc's
Modified:
llvm/trunk/lib/CodeGen/MachineLoopInfo.cpp
Modified: llvm/trunk/lib/CodeGen/MachineLoopInfo.cpp
URL:
http://llvm.org
On Jan 5, 2008, at 2:49 AM, Evan Cheng wrote:
> On Jan 5, 2008, at 1:18 AM, Bill Wendling <[EMAIL PROTECTED]> wrote:
>
>> --- llvm/trunk/lib/Target/X86/X86InstrInfo.cpp (original)
>> +++ llvm/trunk/lib/Target/X86/X86InstrInfo.cpp Sat Jan 5 03:18:04
>> 2008
>> @@ -171,12 +171,15 @@
>> case X86::M
Author: lattner
Date: Sat Jan 5 18:49:05 2008
New Revision: 45640
URL: http://llvm.org/viewvc/llvm-project?rev=45640&view=rev
Log:
split enum emission out from InstrInfoEmitter into it's own tblgen backend.
Added:
llvm/trunk/utils/TableGen/InstrEnumEmitter.cpp
llvm/trunk/utils/TableGen/I
Author: lattner
Date: Sat Jan 5 19:20:13 2008
New Revision: 45643
URL: http://llvm.org/viewvc/llvm-project?rev=45643&view=rev
Log:
further simplifications and cleanup
Modified:
llvm/trunk/utils/TableGen/InstrInfoEmitter.cpp
llvm/trunk/utils/TableGen/InstrInfoEmitter.h
Modified: llvm/tru
Author: lattner
Date: Sat Jan 5 19:21:51 2008
New Revision: 45644
URL: http://llvm.org/viewvc/llvm-project?rev=45644&view=rev
Log:
final cleanups.
Modified:
llvm/trunk/utils/TableGen/InstrInfoEmitter.cpp
llvm/trunk/utils/TableGen/InstrInfoEmitter.h
Modified: llvm/trunk/utils/TableGen/In
Author: lattner
Date: Sat Jan 5 19:12:44 2008
New Revision: 45642
URL: http://llvm.org/viewvc/llvm-project?rev=45642&view=rev
Log:
simplify some code
Modified:
llvm/trunk/utils/TableGen/InstrInfoEmitter.cpp
llvm/trunk/utils/TableGen/InstrInfoEmitter.h
Modified: llvm/trunk/utils/TableGe
Author: lattner
Date: Sat Jan 5 19:10:31 2008
New Revision: 45641
URL: http://llvm.org/viewvc/llvm-project?rev=45641&view=rev
Log:
rename CodegenDAGPatterns -> CodeGenDAGPatterns
Modified:
llvm/trunk/utils/TableGen/CodeGenDAGPatterns.cpp
llvm/trunk/utils/TableGen/CodeGenDAGPatterns.h
Author: lattner
Date: Sat Jan 5 19:53:37 2008
New Revision: 45647
URL: http://llvm.org/viewvc/llvm-project?rev=45647&view=rev
Log:
rearrange some code to allow inferring instr info from the pattern of the
instr, but don't do so yet.
Modified:
llvm/trunk/utils/TableGen/InstrInfoEmitter.cpp
Author: lattner
Date: Sat Jan 5 19:52:22 2008
New Revision: 45646
URL: http://llvm.org/viewvc/llvm-project?rev=45646&view=rev
Log:
improve const correctness.
Modified:
llvm/trunk/utils/TableGen/CodeGenDAGPatterns.cpp
llvm/trunk/utils/TableGen/CodeGenDAGPatterns.h
llvm/trunk/utils/Tab
Author: lattner
Date: Sat Jan 5 20:16:26 2008
New Revision: 45648
URL: http://llvm.org/viewvc/llvm-project?rev=45648&view=rev
Log:
remove some old hacky code that tried to infer whether a store
occured in a pattern, but failed miserably. The new code works for
any instruction that has a store i
Author: lattner
Date: Sat Jan 5 19:35:39 2008
New Revision: 45645
URL: http://llvm.org/viewvc/llvm-project?rev=45645&view=rev
Log:
Split the impl of CodeGenInstruction out to its own .cpp file, add a getName()
accessor.
Added:
llvm/trunk/utils/TableGen/CodeGenInstruction.cpp
Modified:
l
Author: nicholas
Date: Sat Jan 5 21:12:44 2008
New Revision: 45649
URL: http://llvm.org/viewvc/llvm-project?rev=45649&view=rev
Log:
Accept both %y, %x and %x, %y as valid answers.
Modified:
llvm/trunk/test/Analysis/BasicAA/2007-08-01-NoAliasAndCalls.ll
Modified: llvm/trunk/test/Analysis/Bas
Author: lattner
Date: Sat Jan 5 23:36:50 2008
New Revision: 45650
URL: http://llvm.org/viewvc/llvm-project?rev=45650&view=rev
Log:
set the 'isstore' flag for instructions whose pattern is an
intrinsic that writes to memory.
Modified:
llvm/trunk/utils/TableGen/CodeGenDAGPatterns.cpp
llvm
Author: lattner
Date: Sat Jan 5 23:43:21 2008
New Revision: 45651
URL: http://llvm.org/viewvc/llvm-project?rev=45651&view=rev
Log:
describe isStore and simplify the implementation of hasUnmodelledSideEffects.
No functionality change.
Modified:
llvm/trunk/include/llvm/Target/TargetInstrInfo.
Author: lattner
Date: Sat Jan 5 23:53:26 2008
New Revision: 45652
URL: http://llvm.org/viewvc/llvm-project?rev=45652&view=rev
Log:
remove some isStore flags that are now inferred automatically.
Modified:
llvm/trunk/lib/Target/PowerPC/PPCInstr64Bit.td
llvm/trunk/lib/Target/PowerPC/PPCInst
Author: lattner
Date: Sat Jan 5 23:55:01 2008
New Revision: 45653
URL: http://llvm.org/viewvc/llvm-project?rev=45653&view=rev
Log:
remove explicit isStore flags that are now inferrable.
Modified:
llvm/trunk/lib/Target/ARM/ARMInstrInfo.td
llvm/trunk/lib/Target/ARM/ARMInstrThumb.td
llv
Author: lattner
Date: Sun Jan 6 00:44:58 2008
New Revision: 45654
URL: http://llvm.org/viewvc/llvm-project?rev=45654&view=rev
Log:
Change the 'isStore' inferrer to look for 'SDNPMayStore'
instead of "ISD::STORE". This allows us to mark target-specific dag
nodes as storing (such as ppc byteswap
Author: baldrick
Date: Sun Jan 6 01:43:13 2008
New Revision: 45655
URL: http://llvm.org/viewvc/llvm-project?rev=45655&view=rev
Log:
Remove an unused variable.
Modified:
llvm/trunk/lib/CodeGen/PrologEpilogInserter.cpp
Modified: llvm/trunk/lib/CodeGen/PrologEpilogInserter.cpp
URL:
http://llv
40 matches
Mail list logo
