https://github.com/orcguru closed
https://github.com/llvm/llvm-project/pull/84132
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
orcguru wrote:
> It looks like a few of the regression tests are failing on the
> reverse-iteration buildbot
> (https://lab.llvm.org/buildbot/#/builders/54/builds/9683)
Thank you! Posted: https://github.com/llvm/llvm-project/pull/89714
https://github.com/llvm/llvm-project/pull/66316
https://github.com/orcguru edited
https://github.com/llvm/llvm-project/pull/84132
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/orcguru closed
https://github.com/llvm/llvm-project/pull/88829
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/orcguru updated
https://github.com/llvm/llvm-project/pull/88829
>From 35bf12e2fcb499c8953e4207cea2296d9ddcbf1f Mon Sep 17 00:00:00 2001
From: Ting Wang
Date: Mon, 15 Apr 2024 21:52:02 -0400
Subject: [PATCH 1/2] Following Amy's b1922e55ab3b35dff99238fd0b74be00df0472e7,
and
@@ -3362,6 +3362,64 @@ SDValue PPCTargetLowering::LowerGlobalTLSAddress(SDValue
Op,
return LowerGlobalTLSAddressLinux(Op, DAG);
}
+/// updateForAIXShLibTLSModelOpt - Helper to initialize TLS model opt settings,
+/// and then apply the update.
+static void
@@ -3362,6 +3362,64 @@ SDValue PPCTargetLowering::LowerGlobalTLSAddress(SDValue
Op,
return LowerGlobalTLSAddressLinux(Op, DAG);
}
+/// updateForAIXShLibTLSModelOpt - Helper to initialize TLS model opt settings,
+/// and then apply the update.
+static void
@@ -5019,6 +5019,10 @@ def maix_small_local_exec_tls : Flag<["-"],
"maix-small-local-exec-tls">,
"where the offset from the TLS base is encoded as an "
"immediate operand (AIX 64-bit only). "
"This access sequence is not used for variables
https://github.com/orcguru created
https://github.com/llvm/llvm-project/pull/88829
This patch adds the clang portion of an AIX-specific option to inform the
compiler that it can use a faster access sequence for the local-dynamic
TLS model (formally named aix-small-local-dynamic-tls).
This
orcguru wrote:
Since https://github.com/llvm/llvm-project/pull/86641 has been merged, I will
continue work on this patch to handle the interactions of these two flags.
Mean while I'm waiting for more review comments from experts. Thank you!
https://github.com/llvm/llvm-project/pull/84132
https://github.com/orcguru deleted
https://github.com/llvm/llvm-project/pull/84132
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -3369,6 +3369,59 @@ SDValue
PPCTargetLowering::LowerGlobalTLSAddressAIX(SDValue Op,
bool Is64Bit = Subtarget.isPPC64();
bool HasAIXSmallLocalExecTLS = Subtarget.hasAIXSmallLocalExecTLS();
TLSModel::Model Model = getTargetMachine().getTLSModel(GV);
+ // Initialize
@@ -3369,6 +3369,59 @@ SDValue
PPCTargetLowering::LowerGlobalTLSAddressAIX(SDValue Op,
bool Is64Bit = Subtarget.isPPC64();
bool HasAIXSmallLocalExecTLS = Subtarget.hasAIXSmallLocalExecTLS();
TLSModel::Model Model = getTargetMachine().getTLSModel(GV);
+ // Initialize
@@ -329,6 +329,12 @@ def FeatureAIXLocalExecTLS :
"Produce a TOC-free local-exec TLS sequence for this
function "
"for 64-bit AIX">;
+def FeatureAIXSharedLibraryTLSModelHeuristic :
+
@@ -3369,6 +3369,48 @@ SDValue
PPCTargetLowering::LowerGlobalTLSAddressAIX(SDValue Op,
bool Is64Bit = Subtarget.isPPC64();
bool HasAIXSmallLocalExecTLS = Subtarget.hasAIXSmallLocalExecTLS();
TLSModel::Model Model = getTargetMachine().getTLSModel(GV);
+ // Initialize
@@ -80,6 +80,7 @@ class LLVM_LIBRARY_VISIBILITY PPCTargetInfo : public
TargetInfo {
bool IsISA3_0 = false;
bool IsISA3_1 = false;
bool HasQuadwordAtomics = false;
+ bool HasAIXShLibTLSModelHeuristic = false;
orcguru wrote:
Hi Kai, Thank you for
https://github.com/orcguru edited
https://github.com/llvm/llvm-project/pull/84132
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/orcguru edited
https://github.com/llvm/llvm-project/pull/84132
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/orcguru updated
https://github.com/llvm/llvm-project/pull/84132
>From 92ab84000b638e206106bf5b1e9e2e842f5bcf2b Mon Sep 17 00:00:00 2001
From: Ting Wang
Date: Wed, 6 Mar 2024 02:51:40 -0500
Subject: [PATCH 1/5] [PowerPC] Rename symbols references by tls-local-dynamic
model
https://github.com/orcguru updated
https://github.com/llvm/llvm-project/pull/84132
>From 92ab84000b638e206106bf5b1e9e2e842f5bcf2b Mon Sep 17 00:00:00 2001
From: Ting Wang
Date: Wed, 6 Mar 2024 02:51:40 -0500
Subject: [PATCH 1/4] [PowerPC] Rename symbols references by tls-local-dynamic
model
https://github.com/orcguru closed
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
orcguru wrote:
Thank you all for your efforts reviewing the patch!
If no objections, I will land it before the weekend. Thank you!
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
orcguru wrote:
> A question: For `PPCTLSDynamicCall.cpp`, would it be better if we instead
> reworded some of the sentences and not use `Load*@toc`?
>
> I thought about this for a bit and I think the full sentences may be a bit
> better, so I added some sample comment suggestions, but if you
@@ -370,11 +370,22 @@ namespace llvm {
/// G8RC = TLSGD_AIX, TOC_ENTRY, TOC_ENTRY
/// Op that combines two register copies of TOC entries
/// (region handle into R3 and variable offset into R4) followed by a
-/// GET_TLS_ADDR node which will be expanded to a
https://github.com/orcguru closed
https://github.com/llvm/llvm-project/pull/79968
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
orcguru wrote:
> Since this patch relies on the local-dynamic patch, I think the local-dynamic
> changes are already in here, right? Do we know if there is a way yet to make
> dependent pull requests, just so reviewing this patch may be a bit easier
> when the local-dynamic changes are
@@ -145,9 +164,94 @@ namespace {
.addImm(0);
if (IsAIX) {
- // The variable offset and region handle are copied in r4 and r3. The
- // copies are followed by
orcguru wrote:
> I believe some of the test cases also require updates as a result of #80162.
Sure. Updated!
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
@@ -145,9 +164,94 @@ namespace {
.addImm(0);
if (IsAIX) {
- // The variable offset and region handle are copied in r4 and r3. The
- // copies are followed by
@@ -145,9 +164,94 @@ namespace {
.addImm(0);
if (IsAIX) {
- // The variable offset and region handle are copied in r4 and r3. The
- // copies are followed by
@@ -145,9 +164,89 @@ namespace {
.addImm(0);
if (IsAIX) {
- // The variable offset and region handle are copied in r4 and r3. The
- // copies are followed by
@@ -145,9 +164,89 @@ namespace {
.addImm(0);
if (IsAIX) {
- // The variable offset and region handle are copied in r4 and r3. The
- // copies are followed by
@@ -145,9 +164,89 @@ namespace {
.addImm(0);
if (IsAIX) {
- // The variable offset and region handle are copied in r4 and r3. The
- // copies are followed by
https://github.com/orcguru deleted
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -2645,10 +2653,14 @@ MCSection
*TargetLoweringObjectFileXCOFF::getSectionForTOCEntry(
const MCSymbol *Sym, const TargetMachine ) const {
// Use TE storage-mapping class when large code model is enabled so that
// the chance of needing -bbigtoc is decreased.
+ //
@@ -2645,10 +2653,14 @@ MCSection
*TargetLoweringObjectFileXCOFF::getSectionForTOCEntry(
const MCSymbol *Sym, const TargetMachine ) const {
// Use TE storage-mapping class when large code model is enabled so that
// the chance of needing -bbigtoc is decreased.
+ //
@@ -2058,6 +2089,10 @@ void PPCAIXAsmPrinter::emitLinkage(const GlobalValue *GV,
}
}
+ // Do not emit the _$TLSML symbol.
+ if (GVSym->getName() == "_Renamed..5f24__TLSML[TC]")
orcguru wrote:
Thank you! I will put some flag check before the string
https://github.com/orcguru deleted
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -2058,6 +2089,10 @@ void PPCAIXAsmPrinter::emitLinkage(const GlobalValue *GV,
}
}
+ // Do not emit the _$TLSML symbol.
+ if (GVSym->getName() == "_Renamed..5f24__TLSML[TC]")
orcguru wrote:
I think `MCContext::createXCOFFSymbolImpl` is the logic
@@ -2058,6 +2089,10 @@ void PPCAIXAsmPrinter::emitLinkage(const GlobalValue *GV,
}
}
+ // Do not emit the _$TLSML symbol.
+ if (GVSym->getName() == "_Renamed..5f24__TLSML[TC]")
orcguru wrote:
I think `MCContext::createXCOFFSymbolImpl` is the logic
@@ -846,6 +859,19 @@ void PPCAsmPrinter::emitInstruction(const MachineInstr
*MI) {
return MCSymbolRefExpr::VariantKind::VK_PPC_AIX_TLSGDM;
if (MO.getTargetFlags() & PPCII::MO_TLSGD_FLAG)
return MCSymbolRefExpr::VariantKind::VK_PPC_AIX_TLSGD;
+if
@@ -649,6 +649,14 @@ void XCOFFObjectWriter::recordRelocation(MCAssembler ,
uint64_t ) {
auto getIndex = [this](const MCSymbol *Sym,
const MCSectionXCOFF *ContainingCsect) {
+// Fixup relocation flag for
@@ -649,6 +649,14 @@ void XCOFFObjectWriter::recordRelocation(MCAssembler ,
uint64_t ) {
auto getIndex = [this](const MCSymbol *Sym,
const MCSectionXCOFF *ContainingCsect) {
+// Fixup relocation flag for
@@ -649,6 +649,14 @@ void XCOFFObjectWriter::recordRelocation(MCAssembler ,
uint64_t ) {
auto getIndex = [this](const MCSymbol *Sym,
const MCSectionXCOFF *ContainingCsect) {
+// Fixup relocation flag for
@@ -660,14 +671,16 @@ void PPCAsmPrinter::EmitTlsCall(const MachineInstr *MI,
"GETtls[ld]ADDR[32] must read GPR3");
if (Subtarget->isAIXABI()) {
-// On AIX, the variable offset should already be in R4 and the region
handle
-// should already be in R3.
-
@@ -6,6 +6,6 @@
#endif
static __thread int y __attribute((tls_model("global-dynamic"))); // no-warning
-static __thread int y __attribute((tls_model("local-dynamic"))); //
expected-error {{TLS model 'local-dynamic' is not yet supported on AIX}}
+static __thread int y
https://github.com/orcguru resolved
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -145,9 +160,92 @@ namespace {
.addImm(0);
if (IsAIX) {
- // The variable offset and region handle are copied in r4 and r3. The
- // copies are followed by
https://github.com/orcguru resolved
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -145,9 +160,92 @@ namespace {
.addImm(0);
if (IsAIX) {
- // The variable offset and region handle are copied in r4 and r3. The
- // copies are followed by
https://github.com/orcguru resolved
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -145,9 +160,92 @@ namespace {
.addImm(0);
if (IsAIX) {
- // The variable offset and region handle are copied in r4 and r3. The
- // copies are followed by
https://github.com/orcguru resolved
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/orcguru resolved
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/orcguru resolved
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/orcguru resolved
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/orcguru resolved
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/orcguru resolved
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/orcguru resolved
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
orcguru wrote:
> Also, is it intentional that we combined the patch to enable local-dynamic in
> clang within this patch? Just wondering since I thought I saw a separate
> patch for the clang portion before (unless I am mistaken).
Yes, I have to combine all changes into a monolithic patch,
@@ -6,6 +6,6 @@
#endif
static __thread int y __attribute((tls_model("global-dynamic"))); // no-warning
-static __thread int y __attribute((tls_model("local-dynamic"))); //
expected-error {{TLS model 'local-dynamic' is not yet supported on AIX}}
+static __thread int y
orcguru wrote:
> The code formatting check fails.
That is intentional:
https://github.com/llvm/llvm-project/pull/66316#discussion_r1333195506
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
orcguru wrote:
I saw comments being folded, paste here for notice, sorry about the duplication:
So I did have another version which can essentially remove the assert check
with some side effects:
(1) It did generated extra external symbol reference, which looks like no harm
with simple
@@ -3412,13 +3416,23 @@ SDValue
PPCTargetLowering::LowerGlobalTLSAddressAIX(SDValue Op,
return DAG.getNode(PPCISD::ADD_TLS, dl, PtrVT, TLSReg, VariableOffset);
}
- // Only Local-Exec, Initial-Exec and General-Dynamic TLS models are currently
- // supported models. If
https://github.com/orcguru resolved
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/orcguru resolved
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/orcguru resolved
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/orcguru resolved
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/orcguru resolved
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -116,6 +116,10 @@ std::pair
PPCXCOFFObjectWriter::getRelocTypeAndSignSize(
return {XCOFF::RelocationType::R_TLS_IE, SignAndSizeForFKData};
case MCSymbolRefExpr::VK_PPC_AIX_TLSLE:
return {XCOFF::RelocationType::R_TLS_LE, SignAndSizeForFKData};
+case
orcguru wrote:
> Missing an update in `recordRelocation()` similar as the patch:
> https://reviews.llvm.org/D155415.
>
> Working on it now.
Done. Added one test case aix-tls-ld-xcoff-reloc-large.ll to track this kind of
issue. If full set of test case is necessary, please let me know, and I
orcguru wrote:
Missing an update in `recordRelocation()` similar as the patch:
https://reviews.llvm.org/D155415.
Working on it now.
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
orcguru wrote:
When a module contains three TLS LD variables
>From assembly file generated by "as -a64 -many -o"
```
02f8 (idx: 33) a[TC]:
2f8: 00 00 00 00
02f8: R_TLS_LD (idx: 41) a[TL]
2fc: 00 00 00 00
0300 (idx: 35)
orcguru wrote:
Currently obj mode has one issue: define more than one __thread variable, and
all access return the first variable.
I'm investigating the root cause and will update patch accordingly. ASM mode is
fine.
https://github.com/llvm/llvm-project/pull/66316
orcguru wrote:
> For below case:
>
> ```
> __thread extern int x = 42;
> __thread extern int y = 42;
> int main(void) {
>return x + y;
> }
> ```
>
> For now, we get:
>
> ```
> .main:
> # %bb.0:# %entry
> mflr 0
> stwu 1, -32(1)
>
orcguru wrote:
I'm working on environment issue, and will update remaining two LIT cases and
do testing etc.
I'm requesting another round of review in the mean time. Thank you all for your
inputs!
https://github.com/llvm/llvm-project/pull/66316
___
@@ -3412,13 +3416,23 @@ SDValue
PPCTargetLowering::LowerGlobalTLSAddressAIX(SDValue Op,
return DAG.getNode(PPCISD::ADD_TLS, dl, PtrVT, TLSReg, VariableOffset);
}
- // Only Local-Exec, Initial-Exec and General-Dynamic TLS models are currently
- // supported models. If
orcguru wrote:
> > The primary advantage of the local-dynamic access method is that you only
> > make a single function call to __tls_get_mod() and use the returned pointer
> > to get at all TLS variables. This implementation makes a function call for
> > every variable. The only thing you've
@@ -156,10 +156,11 @@ define void @storesTIInit(double %Val) #0 {
; SMALL32: # %bb.0: # %entry
; SMALL32-NEXT:mflr 0
; SMALL32-NEXT:stwu 1, -32(1)
-; SMALL32-NEXT:lwz 3, L..C4(2) # target-flags(ppc-lo) @TIInit
-; SMALL32-NEXT:lwz 4, L..C5(2) #
@@ -3412,13 +3416,23 @@ SDValue
PPCTargetLowering::LowerGlobalTLSAddressAIX(SDValue Op,
return DAG.getNode(PPCISD::ADD_TLS, dl, PtrVT, TLSReg, VariableOffset);
}
- // Only Local-Exec, Initial-Exec and General-Dynamic TLS models are currently
- // supported models. If
@@ -3412,13 +3416,23 @@ SDValue
PPCTargetLowering::LowerGlobalTLSAddressAIX(SDValue Op,
return DAG.getNode(PPCISD::ADD_TLS, dl, PtrVT, TLSReg, VariableOffset);
}
- // Only Local-Exec, Initial-Exec and General-Dynamic TLS models are currently
- // supported models. If
https://github.com/orcguru resolved
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -1771,9 +1771,13 @@ const char
*PPCTargetLowering::getTargetNodeName(unsigned Opcode) const {
case PPCISD::ADDIS_TLSGD_HA: return "PPCISD::ADDIS_TLSGD_HA";
case PPCISD::ADDI_TLSGD_L:return "PPCISD::ADDI_TLSGD_L";
case PPCISD::GET_TLS_ADDR:return
@@ -2826,6 +2877,8 @@ void PPCAIXAsmPrinter::emitInstruction(const MachineInstr
*MI) {
MMI->hasDebugInfo());
break;
}
+ case PPC::GETtlsMOD32AIX:
+ case PPC::GETtlsMOD64AIX:
case PPC::GETtlsTpointer32AIX:
case PPC::GETtlsADDR64AIX:
case
@@ -1362,6 +1402,8 @@ void PPCAsmPrinter::emitInstruction(const MachineInstr
*MI) {
case PPC::GETtlsADDRPCREL:
case PPC::GETtlsADDR32AIX:
case PPC::GETtlsADDR64AIX:
+ case PPC::GETtlsMOD32AIX:
orcguru wrote:
Added following description:
//
@@ -231,12 +231,15 @@ class PPCTargetAsmStreamer : public PPCTargetStreamer {
MCSymbolXCOFF *TCSym =
cast(Streamer.getCurrentSectionOnly())
->getQualNameSymbol();
- // On AIX, we have a region handle (symbol@m) and the variable offset
-
@@ -1622,6 +1629,10 @@ def TLSGDAIX8 :
"#TLSGDAIX8",
[(set i64:$rD,
(PPCTlsgdAIX i64:$offset, i64:$handle))]>;
+// This pseudo is expanded to one copy to put the module handle in R3, then
call
+//
https://github.com/orcguru resolved
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/orcguru resolved
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/orcguru resolved
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/orcguru resolved
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/orcguru resolved
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/orcguru resolved
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/orcguru resolved
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/orcguru resolved
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -156,10 +156,11 @@ define void @storesTIInit(double %Val) #0 {
; SMALL32: # %bb.0: # %entry
; SMALL32-NEXT:mflr 0
; SMALL32-NEXT:stwu 1, -32(1)
-; SMALL32-NEXT:lwz 3, L..C4(2) # target-flags(ppc-lo) @TIInit
-; SMALL32-NEXT:lwz 4, L..C5(2) #
orcguru wrote:
> The primary advantage of the local-dynamic access method is that you only
> make a single function call to __tls_get_mod() and use the returned pointer
> to get at all TLS variables. This implementation makes a function call for
> every variable. The only thing you've gained
orcguru wrote:
> > I tried to parallelize the review effort, but github didn't show the delta
> > for me: #66972
>
> Let's include FE change in this PR.
Thanks! Added clang change.
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits
https://github.com/orcguru edited
https://github.com/llvm/llvm-project/pull/66316
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
99 matches
Mail list logo
