[clang] [llvm] [clangd] Add support for the c2000 architecture (PR #125663)
James Nagurne via llvm-commits
llvm-commits at lists.llvm.org
Tue Feb 4 14:50:29 PST 2025
================
@@ -0,0 +1,356 @@
+#include "C2000.h"
+#include "Targets.h"
+#include "clang/Basic/Builtins.h"
+#include "clang/Basic/Diagnostic.h"
+#include "clang/Basic/MacroBuilder.h"
+#include "clang/Basic/TargetBuiltins.h"
+
+using namespace clang;
+using namespace clang::targets;
+
+const char *const C2000TargetInfo::GCCRegNames[] = {
+ "ACC", "XAR0", "XAR1", "XAR2", "XAR3", "XAR4", "XAR5", "XAR6", "XAR7"};
+
+ArrayRef<const char *> C2000TargetInfo::getGCCRegNames() const {
+ return llvm::ArrayRef(GCCRegNames);
+}
+
+bool C2000TargetInfo::handleTargetFeatures(std::vector<std::string> &Features,
+ DiagnosticsEngine &Diags) {
+
+ for (const auto &Feature : Features) {
+ if (Feature == "+eabi") {
+ eabi = true;
+ continue;
+ }
+ if (Feature == "+strict_ansi") {
+ strict = true;
+ continue;
+ }
+ if (Feature == "+cla_support") {
+ cla_support = true;
+ }
+ if (Feature == "+cla0") {
+ cla0 = true;
+ continue;
+ }
+ if (Feature == "+cla1") {
+ cla1 = true;
+ continue;
+ }
+ if (Feature == "+cla2") {
+ cla2 = true;
+ continue;
+ }
+ if (Feature == "+relaxed") {
+ relaxed = true;
+ continue;
+ }
+ if (Feature == "+fpu64") {
+ fpu64 = true;
+ continue;
+ }
+ if (Feature == "+fpu32") {
+ fpu32 = true;
+ continue;
+ }
+ if (Feature == "+tmu_support") {
+ tmu_support = true;
+ }
+ if (Feature == "+tmu1") {
+ tmu1 = true;
+ continue;
+ }
+ if (Feature == "+idiv0") {
+ idiv0 = true;
+ continue;
+ }
+ if (Feature == "+vcu_support") {
+ vcu_support = true;
+ }
+ if (Feature == "+vcu2") {
+ vcu2 = true;
+ continue;
+ }
+ if (Feature == "+vcrc") {
+ vcrc = true;
+ continue;
+ }
+ if (Feature == "+opt_level") {
+ opt = true;
+ continue;
+ }
+ }
+ return true;
+}
+
+bool C2000TargetInfo::hasFeature(StringRef Feature) const {
+ return llvm::StringSwitch<bool>(Feature)
+ .Case("eabi", eabi)
+ .Case("strict_ansi", strict)
+ .Case("cla-support", cla_support)
+ .Case("cla0", cla0)
+ .Case("cla1", cla1)
+ .Case("cla2", cla2)
+ .Case("relaxed", relaxed)
+ .Case("fpu64", fpu64)
+ .Case("fpu32", fpu32)
+ .Case("tmu-support", tmu_support)
+ .Case("tmu1", tmu1)
+ .Case("vcu-support", vcu_support)
+ .Case("vcu2", vcu2)
+ .Case("vcrc", vcrc)
+ .Case("opt-level", opt)
+ .Default(false);
+}
+
+void C2000TargetInfo::getTargetDefines(const LangOptions &Opts,
+ MacroBuilder &Builder) const {
+ Builder.undefineMacro("__CHAR_BIT__"); // FIXME: Implement 16-bit char
+ Builder.defineMacro("__CHAR_BIT__", "16");
+ Builder.defineMacro("__TMS320C2000__");
+ Builder.defineMacro("_TMS320C2000");
+ Builder.defineMacro("__TMS320C28XX__");
+ Builder.defineMacro("_TMS320C28XX");
+ Builder.defineMacro("__TMS320C28X__");
+ Builder.defineMacro("_TMS320C28X");
+ Builder.defineMacro("__TI_STRICT_FP_MODE__");
+ Builder.defineMacro("__COMPILER_VERSION__");
+ Builder.defineMacro("__TI_COMPILER_VERSION__");
+ Builder.defineMacro("__TI_COMPILER_VERSION__QUAL_ID");
+ Builder.defineMacro("__TI_COMPILER_VERSION__QUAL__", "QUAL_LETTER");
+ Builder.defineMacro("__little_endian__");
+ Builder.defineMacro("__PTRDIFF_T_TYPE__", "signed long");
+ Builder.defineMacro("__SIZE_T_TYPE__", "unsigned long");
+ Builder.defineMacro("__WCHAR_T_TYPE__", "long unsigned");
+ Builder.defineMacro("__TI_WCHAR_T_BITS", "16");
+ Builder.defineMacro("__TI_C99_COMPLEX_ENABLED");
+ Builder.defineMacro("__TI_GNU_ATTRIBUTE_SUPPORT__");
+ Builder.defineMacro("__LARGE_MODEL__");
+ Builder.defineMacro("__signed_chars__");
+ Builder.defineMacro("__OPTIMIZE_FOR_SPACE");
+
+ if (hasFeature("eabi"))
+ Builder.defineMacro("__TI_EABI__");
+ if (hasFeature("strict_ansi"))
+ Builder.defineMacro("__TI_STRICT_ANSI_MODE__");
+ if (hasFeature("cla-support"))
+ Builder.defineMacro("__TMS320C28XX_CLA__");
+
+ if (hasFeature("cla0"))
+ Builder.defineMacro("__TMS320C28XX_CLA0__");
+ else if (hasFeature("cla1"))
+ Builder.defineMacro("__TMS320C28XX_CLA1__");
+ else if (hasFeature("cla2"))
+ Builder.defineMacro("__TMS320C28XX_CLA2__");
+
+ if (hasFeature("fpu64")) {
+ Builder.defineMacro("__TMS320C28XX_FPU64__");
+ Builder.defineMacro("__TMS320C28XX_FPU32__");
+ } else if (hasFeature("fpu32"))
+ Builder.defineMacro("__TMS320C28XX_FPU32__");
+ if (hasFeature("idiv0"))
+ Builder.defineMacro("__TMS320C28XX_IDIV__");
+ if (hasFeature("tmu1"))
+ Builder.defineMacro("__TMS320C28XX_TMU1__");
+ if (hasFeature("tmu-support")) {
+ Builder.defineMacro("__TMS320C28XX_TMU0__");
+ Builder.defineMacro("__TMS320C28XX_TMU__");
+ }
+ if (hasFeature("vcu-support"))
+ Builder.defineMacro("__TMS320C28XX_VCU0__");
+ if (hasFeature("vcu2"))
+ Builder.defineMacro("__TMS320C28XX_VCU2__");
+ else if (hasFeature("vcrc"))
+ Builder.defineMacro("__TMS320C28XX_VCRC__");
+ if (hasFeature("opt-level"))
+ Builder.defineMacro("_INLINE");
+ if (hasFeature("relaxed"))
+ Builder.undefineMacro("__TI_STRICT_FP_MODE__");
+
+ Builder.defineMacro("__cregister", "");
+ Builder.defineMacro("interrupt", "");
+ Builder.defineMacro("__interrupt", "");
+
+ // Assembly Instrinsics
+
+ Builder.append("int __abs16_sat( int src );");
+ Builder.append("void __add( int *m, int b );");
+ Builder.append("long __addcu( long src1, unsigned int src2 );");
+ Builder.append("void __addl( long *m, long b );");
+ Builder.append("void __and(int *m, int b);");
+ Builder.append("int *__byte_func( int *array, unsigned int byte_index );");
+ Builder.defineMacro("__byte(array, byte_index)",
+ "*__byte_func(array, byte_index)");
+ Builder.append("unsigned long *__byte_peripheral_32_func(unsigned long *x);");
+ Builder.defineMacro("__byte_peripheral_32(x)",
+ "*__byte_peripheral_32_func(x)");
+ Builder.append("void __dec( int *m );");
+
+ // dmac needs macro magic
+ Builder.append("unsigned int __disable_interrupts( );");
+ Builder.append("void __eallow( void );");
+ Builder.append("void __edis( void );");
+ Builder.append("unsigned int __enable_interrupts( );");
+ Builder.append("int __flip16(int src);");
+ Builder.append("long __flip32(long src);");
+ Builder.append("long long __flip64(long long src);");
+ Builder.append("void __inc( int *m );");
+ Builder.append("long __IQ( long double A , int N );");
+ Builder.append("long __IQmpy( long A, long B , int N );");
+ Builder.append("long __IQsat( long A, long max, long min );");
+ Builder.append("long __IQxmpy(long A , long B, int N);");
+ Builder.append("long long __llmax(long long dst, long long src);");
+ Builder.append("long long __llmin(long long dst, long long src);");
+ Builder.append("long __lmax(long dst, long src);");
+ Builder.append("long __lmin(long dst, long src);");
+ Builder.append("int __max(int dst, int src);");
+ Builder.append("int __min(int dst, int src);");
+ Builder.append("int __mov_byte( int *src, unsigned int n );");
+ Builder.append("long __mpy( int src1, int src2 );");
+ Builder.append("long __mpyb( int src1, unsigned int src2 );");
+ Builder.append("long __mpy_mov_t( int src1, int src2, int * dst2 );");
+ Builder.append("unsigned long __mpyu(unsigned int src2, unsigned int srt2);");
+ Builder.append("long __mpyxu( int src1, unsigned int src2 );");
+ Builder.append("long __norm32(long src, int * shift );");
+ Builder.append("long long __norm64(long long src, int * shift );");
+ Builder.append("void __or(int *m, int b);");
+ Builder.append("long __qmpy32( long src32a, long src32b, int q );");
+ Builder.append("long __qmpy32by16(long src32, int src16, int q);");
+ Builder.append("void __restore_interrupts(unsigned int val );");
+ Builder.append("long __rol( long src );");
+ Builder.append("long __ror( long src );");
+ Builder.append("void * __rpt_mov_imm(void * dst , int src ,int count );");
+ Builder.append("int __rpt_norm_inc( long src, int dst, int count );");
+ Builder.append("int __rpt_norm_dec(long src, int dst, int count);");
+ Builder.append("long __rpt_rol(long src, int count);");
+ Builder.append("long __rpt_ror(long src, int count);");
+ Builder.append("long __rpt_subcu(long dst, int src, int count);");
+ Builder.append("unsigned long __rpt_subcul(unsigned long num, unsigned long "
+ "den, unsigned long remainder, int count);");
+ Builder.append("long __sat( long src );");
+ Builder.append("long __sat32( long src, long limit );");
+ Builder.append("long __sathigh16(long src, int limit);");
+ Builder.append("long __satlow16( long src );");
+ Builder.append("long __sbbu( long src1 , unsigned int src2 );");
+ Builder.append("void __sub( int * m, int b );");
+ Builder.append("long __subcu( long src1, int src2 );");
+ Builder.append("unsigned long __subcul(unsigned long num, unsigned long den, "
+ "unsigned long remainder);");
+ Builder.append("void __subl( long * m, long b );");
+ Builder.append("void __subr( int * m , int b );");
+ Builder.append("void __subrl( long * m , long b );");
+ Builder.append("int __tbit( int src , int bit );");
+ Builder.append("void __xor( int * m, int b );");
+
+ // FPU Intrinsics
+ if (hasFeature("fpu64")) {
+ Builder.append("double __einvf64( double x );");
+ Builder.append("double __eisqrtf64( double x );");
+ Builder.append("void __f64_max_idx( double dst, double src, double "
+ "idx_dst, double idx_src );");
+ Builder.append("void __f64_min_idx( double dst, double src, double "
+ "idx_dst, double idx_src );");
+ Builder.append("double __fmax64( double x, double y );");
+ Builder.append("double __fmin64( double x, double y );");
+ Builder.append("double __fracf64(double src );");
+ Builder.append("void __swapff( double &a, double &b );");
+ } else {
+ Builder.append("float __eisqrtf32( float x );");
+ Builder.append("void __f32_max_idx( float dst, float src, float idx_dst, "
+ "float idx_src );");
+ Builder.append("void __f32_min_idx( float dst, float src, float idx_dst, "
+ "float idx_src );");
+ Builder.append("int __f32toi16r(float src );");
+ Builder.append("unsigned int __f32toui16r(float src );");
+ Builder.append("float __fmax( float x, float y );");
+ Builder.append("float __fmin( float x, float y );");
+ Builder.append("float __fracf32(float src );");
+ Builder.append("float __fsat(float val, float max, float min );");
+ Builder.append("void __swapf( float &a, float &b );");
+ Builder.append("void __swapff( float &a, float &b );");
+ }
+
+ // Trigonometric Math Unit (TMU) Intrinsics
+ if (hasFeature("tmu-support")) {
+ Builder.append("float __atan( float src );");
+ Builder.append("float __atan2( float y , float x );");
+ Builder.append("float __atanpuf32( float src );");
+ Builder.append("float __atan2puf32( float x, float y );");
+ Builder.append("float __cos( float src );");
+ Builder.append("float __cospuf32( float src );");
+ Builder.append("float __divf32( float num , float denom );");
+ Builder.append("float __div2pif32( float src );");
+ Builder.append("float __mpy2pif32( float src );");
+ Builder.append("float __quadf32( float ratio, float y, float x );");
+ Builder.append("float __sin( float src );");
+ Builder.append("float __sinpuf32( float src );");
+ Builder.append("float __sqrt( float src );");
+ }
+
+ // Fast Integer Division Intrinsics
+ if (hasFeature("idiv-support")) {
+ Builder.append(
+ "ldiv_t __traditional_div_i16byi16( int dividend, int divisor );");
----------------
DragonDisciple wrote:
This declaration and everything below use some flavor (standard and otherwise) of div_t structure. However, nothing in this pseudo-source would declare those objects.
In the library implementation, available in the installation of stdlib.h, I find that all the div_t flavors are defined in stdlib.h guarded by the macro "\_\_TMS320C28XX_IDIV\_\_'.
This brings up another question: which libc will clangd use? If it uses the host's version, then I foresee many more issues popping up related to missing declarations or declarations that don't quite match.
If it uses the TI C library, then I foresee many more issues popping up related to the assertion that this target **is** 'cl2000'. Take for example the macro _TI_PROPRIETARY_PRAGMA, which is defined in such a way as to be ignored by clang-based compilers but become a TI-only pragma when '\_\_TI_COMPILER_VERSION\_\_" is defined, as it is above. Perhaps clangd will ignore it silently? I don't know...
There are many more uses of "\_\_TI_COMPILER_VERSION\_\_" in our libc that may cause issue, but I haven't dug so deep on those.
https://github.com/llvm/llvm-project/pull/125663
More information about the llvm-commits
mailing list