OpenBLAS/lapack-netlib/SRC/dtgexc.c

#include <math.h>
#include <stdlib.h>
#include <string.h>
#include <stdio.h>
#include <complex.h>
#ifdef complex
#undef complex
#endif
#ifdef I
#undef I
#endif

#if defined(_WIN64)
typedef long long BLASLONG;
typedef unsigned long long BLASULONG;
#else
typedef long BLASLONG;
typedef unsigned long BLASULONG;
#endif

#ifdef LAPACK_ILP64
typedef BLASLONG blasint;
#if defined(_WIN64)
#define blasabs(x) llabs(x)
#else
#define blasabs(x) labs(x)
#endif
#else
typedef int blasint;
#define blasabs(x) abs(x)
#endif

typedef blasint integer;

typedef unsigned int uinteger;
typedef char *address;
typedef short int shortint;
typedef float real;
typedef double doublereal;
typedef struct { real r, i; } complex;
typedef struct { doublereal r, i; } doublecomplex;
#ifdef _MSC_VER
static inline _Fcomplex Cf(complex *z) {_Fcomplex zz={z->r , z->i}; return zz;}
static inline _Dcomplex Cd(doublecomplex *z) {_Dcomplex zz={z->r , z->i};return zz;}
static inline _Fcomplex * _pCf(complex *z) {return (_Fcomplex*)z;}
static inline _Dcomplex * _pCd(doublecomplex *z) {return (_Dcomplex*)z;}
#else
static inline _Complex float Cf(complex *z) {return z->r + z->i*_Complex_I;}
static inline _Complex double Cd(doublecomplex *z) {return z->r + z->i*_Complex_I;}
static inline _Complex float * _pCf(complex *z) {return (_Complex float*)z;}
static inline _Complex double * _pCd(doublecomplex *z) {return (_Complex double*)z;}
#endif
#define pCf(z) (*_pCf(z))
#define pCd(z) (*_pCd(z))
typedef int logical;
typedef short int shortlogical;
typedef char logical1;
typedef char integer1;

#define TRUE_ (1)
#define FALSE_ (0)

/* Extern is for use with -E */
#ifndef Extern
#define Extern extern
#endif

/* I/O stuff */

typedef int flag;
typedef int ftnlen;
typedef int ftnint;

/*external read, write*/
typedef struct
{	flag cierr;
	ftnint ciunit;
	flag ciend;
	char *cifmt;
	ftnint cirec;
} cilist;

/*internal read, write*/
typedef struct
{	flag icierr;
	char *iciunit;
	flag iciend;
	char *icifmt;
	ftnint icirlen;
	ftnint icirnum;
} icilist;

/*open*/
typedef struct
{	flag oerr;
	ftnint ounit;
	char *ofnm;
	ftnlen ofnmlen;
	char *osta;
	char *oacc;
	char *ofm;
	ftnint orl;
	char *oblnk;
} olist;

/*close*/
typedef struct
{	flag cerr;
	ftnint cunit;
	char *csta;
} cllist;

/*rewind, backspace, endfile*/
typedef struct
{	flag aerr;
	ftnint aunit;
} alist;

/* inquire */
typedef struct
{	flag inerr;
	ftnint inunit;
	char *infile;
	ftnlen infilen;
	ftnint	*inex;	/*parameters in standard's order*/
	ftnint	*inopen;
	ftnint	*innum;
	ftnint	*innamed;
	char	*inname;
	ftnlen	innamlen;
	char	*inacc;
	ftnlen	inacclen;
	char	*inseq;
	ftnlen	inseqlen;
	char 	*indir;
	ftnlen	indirlen;
	char	*infmt;
	ftnlen	infmtlen;
	char	*inform;
	ftnint	informlen;
	char	*inunf;
	ftnlen	inunflen;
	ftnint	*inrecl;
	ftnint	*innrec;
	char	*inblank;
	ftnlen	inblanklen;
} inlist;

#define VOID void

union Multitype {	/* for multiple entry points */
	integer1 g;
	shortint h;
	integer i;
	/* longint j; */
	real r;
	doublereal d;
	complex c;
	doublecomplex z;
	};

typedef union Multitype Multitype;

struct Vardesc {	/* for Namelist */
	char *name;
	char *addr;
	ftnlen *dims;
	int  type;
	};
typedef struct Vardesc Vardesc;

struct Namelist {
	char *name;
	Vardesc **vars;
	int nvars;
	};
typedef struct Namelist Namelist;

#define abs(x) ((x) >= 0 ? (x) : -(x))
#define dabs(x) (fabs(x))
#define f2cmin(a,b) ((a) <= (b) ? (a) : (b))
#define f2cmax(a,b) ((a) >= (b) ? (a) : (b))
#define dmin(a,b) (f2cmin(a,b))
#define dmax(a,b) (f2cmax(a,b))
#define bit_test(a,b)	((a) >> (b) & 1)
#define bit_clear(a,b)	((a) & ~((uinteger)1 << (b)))
#define bit_set(a,b)	((a) |  ((uinteger)1 << (b)))

#define abort_() { sig_die("Fortran abort routine called", 1); }
#define c_abs(z) (cabsf(Cf(z)))
#define c_cos(R,Z) { pCf(R)=ccos(Cf(Z)); }
#ifdef _MSC_VER
#define c_div(c, a, b) {Cf(c)._Val[0] = (Cf(a)._Val[0]/Cf(b)._Val[0]); Cf(c)._Val[1]=(Cf(a)._Val[1]/Cf(b)._Val[1]);}
#define z_div(c, a, b) {Cd(c)._Val[0] = (Cd(a)._Val[0]/Cd(b)._Val[0]); Cd(c)._Val[1]=(Cd(a)._Val[1]/df(b)._Val[1]);}
#else
#define c_div(c, a, b) {pCf(c) = Cf(a)/Cf(b);}
#define z_div(c, a, b) {pCd(c) = Cd(a)/Cd(b);}
#endif
#define c_exp(R, Z) {pCf(R) = cexpf(Cf(Z));}
#define c_log(R, Z) {pCf(R) = clogf(Cf(Z));}
#define c_sin(R, Z) {pCf(R) = csinf(Cf(Z));}
//#define c_sqrt(R, Z) {*(R) = csqrtf(Cf(Z));}
#define c_sqrt(R, Z) {pCf(R) = csqrtf(Cf(Z));}
#define d_abs(x) (fabs(*(x)))
#define d_acos(x) (acos(*(x)))
#define d_asin(x) (asin(*(x)))
#define d_atan(x) (atan(*(x)))
#define d_atn2(x, y) (atan2(*(x),*(y)))
#define d_cnjg(R, Z) { pCd(R) = conj(Cd(Z)); }
#define r_cnjg(R, Z) { pCf(R) = conjf(Cf(Z)); }
#define d_cos(x) (cos(*(x)))
#define d_cosh(x) (cosh(*(x)))
#define d_dim(__a, __b) ( *(__a) > *(__b) ? *(__a) - *(__b) : 0.0 )
#define d_exp(x) (exp(*(x)))
#define d_imag(z) (cimag(Cd(z)))
#define r_imag(z) (cimagf(Cf(z)))
#define d_int(__x) (*(__x)>0 ? floor(*(__x)) : -floor(- *(__x)))
#define r_int(__x) (*(__x)>0 ? floor(*(__x)) : -floor(- *(__x)))
#define d_lg10(x) ( 0.43429448190325182765 * log(*(x)) )
#define r_lg10(x) ( 0.43429448190325182765 * log(*(x)) )
#define d_log(x) (log(*(x)))
#define d_mod(x, y) (fmod(*(x), *(y)))
#define u_nint(__x) ((__x)>=0 ? floor((__x) + .5) : -floor(.5 - (__x)))
#define d_nint(x) u_nint(*(x))
#define u_sign(__a,__b) ((__b) >= 0 ? ((__a) >= 0 ? (__a) : -(__a)) : -((__a) >= 0 ? (__a) : -(__a)))
#define d_sign(a,b) u_sign(*(a),*(b))
#define r_sign(a,b) u_sign(*(a),*(b))
#define d_sin(x) (sin(*(x)))
#define d_sinh(x) (sinh(*(x)))
#define d_sqrt(x) (sqrt(*(x)))
#define d_tan(x) (tan(*(x)))
#define d_tanh(x) (tanh(*(x)))
#define i_abs(x) abs(*(x))
#define i_dnnt(x) ((integer)u_nint(*(x)))
#define i_len(s, n) (n)
#define i_nint(x) ((integer)u_nint(*(x)))
#define i_sign(a,b) ((integer)u_sign((integer)*(a),(integer)*(b)))
#define pow_dd(ap, bp) ( pow(*(ap), *(bp)))
#define pow_si(B,E) spow_ui(*(B),*(E))
#define pow_ri(B,E) spow_ui(*(B),*(E))
#define pow_di(B,E) dpow_ui(*(B),*(E))
#define pow_zi(p, a, b) {pCd(p) = zpow_ui(Cd(a), *(b));}
#define pow_ci(p, a, b) {pCf(p) = cpow_ui(Cf(a), *(b));}
#define pow_zz(R,A,B) {pCd(R) = cpow(Cd(A),*(B));}
#define s_cat(lpp, rpp, rnp, np, llp) { 	ftnlen i, nc, ll; char *f__rp, *lp; 	ll = (llp); lp = (lpp); 	for(i=0; i < (int)*(np); ++i) {         	nc = ll; 	        if((rnp)[i] < nc) nc = (rnp)[i]; 	        ll -= nc;         	f__rp = (rpp)[i]; 	        while(--nc >= 0) *lp++ = *(f__rp)++;         } 	while(--ll >= 0) *lp++ = ' '; }
#define s_cmp(a,b,c,d) ((integer)strncmp((a),(b),f2cmin((c),(d))))
#define s_copy(A,B,C,D) { int __i,__m; for (__i=0, __m=f2cmin((C),(D)); __i<__m && (B)[__i] != 0; ++__i) (A)[__i] = (B)[__i]; }
#define sig_die(s, kill) { exit(1); }
#define s_stop(s, n) {exit(0);}
static char junk[] = "\n@(#)LIBF77 VERSION 19990503\n";
#define z_abs(z) (cabs(Cd(z)))
#define z_exp(R, Z) {pCd(R) = cexp(Cd(Z));}
#define z_sqrt(R, Z) {pCd(R) = csqrt(Cd(Z));}
#define myexit_() break;
#define mycycle() continue;
#define myceiling(w) {ceil(w)}
#define myhuge(w) {HUGE_VAL}
//#define mymaxloc_(w,s,e,n) {if (sizeof(*(w)) == sizeof(double)) dmaxloc_((w),*(s),*(e),n); else dmaxloc_((w),*(s),*(e),n);}
#define mymaxloc(w,s,e,n) {dmaxloc_(w,*(s),*(e),n)}

/* procedure parameter types for -A and -C++ */

#define F2C_proc_par_types 1
#ifdef __cplusplus
typedef logical (*L_fp)(...);
#else
typedef logical (*L_fp)();
#endif

static float spow_ui(float x, integer n) {
	float pow=1.0; unsigned long int u;
	if(n != 0) {
		if(n < 0) n = -n, x = 1/x;
		for(u = n; ; ) {
			if(u & 01) pow *= x;
			if(u >>= 1) x *= x;
			else break;
		}
	}
	return pow;
}
static double dpow_ui(double x, integer n) {
	double pow=1.0; unsigned long int u;
	if(n != 0) {
		if(n < 0) n = -n, x = 1/x;
		for(u = n; ; ) {
			if(u & 01) pow *= x;
			if(u >>= 1) x *= x;
			else break;
		}
	}
	return pow;
}
#ifdef _MSC_VER
static _Fcomplex cpow_ui(complex x, integer n) {
	complex pow={1.0,0.0}; unsigned long int u;
		if(n != 0) {
		if(n < 0) n = -n, x.r = 1/x.r, x.i=1/x.i;
		for(u = n; ; ) {
			if(u & 01) pow.r *= x.r, pow.i *= x.i;
			if(u >>= 1) x.r *= x.r, x.i *= x.i;
			else break;
		}
	}
	_Fcomplex p={pow.r, pow.i};
	return p;
}
#else
static _Complex float cpow_ui(_Complex float x, integer n) {
	_Complex float pow=1.0; unsigned long int u;
	if(n != 0) {
		if(n < 0) n = -n, x = 1/x;
		for(u = n; ; ) {
			if(u & 01) pow *= x;
			if(u >>= 1) x *= x;
			else break;
		}
	}
	return pow;
}
#endif
#ifdef _MSC_VER
static _Dcomplex zpow_ui(_Dcomplex x, integer n) {
	_Dcomplex pow={1.0,0.0}; unsigned long int u;
	if(n != 0) {
		if(n < 0) n = -n, x._Val[0] = 1/x._Val[0], x._Val[1] =1/x._Val[1];
		for(u = n; ; ) {
			if(u & 01) pow._Val[0] *= x._Val[0], pow._Val[1] *= x._Val[1];
			if(u >>= 1) x._Val[0] *= x._Val[0], x._Val[1] *= x._Val[1];
			else break;
		}
	}
	_Dcomplex p = {pow._Val[0], pow._Val[1]};
	return p;
}
#else
static _Complex double zpow_ui(_Complex double x, integer n) {
	_Complex double pow=1.0; unsigned long int u;
	if(n != 0) {
		if(n < 0) n = -n, x = 1/x;
		for(u = n; ; ) {
			if(u & 01) pow *= x;
			if(u >>= 1) x *= x;
			else break;
		}
	}
	return pow;
}
#endif
static integer pow_ii(integer x, integer n) {
	integer pow; unsigned long int u;
	if (n <= 0) {
		if (n == 0 || x == 1) pow = 1;
		else if (x != -1) pow = x == 0 ? 1/x : 0;
		else n = -n;
	}
	if ((n > 0) || !(n == 0 || x == 1 || x != -1)) {
		u = n;
		for(pow = 1; ; ) {
			if(u & 01) pow *= x;
			if(u >>= 1) x *= x;
			else break;
		}
	}
	return pow;
}
static integer dmaxloc_(double *w, integer s, integer e, integer *n)
{
	double m; integer i, mi;
	for(m=w[s-1], mi=s, i=s+1; i<=e; i++)
		if (w[i-1]>m) mi=i ,m=w[i-1];
	return mi-s+1;
}
static integer smaxloc_(float *w, integer s, integer e, integer *n)
{
	float m; integer i, mi;
	for(m=w[s-1], mi=s, i=s+1; i<=e; i++)
		if (w[i-1]>m) mi=i ,m=w[i-1];
	return mi-s+1;
}
static inline void cdotc_(complex *z, integer *n_, complex *x, integer *incx_, complex *y, integer *incy_) {
	integer n = *n_, incx = *incx_, incy = *incy_, i;
#ifdef _MSC_VER
	_Fcomplex zdotc = {0.0, 0.0};
	if (incx == 1 && incy == 1) {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc._Val[0] += conjf(Cf(&x[i]))._Val[0] * Cf(&y[i])._Val[0];
			zdotc._Val[1] += conjf(Cf(&x[i]))._Val[1] * Cf(&y[i])._Val[1];
		}
	} else {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc._Val[0] += conjf(Cf(&x[i*incx]))._Val[0] * Cf(&y[i*incy])._Val[0];
			zdotc._Val[1] += conjf(Cf(&x[i*incx]))._Val[1] * Cf(&y[i*incy])._Val[1];
		}
	}
	pCf(z) = zdotc;
}
#else
	_Complex float zdotc = 0.0;
	if (incx == 1 && incy == 1) {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc += conjf(Cf(&x[i])) * Cf(&y[i]);
		}
	} else {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc += conjf(Cf(&x[i*incx])) * Cf(&y[i*incy]);
		}
	}
	pCf(z) = zdotc;
}
#endif
static inline void zdotc_(doublecomplex *z, integer *n_, doublecomplex *x, integer *incx_, doublecomplex *y, integer *incy_) {
	integer n = *n_, incx = *incx_, incy = *incy_, i;
#ifdef _MSC_VER
	_Dcomplex zdotc = {0.0, 0.0};
	if (incx == 1 && incy == 1) {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc._Val[0] += conj(Cd(&x[i]))._Val[0] * Cd(&y[i])._Val[0];
			zdotc._Val[1] += conj(Cd(&x[i]))._Val[1] * Cd(&y[i])._Val[1];
		}
	} else {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc._Val[0] += conj(Cd(&x[i*incx]))._Val[0] * Cd(&y[i*incy])._Val[0];
			zdotc._Val[1] += conj(Cd(&x[i*incx]))._Val[1] * Cd(&y[i*incy])._Val[1];
		}
	}
	pCd(z) = zdotc;
}
#else
	_Complex double zdotc = 0.0;
	if (incx == 1 && incy == 1) {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc += conj(Cd(&x[i])) * Cd(&y[i]);
		}
	} else {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc += conj(Cd(&x[i*incx])) * Cd(&y[i*incy]);
		}
	}
	pCd(z) = zdotc;
}
#endif	
static inline void cdotu_(complex *z, integer *n_, complex *x, integer *incx_, complex *y, integer *incy_) {
	integer n = *n_, incx = *incx_, incy = *incy_, i;
#ifdef _MSC_VER
	_Fcomplex zdotc = {0.0, 0.0};
	if (incx == 1 && incy == 1) {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc._Val[0] += Cf(&x[i])._Val[0] * Cf(&y[i])._Val[0];
			zdotc._Val[1] += Cf(&x[i])._Val[1] * Cf(&y[i])._Val[1];
		}
	} else {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc._Val[0] += Cf(&x[i*incx])._Val[0] * Cf(&y[i*incy])._Val[0];
			zdotc._Val[1] += Cf(&x[i*incx])._Val[1] * Cf(&y[i*incy])._Val[1];
		}
	}
	pCf(z) = zdotc;
}
#else
	_Complex float zdotc = 0.0;
	if (incx == 1 && incy == 1) {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc += Cf(&x[i]) * Cf(&y[i]);
		}
	} else {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc += Cf(&x[i*incx]) * Cf(&y[i*incy]);
		}
	}
	pCf(z) = zdotc;
}
#endif
static inline void zdotu_(doublecomplex *z, integer *n_, doublecomplex *x, integer *incx_, doublecomplex *y, integer *incy_) {
	integer n = *n_, incx = *incx_, incy = *incy_, i;
#ifdef _MSC_VER
	_Dcomplex zdotc = {0.0, 0.0};
	if (incx == 1 && incy == 1) {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc._Val[0] += Cd(&x[i])._Val[0] * Cd(&y[i])._Val[0];
			zdotc._Val[1] += Cd(&x[i])._Val[1] * Cd(&y[i])._Val[1];
		}
	} else {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc._Val[0] += Cd(&x[i*incx])._Val[0] * Cd(&y[i*incy])._Val[0];
			zdotc._Val[1] += Cd(&x[i*incx])._Val[1] * Cd(&y[i*incy])._Val[1];
		}
	}
	pCd(z) = zdotc;
}
#else
	_Complex double zdotc = 0.0;
	if (incx == 1 && incy == 1) {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc += Cd(&x[i]) * Cd(&y[i]);
		}
	} else {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc += Cd(&x[i*incx]) * Cd(&y[i*incy]);
		}
	}
	pCd(z) = zdotc;
}
#endif
/*  -- translated by f2c (version 20000121).
   You must link the resulting object file with the libraries:
	-lf2c -lm   (in that order)
*/




/* Table of constant values */

static integer c__1 = 1;
static integer c__2 = 2;

/* > \brief \b DTGEXC */

/*  =========== DOCUMENTATION =========== */

/* Online html documentation available at */
/*            http://www.netlib.org/lapack/explore-html/ */

/* > \htmlonly */
/* > Download DTGEXC + dependencies */
/* > <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/dtgexc.
f"> */
/* > [TGZ]</a> */
/* > <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/dtgexc.
f"> */
/* > [ZIP]</a> */
/* > <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/dtgexc.
f"> */
/* > [TXT]</a> */
/* > \endhtmlonly */

/*  Definition: */
/*  =========== */

/*       SUBROUTINE DTGEXC( WANTQ, WANTZ, N, A, LDA, B, LDB, Q, LDQ, Z, */
/*                          LDZ, IFST, ILST, WORK, LWORK, INFO ) */

/*       LOGICAL            WANTQ, WANTZ */
/*       INTEGER            IFST, ILST, INFO, LDA, LDB, LDQ, LDZ, LWORK, N */
/*       DOUBLE PRECISION   A( LDA, * ), B( LDB, * ), Q( LDQ, * ), */
/*      $                   WORK( * ), Z( LDZ, * ) */


/* > \par Purpose: */
/*  ============= */
/* > */
/* > \verbatim */
/* > */
/* > DTGEXC reorders the generalized real Schur decomposition of a real */
/* > matrix pair (A,B) using an orthogonal equivalence transformation */
/* > */
/* >                (A, B) = Q * (A, B) * Z**T, */
/* > */
/* > so that the diagonal block of (A, B) with row index IFST is moved */
/* > to row ILST. */
/* > */
/* > (A, B) must be in generalized real Schur canonical form (as returned */
/* > by DGGES), i.e. A is block upper triangular with 1-by-1 and 2-by-2 */
/* > diagonal blocks. B is upper triangular. */
/* > */
/* > Optionally, the matrices Q and Z of generalized Schur vectors are */
/* > updated. */
/* > */
/* >        Q(in) * A(in) * Z(in)**T = Q(out) * A(out) * Z(out)**T */
/* >        Q(in) * B(in) * Z(in)**T = Q(out) * B(out) * Z(out)**T */
/* > */
/* > \endverbatim */

/*  Arguments: */
/*  ========== */

/* > \param[in] WANTQ */
/* > \verbatim */
/* >          WANTQ is LOGICAL */
/* >          .TRUE. : update the left transformation matrix Q; */
/* >          .FALSE.: do not update Q. */
/* > \endverbatim */
/* > */
/* > \param[in] WANTZ */
/* > \verbatim */
/* >          WANTZ is LOGICAL */
/* >          .TRUE. : update the right transformation matrix Z; */
/* >          .FALSE.: do not update Z. */
/* > \endverbatim */
/* > */
/* > \param[in] N */
/* > \verbatim */
/* >          N is INTEGER */
/* >          The order of the matrices A and B. N >= 0. */
/* > \endverbatim */
/* > */
/* > \param[in,out] A */
/* > \verbatim */
/* >          A is DOUBLE PRECISION array, dimension (LDA,N) */
/* >          On entry, the matrix A in generalized real Schur canonical */
/* >          form. */
/* >          On exit, the updated matrix A, again in generalized */
/* >          real Schur canonical form. */
/* > \endverbatim */
/* > */
/* > \param[in] LDA */
/* > \verbatim */
/* >          LDA is INTEGER */
/* >          The leading dimension of the array A. LDA >= f2cmax(1,N). */
/* > \endverbatim */
/* > */
/* > \param[in,out] B */
/* > \verbatim */
/* >          B is DOUBLE PRECISION array, dimension (LDB,N) */
/* >          On entry, the matrix B in generalized real Schur canonical */
/* >          form (A,B). */
/* >          On exit, the updated matrix B, again in generalized */
/* >          real Schur canonical form (A,B). */
/* > \endverbatim */
/* > */
/* > \param[in] LDB */
/* > \verbatim */
/* >          LDB is INTEGER */
/* >          The leading dimension of the array B. LDB >= f2cmax(1,N). */
/* > \endverbatim */
/* > */
/* > \param[in,out] Q */
/* > \verbatim */
/* >          Q is DOUBLE PRECISION array, dimension (LDQ,N) */
/* >          On entry, if WANTQ = .TRUE., the orthogonal matrix Q. */
/* >          On exit, the updated matrix Q. */
/* >          If WANTQ = .FALSE., Q is not referenced. */
/* > \endverbatim */
/* > */
/* > \param[in] LDQ */
/* > \verbatim */
/* >          LDQ is INTEGER */
/* >          The leading dimension of the array Q. LDQ >= 1. */
/* >          If WANTQ = .TRUE., LDQ >= N. */
/* > \endverbatim */
/* > */
/* > \param[in,out] Z */
/* > \verbatim */
/* >          Z is DOUBLE PRECISION array, dimension (LDZ,N) */
/* >          On entry, if WANTZ = .TRUE., the orthogonal matrix Z. */
/* >          On exit, the updated matrix Z. */
/* >          If WANTZ = .FALSE., Z is not referenced. */
/* > \endverbatim */
/* > */
/* > \param[in] LDZ */
/* > \verbatim */
/* >          LDZ is INTEGER */
/* >          The leading dimension of the array Z. LDZ >= 1. */
/* >          If WANTZ = .TRUE., LDZ >= N. */
/* > \endverbatim */
/* > */
/* > \param[in,out] IFST */
/* > \verbatim */
/* >          IFST is INTEGER */
/* > \endverbatim */
/* > */
/* > \param[in,out] ILST */
/* > \verbatim */
/* >          ILST is INTEGER */
/* >          Specify the reordering of the diagonal blocks of (A, B). */
/* >          The block with row index IFST is moved to row ILST, by a */
/* >          sequence of swapping between adjacent blocks. */
/* >          On exit, if IFST pointed on entry to the second row of */
/* >          a 2-by-2 block, it is changed to point to the first row; */
/* >          ILST always points to the first row of the block in its */
/* >          final position (which may differ from its input value by */
/* >          +1 or -1). 1 <= IFST, ILST <= N. */
/* > \endverbatim */
/* > */
/* > \param[out] WORK */
/* > \verbatim */
/* >          WORK is DOUBLE PRECISION array, dimension (MAX(1,LWORK)) */
/* >          On exit, if INFO = 0, WORK(1) returns the optimal LWORK. */
/* > \endverbatim */
/* > */
/* > \param[in] LWORK */
/* > \verbatim */
/* >          LWORK is INTEGER */
/* >          The dimension of the array WORK. */
/* >          LWORK >= 1 when N <= 1, otherwise LWORK >= 4*N + 16. */
/* > */
/* >          If LWORK = -1, then a workspace query is assumed; the routine */
/* >          only calculates the optimal size of the WORK array, returns */
/* >          this value as the first entry of the WORK array, and no error */
/* >          message related to LWORK is issued by XERBLA. */
/* > \endverbatim */
/* > */
/* > \param[out] INFO */
/* > \verbatim */
/* >          INFO is INTEGER */
/* >           =0:  successful exit. */
/* >           <0:  if INFO = -i, the i-th argument had an illegal value. */
/* >           =1:  The transformed matrix pair (A, B) would be too far */
/* >                from generalized Schur form; the problem is ill- */
/* >                conditioned. (A, B) may have been partially reordered, */
/* >                and ILST points to the first row of the current */
/* >                position of the block being moved. */
/* > \endverbatim */

/*  Authors: */
/*  ======== */

/* > \author Univ. of Tennessee */
/* > \author Univ. of California Berkeley */
/* > \author Univ. of Colorado Denver */
/* > \author NAG Ltd. */

/* > \date December 2016 */

/* > \ingroup doubleGEcomputational */

/* > \par Contributors: */
/*  ================== */
/* > */
/* >     Bo Kagstrom and Peter Poromaa, Department of Computing Science, */
/* >     Umea University, S-901 87 Umea, Sweden. */

/* > \par References: */
/*  ================ */
/* > */
/* > \verbatim */
/* > */
/* >  [1] B. Kagstrom; A Direct Method for Reordering Eigenvalues in the */
/* >      Generalized Real Schur Form of a Regular Matrix Pair (A, B), in */
/* >      M.S. Moonen et al (eds), Linear Algebra for Large Scale and */
/* >      Real-Time Applications, Kluwer Academic Publ. 1993, pp 195-218. */
/* > \endverbatim */
/* > */
/*  ===================================================================== */
/* Subroutine */ void dtgexc_(logical *wantq, logical *wantz, integer *n, 
	doublereal *a, integer *lda, doublereal *b, integer *ldb, doublereal *
	q, integer *ldq, doublereal *z__, integer *ldz, integer *ifst, 
	integer *ilst, doublereal *work, integer *lwork, integer *info)
{
    /* System generated locals */
    integer a_dim1, a_offset, b_dim1, b_offset, q_dim1, q_offset, z_dim1, 
	    z_offset, i__1;

    /* Local variables */
    integer here, lwmin;
    extern /* Subroutine */ void dtgex2_(logical *, logical *, integer *, 
	    doublereal *, integer *, doublereal *, integer *, doublereal *, 
	    integer *, doublereal *, integer *, integer *, integer *, integer 
	    *, doublereal *, integer *, integer *);
    extern int xerbla_(char *, integer *, ftnlen);
    integer nbnext;
    logical lquery;
    integer nbf, nbl;


/*  -- LAPACK computational routine (version 3.7.0) -- */
/*  -- LAPACK is a software package provided by Univ. of Tennessee,    -- */
/*  -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..-- */
/*     December 2016 */


/*  ===================================================================== */


/*     Decode and test input arguments. */

    /* Parameter adjustments */
    a_dim1 = *lda;
    a_offset = 1 + a_dim1 * 1;
    a -= a_offset;
    b_dim1 = *ldb;
    b_offset = 1 + b_dim1 * 1;
    b -= b_offset;
    q_dim1 = *ldq;
    q_offset = 1 + q_dim1 * 1;
    q -= q_offset;
    z_dim1 = *ldz;
    z_offset = 1 + z_dim1 * 1;
    z__ -= z_offset;
    --work;

    /* Function Body */
    *info = 0;
    lquery = *lwork == -1;
    if (*n < 0) {
	*info = -3;
    } else if (*lda < f2cmax(1,*n)) {
	*info = -5;
    } else if (*ldb < f2cmax(1,*n)) {
	*info = -7;
    } else if (*ldq < 1 || *wantq && *ldq < f2cmax(1,*n)) {
	*info = -9;
    } else if (*ldz < 1 || *wantz && *ldz < f2cmax(1,*n)) {
	*info = -11;
    } else if (*ifst < 1 || *ifst > *n) {
	*info = -12;
    } else if (*ilst < 1 || *ilst > *n) {
	*info = -13;
    }

    if (*info == 0) {
	if (*n <= 1) {
	    lwmin = 1;
	} else {
	    lwmin = (*n << 2) + 16;
	}
	work[1] = (doublereal) lwmin;

	if (*lwork < lwmin && ! lquery) {
	    *info = -15;
	}
    }

    if (*info != 0) {
	i__1 = -(*info);
	xerbla_("DTGEXC", &i__1, (ftnlen)6);
	return;
    } else if (lquery) {
	return;
    }

/*     Quick return if possible */

    if (*n <= 1) {
	return;
    }

/*     Determine the first row of the specified block and find out */
/*     if it is 1-by-1 or 2-by-2. */

    if (*ifst > 1) {
	if (a[*ifst + (*ifst - 1) * a_dim1] != 0.) {
	    --(*ifst);
	}
    }
    nbf = 1;
    if (*ifst < *n) {
	if (a[*ifst + 1 + *ifst * a_dim1] != 0.) {
	    nbf = 2;
	}
    }

/*     Determine the first row of the final block */
/*     and find out if it is 1-by-1 or 2-by-2. */

    if (*ilst > 1) {
	if (a[*ilst + (*ilst - 1) * a_dim1] != 0.) {
	    --(*ilst);
	}
    }
    nbl = 1;
    if (*ilst < *n) {
	if (a[*ilst + 1 + *ilst * a_dim1] != 0.) {
	    nbl = 2;
	}
    }
    if (*ifst == *ilst) {
	return;
    }

    if (*ifst < *ilst) {

/*        Update ILST. */

	if (nbf == 2 && nbl == 1) {
	    --(*ilst);
	}
	if (nbf == 1 && nbl == 2) {
	    ++(*ilst);
	}

	here = *ifst;

L10:

/*        Swap with next one below. */

	if (nbf == 1 || nbf == 2) {

/*           Current block either 1-by-1 or 2-by-2. */

	    nbnext = 1;
	    if (here + nbf + 1 <= *n) {
		if (a[here + nbf + 1 + (here + nbf) * a_dim1] != 0.) {
		    nbnext = 2;
		}
	    }
	    dtgex2_(wantq, wantz, n, &a[a_offset], lda, &b[b_offset], ldb, &q[
		    q_offset], ldq, &z__[z_offset], ldz, &here, &nbf, &nbnext,
		     &work[1], lwork, info);
	    if (*info != 0) {
		*ilst = here;
		return;
	    }
	    here += nbnext;

/*           Test if 2-by-2 block breaks into two 1-by-1 blocks. */

	    if (nbf == 2) {
		if (a[here + 1 + here * a_dim1] == 0.) {
		    nbf = 3;
		}
	    }

	} else {

/*           Current block consists of two 1-by-1 blocks, each of which */
/*           must be swapped individually. */

	    nbnext = 1;
	    if (here + 3 <= *n) {
		if (a[here + 3 + (here + 2) * a_dim1] != 0.) {
		    nbnext = 2;
		}
	    }
	    i__1 = here + 1;
	    dtgex2_(wantq, wantz, n, &a[a_offset], lda, &b[b_offset], ldb, &q[
		    q_offset], ldq, &z__[z_offset], ldz, &i__1, &c__1, &
		    nbnext, &work[1], lwork, info);
	    if (*info != 0) {
		*ilst = here;
		return;
	    }
	    if (nbnext == 1) {

/*              Swap two 1-by-1 blocks. */

		dtgex2_(wantq, wantz, n, &a[a_offset], lda, &b[b_offset], ldb,
			 &q[q_offset], ldq, &z__[z_offset], ldz, &here, &c__1,
			 &c__1, &work[1], lwork, info);
		if (*info != 0) {
		    *ilst = here;
		    return;
		}
		++here;

	    } else {

/*              Recompute NBNEXT in case of 2-by-2 split. */

		if (a[here + 2 + (here + 1) * a_dim1] == 0.) {
		    nbnext = 1;
		}
		if (nbnext == 2) {

/*                 2-by-2 block did not split. */

		    dtgex2_(wantq, wantz, n, &a[a_offset], lda, &b[b_offset], 
			    ldb, &q[q_offset], ldq, &z__[z_offset], ldz, &
			    here, &c__1, &nbnext, &work[1], lwork, info);
		    if (*info != 0) {
			*ilst = here;
			return;
		    }
		    here += 2;
		} else {

/*                 2-by-2 block did split. */

		    dtgex2_(wantq, wantz, n, &a[a_offset], lda, &b[b_offset], 
			    ldb, &q[q_offset], ldq, &z__[z_offset], ldz, &
			    here, &c__1, &c__1, &work[1], lwork, info);
		    if (*info != 0) {
			*ilst = here;
			return;
		    }
		    ++here;
		    dtgex2_(wantq, wantz, n, &a[a_offset], lda, &b[b_offset], 
			    ldb, &q[q_offset], ldq, &z__[z_offset], ldz, &
			    here, &c__1, &c__1, &work[1], lwork, info);
		    if (*info != 0) {
			*ilst = here;
			return;
		    }
		    ++here;
		}

	    }
	}
	if (here < *ilst) {
	    goto L10;
	}
    } else {
	here = *ifst;

L20:

/*        Swap with next one below. */

	if (nbf == 1 || nbf == 2) {

/*           Current block either 1-by-1 or 2-by-2. */

	    nbnext = 1;
	    if (here >= 3) {
		if (a[here - 1 + (here - 2) * a_dim1] != 0.) {
		    nbnext = 2;
		}
	    }
	    i__1 = here - nbnext;
	    dtgex2_(wantq, wantz, n, &a[a_offset], lda, &b[b_offset], ldb, &q[
		    q_offset], ldq, &z__[z_offset], ldz, &i__1, &nbnext, &nbf,
		     &work[1], lwork, info);
	    if (*info != 0) {
		*ilst = here;
		return;
	    }
	    here -= nbnext;

/*           Test if 2-by-2 block breaks into two 1-by-1 blocks. */

	    if (nbf == 2) {
		if (a[here + 1 + here * a_dim1] == 0.) {
		    nbf = 3;
		}
	    }

	} else {

/*           Current block consists of two 1-by-1 blocks, each of which */
/*           must be swapped individually. */

	    nbnext = 1;
	    if (here >= 3) {
		if (a[here - 1 + (here - 2) * a_dim1] != 0.) {
		    nbnext = 2;
		}
	    }
	    i__1 = here - nbnext;
	    dtgex2_(wantq, wantz, n, &a[a_offset], lda, &b[b_offset], ldb, &q[
		    q_offset], ldq, &z__[z_offset], ldz, &i__1, &nbnext, &
		    c__1, &work[1], lwork, info);
	    if (*info != 0) {
		*ilst = here;
		return;
	    }
	    if (nbnext == 1) {

/*              Swap two 1-by-1 blocks. */

		dtgex2_(wantq, wantz, n, &a[a_offset], lda, &b[b_offset], ldb,
			 &q[q_offset], ldq, &z__[z_offset], ldz, &here, &
			nbnext, &c__1, &work[1], lwork, info);
		if (*info != 0) {
		    *ilst = here;
		    return;
		}
		--here;
	    } else {

/*             Recompute NBNEXT in case of 2-by-2 split. */

		if (a[here + (here - 1) * a_dim1] == 0.) {
		    nbnext = 1;
		}
		if (nbnext == 2) {

/*                 2-by-2 block did not split. */

		    i__1 = here - 1;
		    dtgex2_(wantq, wantz, n, &a[a_offset], lda, &b[b_offset], 
			    ldb, &q[q_offset], ldq, &z__[z_offset], ldz, &
			    i__1, &c__2, &c__1, &work[1], lwork, info);
		    if (*info != 0) {
			*ilst = here;
			return;
		    }
		    here += -2;
		} else {

/*                 2-by-2 block did split. */

		    dtgex2_(wantq, wantz, n, &a[a_offset], lda, &b[b_offset], 
			    ldb, &q[q_offset], ldq, &z__[z_offset], ldz, &
			    here, &c__1, &c__1, &work[1], lwork, info);
		    if (*info != 0) {
			*ilst = here;
			return;
		    }
		    --here;
		    dtgex2_(wantq, wantz, n, &a[a_offset], lda, &b[b_offset], 
			    ldb, &q[q_offset], ldq, &z__[z_offset], ldz, &
			    here, &c__1, &c__1, &work[1], lwork, info);
		    if (*info != 0) {
			*ilst = here;
			return;
		    }
		    --here;
		}
	    }
	}
	if (here > *ilst) {
	    goto L20;
	}
    }
    *ilst = here;
    work[1] = (doublereal) lwmin;
    return;

/*     End of DTGEXC */

} /* dtgexc_ */