Implement Debian patch for QR decomposition and modify it

The QR decomposition test program can now called with parameters for
matrix size, nev, and block size.
This closses issue #20

Makefile.am

@@ -317,9 +317,9 @@ TESTS = $(check_SCRIPTS)
 	chmod +x $@
 
 ## this one does not want any arguments
-elpa2_test_real_default_kernel_qr_decomposition@SUFFIX@.sh:
-	echo '$(wrapper)./elpa2_test_real_default_kernel_qr_decomposition@SUFFIX@' > $@
-	chmod +x $@
+#elpa2_test_real_default_kernel_qr_decomposition@SUFFIX@.sh:
+#	echo '$(wrapper)./elpa2_test_real_default_kernel_qr_decomposition@SUFFIX@' > $@
+#	chmod +x $@
 
 #elpa2_test_real_default_kernel_qr_decomposition_single_precision@SUFFIX@.sh:
 #	echo '$(wrapper)./elpa2_test_real_default_kernel_qr_decomposition_single_precision@SUFFIX@' > $@

src/elpa2.F90

@@ -193,7 +193,7 @@ function solve_evp_real_2stage(na, nev, a, lda, ev, q, ldq, nblk,        &
    endif
 
    if (useQRActual) then
-     if (mod(na,nblk) .ne. 0) then
+     if (mod(na,2) .ne. 0) then
        if (wantDebug) then
          write(error_unit,*) "solve_evp_real_2stage: QR-decomposition: blocksize does not fit with matrixsize"
        endif

src/elpa_qr/elpa_pdlarfb.F90

@@ -581,6 +581,7 @@ subroutine qr_tmerge_pdlarfb_1dcomm(m,mb,n,oldk,k,v,ldv,t,ldt,a,lda,baseidx,rev,
         else
             ! do not calculate parts for T merge as there is nothing to merge
 
+            mergeoffset  = 0
             updateoffset = 0
 
             tgenoffset = updateoffset + updatesize

test/fortran_test_programs/test_real2_default_kernel_qr_decomposition.F90

@@ -136,19 +136,26 @@ program test_real2
 
    success = .true.
    !write_to_file = .false.
+   call read_input_parameters(na, nev, nblk, write_to_file)
 
-   if (COMMAND_ARGUMENT_COUNT() /= 0) then
-     write(error_unit,*) "This program does not support any command-line arguments"
-     stop 1
-   endif
+   !if (COMMAND_ARGUMENT_COUNT() /= 0) then
+   !  write(error_unit,*) "This program does not support any command-line arguments"
+   !  stop 1
+   !endif
 
    ! override nblk
-   nblk = 2
-   na   = 4000
-   nev  = 1500
+   !   nblk = 2
+   !   na   = 4000
+   !   nev  = 1500
+
+   ! make sure na, nbl is even
+   if (mod(nblk,2 ) .ne. 0) then
+     nblk = nblk - 1
+   endif
+
 
    ! make sure na is even
-   if (mod(na,nblk) .ne. 0) then
+   if (mod(na,2) .ne. 0) then
      na = na - 1
    endif
    ! make sure na is at least 34