BlockArray errors with generic linear algebra functions

When using the generic fallback functions provided by LinearAlgebra, BlockArray type matrices don't work properly. I'm not sure what the best way of handling this would be, but I guess we could theoretically generate versions of the linear algebra functions that fail and have them take the block array and do a conversion to a normal array when calling the standard linear algebra function.

For example, the cond function fails due to a missing bidiagonalize_tall! function. Of course, explicitly converting the block array to a normal array allows the computation to succeed (since it no longer needs the specialization of bidiagonalize_tall!).

julia> D = BlockArray( randn(4,4), [2, 2], [2, 2] )                                                                                                                                                                                                       
2×2-blocked 4×4 BlockArray{Float64,2}:
  0.923874   -0.0829803  │  0.286211   0.46231 
  0.195939    0.14953    │  1.8477    -0.321445
 ────────────────────────┼─────────────────────
 -0.0989605  -1.26688    │  1.42871   -0.659864
  0.403813   -0.86728    │  0.72237   -0.621602

julia> cond( D )                                                                                                                                                                                                                                          
ERROR: MethodError: no method matching bidiagonalize_tall!(::BlockArray{Float64,2,Array{Array{Float64,2},2},Tuple{BlockedUnitRange{Array{Int64,1}},BlockedUnitRange{Array{Int64,1}}}})
Closest candidates are:
  bidiagonalize_tall!(::Array{T,2}) where T at /home/imcinerney/.julia/packages/GenericSVD/cT5Cu/src/bidiagonalize.jl:51
  bidiagonalize_tall!(::Array{T,2}, ::Bidiagonal) where T at /home/imcinerney/.julia/packages/GenericSVD/cT5Cu/src/bidiagonalize.jl:19
  bidiagonalize_tall!(::Adjoint{T2,Array{T,2}}) where {T, T2} at /home/imcinerney/.julia/packages/GenericSVD/cT5Cu/src/bidiagonalize.jl:47
Stacktrace:
 [1] generic_svdvals!(::BlockArray{Float64,2,Array{Array{Float64,2},2},Tuple{BlockedUnitRange{Array{Int64,1}},BlockedUnitRange{Array{Int64,1}}}}; sorted::Bool) at /home/imcinerney/.julia/packages/GenericSVD/cT5Cu/src/GenericSVD.jl:57
 [2] generic_svdvals!(::BlockArray{Float64,2,Array{Array{Float64,2},2},Tuple{BlockedUnitRange{Array{Int64,1}},BlockedUnitRange{Array{Int64,1}}}}) at /home/imcinerney/.julia/packages/GenericSVD/cT5Cu/src/GenericSVD.jl:53
 [3] svdvals!(::BlockArray{Float64,2,Array{Array{Float64,2},2},Tuple{BlockedUnitRange{Array{Int64,1}},BlockedUnitRange{Array{Int64,1}}}}) at /home/imcinerney/.julia/packages/GenericSVD/cT5Cu/src/GenericSVD.jl:21
 [4] svdvals(::BlockArray{Float64,2,Array{Array{Float64,2},2},Tuple{BlockedUnitRange{Array{Int64,1}},BlockedUnitRange{Array{Int64,1}}}}) at /home/imcinerney/dev/julia/main/1.4/usr/share/julia/stdlib/v1.4/LinearAlgebra/src/svd.jl:194
 [5] cond(::BlockArray{Float64,2,Array{Array{Float64,2},2},Tuple{BlockedUnitRange{Array{Int64,1}},BlockedUnitRange{Array{Int64,1}}}}, ::Int64) at /home/imcinerney/dev/julia/main/1.4/usr/share/julia/stdlib/v1.4/LinearAlgebra/src/dense.jl:1419
 [6] cond(::BlockArray{Float64,2,Array{Array{Float64,2},2},Tuple{BlockedUnitRange{Array{Int64,1}},BlockedUnitRange{Array{Int64,1}}}}) at /home/imcinerney/dev/julia/main/1.4/usr/share/julia/stdlib/v1.4/LinearAlgebra/src/dense.jl:1418
 [7] top-level scope at REPL[15]:1
 [8] eval(::Module, ::Any) at ./boot.jl:331
 [9] eval_user_input(::Any, ::REPL.REPLBackend) at /home/imcinerney/dev/julia/main/1.4/usr/share/julia/stdlib/v1.4/REPL/src/REPL.jl:86
 [10] run_backend(::REPL.REPLBackend) at /home/imcinerney/.julia/packages/Revise/XFtoQ/src/Revise.jl:1162
 [11] top-level scope at REPL[1]:0

julia> cond( Array( D ) )                                                                                                                                                                                                                                 
8.508044778284596

So far the functions that I have found fail are (but there are probably more):

• cond
• svd
• svdvals
• eigen

The failure enters through LinearAlgebra.cond and GenericSVD.generic_svdvals! in GenericSVD.jl at line 57; svd, svdvals, and eigen are also reported. Reproduce the BlockArray example and compare it with Array(D); done means these generic functions work on BlockArray without requiring an explicit caller-side conversion.