Call lmul!/rmul! recursively in scalings

[jishnub]

Avoiding broadcasting helps with performance in certain cases:

julia> using LinearAlgebra, SparseArrays

julia> D = Diagonal(sprand(1000, 0.01));

julia> @b D lmul!(2.0, _)
88.674 ns (4 allocs: 384 bytes) # master
10.025 ns # this PR

Similar performance boosts for the other banded matrices.

[dkarrasch]

Nice!

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 91.96%. Comparing base (8cc4216) to head (a218816).
Report is 1 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1236      +/-   ##
==========================================
- Coverage   91.99%   91.96%   -0.04%     
==========================================
  Files          34       34              
  Lines       15480    15475       -5     
==========================================
- Hits        14241    14231      -10     
- Misses       1239     1244       +5     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.