Regression analysis - normal distribution and linearity

I am using a scale for a paper. The scores are based on 7 questions with Likert-type responses. I want to use it as a dependent variable in a multiple linear regression. From what I've read this can be ok even though there's disagreement wether this type of scale can be used as a continous measure?

I am a bit confused as to whether the data needs to be normally distributed (which they are not) on the scale when using it in a linear regression analysis. Do they?

What is the best way (in SPSS) to figure out whether the relationship with the independent variables is linear? Do I check this seperately for each variable compared to the dependent variable? If I do a hierarcical regression analysis with one main independent variable in one block and several independent variables that i want to control for in another block, do all the variables need to have a linear relationship to the dependent variable?

If it is not a good idea to use a linear regression I could also use transform the dependent variable into four values on an ordinal scale. What type of test would be the best to compare this to several independent variables?

Any response would be highly appreciated