RNASeq is very sensitive to batch effects. If we compared normal kidneys from your lab and my lab, there would be significant differences.
If you get control tissues from another experiment, any differences you see might be due entirely to batch, not biology. The most you can do is exploratory work, meaning step one of any further experiments will be to try and verify that what you found is biology, and not batch.
GDC has matched normal tissue controls for kidney cancer (TCGA-KIRC). You can use those.
Also, since batch effects were mentioned in another post: You can do some comparisons between tumor vs. normal from separate experiments -- you just have to be careful about how you do so. You can't do a robust differential gene expression analysis when your tumor/control condition is completely confounded by batches. But that doesn't make the data useless for comparisons: You can do gene sets, pathways, correlations, etc. analyses separately for the normal tissue and the tumor tissue. For example: Regardless of batch effects, kidney markers should generally be ranked highly in kidney normal tissue while tumor markers should generally be ranked highly in tumor tissue, if you rank all genes by expression within individual samples.