As Shalani said, breeding to a double dilute will guarantee that the foal inherits a dilute gene.
Breeding a sorrell/chestnut (interchangeable terms as far as genetics go) to a cremello will yield palomino 100% of the time. Breeding to a perlino gives the possibility of buckskin or smoky black, or palomino, all depending on the extension (E) status of the perlino, and the agouti (A) status of both the sire and dam.
With a grey sire, the foal could be born chestnut depending on the sire's E, but if the grey is heterozygous, the foal will have a 50/50 chance of inheriting the grey gene and it will gradually get white hairs as it ages-- ie. "turn grey". If the sire is homozygous for grey, the foal will turn grey 100% of the time.
With a sire that is homozygous black, (meaning EE at extension) you will not get a chestnut foal-- depending on the agouti (A) status of the parents, you can get black or bay or brown.
(OH also a grey sire could also be homozygous black-- the gene for grey does not replace other genes for color a horse has, it just causes white hairs to progressively replace whatrever color the horse was born.)
Both E and A are testable, if you are curious. You would not need to test a sorrell/chestnut mare for E (by her color she is ee) but you could test for A for more info on the possibilities. If you know what color her parents and better yet grandparents and other ancestors were, it might be possible to determine her A staus without a test, by analyzing her pedigree.