As far as I can see, free_irq does nothing if its second argument is not the same as the last argument of the corresponding call to request_irq. These were found using the semantic match below (http://coccinelle.lip6.fr/). This semantic match finds a number of other cases, but they are mostly in platform driver probe functions, so the functions should be just converted to use devm functions, eliminating the need to call free_irq at all. // <smpl> @r exists@ expression e,e1,e2,e3,e4,e5; type T; position p1,p2; @@ request_irq@p1(e1,e2,e3,e4,e5) ... ( free_irq(e1,(T)e5); | free_irq@p2(e1,e); ) @bad1 exists@ position r.p1,r.p2; expression e1,e2,e3,e4,e5; @@ request_irq@p1(e1,e2,e3,e4,(void *)e5) ... free_irq@p2(e1,e5); @bad2 exists@ position r.p1,r.p2; statement S; @@ if (request_irq@p1(...)) S else { <+... free_irq@p2(...); ...+> } @script:python depends on !bad1 && !bad2@ p1 << r.p1; p2 << r.p2; @@ cocci.print_main("",p1) cocci.print_secs("",p2) // </smpl> -- To unsubscribe from this list: send the line "unsubscribe kernel-janitors" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
