On Tuesday 22 April 2008 03:01:58 pm Rene Herman wrote: > It seems you designed the list to be basically in any order, judging by > things such as pnp_new_resource which'll happily reuse resources of the > correct type at any position in the list. Yet, pnp_assign_foo() and friends > retrieve resources (through pnp_get_resource) by position in the list and > not by the index. I'm not overly sure of failure scenarios but isn't this > mixing up position and index in a bad way? Yes, I think you're right. My idea is that the list should end up in the same order as the resource list from the BIOS, e.g., the _CRS and _PRS lists for ACPI. In the case of ISAPNP, it'll be in the order that we read things from the hardware registers. I think I should probably make pnp_new_resource() always allocate a new resource and get rid of the idea of reusing something already in the list. > >> (do note that pnp_assign_foo are the only callers of pnp_check_foo and they > >> could be either merged together or at least not communicate via "idx" but > >> simply by passing the res/pnp_res). > > > > Yes, I'd like to do that. But I think I'd better wait or I'll never > > get anything finished :-) > > Well, the idea here was that getting rid of one "idx" here so that things > communicate directly removes at least one possible ordering artifact... Yup. I ended up doing this after all. I updated the patches and my ACPI box with the inactive devices now boots (it failed with the v3 patches). I'll work on the pnp_new_resource() thing tomorrow, but I'll send you the current patches now in case you have a chance to try them on your ISA box. Bjorn -- To unsubscribe from this list: send the line "unsubscribe linux-acpi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
