Per-device parameter support (11/16)
dp_11_module_param_arr.diff
This is the 11st patch of 16 patches for devparam.
The unsigned int * @nump of module_param_array is changed back to
unsigned int @num, and new sets of macros named module_param_arr*() are added. These new macros don't take the num argument. This change is made for two reasons
1. To be consistent with devparam macros. In devparam, we'll be using field name of struct elements, so we won't be able to use pointer argument.
2. It's more consistent with other moduleparam macros.
This patch only modifies moduleparam.h and doesn't modify the users
of the modified macros. The next patch takes care of that. This and the next patch (dp_12_module_param_arr_apply.diff) are optional.
include/linux/moduleparam.h +25 -8
@@ -157,22 +157,39 @@ extern int param_set_invbool(const charextern int param_get_invbool (char *buffer,struct kernel_param *kp);#define param_check_invbool (name, p) __param_check(name, p, int) -/* Comma-separated array: *nump is set to number they actually specified. */ -#define module_param_array_named_ranged (name, array, type, nump, min, max, perm)\ +/* Array fundamental function */ +#define __module_param_array (name, array, type, nump, min, max, perm) \ static struct kparam_array __param_arr_##name \ = { ARRAY_SIZE(array), nump, param_set_##type, param_get_##type,\ sizeof(array[0]), array }; \ module_param_call_ranged(name, param_array_set, param_array_get,\ &__param_arr_##name, min, max, perm) -#define module_param_array_ranged (name, type, nump, min, max, perm) \ - module_param_array_named_ranged(name, name, type, nump, min, max, perm) +/* Comma-separated array: num is set to number they actually specified. */ +#define module_param_array_named_ranged (name, array, type, num, min, max, perm)\ + __module_param_array(name, array, type, &(num), min, max, perm) -#define module_param_array_named (name, array, type, nump, perm) \ - module_param_array_named_ranged(name, array, type, nump, 1, 0, perm) +#define module_param_array_ranged (name, type, num, min, max, perm) \ + module_param_array_named_ranged(name, name, type, num, min, max, perm) -#define module_param_array (name, type, nump, perm) \ - module_param_array_named(name, name, type, nump, perm) +#define module_param_array_named (name, array, type, num, perm) \ + module_param_array_named_ranged(name, array, type, num, 1, 0, perm) + +#define module_param_array (name, type, num, perm) \ + module_param_array_named(name, name, type, num, perm) + +/* Comma-separated array without len. */ +#define module_param_arr_named_ranged (name, array, type, min, max, perm)\ + __module_param_array(name, array, type, NULL, min, max, perm) + +#define module_param_arr_ranged (name, type, min, max, perm) \ + module_param_arr_named_ranged(name, name, type, min, max, perm) + +#define module_param_arr_named (name, array, type, perm) \ + module_param_arr_named_ranged(name, array, type, 1, 0, perm) + +#define module_param_arr (name, type, perm) \ + module_param_arr_named(name, name, type, perm)extern int param_array_set (const char *val,struct kernel_param *kp);extern int param_array_get (char *buffer,struct kernel_param *kp);
Quoted 9 lines · Tejun Heo
dp_11_module_param_arr.diff
This is the 11st patch of 16 patches for devparam.
The unsigned int * @nump of
module_param_arrayis changed back to
unsigned int @num, and new sets of macros named module_param_arr*() are added. These new macros don't take the num argument. This change is made for two reasons
Quoted 8 lines
1. To be consistent with devparam macros. In devparam, we'll be using field name of struct elements, so we won't be able to use pointer argument.
2. It's more consistent with other moduleparam macros.This patch only modifies
moduleparam.hand doesn't modify the users
of the modified macros. The next patch takes care of that. This and the next patch (dp_12_module_param_arr_apply.diff) are optional.
It's finely balanced, but the module_param macro only skips the "&" because it turns it into a name. module_param_named() should probably be changed to take a ptr too. But you'll get a warning if you use the interface wrong, so it's not a huge issue.
The balance is tipped here I think because the massive number of functions we're starting to sprout into.
Rusty.
Signature · 1 line
A bad analogy is like a leaky screwdriver -- Richard Braakman