On Thu, Jan 03, 2019 at 10:47:14PM +1100, Jonathan Gray wrote:
> On Wed, Jan 02, 2019 at 04:08:53PM +1000, David Gwynne wrote:
> > This makes it more obvious what interesting things there are to hack on.
> >
> > Thoughts? ok?
> >
> > For example, from an od1000:
> >
> > dlg@o1000 fdt$ dmesg | grep 'not configured'
>
> The whitespace is wrong on 'return (UNCONF)'
>
> The same diff for armv7 on cubox:
And octeon which I can't test:
Index: simplebus.c
===================================================================
RCS file: /cvs/src/sys/arch/octeon/dev/simplebus.c,v
retrieving revision 1.1
diff -u -p -r1.1 simplebus.c
--- simplebus.c 8 Dec 2016 16:24:51 -0000 1.1
+++ simplebus.c 5 Jan 2019 08:27:04 -0000
@@ -121,6 +121,26 @@ simplebus_submatch(struct device *self,
return 0;
}
+int
+simplebus_print(void *aux, const char *pnp)
+{
+ struct fdt_attach_args *fa = aux;
+ char name[32];
+
+ if (!pnp)
+ return (QUIET);
+
+ if (OF_getprop(fa->fa_node, "name", name, sizeof(name)) > 0) {
+ name[sizeof(name) - 1] = 0;
+ printf("\"%s\"", name);
+ } else
+ printf("node %u", fa->fa_node);
+
+ printf(" at %s", pnp);
+
+ return (UNCONF);
+}
+
/*
* Look for a driver that wants to be attached to this node.
*/
@@ -186,7 +206,8 @@ simplebus_attach_node(struct device *sel
OF_getpropintarray(node, "interrupts", fa.fa_intr, len);
}
- config_found_sm(self, &fa, NULL, simplebus_submatch);
+ config_found_sm(self, &fa, sc->sc_early ? NULL : simplebus_print,
+ simplebus_submatch);
free(fa.fa_reg, M_DEVBUF, fa.fa_nreg * sizeof(struct fdt_reg));
free(fa.fa_intr, M_DEVBUF, fa.fa_nintr * sizeof(uint32_t));
