Dear Komuro,
On Sun, Nov 30, 2008 at 05:54:02PM +0900, Komuro wrote:
> Multi-function card can't get the correct ConfigBase.
>
> unfortunately, pccard_read_tuple always get the ConfigBase for 1st device.
Is this still an issue?
> The reason of this problem is pccard_read_tuple set
> tuple.Attributes = TUPLE_RETURN_COMMON.
>
> Any idea to fix this?
Well, we could add something like that:
diff --git a/drivers/pcmcia/cistpl.c b/drivers/pcmcia/cistpl.c
index 4a110b7..0141063 100644
--- a/drivers/pcmcia/cistpl.c
+++ b/drivers/pcmcia/cistpl.c
@@ -1463,7 +1463,10 @@ int pccard_read_tuple(struct pcmcia_socket *s, unsigned
int function, cisdata_t
return -ENOMEM;
}
tuple.DesiredTuple = code;
- tuple.Attributes = TUPLE_RETURN_COMMON;
+ if (function == 0)
+ tuple.Attributes = TUPLE_RETURN_COMMON;
+ else:
+ tuple.Attributes = TUPLE_RETURN_LINK;
ret = pccard_get_first_tuple(s, function, &tuple);
if (ret != 0)
goto done;
though I would be worried whether this breaks other accesses to
pccard_read_tuple with function > 0, e.g. when looking up the MANF_ID?
Best,
Dominik
_______________________________________________
Linux PCMCIA reimplementation list
http://lists.infradead.org/mailman/listinfo/linux-pcmcia
