On 2010-03-09 12:28 PM, Jonas Gorski wrote: > While trying to add ar8316 support to the ar8216 driver, I noticed > several small issues in the driver itself. Since I lack an ar8216 > to test, somebody please verify I did not break it. > > This patch makes the following changes: > > * Create defines for some magic values/masks. > > * Change vlan_id to u16, to allow VIDs > 255. > > * Add a range check to set_pvid as it isn't a VID, but the index > in the vlan table. > > * Set the max VID to 4094, since 4095 is a reserved value and > should not be used. > > * In mangle_rx replace the provided VID with the VID of the table > entry of the port, not the index of the table (this one needs > review). > > * In hw_apply, remove a redundant emptiness check (was already > checked several lines above). > > * In no vlan mode do not set the ingress mode to secure, as there > are no vlan table entries, but to use the port's destination > masks. Otherwise the switch won't forward anything. > > * In read_status tell that the phy is up (taken from the rtl8306 > driver). > > The change in mange_rx needs to be verified. Perhaps I missed > some assumption somewhere else, but it looks like it should be that > way. This change has only any effect in situations where the vlan's > VIDs aren't the same as their position in the table. Applied in r20083, with one exception. I left out this change: -#define AR8216_GCTRL_MTU BITS(0, 10) +#define AR8216_GCTRL_MTU BITS(0, 14) as it does not match my hardware info for AR8216. If the bitfield is different on AR8316, then we should add a separate definition for that.
Thanks, - Felix _______________________________________________ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
