
VIA Technologies, Inc.
Preliminary VT6508 Datasheet
-18-
(3)
This is the bit mask of the destination ports corresponding to the forwarding table slot.
The bits 49 and 48, corresponding to CPU port and MII port respectively, should be
zero in VT6508.
(4)
bits 57..56: Age Count
(5)
If this field is equal to zero, it means this slot is an empty one; otherwise, it is valid.
When the auto-aging is enabled (by default), every valid slot’ age count will be
decreased by one per 100 seconds.
(6)
bits 58: Static Flag
(7)
This flag is used to program a static MAC slot, that will not be overwritten by dynamic
learning mechanism, with specified forwarding port mask for special purpose.
(8)
bits 59: Over-Write Flag
During learning, if both slots are used, the 3rd address will replace one of the slots
according to current value of the concatenated over-write order pointer and their static
bits regardless of ages of the two slots.
Let the concatenated over-write pointer ow[1:0] = (over-write pointer of slot 1,
over-write pointer of slot 0).
Let static[1:0] = {static bit of slot 1, static bit of slot 0}.
The possible mapping of ow[1:0] and static[1:0] are as follows:
static[1:0]
ow[1:0]
(0,0)
(0,1)
(1,1)
(1,0)
(0,0)
0
1
x
0
(0,1)
1
1
x
0
(1,1)
0
1
x
0
(1,0)
1
1
x
0
The internal operation rules for the over-write pointer are as follows:
(1) When a slot is over-written, its over-write pointer is toggled.
(2) When one of the slots is a static entry, the static entry must always be placed in slot
0, and slot 1 is always the entry over-written. Over-write pointer has no effect what
so ever here.
(3) When both slots are static, no entry can be written to either slot by the learning
process.
(4) When update an entry, value of the over-write pointer must not be changed unless
indicated otherwise, i.e. over-writing an existing slot.
Another data structures in SRAM memory layout is the packet buffer pool, located at the
upper address space. Every packet buffer is 1.5K bytes with a corresponding embedded link
node in the tailing 8 bytes. The embedded link nodes are used to construct output queues, one
queue for each output port. Its structure is shown in Figure 5, composed of:
(1) bits 9..0: Next Packet Buffer ID
(2) bits 19..10: Destination Ports Bit Mask