zl50409 Zarlink Semiconductor, zl50409 Datasheet - Page 46

• Manufacturer Part Number: zl50409
• Description: Managed 9-port 10/100m Ethernet Switch
• Manufacturer: Zarlink Semiconductor

In addition, there is a shared pool, which can store any type of frame. The frame engine allocates the frames first in
the three priority sections. When the priority section is full or the packet has priority 1 or 0, the frame is allocated in
the shared pool. Once the shared pool is full the frames are allocated in the section reserved for the source port.
The following registers define the size of each section of the Frame data Buffer:
See Buffer Allocation application note, ZLAN-47, for more information.
7.6.1
As already discussed, the WRED mechanism may drop frames on output queue status. In addition to these
reasons for dropping, we also drop frames when global buffer space becomes scarce. The function of buffer
management is to make sure that such dropping causes as little blocking as possible. If a received frame is
dispatched to the best effort queue, the buffer management will check on the overall buffer situation plus the output
queue status to decide the frame drop condition. If the source port has not enough buffer for it, the frame will be
dropped. If the output queue reach the UCC (unicast congest control) and the shared buffer has run out, the frame
will be dropped by b%. If the output queue reach the UCC and the source port reservation is lower than the buffer
low threshold, the frame will be dropped. All the dropping functions are disabled if the source port is flow control
capable.
7.7
Because frame loss is unacceptable for some applications, the ZL50409 provides a flow control option. When flow
control is enabled, scarcity of source port buffer space may trigger a flow control signal; this signal tells a source
port sending a packet to this switch, to temporarily hold off.
While flow control offers the clear benefit of no packet loss, it also introduces a problem for quality of service. When
a source port receives an Ethernet flow control signal, all microflows originating at that port, well-behaved or not,
are halted. A single packet destined for a congested output can block other packets destined for un-congested
outputs. The resulting head-of-line blocking phenomenon means that quality of service cannot be assured with high
confidence when flow control is enabled.
-
-
-
-
-
-
-
Per Source Port
Flow Control Basics
PR100_N - Port Reservation for RMAC Ports
PR100_CPU - Port Reservation for CPU Port
PRM - Port Reservation for MMAC Port
SFCB - Share FCB Size
C1RS - Class 1 Reserve Size (priority 2 & 3)
C2RS - Class 2 Reserve Size (priority 4 & 5)
C3RS - Class 3 Reserve Size (priority 6 & 7)
Reservation
Reservation
Dropping When Buffers Are Scarce
Per Class
R
R
p0
pri1
Figure 11 - Buffer Partition Scheme
R
p1
R
Zarlink Semiconductor Inc.
pri2
R
ZL50409
p2
46
R
R
pri3
Temporary reservation
p3
R
p4
R
p5
Shared Pool S
R
p6
R
p7
(CPU)
R
p8
Data Sheet
R
p9