VSC7511 datasheet(204/324 Pages) MICROSEMI

204 / 324 page

VCore-Ie System and CPU Interfaces

VMDS-10488 VSC7511 Datasheet Revision 4.2

189

Tokens can be added to the token counter by writing the FDMA_CH_INJ_TOKEN_CNT[ich] register.

Tokens can also be added automatically (with a fixed interval) by using the dedicated token tick counter.

Setting FDMA_CH_INJ_TOKEN_TICK_RLD[ich] to a value n (different from 0) will cause one token to be

added to that injection channel every n x 200 ns.

4.7.7

Manual Mode

The decision making logic of the FDMA extraction path and/or injection paths can be disabled to give

control of the FDMA’s extraction and/or injection buffers directly to any master attached to the Shared

Bus Architecture. When operating in manual mode DCB structure, counters and most of the interrupts do

not apply.

Manual extraction and injection use hard-coded channel numbers.

•

Manual extraction mode uses FDMA channel 1 (port 12 extraction direction).

•

Manual injection mode uses FDMA channel 9 (port 12 injection direction).

To enable manual extraction, set MANUAL_CFG.XTR_ENA. The FDMA must not be enabled for FDMA

extraction when manual extraction mode is active. To enable manual injection, set

MANUAL_CFG.INJ_ENA and FDMA_CH_CFG[9].CH_INJ_CRP = 1. The FDMA must not be enabled for

FDMA injection when manual injection mode is active. Extraction and injection paths are independent.

For example, FDMA can control extraction at the same time as doing manual injection by means of the

CPU.

4.7.7.1

Manual Extraction

Extraction is done by reading one or more blocks of data from the extraction memory area. A block of

data is defined by reading one or more data words followed by reading of the extraction status word. The

extraction memory area is 16 kilobytes and is implemented as a replicated register region

MANUAL_XTR. The highest register in the replication (0xFFF) maps to the extraction status word. The

status word is updated by the extraction logic and shows the number of frame bytes read, SOF and EOF

indications, and PRUNED/ABORT indications.

Note During frame extraction, the CPU does not know the frame length. This means that the CPU must

check for EOF regularly during frame extraction. When reading a data block from the device, the CPU

can burst read from the memory area in such a way that the extraction status word is read as the last

word of the burst.

Figure 69 •

Extraction Status Word Encoding

The extraction logic updates the extraction status word while the CPU is reading out data for a block.

Prior to starting a new data block, the CPU can write the two least significant bits of the BLOCKO field.

The BLOCKO value is stored in the extraction logic and takes effect when the new data block is started.

Reading the status field always returns the BLOCKO value that applies to the current data block. Unless

[31:20]

[15:0]

BLOCKL: The amount of frame data (bytes) that

has been read in this block. The CPU might have

read more data than this, only valid frame data

(including Extraction Header) is counted here.

SOF: Start of frame, set if this block contains

start of a frame.

EOF: End of frame, set if this block contains end

of a frame.

ABORT: Set if this frame has been aborted by

the Queuing System. Write ‘1’ to discard the

remainder of the active frame.

PD: Pruned indication, set if the frame has been

pruned to less than original length.

BLOCKO: The amount of dummy bytes (non-frame

data) read before start of frame data.

BLOCKL

BLOCKO

VSC7511 Datasheet(PDF) 204 Page - Microsemi Corporation

VSC7511 Datasheet(HTML) 204 Page - Microsemi Corporation

Html Pages

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 ...More

Datasheet Download

Link URL