
Philips Semiconductors
PNX15xx Series
Volume 1 of 1
Chapter 29: Endian Mode
PNX15XX_SER_3
Koninklijke Philips Electronics N.V. 2006. All rights reserved.
Product data sheet
Rev. 3 — 17 March 2006
29-15
The 32-bit PCI bus uses byte address conventions identical to the DTL interface and
MTL Memory Bus Interface: Refer to
Table 11.
The DCS Network uses byte address conventions given in
Table 8.
The MTL Memory Bus uses the byte address conventions given in
Table 9.With the above byte address conventions on the three sides of the bridge and the
byte address invariance rule for bridges, the swap modes can be derived. Since the
convention on the PCI bus closely matches those on the MTL Bus.
7.
Detailed Example
This section describes all steps involved in how a big-endian mode external CPU
(e.g., a Power Macintosh), paints an RGB-565 pixel format frame buffer in the
PNX15xx Series SDRAM and how this is displayed on the QVCP. This example
illustrates the following:
The Power Macintosh PCI bridge and its address invariance rule based swapper
The BIG-endian PCI pixel transfer
How data arrives correct in SDRAM in native RGB565 pixel format
How the QVCP takes it and displays it
How the TM32 CPU core sees the data
The Power Macintosh was the rst platform that successfully demonstrated big-
endian operations across the PCI bus. Details of how this works can be found in the
Apple document “Designing PCI Cards and Drivers for Power MacIntosh Computers.”
Suppose that the big-endian CPU in the Power Macintosh uses a 32-bit store
operation to create two RGB565 pixels. Pixel 1, the left-most pixel, has (byte) address
“A” and pixel 2 has address “A+2.” Since these two pixels are transferred in a single
32-bit word, “A” is a multiple of 4.
The intermediate stages that the data goes through can be found in
Figure 7Table 11: 32 Bit PCI Interface Byte Address
PCI-AD[31:24]
PCI-AD[23:16]
PCI-AD[15:8]
PCI-AD[7:0]
4n+3
4n+2
4n+1
4n+0