System Address Map
3.9
Memory Shadowing
Any block of memory that can be designated as read-only or write-only can be
“shadowed” into GMCH DRAM memory. Typically this is done to allow ROM code to
execute more rapidly out of main DRAM. ROM is used as a read-only during the copy
process while DRAM at the same time is designated write-only. After copying, the
DRAM is designated read-only so that ROM is shadowed. Processor bus transactions
are routed accordingly.
3.10
I/O Address Space
The GMCH does not support the existence of any other I/O devices beside itself on the
processor bus. The GMCH generates either DMI Interface or PCI Express bus cycles for
all processor I/O accesses that it does not claim. Within the host bridge, the GMCH
contains two internal registers in the processor I/O space, Configuration Address
Register (CONFIG_ADDRESS) and the Configuration Data Register (CONFIG_DATA).
These locations are used to implement configuration space access mechanism.
The processor allows 64K+3 bytes to be addressed within the I/O space. The GMCH
propagates the processor I/O address without any translation on to the destination
bus and therefore provides addressability for 64K+3 byte locations. Note that the
upper 3 locations can be accessed only during I/O address wrap-around when
processor bus HA16# address signal is asserted. HA16# is asserted on the processor
bus when an I/O access is made to 4 bytes from address 0FFFDh, 0FFFEh, or 0FFFFh.
HA16# is also asserted when an I/O access is made to 2 bytes from address 0FFFFh.
A set of I/O accesses (other than ones used for configuration space access) are
consumed by the internal graphics device if it is enabled. The mechanisms for internal
graphics IO decode and the associated control is explained later.
The I/O accesses (other than ones used for configuration space access) are forwarded
normally to the DMI Interface bus unless they fall within the PCI Express I/O address
range as defined by the mechanisms explained below. I/O writes are NOT posted.
Memory writes to the ICH or PCI Express are posted. The PCICMD1 register can
disable the routing of I/O cycles to the PCI Express.
The GMCH responds to I/O cycles initiated on PCI Express or DMI with an UR status.
Upstream I/O cycles and configuration cycles should never occur. If one does occur,
the request will route as a read to Memory address 000C_0000h so a completion is
naturally generated (whether the original request was a read or write). The
transaction will complete with an UR completion status.
For Intel microprocessors, I/O reads that lie within 8-byte boundaries but cross 4-byte
boundaries are issued from the processor as 1 transaction. The GMCH will break this
into 2 separate transactions. I/O writes that lie within 8-byte boundaries but cross 4-
byte boundaries are assumed to be split into 2 transactions by the processor.
62
Datasheet
INTEL [ INTEL ]
