grosen@amadeus.ucsb.edu (Mark D. Grosen) (06/05/90)
Well, still no sign of the 3.0 SDK. Meanwhile, I am curious to know how memory mapped peripherals are accessed in 3.0's 386 protected mode. Judging from the behavior of my old programs that access a hunk of shared memory at 0xd0000, making a far pointer in 3.0 is a no-no. Is there something similar to Unix's "physical I/O" functions that let you specify an absolute address? Anybody with a beta 3.0 SDK care to comment? Mark Mark D. Grosen ARPA: grosen@amadeus.ucsb.edu Signal Processing Lab / Communications Research Lab ECE Dept. University of California Santa Barbara, CA 93106
brianf@umd5.umd.edu (Brian Farmer) (06/05/90)
In article <5660@hub.ucsb.edu> grosen@amadeus.ucsb.edu (Mark D. Grosen) writes: > >Well, still no sign of the 3.0 SDK. Meanwhile, I am curious to know how >memory mapped peripherals are accessed in 3.0's 386 protected mode. >Judging from the behavior of my old programs that access a hunk of >shared memory at 0xd0000, making a far pointer in 3.0 is a no-no. Is there >something similar to Unix's "physical I/O" functions that let you specify >an absolute address? > A few global selectors have been defined to directly access the memory. They are __A000H, __B000H, __B800H, __C000H, __D000H, __E000H, __F000H.