Messages in this thread | | | Date | Tue, 1 Jun 1999 16:01:25 +0200 | From | Jamie Lokier <> | Subject | Re: ptrace question ... |
| |
Ralf Baechle wrote: > Is there any reason why we don't attempt to handle requests for addresses > with MAP_NR(page) >= MAP_NR(high_memory) correctly? I've got a pending > bug report for the MIPS port where somebody complains that the respective > code for MIPS prevents accessing the hardware directly from GDB.
Yes: if you remove the test, accessing the hardware still doesn't work!
> *(unsigned long *) (page + (addr & ~PAGE_MASK)) = data;
This line only works for pages where MAP_NR(page) < MAP_NR(high_memory). To access other hardware addresses from kernel space requires a call to ioremap() first.
Well ok, _some_ addresses will still work if the phase of your moon is right, but not reliably.
However I don't know why the test isn't simply `page < high_memory'.
-- Jamie
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |