3bd323a1da
When probing the keyboard controller to enable A20, if we get FF back (which is *possible* as a valid status word, but is extremely unlikely) then bail after much fewer iterations than we otherwise would, and abort the attempt to access the KBC. This hopefully should make it work a lot better for embedded platforms which don't have a KBC and where the BIOS doesn't implement INT 15h AX=2401h (and doesn't boot with A20 already enabled.) If this works, it will be the one remaining use of CONFIG_X86_ELAN as anything other than a processor type optimization option. Signed-off-by: H. Peter Anvin <hpa@linux.intel.com> |
||
---|---|---|
.. | ||
compressed | ||
tools | ||
.gitignore | ||
a20.c | ||
apm.c | ||
bitops.h | ||
boot.h | ||
cmdline.c | ||
code16gcc.h | ||
copy.S | ||
cpu.c | ||
cpucheck.c | ||
edd.c | ||
header.S | ||
install.sh | ||
main.c | ||
Makefile | ||
mca.c | ||
memory.c | ||
mkcpustr.c | ||
mtools.conf.in | ||
pm.c | ||
pmjump.S | ||
printf.c | ||
setup.ld | ||
string.c | ||
tty.c | ||
version.c | ||
vesa.h | ||
video-bios.c | ||
video-mode.c | ||
video-vesa.c | ||
video-vga.c | ||
video.c | ||
video.h | ||
voyager.c |