Skip to contentSkip to conversation

question on i386 very early memory detection cleanup patch

DISCUSSION Jan Beulich 1 participant 1 message Sep 2, 2004

lore
Share & tools
lore
#0 lore raw

If you refer to the 4G/4G split patch - this isn't part of mainline, and if it needs it would seem it should do this adjustment, not an unrelated patch. Jan

Quoted 1 line
Quoted 1 line
Quoted 1 line

"H. Peter Anvin" <[email protected]> 01.09.04 18:20:01 >>>

Quoted 3 lines · Jan Beulich

Is there a particular reason why this patch changes the alignment of cpu_gdt_table to be page rather than cache line aligned? This is particulary strange to me because the alignment is guaranteed only

for

Quoted 2 lines

the boot processor, but not for any of the APs; for the latter ones there isn't even a string guarantee that the table would be cache

line

Quoted 1 line

aligned (which it really should be); the weak guarantee only is

through

Quoted 1 line

an appearant assumption of GDT_ENTRIES being a sufficiently large

power

Quoted 3 lines

of two.

Thanks, Jan

The 4+4 GB patch apparently needs it.