* i386 subarchitectures: boot page table flags
@ 2006-05-31 1:57 H. Peter Anvin
2006-05-31 2:21 ` James Bottomley
0 siblings, 1 reply; 3+ messages in thread
From: H. Peter Anvin @ 2006-05-31 1:57 UTC (permalink / raw)
To: Linux Kernel Mailing List, James.Bottomley, Martin J. Bligh
Hi all,
Does any of the i386 subarchitectures actually care about the Accessed and Dirty bits in
the bootup pagetables (the ones that start at pg0, used before the mm is initialized?) If
not, I'd like to speed up booting by setting those bits at initialization time.
-hpa
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: i386 subarchitectures: boot page table flags
2006-05-31 1:57 i386 subarchitectures: boot page table flags H. Peter Anvin
@ 2006-05-31 2:21 ` James Bottomley
2006-05-31 2:49 ` H. Peter Anvin
0 siblings, 1 reply; 3+ messages in thread
From: James Bottomley @ 2006-05-31 2:21 UTC (permalink / raw)
To: H. Peter Anvin; +Cc: Linux Kernel Mailing List, Martin J. Bligh
On Tue, 2006-05-30 at 18:57 -0700, H. Peter Anvin wrote:
> Does any of the i386 subarchitectures actually care about the Accessed and Dirty bits in
> the bootup pagetables (the ones that start at pg0, used before the mm is initialized?) If
> not, I'd like to speed up booting by setting those bits at initialization time.
Depends what you mean by "care". I do hijack pg0 in
voyager_memory_detect() to access the clickmap for ascertaining the
memory layout, but I don't use the accessed or dirty bits.
James
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: i386 subarchitectures: boot page table flags
2006-05-31 2:21 ` James Bottomley
@ 2006-05-31 2:49 ` H. Peter Anvin
0 siblings, 0 replies; 3+ messages in thread
From: H. Peter Anvin @ 2006-05-31 2:49 UTC (permalink / raw)
To: James Bottomley; +Cc: Linux Kernel Mailing List, Martin J. Bligh
James Bottomley wrote:
> On Tue, 2006-05-30 at 18:57 -0700, H. Peter Anvin wrote:
>> Does any of the i386 subarchitectures actually care about the Accessed and Dirty bits in
>> the bootup pagetables (the ones that start at pg0, used before the mm is initialized?) If
>> not, I'd like to speed up booting by setting those bits at initialization time.
>
> Depends what you mean by "care". I do hijack pg0 in
> voyager_memory_detect() to access the clickmap for ascertaining the
> memory layout, but I don't use the accessed or dirty bits.
Okay...
Leaving the A and D bits clear means the CPU has to trap to microcode to
set those bits before it is allowed to load the entry into the TLB.
Furthermore, and this is the real killer on some CPUs, it is not allowed
to load those entries speculatively.
As far as I can tell, this is completely pointless for pg0.
(And no, the Voyager code wouldn't be affected.)
-hpa
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2006-05-31 2:50 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2006-05-31 1:57 i386 subarchitectures: boot page table flags H. Peter Anvin
2006-05-31 2:21 ` James Bottomley
2006-05-31 2:49 ` H. Peter Anvin
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).