* [MODERATED] Change conditional to always?
@ 2018-08-10 20:28 Konrad Rzeszutek Wilk
2018-08-10 20:29 ` [MODERATED] " Konrad Rzeszutek Wilk
2018-08-10 22:04 ` Andi Kleen
0 siblings, 2 replies; 4+ messages in thread
From: Konrad Rzeszutek Wilk @ 2018-08-10 20:28 UTC (permalink / raw)
To: speck
Hi,
With the combination of spectrev1 to bring various host and guest date
in the L1 and then L1TF slurping is the choice of using conditional
good?
That is I could see the guest trapping in the emulation code - and
the condional would stay at false - which means we don't do IA32_PRED_CMD
flush and end up providing a nice way for the guest to slurp up data.
Thoughts?
^ permalink raw reply [flat|nested] 4+ messages in thread
* [MODERATED] Re: Change conditional to always?
2018-08-10 20:28 [MODERATED] Change conditional to always? Konrad Rzeszutek Wilk
@ 2018-08-10 20:29 ` Konrad Rzeszutek Wilk
2018-08-13 13:34 ` Paolo Bonzini
2018-08-10 22:04 ` Andi Kleen
1 sibling, 1 reply; 4+ messages in thread
From: Konrad Rzeszutek Wilk @ 2018-08-10 20:29 UTC (permalink / raw)
To: speck
[once more but this time encrypting this]
Hi,
With the combination of spectrev1 to bring various host and guest date
in the L1 and then L1TF slurping is the choice of using conditional
good?
That is I could see the guest trapping in the emulation code - and
the condional would stay at false - which means we don't do IA32_PRED_CMD
flush and end up providing a nice way for the guest to slurp up data.
Thoughts?
^ permalink raw reply [flat|nested] 4+ messages in thread
* [MODERATED] Re: Change conditional to always?
2018-08-10 20:28 [MODERATED] Change conditional to always? Konrad Rzeszutek Wilk
2018-08-10 20:29 ` [MODERATED] " Konrad Rzeszutek Wilk
@ 2018-08-10 22:04 ` Andi Kleen
1 sibling, 0 replies; 4+ messages in thread
From: Andi Kleen @ 2018-08-10 22:04 UTC (permalink / raw)
To: speck
On Fri, Aug 10, 2018 at 04:28:31PM -0400, speck for Konrad Rzeszutek Wilk wrote:
> Hi,
>
> With the combination of spectrev1 to bring various host and guest date
> in the L1 and then L1TF slurping is the choice of using conditional
> good?
>
> That is I could see the guest trapping in the emulation code - and
> the condional would stay at false - which means we don't do IA32_PRED_CMD
> flush and end up providing a nice way for the guest to slurp up data.
>
> Thoughts?
I would wait for exploits for this. Combinations of such fragile exploits
tends to be very hard and it may very well not be practical.
There are ongoing efforts to reduce the gadgets too.
Far too early to change any defaults I think.
-Andi
^ permalink raw reply [flat|nested] 4+ messages in thread
* [MODERATED] Re: Change conditional to always?
2018-08-10 20:29 ` [MODERATED] " Konrad Rzeszutek Wilk
@ 2018-08-13 13:34 ` Paolo Bonzini
0 siblings, 0 replies; 4+ messages in thread
From: Paolo Bonzini @ 2018-08-13 13:34 UTC (permalink / raw)
To: speck
[-- Attachment #1: Type: text/plain, Size: 534 bytes --]
On 10/08/2018 22:29, speck for Konrad Rzeszutek Wilk wrote:
> With the combination of spectrev1 to bring various host and guest date
> in the L1 and then L1TF slurping is the choice of using conditional
> good?
>
> That is I could see the guest trapping in the emulation code - and
> the condional would stay at false - which means we don't do IA32_PRED_CMD
> flush and end up providing a nice way for the guest to slurp up data.
Emulation should cause the cache flush to happen (see
x86_emulate_instruction).
Paolo
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2018-08-13 13:34 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-08-10 20:28 [MODERATED] Change conditional to always? Konrad Rzeszutek Wilk
2018-08-10 20:29 ` [MODERATED] " Konrad Rzeszutek Wilk
2018-08-13 13:34 ` Paolo Bonzini
2018-08-10 22:04 ` Andi Kleen
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.