Re: [RFC][PATCH v2 00/21] x86/pti: Defer CR3 switch to C code

From: Alexandre Chartre
Date: Wed Nov 18 2020 - 14:38:37 EST



On 11/18/20 12:29 PM, Borislav Petkov wrote:
On Wed, Nov 18, 2020 at 08:41:42AM +0100, Alexandre Chartre wrote:
Well, it looks like I wrongfully assume that KPTI was a well known performance
overhead since it was introduced (because it adds extra page-table switches),
but you are right I should be presenting my own numbers.

Here's one recipe, courtesy of Mel:

https://github.com/gormanm/mmtests


Thanks for the detailed information, I have run the test and I see the same difference
as with the tools/perf and libMICRO I already sent: there's a 150% difference for
getpid() with and without pti.

alex.

-----

# ../../compare-kernels.sh --baseline test-nopti --compare test-pti

poundsyscall
test test
nopti pti
Min 2 1.99 ( 0.00%) 5.08 (-155.28%)
Min 4 1.02 ( 0.00%) 2.60 (-154.90%)
Min 6 0.94 ( 0.00%) 2.07 (-120.21%)
Min 8 0.81 ( 0.00%) 1.60 ( -97.53%)
Min 12 0.85 ( 0.00%) 1.65 ( -94.12%)
Min 18 0.82 ( 0.00%) 1.61 ( -96.34%)
Min 24 0.81 ( 0.00%) 1.60 ( -97.53%)
Min 30 0.81 ( 0.00%) 1.60 ( -97.53%)
Min 32 0.81 ( 0.00%) 1.60 ( -97.53%)
Amean 2 2.02 ( 0.00%) 5.10 *-151.83%*
Amean 4 1.03 ( 0.00%) 2.61 *-151.98%*
Amean 6 0.96 ( 0.00%) 2.07 *-116.74%*
Amean 8 0.82 ( 0.00%) 1.60 * -96.56%*
Amean 12 0.87 ( 0.00%) 1.67 * -91.73%*
Amean 18 0.82 ( 0.00%) 1.63 * -97.94%*
Amean 24 0.81 ( 0.00%) 1.60 * -97.41%*
Amean 30 0.82 ( 0.00%) 1.60 * -96.93%*
Amean 32 0.82 ( 0.00%) 1.60 * -96.56%*
Stddev 2 0.02 ( 0.00%) 0.02 ( 33.78%)
Stddev 4 0.01 ( 0.00%) 0.01 ( 7.18%)
Stddev 6 0.01 ( 0.00%) 0.00 ( 68.77%)
Stddev 8 0.01 ( 0.00%) 0.01 ( 10.56%)
Stddev 12 0.01 ( 0.00%) 0.02 ( -12.69%)
Stddev 18 0.01 ( 0.00%) 0.01 (-107.25%)
Stddev 24 0.00 ( 0.00%) 0.00 ( -14.56%)
Stddev 30 0.01 ( 0.00%) 0.01 ( 0.00%)
Stddev 32 0.01 ( 0.00%) 0.00 ( 20.00%)
CoeffVar 2 1.17 ( 0.00%) 0.31 ( 73.70%)
CoeffVar 4 0.82 ( 0.00%) 0.30 ( 63.16%)
CoeffVar 6 1.41 ( 0.00%) 0.20 ( 85.59%)
CoeffVar 8 0.87 ( 0.00%) 0.39 ( 54.50%)
CoeffVar 12 1.66 ( 0.00%) 0.98 ( 41.23%)
CoeffVar 18 0.85 ( 0.00%) 0.89 ( -4.71%)
CoeffVar 24 0.52 ( 0.00%) 0.30 ( 41.97%)
CoeffVar 30 0.65 ( 0.00%) 0.33 ( 49.22%)
CoeffVar 32 0.65 ( 0.00%) 0.26 ( 59.30%)
Max 2 2.04 ( 0.00%) 5.13 (-151.47%)
Max 4 1.04 ( 0.00%) 2.62 (-151.92%)
Max 6 0.98 ( 0.00%) 2.08 (-112.24%)
Max 8 0.83 ( 0.00%) 1.62 ( -95.18%)
Max 12 0.89 ( 0.00%) 1.70 ( -91.01%)
Max 18 0.84 ( 0.00%) 1.66 ( -97.62%)
Max 24 0.82 ( 0.00%) 1.61 ( -96.34%)
Max 30 0.82 ( 0.00%) 1.61 ( -96.34%)
Max 32 0.82 ( 0.00%) 1.61 ( -96.34%)
BAmean-50 2 2.01 ( 0.00%) 5.09 (-153.39%)
BAmean-50 4 1.03 ( 0.00%) 2.60 (-152.62%)
BAmean-50 6 0.95 ( 0.00%) 2.07 (-118.82%)
BAmean-50 8 0.81 ( 0.00%) 1.60 ( -97.53%)
BAmean-50 12 0.86 ( 0.00%) 1.66 ( -92.79%)
BAmean-50 18 0.82 ( 0.00%) 1.62 ( -97.56%)
BAmean-50 24 0.81 ( 0.00%) 1.60 ( -97.53%)
BAmean-50 30 0.81 ( 0.00%) 1.60 ( -97.53%)
BAmean-50 32 0.81 ( 0.00%) 1.60 ( -97.53%)
BAmean-95 2 2.02 ( 0.00%) 5.09 (-151.87%)
BAmean-95 4 1.03 ( 0.00%) 2.61 (-151.99%)
BAmean-95 6 0.95 ( 0.00%) 2.07 (-117.25%)
BAmean-95 8 0.81 ( 0.00%) 1.60 ( -96.72%)
BAmean-95 12 0.87 ( 0.00%) 1.67 ( -91.82%)
BAmean-95 18 0.82 ( 0.00%) 1.63 ( -97.97%)
BAmean-95 24 0.81 ( 0.00%) 1.60 ( -97.53%)
BAmean-95 30 0.81 ( 0.00%) 1.60 ( -97.00%)
BAmean-95 32 0.81 ( 0.00%) 1.60 ( -96.59%)
BAmean-99 2 2.02 ( 0.00%) 5.09 (-151.87%)
BAmean-99 4 1.03 ( 0.00%) 2.61 (-151.99%)
BAmean-99 6 0.95 ( 0.00%) 2.07 (-117.25%)
BAmean-99 8 0.81 ( 0.00%) 1.60 ( -96.72%)
BAmean-99 12 0.87 ( 0.00%) 1.67 ( -91.82%)
BAmean-99 18 0.82 ( 0.00%) 1.63 ( -97.97%)
BAmean-99 24 0.81 ( 0.00%) 1.60 ( -97.53%)
BAmean-99 30 0.81 ( 0.00%) 1.60 ( -97.00%)
BAmean-99 32 0.81 ( 0.00%) 1.60 ( -96.59%)

test test
nopti pti
Duration User 150.13 432.03
Duration System 372.10 657.69
Duration Elapsed 94.17 199.27