[OTBN] vertical attack initial work #115

m-temp · 2023-01-25T09:48:23Z

This is a PR to bring together the works of Jade, Bilgiday and me

Co-authored-by: Jade Philipoom [email protected]
Signed-off-by: Michael Tempelmeier [email protected]

Things to do:

Add TVLA support
- stimuli
- analysis script
~~[ ] Restructure code to support different code snippets~~
- ~~[ ] add select function to simple serial (python + c)~~
- ~~[ ] put "golden model" in a function to support different calculations~~
Prepare and merge PR in opentitan repo ==> add binaries with specific hash id

cw/capture.py

vogelpi

Hi @m-temp ,
thanks for the PR!
I had a first brief look now. I think we should try to keep this capture setup as generic as possible because we're going to use it for other code snippets than keygen as well. Also, since the captures are short we don't need some of the infrastructure used in for ECDSA captures. The whole capture and analysis is much more similar to KMAC/AES. Maybe it's best to discuss this tomorrow, I've set up a little sync meeting.

cw/capture.py

cw/util/device.py

m-temp · 2023-01-30T17:07:38Z

I've added a first fixed-vs-random tvla version.
Major points are:

finding good parameter sets (out of memory vs enough data)
add a batch version
restructure the code

m-temp · 2023-02-10T11:40:22Z

To expedite cooperation, I mark this PR as ready for review

It adds:

basic fvsr support for otbn vertical
ctrl+c to store capture results
config outputs to tvla plots

A brief test confirmed that sha3, kmac, aes are not broken.

What is still missing:

doc for otbn vertical

abdullahvarici

Hello! Thanks for the PR. Sorry, I don't have any hardware to try it out for now. I think, in general it looks good. I've added a couple of feedbacks.

cw/capture.py

cw/tvla.py

m-temp · 2023-02-27T10:58:19Z

Thanks @abdullahvarici for your review and sry for all those typos in the code. I will address them.

vogelpi

Overall this looks good and we should move forward. Many thanks for all the work you've put into this @m-temp and @wettermo .

In addition to the OTBN vertical support, this PR contains a few very nice additions that should be factored out into separate PRs. These additions are uncontroversial and are unrelated to OTBN vertical support.

cw/capture.py

vogelpi · 2023-02-27T13:38:20Z

cw/capture.py

+ ot.scope.adc.bits_per_sample = 12
+ ot.scope.adc.samples = capture_cfg["num_samples"]
+ else:
+ # TODO: Add cw-lite support


We won't add CW-Lite support. There is actually an issue to remove CW-Lite support ;-)

@vogelpi, should we then remove the TODO with the cleaning PR proposed in this comment above? probably, keeping the else statement and the message is still OK.

Hm, I am not sure I understand your comment. I think we don't want to have a TODO to Add CW-Lite support, because we're never going to add it but to error out if CW-Lite is detected is the best thing to do here.

cw/tvla_cfg_otbn.yaml

vogelpi · 2023-02-27T13:45:26Z

cw/tvla.py

+ traces[i_trace] = project.waves[i_trace +
+ trace_start][sample_start:sample_start +
+ num_samples]
+ else: # FIXME: The wave slicing is only tested for otbn


This is a very nice feature. Thanks for adding this!

It will also be very useful for other targets. It's not expected to test this for all different targets (there are simply too many options and we have other priorities). However, I would appreciate if you changed the code such that there is simply an assertion to check that when sample_start and sample_end are set, also otbn is selected and provide a comment where this check is done. This way we have to change much less code later on when enabling it e.g. for AES or KMAC.

vogelpi · 2023-02-27T13:47:03Z

cw/tvla.py

+ xaxs = range(sample_start, sample_start + num_samples)
 fig, axs = plt.subplots(3, sharex=True)
- axs[0].plot(single_trace, "k")
+ axs[0].set_title(
+ "TVLA of " + cfg["project_file"] + '\n' + "No. of traces: " + str(num_traces))
+ axs[0].plot(xaxs, single_trace, "k")
+ axs[0].set_ylabel("trace")
 for i_order in range(num_orders):
- axs[1 + i_order].plot(ttest_trace[i_order, 0, 0], "k")
- axs[1 + i_order].plot(c * threshold, "r")
- axs[1 + i_order].plot(-threshold * c, "r")
+ axs[1 + i_order].plot(xaxs, ttest_trace[i_order, 0, 0], "k")
+ axs[1 + i_order].plot(xaxs, c * threshold, "r")
+ axs[1 + i_order].plot(xaxs, -threshold * c, "r")


This commit needs to be merged with the one adding support for sample_start and num_samples. Also these two things should go into a separate PR.

vogelpi · 2023-02-27T13:49:05Z

cw/tvla.py

+ axs[0].set_title(
+ "TVLA of " + cfg["project_file"] + '\n' + "No. of traces: " + str(num_traces))


I think just this line here actually belongs into this very commit.

Hmm let's put it also in the other PR, where we change the plotting

vogelpi · 2023-02-27T13:51:53Z

Also updating the bitstreams and existing binaries can be factored out. Remember the goal is always to have as small PRs and commit as possible. This speeds up the review and allows us to progress faster as a team.

cw/capture.py

m-temp · 2023-02-28T09:04:02Z

Thx all for the reviews. I changed it accordingly and will file a couple of PRs for the remaining features.

vogelpi

Thanks for addressing the comments. This looks good!

Two minor nits:

There is still a TODO for the CW-Lite, but we won't add support for this. Having the script to error out for something else than Husky still desirable though.
The PR also adds a new bitstream in the first commit. Is this really needed? We wanted to add the new bistream with the different URND behavior in a separate PR I believe.

vogelpi · 2023-02-28T09:25:53Z

cw/capture.py

+ ot.scope.adc.bits_per_sample = 12
+ ot.scope.adc.samples = capture_cfg["num_samples"]
+ else:
+ # TODO: Add cw-lite support


Hm, I am not sure I understand your comment. I think we don't want to have a TODO to Add CW-Lite support, because we're never going to add it but to error out if CW-Lite is detected is the best thing to do here.

m-temp · 2023-02-28T09:42:56Z

1. There is still a TODO for the CW-Lite, but we won't add support for this. Having the script to error out for something else than Husky still desirable though.

Sry. I missed that. This was a leftover from Jade. I'll remove the comment, but keep the if-else-branch

2. The PR also adds a new bitstream in the first commit. Is this really needed? We wanted to add the new bistream with the different URND behavior in a separate PR I believe.

Ah okay, I thought this was only related to the second bins, but you are right, I delete the old bins and file a separate PR for the new ones.

Co-authored-by: Jade Philipoom <[email protected]> Signed-off-by: Michael Tempelmeier <[email protected]>

Signed-off-by: Michael Tempelmeier <[email protected]>

This is a initial work to add fixed-vs.-random tvla tests for otbn. Signed-off-by: Michael Tempelmeier <[email protected]>

Signed-off-by: Michael Tempelmeier <[email protected]>

m-temp commented Jan 25, 2023

View reviewed changes

cw/capture.py Outdated Show resolved Hide resolved

m-temp force-pushed the otbn-vertical branch 5 times, most recently from bed7225 to df6da38 Compare January 25, 2023 14:34

vogelpi reviewed Jan 26, 2023

View reviewed changes

cw/capture.py Outdated Show resolved Hide resolved

cw/capture.py Outdated Show resolved Hide resolved

cw/util/device.py Outdated Show resolved Hide resolved

cw/util/device.py Outdated Show resolved Hide resolved

m-temp force-pushed the otbn-vertical branch from 063a3ca to 6baf71d Compare January 27, 2023 15:52

m-temp mentioned this pull request Jan 30, 2023

ktp.next() for keys #116

Open

m-temp force-pushed the otbn-vertical branch 2 times, most recently from 85bb10b to 652ce81 Compare January 30, 2023 13:53

m-temp force-pushed the otbn-vertical branch 4 times, most recently from f48d61f to 2e51a0a Compare February 3, 2023 15:13

wettermo force-pushed the otbn-vertical branch 3 times, most recently from 9f40120 to 63ec147 Compare February 8, 2023 10:33

m-temp force-pushed the otbn-vertical branch from 63ec147 to 049f5f4 Compare February 10, 2023 11:34

m-temp marked this pull request as ready for review February 10, 2023 11:40

m-temp requested review from abdullahvarici, alphan, andreaskurth and vrozic as code owners February 10, 2023 11:40

m-temp changed the title ~~WIP [OTBN] vertical attack initial work~~ [OTBN] vertical attack initial work Feb 10, 2023

m-temp requested a review from vogelpi February 10, 2023 13:20

wettermo mentioned this pull request Feb 13, 2023

[crypto/otbn/sca] ecc256 device software for SCA lowRISC/opentitan#17273

Closed

m-temp force-pushed the otbn-vertical branch from da4df2c to 0f50aca Compare February 14, 2023 10:14

wettermo force-pushed the otbn-vertical branch 2 times, most recently from 7d3ea70 to f80f968 Compare February 20, 2023 10:13

m-temp mentioned this pull request Feb 24, 2023

[tvla] Usability and performance improvements, code cleanup #72

Open

22 tasks

abdullahvarici reviewed Feb 26, 2023

View reviewed changes

wettermo force-pushed the otbn-vertical branch from 26dff15 to 9cb2d3c Compare February 27, 2023 11:08

m-temp force-pushed the otbn-vertical branch from 9cb2d3c to e813226 Compare February 27, 2023 13:10

wettermo force-pushed the otbn-vertical branch from e813226 to cfcd269 Compare February 27, 2023 13:37

vogelpi reviewed Feb 27, 2023

View reviewed changes

vrozic reviewed Feb 27, 2023

View reviewed changes

cw/capture.py Outdated Show resolved Hide resolved

m-temp force-pushed the otbn-vertical branch 2 times, most recently from 94dcfd1 to a0195f5 Compare February 28, 2023 09:00

vogelpi approved these changes Feb 28, 2023

View reviewed changes

m-temp force-pushed the otbn-vertical branch from a0195f5 to b5469cb Compare February 28, 2023 09:49

wettermo mentioned this pull request Feb 28, 2023

[otbn/tvla] Add bitstream + device binary for OTBN vertical TVLA #123

Merged

m-temp and others added 4 commits February 28, 2023 12:56

[otbn] add vertical attack - inital work

8743df7

Co-authored-by: Jade Philipoom <[email protected]> Signed-off-by: Michael Tempelmeier <[email protected]>

[sca capture.py] workaround for lowRISC#116

0f9f8fc

Signed-off-by: Michael Tempelmeier <[email protected]>

[tvla/otbn] add fixed-vs-random tvla for otbn

957bb29

This is a initial work to add fixed-vs.-random tvla tests for otbn. Signed-off-by: Michael Tempelmeier <[email protected]>

[otbn/sca] add config option to disable mask

ab9d22a

Signed-off-by: Michael Tempelmeier <[email protected]>

m-temp force-pushed the otbn-vertical branch from b5469cb to ab9d22a Compare February 28, 2023 11:58

m-temp merged commit 4ede900 into lowRISC:master Feb 28, 2023

m-temp deleted the otbn-vertical branch February 28, 2023 12:02

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[OTBN] vertical attack initial work #115

[OTBN] vertical attack initial work #115

m-temp commented Jan 25, 2023 •

edited

Loading

vogelpi left a comment

m-temp commented Jan 30, 2023 •

edited

Loading

m-temp commented Feb 10, 2023 •

edited

Loading

abdullahvarici left a comment

m-temp commented Feb 27, 2023

vogelpi left a comment

vogelpi Feb 27, 2023

bilgiday Feb 27, 2023

vogelpi Feb 28, 2023

vogelpi Feb 27, 2023

m-temp Feb 28, 2023

vogelpi Feb 27, 2023

m-temp Feb 28, 2023

vogelpi Feb 27, 2023

m-temp Feb 28, 2023

vogelpi commented Feb 27, 2023

m-temp commented Feb 28, 2023

vogelpi left a comment

vogelpi Feb 28, 2023

m-temp commented Feb 28, 2023

		axs[0].set_title(
		"TVLA of " + cfg["project_file"] + '\n' + "No. of traces: " + str(num_traces))

[OTBN] vertical attack initial work #115

[OTBN] vertical attack initial work #115

Conversation

m-temp commented Jan 25, 2023 • edited Loading

vogelpi left a comment

Choose a reason for hiding this comment

m-temp commented Jan 30, 2023 • edited Loading

m-temp commented Feb 10, 2023 • edited Loading

abdullahvarici left a comment

Choose a reason for hiding this comment

m-temp commented Feb 27, 2023

vogelpi left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

vogelpi commented Feb 27, 2023

m-temp commented Feb 28, 2023

vogelpi left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

m-temp commented Feb 28, 2023

m-temp commented Jan 25, 2023 •

edited

Loading

m-temp commented Jan 30, 2023 •

edited

Loading

m-temp commented Feb 10, 2023 •

edited

Loading