[ecdsa-sca] Enable ECDSA-256 secret shares for capture by bilgiday · Pull Request #109 · lowRISC/ot-sca

bilgiday · 2022-12-20T03:20:42Z

Enable sending of all secret shares (k0, k1, d0, d1) from the Python code (capture.py).

Signed-off-by: Bilgiday Yuce bilgiday@opentitan.org

bilgiday · 2022-12-20T03:23:19Z

./capture.py --cfg-file capture_ecdsa_cw310.yaml capture ecdsa-simple --num-traces 2 --plot-traces 1 can be used to run the experiments locally.

The C counterpart is here: opentitan-pr16899.

cw/capture.py

vogelpi

Thanks @bilgiday for working on this. I've left a few comments and also have some questions.

Please move all the binaries related to ECDSA captures (bitstreams + target binaries) to Git LFS. We should have done this right from the beginning but it seems I wasn't fast enough in reviewing your last/initial ECDSA PR. For this, you'll have to add entries for these binaries in .gitattributes in the top directory of the repo.

cw/capture_ecdsa_cw310.yaml

cw/capture.py

m-temp · 2022-12-23T09:55:54Z

Please move all the binaries related to ECDSA captures (bitstreams + target binaries) to Git LFS. We should have done this right from the beginning but it seems I wasn't fast enough in reviewing your last/initial ECDSA PR. For this, you'll have to add entries for these binaries in .gitattributes in the top directory of the repo.

@vogelpi should we add wildcards to the .gitattributes? Is there any reason why we might not want to have a .bin or .bit file in the lfs but rather in the normal git? I'm asking because I also messed this up a couple of times in the past....

vogelpi · 2023-01-17T17:23:43Z

.gitattributes

+cw/objs/ecc_serial_fpga_cw310.bin filter=lfs diff=lfs merge=lfs -text
+cw/objs/ecc384_serial_fpga_cw310.bin filter=lfs diff=lfs merge=lfs -text


I don't think just declaring these files in the .gitattributes here is enough. As you can see further below in the diff, Git still thinks you've pushed new binary files. You probably have to first delete the files using git rm, then add them to Git LFS and last but not least squash the commits into one. For reference this is how the diff should look like: https://github.com/lowRISC/ot-sca/pull/4/files

Thanks for the detailed steps. I updated the files by following these instructions.

vogelpi · 2023-01-17T17:26:58Z

cw/capture_ecdsa_cw310.yaml

+  # https://github.com/lowRISC/ot-sca/issues/106#issuecomment-1359530023
+  # https://rtfm.newae.com/Capture/ChipWhisperer-Husky/#streaming-mode
+  # A diagram showing the clocking structure for optentitan CW310:
+  # http://drawings/1oVABuYbeXq5lwneQMZ49D5MBFfBJoSxz6V1A17j4XDo


This link doesn't work. What you can do is include it in this PR as .png and then link to that. On the other hand, you don't really refer to this in the file or explain what is meant. So I think it's fine to skip this link in the source file. But it would be worth to add it to the PR for discussion.

I removed the link from the source code. Added it to PR (doc/img) for discussion and future reference.

vogelpi · 2023-01-17T17:27:24Z

cw/capture_ecdsa_cw310.yaml

-  # Target operating frequency
-  pll_frequency: 22000000
+  # Target operating frequency: 
+  # for ecdsa-simple: 25-50MHz (alignment issues above 66MHZ)


Nit: 66MHZ -> 66MHz

cw/capture_ecdsa_cw310.yaml

vogelpi

Thanks for implementing the suggested changes, this looks mostly good now.

What I would like to happen before we merge this:

@vrozic was planning to also try out the PR on the FPGA
There still seems to be an issue with Git LFS, I left a more detailed comment.
It would be good if also @m-temp could have a look at the PR because he will also build on top of this work.

I am not hitting "request changes" because I don't want to block progress (OoO tomorrow) but please make sure to the points above are addressed before merging this.

cw/capture_ecdsa_cw310.yaml

vogelpi · 2023-01-17T17:36:57Z

Please move all the binaries related to ECDSA captures (bitstreams + target binaries) to Git LFS. We should have done this right from the beginning but it seems I wasn't fast enough in reviewing your last/initial ECDSA PR. For this, you'll have to add entries for these binaries in .gitattributes in the top directory of the repo.

@vogelpi should we add wildcards to the .gitattributes? Is there any reason why we might not want to have a .bin or .bit file in the lfs but rather in the normal git? I'm asking because I also messed this up a couple of times in the past....

Good point @m-temp , I don't see a reason for not putting such files into LFS. I just wasn't aware this was possible. Would you mind giving this a try as a follow-up for this PR?

vogelpi · 2023-01-17T17:38:20Z

Ah, and you'll need to rebase this PR on top of master for CI to pass. There was a Python-related issue I had to fix.

bilgiday · 2023-01-17T17:47:48Z

Thanks for implementing the suggested changes, this looks mostly good now.

What I would like to happen before we merge this:

@vrozic was planning to also try out the PR on the FPGA

There still seems to be an issue with Git LFS, I left a more detailed comment.

It would be good if also @m-temp could have a look at the PR because he will also build on top of this work.

I am not hitting "request changes" because I don't want to block progress (OoO tomorrow) but please make sure to the points above are addressed before merging this.

Thanks for the comments @vogelpi. I will fix the Git LFS issue :)

bilgiday · 2023-01-17T17:48:43Z

Ah, and you'll need to rebase this PR on top of master for CI to pass. There was a Python-related issue I had to fix.

Thanks for fixing it. I will rebase before pushing the next version.

m-temp · 2023-01-18T08:40:13Z

Good point @m-temp , I don't see a reason for not putting such files into LFS. I just wasn't aware this was possible. Would you mind giving this a try as a follow-up for this PR?

Will do so in #111

m-temp

LGTM. I've added a few comments that could improve understanding the code in the future. Would be nice to have to address these.

m-temp · 2023-01-18T09:21:01Z

cw/capture.py

    # Currently, we need to reprogram the firmware before every sign operation
    # TODO: Investigate the C code (ecc384_serial) to check if we can run multiple
    #       signing operations without reloading the firmware.
-    reset_firmware = True
+    reset_firmware = False


The comment says we need to reprogram, but reset_firmware = False. Could you explain what's the actual TODO? Is this only a TODO for 384 and set to false for 256?

Thanks Michael! It seems that I forgot to update these comments.

At the beginning of the project, there was an issue in the C code, and we needed to reflash the firmware for each signing operation. Now, that problem has been solved and we don't need to reflash the firmware again and again.

I just wanted to keep this reflashing option there in case we need it in the future. I updated the comments accordingly to reflect the current state.

m-temp · 2023-01-18T09:22:56Z

cw/capture.py

+
        # Create an arrey to keep the traces
        waves = np.array([])
        # Currently, we have to reload the firmware to be able to send a


Although this is not port of the current PR could you add a comment, when this branch/condition needs to be taken? See also comment above, where reset_firmware is set

edit: Hmm there's something wrong with the associated lines. This part belongs tho the reset_firmeware part.

Thank you, I updated the reset_firmware related comments.

cw/capture.py

vrozic

This works well on my setup for 256b version.
Capture rates: ~2.5s per trace (simple mode) and 1.5s per trace (stream mode). I assume these are the expected values.

Thanks for the great work @bilgiday

- Enables sending of all secret shares (k0, k1, d0, d1) from the Python code (capture.py). - Enables the trace collection without needing to reflash the firmware at every iteration. Signed-off-by: Bilgiday Yuce <bilgiday@opentitan.org>

bilgiday · 2023-01-18T14:08:21Z

This works well on my setup for 256b version. Capture rates: ~2.5s per trace (simple mode) and 1.5s per trace (stream mode). I assume these are the expected values.

Thanks for the great work @bilgiday

Thanks for testing it @vrozic, the capture rates seems right to me.

m-temp

Thanks @bilgiday for your work and addressing/clarifying the issues!
LGTM

bilgiday · 2023-01-19T11:03:15Z

It seems everyone is happy with the current status of the PR. I am merging this PR not to block the others.

On the opentitan side, lowRISC/opentitan#16899, I am still waiting for one last nod to go.

bilgiday mentioned this pull request Dec 20, 2022

[otbn/sca] Add new sca features for otbn_ecdsa_256 lowRISC/opentitan#16899

Merged

bilgiday commented Dec 20, 2022

View reviewed changes

cw/capture.py Show resolved Hide resolved

bilgiday commented Dec 20, 2022

View reviewed changes

cw/capture.py Outdated Show resolved Hide resolved

bilgiday mentioned this pull request Dec 20, 2022

[ecdsa] Extend setup to allow for random and fixed data #106

Open

bilgiday force-pushed the update-ecdsa-256 branch 4 times, most recently from b964204 to c681878 Compare December 21, 2022 12:11

vogelpi reviewed Dec 22, 2022

View reviewed changes

bilgiday force-pushed the update-ecdsa-256 branch from c681878 to 0b7b862 Compare January 17, 2023 14:52

bilgiday marked this pull request as ready for review January 17, 2023 15:00

bilgiday requested review from abdullahvarici, alphan, andreaskurth and vrozic as code owners January 17, 2023 15:00

bilgiday requested review from vogelpi and removed request for abdullahvarici, alphan, andreaskurth and vrozic January 17, 2023 15:00

vogelpi reviewed Jan 17, 2023

View reviewed changes

cw/capture_ecdsa_cw310.yaml Show resolved Hide resolved

vogelpi requested review from m-temp and vrozic January 17, 2023 17:37

bilgiday force-pushed the update-ecdsa-256 branch 2 times, most recently from 2a9a026 to 8fec21a Compare January 17, 2023 19:53

m-temp approved these changes Jan 18, 2023

View reviewed changes

m-temp mentioned this pull request Jan 18, 2023

Add wildcards to .gitattributes to include all bin files in git lfs #111

Closed

vrozic approved these changes Jan 18, 2023

View reviewed changes

bilgiday force-pushed the update-ecdsa-256 branch from 8fec21a to b9c113d Compare January 18, 2023 14:02

bilgiday requested a review from m-temp January 18, 2023 14:28

m-temp approved these changes Jan 18, 2023

View reviewed changes

bilgiday merged commit ef01775 into lowRISC:master Jan 19, 2023

		cw/objs/ecc_serial_fpga_cw310.bin filter=lfs diff=lfs merge=lfs -text
		cw/objs/ecc384_serial_fpga_cw310.bin filter=lfs diff=lfs merge=lfs -text

Comments

Conversation

bilgiday commented Dec 20, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

bilgiday commented Dec 20, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Uh oh!

vogelpi left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

m-temp commented Dec 23, 2022

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

vogelpi left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

vogelpi commented Jan 17, 2023

Uh oh!

vogelpi commented Jan 17, 2023

Uh oh!

bilgiday commented Jan 17, 2023

Uh oh!

bilgiday commented Jan 17, 2023

Uh oh!

m-temp commented Jan 18, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

m-temp left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

m-temp Jan 18, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

vrozic left a comment

Choose a reason for hiding this comment

Uh oh!

bilgiday commented Jan 18, 2023

Uh oh!

m-temp left a comment

Choose a reason for hiding this comment

Uh oh!

bilgiday commented Jan 19, 2023

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

bilgiday commented Dec 20, 2022 •

edited

Loading

bilgiday commented Dec 20, 2022 •

edited

Loading

m-temp commented Jan 18, 2023 •

edited

Loading

m-temp Jan 18, 2023 •

edited

Loading