pkg/mbedtls: initial pkg import to use entropy module by PeterKietzmann · Pull Request #15671 · RIOT-OS/RIOT

PeterKietzmann · 2020-12-17T18:07:35Z

Contribution description

This PR adds mbed TLS as a package to be used for entropy generation, and it is the only module that has been tested in RIOT so far. However, this PR is an enabler for additional mbed TLS features in RIOT.

The PR includes a simple RIOT API to access the entropy generated by mbed TLS. Entropy sources that feed the module are currently ADC Noise (#14324) and the HWRNG. Further information can be found in the package documentation.

Testing procedure

Carefully read and understand the high level documentation (and of cause all other code & doc :) )
Build doxygen (make doc).
Execute the provided test on boards and native.

Issues/PRs references

Successor of sys/entropy_source: add new module for entropy sources #14324
Affected by dist/tools/ci or tests: EXCLUDE not assigned correctly? #15660

miri64 · 2020-12-17T18:47:04Z

pkg/mbedtls/include/riot_config.h

@@ -0,0 +1,3344 @@
+/**
+ * \file config.h


I might be mistaking, but this files seems similar to what I did with lwipopts.h in lwIP. Does this file has to be exactly this format or can we adapt it to be more RIOT-y? The fact, that this file is 96% comments, makes me think the latter.

I also don't think that you have to be that verbose on documentation, as the documentation (hopefully) lies with mbedtls (which you could just link here).

The file is a copy-paste from mbed TLS. For sure I could write a similar RIOT(-y) file, but for further integration of the package, I think it simplifies the development process to have a (almost) copy available. With that, a developer (i) sees immediately all available configuration options (&doc) and (ii) can easily look at diffs to the original file during version bump or the like...

Mh from the other point of view, I might be interested as a user, which configuration options of mbedtls are actually used. Scrolling though this 3300 line file to filter out the 142 lines that are actually important seems cumbersome, especially given, that you want to exclude this file from out doxygen build. Furthermore, we might want to make options configurable from the app (like we do e.g. with lwIP...), which seems a bit more complicated when you need to comment out / comment in defines.

I also don't see the maintenance advantage you propose. If you want to see which options are available in mbedtls: go to the mbedtls documentation. If there is a version bump and config options change this will usually lead to failing compilation, giving you exactly the names of the macros that are new (and thus missing).

From #15671 (comment)

Lets give it some more time and if no one opposes, we'll go with your proposal.

Yes, please @RIOT-OS/maintainers, can we have more thoughts and comments on this?

(I haven't checked the code, just followed the discussion)
If we can avoid the 3300 lines of configuration code, that would be better. If it's added, there's no reason to skip it from doxygen, or any other checks.
IIUC, the goal of this file is to provide the configuration of the package. Why not link it to the module mechanism of RIOT to define flags (like MBEDTLS_SHA256_ALT) or via a Kconfig file or both (depending on the cases) ?

for the module version:

#if IS_USED(MODULE_MBEDTLS_CIPHER_MODE_CBC) #define MBEDTLS_CIPHER_MODE_CBC #endif

for the Kconfig version (assuming there's a corresponding Kconfig boolean option):

#if IS_ACTIVE(CONFIG_MBEDTLS_ECP_DP_SECP192R1_ENABLED) #define MBEDTLS_ECP_DP_SECP192R1_ENABLED #endif

PeterKietzmann · 2020-12-17T18:52:22Z

To answer @miri64's question in this comment:

And then there is the question why you want to exclude it in the first place? pkg/mbedtls/include/riot_config.h seems very RIOT-specific to me, so I guess it is not an external import. If it is: since you are in a pkg, can't you import it via the pkg mechanism as well?

It is an mbed TLS config file which I slightly adjusted for RIOT, by en-/disabling some options. By enabling additional library features, further changes might be required. To be still mostly "copy-paste-able" I've tried to keep changes at minimum.

Technically I could import & patch it. Practically, as a user, I would like the have the file directly available instead of a patch.

miri64 · 2020-12-17T18:54:12Z

pkg/mbedtls/contrib/entropy_mbedtls_riot.c

@@ -0,0 +1,73 @@
+/*


Shouldn't this be its on module so in the future people can use mbedtls without the entropy source (if they so desire)?

So pkg/mbedtls/contrib/mbedtls_entropy_source/.... This way you also might not need to define pseudo-modules.

Again: see lwIP how to implement actual modules within the contrib directory of a package.

Good point. I have tried this in my last commit, mind to have another look?

miri64 · 2020-12-17T18:57:22Z

It is an mbed TLS config file which I slightly adjusted for RIOT, by en-/disabling some options. By enabling additional library features, further changes might be required. To be still mostly "copy-paste-able" I've tried to keep changes at minimum.

Why do you want to have it "copy-paste-able"? Copy-paste from where? Is configuration so complicated within mbedtls that you need to copy files around to keep track of your configuration? I highly doubt any of this, given the popularity and porting status of the library.

Technically I could import & patch it. Practically, as a user, I would like the have the file directly available instead of a patch.

PeterKietzmann · 2020-12-17T19:02:36Z

Is configuration so complicated within mbedtls

Well, I guess it strongly depends on the point of view. A look at the amount of configuration options in that file gives you a first impression. Personally (proposing the library import here) I don't have a problem providing a specific configuration file. Still, I could think of other developers seeing this differently. Lets give it some more time and if no one opposes, we'll go with your proposal.

aabadie

Went through this PR and have some comment related to the build system and other minor things.
I also think that the riot_config.h should be rewritten and not excluded from the doxygen build (since it might contain important configuration information) and the CI checks.

aabadie · 2020-12-18T13:05:28Z

pkg/mbedtls/Makefile

@@ -0,0 +1,10 @@
+PKG_NAME=mbedtls
+PKG_URL=https://github.com/ARMmbed/mbedtls.git
+PKG_VERSION=2a1d9332d55d1270084232e42df08fdb08129f1b


If this commit hash correspond to a tagged version, please add it in a comment at the end of the line

aabadie · 2020-12-18T13:07:37Z

pkg/mbedtls/Makefile

+
+all:
+	"$(MAKE)" -C $(PKGDIRBASE)/mbedtls/library/ \
+			  -f $(RIOTPKG)/mbedtls/Makefile.$(PKG_NAME)


Makefile.mbedtls only contains a minimal setup, you could just use Makefile.base directly and pass the MODULE variable in the command line:

Suggested change

-f $(RIOTPKG)/mbedtls/Makefile.$(PKG_NAME)

-f $(RIOTBASE)/Makefile.base MODULE=$(PKG_NAME)

And Makefile.mbedtls can be removed.

aabadie · 2020-12-18T13:08:31Z

pkg/mbedtls/Makefile.dep

+  # include sources if hardware is available
+  FEATURES_OPTIONAL += periph_hwrng
+  ifneq (,$(filter periph_hwrng,$(FEATURES_USED)))
+	  DEFAULT_MODULE += mbedtls_entropy_source_hwrng


indentation is off, it should be 2 spaces inside the ifneq block. Maybe it's a tab ?
Same comment applies in the rest of this file.

aabadie · 2020-12-18T13:09:13Z

pkg/mbedtls/Makefile.dep

+  # include sources if hardware is available
+  FEATURES_OPTIONAL += periph_hwrng
+  ifneq (,$(filter periph_hwrng,$(FEATURES_USED)))
+	  DEFAULT_MODULE += mbedtls_entropy_source_hwrng
+  endif
+
+  FEATURES_OPTIONAL += periph_adc
+  ifneq (,$(filter periph_adc,$(FEATURES_USED)))
+	  DEFAULT_MODULE += mbedtls_entropy_source_adc
+  endif


What happens if the hardware provides both ? What is the entropy source selected in this case ?

I think those are not mutually exclusive, @PeterKietzmann ?

Then both are used (if not manually disabled). That is actually the intent of the entropy module, to accumulate multiple entropy sources.

aabadie · 2020-12-18T13:11:26Z

pkg/mbedtls/Makefile.dep

+		  CFLAGS += -DCONFIG_ENTROPY_SOURCE_ADC_COND=0
+		  CFLAGS += -DCONFIG_ENTROPY_SOURCE_ADC_HEALTH_TEST=1


Those 2 constants seem to be unused.

Nope, they configure the ADC noise source which was added in #14324.

Shouldn't this be better placed in the Makefile.include?

aabadie · 2020-12-18T13:22:09Z

pkg/mbedtls/doc.txt

+ *
+ * ```
+ * FEATURES_PROVIDED += periph_hwrng
+ * FEATURES_PROVIDED += entropy_source_adc_noise


Suggested change

* FEATURES_PROVIDED += entropy_source_adc_noise

* FEATURES_PROVIDED += periph_adc

?

aabadie · 2020-12-18T13:25:46Z

pkg/mbedtls/contrib/entropy/entropy_sources_mbedtls_riot.c

+ */
+
+/**
+ * @ingroup     pkg_mbedtls


Suggested change

* @ingroup pkg_mbedtls

* @ingroup pkg_mbedtls_entropy

?

Yes, probably a good idea to spend an own doxygen group now that we moved the entropy support to its own module.

aabadie · 2020-12-18T13:26:01Z

pkg/mbedtls/contrib/entropy/entropy_mbedtls_riot.c

+ */
+
+/**
+ * @ingroup     pkg_mbedtls


Suggested change

* @ingroup pkg_mbedtls

* @ingroup pkg_mbedtls_entropy

?

aabadie · 2020-12-18T13:27:55Z

pkg/mbedtls/include/sha256_alt.h

+extern "C" {
+#endif
+
+#if defined (MBEDTLS_SHA256_ALT)


Suggested change

#if defined (MBEDTLS_SHA256_ALT)

#if IS_ACTIVE(MBEDTLS_SHA256_ALT)

aabadie · 2020-12-18T13:29:17Z

tests/pkg_mbedtls/Makefile

+ifneq (,$(filter mbedtls_entropy_source_adc,$(USEMODULE)))
+	ifndef CONFIG_KCONFIG_USEMODULE_ENTROPY_SOURCE_ADC_NOISE
+		CFLAGS += -DCONFIG_ENTROPY_SOURCE_ADC_HMIN=65536
+	endif
+endif


Why don't you put this in the package Makefile.include ?

To avoid mis-configuration of this entropy source, we set an invalid default value. Hence, a user must evaluate and configure the module according to the hardware in use and the deployment. This line prevents failures during testing.

PeterKietzmann · 2020-12-18T14:07:37Z

@aabadie many thanks for your review. I have left some answers inline and will work on the other comments. Regarding the configuration file there seems to be a favor for the RIOT-style option. Lets give it some time anyway, and I will provide an alternative during the next couple of days if no one objects.

PeterKietzmann · 2020-12-18T15:18:13Z

@aabadie I have addressed your comments (despite the config header file as described above)

fjmolinas · 2021-03-12T07:14:18Z

ping @aabadie

PeterKietzmann · 2021-07-08T10:03:02Z

Added Kconfig options, enabled config header for doxygen again, tested on native and nrf52840dk, cleaned up minor stuff...

@aabadie, mind to give it an other look to remove the change request?

PeterKietzmann · 2021-07-09T13:02:01Z

Btw: all tests now pass

fjmolinas · 2021-11-18T09:52:04Z

@leandrolanzieri is this one ready to go?

leandrolanzieri · 2021-11-18T10:55:45Z

pkg/mbedtls/Makefile.dep

+		  CFLAGS += -DCONFIG_ENTROPY_SOURCE_ADC_COND=0
+		  CFLAGS += -DCONFIG_ENTROPY_SOURCE_ADC_HEALTH_TEST=1


Shouldn't this be better placed in the Makefile.include?

pkg/mbedtls/contrib/entropy/entropy_sources_mbedtls_riot.c

leandrolanzieri · 2021-11-18T11:05:04Z

pkg/mbedtls/contrib/entropy/entropy_sources_mbedtls_riot.c

+#if defined(MODULE_MBEDTLS_ENTROPY_SOURCE_ADC)
+    entropy_source_adc_init();
+#endif


Would this init be needed for additional sources? Would it make sense to add some interface to initialize the selected sources?

It depends on the source, but seems likely that most sources have some kind of an init function. Technically, we could add .init function to entropy_source_mbedtls_riot_t but I would like to delay that, since other sources (e.g. PUF) might have very different requirements from ADC sampling or HWRNG readout.

pkg/mbedtls/contrib/sha256_alt.c

pkg/mbedtls/Kconfig

pkg/mbedtls/include/entropy_mbedtls_riot.h

pkg/mbedtls/include/entropy_sources_mbedtls_riot.h

pkg/mbedtls/contrib/entropy/entropy_sources_mbedtls_riot.c

tests/pkg_mbedtls/Makefile

tests/pkg_mbedtls/Makefile.ci

pkg/mbedtls/Makefile

pkg/mbedtls/contrib/entropy/entropy_mbedtls_riot.c

aabadie · 2021-11-23T08:39:44Z

pkg/mbedtls/contrib/entropy/entropy_sources_mbedtls_riot.c

+    for (unsigned i = 1; i < NUM_ENTROPY_FUNCS; i++) {
+        ret = mbedtls_entropy_add_source(ctx, entropy_funcs[i].func, NULL,
+                                1, entropy_funcs[i].strong);
+    }


I think there's something wrong with the ret usage. For example, if something goes wrong with the first entropy source (ret != 0) but works with the last, the whole function return code will be 0. I don't think this is expected.

Maybe it's better to check the return value in the for loop and return with an error if any of source cannot be added ?

The loop adds entropy sources to a list which will be polled during entropy gathering. The error case occurs when a maximum number of sources have been added. Hence, the case of failure followed by success in not true. Nevertheless, we can save some redundant calls by returning on the first error.

PeterKietzmann · 2021-11-23T16:26:15Z

@aabadie , @leandrolanzieri your latest comments should be addressed in this commit aef6196

benpicco · 2021-11-23T19:52:35Z

pkg/mbedtls/Makefile

@@ -0,0 +1,10 @@
+PKG_NAME=mbedtls
+PKG_URL=https://github.com/ARMmbed/mbedtls.git
+PKG_VERSION=v2.26.0


Why not use v3.0.0 here? (or at least v2.27.0)

v2.26.0 was up to date when I opened this PR. I worked and tested the most with this version. Sticking with it seemed better to me than "blindly" bumping the version.

pkg/mbedtls/contrib/entropy/entropy_sources_mbedtls_riot.c

tests/pkg_mbedtls/Makefile

pkg/mbedtls/include/entropy_sources_mbedtls_riot.h

benpicco

Please squash!

fjmolinas · 2022-01-24T13:37:26Z

@PeterKietzmann the test fails on samr-21xpro (and a couple of other BOARDs I tested locally, some report a "failed", others hang), is that expected? The PR only mentioned testing on native, is it supposed to work on hardware?

PeterKietzmann · 2022-01-24T14:11:19Z

is it supposed to work on hardware

Yes it is. I ran the test on different platforms, back then. Cannot remember any high-ups. Just wondering: who has tested this PR before merging? Anyway, I will look into this as soon as I can!

PeterKietzmann · 2022-01-24T14:12:14Z

some report a "failed", others hang),

@fjmolinas can you name these boards so I have a starting point?

fjmolinas · 2022-01-25T08:21:33Z

some report a "failed", others hang),

@fjmolinas can you name these boards so I have a starting point?

Nightlies are failing on samr21-xpro, I got a hang on nucleo-l152re. I'll run the test on all the BOARDs I have to give a complete list, will do on this PR and master.

fjmolinas · 2022-01-25T08:51:51Z

I ran the tests on the PR and on master.

Results (ignore failed compilations, there is some kind of raise condition with pkg downloads...):

master

### nucleo-g071rb/failuresummary

Failures during compilation:
- [tests/pkg_mbedtls](nucleo-g071rb/tests/pkg_mbedtls/compilation.failed)

### nucleo-f207zg/failuresummary


### nucleo-f334r8/failuresummary


### nucleo-g474re/failuresummary

Failures during compilation:
- [tests/pkg_mbedtls](nucleo-g474re/tests/pkg_mbedtls/compilation.failed)

### nrf52dk/failuresummary


### frdm-k64f/failuresummary


### nucleo-l152re/failuresummary

Failures during test:
- [tests/pkg_mbedtls](nucleo-l152re/tests/pkg_mbedtls/test.failed)

### arduino-uno/failuresummary


### nucleo-f767zi/failuresummary

Failures during compilation:
- [tests/pkg_mbedtls](nucleo-f767zi/tests/pkg_mbedtls/compilation.failed)

### samr21-xpro/failuresummary

Failures during test:
- [tests/pkg_mbedtls](samr21-xpro/tests/pkg_mbedtls/test.failed)

### p-nucleo-wb55/failuresummary


### nucleo-l496zg/failuresummary


### nucleo-l452re/failuresummary


### frdm-kw41z/failuresummary

Failures during test:
- [tests/pkg_mbedtls](frdm-kw41z/tests/pkg_mbedtls/test.failed)

### nucleo-l433rc/failuresummary

Failures during compilation:
- [tests/pkg_mbedtls](nucleo-l433rc/tests/pkg_mbedtls/compilation.failed)

### i-nucleo-lrwan1/failuresummary


### openmote-b/failuresummary


### nucleo-wl55jc/failuresummary


### nucleo-l4r5zi/failuresummary


### z1/failuresummary


### nucleo-f091rc/failuresummary

Failures during test:
- [tests/pkg_mbedtls](nucleo-f091rc/tests/pkg_mbedtls/test.failed)

### arduino-zero/failuresummary

Failures during test:
- [tests/pkg_mbedtls](arduino-zero/tests/pkg_mbedtls/test.failed)

### nucleo-l552ze-q/failuresummary


### nucleo-f303k8/failuresummary


### nucleo-f103rb/failuresummary


### b-l072z-lrwan1/failuresummary


### atmega256rfr2-xpro/failuresummary

Failures during test:
- [tests/pkg_mbedtls](atmega256rfr2-xpro/tests/pkg_mbedtls/test.failed)

### nucleo-f303re/failuresummary


### slstk3402a/failuresummary


### nucleo-f030r8/failuresummary

Failures during test:
- [tests/pkg_mbedtls](nucleo-f030r8/tests/pkg_mbedtls/test.failed)

### nucleo-l073rz/failuresummary

pr


### nucleo-g071rb/failuresummary

Failures during test:
- [tests/pkg_mbedtls](nucleo-g071rb/tests/pkg_mbedtls/test.failed)

### nucleo-f207zg/failuresummary

Failures during compilation:
- [tests/pkg_mbedtls](nucleo-f207zg/tests/pkg_mbedtls/compilation.failed)

### nucleo-f334r8/failuresummary


### nucleo-g474re/failuresummary

Failures during compilation:
- [tests/pkg_mbedtls](nucleo-g474re/tests/pkg_mbedtls/compilation.failed)

### nrf52dk/failuresummary


### frdm-k64f/failuresummary


### nucleo-l152re/failuresummary

Failures during test:
- [tests/pkg_mbedtls](nucleo-l152re/tests/pkg_mbedtls/test.failed)

### arduino-uno/failuresummary


### nucleo-f767zi/failuresummary

Failures during compilation:
- [tests/pkg_mbedtls](nucleo-f767zi/tests/pkg_mbedtls/compilation.failed)

### samr21-xpro/failuresummary

Failures during test:
- [tests/pkg_mbedtls](samr21-xpro/tests/pkg_mbedtls/test.failed)

### p-nucleo-wb55/failuresummary


### nucleo-l496zg/failuresummary

Failures during compilation:
- [tests/pkg_mbedtls](nucleo-l496zg/tests/pkg_mbedtls/compilation.failed)

### nucleo-l452re/failuresummary


### frdm-kw41z/failuresummary

Failures during test:
- [tests/pkg_mbedtls](frdm-kw41z/tests/pkg_mbedtls/test.failed)

### nucleo-l433rc/failuresummary


### i-nucleo-lrwan1/failuresummary


### openmote-b/failuresummary


### nucleo-wl55jc/failuresummary


### nucleo-l4r5zi/failuresummary


### z1/failuresummary


### nucleo-f091rc/failuresummary

Failures during compilation:
- [tests/pkg_mbedtls](nucleo-f091rc/tests/pkg_mbedtls/compilation.failed)

### arduino-zero/failuresummary

Failures during compilation:
- [tests/pkg_mbedtls](arduino-zero/tests/pkg_mbedtls/compilation.failed)

### nucleo-l552ze-q/failuresummary


### nucleo-f303k8/failuresummary


### nucleo-f103rb/failuresummary


### b-l072z-lrwan1/failuresummary


### atmega256rfr2-xpro/failuresummary

Failures during test:
- [tests/pkg_mbedtls](atmega256rfr2-xpro/tests/pkg_mbedtls/test.failed)

### nucleo-f303re/failuresummary


### slstk3402a/failuresummary


### nucleo-f030r8/failuresummary

Failures during test:
- [tests/pkg_mbedtls](nucleo-f030r8/tests/pkg_mbedtls/test.failed)

### nucleo-l073rz/failuresummary

PeterKietzmann · 2022-01-26T16:20:31Z

Just a quick heads-up:

The failure on samr21-xpro is actually a good thing. It has insufficient entropy sources and configures only a fallback ADC line to gather noise using the adc_noise module. This detects failures during the runtime tests which lets the (mbedtls) entropy module fail. I will improve the test to better reflect this behavior.
Hanging on nucleo-l152re is due to a missing ADC initialization before use. Some ADC drivers block on sampling before initialization. Turns our, I forgot to call the initialization function in auto_init. Will add it.
Lets see if these changes will already let the other boards pass.

PeterKietzmann requested review from MichelRottleuthner, aabadie, fjmolinas, jia200x, leandrolanzieri, miri64 and smlng as code owners December 17, 2020 18:07

PeterKietzmann requested review from maribu and mtausig December 17, 2020 18:07

PeterKietzmann added Area: crypto Area: Cryptographic libraries Area: pkg Area: External package ports Area: security Area: Security-related libraries and subsystems Type: new feature The issue requests / The PR implemements a new feature for RIOT labels Dec 17, 2020

PeterKietzmann mentioned this pull request Dec 17, 2020

dist/tools/ci or tests: EXCLUDE not assigned correctly? #15660

Closed

miri64 reviewed Dec 17, 2020

View reviewed changes

PeterKietzmann force-pushed the pr_add_pkg_mbedtls branch from 01539b8 to 36bccad Compare December 18, 2020 12:40

aabadie previously requested changes Dec 18, 2020

View reviewed changes

PeterKietzmann force-pushed the pr_add_pkg_mbedtls branch 3 times, most recently from d11ef29 to f7a7105 Compare December 18, 2020 15:11

MrKevinWeiss added this to the Release 2021.07 milestone Jun 21, 2021

PeterKietzmann force-pushed the pr_add_pkg_mbedtls branch from f7a7105 to a3a2166 Compare July 5, 2021 17:52

github-actions bot added the Area: doc Area: Documentation label Jul 5, 2021

PeterKietzmann mentioned this pull request Nov 12, 2021

sys/random: default to musl LCG instead of TinyMT #17188

Merged

leandrolanzieri reviewed Nov 18, 2021

View reviewed changes

aabadie reviewed Nov 23, 2021

View reviewed changes

pkg/mbedtls/Makefile Outdated Show resolved Hide resolved

aabadie reviewed Nov 23, 2021

View reviewed changes

pkg/mbedtls/contrib/entropy/entropy_mbedtls_riot.c Outdated Show resolved Hide resolved

aabadie reviewed Nov 23, 2021

View reviewed changes

pkg/mbedtls/contrib/entropy/entropy_mbedtls_riot.c Outdated Show resolved Hide resolved

aabadie reviewed Nov 23, 2021

View reviewed changes

benpicco reviewed Nov 23, 2021

View reviewed changes

mariemC mentioned this pull request Jan 14, 2022

pkg/mbedtls: add TLS support for LWIP #17519

Open

benpicco approved these changes Jan 14, 2022

View reviewed changes

pkg/mbedtls:initial port of the entropy module

b05f934

This was referenced Jan 28, 2022

tests/pkg_mbedtls-blacklist: blacklist samr21-xpro #17588

Merged

tests/pkg_mbedtls: improve entropy test case and doc #17606

Merged

Einhornhool mentioned this pull request Sep 19, 2022

prng: how to seed the random number generator? #5150

Open

	-f $(RIOTPKG)/mbedtls/Makefile.$(PKG_NAME)
	-f $(RIOTBASE)/Makefile.base MODULE=$(PKG_NAME)

		CFLAGS += -DCONFIG_ENTROPY_SOURCE_ADC_COND=0
		CFLAGS += -DCONFIG_ENTROPY_SOURCE_ADC_HEALTH_TEST=1

	* FEATURES_PROVIDED += entropy_source_adc_noise
	* FEATURES_PROVIDED += periph_adc

	#if defined (MBEDTLS_SHA256_ALT)
	#if IS_ACTIVE(MBEDTLS_SHA256_ALT)

Conversation

PeterKietzmann commented Dec 17, 2020

Contribution description

Testing procedure

Issues/PRs references

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!

aabadie Dec 17, 2020 • 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!

PeterKietzmann commented Dec 17, 2020

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

miri64 commented Dec 17, 2020

Uh oh!

PeterKietzmann commented Dec 17, 2020

Uh oh!

aabadie 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!

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!

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!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

PeterKietzmann commented Dec 18, 2020

Uh oh!

PeterKietzmann commented Dec 18, 2020

Uh oh!

fjmolinas commented Mar 12, 2021

Uh oh!

PeterKietzmann commented Jul 8, 2021

Uh oh!

PeterKietzmann commented Jul 9, 2021

Uh oh!

fjmolinas commented Nov 18, 2021

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

aabadie Dec 17, 2020 •

edited

Loading

PeterKietzmann commented Nov 23, 2021 •

edited

Loading