Add TRNG API test #7057
Add TRNG API test #7057
Conversation
TESTS/mbed_hal/trng/main.cpp
Outdated
| else | ||
| { | ||
| printf("compression for concatenated buffer after reset was unsuccessful"); | ||
| TEST_ASSERT(false); |
There was a problem hiding this comment.
This can be ASSERT with msg (then printf not needed). I think this entire if can be turned into one assert.
Similar to: TEST_ASSERT_TRUE_MSG(compr_res < buffer_len , "failed......msg....here")
There was a problem hiding this comment.
Because the test contains several parts I wanted the user to have a clear view of the progress, the assert message is only shown when the test fails.
TESTS/mbed_hal/trng/main.cpp
Outdated
| (unsigned char **)htab); | ||
|
|
||
| if (comp_res >= BUFFER_LEN) | ||
| { |
There was a problem hiding this comment.
Can you fix formatting (follow mbed OS code style) , this one ({) should be attached
There was a problem hiding this comment.
Sorry for that, all formatting will be fixed
| @@ -0,0 +1,16 @@ | |||
| #include <stdint.h> | |||
There was a problem hiding this comment.
license header missing in some files (please review), not only this one
There was a problem hiding this comment.
Sorry for that, all license header will be added.
|
From the changes, looks like mbed tls and strorage teams should review this PR? |
|
This appears to have a GPL on it. @0xc0170 |
ghost
added
the
|
@ARMmbed/mbed-os-storage |
|
GPL is still present in this PR. We cannot integrate GPL code with Mbed OS. |
|
Compression lib was changed to pithy (with BSD license) |
|
@ARMmbed/mbed-os-crypto Mind reviewing? |
|
@cmonr We'll review it shortly! Could you please add the "TLS" label to the PR? |
There was a problem hiding this comment.
Looks good to me!
@mohammad1603 will be giving a review too and we need his approval before merging.
|
Thansk for the heads up @yanesca |
TESTS/host_tests/trng_reset.py
Outdated
| """ | ||
| This script is the host script for trng test sequence, it send the | ||
| step signaling sequence and receive and transmit data to the device after | ||
| reset if necesarry (default lading and storing while reseting the device |
There was a problem hiding this comment.
typo: lading -> loading
TESTS/host_tests/trng_reset.py
Outdated
| This script is the host script for trng test sequence, it send the | ||
| step signaling sequence and receive and transmit data to the device after | ||
| reset if necesarry (default lading and storing while reseting the device | ||
| is NVstore, in case NVstore isn't enabled we'll use current infrastructure, |
There was a problem hiding this comment.
what do you mean "device is NVstore" ?
There was a problem hiding this comment.
"default lading and storing while reseting the device is NVstore", the device is the device used for testing, the NVstore is one way of storing and loading data
TESTS/host_tests/trng_reset.py
Outdated
|
|
||
| wait_for_communication = yield | ||
|
|
||
| if self.reset == False: |
There was a problem hiding this comment.
do you expect the device to reset after step1 (dummy message) ?
There was a problem hiding this comment.
Yes, If not an error will occur.
| return BASE64_INVALID_PARAMETER; | ||
| } | ||
|
|
||
| if ((base64 >= 'A') && (base64 <= 'Z')) |
There was a problem hiding this comment.
Style (minor): mixed style of using {.
if (a){
xx;
}
AND
if (a)
{ xx; }
| } | ||
|
|
||
| *writePtr = 0; | ||
| while (( string[currPos] != 0 ) && |
There was a problem hiding this comment.
i think you should switch between the first two conditions - the current code can cause an out of bounds access.
in case the string doesn't end with 0 you can see this behavior.
should be:
while (( currPos < stringMaxSize ) &&
( string[currPos] != 0 ) &&
( writePtr < bufferEnd ) &&
( !isEndOfString )) {
| ( !isEndOfString )) { | ||
| uint8_t val; | ||
|
|
||
| if (string[currPos] == 0 || currPos >= stringMaxSize) |
There was a problem hiding this comment.
in the while you validate this, you can avoid this checks (they will always be true)
| // Check if additional bytes should have been processed but buffer isn't sufficient. | ||
| if (( result == BASE64_SUCCESS ) && | ||
| ( !isEndOfString ) && | ||
| ( string[currPos] != '=' ) && |
There was a problem hiding this comment.
if currPos == stringMaxSize this can be a problem as above
TESTS/mbed_hal/trng/main.cpp
Outdated
| /* | ||
| * The test is based on the assumption that trng will generate random data, random so | ||
| * there will not be any similar patterns in it, that kind of data will be impossible to | ||
| * compress, if compression will acuur the test will result in failure. |
|
@dannybenor TPIP? |
|
@cmonr - TPIP == Third Party IP. This looks to be BSD, which is compatible. I'll run it past legal for a TPIP review just to be sure. |
|
@ChiefBureaucraticOfficer Is this good to progress? |
|
Hi @0xc0170 , |
|
/morph test |
Test : FAILUREBuild number : 2838 |
|
Oh, I feel silly. This needed to have been a build, not a re-test. /morph build |
Build : SUCCESSBuild number : 3067 Triggering tests/morph test |
Exporter Build : SUCCESSBuild number : 2679 |
|
/morph build |
Build : SUCCESSBuild number : 3073 Triggering tests/morph test |
Test : FAILUREBuild number : 2864 |
|
TRNG test seem to fail on K22F. @theamirocohen please take a look. |
Exporter Build : SUCCESSBuild number : 2686 |
|
/morph build |
Build : SUCCESSBuild number : 3156 Triggering tests/morph test |
Exporter Build : SUCCESSBuild number : 2760 |
Test : FAILUREBuild number : 2966 |
|
sigh It looks like failure was an intermittent network issue. /morph test |
Test : SUCCESSBuild number : 2967 |
Use TRNG api to construct different buffers and try to compress them, If the random data is not really random, the test will be able to compress it.
We also check that after device reset when we use TRNG the API will create different data.
We do use lzf lib to compress the random buffer, the library should be checked for copyrights violation by @ChiefBureaucraticOfficer , all files are documented and under the lzflib library.