[toup] zephyr: crypto: port crypto mbedtls wrapper

[fengming-ye]

Port crypto mbedtls wrapper for enterprise and DPP.

Porting from
https://github.com/gstrauss/hostap/blob/mbedtls/src/crypto/crypto_mbedtls.c

[fengming-ye]

Hi @jukkar @krish2718
crypto_mbedtls.c and tls_mbedtls.c are both total re-write and replace.
is this okay to merge?

[fengming-ye]

Update as previous comments.
Add crypto_mbedtls_alt.c and tls_mbedtls_alt.c enabled by CONFIG_WIFI_NM_WPA_SUPPLICANT_CRYPTO_ALT
in this PR
zephyrproject-rtos/zephyr#73249

[dleach02]

Are we good with the copyright concerns now?

[jukkar]

Are we good with the copyright concerns now?

Yes, looks much better now, thanks for the fixes.

[jukkar]

@krish2718 please take a look

[krish2718]

LGTM, except for last commit, can you please elaborate on Thus when DPP enabled, unfixed warning will pop out? The fix for warnings are essential to pass tests with -Werror and the data type fix was found using coverity (downstream NCS).

[fengming-ye]

LGTM, except for last commit, can you please elaborate on Thus when DPP enabled, unfixed warning will pop out? The fix for warnings are essential to pass tests with -Werror and the data type fix was found using coverity (downstream NCS).

Sry I didn't catch your point. The last commit is to fix warnings below:

modules/lib/hostap/src/common/dpp.c:3293:37: warning: unsigned conversion from 'int' to 'enum dpp_status_error' changes value from '256' to '0' [-Woverflow]
modules/lib/hostap/src/common/wpa_ctrl.h:208:24: warning: format '%x' expects argument of type 'unsigned int', but argument 4 has type 'long unsigned int' [-Wformat=]

[krish2718]

LGTM, except for last commit, can you please elaborate on Thus when DPP enabled, unfixed warning will pop out? The fix for warnings are essential to pass tests with -Werror and the data type fix was found using coverity (downstream NCS).

Sry I didn't catch your point. The last commit is to fix warnings below:

modules/lib/hostap/src/common/dpp.c:3293:37: warning: unsigned conversion from 'int' to 'enum dpp_status_error' changes value from '256' to '0' [-Woverflow] modules/lib/hostap/src/common/wpa_ctrl.h:208:24: warning: format '%x' expects argument of type 'unsigned int', but argument 4 has type 'long unsigned int' [-Wformat=]

My bad not the last commit, it's this 00c11d0 revert.

[fengming-ye]

LGTM, except for last commit, can you please elaborate on Thus when DPP enabled, unfixed warning will pop out? The fix for warnings are essential to pass tests with -Werror and the data type fix was found using coverity (downstream NCS).

Sry I didn't catch your point. The last commit is to fix warnings below:
modules/lib/hostap/src/common/dpp.c:3293:37: warning: unsigned conversion from 'int' to 'enum dpp_status_error' changes value from '256' to '0' [-Woverflow] modules/lib/hostap/src/common/wpa_ctrl.h:208:24: warning: format '%x' expects argument of type 'unsigned int', but argument 4 has type 'long unsigned int' [-Wformat=]

My bad not the last commit, it's this 00c11d0 revert.

Oh this is because except from some printfs %ld changed to %lld. There are more %ld unchanged. And I don't want to change much DPP original code for this. I have doubt that if %ld is changed to %lld, won't it cause type warning on Linux?

On the other hand, why would os_time_t defined in unsigned long has warnings as Linux is using unsigned long?
Could you pls share the warning?

BTW my local build has no warning with these commits.

[krish2718]

LGTM, except for last commit, can you please elaborate on Thus when DPP enabled, unfixed warning will pop out? The fix for warnings are essential to pass tests with -Werror and the data type fix was found using coverity (downstream NCS).

Sry I didn't catch your point. The last commit is to fix warnings below:
modules/lib/hostap/src/common/dpp.c:3293:37: warning: unsigned conversion from 'int' to 'enum dpp_status_error' changes value from '256' to '0' [-Woverflow] modules/lib/hostap/src/common/wpa_ctrl.h:208:24: warning: format '%x' expects argument of type 'unsigned int', but argument 4 has type 'long unsigned int' [-Wformat=]

My bad not the last commit, it's this 00c11d0 revert.

Oh this is because except from some printfs %ld changed to %lld. There are more %ld unchanged. And I don't want to change much DPP original code for this. I have doubt that if %ld is changed to %lld, won't it cause type warning on Linux?

The fixes weren't upstreamed but this all depends on the host machine that you are building for.

On the other hand, why would os_time_t defined in unsigned long has warnings as Linux is using unsigned long? Could you pls share the warning?

This was warning from coverity, and it should be signed 64bit to capture full times resolution.

I need to dig my mailbox for the warnings, but I would suggest to remove this commit from the PR and we can handle it separately as it needs a discussion. Rest of the commits are good to go. WDYT?

BTW my local build has no warning with these commits.

Yeah, might be toolchain dependent and which flags are enabled by default.

[fengming-ye]

LGTM, except for last commit, can you please elaborate on Thus when DPP enabled, unfixed warning will pop out? The fix for warnings are essential to pass tests with -Werror and the data type fix was found using coverity (downstream NCS).

Sry I didn't catch your point. The last commit is to fix warnings below:
modules/lib/hostap/src/common/dpp.c:3293:37: warning: unsigned conversion from 'int' to 'enum dpp_status_error' changes value from '256' to '0' [-Woverflow] modules/lib/hostap/src/common/wpa_ctrl.h:208:24: warning: format '%x' expects argument of type 'unsigned int', but argument 4 has type 'long unsigned int' [-Wformat=]

My bad not the last commit, it's this 00c11d0 revert.

Oh this is because except from some printfs %ld changed to %lld. There are more %ld unchanged. And I don't want to change much DPP original code for this. I have doubt that if %ld is changed to %lld, won't it cause type warning on Linux?

The fixes weren't upstreamed but this all depends on the host machine that you are building for.

On the other hand, why would os_time_t defined in unsigned long has warnings as Linux is using unsigned long? Could you pls share the warning?

This was warning from coverity, and it should be signed 64bit to capture full times resolution.

I need to dig my mailbox for the warnings, but I would suggest to remove this commit from the PR and we can handle it separately as it needs a discussion. Rest of the commits are good to go. WDYT?

BTW my local build has no warning with these commits.

Yeah, might be toolchain dependent and which flags are enabled by default.

Okay I removed this commit.