tf-a-tools: fix building the native target#41
tf-a-tools: fix building the native target#41quaresmajose wants to merge 0 commit intoSTMicroelectronics:kirkstonefrom
Conversation
|
Thanks for your contribution Jose. |
Thanks, It is a pleasure to be able to contribute to your project |
|
Please old on a bit as I need to confirm an issue that I see in our CI So my affirmation |
09c28db to
90673b3
Compare
the patch tf-a-tools: fix the RPATH for the native target fixes the issue so it is ready to review. |
|
Hi Jose, I discussed with dev team and ST policy is to stick with tf-a 2.6 (as a whole) and not upgrading some parts of tf-a and not others (like here for tf-a tools). Could you please elaborate a bit why do you want to upgrade tf-a tools to 2.7 version (what is your motivation) ? |
|
I am fine in not upgrading the tf-a tools to 2.7 for now and only fix the native target so I will drop the update patch |
90673b3 to
6c5464b
Compare
|
We're using upstream 2.7 (instead of the 2.6 from ST BSP), but I guess it is easier for us to do the bump at our own layer instead. |
|
ok Clear Ricardo. |
|
The fix are merged on top of kirkstone branche |
6c5464b to
96bffa5
Compare
|
The current 2.6 version in Kirkstone fails to build when creating a toolchain/SDK.
First, there are a bunch of warnings converted to errors because of '-Werror':
Then it also fails at link time:
I guess all this is because Kirkstone does only support OpenSSL 3.0.x, while the tf-a-tools version 2.6 seems to expect OpenSSL 1.x. Is there a fix for Thanks. |
|
@jviguera have you tried build the
if your suspicion is correct it was already broken before. |
|
@quaresmajose and @jviguera , I just pushed a pach to correct the compilation with openssl3. I have test with yocto build for native and native-sdk. |
|
@cpriouzeau but with the pushed patch the native target is broked again 766634e#diff-8b1d74ca8a6425a18839337d0994329c2a7b4f093434a07238f198df3641a3baL23-L30 as this is needed for the native target as you can see in 96bffa5 96bffa5 doesn't fix any build issue but instead it fixes a runtime issues. this happens because the native So as an example if we build this
|
|
@cpriouzeau are you sure this recipe works on the target? 766634e#diff-8b1d74ca8a6425a18839337d0994329c2a7b4f093434a07238f198df3641a3baL14 |
|
this patch does the job for me and don't break the native fiptool, don't haven know if the nativesdk fiptool runs fine but already builds |
- The addiction of the HOSTCC breaks the nativesdk target build - The RPATH only needs to be fixed for the native target regression introduced in: STMicroelectronics#41 Signed-off-by: Jose Quaresma <[email protected]>
No description provided.