Hi Michel,

We will pick up this and try to push a patch to fix it soon.

Hi Alan, Thanks for your workaround, and it is very helpful. We will update the T234 as soon as possible after the patch ready.

Thanks, Edison

-----Original Message----- From: TF-M tf-m-bounces@lists.trustedfirmware.org On Behalf Of DeMars, Alan via TF-M Sent: Wednesday, April 3, 2019 9:32 PM To: Ken Liu (Arm Technology China) Ken.Liu@arm.com Cc: tf-m@lists.trustedfirmware.org Subject: Re: [TF-M] build with GNUARM with option -Os

I assumed there was an outstanding pull request for this ticket:

https://developer.trustedfirmware.org/T234

Alan

-----Original Message----- From: TF-M [mailto:tf-m-bounces@lists.trustedfirmware.org] On Behalf Of Ken Liu (Arm Technology China) via TF-M Sent: Tuesday, April 02, 2019 6:59 PM To: tf-m@lists.trustedfirmware.org Cc: nd Subject: [EXTERNAL] Re: [TF-M] build with GNUARM with option -Os

Hi Alan, Is it OK for you to commit this change?

-Ken

-----Original Message----- From: TF-M tf-m-bounces@lists.trustedfirmware.org On Behalf Of DeMars, Alan via TF-M Sent: Wednesday, April 3, 2019 5:50 AM To: Michel JAOUEN michel.jaouen@st.com Cc: tf-m@lists.trustedfirmware.org Subject: Re: [TF-M] build with GNUARM with option -Os

I had a similar problem. Upon advice in this email list, I fixed it by changing the implementation of "tfm_core_ns_ipc_request()" in secure_fw/core/tfm_psa_api_client.c.

Add 'volatile' to the declaration of these variables:

struct tfm_sfn_req_s desc, *desc_ptr = &desc;

change to:

volatile struct tfm_sfn_req_s desc, *desc_ptr = &desc;

After this change, I am able to build and run with -O3 as well as -Os.

I don't know why this fix hasn't been added to the master branch as this problem has already been identified.

Alan

-----Original Message----- From: TF-M [mailto:tf-m-bounces@lists.trustedfirmware.org] On Behalf Of Michel JAOUEN via TF-M Sent: Tuesday, April 02, 2019 4:11 AM To: tf-m@lists.trustedfirmware.org Subject: [EXTERNAL] [TF-M] build with GNUARM with option -Os

Hello, I tested my board port on top of 1c266ae74bd93c2ef290e9aac0caecf92b06b93d Without option -Os , the tests with ConfigCoreIPC.cmake are passed . When I put the option -Os , 1st test is failing in Hardware Fault.

For info, the test with ConfigRegression.cmake and option -Os is passed .

With the configuration -Os , code footprint is much better.

Is it plan to activate this option in master ? Is the same issue reproduced on another board ?

Best regards

-- TF-M mailing list TF-M@lists.trustedfirmware.org https://lists.trustedfirmware.org/mailman/listinfo/tf-m -- TF-M mailing list TF-M@lists.trustedfirmware.org https://lists.trustedfirmware.org/mailman/listinfo/tf-m

-- TF-M mailing list TF-M@lists.trustedfirmware.org https://lists.trustedfirmware.org/mailman/listinfo/tf-m -- TF-M mailing list TF-M@lists.trustedfirmware.org https://lists.trustedfirmware.org/mailman/listinfo/tf-m