Re: [TF-M] Inline asm syntax unified and symbol reference in IAR toolchain

Hi Ken,

I see you implemented this change anyway  in psa_interface_thread.c and other places. This breaks IAR builds unfortunately: --- ...           :: "i" (spm_interface_thread_dispatcher)              ^ "C:\Users\thomasto\Projects\tf-m13\trusted-firmware-m\secure_fw\spm\cmsis_psa\psa_interface_thread.c",29 Error[Ta094]:           Operand modifier a is not valid for operand           #spm_interface_thread_dispatcher

:: "i" (tfm_spm_client_psa_framework_version),              ^ "C:\Users\thomasto\Projects\tf-m13\trusted-firmware-m\secure_fw\spm\cmsis_psa\psa_interface_thread.c",43 Error[Ta094]:           Operand modifier a is not valid for operand           #tfm_spm_client_psa_framework_version ... ---

For IAR we need to use the #pragma required method previously used.

/Thomas

Den 2021-10-19 kl. 11:11, skrev Ken Liu via TF-M:

Hi Thomas,

Thanks for the confirm 😉

Let’s see if we can refine this part from a coding perspective.

/Ken

*From:* TF-M tf-m-bounces@lists.trustedfirmware.org *On Behalf Of *Thomas Törnblom via TF-M *Sent:* Tuesday, October 19, 2021 4:53 PM *To:* tf-m@lists.trustedfirmware.org *Subject:* Re: [TF-M] Inline asm syntax unified and symbol reference in IAR toolchain

Hi Ken,

We have an open ticket for just consuming the ".syntax unified", which is set to be fixed in a future update.

I've tested your suggestion for "#pragma required" and it doesn't seem to work.

Cheers, Thomas

Den 2021-10-19 kl. 09:22, skrev Ken Liu via TF-M:

Hi Thomas and all,

I noticed there are some

#if !defined(__ICCARM__)

        ".syntax unified \n"

#endif

In source code, looks like �.syntax unified� is not support in
IAR, is that true? If it could not be supported in a short term,
we can define some wrapper such as:

#ifdef __ICCARM__

#define CLAIM_SYNTAX_UNIFIED �\n�

#else

#define CLAIM_SYNTAX_UNIFIED ".syntax unified                  \n"

#endif

Another question is about the:

#if defined(__ICCARM__)

#pragma required = do_schedule

#endif

If we claim do_schedule in the constraints, is the above
�#pragma required� still needed?

__asm (��.  :: �i�(do_schedule));

We can create a patch to test this � using a constraint looks
more proper.

Thanks.

/Ken

--

*Thomas T�rnblom*, /Product Engineer/ IAR Systems AB Box 23051, Strandbodgatan 1 SE-750 23 Uppsala, SWEDEN Mobile: +46 76 180 17 80 Fax: +46 18 16 78 01 E-mail: thomas.tornblom@iar.comWebsite: www.iar.com http://www.iar.com Twitter: www.twitter.com/iarsystems http://www.twitter.com/iarsystems