DRA821U: Regarding the change in SDL's make rules.

tomitama

Part Number: DRA821U

Regarding the change in SDL's make rules.

I am using "ti-processor-sdk-rtos-j7200-evm-08_04_00_04".

I would like to create and use an SDL library that matches the settings of my environment.
For that, I changed the make rule "<SDK>/sdl/build/makerules/rules_r5f.mk".
Will this change affect the functionality of SDL?

Here are the changes:
---------------------------------------------------
# Internal CFLAGS - normally doesn't change
#CFLAGS_INTERNAL = -c -Wall -Werror $(SUPRESS_WARNINGS_FLAG) -$(RTSLIB_ENDIAN) -eo.$(OBJEXT) -ea.$(ASMEXT) -g -mfloat-abi=hard
CFLAGS_INTERNAL = -c -Wall -Werror $(SUPRESS_WARNINGS_FLAG) -$(RTSLIB_ENDIAN) -eo.$(OBJEXT) -ea.$(ASMEXT) -g -mfloat-abi=soft

# Vector Floating-Point co-processor instructions, vfpv3d16 is supported in R4/R5
#CFLAGS_INTERNAL += -mfpu=vfpv3-d16 -mcpu=cortex-r5 -march=armv7-r
CFLAGS_INTERNAL += -mcpu=cortex-r5 -march=armv7-r
---------------------------------------------------

over 2 years ago

0 Karthik Ramanan over 2 years ago

TI__Guru** 113810 points

Tomitama,

Can you please explain the need for the change to allow us to understand the background and respond back to you?

Regards
Karthik

0 tomitama over 2 years ago in reply to Karthik Ramanan

Expert 1090 points

Hi Karthik,

I'm getting a "uses VFP register arguments" error when using the SDL library in a project I'm using.
I think this is because the VPF setting is disabled in the compiler options.
See "VPF_OFF_1.png".

But I get the same error even with the VPF setting enabled.
See "VPF_ON_1.png" and "VPF_ON_2.png".

To avoid this error, change the make rule "<SDK>/sdl/build/makerules/rules_r5f.mk".

Regards.

0 Karthik Ramanan over 2 years ago in reply to tomitama

TI__Guru** 113810 points

Hi Tomitama,

The team is once again starting to look at this issue, we will have some progress in the next week or so.

Apologies for the delay.

Regards

Karthik

0 Josiitaa RL over 2 years ago in reply to tomitama

TI__Mastermind 40846 points

Hi Tomitama,

Using the hard-float ABI is beneficial while running applications relying on floating-point operations as it provides higher performance. This setting must be consistent for all objects that are to be linked - the application and the SDL library.

This link might help - embeddedartistry.com/.../

I will check with our SDL expert and get back to you regarding any functionality changes due to this modification within a few days.

Thanks,

Josiitaa

0 Josiitaa RL over 2 years ago in reply to tomitama

TI__Mastermind 40846 points

Hi Tomitama,

Since SDL does not have any floating point functions, this change in the make rule should not affect any of the SDL functionalities.

Thanks,

Josiitaa

Processors

Processors forum

DRA821U: Regarding the change in SDL's make rules.