This discussion has been locked.

You can no longer post new replies to this discussion. If you have a question you can start a new discussion

Error L6244E Region Not Aligned on CM33 Secure World with Arm v6.19 Compiler

Murat Cakmak over 1 year ago

Hello,

We are working on a CM33 device and developing a code for the Secure World.

We are using Arm v6.19 Compiler, and our scatter file is like this

LR_IROM1 0x0C010100 0xFF00 { ; load region size_region

...

}

We need to leave 0x100 bytes before the image to add a metadata as post-build step.

We need to align a variable with 512

uint8_t globalBuffer[512] __attribute(align(512));

But when we align any symbol, we are getting the following error

.\Objects\OSTests_S1.axf: Error: L6244E: Load region LR_IROM1 address (0x0c010100) not aligned on a 512 byte boundary.

If we align with 256, there is no problem as LR_IROM1 offset aligned with 256.

The same code is used by different architectures, so we need to keep the alignment as is.

But it does not make sense, why a variable alignment affects the image alignment.

And there is no documentation (seems empty) about this error

Any idea?

Thanks

Parents

0 Murat Cakmak over 1 year ago in reply to Ronan Synnott

Hi @ronans,

Thank you for the details.

Let me give a bit more about our scatter file. We have a section to collect all RO, called ER_IROM1 inside the LR_IROM1.

LR_IROM1 0x0C010100 0xFF00 {
ER_IROM1 0x0C010100 0xFF00 {
      *.o (RESET, +First)
      *(InRoot$$Sections)
      .ANY (+RO)
}

  RW_IRAM1 0x30008000 0x00002000 { ; RW data
      .ANY (+RW +ZI)
*(STACK)
  }
}

Yes, we tried to add a section for 0x100 size for metadata just before the ER_IROM1 section, (And we set the LR_IROM offset to 0x0C010000 from 0x0C010100 to include the Metadata)

but then we started to get the same error for ER_IROM1, not for LR_IROM.

But still does not make sense, why a variable alignment is related to the Code section?

As the online documentation is empty, where to get the a full documentation about the linker behaviour.

PS: I can be wrong but we did not have this error before, we have a new setup with the latest MDK 3.28. I am wondering does it come from the toolchain?
Cancel
Up 0 Down

Cancel

Reply

0 Murat Cakmak over 1 year ago in reply to Ronan Synnott

Hi @ronans,

Thank you for the details.

Let me give a bit more about our scatter file. We have a section to collect all RO, called ER_IROM1 inside the LR_IROM1.

LR_IROM1 0x0C010100 0xFF00 {
ER_IROM1 0x0C010100 0xFF00 {
      *.o (RESET, +First)
      *(InRoot$$Sections)
      .ANY (+RO)
}

  RW_IRAM1 0x30008000 0x00002000 { ; RW data
      .ANY (+RW +ZI)
*(STACK)
  }
}

Yes, we tried to add a section for 0x100 size for metadata just before the ER_IROM1 section, (And we set the LR_IROM offset to 0x0C010000 from 0x0C010100 to include the Metadata)

but then we started to get the same error for ER_IROM1, not for LR_IROM.

But still does not make sense, why a variable alignment is related to the Code section?

As the online documentation is empty, where to get the a full documentation about the linker behaviour.

PS: I can be wrong but we did not have this error before, we have a new setup with the latest MDK 3.28. I am wondering does it come from the toolchain?
Cancel
Up 0 Down

Cancel

Children

0 Ronan Synnott over 1 year ago in reply to Murat Cakmak

Hi Murak,

I've built a test case, and I can replicate the issue. I will discuss with colleagues internally.

Ronan
Cancel
Up 0 Down

Cancel