RE: ARM64 EL3 Exception SP_EL3 Issue?

[Chad Karaginides (QUIC)]

Sorry Peter you are right.  I'm not sure why I thought I saw different behavior 
on a real target.  Please disregard my email.  Thank you for the quick reply.

Take care,
Chad

-----Original Message-----
From: Peter Maydell <peter.maydell@linaro.org> 
Sent: Thursday, August 3, 2023 2:39 AM
To: Chad Karaginides (QUIC) <quic_chadk@quicinc.com>
Cc: qemu-arm@nongnu.org
Subject: Re: ARM64 EL3 Exception SP_EL3 Issue?

On Wed, 2 Aug 2023 at 21:04, Chad Karaginides (QUIC) <quic_chadk@quicinc.com> 
wrote:
>
> Hello,
>
>
>
> I am executing the Zephyr RTOS on a QEMU A53 core with Trace32 as the 
> frontend.  I am seeing a behavior difference between the QEMU A53 core and 
> real hardware.  Execution is in EL1h and an SMC call is made.  Execution 
> transitions to EL3h but the EL1 stack pointer is not propagated to SP_EL3.  
> Is this a potential bug or is this functionality covered in a test case 
> somewhere?  If it is a bug I have no problem diving in and trying to help fix 
> it.  I am new to QEMU but willing to help.  Just wanted to make sure I am not 
> doing something wrong first.

Why do you expect EL3 to execute with EL1's stack pointer?
On entry to EL3 the CPU should use the SP_EL3 (which EL3 has set up before it 
dropped down to EL2 or EL1).

thanks
-- PMM