We are running a survey to help us improve the experience for all of our members. If you see the survey appear, please take the time to tell us about your experience if you can.
HI Team,
We are encountering a synchronous data abort (memory abort) immediately after executing the ISB instruction in the enable_mmu_direct_el3() function in TF-A BL1 image. We are using neoverse v2 reference design. How we can find out the root cause of this issue. The following is the reference snippet code.
<enable_mmu_direct_el3 MRS x4,SCTLR_EL3<enable_mmu_direct_el3 MOV x5,#5<enable_mmu_direct_el3 MOVK x5,#8,LSL #16<enable_mmu_direct_el3 ORR x4,x4,x5<enable_mmu_direct_el3 AND x5,x4,#0xfffffffffffffffb<enable_mmu_direct_el3 TST x7,#1<enable_mmu_direct_el3 CSEL x4,x5,x4,NE<enable_mmu_direct_el3 MSR SCTLR_EL3,x4<enable_mmu_direct_el3 ISB<enable_mmu_direct_el3 RET
Register values when issue occured.
ESR_EL3 : 0x0000000086000014SCTLR_EL3: 0x0000000030ed183f
FAR_EL3:0x0000000000010200
The ISB ensures that previous changes to context (in this case, the write to SCTLR) are visible to following instructions. Without knowing what SCTLR_EL3 was previously set to, my guess is that it's the effect of enabling the MMU which is causing the problem.
If I've got my mental hex decoding correct, ESR_EL3.EC is reporting an Instruction Abort and ISS is reporting sync External abort on table walk. Meaning that it was unable to fetch the needed translation table entry to process an instruction fetch.
You'll need to check the rest of the MMU set up to make sure the translation tables are in the correct place.