On QNX ARM, 64-bit float arguments must be aligned. The call to operationArrayPushDouble leads to unpredictable results, because the J_DFGOperation_EDA operation doesn't currently take this into consideration.
A patch will follow.
Created attachment 200527 [details]
Comment on attachment 200527 [details]
I am clearing the r?/cq? flags at this moment, as I realized that alignment of FPRReg by the means of EABI_32BIT_DUMMY_ARG might require different handling between QNX and Linux.
Moreover, I realized that V_DFGOperation_EOZD (calling operationPutDoubleByValBeyondArrayBoundsStrict and operationPutDoubleByValBeyondArrayBoundsNonStrict) needs to be fixed in the same way as J_DFGOperation_EDA.
This is too old to be actionable.