On 8/19/21 5:39 AM, Peter Maydell wrote:
cpu_ldl_code() in the unaligned-access handler is strictly speaking bogus, because the page might have been unmapped after translation but before we got round to actually running it. Better would be to stash the relevant bits of info from the insn in the insn_start param, the way Arm does with the syndrome info.
Yep. That was more than I was prepared to do here. r~
