|
From: | Albrecht Frenzel |
Subject: | Re: [Simulavr-devel] Whats to do on execution of "illegal instruction" |
Date: | Sat, 13 Feb 2016 18:36:25 +0100 |
User-agent: | Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101 Thunderbird/38.5.1 |
Most invalid instructions should default to break
An optional break ist very useful for debug, but "most invalid instructions" implies, that there should be exceptions. Which and why?
That means, a behaviour, that will never be observed on real hardware. Can you explayn _why_ this should be done?Some invalid instructions, e.g. 0xFFFF, should have an instruction-specific default, e.g. NOP.
Why not write a shell script or a batch file, which invokes simulavr with options?Use the GNU guys' @FILE mechanism to deal with a multitude of command-line options.
What does simulavr currently do with LPM R30, Z+?
simulavr terminates - which real hardware never does. Albrecht On 13.02.2016 18:13, Michael Hennebry wrote:
Most invalid instructions should default to break Some invalid instructions, e.g. 0xFFFF, should have an instruction-specific default, e.g. NOP. A config file or command line option should allow a user to change the global default to a compiled-in option, e.g. NOP or abort. A config file or command line option should allow the user to change instruction-specific behavior to a compiled-in option, e.g. change LPM R30, Z+ to NOP or to R30 = *(Z++) . Use the GNU guys' @FILE mechanism to deal with a multitude of command-line options. For the brave and industrious, a mechanism for run-time loading from a dynamic library might be useful. What does simulavr currently do with LPM R30, Z+? Michael Hennebry
[Prev in Thread] | Current Thread | [Next in Thread] |