Looks good over all, just some nits on errx vs err when it appears that there is a proper errno avaliable.

I am not so sure on using sysexits.h either, would rather have a bhyve specific list of error codes as is currently implemented (1 to 4), only expanded to include the any many error conditons that it can exit on.

I have hated tracing down why I get an exit(1) in the code, though this does improve on that some, most of them are just going to be a small set of sysexits codes, and it would greatly help if each exit code was unique and pointed to a very specific problem.

Thank you! Looks good to me, aside of mentioned by Rod err()/errx() issues. All perror()'s should probably turn into err() instead of errx().

This is also going to require a man page update, as now there are more than the 4 exit codes listed in the man page.

In D16161#342752, @rgrimes wrote:

Looks good over all, just some nits on errx vs err when it appears that there is a proper errno avaliable.

I am not so sure on using sysexits.h either, would rather have a bhyve specific list of error codes as is currently implemented (1 to 4), only expanded to include the any many error conditons that it can exit on.

I have hated tracing down why I get an exit(1) in the code, though this does improve on that some, most of them are just going to be a small set of sysexits codes, and it would greatly help if each exit code was unique and pointed to a very specific problem.

The idea of this patch is make the exit code 1 be more unique to 'powered off', although I agree that we need more unique exit codes for each specific problem. I think to achieve a full and more useful exit code we shall need a generic error interface implementation where we can use among all bhyve drivers. However to make that happens it will takes me some time to audit each driver and make them more consistent with error codes.

In D16161#343293, @rgrimes wrote:

This is also going to require a man page update, as now there are more than the 4 exit codes listed in the man page.

I don't think we really need that, because those are the main error codes that users should look at. I do believe later as soon as we have a better and full error code approach then we deserve a manpage update.

Best,

err(errno) is probably not going to work well. For example ENOENT is 2 which maps to "2 halted" in the manpage.

I would rather avoid the use of sysexits(3) as it is not portable and it is actually mostly not used in FreeBSD. I would suggest just adding a new exit code for these errors that is documented in the manpage, e.g. "4 failed to start" and change all these to use that.

In D16161#343831, @jhb wrote:

Oof, I failed to realize that some of these errors are not startup related, so maybe just '4 exited due to an error' in the manpage. I do think the earlier changes to use err() were also good changes, I would just stick with '4' for the exit value. However, if you want to commit this as-is for now and adopt err() in place of fprintf/perror + exit as a later change that is ok with me as well. This is fine with me with the manpage tweak (or some similar tweak).

OK, I will uppate the manpage and commit it as-is just to make the exit(1) more unique for now. I have intention to revisit it and improve all the exit returns. I will replace fprintf/perror + exit as soon as I finish the drivers analysis to make sure what can be improved as error, exit and return code.

Thanks @jhb