[prev in list] [next in list] [prev in thread] [next in thread] 

List:       illumos-developer
Subject:    Re: [developer] About _LP64 ifdefs (was: 9641 want stack-clash mitigation)
From:       John Levon <levon () movementarian ! org>
Date:       2018-07-16 13:17:14
Message-ID: 20180716131714.GA13865 () movementarian ! org
[Download RAW message or body]

On Mon, Jul 16, 2018 at 08:49:03AM -0400, Gordon Ross via illumos-developer wrote:

> >> libfakekernel etc. that I've found very useful).
> >
> > Why do we need a 32-bit libfakekernel?
> 
> There are cases where 32-bit is just easier, i.e. linking
> "kernel" (+libfakekernel) and a user-level consumer...

This doesn't sound like a very strong argument IMHO.

> > Keeping around a half-and-half system, with huge amounts of dead code,
> > is just going to get horribly messy. So while I agree there's no need
> > for a mission, it seems like a good idea to remove when suitable.
> 
> Another viewpoint is it's just good portability, keeping code usable
> in both 64-bit and 32-bit environments.   I guess you presume to
> know that there will never be another 32-bit environment that
> might want to compile that common code?  I'm not sure how
> you'd be so certain of that.  (I just pointed out one such env.)

Unbuilt and untested code is code that doesn't work. I've rarely come
across counter-examples to this. Or are you volunteering to keep this
code at least sanity built somehow?

To say nothing of the considerable burden of all this code.

regards
john

------------------------------------------
illumos: illumos-developer
Permalink: https://illumos.topicbox.com/groups/developer/Tfab2cda109ec0703-M7ae44a16e5bd5d598a4009d1
Delivery options: https://illumos.topicbox.com/groups
[prev in list] [next in list] [prev in thread] [next in thread]