Re: [PATCH] pcie_root_port: Add disable_hotplug option

["Michael S. Tsirkin" <mst@redhat.com>]

On Wed, Feb 19, 2020 at 01:21:10PM +0100, Julia Suvorova wrote:
> On Wed, Feb 19, 2020 at 4:47 AM Michael S. Tsirkin <mst@redhat.com> wrote:
> >
> > On Tue, Feb 18, 2020 at 10:02:19PM -0500, Laine Stump wrote:
> > > Also, is there a rhyme/reason for some options having true/false, and some
> > > being off/on? disable-acs seems to be true/false, but disable-modern is
> > > on/off. Doesn't make any difference to me in the end, but just thought I'd
> > > bring it up in case there might be a reason to use on/off instead of
> > > true/false for this one.
> >
> > Some places accept on/off, some true/false, some on/off/true/false
> > others on/off/yes/no and others on/off/true/false/yes/no.
> >
> > In this case both user visitor machinery. Which I *think*
> > means on/off is the safe choice and true/false can be
> > broken in some places.
> >
> > We really should clean up this mess ... Julia, what do you think?
> > Let's make them all support all options?
> 
> Options already support all of on/off/true/false/yes/no as long as
> they are defined as boolean (look at parse_type_bool()). That is, you
> can use disable-modern with yes/no/true/false too.
> The only problem is with types OnOffSplit and OnOffAuto (as in disable-legacy).

Right. Not only that though - parse_option_bool can only handle
on/off. target/sparc/cpu.c can handle on/off and true/false.
JSON bool is only true/false.
As you said OnOffSplit and OnOffAuto are on/off.

And from documentation POV, it's all over the place.



-- 
MST