svn commit: r297190 - head/sys/kern

[Warner Losh]

On Tue, Mar 22, 2016 at 7:50 PM, Ian Lepore <ian at freebsd.org> wrote:

> On Tue, 2016-03-22 at 19:28 -0600, Warner Losh wrote:
> > On Tue, Mar 22, 2016 at 7:13 PM, Rick Macklem <rmacklem at uoguelph.ca>
> > wrote:
> >
> > > Ian Lepore wrote:
> > > > On Tue, 2016-03-22 at 13:46 +0000, Edward Tomasz Napierala wrote:
> > > > > Author: trasz
> > > > > Date: Tue Mar 22 13:46:01 2016
> > > > > New Revision: 297190
> > > > > URL: https://svnweb.freebsd.org/changeset/base/297190
> > > > >
> > > > > Log:
> > > > >   Wait for root mount tokens before showing the root mount
> > > > > prompt.
> > > > >   This restores the pre-r290196 behaviour, eliminating the need
> > > > > to
> > > manually
> > > > >   press '.' a couple of times to get USB to finish probing.
> > > > >
> > > > >   Note that there's still something wrong with the console
> > > > > (character
> > > > >   echoing doesn't quite work), and there's also a reported
> > > > > problem with
> > > > >   BHyVe, but those two don't seem related to the problem above.
> > > >
> > > > Just a datapoint on the echoing... it works fine on a serial
> > > > console,
> > > > it's been years since I've seen glitches at the mountroot prompt.
> > > >   So
> > > > the problem may be in vt or kbdmux.
> > > >
> > > When I had a console character echo problem (not serial), changing
> > > the
> > > clock source to RTC fixed it.
> > >
> >
> > I fixed this for the UART code a while ago in r260890. Maybe a
> > similar fix
> > is needed here? Basically, is the data being consumed by the
> > interrupt
> > handler before we can poll-read it...
> >
> > Warner
>
> Your cngrab fixes were only for uart?  Or was it some other fix?
>

Yes, I only fixed uart. But the cngrab interface is generic. I didn't have
issues with the console code at the time, but there's been two
elephants of change in that code since then, including a rewrite.
Perhaps some of that change broke what was working at the time
and needs to be fixed. Perhaps it was working only by accident, like
the UARTs were for years before I made the fix (well, working on and
off... something broke them a bit before I made the change). mountroot>
is so little used that things can break there and be broken for years
before people notice.

Warner