Re: port-amd64/56987 (Certain usb devices can no longer be mounted on -current)

[Robert Elz <kre%munnari.OZ.AU@localhost>]

The following reply was made to PR port-amd64/56987; it has been noted by GNATS.

From: Robert Elz <kre%munnari.OZ.AU@localhost>
To: gnats-bugs%netbsd.org@localhost
Cc: 
Subject: Re: port-amd64/56987 (Certain usb devices can no longer be mounted on -current)
Date: Wed, 31 Aug 2022 18:19:55 +0700

     Date:        Tue, 30 Aug 2022 16:50:02 +0000 (UTC)
     From:        mlh%goathill.org@localhost (MLH)
     Message-ID:  <20220830165002.2D1E41A923A%mollari.NetBSD.org@localhost>
 
 
   |  skrll%NetBSD.org@localhost wrote:
   |  > Please bisect it further.
   |  
   |  ?
 
 What Nick (skrll@) asked you to do, is to try running kernels from
 the period between when you know it worked, and when you are seeing
 the problem, and find out exactly when things changed.
 
 Since you are seeing an issue, and most of us are not, you might be
 the only person who can do that.
 
 It worked in mid January, fails in late August, so try something from
 early May.  If that works, then try something from late June or
 early July, and if that works, from early August, ... keep getting
 to a narrower and narrower time range until you are comparing systems
 from sources just an hour or so apart, where the earlier one works,
 and the later one fails.
 
 Obviously (I hope) if early May failed, you instead pick a time between
 Mid Jan and early May - so probably about the end of the first week in
 March (you don't have to be highly accurate in picking the mid point
 of the time interval - but the closer you are the less iterations you're
 likely to need to make, unless you get lucky, and that rarely seems to
 happen).
 
 So each time a mid point test succeeds (sd0 works) you pick half way
 between that time, and the current end (a known failed system).  Each
 time it fails, you pick the mid point between the current beginning
 (a known working version) and the current one (now known to fail).
 You now have a new (begin, end) time pair (where one of those will be
 the mid point of the last test, and the other is unchanged from that
 one), you pick somewhere approximately at the middle of that range,
 and try again.   This is bisection.   (A similar technique is sometimes
 used in mathematics to solve problems that are difficult other ways.)
 
   |  Why does it no longer configure to use sd0 as it used to :
 
 This we don't know - but I'm running 9.99.99 and my sd devices all
 work just fine.   So it is probable that only you can work this out.
 
 Once you find just when (as close as you can get) the breaking change
 happened, let us know, and someone might be able to tell what the problem
 might be, just from looking at what changed about that time.
 
 kre