Re: HEADSUP usb2 (usb4bsd) to become default in 2 weeks.

In message: <20081229181044.GA78575_at_dragon.NUXI.org>
            "David O'Brien" <obrien_at_freebsd.org> writes:
: On Mon, Dec 29, 2008 at 03:19:44AM -0800, Alfred Perlstein wrote:
: > * David O'Brien <obrien_at_freebsd.org> [081229 03:10] wrote:
: > > [ Note: replies directed to list ]
: > > 
: > > On Fri, Dec 26, 2008 at 05:23:58PM +0100, Hans Petter Selasky wrote:
: > > > There are some debugging sysctls which you can try to tune/increase:
: > > > 
: > > > hw.usb2.pr_recovery_delay: 250
: > > > hw.usb2.ss_delay: 0
: > > > hw.usb2.ehci.no_hs: 0 // this value is a boolean
: > > 
: > > Please.  Can we quickly rename this before it gets too ingrained?
: > > 
: > > "usb2" just implies too much the "USB 2.0 specification of April 2000"
: > > vs. the 2nd USB implementation within FreeBSD.
: > > 
: > > To stop confusion between USB 2.0 spec and this work, can it be renamed
: > > to "hpsusb", "usbhps", "usb4bsd", or even the dreaded over-used "usbNG"?
: > 
: > David, in about two weeks we're going to make "usb2" or "hpsusb" ( :) )
: > the default, if things go well we're planning on cleaning it up
: > and making it just "usb" within a few weeks after that.

I'm very nervous about this.  I have several disk enclosure boxes that
flat out don't work on usb2 that work flawlessly on the current usb
stack.  I've not had time to totally characterize things.

Also, the rate of patches into perforce seems high and good.  However,
they aren't making it into svn quickly.  I'd rather see them in the
tree and tested well in advance of making usb2 just usb.

: > Let's not do this now though as we'll just wind up having to
: > do it twice.  Rest assured we will get to it shortly.
: 
: Thanks Alfred.
: Of course I knew it was about to become the default, but wasn't clear as
: to what it would look like as the default.
: 
: Like others, I tried to move to a machine to "hpsusb" and didn't find it
: so turn-key.

I've found that some things work a lot better, while others are a
regression.  When your object is to burn DVDs, it can be too large a
detour to catalog all the problems that I've hit.  I should start
doing that, however, since it is clear things are moving ahead at
breakneck speeds :(

Warner