From mboxrd@z Thu Jan 1 00:00:00 1970 From: Felipe Balbi Subject: Re: Enabling MUSB support Date: Fri, 29 Aug 2008 23:59:21 +0300 Message-ID: <20080829205918.GC14130@frodo> References: <1f11a5490808270626t5b4aa599i18b730d12a6c4667@mail.gmail.com> <200808291305.31639.david-b@pacbell.net> <20080829202114.GA14130@frodo> <200808291350.17181.david-b@pacbell.net> Reply-To: me@felipebalbi.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 8BIT Return-path: Received: from ns1.siteground211.com ([209.62.36.12]:55290 "EHLO serv01.siteground211.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751281AbYH2U7n convert rfc822-to-8bit (ORCPT ); Fri, 29 Aug 2008 16:59:43 -0400 Content-Disposition: inline In-Reply-To: <200808291350.17181.david-b@pacbell.net> Sender: linux-omap-owner@vger.kernel.org List-Id: linux-omap@vger.kernel.org To: David Brownell Cc: me@felipebalbi.com, felipe.balbi@nokia.com, ext Ashwin Bihari , linux-omap@vger.kernel.org On Fri, Aug 29, 2008 at 01:50:16PM -0700, David Brownell wrote: > Right, that's why I thought I'd chase the easier host-side issue > first. Must have "apt-get install ..." working after all!! hmm... sure. > Were those self-powered devices, or bus-powered? Descriptors > are not entirely trustworthy about current usage. Sometimes > they represent worst-case; sometimes it's best-case. bus-powered. > > Do you have the BLACKLIST_HUB set? (just to be sure). > > No, that's not set. yet another new bug so, is it a full or high speed hub ? > IMO it's the silicon that's buggy. The specific window I kept seeing > was between VBUS_ON and the 100msec later where it's expected to be > stable. During that window, VBUS is allowed to not be stable ... which > is good, because that's when current spikes happen while charging the > various capacitances involved. (Cable, caps at either end, etc.) > > I saw TUSB6010 reporting VBUS_ERR during that 100msec window ... and > same thing with DaVinci. Both needed retry logic to cope with the > inappropriate VBUS fault reports. The thing in common is the silicon > IP, not the hardware. Sure, musb is damn buggy and looks like it won't be fixed after all. At least we're sure it's a silicon bug. I wonder if mentor has plans to fix it ?!? A better dma engine would also be welcome, the whole "mode" thingy is a mess. btw, do you need me to help you debugging anything ? I can to the office tomorrow and debug some stuff there if it's needed, otherwise I'll get back to this only on monday. -- balbi