Three quick points, mostly because I'm busy with my day job right now: (1) I'm trying to plan 2.5 changes which will remove the need for many of the unusual_devs.h entries. It's an uphill battle, but should improve interop by more closely following the behavior of Windows, which is what most of these units are tested against.
(2) I do not have an especially large crystal ball. For most devices, enumeration through all 15 (or so) combinations is the only way to determine the SubClass and Protocol overrides. It's highly risky, crash-inducing, and not for the faint of heart. I do not advocate it to anyone who would mind re-installing their OS. (3) Post the logs. Why does nobody turn on debugging and post the logs until they've been bugged about it for a while? And why does nobody look at the descriptive text strings in unusual_devs.h? And why don't people search with Google to discover that many Olympus cameras are very broken, and will only work if you comment out a key sanity check which is used to prevent data corruption? Folks, the data is all online... much of it is even in the list archives. Matt On Wed, Mar 06, 2002 at 03:04:03PM -0800, Stephen J. Gowdy wrote: > Hi David & Rich, > There is indeed a very large mix of users out there. I should also > say that I'm not involved (any more) with usb-storage development so Matt > should say what he wants. If he wishes to add a section to the web site > describing the various things you can do with the unusual device list I'll > add that. However, having someone iterate through all the options by > themselves is (perhaps) a waste of their time. If they post the debug log > files Matt (and others) can usually say what is wrong with a device. > Adding entries to this file (as you infer) doesn't help everyone > as it takes time for those changes (the good ones) to propagate to kernels > that are actually used. Of course, in a perfect world all devices would > follow the specs... > > regards, > > Stephen. > > On Wed, 6 Mar 2002, David Both wrote: > > > > > Let me tell you MY thoughts since I am the "user" in this case. > > > > The setup is this: I am an advanced user. I have certifications in OS/2, > > windoze, Solaris 7, and am an RHCE. I taught the RHCE course for the time > > I worked for Red Hat. > > > > I do shell and Perl and cgi scripting. I program occasionally in C or C++, > > but I am NOT a programmer. I have a little knowledge of programming > > structure and header files. I have recompiled the Linux kernel to remove > > unneeded modules. I have never attempted to change anything in the kernel > > code and then recompile to achieve a particular objective such as making a > > device work. > > > > All of that said, I could probably figure out how to do the necessary > > steps - with some help from my friends - without screwing up my system to > > the point I could not reboot and get back my original kernel. I think it > > would be a wonderful opportunity for me to learn new stuff. > > > > It would be great if there were an RPM or tarball I could install and > > forget about, but there isn't and aren't we talking about what makes Linux > > great right here - the ability for those of us who have a need and the > > time and the desire to get in and do what is necessary and learn something > > and to perhaps sooner or later contribute something back to the community? > > > > Now I realize from having spent many years at IBM as the lead OS/2 support > > person, that not all users should have the advice that has been dispensed > > to me. I know some folks who should be legally barred from having or using > > a computer. But if we found our way here and the advice is offered - > > perhaps with a caveat or two if there are gotchas - where is the harm. I > > have destroyed a whole lot of systems by doing less than compiling a > > module - and will certainly do so again, whether though fatfingers, > > ignorance, temporary insanity, or just plain stupidity. And I won't come > > back here to blame the person who gave me the advice because I made a > > misteak. > > > > Most of the problems I have had in the nearly 30 years I have been in this > > business have been of my own making and I have always been able to recover > > one way or another. > > > > I really do appreciate all of the advice and information. If I wished to > > wait, it looks like I could get a kernel upgrade from Red Hat at some > > point in the future and do this the easy way. I think I will stick around > > and try to learn something. > > > > On Wed, 6 Mar 2002 [EMAIL PROTECTED] wrote: > > > > > Date: Wed, 06 Mar 2002 16:27:17 -0500 > > > From: [EMAIL PROTECTED] > > > To: [EMAIL PROTECTED], [EMAIL PROTECTED] > > > Subject: Re: Re: [Linux-usb-users] RE: Olympus Camera > > > > > > I think I might be able to guess why you feel this way but could you expand on >your opinion? > > > > > > I'm looking at it from an advanced-user/non-developer stand point, I have a >piece of hardware I need to make work. I don't know enough to contribute to the code >but I can figure out enough to make it work within the framework of what the >developer's have built. I don't think the type of user who says, "How do I use usb?" >really poses a problem since they probably have no idea what a kernel header file >even is. However there are dangerous users like myself, who know just enough to >reverse engineer well documented kernel code. ;) > > > > > > So are users like myself doing a service by sorting out previously unknown >hardware and reporting the results or are we doing a disservice? I admit I don't >know the planning framework in usb-storage, is there an attempt to be device agnostic >or is the variability in products going to an ability to add new devices as they show >up? > > > > > > I'm not trying to start any flame war here I really do want to know so that I >don't unintentionally undermine the developers efforts. > > > > > > Rich > > > [EMAIL PROTECTED] wrote: > > > > I don't think we should be encouraging people to add the the unusual device >list. > > > > > > On Wed, 6 Mar 2002 [EMAIL PROTECTED] wrote: > > > > > > > I believe the faq and guide need a bit of expansion. I finally stumbled onto >what has to be done regarding many usb-storage devices on the working devices list. >Namely, you probably will have to add an entry to unusual_devs.h in the usb-storage >module code, then recompile and reinstall the module. I had to do this for a USB HD >I unwisely purchased without checking the list. The things that need to be included >in the new entry are the vendor and device id's, the minor and major device revision >numbers [I used 0x0001 and 0xffff - ie every possible one], give names for the driver >to post, then you have to define protocols and flags - that's the hard part since if >you know as little as I do, you have to just empirically figure them out. Look for >examples on the working devices list by searching for working usb-storage devices. >The protocol options are in usb.c, the flag options are in usb.h > > > > > > > CLIP > > > > > > > > > > > > > > > > -- > /------------------------------------+-------------------------\ > |Stephen J. Gowdy | SLAC, MailStop 34, | > |http://www.slac.stanford.edu/~gowdy/ | 2575 Sand Hill Road, | > | | Menlo Park CA 94025, USA | > |EMail: [EMAIL PROTECTED] | Tel: +1 650 926 3144 | > \------------------------------------+-------------------------/ > > > _______________________________________________ > [EMAIL PROTECTED] > To unsubscribe, use the last form field at: > https://lists.sourceforge.net/lists/listinfo/linux-usb-users -- Matthew Dharm Home: [EMAIL PROTECTED] Maintainer, Linux USB Mass Storage Driver YOU SEE!!?? It's like being born with only one nipple! -- Erwin User Friendly, 10/19/1998
msg02868/pgp00000.pgp
Description: PGP signature
