Tim Jones wrote:
> Wayne,
>
> I've interspersed my new comments within your reply, but I want you to
> understand that my original answers were designed to help you get your
> HP drive working under Linux and to help you become satisfied with your
> purchase. I am sorry that you seem to feel that it was a personal
> attack.
>
> Wayne Horner wrote:
> >
> > Tim Jones wrote:
> >
> > > Wayne Horner wrote:
> > > >
> > > > I have an hp t20i attached to linux mandrake 6.0 kernel 2.2.9.
> > > > The tape drive appears to work. However it wont accept any change in
> > > > blocksize other than 512.
> > > > setblk 0 gives Illegal bits in request field
> > > > setblk 512 works
> > > > setblk 1024 or more gets illegal block size.
> > >
> > > The t20i is NOT a variable block capable drive. It works with 512 byte
> > > blocks - period. This drive is a QIC device and gives you what it gives
> > > you - nothing more. So, all setblk commands will fail except 512.
> > >
> >
> > Without the docs I don't know what the drive 'gives me'. It could be some
> > proprietary bits in the scsi command that is causing the problem. Some drives
> > won't change blocksize unless you are at the beginning of a blank tape. One
> > doesn't know without the scsi spec.
>
> Actually, the QIC-157 specs are publically available from
> http://www.qic.org/. All of the functions of a TR-4 or TR-5 drive are
> defined there.
>
> > > > So HP is not interested in supporting us. I will return their $500 tape
> > > > drive.
> > > > Beware....
> > >
> > > Actually, HP is very interested in supporting us. It's just that they
> > > don't give away tech docs to just anybody.
> >
> > I am not 'just anybody'.... I am the customer... the guy who gave them $500.
> > I wanted to purchase the scsi spec. I didn't ask for the source code of the
> > firmware.
>
> Again, you don't have to purchase a SCSI spec - it's available at
> qic.org.
>
> To a company like HP (and others), you are not a recognized developer
> entity. I wasn't trying to imply that you were "nobody," just that you
> were not someone that HP recognizes as a developer. Please re-read my
> final comments. HP is coming around, it's just that they are like a
> very large ship - takes it a while to change course safely.
>
The qic-157 spec is a generic spec for the common-denominator commands. It doesn't
tell vendor specific information like whether the HP T20 supports variable block
mode or not. Most of the information is commonly available so why does the HP
support department consider this classified information? Would you buy a car from a
company with this attitude?
"Sorry sir, all you need to know is where the gas goes in. If you want to
change the spark plugs then you will have
to take it to one of our 'recognized mechanics'. I am sure he will give you a
fair deal"
>
> >
> > > These drives work great with
> > > the existing drivers, so there's really no requirement for you to write
> > > a new driver.
> >
> > Sounds like a true MICROSOFT-APPLE "closed source" point of view..... your
> > computer is just a toaster plug it in... you can buy anything you (the
> > market) need from us.
>
> Not my point - since Kai and team (see /usr/src/linux/drivers/scsi/st.c)
> have done a great job of implementing the SCSI spec in the driver, the
> device will work. I was not implying that you shouldn't be able to
> write a driver, just that you don't have to in this case. Heck, you can
> rewrite the whole of Linux, if you like. Just not what I feel is a good
> way to spend time. Hey, but it's your time - just don't make generic
> statements for all of us.
>
> > Keep in mind the whole "Open Source" environment means that I decide if I
> > need a new driver not you!
> > Its arrogant to suggest otherwise. Whose side are you on?
>
> I *WAS* on the side of helping you get the drive working. Now, I'm not
> so sure. I was not saying that you shouldn't write a new driver, just
> that it's not required to use the HP T20I.
>
> >
> > > Keep in mind that this whole "Open Source" environment is
> > > totally new to them. They have made leaps and bounds in accepting the
> > > Linux OS,
> >
> > HP may accept Linux.... but Linux wont accept closed, proprietary products.
>
> Again, I stand by my comments below. Please re-read them and realize
> that great steps have been taken and that one bad attitude can actually
> damage the good will that's been developed.
>
> And, don't be so bold as to assume for the entire Linux community as you
> have - I am part of that community and I see the place for closed source
> elements.
>
> > > but there is still a long road ahead until all of the support
> > > groups are fully online with Linux.
> > >
> > > Additionally, we've just opened channels with HP's Colorado drive group
> > > to get all of the TR-4 and TR-5 drives certified under Linux.
> > >
> > > Please be patient and remember that getting a large corporation like HP
> > > to move into the Linux arena is a lengthy process. We'll get 'em there,
> > > but it takes time.
> > >
> > > Tim
