Bug#296687: Problem Fixed
On Tue, Mar 01, 2005 at 11:55:42AM +0900, Horms wrote:
> On Fri, Feb 25, 2005 at 12:07:20AM -0800, David Lawyer wrote:
> > I fixed the problem by setting the parallel port in the BIOS from ECP to
> > SPP. Then the driver does software handshaking and my old printer works
> > OK. So this is a problem with documentation. I just emailed the
> > authors of the kernel docs on the parallel port and told them about this
> > fix. So you could close this bug if you want to, although it would be
> > nice if somehow the software configured it all.
>
> Hi David,
>
> are you suggesting that there should be a kernel option
> to restrict the modes that the parallel port driver will
> operate in - so that if you have older hardware it will only
> use modes supported by the other end. Or are you suggesting
> that the driver should autodetect this somehow. My knowledge
> of the relevant specs are weak (non-existant), but the latter
> sounds like it might not be possible.
I think the latter may be possible if one has a modern (not over 20
years old) parrallel port on their PC (like I do). There's a spec for
negotiation between the two ports over the parallel cable. In my case,
there would be no response from my printer and it would then be assumed
by my PC that the printer port doesn't meet IEEE 1284 specs and thus use
the old Centronics protocol known as SPP. But I don't know how the port
tells that to Linux. The old protocol requires driver handshaking for
every byte sent.
I think that the BIOS allows setting SPP for cases where the software
doesn't know about ECP. Since Linux knows about ECP and since ECP can
fallback to SPP mode, it should have worked in ECP mode. So I now think
it's a bug and needs to be fixed even though I found a work around to
get my printer printing.
> In any case, have you considered reporting this to LKML and
> the maintainers?
How do I do this?
> --
> Horms
>
David Lawyer
Reply to: