Jesse Pollard wrote:
> Originally, (wayback machine on) this was handled by a pull-up resistor
> in the parallel interface, on the "off-line" signal. ANY time the printer
> was powered off, set offline, or cable unplugged, the "off-line" signal
> was raised by the pull-up. No data lost.
>
> Now the parallel interface is bidirectional, and can have multiple devices
> attached - this "fix" cannot be used. The interface is now more of a
> buss than a single attached interface, and signals from a missing device
> (powered off or disconnected) are floating. They may float high or low,
> and depending on the environment (and which end of the cable is unplugged)
> any thing in between.
Not true. Electrical characteristics for parallel port implementations/cards
differ wildly, nevertheless most implementations have:
- data lines: bidirectional (see datasheets)
- signal lines: see datasheets, never floating !
Floating signal lines are a silicon bug/bad engineering and have nothing
to do with bidirectional interfaces !
Nowadays most integrated chips have internal signal line pull-ups internally, e.g.
W83877TF says:
-BUSY, ACK, PE, SLCT, ERR:
TTL level input pin. This pin is pulled high internally.
-AFD, STB, INIT, SLIN
Open-drain output pin with 12 mA sink capability. Pulled up internally.
-Data lines:
TTL level bi-directional with 24 mA source-sink capability.
Of course I would expect add-in cards to exist, with not so sophisticated chipsets
and makers that have "forgotten" external pull-ups for economical reasons (2 cents :-)
We should NOT care for broken hardware !!! I haven't seen any of these yet, even.
On the other hand printer implmentations vary wildly, too.
LJ1100: leave signal lines alone if powered off (0x7f)
i.e. signal printer-not-ready ack-active out-of-paper
DJ500: signal printer-error and off-line when powered off (0x87) !!!
=> Linux would dump data on this printer, if switched off.
I think the current linux lp code tries to handle exotic/weird printers
gracefully and leaves mainstream printers and users alone.
-
Gunther
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sun Jan 07 2001 - 21:00:19 EST