Nicolas Souchu (
Fri, 11 Dec 1998 22:38:26 +0000

On Thu, Dec 10, 1998 at 07:09:12AM +0800, David Campbell wrote:
>> >The port->ops structure has changed. epp_write_block and epp_read_block
>> >now provide a callback, whereas previously they didn't. ('ECP+EPP'
>> >chipsets can DMA EPP.)
>> ^^^^^^^
>> Ouah! I haven't ever thought about this. It would really improve the
>> ZIP driver. Is this particular to some SMC like chips or just to one or more
>> chips?
>> David C.: did you ever try this before?
>Nope, lack of suitable test hardware.
>I *think* my new mainboard is capable of ECP+EPP, but as Nicolas should know
>ECP+EPP is one of those gray areas of parallel port hardware where no-one
>really knows what is truely going on. It took me several months before I
>figured out what a little cryptic "post-it" note that SMC left in one of their
>pdf files actually meant (EPP writes go via the FIFO, EPP reads are direct).
>Without the ECP sync routine, EPP under ECP+EPP is impossible.
>The latest pdfs I have from SMC are lacking the "post-it" note or any reference
>to a link between EPP and the FIFO. Could SMC have changed the EPP so it does
>not use the FIFO? Or more likely forgot to add it to the current pdfs.

I've SMC66x datasheets here, nothing about this.

>At this stage someone (probably myself) need to check and verify if EPP DMA is
>possible for the ZIP drive.

I'm not keen on undocumented features. How would you test this?

-- /
FreeBSD - Turning PCs into workstations -

-- To unsubscribe, send mail to: -- -- with the single word "unsubscribe" in the body of the message. --

This archive was generated by hypermail 2.0b3 on Wed 30 Dec 1998 - 10:18:53 EST