kopia lustrzana https://gitlab.com/sane-project/website
65 wiersze
3.2 KiB
HTML
65 wiersze
3.2 KiB
HTML
<!-- received="Wed Jul 22 04:08:44 1998 PDT" -->
|
||
<!-- sent="Wed, 22 Jul 1998 12:04:54 +0100 (BST)" -->
|
||
<!-- name="Jon Knight" -->
|
||
<!-- email="jon@net.lut.ac.uk" -->
|
||
<!-- subject="Re: developing plustek driver-- questions" -->
|
||
<!-- id="" -->
|
||
<!-- inreplyto="Pine.BSF.4.00.9807211607540.28577-100000@weevil.nac.net" -->
|
||
<title>sane-devel: Re: developing plustek driver-- questions</title>
|
||
<h1>Re: developing plustek driver-- questions</h1>
|
||
<b>Jon Knight</b> (<a href="mailto:jon@net.lut.ac.uk"><i>jon@net.lut.ac.uk</i></a>)<br>
|
||
<i>Wed, 22 Jul 1998 12:04:54 +0100 (BST)</i>
|
||
<p>
|
||
<ul>
|
||
<li> <b>Messages sorted by:</b> <a href="date.html#124">[ date ]</a><a href="index.html#124">[ thread ]</a><a href="subject.html#124">[ subject ]</a><a href="author.html#124">[ author ]</a>
|
||
<!-- next="start" -->
|
||
<li> <b>Next message:</b> <a href="0125.html">Mr Ada: "Re: ANNOUNCE: Plustek backend"</a>
|
||
<li> <b>Previous message:</b> <a href="0123.html">Abdelmajid Kamoun: "abdelmajid.kamoun@telnet.com.tn"</a>
|
||
<!-- nextthread="start" -->
|
||
<!-- reply="end" -->
|
||
</ul>
|
||
<!-- body="start" -->
|
||
On Tue, 21 Jul 1998, Antal Novak wrote:<br>
|
||
<i>> I don't like low-level OS-dependent... I've got FreeBSD and Linux, and I'm</i><br>
|
||
<i>> sure many others use different OSes; it would be too hard to develop</i><br>
|
||
<i>> low-level drivers for every OS. It's much simpler to have to manually</i><br>
|
||
<i>> specify an I/O port (and in the future, it might be auto-detectable)...</i><br>
|
||
<p>
|
||
But surely the low level driver code has to live in kernel space in order<br>
|
||
to make use of interrupts and not busy wait on the port? If the driver<br>
|
||
runs in user space, even if it is run as root and niced to a high priority<br>
|
||
it isn't going to get the snappy response that a kernel module could get?<br>
|
||
Especially seeing as the standard parallel port doesn't come with DMA so<br>
|
||
the processor has to be involved in every byte transfered? Whilst<br>
|
||
effectively hanging the machine whilst a scan is in progress is OK in the<br>
|
||
Windoze world, I don't think it sounds terribly nice for the UNIX<br>
|
||
environment. Or is there a cunning way of doing an efficient, platform<br>
|
||
independent, user space, low level driver that I've missed?<br>
|
||
<p>
|
||
Tatty bye,<br>
|
||
<p>
|
||
Jim'll<br>
|
||
<p>
|
||
#!/usr/bin/perl -- -Whois++-client-in-6-lines-of-Perl -Beat-that-Z39.50! <br>
|
||
use IO::Socket;sub w{$f=shift;$a{$f}=1;($h,$p,$q)=split("/",$f);$s=<br>
|
||
IO::Socket::INET->new(PeerAddr=>"$h:$p")||return;print $s "$q\r\n";while(<$s>)<br>
|
||
{next if(/^%/);if(/^# SERVER-TO-ASK/){while(<$s>){$x=$1 if/Name: (.*)\r\n$/;$y<br>
|
||
=$1 if/Port: (.*)\r\n$/;$f="$x/$y/$q";@j=(@j,$f)if(/^# END/&&!$a{$f})}}else{<br>
|
||
print}}close($s)}@j=shift;while(@j){w(pop(@j))}# whois++.pl host/port/query<br>
|
||
<p>
|
||
<p>
|
||
<pre>
|
||
--
|
||
Source code, list archive, and docs: <a href="http://www.mostang.com/sane/">http://www.mostang.com/sane/</a>
|
||
To unsubscribe: echo unsubscribe sane-devel | mail <a href="mailto:majordomo@mostang.com">majordomo@mostang.com</a>
|
||
</pre>
|
||
<!-- body="end" -->
|
||
<p>
|
||
<ul>
|
||
<!-- next="start" -->
|
||
<li> <b>Next message:</b> <a href="0125.html">Mr Ada: "Re: ANNOUNCE: Plustek backend"</a>
|
||
<li> <b>Previous message:</b> <a href="0123.html">Abdelmajid Kamoun: "abdelmajid.kamoun@telnet.com.tn"</a>
|
||
<!-- nextthread="start" -->
|
||
<!-- reply="end" -->
|
||
</ul>
|