kopia lustrzana https://gitlab.com/sane-project/website
88 wiersze
4.7 KiB
HTML
88 wiersze
4.7 KiB
HTML
<!-- received="Sun Sep 7 04:02:06 1997 MST" -->
|
|
<!-- sent="Sun, 7 Sep 1997 09:06:37 +0200 (MET DST)" -->
|
|
<!-- name="Rogier Wolff" -->
|
|
<!-- email="R.E.Wolff@BitWizard.nl" -->
|
|
<!-- subject="Re: Backtracking in color scanning" -->
|
|
<!-- id="199709070706.JAA00963@cave.BitWizard.nl" -->
|
|
<!-- inreplyto="199709061845.LAA00773@hopper.mosberger" -->
|
|
<title>sane-devel: Re: Backtracking in color scanning</title>
|
|
<h1>Re: Backtracking in color scanning</h1>
|
|
<b>Rogier Wolff</b> (<a href="mailto:R.E.Wolff@BitWizard.nl"><i>R.E.Wolff@BitWizard.nl</i></a>)<br>
|
|
<i>Sun, 7 Sep 1997 09:06:37 +0200 (MET DST)</i>
|
|
<p>
|
|
<ul>
|
|
<li> <b>Messages sorted by:</b> <a href="date.html#141">[ date ]</a><a href="index.html#141">[ thread ]</a><a href="subject.html#141">[ subject ]</a><a href="author.html#141">[ author ]</a>
|
|
<!-- next="start" -->
|
|
<li> <b>Next message:</b> <a href="0142.html">Rogier Wolff: "Re: SANE and Mustek MFC-600S"</a>
|
|
<li> <b>Previous message:</b> <a href="0140.html">Jim McBeath: "UMAX on Sparc/Solaris 2.5"</a>
|
|
<li> <b>In reply to:</b> <a href="0137.html">David Mosberger-Tang: "Re: Backtracking in color scanning"</a>
|
|
<!-- nextthread="start" -->
|
|
<li> <b>Next in thread:</b> <a href="0113.html">Mathias Weigt: "Re: Backtracking in color scanning"</a>
|
|
<!-- reply="end" -->
|
|
</ul>
|
|
<!-- body="start" -->
|
|
David Mosberger-Tang wrote:<br>
|
|
<i>> </i><br>
|
|
<i>> >>>>> On Mon, 1 Sep 1997 13:55:42 +0200 (MET DST), <a href="mailto:R.E.Wolff@BitWizard.nl">R.E.Wolff@BitWizard.nl</a> (Rogier Wolff) said:</i><br>
|
|
<i>> </i><br>
|
|
<i>> Rogier> We could try to issue the next "read" AS FAST AS POSSIBLE,</i><br>
|
|
<i>> Rogier> discarding the data just read. This probably requires some</i><br>
|
|
<i>> Rogier> serious hacking.</i><br>
|
|
<i>> </i><br>
|
|
<i>> As of sane-0.62, this is already what's happening for Linux. That's</i><br>
|
|
<i>> the reason I introduced the asynchronous scsi interface to SANE. I'm</i><br>
|
|
<i>> sure it would be possible to shave off a few cycles here and there,</i><br>
|
|
<i>> but control-flow wise, the right thing is happening. If somebody gets</i><br>
|
|
<i>> serious about this, I'd recommend to first find out how much slack one</i><br>
|
|
<i>> has with issueing the next "read data" command. E.g., modify the</i><br>
|
|
<i>> backend to not transfer any data at all to the frontend and see</i><br>
|
|
<i>> whether that helps. Once you get to the point where no backtracking</i><br>
|
|
<i>> occurs, you can start inserting delays until backtracking starts</i><br>
|
|
<i>> occuring again. At that point, you should have a pretty good idea of</i><br>
|
|
<i>> what should be and can be done.</i><br>
|
|
<p>
|
|
The way I understand the Linux sg device, <br>
|
|
<p>
|
|
you issue the read-data command, write system call<br>
|
|
the low level driver transfers somewhere: read system call<br>
|
|
the data to the kernel-buffer. <br>
|
|
when done: transfer data to user-space read completes<br>
|
|
This is the first opportunity<br>
|
|
to issue the next "write call"<br>
|
|
with the next "read-data"<br>
|
|
command.<br>
|
|
<p>
|
|
At that moment there already has been a kernel->user copy. The time<br>
|
|
involved could already be enough, and disabeling that would require<br>
|
|
a few lines of change in the kernel. <br>
|
|
<p>
|
|
Roger.<br>
|
|
<p>
|
|
P.S. David: I remember replying to messages posted by you last week,<br>
|
|
and they went to the list. Nowadays they would go to you<br>
|
|
personally....<br>
|
|
<p>
|
|
<pre>
|
|
--
|
|
** <a href="mailto:R.E.Wolff@BitWizard.nl">R.E.Wolff@BitWizard.nl</a> ** +31-15-2137555 ** <a href="http://www.BitWizard.nl/">http://www.BitWizard.nl/</a> **
|
|
Florida -- A 39 year old construction worker woke up this morning when a
|
|
109-car freigt train drove over him. According to the police the man was
|
|
drunk. The man himself claims he slipped while walking the dog. 080897
|
|
<p>
|
|
<pre>
|
|
--
|
|
Source code, list archive, and docs: <a href="http://www.azstarnet.com/~axplinux/sane/">http://www.azstarnet.com/~axplinux/sane/</a>
|
|
To unsubscribe: mail -s unsubscribe <a href="mailto:sane-devel-request@listserv.azstarnet.com">sane-devel-request@listserv.azstarnet.com</a>
|
|
</pre>
|
|
<!-- body="end" -->
|
|
<p>
|
|
<ul>
|
|
<!-- next="start" -->
|
|
<li> <b>Next message:</b> <a href="0142.html">Rogier Wolff: "Re: SANE and Mustek MFC-600S"</a>
|
|
<li> <b>Previous message:</b> <a href="0140.html">Jim McBeath: "UMAX on Sparc/Solaris 2.5"</a>
|
|
<li> <b>In reply to:</b> <a href="0137.html">David Mosberger-Tang: "Re: Backtracking in color scanning"</a>
|
|
<!-- nextthread="start" -->
|
|
<li> <b>Next in thread:</b> <a href="0113.html">Mathias Weigt: "Re: Backtracking in color scanning"</a>
|
|
<!-- reply="end" -->
|
|
</ul>
|