On Mon, 24 Aug 1998, Jon Wikne wrote:
> I recently discovered that the initial "fast processor hang" problem and
> color problems with my Mustek Paragon 600 II N scanner / AB306N interface
> card reappeared in Sane 0.74.
I noticed this also, since I built a new SANE (0.74, the last one
I had was 0.68 !) a few days ago. I will send in a patch for this
to be incorporated into the next version (0.75 I hope).
> The files backend/mustek.c and sanei/sanei_ab306.c still come without
> the nice patches provided by Henning Meier-Geinitz in April.
> Shouldn't these, or something similar, come into the official release
> soon? New users are likely to have a headache over these bugs, like I
> originally did before Henning had the same problem, and eventually
> posted the solution....
The fast-processor-hang actually is a bug in the firmware:
the controller sets the "I am ready" line just before doing some
other setup (setting internal flags that tell the interrupt
routine (of the u-controller) what to do), so we have to wait
for 10ms (some 10us would do) before actually transmitting/receiving
data.
With my firmware (labelled "2.00" but surely older than the "1.xx"
that most other people have), the color correction works fine
(with line-distance-fix enabled). Possibly, another subdivision
and case handling will be necessary here :-(. So Please, if
anyone with a 600 II N reads this:
--> Can anybody with a 600 II N with Firmware version 1.xx <--
--> send me a binary image of the EPROM in the scanner ? <--
--> So I could try both versions in my scanner. Thanks ! <--
> Another thing: When using xscanimage with suid root to get access
> to the port, it seems that file ownership for the output files created
> also defaults to root. This is sometimes annoying, although not serious.
Is it still the case with 0.74 ? I think I had to remove the
file /tmp/out.pnm as root before being able to scan into that file,
but may be I have missed to set the mode 4755 before doing this...
> Does anyone know about an easy cure, or do I have to plunge into the
> source code myself when (if!) I find the time?
The process writing the output file should setuid() to the user
before opening the file, but I fear that this is the parent and
so it could probably no longer use inb/outb to access the interface
- or are the ioperm()'s that were set before still valid ?
Greetings,
Andreas.
-- Source code, list archive, and docs: http://www.mostang.com/sane/ To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com