My Microtek backend (for a Scanmaker E6) keeps crashing, and has for a
long time. It's been a long time because I rarely use the scanner. For
the same reason, I can't pinpoint when it started failing.
Here's the GDB backtrace of a simple "scanimage" command:
...bla...bla...bla...
Reading symbols from /usr/local/lib/sane/libsane-net.so.1...done.
Reading symbols from /lib/libnss_files.so.2...done.
#0 0x40207ed6 in sort_values (result=0xbfff7178, scanline=0xbfff7218,
pix=0)
at microtek.c:2147
2147 for (i=0; i<STRIPS; i++) result[i] = (scanline[i])[pix];
(gdb) bt
#0 0x40207ed6 in sort_values (result=0xbfff7178, scanline=0xbfff7218,
pix=0)
at microtek.c:2147
#1 0x40207f68 in calc_calibration (caldata=0x805f618 "",
scanline=0xbfff7218,
pixels=5110) at microtek.c:2177
#2 0x40208359 in do_real_calibrate (s=0x805a7a0) at microtek.c:2276
#3 0x40208696 in finagle_precal (handle=0x805a7a0) at microtek.c:2387
#4 0x4020b5d4 in sane_start_guts (handle=0x805a7a0) at microtek.c:3688
#5 0x4020bc0a in sane_microtek_start (handle=0x805a7a0) at
microtek.c:3853
#6 0x400177aa in sane_dll_start (handle=0x805ef20) at dll.c:824
#7 0x40017a0e in sane_start (h=0x805ef20) at dll-s.c:46
#8 0x804a1d2 in scan_it () at scanimage.c:795
#9 0x804b6a4 in main (argc=1, argv=0xbffff894) at scanimage.c:1458
It's a memory access problem. "print i" says 2, "print pix" says 0, but
"print scanline[i][pix]" says "Cannot access memory at address 0x17".
(scanline[i] == 0x17)
System vital stats:
Linux 2.2.16 SMP (2 Celeron processors, not overclocked)
Scanner attached to an Adaptec AHA-2930U2
SANE 1.0.3 built from source, default configure settings
Problem happens with scanimage, xscanimage and xsane
Another scanner (a film scanner not compatible with SANE) on the same
SCSI bus is working just fine.
I can hear the scanner's lamp mechanism move just a bit right before the
crash. Just a little bump, then it returns home.
I don't have any way of testing it under Windows, as this is a
Linux-only machine, and I don't wanna move the scanner to another
machine. If it comes down to it I guess I can, but it'd be a last
resort.
Clues anyone?
-- = Warren -- ICBM Address: 36.8274040 N, 108.0204086 W, alt. 1714m-- Source code, list archive, and docs: http://www.mostang.com/sane/ To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com
This archive was generated by hypermail 2b29 : Tue Aug 15 2000 - 18:52:53 PDT