[sane-devel] Scanner Button Daemon [scanbd]: high CPU utilization
Rastislav David
rasto at ynet.sk
Tue Nov 29 08:13:11 UTC 2011
Hello Wilhelm,
correction to copy section in action.script
I did last minute change to copy secion and did not test copy before posting thus copy will not work.
There is missing space "--depth 16--format=tiff" and lp command does not process tiff images.
I'm not sure whether it's only my debian box or general issue.
However copy section which works and is tested looks like this:
copy)
logger -t "scanbd: $0" "$SCANBD_ACTION - scanning"
# scanning with 300dpi grayscale 16bit and converting into postscript
scanimage -d $SCANBD_DEVICE --resolution 300 --depth 16 | convert -density 300 - $scan_dir/$filename.ps
logger -t "scanbd: $0" "$SCANBD_ACTION - printing"
# printing postscript
lp $scan_dir/$filename.ps
# deleting postscript
rm $scan_dir/$filename.ps
I'm contemplating, that conversion and printing will be caried out outside of this script as it is blocking scanning.
My debian NAS has ARM 500Mhz, which is not much for conversion tasks and usually takes about 5 minutes to copy A4 size paper.
Rastislav
----- Wilhelm <wilhelm.meier at fh-kl.de> wrote:
> Hello Rastislav,
>
> just added your script to a contrib-directory of the repository.
>
> Thank you,
>
> Wilhelm
>
> Am 25.11.2011 23:07, schrieb Rastislav David:
> > Hello Wilhelm,
> >
> > scanbd starts fine and looks to be ok now on ARM and Debian Wheezy.
> >
> > I've unplugged and plugged back scanner and it now reacts to button presses.
> >
> > Thank You
> >
> > Rastislav
> >
> > I'm including my action.script which performs different tasks based on
> > button for somebody to reuse.
> >
> > scan_dir=/var/spool/scan
> > datetime=`date +%F_%H%M%S`
> > filename=scan-$datetime
> > send_to=machanska at seznam.cz
> > #send_to=rasto at ynet.sk
> >
> > #logger -t "scanbd: $0" "Begin of $SCANBD_ACTION"
> >
> > case $SCANBD_ACTION in
> > pdf)
> > logger -t "scanbd: $0" "$SCANBD_ACTION - scanning and converting to ps"
> > scanimage -d $SCANBD_DEVICE --mode Color --resolution 150 --depth 16 |
> > convert -density 150 - $scan_dir/$filename.ps
> > logger -t "scanbd: $0" "$SCANBD_ACTION - converting ps to pdf"
> > ps2pdf -sPAPERSIZE=a4 $scan_dir/$filename.ps $scan_dir/$filename.pdf
> > rm $scan_dir/$filename.ps
> > ;;
> > scan)
> > scanimage -d $SCANBD_DEVICE --resolution 150 --mode Color --depth 16
> > --format=tiff> $scan_dir/$filename.tiff
> > ;;
> > copy)
> > logger -t "scanbd: $0" "$SCANBD_ACTION - scanning"
> > scanimage -d $SCANBD_DEVICE --resolution 300 --depth 16--format=tiff>
> > $scan_d
> > ir/$filename.tiff
> > # | convert -density 300 - $scan_dir/$filename.ps
> > logger -t "scanbd: $0" "$SCANBD_ACTION - printing"
> > lp $scan_dir/$filename.tiff
> > rm $scan_dir/$filename.ps
> > ;;
> > email)
> > logger -t "scanbd: $0" "$SCANBD_ACTION - scanning and converting to ps"
> > scanimage -d $SCANBD_DEVICE --mode Color --resolution 150 --depth 16|
> > convert -density 150 - $scan_dir/$filename.ps
> > logger -t "scanbd: $0" "$SCANBD_ACTION - converting ps to pdf"
> > ps2pdf -sPAPERSIZE=a4 $scan_dir/$filename.ps $scan_dir/$filename.pdf
> > logger -t "scanbd: $0" "$SCANBD_ACTION - sending pdf"
> > (cat<<EOCAT
> > Document scanned on Canon LiDE 110
> > EOCAT
> > uuencode $scan_dir/$filename.pdf $scan_dir/$filename.pdf ) | mail -a
> > "From: Canon scanner<some at email.sk>" -s "Document $filename" $send_to
> > rm $scan_dir/$filename.ps
> > ;;
> > *)
> > logger -t "scanbd: $0" "Unhandled action request."
> > ;;
> > esac
> >
> > On 25/Nov/11 5:27 PM, Wilhelm wrote:
> >> Hi Rastislav,
> >>
> >> Am 25.11.2011 13:49, schrieb Rastislav David:
> >>>
> >>> ----- Wilhelm<wilhelm.meier at fh-kl.de> wrote:
> >>>> Hi Rastislav,
> >>>>
> >>>> (sorry for calling you wrong last time)
> >>>>
> >>>> Am 25.11.2011 10:35, schrieb Rastislav David:
> >>>>> Hello Wilhelm,
> >>>>>
> >>>>> thank You for responding so quickly.
> >>>>
> >>>> you're welcome
> >>>>
> >>>>>
> >>>>> I've closed point 2) where I've found that button called Pdf is
> >>>>> found with filter file.
> >>>>
> >>>> good to here!
> >>>>>
> >>>>> I've attached scanbd.conf in my last email but I guess it didn't go
> >>>>> through.
> >>>>
> >>>> looks good.
> >>>>
> >>>> I think there problem is, that with debian-testing we get a new libudev
> >>>> with a non-blocking behaviour for the udev_monitor_receive_device()
> >>>> function.
> >>>>
> >>>> I'll fix that upstream, when I discover how to switch back to the
> >>>> blocking behaviour.
> >>>>
> >>>> In the mean time you can compile sanbd without libudev as said in the
> >>>> last email.
> >>>
> >>> Yes.
> >>> Compiled without UDEV and HAL now scanbd works as expected.
> >>> scanbd is now behaving nicely with prety low CPU utilization :)
> >>>
> >>> Your guess at libdev was most likely correct.
> >>
> >> ok, sounds good.
> >>
> >> I just committed a fix regarding this problem. I can confirm that this
> >> problem exists on debian-testing using libudev-dev-173 (but not on
> >> ubuntu-oneiric using libudev-dev-172).
> >>
> >> Please test this new version on your system too with your scanner and
> >> please send me a feedback.
> >>
> >> Is there anyone to test this on ubuntu precise?
> >>
> >>>
> >>>>
> >>>> So, stay tuned.
> >>>
> >>> Fingers crossed.
> >>>
> >>>>
> >>>>>
> >>>>> Here is scanbd.conf I'm using now:
> >>>>>
> >>>>> global {
> >>>>> debug = true
> >>>>> debug-level = 1
> >>>>> user = saned
> >>>>> group = scanner
> >>>>> saned = "/usr/sbin/saned"
> >>>>> saned_opt = {}
> >>>>> saned_env = { "SANE_CONFIG_DIR=/etc/scanbd" }
> >>>>> scanbuttond_backends_dir =
> >>>>> "/usr/local/etc/scanbd/scanbuttond/backends"
> >>>>> timeout = 400
> >>>>> pidfile = "/var/run/scanbd.pid"
> >>>>> environment {
> >>>>> device = "SCANBD_DEVICE"
> >>>>> action = "SCANBD_ACTION"
> >>>>> }
> >>>>> function function_knob {
> >>>>> filter = "^message.*"
> >>>>> desc = "The value of the function knob / wheel /
> >>>>> selector"
> >>>>> env = "SCANBD_FUNCTION"
> >>>>> }
> >>>>> function function_mode {
> >>>>> filter = "^mode.*"
> >>>>> desc = "Color mode"
> >>>>> env = "SCANBD_FUNCTION_MODE"
> >>>>> }
> >>>>> multiple_actions = true
> >>>>> action scan {
> >>>>> filter = "^scan.*"
> >>>>> numerical-trigger {
> >>>>> from-value = 1
> >>>>> to-value = 0
> >>>>> }
> >>>>> desc = "Scan to file"
> >>>>> script = "/etc/scanbd/action.script"
> >>>>> }
> >>>>> action email {
> >>>>> filter = "^email$"
> >>>>> string-trigger {
> >>>>> from-value = ""
> >>>>> to-value = "^email.*"
> >>>>> }
> >>>>> desc = "Scan to email"
> >>>>> script = "/etc/scanbd/action.script"
> >>>>> }
> >>>>> action copy {
> >>>>> filter = "^copy$"
> >>>>> string-trigger {
> >>>>> from-value = ""
> >>>>> to-value = "^copy.*"
> >>>>> }
> >>>>> desc = "Copy to printer"
> >>>>> script = "/etc/scanbd/test.script"
> >>>>> }
> >>>>> action pdf {
> >>>>> filter = "^file$"
> >>>>> numerical-trigger {
> >>>>> from-value = 1
> >>>>> to-value = 0
> >>>>> }
> >>>>> desc = "Pdf"
> >>>>> script = "/etc/scanbd/action.script"
> >>>>> }
> >>>>> action preview {
> >>>>> filter = "^preview$"
> >>>>> numerical-trigger {
> >>>>> from-value = 1
> >>>>> to-value = 0
> >>>>> }
> >>>>> desc = "Preview"
> >>>>> script = "/etc/scanbd/action.script"
> >>>>> }
> >>>>> action globaltest {
> >>>>> filter = "^message.*"
> >>>>> desc = "Test (print all env vars)"
> >>>>> script = "/etc/scanbd/test.script"
> >>>>> }
> >>>>> }
> >>>>> device fujitsu {
> >>>>> filter = "^fujitsu.*"
> >>>>> desc = "Fujitsu: Test"
> >>>>> function function_knob {
> >>>>> filter = "^function.*"
> >>>>> desc = "The value of the function knob / wheel /
> >>>>> selector"
> >>>>> env = "SCANBD_FUNCTION"
> >>>>> }
> >>>>> action paperload {
> >>>>> filter = "^page-loaded.*"
> >>>>> desc = "page loaded"
> >>>>> numerical-trigger {
> >>>>> from-value = 0
> >>>>> to-value = 1
> >>>>> }
> >>>>> script = ""
> >>>>> }
> >>>>> action paperunload {
> >>>>> filter = "^page-loaded.*"
> >>>>> desc = "page unloaded"
> >>>>> numerical-trigger {
> >>>>> from-value = 1
> >>>>> to-value = 0
> >>>>> }
> >>>>> script = ""
> >>>>> }
> >>>>> action coveropen {
> >>>>> filter = "^cover-open.*"
> >>>>> desc = "cover open"
> >>>>> numerical-trigger {
> >>>>> from-value = 0
> >>>>> to-value = 1
> >>>>> }
> >>>>> script = ""
> >>>>> }
> >>>>> action powersave {
> >>>>> filter = "^power-save.*"
> >>>>> desc = "power save"
> >>>>> numerical-trigger {
> >>>>> from-value = 0
> >>>>> to-value = 1
> >>>>> }
> >>>>> script = ""
> >>>>> }
> >>>>> }
> >>>>> device snapscan {
> >>>>> filter = "(.*snapscan.*|.*Perfection.*)"
> >>>>> desc = "Epson"
> >>>>> action test {
> >>>>> filter = "^message.*"
> >>>>> desc = "Test"
> >>>>> script = "/usr/local/etc/scanbd/example.script"
> >>>>> }
> >>>>> action web {
> >>>>> filter = "^web.*"
> >>>>> desc = "Web"
> >>>>> script = "/usr/local/etc/scanbd/example.script"
> >>>>> }
> >>>>> action send {
> >>>>> filter = "^send.*"
> >>>>> desc = "Send"
> >>>>> script = "/usr/local/etc/scanbd/example.script"
> >>>>> }
> >>>>> }
> >>>>> device hp {
> >>>>> filter = "^hpaio.*"
> >>>>> desc = "HP"
> >>>>> action test {
> >>>>> filter = ".*scan.*"
> >>>>> desc = "Test"
> >>>>> script = "/usr/local/etc/scanbd/example.script"
> >>>>> }
> >>>>> }
> >>>>> device avision {
> >>>>> filter = "^avision.*"
> >>>>> desc = "Avision Scanner"
> >>>>> action duplex-test {
> >>>>> filter = "^message.*"
> >>>>> desc = "Avision: Duplex Test"
> >>>>> script = "/usr/local/etc/scanbd/test.script"
> >>>>> string-trigger {
> >>>>> from-value = ""
> >>>>> to-value = "^.:duplex"
> >>>>> }
> >>>>> }
> >>>>> }
> >>>>>
> >>>>> ----- Original Message -----
> >>>>> From: Wilhelm<wilhelm.meier at fh-kl.de>
> >>>>> To: Rastislav David<rasto at ynet.sk>
> >>>>> Cc: sane-devel at lists.alioth.debian.org
> >>>>> Sent: Fri, 25 Nov 2011 07:45:58 +0100 (CET)
> >>>>> Subject: Re: [sane-devel] Scanner Button Daemon [scanbd]: high CPU
> >>>>> utilization
> >>>>>
> >>>>> Hi David,
> >>>>>
> >>>>> Am 24.11.2011 21:49, schrieb Rastislav David:
> >>>>>> Hello all,
> >>>>>>
> >>>>>> I've succesfully compiled scanbd for my Canon CanoScan LiDE 110 on my
> >>>>>> ARM powered NAS Linksys DNS-323 running Debian Wheezy.
> >>>>>>
> >>>>>> It works :) however there are few issues.
> >>>>>> 1) scanbd is using 100% of CPU load and thus misses some button
> >>>>>> presses
> >>>>>
> >>>>> looks like a config problem: please post your scanbd.conf
> >>>>>
> >>>>>> 2) does not react on Pdf button (nothing logged into console when
> >>>>>> pressed) There are 4 buttons. [Pdf Auto-Scan Copy Email]
> >>>>>
> >>>>> we must check the log-file to find out the name of the pdf-button
> >>>>>
> >>>>>>
> >>>>>> If I start daemon with: ./scanbd -df -c scanbd.conf
> >>>>>>
> >>>>>> I see:
> >>>>>>
> >>>>>> ./scanbd: debug on: level: 7
> >>>>>> ./scanbd: dropping privs to uid saned
> >>>>>> ./scanbd: dropping privs to uid scanner
> >>>>>> ./scanbd: group scanner has member:
> >>>>>> ./scanbd: saned
> >>>>>> ./scanbd: drop privileges to gid: 104
> >>>>>> ./scanbd: Running as effective gid 104
> >>>>>> ./scanbd: drop privileges to uid: 110
> >>>>>> ./scanbd: Running as effective uid 110
> >>>>>> ./scanbd: dbus_init
> >>>>>> ./scanbd: dbus match
> >>>>>> type='signal',interface='org.freedesktop.Hal.Manager'
> >>>>>> ./scanbd: sane version 1.0
> >>>>>> ./scanbd: Scanning for local-only devices
> >>>>>> ./scanbd: found device: genesys:libusb:001:011 Canon LiDE 110 flatbed
> >>>>>> scanner
> >>>>>> ./scanbd: start_sane_threads
> >>>>>> ./scanbd: Starting poll thread for genesys:libusb:001:011
> >>>>>> ./scanbd: Thread started for device genesys:libusb:001:011
> >>>>>> ./scanbd: sane_poll
> >>>>>> ./scanbd: start udev thread
> >>>>>> ./scanbd: timeout: 400 ms
> >>>>>> ./scanbd: udev thread started
> >>>>>> ./scanbd: Iteration on dbus call
> >>>>>> ./scanbd: no device from udev
> >>>>>> ./scanbd: no device from udev
> >>>>>> ....
> >>>>>> 300x / second
> >>>>>> ./scanbd: no device from udev
> >>>>>> ./scanbd: no device from udev
> >>>>>> ./scanbd: no device from udev
> >>>>>> ./scanbd: no device from udev
> >>>>>> ./scanbd: no device from udev
> >>>>>> ./scanbd: found 41 options for device genesys:libusb:001:011
> >>>>>> ./scanbd: sane_find_matching_options
> >>>>>> ./scanbd: found 5 actions in section (null)
> >>>>>> ./scanbd: checking action scan with filter: ^scan.*
> >>>>>> ./scanbd: found active option[2] mode (type: 3) for device
> >>>>>> genesys:libusb:001:01
> >>>>>> 1
> >>>>>> ./scanbd: found active option[4] preview (type: 0) for device
> >>>>>> genesys:libusb:001
> >>>>>> :011
> >>>>>> ./scanbd: found active option[5] depth (type: 1) for device
> >>>>>> genesys:libusb:001:0
> >>>>>> 11
> >>>>>> ./scanbd: found active option[6] resolution (type: 1) for device
> >>>>>> genesys:libusb:
> >>>>>> 001:011
> >>>>>> ./scanbd: found active option[8] tl-x (type: 2) for device
> >>>>>> genesys:libusb:001:01
> >>>>>> 1
> >>>>>> ./scanbd: found active option[9] tl-y (type: 2) for device
> >>>>>> genesys:libusb:001:01
> >>>>>> 1
> >>>>>> ./scanbd: found active option[10] br-x (type: 2) for device
> >>>>>> genesys:libusb:001:011
> >>>>>> ./scanbd: found active option[11] br-y (type: 2) for device
> >>>>>> genesys:libusb:001:011
> >>>>>>
> >>>>>>
> >>>>>> these "no device from udev" messages are generated about 300/s which I
> >>>>>> think is causing high CPU utilization.
> >>>>>
> >>>>> That's strange: did not see this before :-(
> >>>>>
> >>>>>>
> >>>>>> in scandb.conf I left all devices as they were and for actions scan
> >>>>>> email and copy I've changed script to /etc/scanbd/action.script.
> >>>>>> This script works fine for said 3 buttons.
> >>>>>>
> >>>>>> Can You help where could be the issue with High CPU utilization?
> >>>>>> Is there some problem with configuration?
> >>>>>
> >>>>> I think this has to to with the libudev. You can disable libudev at
> >>>>> compile-time (and using hald if you want) loosing the capability to
> >>>>> dynamically add scanners. Can you give this a try, please?
> >>>>>
> >>>>>>
> >>>>>> Rastislav
> >>>>>>
> >>>>>> P.S: for scanbd to compile I had to change
> >>>>>> -I/usr/local/include/dbus-1.0/include to
> >>>>>> -I/usr/lib/arm-linux-gnueabi/dbus-1.0/include
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>
> >>>>>
> >>>>
> >>>>
> >>>> --
> >>>> Wilhelm
> >>>>
> >>>
> >>
> >>
> >
>
>
> --
> Wilhelm
>
More information about the sane-devel
mailing list