In the 2.6.3 kernel, the old USB scanner module had vanished (I don't know why). Apparently one is supposed to use the libusb code instead, and I eventually figured out how to get that set up. Unfortunately, when I do that, while the scanner works for previews and small scans, on larger scans, quite likely the scanner hangs, and I find this in /var/log/messages: Mar 7 12:59:40 localhost kernel: usb 2-1: control timeout on ep0out At this point, the only thing I've found that will clear the problem is rebooting; if I turn the scanner on and off, I get something like this: Mar 7 13:01:40 localhost kernel: usb 2-1: USB disconnect, address 2 Mar 7 13:01:56 localhost kernel: usb 2-1: new full speed USB device using address 3 Mar 7 13:02:01 localhost kernel: usb 2-1: control timeout on ep0out Here si a forum posts(possibly) reporting similar problems: <http://forums.gentoo.org/viewtopic.php?t=139874> But neither this not anything else I found searching on google (for "control timeout on ep0out" among other things), have any solution to the problem (other than the one person who suggested putting back the scanner module). The scanner is question is an Epson Perfection 1650, and it works fine with the 2.4 series (and even 2.6.2) with the scanner module. Let me know if you need more information. For now I've reverted to the 2.4 series as this is blocker for me to go to 2.6.3+. Reproducible: Sometimes Steps to Reproduce: 1. Install kernel 2.6.3 2. Emerge libusb and re-emerge sane backends to get the scanner to work at all. 3. Figure out hotplug to get the scanner to work for non-root users. 4. Scan a bunch of stuff Actual Results: Eventually the scanner will probably wedge as indicated above. Expected Results: Scanner should work. Portage 2.0.50-r1 (default-x86-1.4, gcc-3.3.2, glibc-2.3.2-r9, 2.6.3-gentoo-r1) ================================================================= System uname: 2.6.3-gentoo-r1 i686 AMD Athlon(tm) XP 1900+ Gentoo Base System version 1.4.3.13 Autoconf: sys-devel/autoconf-2.58-r1 Automake: sys-devel/automake-1.7.7 ACCEPT_KEYWORDS="x86" AUTOCLEAN="yes" CFLAGS="-g -O2 -fstack-protector" CHOST="i686-pc-linux-gnu" COMPILER="gcc3" CONFIG_PROTECT="/etc /usr/X11R6/lib/X11/xkb /usr/kde/2/share/config /usr/kde/3.1/share/config /usr/kde/3.2/share/config /usr/kde/3/share/config /usr/share/config /usr/share/texmf/dvipdfm/config/ /usr/share/texmf/dvips/config/ /usr/share/texmf/tex/generic/config/ /usr/share/texmf/tex/platex/config/ /usr/share/texmf/xdvi/ /var/qmail/control" CONFIG_PROTECT_MASK="/etc/gconf /etc/env.d" CXXFLAGS="-g -O2 -fstack-protector" DISTDIR="/usr/portage/distfiles" FEATURES="autoaddcvs fixpackages nostrip sandbox userpriv" GENTOO_MIRRORS="ftp://mirrors.tds.net/gentoo http://mirror.tucdemonic.org/gentoo/ ftp://gentoo.mirrors.pair.com/ http://gentoo.mirrors.pair.com/ http://gentoo.ccccom.com ftp://gentoo.ccccom.com http://gentoo.oregonstate.edu http://www.ibiblio.org/pub/Linux/distributions/gentoo" MAKEOPTS="-j2" PKGDIR="/usr/portage/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="/usr/local/portage" SYNC="rsync://rsync.gentoo.org/gentoo-portage" USE="3dnow X Xaw3d aalib apm arts avi berkdb canna cdr cjk crypt cups debug doc dvd encode esd foomaticdb freetype freewnn gdbm gif gnome gpm gtk gtk2 guile imlib java joystick jpeg kde libg++ libwww mad maildir matrox mbox mikmod mozilla mozsvg mpeg mule ncurses nls oggvorbis opengl oss pam pdflib perl png python qt quicktime readline ruby scanner sdl slang spell sse ssl svga tcltk tcpd tetex tiff truetype usb video_cards_matrox x86 xinerama xml2 xmms xv zlib"
Created attachment 27002 [details] Config file for kernel This is the .config file from the kenrel exhibiting the problem.
Created attachment 27003 [details] /var/log/messages This is the contents of /ver/log/messages for a day when I had the problem.
OK; this looks like a possible upstream bug in 2.6.3. Can you try out development-sources-2.6.3 and see if you get it there? Try this as a possible temporary fix: ``umount /proc/bus/usb''; reboot the scanner; and ``mount /proc/bus/usb'' again.
Okay, I'll give those a try, though it may be a few days before I have a chance. I also noticed that there was a USB bug fixed in 2.6.4, so maybe that will fix the problem.
Thanks for the reply and suggestions. I tried development-sources-2.6.4 (not 2.6.3 since I took so long to try), but had the same problem. (Though it looks like my guess as it being the size of the image causing the problem was wrong; I have an image that seems to cause the problem most of the time, but some bigger scans work okay. Again, no problems with the 2.4 series using the kernel usb scanner modules.) Also, I tried the workaround you suggested, but it didn't help; I get the same sort of messages when I remount /proc/bus/usb and plug the scanner back in: Mar 27 18:22:15 localhost kernel: usb 2-1: new full speed USB device using address 6 Mar 27 18:22:20 localhost kernel: usb 2-1: control timeout on ep0out Note that I have a USB mouse, if that matters. Another google search again turned up others with the problem (such as <http://www.mail-archive.com/newbie@linux-mandrake.com/msg153711.html> on a mandrake system) but no one with a solution.
Sorry, that should be development-sources-2.6.4-r1.
Sorry, development-sources-2.6.4-rc1 (not -r1).
Would you mind trying to pass pci=noacpi on the knerel line in {grub|lilo}.conf file and see what it does?
I have ACPI turned off in the kernel configuration: # CONFIG_ACPI is not set So would pci=noacpi make any difference?
You're right, passing pci=noacpi on the kernel line wouldn't make any difference. Could you try enabling acpi and see if that sorts things out?
this was a kernel bug in usbfs, and is fixed in the latest -mm releases. Can you try it to see if it is fixed in there?
Okay, I'll give those a try (starting with the mm kernel).
I tried mm-sources-2.6.5-r1. It was worse: any attempt to access the scanner (e.g., scanimage -L) hangs the process, and nothing short of a reboot would get rid of it. There aren't any helpful USB messages in /var/log/messages (though I'll attach it anyway), even though I had CONFIG_USB_DEBUG=y in my .config. Oddly, during all this, the USB mouse works fine. (Note that there seem to be two USB hubs; the mouse and the scanner are on different hubs.) I'll try turning on ACPI when I have a change, hopefully this weekend.
I tried swapping where the scanner was plugged in, so it was on the same hub as the mouse. That made no difference. Also, I keep forgetting to mention: after one of the these USB hangs (with any of the kernels), shutdown hangs when its shutting down the USB subsystem. "rmmod uhci_hcd" also hangs; I get this message (either in shutdown or rmmod): Apr 15 20:34:21 localhost kernel: uhci_hcd 0000:00:11.2: remove, state 1 Apr 15 20:34:21 localhost kernel: usb usb1: USB disconnect, address 1 Apr 15 20:34:21 localhost kernel: usb 1-1: USB disconnect, address 2
Created attachment 29391 [details] /var/log/messages from the -mm kernel
According to this, <http://bugme.osdl.org/show_bug.cgi?id=2342>, the mm2 version had the second USB problem, so maybe the -mm1 did too. I see there's an -mm6 now in portage, so I'll give that a try.
It looks like with the mm-sources-2.6.5-r6, my scanner is functioning without either of the problems I reported above. (Since the first problem is sporatic, I can't be completely sure, but I've done more scans than I could before.) I'm therefore reducing the severity level of the bug, but it shouldn't be closed until the fix makes it into the gentoo kernel.
Should be now in the 2.6.6-rc2 kernel.
Thanks, I'll check that out as soon as it gets into Portage.
It's in Portage now - is the problem fixed?
Yes, the bug appears to be fixed in development-sources-2.6.6_rc2.
Yeah, it's fixed then with no work from us :)