Red Hat Bugzilla – Full Text Bug Listing
|Summary:||Scanner permissions not set if connected after user logs in|
|Product:||[Fedora] Fedora||Reporter:||W. Michael Petullo <redhat>|
|Component:||udev||Assignee:||Harald Hoyer <harald>|
|Status:||CLOSED RAWHIDE||QA Contact:|
|Version:||rawhide||CC:||bugzilla.redhat.com, harald, james, kdekorte, tmraz|
|Fixed In Version:||Doc Type:||Bug Fix|
|Doc Text:||Story Points:||---|
|Last Closed:||2006-04-30 13:31:52 EDT||Type:||---|
|oVirt Team:||---||RHEL 7.3 requirements from Atomic Host:|
Description W. Michael Petullo 2006-02-25 16:22:57 EST
Description of problem: When a USB scanner is plugged in, udev sets up the proper device files (see closed bug #177650.) Upon logging in, the permissions on this device file are set to the console owner by pam_console. However, if the scanner is plugged in after the user has already logged in, then the permissions of its device file will not be set. Version-Release number of selected component (if applicable): udev-084-4 How reproducible: Every time Steps to Reproduce: 1. Log in, obtaining console ownership. 2. Plug in a USB scanner. Actual results: Notice that the logged in user does not own /dev/scanner-... Expected results: /dev/scanner... should be owned by the user that ownes the console. Additional info: At one time /etc/dev.d/default/05-pam_console.dev set the permissions on newly created devices when a user was logged in. This seems to be missing from recent udev packages. I don't see anything in the RPM changelog explaining the script's absence.
Comment 1 Takaki Hiroshi 2006-03-05 12:13:49 EST
*** Bug 175005 has been marked as a duplicate of this bug. ***
Comment 2 Kevin DeKorte 2006-03-20 21:01:41 EST
I have an HP 5400c scanner. I login as my limited user and I plugin the scanner. The device /dev/scanner-usbdev1.7 is created as a simlink to lrwxrwxrwx 1 root root 15 Mar 20 18:49 scanner-usbdev1.7 -> bus/usb/001/007 ls -la bus/usb/001/007 crw-r--r-- 1 root root 189, 6 Mar 20 18:54 bus/usb/001/007 So the user cannot open the scanner. If I run xscan under sudo then the scanner works correctly. I DO NOT have SELinux enabled. It looks to me that either the scanner should be group owned by the "scanner" group and the user needs to be added to that group or some other privledge change should be made. I'm on FC5Test3 and upgrading to FC5 Final at the moment.
Comment 3 Harald Hoyer 2006-03-21 07:33:40 EST
should be fixed with udev-084-13
Comment 4 Takaki Hiroshi 2006-03-21 08:01:12 EST
I'm useing FC5 final. udev-084-13 not works for me. Should I reopen this bug, or wait for next release of udev?
Comment 5 Harald Hoyer 2006-03-21 08:54:31 EST
hmm... that's odd.. could you set udev_log=debug in /etc/udev/udev.conf and post the entries in /var/log/messages regarding udev when you plug in the scanner?
Comment 6 Takaki Hiroshi 2006-03-21 09:23:31 EST
Created attachment 126396 [details] /var/log/messages when I power on the scanner. I hope this is help.
Comment 7 Kevin DeKorte 2006-03-21 09:40:24 EST
I tried udev-084-13 and after a reboot everything worked fine for me. So this issue is fixed for me.
Comment 8 Harald Hoyer 2006-03-21 09:41:34 EST
hmm Mar 21 23:15:20 tdvia udevd-event: run_program: '/sbin/pam_console_apply /dev/bus/usb/001/002 /dev/scanner-usbdev1.2' should have changed the permissions.... reassigning to pam. you could try to run: # /sbin/pam_console_apply /dev/bus/usb/001/002 /dev/scanner-usbdev1.2 but remember, you should be logged in with your normal user first to own the console.
Comment 9 Tomas Mraz 2006-03-21 10:05:04 EST
The filename argument of pam_console_apply must match device patterns in /etc/security/console.perms.d. However it shouldn't matter much if /dev/scanner-usbdev1.2 points to /dev/bus/usb/001/002 anyway.
Comment 10 Harald Hoyer 2006-03-21 10:07:21 EST
/dev/scanner-usbdev1.2 should match /dev/scanner* :)
Comment 11 Takaki Hiroshi 2006-03-21 10:14:46 EST
I tried a little. 1. login as user named takaki. [takaki@tdvia home]$ ls -l /dev/bus/usb/001/002 crw------- 1 takaki root 189, 1 Mar 21 23:15 /dev/bus/usb/001/002 2. power off and then power on the scanner. [takaki@tdvia home]$ ls -l /dev/bus/usb/001/003 crw-r--r-- 1 root root 189, 2 Mar 21 23:57 /dev/bus/usb/001/003 3. pam_console_apply with normal user. [takaki@tdvia home]$ /sbin/pam_console_apply /dev/bus/usb/001/003 /dev/scanner-usbdev1.3 [takaki@tdvia home]$ ls -l /dev/bus/usb/001/003 crw-r--r-- 1 root root 189, 2 Mar 21 23:57 /dev/bus/usb/001/003 4. pam_console_apply with root user. [takaki@tdvia home]$ sudo /sbin/pam_console_apply /dev/bus/usb/001/003 /dev/scanner-usbdev1.3 [takaki@tdvia home]$ ls -l /dev/bus/usb/001/003 crw------- 1 takaki root 189, 2 Mar 21 23:57 /dev/bus/usb/001/003
Comment 12 Harald Hoyer 2006-03-21 10:24:27 EST
udev calls /sbin/pam_console_apply with root perms... so it should have worked (for Kevin DeKorte it does :)
Comment 13 Tomas Mraz 2006-03-21 10:29:26 EST
I don't think it's pam_console_apply problem as it clearly works fine when it's run from command line. Couldn't there be some race condition so the pam_console_apply is run before the symlink is created or something like that?
Comment 14 Takaki Hiroshi 2006-03-21 10:31:59 EST
(In reply to comment #7) > I tried udev-084-13 and after a reboot everything worked fine for me. So this > issue is fixed for me. Is it works for you, if you power off & on the scanner?
Comment 15 Kevin DeKorte 2006-03-21 11:08:33 EST
My situation is perhaps a little different than yours. My scanner is across the room so I have a long cable to hook it up (it is not always plugged into my machine). And if the scannner is off, it does not show up on the USB bus. So am logged in and plugin the cable with the scanner off. I then power the scanner and everything comes up right.
Comment 16 Takaki Hiroshi 2006-03-21 11:48:55 EST
Thanks for reply, Kevin. It's works for me, if I tried another box. My problem occur on particular box. I'll check up more it. Please, close here, Harald. I'm sorry for bothering.
Comment 17 Takaki Hiroshi 2006-03-25 15:31:50 EST
I'v found another box which has same problem. Two machines are also smp system. This issue may be software problem but not depend on a particular hardware. machine A: RIOWORKS TDVIA machine B: Dell PowerEdge SC430
Comment 18 W. Michael Petullo 2006-04-30 13:31:52 EDT
I can confirm that this has been fixed in Rawhide. It works for me with and without SELinux enforcing its strict policy. If there are issues with individual pieces of hardware, then individual bugs should be opened. In general, the system works.
Comment 19 W. Michael Petullo 2006-04-30 13:34:30 EDT
(In reply to comment #18) > I can confirm that this has been fixed in Rawhide. It works for me with and > without SELinux enforcing its strict policy. If there are issues with > individual pieces of hardware, then individual bugs should be opened. In > general, the system works. Please replace strict with targeted in the above comment. I have not tested the strict policy. However, if there is a problem with the strict policy, then a seperate bug should be opened against the policy.