Description of problem: Case 1: There seems to be a mixture of 'verbose' and 'verbosity' option of mount.ecryptfs command. 'man ecryptfs' says that verbosity option is used to get more verbose output from ecryptfs mount helper in system logs. When I use 'verbosity=1' the output in syslog is still the same. When I use 'verbose=1' instead syslog is filled with a lots of information from ecryptfs. Case 2: Furthermore I've been using 'verbosity=0' to stop mount helper from prompting me when all required options (key method, passwd, cipher, key_bytes) are specified in the option list and use defaults for the rest (plaintext passthrough, xattr). If I don't use this option I'm being asked for plaintext passthrough which should be off by default. Also following message is written in syslog: ecryptfs_parse_options: eCryptfs: unrecognized option 'verbosity=0' Version-Release number of selected component (if applicable): ecryptfs-utils-41-1.el5 How reproducible: 100% Steps to Reproduce: 0. look at man pages - search for 'verbosity' option Do following as root: Case 1: verbosity/verbose mixture just mount with these commands and compare /var/log/messages output ... # mount -t ecryptfs .secret .secret -o key=passphrase:passwd=secret,ecryptfs_cipher=aes,ecryptfs_key_bytes=16,no_sig_cache,verbosity=1 -- not lots of information in /var/log/messages -- and unrecognized option verbosity in logs ... # mount -t ecryptfs .secret .secret -o key=passphrase:passwd=secret,ecryptfs_cipher=aes,ecryptfs_key_bytes=16,no_sig_cache,verbose=1 -- LOTS of information from ecryptfs in /var/log/messages Case 2: verbosity=0 influences further prompts ... # mount -t ecryptfs .secret .secret -o key=passphrase:passwd=secret,ecryptfs_cipher=aes,ecryptfs_key_bytes=16,no_sig_cache -- prompt for plaintext passthrough (why? when it's off by default) # mount -t ecryptfs .secret .secret -o key=passphrase:passwd=secret,ecryptfs_cipher=aes,ecryptfs_key_bytes=16,no_sig_cache,verbosity=0 -- no prompt and fs mounted, but /var/log/messages shows that the option is not recognized! Actual results: included in test cases Expected results: Case 1: man pages should change word 'verbosity' to 'verbose' Case 2: Prompt for 'plaintext passthrough' is not shown when all required options are specified. If user wants that feature he has to specify that explicitly. The 'verbosity' option should be removed and should not influence mount helper behaviour.
Sounds like another doable fix. Proposing for RHEL-5.3 and granting Devel ACK. Read ya, Phil
Proposing bug for RHEL-5.3 FasTrack. Read ya, Phil
Case 1: verbose isn't the same as verbosity. verbosity takes an argument, verbose always switches verbose output on and verbose=0 is the same as verbose=1 or verboseXXXX verbosity=1 isn't parsed anywhere in the sources Case 2: Here are several bugs: - verbosity isn't removed from the list of options passed to systems mount command. This causes the 'unrecognized option' error. - verbosity=0 apparently does the same as no_sig_cache, this looks like a cut¬paste error in the sources. I still need to look why and where it turns of the prompts, though
approved comp, clearing fast flag
I think I need to clarify this: mount.ecryptfs has two distinct parameters 'verbose' and 'verbosity={0,1}' 'verbose' enables a very verbose logging in /var/log/messages and shouldn't be used in production environments unless you're debugging something. 'verbosity=0' uses default values for required parameters which aren't specified on the commandline, p.e. it doesn't ask for 'plaintext passthrough (y/n)' anymore with the commandline from comment #7 and just uses plaintext passthrough. 'verbosity=1' asks for every required parameter which hasn't been specified on the commandline. That's the same as not using the verbosity option at all. Older versions had a bug which passed the verbosity option to /sbin/mount instead of keeping it in /sbin/mount.ecryptfs. This resulted in a message "ecryptfs_parse_options: eCryptfs: unrecognized option 'verbosity=0'" in the syslog. This is fixed in our current package. Nonetheless there's a bug in the current package which misinterpretes 'verbosity=1' to be the same as 'verbose'. This needs to be fixed and I already have a patch for that.
An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on therefore solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHBA-2009-0203.html