Bug 141277 – app-crypt/gnupg - gpg --disable-ccid option when using pcmcia based cardreader

• Product: Gentoo Linux
• Status: RESOLVED NEEDINFO

[Description Harvey Muller 2006-07-21 06:03:57 UTC]

PROBLEM:
gpg barfs =) when attempting to generate keys if the --disable-ccid option is not set.  The error generally looks like this:

gpg: apdu_send_simple(0) failed: unknown status error
gpg: generating key failed
gpg: key generation failed: general error
Key generation failed: general error
gpg: apdu_send_simple(0) failed: unknown status error
gpg: failed to set `CHV-STATUS-1': general error
gpg: error setting forced signature PIN flag: general error

WORKAROUND:
The above error disappears when using the --disable-option at the command prompt.  An alias statement in .bashrc, is obviously a little better than that.

PROPOSED SOLUTION:
I preface the proposal, by stating I know very little about ebuilds at this point in time.  But would it be possible to set this permanently during compile time, if a pcmcia base card reader is detected?

Thanks for all you do.

[Comment 1 Jakub Moc (RETIRED) 2006-07-21 06:23:12 UTC]

Please, always include ebuild name and version plus emerge --info when reporting bugs.

[Comment 2 Harvey Muller 2006-07-21 07:32:21 UTC]

Created attachment 92417 [details]
emerge --info data for initial report

As requested, and will do in the future.

[Comment 3 Harvey Muller 2006-07-21 07:33:35 UTC]

ebuild name / version / USE flags:

app-crypt/gnupg-1.9.20-r3 USE="smartcard -X -caps -gpg2-experimental -ldap -nls"

See attachment 92417 [details] (created 2006-07-21 07:32 PST) for emerge --info

[Comment 4 Harvey Muller 2006-07-24 13:23:56 UTC]

UPDATE:  Although this 'seemed' to resolve the problem (tested on two different cards, only generating primary keys on the card), the problem continued to persist when I generated subkeys to add to the card.  The same general error was present.

Through trial and error, what seemed to help, was typing a bunch of junk into a dummy document to help increase random number generation, then generating a subkey.  I was able to create three subkeys on the card using this approach.  So maybe this is an issue with the application not getting enough random bits?  I don't know enough to be sure, but that's my hunch.

[Comment 5 Alon Bar-Lev (RETIRED) 2006-09-13 10:34:17 UTC]

Hello,

I cannot really understand the problem.

Can you please try to add some more description?

Which cards do you use?
Which readers do you use?
Is there any other application where card work?
Do you use pcsc-lite reader drivers, or relay on gnupg ccid implementation?

Thanks!

[Comment 6 Daniel Black (RETIRED) 2006-09-19 14:23:12 UTC]

As per comment #5 - need a bit more info. Please reopen when you attach it.