Hang in scdaemon or pcsc-wrapper on Yosemite
lukele at gpgtools.org
Sat Jan 17 18:23:36 CET 2015
Thank you very much for this information.
Especially the hint to instruct gdb or lldb to ignore signals might just be what we need to further debug this issue.
On a different not, on Mavericks it appears is not allowed to claim exclusive access to a smart card unless some system files are manipulated.
After reading up a little on possible workaround, it looks like using transactions instead of claiming exclusive access might work.
We’d be happy to add support for this, but wanted to ask you if there are any security or other issues we should be aware of, which make the current solution the better one.
On 14.01.2015, at 17:16, Werner Koch <wk at gnupg.org> wrote:
> On Wed, 14 Jan 2015 12:42, lukele at gpgtools.org said:
>> In order to find out what’s going on we were trying to connect to a
>> debug build of scdaemon and pcsc-wrapper. Unfortunately though,as soon
>> as the debugger is connected, scdaemon as well as the pcsc-wrapper
>> exit due to a SIGPIPE signal.
> scdaemon can't die due to a SIGPIPE because the handler has been set to
> SIG_IGN. pcsc-wrapper may die though.
> I assume your problem is with the pth library. Note that pth uses a
> signal for its own (IIRC, SIGUSR2) and thus you need to tell the
> debugger to pass that signal through. Some debuggers have problems
> with pth; gdb works fine.
>> Also if anyone knows what might lead to a hang in scdaemon, that
>> information would be very valuable as well.
> Use strace/truss etc to debug such problems.
> Die Gedanken sind frei. Ausnahmen regelt ein Bundesgesetz.
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 236 bytes
Desc: Message signed with OpenPGP using GPGMail
More information about the Gnupg-devel