[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]
Re: [tor-bugs] #17789 [Tor]: Add syscall-based crypto seeding for OS X
#17789: Add syscall-based crypto seeding for OS X
--------------------+------------------------------------
Reporter: teor | Owner:
Type: defect | Status: closed
Priority: Medium | Milestone: Tor: 0.2.8.x-final
Component: Tor | Version: Tor: unspecified
Severity: Normal | Resolution: wontfix
Keywords: | Actual Points:
Parent ID: | Points:
Sponsor: |
--------------------+------------------------------------
Changes (by teor):
* status: new => closed
* resolution: => wontfix
Comment:
Replying to [comment:2 nickm]:
> (Which is to say, if they include an internal PRNG I'm not too excited
about them, and if they just wrap open("/dev/random") I'm not too excited
about them.)
In fact, those are the exact two options the Apple APIs support!
They're also not fork()-safe, and they express this by abort()ing on the
first call in a child after a fork() but before an exec(). This causes
issues with (misbehaved) tor unit tests that initialise global random
state, but don't TT_FORK.
There are ways to work around this, and the commit comment in my branch
feature17789-v2 describes them. But they're complex, and pointless since
we're not getting random bytes from the kernel itself.
Closing as wontfix, until a better Apple API comes along. At that time,
the code in feature17789-v2 could be useful as a starting point.
--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/17789#comment:3>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online
_______________________________________________
tor-bugs mailing list
tor-bugs@xxxxxxxxxxxxxxxxxxxx
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs