[tor-bugs] #26040 [Core Tor/Tor]: Improve getrandom handling

[Tor Bug Tracker & Wiki]

#26040: Improve getrandom handling
--------------------------+------------------------------------
 Reporter:  Hello71       |          Owner:  Hello71
     Type:  enhancement   |         Status:  needs_revision
 Priority:  Medium        |      Milestone:  Tor: 0.3.4.x-final
Component:  Core Tor/Tor  |        Version:
 Severity:  Normal        |     Resolution:
 Keywords:                |  Actual Points:
Parent ID:                |         Points:
 Reviewer:  catalyst      |        Sponsor:
--------------------------+------------------------------------

Comment (by cypherpunks):

 Replying to [comment:5 Hello71]:
 > Hm... I see, you are correct. I really really hope that nowadays
 everybody has random seed persistence, but it is of course better to be
 conservative here.

 Unfortunately, the random seed takes quite some time (on the order of
 minutes) to actually take effect. The seed is written to the non-blocking
 character device which triggers the `random_write` file operation which
 uses `write_pool` to send the data to the input pool. Unfortunately it can
 take a while for the secondary pools to receive the seed, since they have
 to wait for the `push_to_pool` workqueue function to be triggered. On
 newer Linux kernels (using ChaCha20 rather than SHA-1 for the non-blocking
 character device), the input pool is queried every 5 minutes, and it only
 reseeds the stream cipher if more than 128 bits of entropy have been
 collected in the input pool since the last reseed.

 If you do not check for `EINTR` (and avoid the blocking behavior
 altogether), then even if you are using a persistent random seed, you will
 end up obtaining potentially predictable random data.

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/26040#comment:12>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online