#23061: crypto_rand_double() should produce all possible outputs on platforms
with
32-bit int
-------------------------------------------------+-------------------------
Reporter: teor | Owner: nickm
Type: defect | Status:
| needs_revision
Priority: Medium | Milestone: Tor:
| 0.3.3.x-final
Component: Core Tor/Tor | Version: Tor:
| 0.2.2.14-alpha
Severity: Normal | Resolution:
Keywords: tor-relay, security-low, privcount, | Actual Points: 0.5
031-backport, 030-backport, 029-backport, 028 |
-backport-maybe, 026-backport-maybe, review- |
group-22 |
Parent ID: #23126 | Points: 0.1
Reviewer: | Sponsor:
| SponsorQ
-------------------------------------------------+-------------------------
Comment (by teor):
Replying to [comment:45 catalyst]:
> Section 5.2 of
https://pdfs.semanticscholar.org/2f2b/7a0d5000a31f7f0713a3d20919f9703c9876.pdf
describes one way to sample uniformly from all the representable floating
point numbers in [0, 1). It's not clear to me whether including the
numbers with ULPs less than 2^-53^ is required for the success of the
snapping mitigation described in that paper.
It isn't required, and might even be counterproductive.
The floating point numbers in the proof in section 5.2 are of the form
N*2^-53^, N integer.
(The proof uses a significand of 52 bits, and refers to 2^-53^
repeatedly.)
--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/23061#comment:46>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online
_______________________________________________
tor-bugs mailing list
[email protected]
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs
