Re: use of cprng_fast vs. random(9) (Was: CVS commit: src/sys/altq)
Robert Elz <[email protected]>
| Newsgroups | gmane.os.netbsd.devel.kernel |
|---|---|
| Message-ID | <[email protected]> |
Date: Wed, 22 Jan 2025 15:03:36 +0100
From: Christoph Badura <[email protected]>
Message-ID: <[email protected]>
No comment on why the change happened in the first place, but:
| The followup commit by kre partiallly undid the change for "rump and
| anything else userland". Why is this necessary for rump?
<sys/cprng.h> isn't installed, so is only available to kernel builds,
that could be changed, but was more than I was willing to do to make
the builds work again, which was my sole objective.
| And what specifically the "anything else userland" is.
Most likely, nothing.
| [1] It seems to me that pampering over over rump failures instead of
| properly fixing the issues as the crop up has become kind of a sport in
| recent years.
By all means, if you have the energy, dive into rump and fix all of its
issues properly.
kre