https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=243669
Gregory Lynn <horroraro...@gmail.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |horroraro...@gmail.com
--- Comment #13 from Gregory Lynn <horroraro...@gmail.com> ---
Therefore, the issue isn't that a limit existed before #f058359ba5 and that one
did after #f058359ba5. I would like to know what kernel version Qemu is
running and if it is configured to utilize a non-default backend for entropy.
As per the information found under https://wiki.qemu.org/Features/VirtIORNG,
the default setting is /dev/random, which can cause issues till later Linux
releases. This is where I'm wondering if our (10hz?) polling is exhausting the
host's /dev/random, which often leaves us in a tight spot:
https://cgit.freebsd.org/src/tree/sys/dev/virtio/virtqueue.c?h=releng/13.0#n605
https://geometrydash-online.io -- Adding some virtio and csprng@ people to the
CC list.
I'm wondering if it would make sense to give virtqueue_poll() an optional
timeout parameter so that, in the event that vtrnd is unable to contribute to
the pool, rng doesn't become stuck.
--
You are receiving this mail because:
You are the assignee for the bug.
You are on the CC list for the bug.
