https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=256781
--- Comment #31 from martin.stoya...@vertalo.com ---
(In reply to Colin Percival from comment #30)
With the terraform test case I attached, I am able to reproduce this fairly
reliably, but only with the 13.0-RELEASE AMIs. As a test, I start
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=256781
--- Comment #32 from martin.stoya...@vertalo.com ---
(In reply to Colin Percival from comment #30)
I guess I spoke too soon - after a reboot of this instance, this is what I get
(with TSC-low):
root@freebsd:~ # uname -a
FreeBSD freebsd 14.0
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=254513
--- Comment #18 from Chris Collins ---
I believe the default on qemu is /dev/random, but I am using /dev/urandom as is
the default on proxmox, they state the reason for using /dev/unrandom is its
non blocking, I could retry on /dev/random t
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=254513
--- Comment #19 from Chris Collins ---
Sorry just to add, I am testing this on a local proxmox, so I can reboot and
the actual kernel version is the 5.11 version, specifically 5.11.22-2-pve.
so using /dev/urandom proxmox default with 1024
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=254513
--- Comment #20 from Conrad Meyer ---
It would be nice to get line numbers on that stack in vtrnd_read.
I wonder if this (misguided) Qemu feature is in play:
> Optional parameters to limit the rate of data sent to the guest are inlucded:
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=254513
--- Comment #21 from Conrad Meyer ---
Current qemu default appears to be infinity:
https://github.com/qemu/qemu/blob/master/hw/virtio/virtio-rng.c#L251
--
You are receiving this mail because:
You are the assignee for the bug.
You are on t
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=254513
--- Comment #22 from Conrad Meyer ---
This arbitrary low limit seems to be imposed by proxmox?
> By default, the limit is set to 1024 bytes per 1000 ms (1 KiB/s). It is
> recommended to always use a limiter to avoid guests using too many
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=254513
--- Comment #23 from Conrad Meyer ---
We read a lot from random sources we expect to be “fast”, like virtio-rng.
https://github.com/freebsd/freebsd-src/blob/main/sys/dev/random/random_harvestq.c#L231
10Hz x 4B x HARVESTSIZE x (max) RANDOM
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=254513
--- Comment #24 from Conrad Meyer ---
Some related history:
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=230808#c1
--
You are receiving this mail because:
You are the assignee for the bug.
You are on the CC list for the bug.
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=254513
--- Comment #25 from Conrad Meyer ---
(Min rate is about 2.5 kB/s, also in excess of the 1kB/s proxmox default.)
--
You are receiving this mail because:
You are the assignee for the bug.
You are on the CC list for the bug.
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=256781
--- Comment #33 from Colin Percival ---
(In reply to martin.stoyanov from comment #32)
Just to confirm: It looks like that instance is having the problem by default,
but the problem goes away when you tell it to use kvmclock?
--
You are r
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=256781
--- Comment #34 from martin.stoya...@vertalo.com ---
(In reply to Colin Percival from comment #33)
Yep, that's exactly right.
--
You are receiving this mail because:
You are the assignee for the bug.
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=256781
--- Comment #35 from Colin Percival ---
(In reply to martin.stoyanov from comment #34)
Ok, great. In that case no need to keep the instance around; we'll get the
default timecounter changed and the issue should be gone next release.
--
Y
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=256781
--- Comment #36 from martin.stoya...@vertalo.com ---
(In reply to Colin Percival from comment #35)
Works for me; thanks a lot for your time and effort!
--
You are receiving this mail because:
You are the assignee for the bug.
14 matches
Mail list logo
