new Ryzen lockup issue ?

[Mike Tancsa]

Not sure if I have found another issue specific to Ryzen, or a bug that
manifests itself on Ryzen systems easier.  I installed the latest
virtualbox from the ports and was doing some network performance tests
between a vm and the hypervisor using iperf3.  The guest is just a
RELENG11 image and the network is an em nic bridged to epair1b

Then running in a loop

#!/bin/sh
while true
do
iperf3 -t 20 -R -c 192.168.19.18
iperf3  -t 20 -c 192.168.19.18
sleep 20
done


where 192.168.17.19 is the guest image, the box locks up solid after
10min to 30min of letting the script run.  By lockup I mean the keyboard
is not responsive or anything.  I can do this on two separate physical
machines (both ASUS MBs).

# ifconfig epair1a
epair1a: flags=8943
metric 0 mtu 1500
description: ---FreeBSD11-generic-VM
options=8
ether 02:ff:20:00:06:0a
hwaddr 02:ff:20:00:06:0a
inet 192.168.19.17 netmask 0xff00 broadcast 192.168.19.255
nd6 options=29
media: Ethernet 10Gbase-T (10Gbase-T )
status: active
groups: epair

# ifconfig epair1b
epair1b: flags=8843 metric 0 mtu
1500
options=8
ether 02:00:60:00:07:0b
hwaddr 02:00:60:00:07:0b
nd6 options=29
media: Ethernet 10Gbase-T (10Gbase-T )
status: active
groups: epair

Any idea how to debug this to see if this is a hardware issue vs the
kernel deadlocking ? If I need to upgrade it to HEAD I can do so if need
be, but it would be nice to try and narrow it down first to a hardware
issue (its 2 different boxes) or software issue.

---Mike


-- 
---
Mike Tancsa, tel +1 519 651 3400 x203
Sentex Communications, m...@sentex.net
Providing Internet services since 1994 www.sentex.net
Cambridge, Ontario Canada
___
freebsd-stable@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

Re: Locale problem updating 10.3 to 11.1

[Eivind Nicolay Evensen]

On Wed, Feb 21, 2018 at 07:16:49AM -0500, Brandon Allbery wrote:
> A locale mapping is basically a lookup table (with complications for things
> like ß). A single-byte lookup table will be 256 entries, each holding one
> or more (because of combining characters) Unicode codepoints representing
> the mapping from the locale character set to the underlying common
> character set (Unicode). (There may also be a reverse lookup table for
> mapping Unicode codepoints to locale codepoints.)

That's fine, it doesn't make my life miserable such as it would when
directly using multibyte character sets, as long as it doesn't
negatively affect performance.

> Without this, every program would have to deal directly with every possible
> character set.

Or only handle what one cares about.

> (Complications include things like: depending on encoding/locale details,
> German lowercase ß will uppercase to either SS or ???.

While German is not my main language, I've never seen a situation
where an uppcase variant of ß would make sense, though I understand
the example.

> And that's one of the
> simpler ones; for some locales, things can get *really* weird. Not to
> mention fun stuff like Arabic having 4 representations of every character:
> initial, medial, final, standalone.)

Complications I don't want or need, nicely points out what I dislike
about unicode, although I can understand some os wanting to support it,
to be useful in more situations.



-- 
Eivind
___
freebsd-stable@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

Re: Locale problem updating 10.3 to 11.1

[Brandon Allbery]

A locale mapping is basically a lookup table (with complications for things
like ß). A single-byte lookup table will be 256 entries, each holding one
or more (because of combining characters) Unicode codepoints representing
the mapping from the locale character set to the underlying common
character set (Unicode). (There may also be a reverse lookup table for
mapping Unicode codepoints to locale codepoints.)

Without this, every program would have to deal directly with every possible
character set. With it, code can use Unicode internally and let the locale
system map to what to display, or in the other direction from what it has
read to the common representation.

(Complications include things like: depending on encoding/locale details,
German lowercase ß will uppercase to either SS or ẞ. And that's one of the
simpler ones; for some locales, things can get *really* weird. Not to
mention fun stuff like Arabic having 4 representations of every character:
initial, medial, final, standalone.)

Locale handling is seriously *nasty*. Unicode is also pretty nasty... but
it mostly manages the superset of individual locale nastinesses in about as
logical a way as possible given that locales are fundamentally illogical:
very few of them were designed, most grew organically and without regard
for rules or logic. (Esperanto locales being an exception... but even
Esperanto has developed some organic extensions with actual usage. It's how
humans work.)

On Wed, Feb 21, 2018 at 7:08 AM, Eivind Nicolay Evensen <
eivi...@terraplane.org> wrote:

> On Wed, Feb 21, 2018 at 01:03:01AM -0500, Brandon Allbery wrote:
> > On Tue, Feb 20, 2018 at 6:08 PM, Eivind Nicolay Evensen <
> > eivi...@terraplane.org> wrote:
> >
> > > However, since it was mentioned in a note starting with
> > > "Add support for unicode collation" I most likely didn't even read it
> > > since I'll never touch unicode.
> > >
> >
> > If you ever use anything other than LANG=C, you *are* touching Unicode.
>
> Well, I don't see multibyte characters with 8859-1, and
> multibyte is what I don't tolerate. I didn't even know
> that unicode could be single-byte character only sets.
>
>
>
>
> --
> Eivind
>



-- 
brandon s allbery kf8nh   sine nomine associates
allber...@gmail.com  ballb...@sinenomine.net
unix, openafs, kerberos, infrastructure, xmonadhttp://sinenomine.net
___
freebsd-stable@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

Re: Locale problem updating 10.3 to 11.1

[Eivind Nicolay Evensen]

On Wed, Feb 21, 2018 at 01:03:01AM -0500, Brandon Allbery wrote:
> On Tue, Feb 20, 2018 at 6:08 PM, Eivind Nicolay Evensen <
> eivi...@terraplane.org> wrote:
> 
> > However, since it was mentioned in a note starting with
> > "Add support for unicode collation" I most likely didn't even read it
> > since I'll never touch unicode.
> >
> 
> If you ever use anything other than LANG=C, you *are* touching Unicode.

Well, I don't see multibyte characters with 8859-1, and
multibyte is what I don't tolerate. I didn't even know
that unicode could be single-byte character only sets.




-- 
Eivind
___
freebsd-stable@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

Re: Locale problem updating 10.3 to 11.1

[Brandon Allbery]

Sorry, it's a bit trickier than that. I said "touching", not "using" as in
"UTF-8 locale".

Any locale system needs a common base to build locale descriptions from.
Unicode, or something functionally equivalent to it -- but given that we
need to support Unicode locales anyway, it makes the most sense ---
provides that common base.

So anything that affects Unicode handling implicitly affects the entire
locale system. LANG=C is the exception, because by the locale specification
it is the null mapping / what you would get if there were no locale system
at all.

On Wed, Feb 21, 2018 at 3:14 AM,  wrote:

> > > However, since it was mentioned in a note starting with
> > > "Add support for unicode collation" I most likely didn't even read it
> > > since I'll never touch unicode.
> > >
> >
> > If you ever use anything other than LANG=C, you *are* touching Unicode.
>
> % echo $LANG
> LANG: Undefined variable.
>
> % echo $LC_CTYPE
> nb_NO.ISO8859-1
>
> Works for me.
>
> But I did use a while to figure out what had happened between 10.3 and
> 11.1, since my Norwegian æøå suddenly stopped working (before changing
> LC_CTYPE to nb_NO.ISO8859-1).
>
> Steinar Haug, Nethelp consulting, sth...@nethelp.no
>



-- 
brandon s allbery kf8nh   sine nomine associates
allber...@gmail.com  ballb...@sinenomine.net
unix, openafs, kerberos, infrastructure, xmonadhttp://sinenomine.net
___
freebsd-stable@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

Re: Locale problem updating 10.3 to 11.1

[sthaug]

> > However, since it was mentioned in a note starting with
> > "Add support for unicode collation" I most likely didn't even read it
> > since I'll never touch unicode.
> >
> 
> If you ever use anything other than LANG=C, you *are* touching Unicode.

% echo $LANG
LANG: Undefined variable.

% echo $LC_CTYPE
nb_NO.ISO8859-1

Works for me.

But I did use a while to figure out what had happened between 10.3 and
11.1, since my Norwegian æøå suddenly stopped working (before changing
LC_CTYPE to nb_NO.ISO8859-1).

Steinar Haug, Nethelp consulting, sth...@nethelp.no
___
freebsd-stable@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"