NetBSD Problem Report #58002
From www@netbsd.org Wed Mar 6 12:20:26 2024
Return-Path: <www@netbsd.org>
Received: from mail.netbsd.org (mail.netbsd.org [199.233.217.200])
(using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits))
(Client CN "mail.NetBSD.org", Issuer "mail.NetBSD.org CA" (not verified))
by mollari.NetBSD.org (Postfix) with ESMTPS id BF7FE1A9239
for <gnats-bugs@gnats.NetBSD.org>; Wed, 6 Mar 2024 12:20:26 +0000 (UTC)
Message-Id: <20240306122025.842721A923A@mollari.NetBSD.org>
Date: Wed, 6 Mar 2024 12:20:25 +0000 (UTC)
From: marc.fege@uni-bonn.de
Reply-To: marc.fege@uni-bonn.de
To: gnats-bugs@NetBSD.org
Subject: NetBSD-VAX network stack or drivers broken since NetBSD 10?
X-Send-Pr-Version: www-1.0
>Number: 58002
>Category: port-vax
>Synopsis: NetBSD-VAX network stack or drivers broken since NetBSD 10?
>Confidential: no
>Severity: critical
>Priority: high
>Responsible: port-vax-maintainer
>State: open
>Class: sw-bug
>Submitter-Id: net
>Arrival-Date: Wed Mar 06 12:25:00 +0000 2024
>Last-Modified: Thu Mar 07 14:30:01 +0000 2024
>Originator: Marc Fege
>Release: 10RC5
>Organization:
>Environment:
NetBSD vax1.familie.fege.local 10.0_RC5 NetBSD 10.0_RC5 (GENERIC) #0: Tue Feb 27
05:27:39 UTC 2024 mkrepro@mkrepro.NetBSD.org:/usr/src/sys/arch/vax/compile/GEN
ERIC vax
>Description:
Dear community,
there seems to be a bug within NetBSDs network stack or drivers for some RCs of NetBSD 10 right now affecting the VAX port (see system messages below). I upgraded my SIMH-emulated VAX-780 from NetBSD 9.3 to 10 Beta without any issue back then over half a year ago, but now for some RCs of NetBSD 10 it doesn't matter which VAX (MicroVAX or VAX-780) and network adapter (de0, qe0) with appropiate configuration in the SIMH configuration I choose, the interface reports always "Inappropriate ioctl". For some reason no automatic IPv4 setting is done at boot time and even a manual ifconfig for setting up IPv4 addresses manually, won't help out. Neither IPv4 or IPv6 networking is possible beyond the localhost loopback device.
The SIMH-emulator and surrounding environment on the other hand never changed over that time. Odd...
I used my network configuration you'll find below for ages now since NetBSD 6, which is similar for other machines of other ports which I use since 2012. So, no configuration issue so far.
Any advices?
Thanks for your efforts!
Best regards,
Marc.
System messages (network addresses randomised):
[...]
Starting network.
Hostname: vax
Configuring network interfaces: de0ifconfig: SIOCGIFMEDIA: Inappropriate ioctl
for device.
Adding interface aliases:.
route: writing to routing socket: Network is unreachable
add net default: gateway 10.0.0.1: Network is unreachable
Waiting for duplicate address detection to finish...
Starting dhcpcd.
dhcpcd-9.4.1 starting
DUID ff:ff:ff:ff:ff:ff:ff:ff:ff:ff:ff:ff:ff:ff
de0: IAID ff:ff:ff:ff
de0: soliciting an IPv6 router
de0: Router Advertisement from fe80::ffff:ffff:ffff:ffff
de0: adding address ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff/64
de0: adding address fd00::ffff:ffff:ffff:ffff/64
de0: adding route to 2003:ffff:ffff:ffff::/64
de0: adding route to fd00::/64
de0: adding default route via fe80::ffff:ffff:ffff:ffff
timed out
[...]
=============
/etc/rc.conf:
hostname="vax"
defaultroute="10.0.0.1"
ip6mode="autohost"
dhcpcd="YES" dhcpcd_flags="-6"
=============
/etc/ifconfig.de0:
up
10.0.0.90 netmask 255.255.255.0 media autoselect
inet6 fd00::ffff:ffff:ffff prefixlen 64
=============
vax# ifconfig
de0: flags=0x8a43<UP,BROADCAST,RUNNING,ALLMULTI,SIMPLEX,MULTICAST> mtu 1500
address: ff:ff:ff:ff:ff:ff
inet6 fe80::ffff:ffff:ffff:ffff%de0/64 flags 0 scopeid 0x1
inet6 fd00::ffff:ffff:ffff/64 flags 0
inet6 2003:ffff:ffff:ffff:ffff:ffff:ffff:ffff/64 flags 0x40<AUTOCONF>
inet6 fd00::ffff:ffff:ffff:ffff/64 flags 0x40<AUTOCONF>
lo0: flags=0x8049<UP,LOOPBACK,RUNNING,MULTICAST> mtu 33176
status: active
inet6 ::1/128 flags 0x20<NODAD>
inet6 fe80::1%lo0/64 flags 0 scopeid 0x2
inet 127.0.0.1/8 flags 0
NetBSD vax1.familie.fege.local 10.0_RC5 NetBSD 10.0_RC5 (GENERIC) #0: Tue Feb 27
05:27:39 UTC 2024 mkrepro@mkrepro.NetBSD.org:/usr/src/sys/arch/vax/compile/GEN
ERIC vax
[...]
Starting network.
Hostname: vax
Configuring network interfaces: de0ifconfig: SIOCGIFMEDIA: Inappropriate ioctl
for device.
Adding interface aliases:.
route: writing to routing socket: Network is unreachable
add net default: gateway 10.0.0.1: Network is unreachable
Waiting for duplicate address detection to finish...
Starting dhcpcd.
dhcpcd-9.4.1 starting
DUID ff:ff:ff:ff:ff:ff:ff:ff:ff:ff:ff:ff:ff:ff
de0: IAID ff:ff:ff:ff
de0: soliciting an IPv6 router
de0: Router Advertisement from fe80::ffff:ffff:ffff:ffff
de0: adding address ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff/64
de0: adding address fd00::ffff:ffff:ffff:ffff/64
de0: adding route to 2003:ffff:ffff:ffff::/64
de0: adding route to fd00::/64
de0: adding default route via fe80::ffff:ffff:ffff:ffff
timed out
[...]
=============
/etc/rc.conf:
hostname="vax"
wpa_supplicant="NO"
defaultroute="10.0.0.1"
ip6mode="autohost"
dhcpcd="YES" dhcpcd_flags="-6"
=============
/etc/ifconfig.de0:
up
10.0.0.90 netmask 255.255.255.0 media autoselect
inet6 fd00::ffff:ffff:ffff prefixlen 64
=============
vax# ifconfig
de0: flags=0x8a43<UP,BROADCAST,RUNNING,ALLMULTI,SIMPLEX,MULTICAST> mtu 1500
address: ff:ff:ff:ff:ff:ff
inet6 fe80::ffff:ffff:ffff:ffff%de0/64 flags 0 scopeid 0x1
inet6 fd00::ffff:ffff:ffff/64 flags 0
inet6 2003:ffff:ffff:ffff:ffff:ffff:ffff:ffff/64 flags 0x40<AUTOCONF>
inet6 fd00::ffff:ffff:ffff:ffff/64 flags 0x40<AUTOCONF>
lo0: flags=0x8049<UP,LOOPBACK,RUNNING,MULTICAST> mtu 33176
status: active
inet6 ::1/128 flags 0x20<NODAD>
inet6 fe80::1%lo0/64 flags 0 scopeid 0x2
inet 127.0.0.1/8 flags 0
>How-To-Repeat:
Try to give set any IP address to a network adapter (with ifconfig).
>Fix:
>Audit-Trail:
From: "Jonathan A. Kollasch" <jakllsch@kollasch.net>
To: gnats-bugs@netbsd.org
Cc:
Subject: Re: port-vax/58002: NetBSD-VAX network stack or drivers broken since
NetBSD 10?
Date: Wed, 6 Mar 2024 08:04:58 -0600
It looks like you have a layer 2 multicast address in place of a layer 2 unicast
MAC address. Can you try again with say fe:ff:ff:ff:ff:ff instead? Or
is this real hardware?
From: "Fege, Marc Daniel" <marc.fege@uni-bonn.de>
To: gnats-bugs@netbsd.org, port-vax-maintainer@netbsd.org, gnats-admin@netbsd.org, netbsd-bugs@netbsd.org
Cc:
Subject: Re: port-vax/58002: NetBSD-VAX network stack or drivers broken since NetBSD 10?
Date: Wed, 6 Mar 2024 16:15:31 +0100
--b1=_carnOWI32t2ojiKliBQwWt2kut4EQPbGUOHIupdQE
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 8bit
Hello Jonathan,
many thanks for your quick reply!
All of those MAC or IPv6 addresses in the presented logs were masked
by me for privacy reasons, if that is it, what you mean. The emulated
hardware MAC is recognised by the OS/drivers as it should be as it did
with NetBSD 9 before. Thus were the IPv6 addresses calculated ind
given accordingly. Even the site-local 'fd00:....' and global
'2003:...' addresses could be obtained somehow.
However: now neither IPv6 traffic is possible through that device,
even if being able automatically set appropriate IPv6 addresses to
that port. Nor is it even possible to set the preset IPv4 address to
the network device "de0" at boot. Should I do so manually with
"ifconfig 10.0.0.90 255.255.255.0" after logon, ifconfig set it so as
requested, but no traffic will be transmitted through the port.
Neither IPv6, nor IPv4. Not even pings.
What I miss in checking the ethernet port with a simple "ifconfig"
command within the emulated/guest NetBSD-10 machine are some lines in
ifconfig's output such as the following that I have for my NetBSD 9
host machine, where the emulated port is connected to and which works
fine as desired. That could be a hint:
"
media: Ethernet autoselect (1000baseT full-duplex)
status: active
"
Neither of those are present since the change from NetBSD 9.3 to
NetBSD 10RC5 of the guest. It seems to be, that the driver is unable
to determin the used medium and thus does not recognize the medium as
"active.
– Marc.
Am Mittwoch, den 06.03.2024 um 15:10 schrieb Jonathan A. Kollasch:
The following reply was made to PR port-vax/58002; it has been noted
by GNATS.
From: "Jonathan A. Kollasch"
To: gnats-bugs@netbsd.org
Cc:
Subject: Re: port-vax/58002: NetBSD-VAX network stack or drivers
broken since
NetBSD 10?
Date: Wed, 6 Mar 2024 08:04:58 -0600
It looks like you have a layer 2 multicast address in place of a layer
2 unicast
MAC address. Can you try again with say fe:ff:ff:ff:ff:ff
instead? Or
is this real hardware?
--b1=_carnOWI32t2ojiKliBQwWt2kut4EQPbGUOHIupdQE
Content-Type: text/html; charset=utf-8
Content-Transfer-Encoding: quoted-printable
<html>
<head>
<style type=3D"text/css" id=3D"groupoffice-email-style">
/*
* Convert transparent color to hex value by given background
*/
h6 {
font-size: 11px;
line-height: 14px;
font-weight: bold;
color: rgba(0, 0, 0, 0.64);
}
h4 {
font-size: 14px;
line-height: 21px;
letter-spacing: 0.4px;
color: rgba(0, 0, 0, 0.87);
font-weight: normal;
}
h5 {
font-size: 12px;
color: rgba(0, 0, 0, 0.64);
font-weight: normal;
}
h3 {
font-size: 16px;
line-height: 21px;
font-weight: normal;
letter-spacing: 0.6px;
color: black;
}
h2 {
font-size: 21px;
line-height: 28px;
font-weight: normal;
letter-spacing: 0.6px;
color: black;
}
h1 {
font-size: 30px;
line-height: 35px;
font-weight: normal;
letter-spacing: 0.6px;
color: black;
}
body, p, span, div {
font-family: Helvetica, Arial, sans-serif;
font-size: 14px;
color: rgba(0, 0, 0, 0.87);
font-weight: normal;
line-height: 21px;
}
code {
border: 1px solid rgba(0, 0, 0, 0.12);
background-color: #fafafa;
padding: 7px;
margin: 14px 0;
display: block;
font-family: "Courier New", Courier, monospace;
color: black;
border-radius: 3.5px;
}
p {
margin: 0;
}
ul {
display: block;
list-style-type: disc;
list-style-position: outside;
margin: 0;
padding: 0 0 0 2em;
}
ol {
display: block;
list-style-type: decimal;
list-style-position: outside;
margin: 0;
padding: 0 0 0 2em;
}
ol > ol {
list-style-type: lower-alpha;
}
ol > ol > ol {
list-style-type: lower-roman;
}
@media (max-device-width: 1200px) {
/*
* Convert transparent color to hex value by given background
*/
h6 {
font-size: 13px;
line-height: 16px;
}
h4 {
font-size: 16px;
line-height: 24px;
}
h5 {
font-size: 14px;
}
h3 {
font-size: 18px;
line-height: 24px;
}
h2 {
font-size: 27px;
line-height: 32px;
}
h1 {
font-size: 24px;
line-height: 40px;
}
body, p, span, div {
font-size: 16px;
}
code {
padding: 8px;
margin: 16px 0;
display: block;
border-radius: 4px;
}
}
</style>
</head>
<body><div>Hello Jonathan,</div><div><br></div><div>many thanks for your qu=
ick reply!<br></div><div><br></div><div>All of those MAC or IPv6 addresses =
in the presented logs were masked by me for privacy reasons, if that is it,=
what you mean. The emulated hardware MAC is recognised by the OS/drivers a=
s it should be as it did with NetBSD 9 before. Thus were the IPv6 addresses=
<style></style>calculated ind given accordingly. Even the site-local 'fd00=
:....' and global '2003:...' addresses could be obtained somehow.</div><div=
>However: now neither IPv6 traffic is possible through that device, even if=
being able automatically set appropriate IPv6 addresses to that port. Nor =
is it even possible to set the preset IPv4 address to the network device "d=
e0" at boot. Should I do so manually with "ifconfig 10.0.0.90 255.255.255.0=
" after logon, ifconfig set it so as requested, but no traffic will be tran=
smitted through the port. Neither IPv6, nor IPv4. Not even pings.</div><div=
><br></div><div>What I miss in checking the ethernet port with a simple "if=
config" command within the emulated/guest NetBSD-10 machine are some lines =
in ifconfig's output such as the following that I have for my NetBSD 9 host=
machine, where the emulated port is connected to and which works fine as d=
esired. That could be a hint:</div><div><style></style></div><div>"</div><d=
iv> media: Ethernet autoselect (1000baseT full-duplex)<br=
> status: active<br></div><div>"<br></div><div><style></s=
tyle></div><div>Neither of those are present since the change from NetBSD 9=
.3 to NetBSD 10RC5 of the guest. It seems to be, that the driver is u=
nable to determin the used medium and thus does not recognize the medium as=
"active.</div><div><br></div><div>=E2=80=93 Marc.<br></div>
<br><br><br>Am Mittwoch, den 06.03.2024 um 15:10 schrieb Jonathan A. Kollas=
ch:<br><blockquote style=3D"border:0;border-left: 2px solid #22437f; paddin=
g:0px; margin:0px; padding-left:5px; margin-left: 5px; "><div class=3D"msg"=
>The following reply was made to PR port-vax/58002; it has been noted by GN=
ATS.<br>
<br>
From: "Jonathan A. Kollasch" <<a href=3D"mailto:jakllsch@kollasch.net" c=
lass=3D"normal-link normal-link-email" target=3D"_blank" rel=3D"noopener no=
referrer">jakllsch@kollasch.net</a>><br>
To: <a class=3D"normal-link" href=3D"mailto:gnats-bugs@netbsd.org">gnats-bu=
gs@netbsd.org</a><br>
Cc: <br>
Subject: Re: port-vax/58002: NetBSD-VAX network stack or drivers broken sin=
ce<br>
NetBSD 10?<br>
Date: Wed, 6 Mar 2024 08:04:58 -0600<br>
<br>
It looks like you have a layer 2 multicast address in place of a layer 2 u=
nicast<br>
MAC address. Can you try again with say fe:ff:ff:ff:ff:ff inste=
ad? Or<br>
is this real hardware?</div></blockquote></body></html>
--b1=_carnOWI32t2ojiKliBQwWt2kut4EQPbGUOHIupdQE--
From: "Fege, Marc Daniel" <marc.fege@uni-bonn.de>
To: gnats-bugs@netbsd.org, port-vax-maintainer@netbsd.org, gnats-admin@netbsd.org, netbsd-bugs@netbsd.org
Cc:
Subject: Re: port-vax/58002: NetBSD-VAX network stack or drivers broken since NetBSD 10?
Date: Thu, 7 Mar 2024 15:05:08 +0100
--b1=_ROvBNlyd4CoeqLhX1tn15l9K0nU1LqO0VWiiRgkLw
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 8bit
Hello folks,
little supplementum on that issue. It has been resolved. The problem
seems to be a config issue:
In the /etc/ifconfig.de0 the line
"10.0.0.90 netmask 255.255.255.0 media autoselect"
is no longer valid. The suffix "media autoselect" needs to be removed
since NetBSD-10. So no break in the network stack, but a break in the
POLA directive for the end user for thier configs might be expected
with NetBSD-10 on VAX. With other ports such as amd64, i386,
evbarm-earm6hf or arm64 I can report that the behaviour there did not
change from NetBSD-9 to 10. So if one might address this issue for the
VAX port, it would be warmly welcomed.
Therefore I consider this bug report solved and closed. Thank you all
for your feedback.
Keep up that good work!
- Marc.
Am Mittwoch, den 06.03.2024 um 16:20 schrieb Fege, Marc Daniel:
The following reply was made to PR port-vax/58002; it has been noted
by GNATS.
From: "Fege, Marc Daniel"
To: gnats-bugs@netbsd.org, port-vax-maintainer@netbsd.org,
gnats-admin@netbsd.org, netbsd-bugs@netbsd.org
Cc:
Subject: Re: port-vax/58002: NetBSD-VAX network stack or drivers
broken since NetBSD 10?
Date: Wed, 6 Mar 2024 16:15:31 +0100
--b1=_carnOWI32t2ojiKliBQwWt2kut4EQPbGUOHIupdQE
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 8bit
Hello Jonathan,
many thanks for your quick reply!
All of those MAC or IPv6 addresses in the presented logs were masked
by me for privacy reasons, if that is it, what you mean. The emulated
hardware MAC is recognised by the OS/drivers as it should be as it did
with NetBSD 9 before. Thus were the IPv6 addresses calculated ind
given accordingly. Even the site-local 'fd00:....' and global
'2003:...' addresses could be obtained somehow.
However: now neither IPv6 traffic is possible through that device,
even if being able automatically set appropriate IPv6 addresses to
that port. Nor is it even possible to set the preset IPv4 address to
the network device "de0" at boot. Should I do so manually with
"ifconfig 10.0.0.90 255.255.255.0" after logon, ifconfig set it so as
requested, but no traffic will be transmitted through the port.
Neither IPv6, nor IPv4. Not even pings.
What I miss in checking the ethernet port with a simple "ifconfig"
command within the emulated/guest NetBSD-10 machine are some lines in
ifconfig's output such as the following that I have for my NetBSD 9
host machine, where the emulated port is connected to and which works
fine as desired. That could be a hint:
"
media: Ethernet autoselect (1000baseT full-duplex)
status: active
"
Neither of those are present since the change from NetBSD 9.3 to
NetBSD 10RC5 of the guest. It seems to be, that the driver is unable
to determin the used medium and thus does not recognize the medium as
"active.
– Marc.
Am Mittwoch, den 06.03.2024 um 15:10 schrieb Jonathan A. Kollasch:
The following reply was made to PR port-vax/58002; it has been noted
by GNATS.
From: "Jonathan A. Kollasch"
To: gnats-bugs@netbsd.org
Cc:
Subject: Re: port-vax/58002: NetBSD-VAX network stack or drivers
broken since
NetBSD 10?
Date: Wed, 6 Mar 2024 08:04:58 -0600
It looks like you have a layer 2 multicast address in place of a layer
2 unicast
MAC address. Can you try again with say fe:ff:ff:ff:ff:ff
instead? Or
is this real hardware?
--b1=_carnOWI32t2ojiKliBQwWt2kut4EQPbGUOHIupdQE
Content-Type: text/html; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Hello Jonathan,
many thanks for your qu=
ick reply!
All of those MAC or IPv6 addresses =
in the presented logs were masked by me for privacy reasons, if that
is it,=
what you mean. The emulated hardware MAC is recognised by the
OS/drivers a=
s it should be as it did with NetBSD 9 before. Thus were the IPv6
addresses=
calculated ind given accordingly. Even the site-local 'fd00=
:....' and global '2003:...' addresses could be obtained somehow.
>However: now neither IPv6 traffic is possible through that device,
even if=
being able automatically set appropriate IPv6 addresses to that
port. Nor =
is it even possible to set the preset IPv4 address to the network
device "d=
e0" at boot. Should I do so manually with "ifconfig 10.0.0.90
255.255.255.0=
" after logon, ifconfig set it so as requested, but no traffic will be
tran=
smitted through the port. Neither IPv6, nor IPv4. Not even pings.
>
What I miss in checking the ethernet port with a simple "if=
config" command within the emulated/guest NetBSD-10 machine are some
lines =
in ifconfig's output such as the following that I have for my NetBSD 9
host=
machine, where the emulated port is connected to and which works
fine as d=
esired. That could be a hint:
" media: Ethernet autoselect (1000baseT full-duplex)
> status: active
"
Neither of those are present since the change from NetBSD 9=
.3 to NetBSD 10RC5 of the guest. It seems to be, that the driver
is u=
nable to determin the used medium and thus does not recognize the
medium as=
"active.
=E2=80=93 Marc.
Am Mittwoch, den 06.03.2024 um 15:10 schrieb Jonathan A. Kollas=
ch:
>The following reply was made to PR port-vax/58002; it has been noted
by GN=
ATS.
From: "Jonathan A. Kollasch" <jakllsch@kollasch.net" c=
lass=3D"normal-link normal-link-email" target=3D"_blank"
rel=3D"noopener no=
referrer">jakllsch@kollasch.net>
To: gnats-bugs@netbsd.org">gnats-bu=
gs@netbsd.org
Cc:
Subject: Re: port-vax/58002: NetBSD-VAX network stack or drivers
broken sin=
ce
NetBSD 10?
Date: Wed, 6 Mar 2024 08:04:58 -0600
It looks like you have a layer 2 multicast address in place of a
layer 2 u=
nicast
MAC address. Can you try again with say
fe:ff:ff:ff:ff:ff inste=
ad? Or
is this real hardware?
--b1=_carnOWI32t2ojiKliBQwWt2kut4EQPbGUOHIupdQE--
--b1=_ROvBNlyd4CoeqLhX1tn15l9K0nU1LqO0VWiiRgkLw
Content-Type: text/html; charset=utf-8
Content-Transfer-Encoding: quoted-printable
<html>
<head>
<style type=3D"text/css" id=3D"groupoffice-email-style">
/*
* Convert transparent color to hex value by given background
*/
h6 {
font-size: 11px;
line-height: 14px;
font-weight: bold;
color: rgba(0, 0, 0, 0.64);
}
h4 {
font-size: 14px;
line-height: 21px;
letter-spacing: 0.4px;
color: rgba(0, 0, 0, 0.87);
font-weight: normal;
}
h5 {
font-size: 12px;
color: rgba(0, 0, 0, 0.64);
font-weight: normal;
}
h3 {
font-size: 16px;
line-height: 21px;
font-weight: normal;
letter-spacing: 0.6px;
color: black;
}
h2 {
font-size: 21px;
line-height: 28px;
font-weight: normal;
letter-spacing: 0.6px;
color: black;
}
h1 {
font-size: 30px;
line-height: 35px;
font-weight: normal;
letter-spacing: 0.6px;
color: black;
}
body, p, span, div {
font-family: Helvetica, Arial, sans-serif;
font-size: 14px;
color: rgba(0, 0, 0, 0.87);
font-weight: normal;
line-height: 21px;
}
code {
border: 1px solid rgba(0, 0, 0, 0.12);
background-color: #fafafa;
padding: 7px;
margin: 14px 0;
display: block;
font-family: "Courier New", Courier, monospace;
color: black;
border-radius: 3.5px;
}
p {
margin: 0;
}
ul {
display: block;
list-style-type: disc;
list-style-position: outside;
margin: 0;
padding: 0 0 0 2em;
}
ol {
display: block;
list-style-type: decimal;
list-style-position: outside;
margin: 0;
padding: 0 0 0 2em;
}
ol > ol {
list-style-type: lower-alpha;
}
ol > ol > ol {
list-style-type: lower-roman;
}
@media (max-device-width: 1200px) {
/*
* Convert transparent color to hex value by given background
*/
h6 {
font-size: 13px;
line-height: 16px;
}
h4 {
font-size: 16px;
line-height: 24px;
}
h5 {
font-size: 14px;
}
h3 {
font-size: 18px;
line-height: 24px;
}
h2 {
font-size: 27px;
line-height: 32px;
}
h1 {
font-size: 24px;
line-height: 40px;
}
body, p, span, div {
font-size: 16px;
}
code {
padding: 8px;
margin: 16px 0;
display: block;
border-radius: 4px;
}
}
</style>
</head>
<body><div>Hello folks,</div><div><br></div><div>little supplementum on tha=
t issue. It has been resolved. The problem seems to be a config issue:<br><=
/div><div>In the /etc/ifconfig.de0 the line</div><div><br></div><div>"10.0.=
0.90 netmask 255.255.255.0 media autoselect"</div><div><br></div><div>is no=
longer valid. The suffix "media autoselect" needs to be removed since NetB=
SD-10. So no break in the network stack, but a break in the POLA directive =
for the end user for thier configs might be expected with NetBSD-10 on VAX.=
With other ports such as amd64, i386, evbarm-earm6hf or arm64 I can report=
that the behaviour there did not change from NetBSD-9 to 10. So if one mig=
ht address this issue for the VAX port, it would be warmly welcomed.</div><=
div><br></div><div>Therefore I consider this bug report solved and closed. =
Thank you all for your feedback.</div><div>Keep up that good work!</div><di=
v><br></div><div>- Marc.<br></div><div><style></style></div><br><br>Am Mitt=
woch, den 06.03.2024 um 16:20 schrieb Fege, Marc Daniel:<br><blockquote sty=
le=3D"border:0;border-left: 2px solid #22437f; padding:0px; margin:0px; pad=
ding-left:5px; margin-left: 5px; "><div class=3D"msg">The following reply w=
as made to PR port-vax/58002; it has been noted by GNATS.<br>
<br>
From: "Fege, Marc Daniel" <<a href=3D"mailto:marc.fege@uni-bonn.de" clas=
s=3D"normal-link normal-link-email" target=3D"_blank" rel=3D"noopener noref=
errer">marc.fege@uni-bonn.de</a>><br>
To: <a href=3D"mailto:gnats-bugs@netbsd.org" class=3D"normal-link normal-li=
nk-email" target=3D"_blank" rel=3D"noopener noreferrer">gnats-bugs@netbsd.o=
rg</a>, <a href=3D"mailto:port-vax-maintainer@netbsd.org" class=3D"normal-l=
ink normal-link-email" target=3D"_blank" rel=3D"noopener noreferrer">port-v=
ax-maintainer@netbsd.org</a>, <a href=3D"mailto:gnats-admin@netbsd.org" cla=
ss=3D"normal-link normal-link-email" target=3D"_blank" rel=3D"noopener nore=
ferrer">gnats-admin@netbsd.org</a>, <a class=3D"normal-link" href=3D"mailto=
:netbsd-bugs@netbsd.org">netbsd-bugs@netbsd.org</a><br>
Cc: <br>
Subject: Re: port-vax/58002: NetBSD-VAX network stack or drivers broken sin=
ce NetBSD 10?<br>
Date: Wed, 6 Mar 2024 16:15:31 +0100<br>
<br>
--b1=3D_carnOWI32t2ojiKliBQwWt2kut4EQPbGUOHIupdQE<br>
Content-Type: text/plain; charset=3Dutf-8<br>
Content-Transfer-Encoding: 8bit<br>
<br>
Hello Jonathan,<br>
<br>
many thanks for your quick reply!<br>
<br>
<br>
All of those MAC or IPv6 addresses in the presented logs were masked<br>
by me for privacy reasons, if that is it, what you mean. The emulated<br>
hardware MAC is recognised by the OS/drivers as it should be as it did<br>
with NetBSD 9 before. Thus were the IPv6 addresses calculated ind<br>
given accordingly. Even the site-local 'fd00:....' and global<br>
'2003:...' addresses could be obtained somehow.<br>
However: now neither IPv6 traffic is possible through that device,<br>
even if being able automatically set appropriate IPv6 addresses to<br>
that port. Nor is it even possible to set the preset IPv4 address to<br>
the network device "de0" at boot. Should I do so manually with<br>
"ifconfig 10.0.0.90 255.255.255.0" after logon, ifconfig set it so as<br>
requested, but no traffic will be transmitted through the port.<br>
Neither IPv6, nor IPv4. Not even pings.<br>
<br>
What I miss in checking the ethernet port with a simple "ifconfig"<br>
command within the emulated/guest NetBSD-10 machine are some lines in<br>
ifconfig's output such as the following that I have for my NetBSD 9<br>
host machine, where the emulated port is connected to and which works<br>
fine as desired. That could be a hint:<br>
<br>
"<br>
media: Ethernet autoselect (1000baseT full-duplex)<br>
status: active<br>
<br>
"<br>
<br>
<br>
Neither of those are present since the change from NetBSD 9.3 to<br>
NetBSD 10RC5 of the guest. It seems to be, that the driver is unable=
<br>
to determin the used medium and thus does not recognize the medium as<br>
"active.<br>
<br>
=E2=80=93 Marc.<br>
<br>
<br>
<br>
Am Mittwoch, den 06.03.2024 um 15:10 schrieb Jonathan A. Kollasch:<br>
<br>
<br>
<br>
The following reply was made to PR port-vax/58002; it has been noted<br>
by GNATS.<br>
<br>
From: "Jonathan A. Kollasch" <br>
To: <a class=3D"normal-link" href=3D"mailto:gnats-bugs@netbsd.org">gnats-b=
ugs@netbsd.org</a><br>
Cc: <br>
Subject: Re: port-vax/58002: NetBSD-VAX network stack or drivers<br>
broken since<br>
NetBSD 10?<br>
Date: Wed, 6 Mar 2024 08:04:58 -0600<br>
<br>
It looks like you have a layer 2 multicast address in place of a layer<br>
2 unicast<br>
MAC address. Can you try again with say fe:ff:ff:ff:ff:ff<br>
instead? Or<br>
is this real hardware?<br>
<br>
--b1=3D_carnOWI32t2ojiKliBQwWt2kut4EQPbGUOHIupdQE<br>
Content-Type: text/html; charset=3Dutf-8<br>
Content-Transfer-Encoding: quoted-printable<br>
<br>
<html><br>
<head><br>
<style type=3D3D"text/css" id=3D3D"groupoffice-email-style"><br>
/*<br>
* Convert transparent color to hex value by given background<br=
>
*/<br>
h6 {<br>
font-size: 11px;<br>
line-height: 14px;<br>
font-weight: bold;<br>
color: rgba(0, 0, 0, 0.64);<br>
}<br>
h4 {<br>
font-size: 14px;<br>
line-height: 21px;<br>
letter-spacing: 0.4px;<br>
color: rgba(0, 0, 0, 0.87);<br>
font-weight: normal;<br>
}<br>
h5 {<br>
font-size: 12px;<br>
color: rgba(0, 0, 0, 0.64);<br>
font-weight: normal;<br>
}<br>
h3 {<br>
font-size: 16px;<br>
line-height: 21px;<br>
font-weight: normal;<br>
letter-spacing: 0.6px;<br>
color: black;<br>
}<br>
h2 {<br>
font-size: 21px;<br>
line-height: 28px;<br>
font-weight: normal;<br>
letter-spacing: 0.6px;<br>
color: black;<br>
}<br>
h1 {<br>
font-size: 30px;<br>
line-height: 35px;<br>
font-weight: normal;<br>
letter-spacing: 0.6px;<br>
color: black;<br>
}<br>
body, p, span, div {<br>
font-family: Helvetica, Arial, sans-serif;<br>
font-size: 14px;<br>
color: rgba(0, 0, 0, 0.87);<br>
font-weight: normal;<br>
line-height: 21px;<br>
}<br>
code {<br>
border: 1px solid rgba(0, 0, 0, 0.12);<br>
background-color: #fafafa;<br>
padding: 7px;<br>
margin: 14px 0;<br>
display: block;<br>
font-family: "Courier New", Courier, monospace;<br>
color: black;<br>
border-radius: 3.5px;<br>
}<br>
p {<br>
margin: 0;<br>
}<br>
ul {<br>
display: block;<br>
list-style-type: disc;<br>
list-style-position: outside;<br>
margin: 0;<br>
padding: 0 0 0 2em;<br>
}<br>
ol {<br>
display: block;<br>
list-style-type: decimal;<br>
list-style-position: outside;<br>
margin: 0;<br>
padding: 0 0 0 2em;<br>
}<br>
ol > ol {<br>
list-style-type: lower-alpha;<br>
}<br>
ol > ol > ol {<br>
list-style-type: lower-roman;<br>
}<br>
@media (max-device-width: 1200px) {<br>
/*<br>
* Convert transparent color to hex value by given background<b=
r>
*/<br>
h6 {<br>
font-size: 13px;<br>
line-height: 16px;<br>
}<br>
h4 {<br>
font-size: 16px;<br>
line-height: 24px;<br>
}<br>
h5 {<br>
font-size: 14px;<br>
}<br>
h3 {<br>
font-size: 18px;<br>
line-height: 24px;<br>
}<br>
h2 {<br>
font-size: 27px;<br>
line-height: 32px;<br>
}<br>
h1 {<br>
font-size: 24px;<br>
line-height: 40px;<br>
}<br>
body, p, span, div {<br>
font-size: 16px;<br>
}<br>
code {<br>
padding: 8px;<br>
margin: 16px 0;<br>
display: block;<br>
border-radius: 4px;<br>
}<br>
}<br>
</style><br>
</head><br>
<body><div>Hello Jonathan,</div><div><br><=
;/div><div>many thanks for your qu=3D<br>
ick reply!<br></div><div><br></div><div&g=
t;All of those MAC or IPv6 addresses =3D<br>
in the presented logs were masked by me for privacy reasons, if that is it=
,=3D<br>
what you mean. The emulated hardware MAC is recognised by the O=
S/drivers a=3D<br>
s it should be as it did with NetBSD 9 before. Thus were the IPv6 addresse=
s=3D<br>
<style></style>calculated ind given accordingly. Ev=
en the site-local 'fd00=3D<br>
:....' and global '2003:...' addresses could be obtained somehow.</div&=
gt;<div=3D<br>
>However: now neither IPv6 traffic is possible through that device, eve=
n if=3D<br>
being able automatically set appropriate IPv6 addresses to that=
port. Nor =3D<br>
is it even possible to set the preset IPv4 address to the network device "=
d=3D<br>
e0" at boot. Should I do so manually with "ifconfig 10.0.0.90 255.255.255.=
0=3D<br>
" after logon, ifconfig set it so as requested, but no traffic will be tra=
n=3D<br>
smitted through the port. Neither IPv6, nor IPv4. Not even pings.</div&=
gt;<div=3D<br>
><br></div><div>What I miss in checking the ethernet =
port with a simple "if=3D<br>
config" command within the emulated/guest NetBSD-10 machine are some lines=
=3D<br>
in ifconfig's output such as the following that I have for my NetBSD 9 hos=
t=3D<br>
machine, where the emulated port is connected to and which work=
s fine as d=3D<br>
esired. That could be a hint:</div><div><style></styl=
e></div><div>"</div><d=3D<br>
iv>&nbsp;&nbsp;&nbsp; media: Ethernet autoselect (1000baseT=
full-duplex)<br=3D<br>
>&nbsp;&nbsp;&nbsp; status: active<br></div><=
;div>"<br></div><div><style></s=3D<br>
tyle></div><div>Neither of those are present since the chan=
ge from NetBSD 9=3D<br>
.3 to NetBSD 10RC5 of the guest.&nbsp; It seems to be, that the driver=
is u=3D<br>
nable to determin the used medium and thus does not recognize the medium a=
s=3D<br>
"active.</div><div><br></div><div>=
;=3DE2=3D80=3D93 Marc.<br></div><br>
<br><br><br>Am Mittwoch, den 06.03.2024 um 15:10 schrieb=
Jonathan A. Kollas=3D<br>
ch:<br><blockquote style=3D3D"border:0;border-left: 2px solid #22=
437f; paddin=3D<br>
g:0px; margin:0px; padding-left:5px; margin-left: 5px; "><div class=
=3D3D"msg"=3D<br>
>The following reply was made to PR port-vax/58002; it has been noted b=
y GN=3D<br>
ATS.<br><br>
<br><br>
From: "Jonathan A. Kollasch" &lt;<a href=3D3D"<a href=3D"mailto:jak=
llsch@kollasch.net" class=3D"normal-link normal-link-email" target=3D"_blan=
k" rel=3D"noopener noreferrer">jakllsch@kollasch.net</a>" c=3D<br>
lass=3D3D"normal-link normal-link-email" target=3D3D"_blank" rel=3D3D"noop=
ener no=3D<br>
referrer"><a href=3D"mailto:jakllsch@kollasch.net" class=3D"normal-link=
normal-link-email" target=3D"_blank" rel=3D"noopener noreferrer">jakllsch@=
kollasch.net</a></a>&gt;<br><br>
To: <a class=3D3D"normal-link" href=3D3D"<a href=3D"mailto:gnats-bugs@n=
etbsd.org" class=3D"normal-link normal-link-email" target=3D"_blank" rel=3D=
"noopener noreferrer">gnats-bugs@netbsd.org</a>">gnats-bu=3D<br>
<a href=3D"mailto:gs@netbsd.org" class=3D"normal-link normal-link-email" t=
arget=3D"_blank" rel=3D"noopener noreferrer">gs@netbsd.org</a></a><=
;br><br>
Cc: <br><br>
Subject: Re: port-vax/58002: NetBSD-VAX network stack or drivers broken si=
n=3D<br>
ce<br><br>
NetBSD 10?<br><br>
Date: Wed, 6 Mar 2024 08:04:58 -0600<br><br>
<br><br>
It looks like you have a layer 2 multicast address in place of =
a layer 2 u=3D<br>
nicast<br><br>
MAC address.&nbsp;&nbsp;Can you try again with say fe:f=
f:ff:ff:ff:ff inste=3D<br>
ad?&nbsp;&nbsp;Or<br><br>
is this real hardware?</div></blockquote></body&=
gt;</html><br>
<br>
--b1=3D_carnOWI32t2ojiKliBQwWt2kut4EQPbGUOHIupdQE--</div></blockquote></bo=
dy></html>
--b1=_ROvBNlyd4CoeqLhX1tn15l9K0nU1LqO0VWiiRgkLw--
From: Martin Husemann <martin@duskware.de>
To: "Fege, Marc Daniel" <marc.fege@uni-bonn.de>
Cc: gnats-bugs@netbsd.org
Subject: Re: port-vax/58002: NetBSD-VAX network stack or drivers broken since
NetBSD 10?
Date: Thu, 7 Mar 2024 15:25:53 +0100
On Thu, Mar 07, 2024 at 03:05:08PM +0100, Fege, Marc Daniel wrote:
> "10.0.0.90 netmask 255.255.255.0 media autoselect"
>
> is no longer valid. The suffix "media autoselect" needs to be removed
> since NetBSD-10.
Whether media selection is supported is a property of the concrete device
driver (some hardware simply can not offer any choice).
Does
ifconfig -m de0
show a list of available media in 9.x? I only have a 8.x MicroVax3900 simh
instance handy and the qt0 in there does not show any media options,
and also
ifconfig qt0 $some.ip media autoselect
causes the same error message from ifconfig (SIOCGIFMEDIA:
Inappropriate ioctl for device).
Martin
(Contact us)
$NetBSD: query-full-pr,v 1.47 2022/09/11 19:34:41 kim Exp $
$NetBSD: gnats_config.sh,v 1.9 2014/08/02 14:16:04 spz Exp $
Copyright © 1994-2024
The NetBSD Foundation, Inc. ALL RIGHTS RESERVED.
Home