SEO

Does the sun be 10/100 sbus ethernet card work on sparc64? It works
just fine more me on sparc32, but on sparc64 I get DMA errors and the
card is unusable. According to the docs, the NIC is supported on both
platforms.

Thanks.

Bob Healey

> Does the sun be 10/100 sbus ethernet card work on sparc64? It works
> just fine more me on sparc32, but on sparc64 I get DMA errors and the
> card is unusable. According to the docs, the NIC is supported on both
> platforms.

It works, but on some PCI cards with on-board PCI bridges, interrupt
swizzling problems prevent the interfaces from working.

This is a known issue for which there is currently no good workaround,
but which might eventually be fixed.

If your card is an SBus card, then this is a new issue; please report
the problem with sendbug.

Miod Vallat wrote:
>>Does the sun be 10/100 sbus ethernet card work on sparc64? It works
>>just fine more me on sparc32, but on sparc64 I get DMA errors and the
>>card is unusable. According to the docs, the NIC is supported on both
>>platforms.
>
>
> It works, but on some PCI cards with on-board PCI bridges, interrupt
> swizzling problems prevent the interfaces from working.
>
> This is a known issue for which there is currently no good workaround,
> but which might eventually be fixed.
>
> If your card is an SBus card, then this is a new issue; please report
> the problem with sendbug.

This is the sbus card - I don't have the most recent flash revision on
my system, so I want to update that first. The machine is an Ultra 1.
The card works just fine on a Sparc-Station 5. I am running 3.6 with
all the erratta applied.

Bob Healey

In article <cteufu$2fg$1@misc-cct.server.rpi.edu>,
Bob Healey <healer@rpi.edu> wrote:
>Miod Vallat wrote:
>>>Does the sun be 10/100 sbus ethernet card work on sparc64? It works
>>>just fine more me on sparc32, but on sparc64 I get DMA errors and the
>>>card is unusable. According to the docs, the NIC is supported on both
>>>platforms.
>>
>>
>> It works, but on some PCI cards with on-board PCI bridges, interrupt
>> swizzling problems prevent the interfaces from working.
>>
>> This is a known issue for which there is currently no good workaround,
>> but which might eventually be fixed.
>>
>> If your card is an SBus card, then this is a new issue; please report
>> the problem with sendbug.
>
>This is the sbus card - I don't have the most recent flash revision on
>my system, so I want to update that first. The machine is an Ultra 1.
>The card works just fine on a Sparc-Station 5. I am running 3.6 with
>all the erratta applied.

Hmm ... are these the ones which identify as hme*? I've got a
four-port one in an Ultra-1/140, and it is working without obvious
problems.

But that system is running OpenBSD 3.4.

Here is the (partial) output of "ifconfig -a", showing only the
two ports currently in service:

================================================== ====================
hme0: flags=8063<UP,BROADCAST,NOTRAILERS,RUNNING,MULTICA ST> mtu 1500
address: 08:00:20:bd:e5:00
media: Ethernet autoselect (100baseTX)
status: active
inet 10.0.0.50 netmask 0xffffff00 broadcast 10.0.0.255
inet6 fe80::a00:20ff:febd:e500%hme0 prefixlen 64 scopeid 0x2
hme1: flags=8863<UP,BROADCAST,NOTRAILERS,RUNNING,SIMPLEX ,MULTICAST> mtu
1500
address: 08:00:20:bd:e5:01
media: Ethernet autoselect (100baseTX full-duplex)
status: active
inet 10.4.14.50 netmask 0xffffff00 broadcast 10.4.14.255
inet6 fe80::a00:20ff:febd:e501%hme1 prefixlen 64 scopeid 0x3
================================================== ====================

And -- I accessed it through hme0 from an Ultra-2 running
Solaris 2.8, FWIW.

Both are internal subnets. The external is on the le0
interface, since it is not capable of being hit with as much bandwidth
anyway.

Could it be something wrong with your specific card? I got one
(totally dead) card from an eBay vendor. He replaced it without
question. It was so bad that it even locked up the system so it could
not show the OpenBoot prompt, let alone actually boot fully.

Good Luck,
DoN.
--
Email: <dnichols@d-and-d.com> | Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
--- Black Holes are where God is dividing by zero ---

DoN. Nichols wrote:
> In article <cteufu$2fg$1@misc-cct.server.rpi.edu>,
> Bob Healey <healer@rpi.edu> wrote:
>
>>Miod Vallat wrote:
>>
>>>>Does the sun be 10/100 sbus ethernet card work on sparc64? It works
>>>>just fine more me on sparc32, but on sparc64 I get DMA errors and the
>>>>card is unusable. According to the docs, the NIC is supported on both
>>>>platforms.
>>>
>>>
>>>It works, but on some PCI cards with on-board PCI bridges, interrupt
>>>swizzling problems prevent the interfaces from working.
>>>
>>>This is a known issue for which there is currently no good workaround,
>>>but which might eventually be fixed.
>>>
>>>If your card is an SBus card, then this is a new issue; please report
>>>the problem with sendbug.
>>
>>This is the sbus card - I don't have the most recent flash revision on
>>my system, so I want to update that first. The machine is an Ultra 1.
>>The card works just fine on a Sparc-Station 5. I am running 3.6 with
>>all the erratta applied.
>
>
> Hmm ... are these the ones which identify as hme*? I've got a
> four-port one in an Ultra-1/140, and it is working without obvious
> problems.
>
> But that system is running OpenBSD 3.4.
>
> Here is the (partial) output of "ifconfig -a", showing only the
> two ports currently in service:
>
> ================================================== ====================
> hme0: flags=8063<UP,BROADCAST,NOTRAILERS,RUNNING,MULTICA ST> mtu 1500
> address: 08:00:20:bd:e5:00
> media: Ethernet autoselect (100baseTX)
> status: active
> inet 10.0.0.50 netmask 0xffffff00 broadcast 10.0.0.255
> inet6 fe80::a00:20ff:febd:e500%hme0 prefixlen 64 scopeid 0x2
> hme1: flags=8863<UP,BROADCAST,NOTRAILERS,RUNNING,SIMPLEX ,MULTICAST> mtu
> 1500
> address: 08:00:20:bd:e5:01
> media: Ethernet autoselect (100baseTX full-duplex)
> status: active
> inet 10.4.14.50 netmask 0xffffff00 broadcast 10.4.14.255
> inet6 fe80::a00:20ff:febd:e501%hme1 prefixlen 64 scopeid 0x3
> ================================================== ====================
>
> And -- I accessed it through hme0 from an Ultra-2 running
> Solaris 2.8, FWIW.
>
> Both are internal subnets. The external is on the le0
> interface, since it is not capable of being hit with as much bandwidth
> anyway.
>
> Could it be something wrong with your specific card? I got one
> (totally dead) card from an eBay vendor. He replaced it without
> question. It was so bad that it even locked up the system so it could
> not show the OpenBoot prompt, let alone actually boot fully.
>
> Good Luck,
> DoN.

I had the card in a sparcstation 5 running as be0 just fine. I replaced
the sparcstation 5 with an ultra-1 and the kernel recognizes the card is
there, but i get a massive string of DMA errors and can't use the card.

Bob Healey

In article <cthfc2$9bl$1@misc-cct.server.rpi.edu>,
Bob Healey <healer@rpi.edu> wrote:
>DoN. Nichols wrote:
>> In article <cteufu$2fg$1@misc-cct.server.rpi.edu>,
>> Bob Healey <healer@rpi.edu> wrote:
>>
>>>Miod Vallat wrote:
>>>
>>>>>Does the sun be 10/100 sbus ethernet card work on sparc64? It works
>>>>>just fine more me on sparc32, but on sparc64 I get DMA errors and the
>>>>>card is unusable. According to the docs, the NIC is supported on both
>>>>>platforms.

[ ... ]

>> Hmm ... are these the ones which identify as hme*? I've got a
>> four-port one in an Ultra-1/140, and it is working without obvious
>> problems.

[ ... ]

>I had the card in a sparcstation 5 running as be0 just fine.

O.K. If it identifies itself as "be0", then it is not the same
card which I have been using, so I don't know. Mine shows up in dmesg as:

================================================== ====================
SUNW,bpp at sbus0 slot 14 offset 0xc800000 vector 22 ipl 2 not configured
hme0 at sbus0 slot 0 offset 0x8c00000 vector 4 ipl 6: address 08:00:20:bd:e5:00
ukphy0 at hme0 phy 1: Generic IEEE 802.3u media interface
ukphy0: OUI 0x00601d, model 0x000c, rev. 1
hme1 at sbus0 slot 0 offset 0x8c10000 vector 4 ipl 6: address 08:00:20:bd:e5:01
ukphy1 at hme1 phy 1: Generic IEEE 802.3u media interface
ukphy1: OUI 0x00601d, model 0x000c, rev. 1
hme2 at sbus0 slot 0 offset 0x8c20000 vector 4 ipl 6: address 08:00:20:bd:e5:02
ukphy2 at hme2 phy 1: Generic IEEE 802.3u media interface
ukphy2: OUI 0x00601d, model 0x000c, rev. 1
hme3 at sbus0 slot 0 offset 0x8c30000 vector 4 ipl 6: address 08:00:20:bd:e5:03
ukphy3 at hme3 phy 1: Generic IEEE 802.3u media interface
ukphy3: OUI 0x00601d, model 0x000c, rev. 1
================================================== ====================

And is either a 501-4302 or a 501-4837/501-5443. I can't tell
you which without bringing the system down and removing the card to read
its barcode at this point -- and since it is serving in a firewall, I
would rather not do that. There are similar ones with a single
interface instead of the four-holer which I am using, and I am using
some of these in other systems (SS-5 and SS-20), but all of those are
running some form of Solaris, not OpenBSD.

What card identifies itself as a "be0"? Can you point to the
page in Sun's FEH (Field Engineer's Handbook)? At least with it moving
from a SS-5 to an Ultra-1, it should be an Sbus card. But I didn't know
that there was another 100BaseT/10BaseT card in their line.

> I replaced
>the sparcstation 5 with an ultra-1 and the kernel recognizes the card is
>there, but i get a massive string of DMA errors and can't use the card.

Since you are apparently talking about a different card from the
one which I am using, I guess that I can't be any help.

Best of luck,
DoN.
--
Email: <dnichols@d-and-d.com> | Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
--- Black Holes are where God is dividing by zero ---

DoN. Nichols wrote:
> In article <cthfc2$9bl$1@misc-cct.server.rpi.edu>,
> Bob Healey <healer@rpi.edu> wrote:
>
>>DoN. Nichols wrote:
>>
>>>In article <cteufu$2fg$1@misc-cct.server.rpi.edu>,
>>>Bob Healey <healer@rpi.edu> wrote:
>>>
>>>
>>>>Miod Vallat wrote:
>>>>
>>>>
>>>>>>Does the sun be 10/100 sbus ethernet card work on sparc64? It works
>>>>>>just fine more me on sparc32, but on sparc64 I get DMA errors and the
>>>>>>card is unusable. According to the docs, the NIC is supported on both
>>>>>>platforms.
>
>
> [ ... ]
>
>
>>> Hmm ... are these the ones which identify as hme*? I've got a
>>>four-port one in an Ultra-1/140, and it is working without obvious
>>>problems.
>
>
> [ ... ]
>
>
>>I had the card in a sparcstation 5 running as be0 just fine.
>
>
> O.K. If it identifies itself as "be0", then it is not the same
> card which I have been using, so I don't know. Mine shows up in dmesg as:
>
> ================================================== ====================
> SUNW,bpp at sbus0 slot 14 offset 0xc800000 vector 22 ipl 2 not configured
> hme0 at sbus0 slot 0 offset 0x8c00000 vector 4 ipl 6: address 08:00:20:bd:e5:00
> ukphy0 at hme0 phy 1: Generic IEEE 802.3u media interface
> ukphy0: OUI 0x00601d, model 0x000c, rev. 1
> hme1 at sbus0 slot 0 offset 0x8c10000 vector 4 ipl 6: address 08:00:20:bd:e5:01
> ukphy1 at hme1 phy 1: Generic IEEE 802.3u media interface
> ukphy1: OUI 0x00601d, model 0x000c, rev. 1
> hme2 at sbus0 slot 0 offset 0x8c20000 vector 4 ipl 6: address 08:00:20:bd:e5:02
> ukphy2 at hme2 phy 1: Generic IEEE 802.3u media interface
> ukphy2: OUI 0x00601d, model 0x000c, rev. 1
> hme3 at sbus0 slot 0 offset 0x8c30000 vector 4 ipl 6: address 08:00:20:bd:e5:03
> ukphy3 at hme3 phy 1: Generic IEEE 802.3u media interface
> ukphy3: OUI 0x00601d, model 0x000c, rev. 1
> ================================================== ====================
>
> And is either a 501-4302 or a 501-4837/501-5443. I can't tell
> you which without bringing the system down and removing the card to read
> its barcode at this point -- and since it is serving in a firewall, I
> would rather not do that. There are similar ones with a single
> interface instead of the four-holer which I am using, and I am using
> some of these in other systems (SS-5 and SS-20), but all of those are
> running some form of Solaris, not OpenBSD.
>
> What card identifies itself as a "be0"? Can you point to the
> page in Sun's FEH (Field Engineer's Handbook)? At least with it moving
> from a SS-5 to an Ultra-1, it should be an Sbus card. But I didn't know
> that there was another 100BaseT/10BaseT card in their line.
>
>
>> I replaced
>>the sparcstation 5 with an ultra-1 and the kernel recognizes the card is
>>there, but i get a massive string of DMA errors and can't use the card.
>
>
> Since you are apparently talking about a different card from the
> one which I am using, I guess that I can't be any help.
>
> Best of luck,
> DoN.

From the sunsolve.sun.com:

http://sunsolve.sun.com/handbook_pub...stEther_1.html


I believe the hme card is version 2.

Dmesg from my sparc 5:
qec0 at sbus0 slot 2 offset 0x20000: 128K memory 1 channel
be0 at qec0 offset 0x10000 pri 7: rev 1 address 08:00:20:71:94:4b
qec1 at sbus0 slot 3 offset 0x20000: 128K memory 1 channel
be1 at qec1 offset 0x10000 pri 7: rev 1 address 08:00:20:71:94:4b

I ran sendbug yesterday with the ultra, but the report hasn't shown up
in the database yet.

Bob Healey

On 31 Jan 2005 01:02:53 -0500, dnichols@d-and-d.com (DoN. Nichols) wrote:

>>>>>>Does the sun be 10/100 sbus ethernet card work on sparc64? It works
>>>>>>just fine more me on sparc32, but on sparc64 I get DMA errors and the
>>>>>>card is unusable. According to the docs, the NIC is supported on both
>>>>>>platforms.
>
> [ ... ]
>
>>> Hmm ... are these the ones which identify as hme*? I've got a
>>> four-port one in an Ultra-1/140, and it is working without obvious
>>> problems.
>
> [ ... ]
>
>>I had the card in a sparcstation 5 running as be0 just fine.
>

> What card identifies itself as a "be0"? Can you point to the
>page in Sun's FEH (Field Engineer's Handbook)? At least with it moving
>from a SS-5 to an Ultra-1, it should be an Sbus card. But I didn't know
>that there was another 100BaseT/10BaseT card in their line.

It's a X1056A.

Pictures: http://sunstuff.org/hardware/partnumbers/X/1/0/5/6/A/

"Not supported by Solaris 7.0 & up, Supported in Sun-4/15/30, SS5, SS10,
SS20, SS600, SS1000, SS2000."

Mit freundlichen Grüßen

Dipl.-Ing. Frank-Christian Krügel
IstDa Kommunikationssysteme

In article <ctl9nk$nua$1@misc-cct.server.rpi.edu>,
Bob Healey <healer@rpi.edu> wrote:
>DoN. Nichols wrote:
>> In article <cthfc2$9bl$1@misc-cct.server.rpi.edu>,
>> Bob Healey <healer@rpi.edu> wrote:
>>
>>>DoN. Nichols wrote:

[ ... ]

>>>> Hmm ... are these the ones which identify as hme*? I've got a
>>>>four-port one in an Ultra-1/140, and it is working without obvious
>>>>problems.
>>
>>
>> [ ... ]
>>
>>
>>>I had the card in a sparcstation 5 running as be0 just fine.
>>
>>
>> O.K. If it identifies itself as "be0", then it is not the same
>> card which I have been using, so I don't know. Mine shows up in dmesg as:
>>
>> ================================================== ====================
>> SUNW,bpp at sbus0 slot 14 offset 0xc800000 vector 22 ipl 2 not configured
>> hme0 at sbus0 slot 0 offset 0x8c00000 vector 4 ipl 6: address 08:00:20:bd:e5:00
>> ukphy0 at hme0 phy 1: Generic IEEE 802.3u media interface
>> ukphy0: OUI 0x00601d, model 0x000c, rev. 1

[ ... ]

>> ukphy3 at hme3 phy 1: Generic IEEE 802.3u media interface
>> ukphy3: OUI 0x00601d, model 0x000c, rev. 1
>> ================================================== ====================
>>
>> And is either a 501-4302 or a 501-4837/501-5443. I can't tell
>> you which without bringing the system down and removing the card to read
>> its barcode at this point -- and since it is serving in a firewall, I
>> would rather not do that. There are similar ones with a single
>> interface instead of the four-holer which I am using, and I am using
>> some of these in other systems (SS-5 and SS-20), but all of those are
>> running some form of Solaris, not OpenBSD.
>>
>> What card identifies itself as a "be0"? Can you point to the
>> page in Sun's FEH (Field Engineer's Handbook)? At least with it moving
>> from a SS-5 to an Ultra-1, it should be an Sbus card. But I didn't know
>> that there was another 100BaseT/10BaseT card in their line.
>>
>>
>>> I replaced
>>>the sparcstation 5 with an ultra-1 and the kernel recognizes the card is
>>>there, but i get a massive string of DMA errors and can't use the card.
>>
>>
>> Since you are apparently talking about a different card from the
>> one which I am using, I guess that I can't be any help.

[ ... ]

> From the sunsolve.sun.com:
>
>http://sunsolve.sun.com/handbook_pub...stEther_1.html
>
>
>I believe the hme card is version 2.

O.K. A quick check shows that all that I have are Version
2.0/2.1, based on the relative orientation of the the MII connector and
the RJ-45 one.

>Dmesg from my sparc 5:
>qec0 at sbus0 slot 2 offset 0x20000: 128K memory 1 channel
>be0 at qec0 offset 0x10000 pri 7: rev 1 address 08:00:20:71:94:4b
>qec1 at sbus0 slot 3 offset 0x20000: 128K memory 1 channel
>be1 at qec1 offset 0x10000 pri 7: rev 1 address 08:00:20:71:94:4b
>
>I ran sendbug yesterday with the ultra, but the report hasn't shown up
>in the database yet.

O.K. Good luck, I can state that the HME cards do work with
OpenBSD 3.4. I don't have 3.5 or 3.6 on any SPARC based systems, yet,
so I don't know whether it is a problem in the OS version or the card
version.

Best of luck,
DoN.

--
Email: <dnichols@d-and-d.com> | Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
--- Black Holes are where God is dividing by zero ---

In article <41fe2ed3.16238000@uaxe.intern.istda.com>,
Frank-Christian Kruegel <usenet1003> wrote:
>On 31 Jan 2005 01:02:53 -0500, dnichols@d-and-d.com (DoN. Nichols) wrote:
>
>>>>>>>Does the sun be 10/100 sbus ethernet card work on sparc64? It works
>>>>>>>just fine more me on sparc32, but on sparc64 I get DMA errors and the
>>>>>>>card is unusable. According to the docs, the NIC is supported on both
>>>>>>>platforms.
>>
>> [ ... ]
>>
>>>> Hmm ... are these the ones which identify as hme*? I've got a
>>>> four-port one in an Ultra-1/140, and it is working without obvious
>>>> problems.
>>
>> [ ... ]
>>
>>>I had the card in a sparcstation 5 running as be0 just fine.
>>
>
>> What card identifies itself as a "be0"? Can you point to the
>>page in Sun's FEH (Field Engineer's Handbook)? At least with it moving
>>from a SS-5 to an Ultra-1, it should be an Sbus card. But I didn't know
>>that there was another 100BaseT/10BaseT card in their line.
>
>It's a X1056A.
>
>Pictures: http://sunstuff.org/hardware/partnumbers/X/1/0/5/6/A/
>
>"Not supported by Solaris 7.0 & up, Supported in Sun-4/15/30, SS5, SS10,
>SS20, SS600, SS1000, SS2000."

And in particular, swept out from that web page, note this line:

This card is supported in most pre-Ultra Suns

which suggests that it does *not* work in the Ultras (64-bit SPARCs). I
do know that the hme versions do work in Ultras -- and in the Ultra-2,
it is even the built-in ethernet interface.

Sorry about that. Move that card back to the SS-5, and look for
one of the hme cards for the Ultra-1. Those *do* work.

Good Luck,
DoN.

--
Email: <dnichols@d-and-d.com> | Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
--- Black Holes are where God is dividing by zero ---