Re: CFT: vr(4)

On Monday 04 February 2008, Pyun YongHyeon wrote:
Dear all,

Here is overhauled vr(4) that shall address all known issues. PR
database showed vr(4) is not stable enough under high load and
link state handling didn't work as expected as well as its poor
performance. I've tried hard to fix the bugs reported in PR
database for several months. It tooked more time than I had
planned when I received donated hardware.
vr(4) will now
- work on all architectures.
- have better Rx performance.
- have more robust error recovery.
- reliably detect link state changes.
- support 32bit perfect multicast filtering for VT6105M.
- support WOL.

It seems that the overhauled vr(4) runs well under my limited test
environments. Several users already reported success. Because there
are three kinds of Rhine family and lots of variants for the
hardware, I'd like to hear feedback from users priror to commit.
You can get the latest vr(4) at the following URL.

http://people.freebsd.org/~yongari/vr/if_vr.c
http://people.freebsd.org/~yongari/vr/if_vrreg.h

I wouldn't be available for 7 ~ 10 days from Feb 5, so please don't
expect quick reply.

Hi,

did anybody test Routerboard 44? It is quad network card, uses VT6105M chips
as network controller (Via) and PCI6152 as PCI-PCI bridge (PLX Technology).
With both stock if_vr and modified from site given above it partially works -
interfaces are created, ifconfig works, but even pinging some host in local
network (I achieve the effect with ping -f in a minue or so) hangs system.
Nothing on my console, nothing in system log. Only hard powerdown restores
system in functioning state.

I tested it with stock vr driver in both 6 and 7 stable on both i386 and
amd64, and now with overhauled vr in 8 freshly cvsupped and two files
replaced on i386.

If anybody has any idea or some patches I could test, I will.

Regards,
Milan

This is on console after kldload if_vr:

Feb 5 14:36:29 tix kernel: vr0: <VIA VT6105M Rhine III 10/100BaseTX> port
0x9c00-0x9cff mem 0xfceff000-0xfceff0ff irq 17 at device 8.0 on pci3
Feb 5 14:36:29 tix kernel: vr0: Quirks: 0x6
Feb 5 14:36:29 tix kernel: vr0: Revision: 0x96
Feb 5 14:36:29 tix kernel: miibus1: <MII bus> on vr0
Feb 5 14:36:29 tix kernel: ukphy0: <Generic IEEE 802.3u media interface> PHY
1 on miibus1
Feb 5 14:36:29 tix kernel: ukphy0: 10baseT, 10baseT-FDX, 100baseTX,
100baseTX-FDX, auto
Feb 5 14:36:29 tix kernel: vr0: Ethernet address: 00:0c:42:02:38:a8
Feb 5 14:36:29 tix kernel: vr0: [ITHREAD]
Feb 5 14:36:29 tix kernel: vr1: <VIA VT6105M Rhine III 10/100BaseTX> port
0x9800-0x98ff mem 0xfcefe000-0xfcefe0ff irq 18 at device 9.0 on pci3
Feb 5 14:36:29 tix kernel: vr1: Quirks: 0x6
Feb 5 14:36:29 tix kernel: vr1: Revision: 0x96
Feb 5 14:36:29 tix kernel: miibus2: <MII bus> on vr1
Feb 5 14:36:29 tix kernel: ukphy1: <Generic IEEE 802.3u media interface> PHY
1 on miibus2
Feb 5 14:36:29 tix kernel: ukphy1: 10baseT, 10baseT-FDX, 100baseTX,
100baseTX-FDX, auto
Feb 5 14:36:29 tix kernel: vr1: Ethernet address: 00:0c:42:02:38:a9
Feb 5 14:36:29 tix kernel: vr1: [ITHREAD]
Feb 5 14:36:29 tix kernel: vr2: <VIA VT6105M Rhine III 10/100BaseTX> port
0x9400-0x94ff mem 0xfcefd000-0xfcefd0ff irq 19 at device 10.0 on pci3
Feb 5 14:36:29 tix kernel: vr2: Quirks: 0x6
Feb 5 14:36:29 tix kernel: vr2: Revision: 0x96
Feb 5 14:36:29 tix kernel: miibus3: <MII bus> on vr2
Feb 5 14:36:29 tix kernel: ukphy2: <Generic IEEE 802.3u media interface> PHY
1 on miibus3
Feb 5 14:36:29 tix kernel: ukphy2: 10baseT, 10baseT-FDX, 100baseTX,
100baseTX-FDX, auto
Feb 5 14:36:29 tix kernel: vr2: Ethernet address: 00:0c:42:02:38:aa
Feb 5 14:36:29 tix kernel: vr2: [ITHREAD]
Feb 5 14:36:29 tix kernel: vr3: <VIA VT6105M Rhine III 10/100BaseTX> port
0x9000-0x90ff mem 0xfcefc000-0xfcefc0ff irq 16 at device 11.0 on pci3
Feb 5 14:36:29 tix kernel: vr3: Quirks: 0x6
Feb 5 14:36:29 tix kernel: vr3: Revision: 0x96
Feb 5 14:36:29 tix kernel: miibus4: <MII bus> on vr3
Feb 5 14:36:29 tix kernel: ukphy3: <Generic IEEE 802.3u media interface> PHY
1 on miibus4
Feb 5 14:36:29 tix kernel: ukphy3: 10baseT, 10baseT-FDX, 100baseTX,
100baseTX-FDX, auto
Feb 5 14:36:29 tix kernel: vr3: Ethernet address: 00:0c:42:02:38:ab
Feb 5 14:36:29 tix kernel: vr3: [ITHREAD]
Feb 5 14:36:29 tix kernel: vr0: link state changed to DOWN
Feb 5 14:36:29 tix kernel: vr1: link state changed to DOWN
Feb 5 14:36:29 tix kernel: vr2: link state changed to DOWN
Feb 5 14:36:29 tix kernel: vr3: link state changed to DOWN


--
Address this mail is sent from is used only for this mailing list.
Do not send any messages to it directly as a response, reply only
to mailing list. For mail to me personally, use milan in address instead.
_______________________________________________
freebsd-current@xxxxxxxxxxx mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any mail to "freebsd-current-unsubscribe@xxxxxxxxxxx"

Relevant Pages

  • Re: CFT: vr(4)
    ... database for several months. ... reliably detect link state changes. ... on miibus1 ... 100baseTX-FDX, auto ...
    (freebsd-current)
  • Problem: How to resize FreeBSD "partitions" on a live system?
    ... I have found that with the auto defaults, ... out of control (heavy database; lots of spam filling up logs, ...
    (comp.unix.bsd.freebsd.misc)
  • RE: Merging 9 stand alone Databases
    ... If you delete the relationship, change the autonumber field to number, ... Then change the autonumber PK fields to Numbers EXCEPT for in the master ... The problem is the auto number it is the primary key and unique Identifier ... Each database structure is identical so ...
    (microsoft.public.access.tablesdbdesign)
  • RE: Merging 9 stand alone Databases
    ... Then change the autonumber PK fields to Numbers EXCEPT for in the master ... Jerry Whittle, Microsoft Access MVP ... The problem is the auto number it is the primary key and unique Identifier ... Each database structure is identical so ...
    (microsoft.public.access.tablesdbdesign)
  • Re: Auto Number
    ... The problem was fixed long ago, but if your client has no ... to run some code to fix the database that already has the problem. ... > I have a database that uses an auto number field type that goes out of ... > auto number 1, and sequentially up to about 20,000. ...
    (comp.databases.ms-access)