I could test this maybe on the night, because we actually have our zvm 4.3
running linux vm's for production (thanks for all the people here :-D)

On Mon, 17 Nov 2003, Post, Mark K wrote:

> Ok, I see.  You're right, z/VM should not be abending, but that's a separate
> problem.
>
> If you specify
> add_parms,0x10,0x503,0x505,portname:NIC503
> qeth2,0x504,0x505,0x503
>
> I imagine you might get this to work properly while IBM is figuring out the
> dump.
>
>
> Mark Post
>
> -----Original Message-----
> From: Alejandro Leyva Rabinovich [mailto:[EMAIL PROTECTED]
> Sent: Monday, November 17, 2003 12:56 PM
> To: [EMAIL PROTECTED]
> Subject: Re: qdio problem
>
>
> The /etc/modutils/0chandev.chandev is:
>
> add_parms,0x10,0x2000,0x2002,primary_router,portname:OSA1
> qeth0,0x2000,0x2001,0x2002
> add_parms,0x10,0x500,0x502,portname:NIC500
> qeth1,0x500,0x501,0x502
> add_parms,0x10,0x503,0x505,portname:NIC503
> qeth2,0x503,0x504,0x505
>
> so, the configuration defines 503, 504 and 505, in that order,
> /proc/chandev say that the addresses are configured in that order:
>
> Forced devices
>   chan defif read   write  data   memory      port         ip    hw   host
> adapter   api
>   type  num  devno  devno  devno  usage(k) protocol no.  chksum stats name
> name     name
> ============================================================================
> ===================
>   0x10    2  0x0503 0x0504 0x0505 default         0       0    0
>
> when i use zvm 4.4, in the section "Initialised Devices" i dont have the
> 503-505 device, but with 4.3 i have:
>
> Initialised Devices
>  read   write  data  read   write  data  chan port  dev     dev
> memory   read msck    write msck    data msck
>  irq     irq    irq  devno  devno  devno type no.   ptr     name
> usage(k)  status       status        status
> ============================================================================
> =========================================
> 0x0008 0x0009 0x0007 0x0504 0x0505 0x0503 0x10  0 0x005dd800 eth2
> 8192   good         good         good
>
> i dont understand why i have the addresses in different order.
>
> anyway, that isnt a reason for zvm to abend.
>
>
> On Mon, 17 Nov 2003, Post, Mark K wrote:
>
> > I'm confused.  How is 0x504 an "odd device address?"  I've done things
> such
> > as
> > 504,505,50A before.  How is 504,505,503 any different?
> >
> >
> > Mark Post
> >
> > -----Original Message-----
> > From: Dennis Musselwhite [mailto:[EMAIL PROTECTED]
> > Sent: Wednesday, November 12, 2003 8:23 AM
> > To: [EMAIL PROTECTED]
> > Subject: Re: qdio problem
> >
> >
> > Hi...
> >
> > The following response:
> >
> > qeth: Trying to use card with devnos 0x504/0x505/0x503
> >  qeth: received an IDX TERMINATE on irq 0x8/0x9 with cause code 0x04
> >  qeth: IDX_ACTIVATE on read channel irq 0x8: negative reply
> >  qeth: There were problems in hard-setting up the card.
> >
> > indicates (cause code 0x04) that the read/write device assignments are
> > backwards.  Normally you can use the even-odd rule for read and write
> > devices, but since this NIC is defined to begin on an odd device address,
> > the even offset is on the odd device.  You can avoid this activation
> > problem by EITHER using an even number on the SPECIAL statement, OR
> listing
> > your devices as 0x503,0x504,0x505 in chandev.conf.
> >
> > Sometimes, when switching from HiperSockets to QDIO type, people forget to
> > switch the DEFINE LAN, but it looks like your NIC at 500 came up ok, so
> > that's probably not a problem.
> >
> > For the system to abend (whether it was caused by QUERY NIC, or just a
> > coincidence) you should contact IBM and open a problem report (as Adam
> > suggested) and send in the dump.
> >
> > Regards,
> > Dennis
> > ----------------------------------------------------------------
> > Dennis Musselwhite ([EMAIL PROTECTED]) +1(607)429-3831
> > z/VM Development -- CP Network Simulation -- IBM Endicott NY
> >
>
> --
> Alejandro Leyva Rabinovich.
> Jefe de la Unidad Departamental de Soporte Ticnico
> (Administracisn de Mainframe).
> Direccisn General de Informatica.
> Secretarma de Finanzas.
> Gobierno del Distrito Federal.
>

--
Alejandro Leyva Rabinovich.
Jefe de la Unidad Departamental de Soporte Ticnico
(Administracisn de Mainframe).
Direccisn General de Informatica.
Secretarma de Finanzas.
Gobierno del Distrito Federal.

Reply via email to