APIC problem or 3com 3c590 driver problem in smp kernel 2.4.x

APIC problem or 3com 3c590 driver problem in smp kernel 2.4.x

[Feng Xian]

Hi, all,

I have one pci device in my dell optiplex gx300 dual pIII box. when this
device in a certain pci slot, it shares the same IRQ with the on mother
board 3com NIC(3c905) (irq=5). when I run the kernel smp-2.4.x, my PCI
device can not receive any interrupt while the /proc/interrupts shows that
3c905 receives over million of interrupts and number grows very fast. then
I moved my pci device to another pci slot, now, 3c905's irq=5, my device's
irq=0xb, my device behaves normal, the interrupt number for 3c905 also
looks normal. 	I also tried to use a uni-processor kernel without APIC
support enabled, my pci device shares irq with 3c905, but works fine.

It looks more like APIC support problem. but it also may be 3c905
driver's ISR's problem. Anyone who owns the code has any idea about this?

Thanks in advance.

Alex


-- 
        Feng Xian
   _o)     .~.      (o_
   /\\     /V\      //\
  _\_V    // \\     V_/_
         /(   )\
          ^^-^^
           ALEX

Re: APIC problem or 3com 3c590 driver problem in smp kernel 2.4.x

[Bogdan Costescu]

On Wed, 30 May 2001, Feng Xian wrote:

> when I run the kernel smp-2.4.x, my PCI
> device can not receive any interrupt while the /proc/interrupts shows that
> 3c905 receives over million of interrupts and number grows very fast.

That's a bit strange as one of the first things done in 3c59x ISR is:

if ((status & IntLatch) == 0)
         goto handler_exit;  /* No interrupt: shared IRQs can cause this */

with which the driver detects if the interrupt was generated by the card.
Are you sure the driver for the other PCI device knows to play nice with
shared interrupts ?

Sincerely,

Bogdan Costescu

IWR - Interdisziplinaeres Zentrum fuer Wissenschaftliches Rechnen
Universitaet Heidelberg, INF 368, D-69120 Heidelberg, GERMANY
Telephone: +49 6221 54 8869, Telefax: +49 6221 54 8868
E-mail: Bogdan.Costescu@IWR.Uni-Heidelberg.De

Re: APIC problem or 3com 3c590 driver problem in smp kernel 2.4.x

[Feng Xian]

The driver for my pci device, I have the SA_SHIRQ set.

Actually what I am thinking it may be APIC support problem. I rebuild my
kernel to use single cpu without APIC support, my device and 3c905 both
work fine. they don't work for SMP kernel (APIC is by default enabled)
Then I configured my uni-processor kernel to enable the APIC support, my
device won't work with the 3c905, just exactly same as it behaves in the
SMP kernel.

regards,

Alex


On Thu, 31 May 2001, Bogdan Costescu wrote:

> On Wed, 30 May 2001, Feng Xian wrote:
>
> > when I run the kernel smp-2.4.x, my PCI
> > device can not receive any interrupt while the /proc/interrupts shows that
> > 3c905 receives over million of interrupts and number grows very fast.
>
> That's a bit strange as one of the first things done in 3c59x ISR is:
>
> if ((status & IntLatch) == 0)
>          goto handler_exit;  /* No interrupt: shared IRQs can cause this */
>
> with which the driver detects if the interrupt was generated by the card.
> Are you sure the driver for the other PCI device knows to play nice with
> shared interrupts ?
>
> Sincerely,
>
> Bogdan Costescu
>
> IWR - Interdisziplinaeres Zentrum fuer Wissenschaftliches Rechnen
> Universitaet Heidelberg, INF 368, D-69120 Heidelberg, GERMANY
> Telephone: +49 6221 54 8869, Telefax: +49 6221 54 8868
> E-mail: Bogdan.Costescu@IWR.Uni-Heidelberg.De
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>

-- 
        Feng Xian
   _o)     .~.      (o_
   /\\     /V\      //\
  _\_V    // \\     V_/_
         /(   )\
          ^^-^^
           ALEX

Re: APIC problem or 3com 3c590 driver problem in smp kernel 2.4.x

[Ingo Oeser]

On Thu, May 31, 2001 at 12:27:07PM -0400, Feng Xian wrote:
> The driver for my pci device, I have the SA_SHIRQ set.

What kind of PCI device do you have? I had this problem once with
an PCI-Matchmaker[1] based board (for which we still have the wrong
PCI-ID btw, but my patch was rejected twice...).

> Actually what I am thinking it may be APIC support problem. I rebuild my
> kernel to use single cpu without APIC support, my device and 3c905 both
> work fine. they don't work for SMP kernel (APIC is by default enabled)
> Then I configured my uni-processor kernel to enable the APIC support, my
> device won't work with the 3c905, just exactly same as it behaves in the
> SMP kernel.

With 2.2 I also had this without APIC. 

I have been flooded with interrupts which have been intended for
the Cyclone card (3c905B 100BaseTX), and exited the ISR quickly
after querying the interrupt register of my Matchmaker board
without any ACKing, but the Cyclone never got these interrupts
anymore.

But is doesn't seem to be a 3c905 based problem, as I have 

 11:   95772726          XT-PIC  es1371, eth0, eth1
 
in /proc/interrupts where eth0 and eth1 are both Cyclones.
Even the vga card has IRQ 11 assigned.

So this is not really unknown ;-)

Regards

Ingo Oeser

[1] class 0b40, vendor id: 10e8, device id: 807d
-- 
To the systems programmer,
users and applications serve only to provide a test load.

Re: APIC problem or 3com 3c590 driver problem in smp kernel 2.4.x

[Feng Xian]

It's our own's card. so it could be the card's problem. does the pci
device have to do some special thing to support APIC? my card won't work
properly on uni-processor with APIC enable kernel or smp kernel when the
card is sharing IRQ with some other pci devices.


Alex

On Fri, 1 Jun 2001, Ingo Oeser wrote:

> On Thu, May 31, 2001 at 12:27:07PM -0400, Feng Xian wrote:
> > The driver for my pci device, I have the SA_SHIRQ set.
>
> What kind of PCI device do you have? I had this problem once with
> an PCI-Matchmaker[1] based board (for which we still have the wrong
> PCI-ID btw, but my patch was rejected twice...).
>
> > Actually what I am thinking it may be APIC support problem. I rebuild my
> > kernel to use single cpu without APIC support, my device and 3c905 both
> > work fine. they don't work for SMP kernel (APIC is by default enabled)
> > Then I configured my uni-processor kernel to enable the APIC support, my
> > device won't work with the 3c905, just exactly same as it behaves in the
> > SMP kernel.
>
> With 2.2 I also had this without APIC.
>
> I have been flooded with interrupts which have been intended for
> the Cyclone card (3c905B 100BaseTX), and exited the ISR quickly
> after querying the interrupt register of my Matchmaker board
> without any ACKing, but the Cyclone never got these interrupts
> anymore.
>
> But is doesn't seem to be a 3c905 based problem, as I have
>
>  11:   95772726          XT-PIC  es1371, eth0, eth1
>
> in /proc/interrupts where eth0 and eth1 are both Cyclones.
> Even the vga card has IRQ 11 assigned.
>
> So this is not really unknown ;-)
>
> Regards
>
> Ingo Oeser
>
> [1] class 0b40, vendor id: 10e8, device id: 807d
>

-- 
        Feng Xian
   _o)     .~.      (o_
   /\\     /V\      //\
  _\_V    // \\     V_/_
         /(   )\
          ^^-^^
           ALEX

Re: APIC problem or 3com 3c590 driver problem in smp kernel 2.4.x

[Alan Cox]

> It's our own's card. so it could be the card's problem. does the pci
> device have to do some special thing to support APIC? my card won't work

Nope. APIC is invisible to a correctly built piece of hardware. It just changes
how INTA-INTD are handled at the host end of things

> properly on uni-processor with APIC enable kernel or smp kernel when the
> card is sharing IRQ with some other pci devices.

That sounds like the driver isnt testing the irq was sourced by that card. You
also see hangups on insmod/rmmod when people do

		writel(ENABLE_MY_IRQ, my_card->control);

before they request the IRQ and are able to clear it

Re: APIC problem or 3com 3c590 driver problem in smp kernel 2.4.x

[Feng Xian]

On Sat, 2 Jun 2001, Alan Cox wrote:

> > It's our own's card. so it could be the card's problem. does the pci
> > device have to do some special thing to support APIC? my card won't work
>
> Nope. APIC is invisible to a correctly built piece of hardware. It just changes
> how INTA-INTD are handled at the host end of things
>
> > properly on uni-processor with APIC enable kernel or smp kernel when the
> > card is sharing IRQ with some other pci devices.
>
> That sounds like the driver isnt testing the irq was sourced by that card. You
> also see hangups on insmod/rmmod when people do
>
> 		writel(ENABLE_MY_IRQ, my_card->control);
>
> before they request the IRQ and are able to clear it
>

I forget to mention. in the same hardware configuration (same slot, sharing
IRQ with other card) my card works perfect if I was using uni-processor
without APIC support kernel (i tried 2.4.5-ac6 with apic disabled
uniprocessor on a dual p3 box). If the driver did something wrong, it
should not work on that system either. Maybe what I thought was wrong.

Alex

-- 
        Feng Xian
   _o)     .~.      (o_
   /\\     /V\      //\
  _\_V    // \\     V_/_
         /(   )\
          ^^-^^
           ALEX

Re: APIC problem or 3com 3c590 driver problem in smp kernel 2.4.x

[Maciej W. Rozycki]

On Sat, 2 Jun 2001, Feng Xian wrote:

> I forget to mention. in the same hardware configuration (same slot, sharing
> IRQ with other card) my card works perfect if I was using uni-processor
> without APIC support kernel (i tried 2.4.5-ac6 with apic disabled
> uniprocessor on a dual p3 box). If the driver did something wrong, it
> should not work on that system either. Maybe what I thought was wrong.

 There is a small difference in interrupt delivery between i8259A and I/O
APIC configurations.  An I/O APIC incurs a slightly larger delivery
latency, both for asserts and for deasserts.  You need to be prepared for
it in the driver, especially for the latter case, i.e. send an IRQ ack
early to the device, so there is enough time for a message to go through
the inter-APIC bus.  Otherwise you risk spurious interrupts. 

-- 
+  Maciej W. Rozycki, Technical University of Gdansk, Poland   +
+--------------------------------------------------------------+
+        e-mail: macro@ds2.pg.gda.pl, PGP key available        +

Re: APIC problem or 3com 3c590 driver problem in smp kernel 2.4.x

[Feng Xian]

I am not sure if it's a VIA-based board. It's Dell's SMP board with Intel
820 chipset on it.


-[00]-+-00.0  Intel Corporation 82820 820 (Camino) Chipset Host Bridge
(MCH)
      +-01.0-[01]----00.0  nVidia Corporation Vanta [NV6]
      +-1e.0-[02-03]--+-0a.0-[03]----00.0  Chrysalis-ITS: c0de
      |               \-0c.0  3Com Corporation 3c905C-TX [Fast Etherlink]
      +-1f.0  Intel Corporation 82801AA ISA Bridge (LPC)
      +-1f.1  Intel Corporation 82801AA IDE
      +-1f.2  Intel Corporation 82801AA USB
      +-1f.3  Intel Corporation 82801AA SMBus
      \-1f.5  Intel Corporation 82801AA AC'97 Audio

Alex

On Thu, 31 May 2001, Mark Hahn wrote:

> > I have one pci device in my dell optiplex gx300 dual pIII box. when this
>
> is that a VIA-based board?  if so, there are known problems
> with VIA SMP irq routing.
>

-- 
        Feng Xian
   _o)     .~.      (o_
   /\\     /V\      //\
  _\_V    // \\     V_/_
         /(   )\
          ^^-^^
           ALEX