Re: Spurious 8259A?
> Listan,
>
> Jag har en burk där jag misstänker något håller på att rasa, bland annat
> låter den annorlunda (disk, fläkt, cpu-fläkt?) än vad jag tror den gjort
> tidigare.
>
> En fsck på disken i samband med boot går igenom, men däremot hittar jag
> meddelanden i loggen: Kernel spurious 8259A interupt; IRQ7 på två
> ställen. Någon som vet vad det gäller? Kan det ha något samband med
> någon fläkt vars lager är på upphällningen? Något med temperaturen? Kan
> jag friskriva disken?
>
> /Anders W
...
Från:
"Intel386TM EX Embedded Microprocessor User' s Manual"
Intel Corporation
1996
Order Number 272485-002
(går förmodligen att ladda ner från nätet)
9.4.3 Spurious Interrupts
For both edge and level-triggered interrupts, a high level must be
maintained on the IR line until after the falling edge of the first
INTA# pulse (see Figure 9-18). A spurious interrupt request is generated
if this stipulation is not met. A spurious interrupt on any IR line
generates the same vector number as an IR7 request. The spurious
interrupt, however, does not set the in-service bit for IR7. Therefore,
an IR7 interrupt service routine must check the in-service register to
determine whether the interrupt source was a valid IR7 (the in-service
bit is set) or a spurious interrupt (the in-service bit is cleared).
(IR = interrupt request)
Dvs. en sk. "surious interrupt" inträffar bara för irq7, men manualen
säger ingenting varför de inträffar.
--------
Från kärnan:
$ sed -ne '272,286p' /usr/src/kernel-source-2.4.20/arch/i386/kernel/i8259.c
/*
* Lightweight spurious IRQ detection. We do not want
* to overdo spurious IRQ handling - it's usually a sign
* of hardware problems, so we only do the checks we can
* do without slowing down good hardware unnecesserily.
*
* Note that IRQ7 and IRQ15 (the two spurious IRQs
* usually resulting from the 8259A-1|2 PICs) occur
* even if the IRQ is masked in the 8259A. Thus we
* can check spurious 8259A IRQs without doing the
* quite slow i8259A_irq_real() call for every IRQ.
* This does not cover 100% of spurious interrupts,
* but should be enough to warn the user that there
* is something bad going on ...
*/
--------
Från loggar:
# zgrep -C -i spurious /var/log/all.log.?.gz
/var/log/all.log.2.gz:Feb 22 10:10:18 swat kernel: ext3: No journal on filesystem on ide0(3,2)
/var/log/all.log.2.gz:Feb 22 10:10:18 swat kernel: Adding Swap: 96384k swap-space (priority -1)
/var/log/all.log.2.gz:Feb 22 10:10:18 swat kernel: spurious 8259A interrupt: IRQ7.
/var/log/all.log.2.gz:Feb 22 10:10:18 swat kernel: Real Time Clock Driver v1.10e
/var/log/all.log.2.gz:Feb 22 10:10:18 swat kernel: Linux Tulip driver version 0.9.15-pre12 (Aug 9, 2002)
/var/log/all.log.3.gz:Feb 21 08:42:38 pyrit ippl: ICMP message type destination unreachable - bad port from 127.0.0.1
/var/log/all.log.3.gz:Feb 21 08:42:42 opal ippl: ICMP message type destination unreachable - bad port from 127.0.0.1
/var/log/all.log.3.gz:Feb 21 08:42:52 swat kernel: spurious 8259A interrupt: IRQ7.
/var/log/all.log.3.gz:Feb 21 08:43:04 opal ippl: port 13327 connection attempt from 192.168.93.38
/var/log/all.log.3.gz:Feb 21 08:43:38 pyrit ippl: ICMP message type destination unreachable - bad port from 127.0.0.1
/var/log/all.log.3.gz:--
/var/log/all.log.3.gz:Feb 21 16:18:48 swat kernel: Adding Swap: 96384k swap-space (priority -1)
/var/log/all.log.3.gz:Feb 21 16:18:48 swat kernel: Real Time Clock Driver v1.10e
/var/log/all.log.3.gz:Feb 21 16:18:48 swat kernel: spurious 8259A interrupt: IRQ7.
/var/log/all.log.3.gz:Feb 21 16:18:48 swat kernel: Linux Tulip driver version 0.9.15-pre12 (Aug 9, 2002)
/var/log/all.log.3.gz:Feb 21 16:18:48 swat kernel: PCI: Found IRQ 9 for device 00:0b.0
/var/log/all.log.4.gz:Feb 20 07:39:24 swat kernel: /dev/ide/host0/bus0/target0/lun0: [PTBL] [1869/255/63] p1 p2 p3 p4 < p5 p6 p7 p8 >
/var/log/all.log.4.gz:Feb 20 07:39:24 swat kernel: ext3: No journal on filesystem on ide0(3,2)
/var/log/all.log.4.gz:Feb 20 07:39:24 swat kernel: spurious 8259A interrupt: IRQ7.
/var/log/all.log.4.gz:Feb 20 07:39:24 swat kernel: Adding Swap: 96384k swap-space (priority -1)
/var/log/all.log.4.gz:Feb 20 07:39:24 swat kernel: Real Time Clock Driver v1.10e
/var/log/all.log.5.gz:Feb 19 08:53:01 swat kernel: /dev/ide/host0/bus0/target0/lun0: [PTBL] [1869/255/63] p1 p2 p3 p4 < p5 p6 p7 p8 >
/var/log/all.log.5.gz:Feb 19 08:53:01 swat kernel: ext3: No journal on filesystem on ide0(3,2)
/var/log/all.log.5.gz:Feb 19 08:53:01 swat kernel: spurious 8259A interrupt: IRQ7.
/var/log/all.log.5.gz:Feb 19 08:53:01 swat kernel: Adding Swap: 96384k swap-space (priority -1)
/var/log/all.log.5.gz:Feb 19 08:53:01 swat kernel: Real Time Clock Driver v1.10e
/var/log/all.log.5.gz:--
/var/log/all.log.5.gz:Feb 19 18:05:44 swat kernel: /dev/ide/host0/bus0/target0/lun0: [PTBL] [1869/255/63] p1 p2 p3 p4 < p5 p6 p7 p8 >
/var/log/all.log.5.gz:Feb 19 18:05:44 swat kernel: ext3: No journal on filesystem on ide0(3,2)
/var/log/all.log.5.gz:Feb 19 18:05:44 swat kernel: spurious 8259A interrupt: IRQ7.
/var/log/all.log.5.gz:Feb 19 18:05:44 swat kernel: Adding Swap: 96384k swap-space (priority -1)
/var/log/all.log.5.gz:Feb 19 18:05:44 swat kernel: Real Time Clock Driver v1.10e
/var/log/all.log.6.gz:Feb 18 10:36:35 swat kernel: Adding Swap: 96384k swap-space (priority -1)
/var/log/all.log.6.gz:Feb 18 10:36:35 swat kernel: Real Time Clock Driver v1.10e
/var/log/all.log.6.gz:Feb 18 10:36:35 swat kernel: spurious 8259A interrupt: IRQ7.
/var/log/all.log.6.gz:Feb 18 10:36:35 swat kernel: Linux Tulip driver version 0.9.15-pre11 (May 11, 2002)
/var/log/all.log.6.gz:Feb 18 10:36:35 swat kernel: PCI: Found IRQ 9 for device 00:0b.0
/var/log/all.log.6.gz:--
/var/log/all.log.6.gz:Feb 18 17:29:19 swat kernel: /dev/ide/host0/bus0/target0/lun0: [PTBL] [1869/255/63] p1 p2 p3 p4 < p5 p6 p7 p8 >
/var/log/all.log.6.gz:Feb 18 17:29:19 swat kernel: ext3: No journal on filesystem on ide0(3,2)
/var/log/all.log.6.gz:Feb 18 17:29:19 swat kernel: spurious 8259A interrupt: IRQ7.
/var/log/all.log.6.gz:Feb 18 17:29:19 swat kernel: Adding Swap: 96384k swap-space (priority -1)
/var/log/all.log.6.gz:Feb 18 17:29:19 swat kernel: Real Time Clock Driver v1.10e
Och det har bara uppträtt på en maskin hos mig och i samband med
uppstart (den senaste veckan).
Hälsningar,
/Karl
-----------------------------------------------------------------------
Karl Hammar Aspö Data karl@kalle.csb.ki.se
Lilla Aspö 2340 Nätverk
S-742 94 Östhammar 0173 140 57 Datorer/Utrustning
Sweden 070 511 97 84 Linux/Unix konsulting
-----------------------------------------------------------------------
Reply to: