* interrupt problem
@ 2001-05-17 18:18 Tom Appermont
0 siblings, 0 replies; 2+ messages in thread
From: Tom Appermont @ 2001-05-17 18:18 UTC (permalink / raw)
To: linux-mips
Howdy,
I named the subject "interrupt problem", although I am not
completely sure it is really a problem with interrupts I'm
having. Anyway, the symptoms are the following:
- a ddb 5074 board with 3com boomerang ethernet pci card, boots
with initial ramdisk (ok)
- serial interrupts work (ok)
- ethernet card interrupts work when doing a "simple" ping
- flooding the card with network packets (ping -f) causes the
ddb board to hang/freeze/stall after an unpredictible amount of
time, without producing any output on the serial console.
- the problem does not seem to exist when I use a realtek ethernet
card (ne2k driver) in stead of the 3com card.
I've tried to remedy this by:
- replacing the low level interrupt handling code for ddb5074
with the new irq.c and time.c in arch/mips/kernel (needs
CONFIG_NEW_TIME_C and CONFIG_NEW_IRQ).
- replacing the timer interrupts with CPU timer interrupts.
- getting the newest 3c59x driver (although I don't have any
probs with this driver on pc).
- banging my head on the table.
No luck so far. I am a bit out of ideas on how to track down the
cause of this problem, so it would be really helpful if somebody
could either indicate a hint of a possible solution, or suggest
ways to debug this situation.
Greetz,
Tom
^ permalink raw reply [flat|nested] 2+ messages in thread
* interrupt problem....
@ 2001-06-26 4:49 mdaljeet
0 siblings, 0 replies; 2+ messages in thread
From: mdaljeet @ 2001-06-26 4:49 UTC (permalink / raw)
To: linux-kernel
I am experiencing a strange problem.
I am doing a continuous DMA for long hours using a card on my system. In my
code I enable interrupts and clear the interrupts in the interrupt handler
which is called on completion of every DMA cycle. Now, the program works
fine for say 16-20 hours but I think when the debug files'(where i put some
debug information for each DMA cycle + system log files etc) size becomes
too large, card stops generating interrupts.
1. There is a bit in card that tells that DMA is complete
2. There is a bit in card that tells card to generate interrupt when DMA is
complete
3. If we clear the above two bits, the interrupt handling is complete
Now, DMA is getting complete as the bits specified in the point 1 and 2 are
getting set. I have added a 'printk' in the interrupt handler to see
whether the card is generating interrupts or not. Card stops generating
interrupt. Even if I delete all the debug files and start my program again,
same thing happens. After I reboot the system and run my program, the card
starts working properly with same code.
I have noted following things:
a) If I delete all the system log files and restrict the size of my debug
file, the code runs fine for more than 48 hours. Basically I stop the
program after having a 48 hour run.
b) If the system log files has large size and I do not delete them before
starting my program, the card stops generating interrupts after 16-20 hours
of run.
Any pointers to the problem......????
Regards,
Daljeet.
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2001-06-26 6:32 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2001-05-17 18:18 interrupt problem Tom Appermont
-- strict thread matches above, loose matches on Subject: below --
2001-06-26 4:49 mdaljeet
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.