Re: Probable module bug in linux-2.6.5-1.358

From: Richard B. Johnson
Date: Thu Oct 07 2004 - 08:08:58 EST


On Thu, 7 Oct 2004, Richard B. Johnson wrote:

On Thu, 7 Oct 2004, Arjan van de Ven wrote:

On Thu, Oct 07, 2004 at 08:26:22AM -0400, Richard B. Johnson wrote:
On Thu, 7 Oct 2004, Arjan van de Ven wrote:

On Thu, Oct 07, 2004 at 08:01:47AM -0400, Richard B. Johnson wrote:
Also, when this driver is running, transferring large volumes
of data, the kernel decides that there have been too many interrupts, and
does:

Message from syslogd@chaos at Wed Oct 6 21:22:57 2004 ...
chaos kernel: Disabling IRQ #18

This, in spite of the fact that interrupts occur only when
DMA completion happens and new data are available, i.e.,
one interrupt every 16 megabytes of data transferred.

Who decided that it had a right to disable my interrupt????

the kernel did because you don't return the proper value for "I handled the
IRQ" from your ISR.

Do you know what that value is? I can't find it. I just returned 0
and it worked for awhile.

IRQ_HANDLED is you handled the irq, IRQ_NONE if you didn't



Okay. I'll fix that.



The kernel calls cleanup_module() and the printk() shows that it
was truly called.

I fail to find where you declare module_exit() in your sources


Well I don't. Is it now required? What does it do? If I put
in module_exit() and have it execute cleanup_module(), it

I found it..........

module_exit() is just a wrapper around cleanup_module().

cleanup_module() gets called. I could rename cleanup_module to
foo() and then use module_exit(foo);, which seems obtuse
since this has to compile on 2.4.x also, clearly not the
right thing to do.

`dmesg` clearly shown that the exit routine was called....

PCI: Enabling device 0000:02:04.0 (0106 -> 0107)
Analogic Corp Datalink Driver : Installed 12d6:8004 IRQ18 slot:0204 DMA:1f401000
Analogic Corp Datalink Driver : Initialization complete
Analogic Corp Datalink Driver : Module removed


Cheers,
Dick Johnson
Penguin : Linux version 2.6.5-1.358-noreg on an i686 machine (5537.79 BogoMips).
Note 96.31% of all statistics are fiction.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/