RE: [Patch] speed up interrupt handling

From: Mallick, Asit K <>
Date: 2004-02-28 20:05:06
The patch is fine but I will be concerned about using disable/enable
(masking and unmasking at IOAPIC) especially for edge triggered
interrupts. There is a possibility of loosing interrupts. The user
handler handles the device interrupt with this interrupt masked. A new
interrupt could be generated by the device while the interrupt is still
masked. This new interrupt if edge triggered will be lost since IOAPIC
does not keep pending interrupts. Level triggered interrupts should be

>-----Original Message-----
>[] On Behalf Of David Mosberger
>Sent: Friday, February 27, 2004 9:48 PM
>To: Peter Chubb
>Cc:; Feldman, Scott; Grant Grundler; Peter 
>Subject: RE: [Patch] speed up interrupt handling
>>>>>> On Sat, 28 Feb 2004 15:22:41 +1100, Peter Chubb 
><> said:
>  Peter> Sorry, I should have been more clear.  For the 
>user-mode driver, we
>  Peter> need to disable the interrupt so that it can be 
>acknowledged to the
>  Peter> device in the user-level code after the in-kernel 
>generic  interrupt
>  Peter> handler has returned.
>Ah, now it makes sense.  OK, I don't think it will have much of a
>practical effect on the normal kernel, but the patch is obvious and
>simple enough that I don't see a reason not to do it.
>	--david
>To unsubscribe from this list: send the line "unsubscribe 
>linux-ia64" in
>the body of a message to
>More majordomo info at
To unsubscribe from this list: send the line "unsubscribe linux-ia64" in
the body of a message to
More majordomo info at
Received on Sat Feb 28 04:05:52 2004

This archive was generated by hypermail 2.1.8 : 2005-08-02 09:20:23 EST