Michael Chan wrote:
"Jeff Garzik" wrote:
I would prefer that a new 'static inline' function was created,
containing the core interrupt handler code.
This avoids the code duplication you have in this patch, in
new function
tg3_msi().
Jeff
There are enough subtle differences between tg3_interrupt() and tg3_msi()
that makes it impractical to create a core inline function to be shared by
the two. To share a meaningful chunk of the code, it will have to check for
MSI mode and skip some of the register reads. The reason I created tg3_msi()
was to avoid all conditionals.
After reviewing the existing tg3.c code again, I agree. Objection
withdrawn.
Jeff
|