Re: mtd_cpl question

[Chuck Tuffli <chuck_tuffli@xxxxxxxxxxx>]

Matthew Dillon wrote:
> :The reason I ask is that I have the bus_alloc_resource() side of MSI 
> :(message signaled interrupts) working and wanted to connect it to the 
> :rest of the interrupt code via bus_setup_intr(). Do you have any 
> :short/long term recommendations? Do you want help on any of the above?
> :
> :Alternatively, how badly would things break if I went around the system 
> :and wrote the IDT entries directly assuming there would be no overlap in 
> :the IRQ/MSI ranges?
> :
> :-- 
> :Chuck Tuffli
> :Agilent Technologies
>
>     I would say that bus_setup_intr() should specifically request them
>     with a flag.  In particular, since we can't support SPL's on those
>     the device drivers that use them should not require the use of SPLs.
>     This means the setup call has to pass an interrupt type of INTR_TYPE_MISC
>     (i.e. no cpl mask involved).
>
>     Right now that is just the EM driver (I am about to commit the EM work
>     along with some major polling work).  I believe once I do this commit
>     today you will then be able to use that as a basis for further work,
>     either to work with the EM driver itself or to see what I did to convert
>     a driver that you have hardware for.  A network driver, please, disk
>     drivers still require SPLBIO to work.

Ok, I will start with the em driver. Are you suggesting a network driver 
because the network stack does not depend on the BGL (I think that's 
what you meant above wrt SPL)?

>     Are MSI interrupts edge triggered or level triggered?  There's more
>     work involved if they are edge triggered.

PCIe defines them to be edge, but I'd have to look at the base spec to 
know what happens on PCI.

>     You should be able to write IDT entries within the bounds of the IDT
>     table, as long as they do not interfere with existing IDT entries for
>     slow and fast interrupt vectors.  I'm presuming that you will use the
>     same API that the interrupt code uses to access those entries, right ?

I'd rather use the existing API unless there is something that prevents it.

--
Chuck Tuffli
Agilent Technologies