DragonFly BSD
DragonFly commits List (threaded) for 2009-11
[Date Prev][Date Next]  [Thread Prev][Thread Next]  [Date Index][Thread Index]

Re: git: Revert "IO APIC: Assign pins dedicated to PCI in the early stage."


From: Matthew Dillon <dillon@xxxxxxxxxxxxxxxxxxxx>
Date: Mon, 9 Nov 2009 19:12:56 -0800 (PST)

:If you are going to do it, I think the first step will be supporting
:256 intr vectors.  And the major work will concern how to properly
:handle gd_fpending.  I was thinking about splitting it into two parts,
:one gd_fpending_group (uint8_t should be enough), and one
:gd_fpenfing[8] (uint32_t mask array), but never find time to do it...
:
:Best Regards,
:sephe
:
:-- 
:Live Free or Die

    That seems to be a reasonable way to do it.  Even better, we could
    probably combine spending, fpending, and ipending all together,
    bump the vectors up to 1024 (32x32), and then run it through an array
    of (C) function pointers for various actions.  That would remove a
    large chunk of the code duplication.

    I'll think about it a bit.

					-Matt
					Matthew Dillon 
					<dillon@backplane.com>



[Date Prev][Date Next]  [Thread Prev][Thread Next]  [Date Index][Thread Index]