On Sat, 12 Dec 2020 at 18:27, BALATON Zoltan <balaton@eik.bme.hu> wrote:
On Sat, 12 Dec 2020, Peter Maydell wrote:
Currently the PPC UIC ("Universal Interrupt Controller") is implemented
as a non-QOM device in ppc4xx_devs.c. Convert it to a proper QOM device
in hw/intc.
The ppcuic_init() function is retained for the moment with its current
interface; in subsequent commits this will be tidied up to avoid the
allocation of an irq array.
This conversion adds VMState support.
It leaves the LOG_UIC() macro as-is to maximise the extent to which
this is simply code-movement rather than a rewrite (in new code it
would be better to use tracepoints).
The default property values for dcr-base and use-vectors are set to
match those use by most of our boards with a UIC.
Would this be easier to review if it's split into two patches, first doing
the conversion in place, then another just doing code moving? Maybe we
could spot some problem that way easier?
That didn't really feel very feasible to me. You can see that
mostly it's code motion using the git diff --color-moved option
if you look at the commit.