Re: [Qemu-devel] [PATCH 4/5] milkymist-softusb: Don't map RAM memory regions in the device itself

[Paolo Bonzini]

Il 15/02/2013 19:38, Peter Maydell ha scritto:
>>> But these memory regions belong to this device -- we own them and
>>> >> they won't go away until the device is destroyed [which is never,
>>> >> as it happens, for this device.] More generally, if it's valid
>>> >> for us to hold a MemoryRegion* and call memory_region_get_ram_ptr()
>>> >> in the read/write function, it's just as valid to keep the ram pointer:
>>> >> the two have exactly matching lifetimes, unless I'm missing something.
>> >
>> > No, you're not: "In practice it should be handled just fine by reference
>> > counting, but I still find it harder to wrap my head around it".
> I'm still confused. Memory regions aren't reference counted.
> We're the device, we own the MemoryRegion, we can happily
> do whatever we like with it for the lifetime of the device.

Yes: reference counting of the device.

> I don't propose to change this patch for v2 of this series,
> since there doesn't seem to be any need to do so.

Fair enough.

Paolo