Re: [Qemu-devel] [PATCH 2/3] e500: Adding CCSR memory region

[Avi Kivity <avi@redhat.com>]

On 10/09/2012 06:57 PM, Bhushan Bharat-R65777 wrote:
> 
> 
>> -----Original Message-----
>> From: Avi Kivity [mailto:avi@redhat.com]
>> Sent: Tuesday, October 09, 2012 10:24 PM
>> To: Bhushan Bharat-R65777
>> Cc: Andreas Färber; agraf@suse.de; qemu-ppc@nongnu.org; qemu-devel@nongnu.org
>> Subject: Re: [PATCH 2/3] e500: Adding CCSR memory region
>> 
>> On 10/09/2012 06:45 PM, Bhushan Bharat-R65777 wrote:
>> >
>> > What about adding a API:
>> > void sysbus_mmio_map_to_mr(SysBusDevice *dev, int n, target_phys_addr_t addr,
>> >                            MemoryRegion *mr) {
>> >     assert(n >= 0 && n < dev->num_mmio);
>> >
>> >     if (dev->mmio[n].addr == addr) {
>> >         /* ??? region already mapped here.  */
>> >         return;
>> >     }
>> >     if (dev->mmio[n].addr != (target_phys_addr_t)-1) {
>> >         /* Unregister previous mapping.  */
>> >         memory_region_del_subregion(mr, dev->mmio[n].memory);
>> >     }
>> >     dev->mmio[n].addr = addr;
>> >     memory_region_add_subregion(mr, addr, dev->mmio[n].memory); }
>> >
>> 
>> I think you can just use sysbus_mmio_get_region().  There are plenty of other
>> users, so there's precedent.
> 
> You mean something like this : memory_region_add_subregion(mr, addr, sysbus_mmio_get_region(dev, 0);
> 
> Ok, but this will still not resolve the issue of not setting the "dev->mmio[n].addr", no ?

Correct.  But there are 20 uses already, so it can't matter much.  If
someone wants to fix them, they can write a new API and do a sweep.

But really, sysbus just needs to go away.  It's pointless to give it
more and more features.


-- 
error compiling committee.c: too many arguments to function