On 06/20/2012 05:26 PM, Peter Maydell wrote:
On 20 June 2012 22:14, Anthony Liguori<anth...@codemonkey.ws>  wrote:
Why not make life easy and fix dma_addr_t to 64-bit?

...for that matter weren't we tossing around the idea of just
making target_phys_addr_t 64 bits for everything? (I actually
want to do this for target-arm anyway; last time I did some
quick smoke-tests of performance it didn't seem to hurt really
even on a 32 bit host, and it avoids having to put the A15 in
a different qemu-system-* binary to the other cores.)

Didn't you whine and moan about the impact to printf()s last time I did this? 
;-)

I can refresh my patch and resubmit..

Regards,

Anthony Liguori


-- PMM


Reply via email to