On 10/12/2010 05:12 PM, malc wrote:
Bottom line: QEMU assumes that long is large enough to hold a pointer,
this assumption is wrong, and your patch is wrong (not because it
miscalculates size of ABI long) but because of the assumptions current
code has. Whether two wrongs makes right is an open question, currently
Win64 isnot supported in any shape or form, even though Filip Navarra
once did a port[1] which never went into mainline.
[1] http://marc.info/?l=qemu-devel&m=124912692531724&w=2
His port apparently is fine with HOST_LONG_BITS == sizeof(void *), I'll
adjust for v2.
Paolo
