On 12/2/21 9:21 PM, Emilio Cota wrote:
On Thu, Dec 2, 2021 at 4:47 AM Vasilev Oleg <vasilev.o...@huawei.com> wrote:
The mentioned paper[4] also describes other possible improvements.
Some of those are already implemented (such as victim TLB and dynamic
size for TLB), but others are not (e.g. TLB lookup uninlining and
set-associative TLB layer). Do you think those improvements
worth trying?
I cannot find the emails, but I do remember that Richard wrote tcg-i386 patches
for uninlining TLB lookups. Unfortunately they resulted in a slowdown on
modern machines.
That code is still around at
https://github.com/rth7680/qemu/tree/tcg-softmmu-ool
r~