Hi Richard, Thanks! I'll test this tomorrow morning and we can choose whether to include your proper fix or the workaround.
I think we have time assuming we send out PRs tomorrow. Given our important fixes have review including either this fix by tomorrow or the workaround, and Philippe has reviewed our other important bugs fixes, then we should be fine. Then after getting the critical and important fixes out of the way, then I perhaps make a PR for the other reviewed changes, although these might best wait until QEMU 2.13 opens. Thanks again, Michael. On Tue, Mar 27, 2018 at 7:22 PM, Richard Henderson < richard.hender...@linaro.org> wrote: > Since it was my patch that broke FP state tracking in the > first place, I feel obligated to fix it again. > > Mark mstatus[fs] as dirty whenever we write to the file. > This can be optimized by only doing so once within a TB > which initially began with a clean file. > > I have not yet put together an environment that can test > this, so I'll need someone else to give it a go. > > > r~ > > > Richard Henderson (2): > target/riscv: Split out mstatus_fs from tb_flags during translation > target/riscv: Mark MSTATUS_FS dirty > > target/riscv/cpu.h | 6 +++--- > target/riscv/op_helper.c | 25 ++++++++++++++++-------- > target/riscv/translate.c | 50 ++++++++++++++++++++++++++++++ > ++++++++++++------ > 3 files changed, 64 insertions(+), 17 deletions(-) > > -- > 2.14.3 > >
