Hi, Paul Burton,
On Mon, Aug 24, 2020 at 3:19 PM Huacai Chen <chenhc@xxxxxxxxxx> wrote:
>
> Hi, Pei,
>
> On Fri, Aug 21, 2020 at 3:26 PM Huang Pei <huangpei@xxxxxxxxxxx> wrote:
> >
> > own_fp_inatomic() does not restore/initialize MSACSR and upper MSA regs, so
> > MSACSR and MSA upper regs's value from previous task on current cpu can leak
> > into current task and cause unpredictable behavior when MSA context not
> > initialized.
> >
> > Signed-off-by: Huang Pei <huangpei@xxxxxxxxxxx>
> > ---
> > arch/mips/kernel/traps.c | 2 ++
> > 1 file changed, 2 insertions(+)
> >
> > diff --git a/arch/mips/kernel/traps.c b/arch/mips/kernel/traps.c
> > index 38aa07ccdbcc..99a8f1b7b522 100644
> > --- a/arch/mips/kernel/traps.c
> > +++ b/arch/mips/kernel/traps.c
> > @@ -1287,6 +1287,8 @@ static int enable_restore_fp_context(int msa)
> > err = own_fpu_inatomic(1);
> > if (msa && !err) {
> > enable_msa();
> > + write_msa_csr(current->thread.fpu.msacsr);
> > + init_msa_upper();
> I think write_msa_csr(current->thread.fpu.msacsr) is needed here, but
> I don't know whether init_msa_upper() is also needed. Maybe you should
> investegate cc97ab235f3fe32401ca198cebe6f42642e9 ("MIPS: Simplify FP
> context initialization") which removes init_msa_upper().
I think I need you to explain why removing init_msa_upper() here.
Huacai
>
> Huacai
> > set_thread_flag(TIF_USEDMSA);
> > set_thread_flag(TIF_MSA_CTX_LIVE);
> > }
> > --
> > 2.17.1
> >
