On Thu, 2016-03-10 at 17:22 +0200, Andy Shevchenko wrote:
> On Thu, Mar 10, 2016 at 4:59 PM, Borislav Petkov <b...@alien8.de>
> wrote:
> > On Thu, Mar 10, 2016 at 03:31:43PM +0200, Andy Shevchenko wrote:
> > > Looks like it lacks that one.
> > >
> > > # grep -i fxsr /proc/cpuinfo; echo $?
> > > 1
> >
> > Ok, so looking at where the warning comes from:
> >
> > [ 14.714533] WARNING: CPU: 0 PID: 823 at
> > arch/x86/include/asm/fpu/internal.h:163 fpu__clear+0x8c/0x160
> >
> > static inline void copy_kernel_to_fxregs(struct fxregs_state *fx)
> > {
> > int err;
> >
> > if (config_enabled(CONFIG_X86_32)) {
> > err = check_insn(fxrstor %[fx], "=m" (*fx), [fx]
> > "m" (*fx));
> > ^^^^^^^^^^^^^^^^^
> > } else {
> >
> > ...
> >
> > /* Copying from a kernel buffer to FPU registers should
> > never fail: */
> > WARN_ON_FPU(err);
> >
> >
> > and the stacktrace is pretty clear:
> >
> > flush_thread
> > > -> fpu__clear(&tsk->thread.fpu);
> > |-> we are eager by default here:
> >
> > if (!use_eager_fpu() || !static_cpu_has(X86_FEATURE_FPU)) {
> > /* FPU state will be reallocated lazily at the
> > first use. */
> > fpu__drop(fpu);
> > } else {
> >
> > --> we're in that branch.
> >
> > copy_init_fpstate_to_fpregs();
> > |-> copy_kernel_to_fxregs()
> >
> >
> > I think we should use FRSTOR on quark, i.e.,
> > copy_kernel_to_fregs().
> >
> > Does this untested wild guess even work?
> >
> > ---
> > diff --git a/arch/x86/kernel/fpu/core.c
> > b/arch/x86/kernel/fpu/core.c
> > index dea8e76d60c6..bbafe5e8a1a6 100644
> > --- a/arch/x86/kernel/fpu/core.c
> > +++ b/arch/x86/kernel/fpu/core.c
> > @@ -474,8 +474,11 @@ static inline void
> > copy_init_fpstate_to_fpregs(void)
> > {
> > if (use_xsave())
> > copy_kernel_to_xregs(&init_fpstate.xsave, -1);
> > - else
> > + else if (static_cpu_has(X86_FEATURE_FXSR))
> > copy_kernel_to_fxregs(&init_fpstate.fxsave);
> > + else
> > + copy_kernel_to_fregs(&init_fpstate.fsave);
> > +
>
> Obviously redundant line, otherwise it indeed works
>
> Tested-by: Andy Shevchenko <andy.shevche...@gmail.com>
>
> > }
> >
> > /*
>
>
>
It works but user-space FPU is broken; something's wrong with the
initial state of the FPU regs - it looks as though they aren't being
properly initialized and FPU context in the signal handler is wrong
too.
Linux 3.8.7:
/root@galileo:~# ./fpu
f is 10.000000 g is 10.100000
Double value is 0.000000
Double value is 0.100000
Double value is 0.200000
^Chandler value of variable is 0.300000
Double value is 0.300000
Double value is 0.400000
Linux-next + Boris' fix:
root@galileo:~# ./fpu
f is -nan g is -nan
Double value is 0.000000
Double value is 0.100000
Double value is 0.200000^C
handler value of variable is -nan
Double value is 0.300000
Double value is 0.400000^Z[1]+ Stopped
root@galileo:~# uname -aLinux galileo 4.5.0-rc7-next-20160310+ #185 Thu
Mar 10 15:11:10 GMT 2016 i586 GNU/Linux
#include <stdio.h>
#include <signal.h>
#include <string.h>
float a = 0;
void handler(int signum, siginfo_t *info, void *ptr)
{
printf("%s value of variable is %f\n", __func__, a);
}
int main(int argc, char *argv[])
{
struct sigaction sig;
float f = 10, g = 10.1f;
memset(&sig, 0x00, sizeof(sig));
sig.sa_sigaction = handler;
sig.sa_flags = SA_SIGINFO;
sigaction(SIGTERM, &sig, NULL);
sigaction(SIGINT, &sig, NULL);
printf("f is %f g is %f\n", f, g);
while(1) {
sleep(1);
printf("Double value is %f\n", a);
a += 0.1f;
}
return 0;
}
