>>>>> On Mon, 11 Oct 2004 09:54:24 -0700, Jun Sun <email@example.com> said:
jsun> I actually don't see which approach is more "simple and robust".
jsun> In terms of being simple, allowing kernel mode FPU trap is
jsun> definitely simpler.
jsun> If you can't find any pitfalls of this approach it is actually
jsun> robust. The new FPU code is already greatly simplified. It is
jsun> possible kernel FPU trap is not that evil anymore (assuming
jsun> kernel continues voluntarily not using FPU).
Hmm... OK, I agree enabling FPU trap in kernel seems simple. I tried
it today but it did not work unfortunately. Just modifying a
following line in traps.c was not enough.
die_if_kernel("do_cpu invoked from kernel context!", regs);
One point I found is do_cpu() must enable CU1 bit in pt_regs also.
Another problem is that resume(), own_fpu() and lose_fpu() manipulate
CU1 bit in only first level kernel stack (KSTK_STATUS(current)).
current->thread.cp0_status may be manipulated also. Modifying
resume() looks too dangerous to me.
Anyway, in math-emu case, we should call lose_fpu() BEFORE running the
emulator. Since the emulator never use real FPU, we naturally call
own_fpu() AFTER the returning from the emulator. It's simple.
Enabling kernel FPU trap is not needed.
>> It is safe? get_user/put_user will fail if preempt disabled. Excerpt
>> from do_page_fault:
>> * If we're in an interrupt or have no user
>> * context, we must not take the fault..
>> if (in_atomic() || !mm)
>> goto bad_area_nosemaphore;
jsun> It should be safe. This might be a bug in kernel. I bet
jsun> in_atomic() is "in_interrupt()" in older version of the code,
jsun> which seems to be the correct code. When you diable preemption
jsun> you still have a process context.
But many archs have above check in do_page_fault() long time (since
2.5.32 at i386). Are these all broken?
jsun> BTW, have you thought about possibly the third approach, which
jsun> is to somehow isolate the code segment where get_user/put_user
jsun> could cause page faults and then make sure all FPU manipulations
jsun> will succeed?
I took this approach to fix sigcontext problem. Please look at
signal.c in my patch.
Thank you for quick comments.