Nouveau - May 2019 - [RFC][PATCH] kernel.h: Add generic roundup_64() macro

On Thu, 23 May 2019 08:10:44 -0700
Linus Torvalds <torvalds at linux-foundation.org> wrote:
> On Thu, May 23, 2019 at 7:00 AM Steven Rostedt <rostedt at
goodmis.org> wrote:
> >
> > +# define roundup_64(x, y) (                            \
> > +{                                                      \
> > +       typeof(y) __y = y;                              \
> > +       typeof(x) __x = (x) + (__y - 1);                \
> > +       do_div(__x, __y);                               \
> > +       __x * __y;                                      \
> > +}                                                      \
> 
> The thing about this is that it absolutely sucks for power-of-two
arguments.
> 
> The regular roundup() that uses division has the compiler at least
> optimize them to shifts - at least for constant cases. But do_div() is
> meant for "we already know it's not a power of two", and the
compiler
> doesn't have any understanding of the internals.
> 
> And it looks to me like the use case you want this for is very much
> probably a power of two. In which case division is all kinds of just
> stupid.
> 
> And we already have a power-of-two round up function that works on
> u64. It's called "round_up()".
> 
> I wish we had a better visual warning about the differences between
> "round_up()" (limited to powers-of-two, but efficient, and works
with
> any size) and "roundup()" (generic, potentially horribly slow,
and
> doesn't work for 64-bit on 32-bit).
> 
> Side note: "round_up()" has the problem that it uses
"x" twice.
> 
> End result: somebody should look at this, but I really don't like the
> "force division" case that is likely horribly slow and nasty.
I haven't yet tested this, but what about something like the following:

# define roundup_64(x, y) (				\
{							\
	typeof(y) __y;					\
	typeof(x) __x;					\
							\
	if (__builtin_constant_p(y) &&			\
	    !(y & (y >> 1))) {				\
		__x = round_up(x, y);			\
	} else {					\
		__y = y;				\
		__x = (x) + (__y - 1);			\
		do_div(__x, __y);			\
		__x = __x * __y;			\
	}						\
	__x;						\
}							\
)

If the compiler knows enough that y is a power of two, it will use the
shift version. Otherwise, it doesn't know enough and would divide
regardless. Or perhaps forget about the constant check, and just force
the power of two check:

# define roundup_64(x, y) (				\
{							\
	typeof(y) __y = y;				\
	typeof(x) __x;					\
							\
	if (!(__y & (__y >> 1))) {			\
		__x = round_up(x, y);			\
	} else {					\
		__x = (x) + (__y - 1);			\
		do_div(__x, __y);			\
		__x = __x * __y;			\
	}						\
	__x;						\
}							\
)

This way even if the compiler doesn't know that this is a power of two,
it will still do the shift if y ends up being one.

-- Steve

On Thu, May 23, 2019 at 8:27 AM Steven Rostedt <rostedt at goodmis.org>
wrote:
>
> I haven't yet tested this, but what about something like the following:
So that at least handles the constant case that the normal "roundup()"
case also handles.

At the same time, in the case you are talking about, I really do
suspect that we have a (non-constant) power of two, and that you
should have just used "round_up()" which works fine regardless of
size, and is always efficient.

On a slight tangent.. Maybe we should have something like this:

#define size_fn(x, prefix, ...) ({                      \
        typeof(x) __ret;                                \
        switch (sizeof(x)) {                            \
        case 1: __ret = prefix##8(__VA_ARGS__); break;  \
        case 2: __ret = prefix##16(__VA_ARGS__); break; \
        case 4: __ret = prefix##32(__VA_ARGS__); break; \
        case 8: __ret = prefix##64(__VA_ARGS__); break; \
        default: __ret = prefix##bad(__VA_ARGS__);      \
        } __ret; })

#define type_fn(x, prefix, ...) ({                              \
        typeof(x) __ret;                                        \
        if ((typeof(x))-1 > 1)                                  \
                __ret = size_fn(x, prefix##_u, __VA_ARGS__);    \
        else                                                    \
                __ret = size_fn(x, prefix##_s, __VA_ARGS__);    \
        __ret; })

which would allow typed integer functions like this. So you could do
something like

     #define round_up(x, y) size_fn(x, round_up_size, x, y)

and then you define functions for round_up_size8/16/32/64 (and you
have toi declare - but not define - round_up_sizebad()).

Of course, you probably want the usual "at least use 'int'"
semantics,
in which case the "type" should be "(x)+0":

     #define round_up(x, y) size_fn((x)+0, round_up_size, x, y)

 and the 8-bit and 16-bit cases will never be used.

We have a lot of cases where we end up using "type overloading" by
size. The most explicit case is perhaps "get_user()" and
"put_user()",
but this whole round_up thing is another example.

Maybe we never really care about "char" and "short", and
always want
just the "int-vs-long-vs-longlong"? That would make the cases simpler
(32 and 64). And maybe we never care about sign. But we could try to
have some unified helper model like the above..

                  Linus

On Thu, 23 May 2019 09:51:29 -0700
Linus Torvalds <torvalds at linux-foundation.org> wrote:
> On Thu, May 23, 2019 at 8:27 AM Steven Rostedt <rostedt at
goodmis.org> wrote:
> >
> > I haven't yet tested this, but what about something like the
following:
> 
> So that at least handles the constant case that the normal
"roundup()"
> case also handles.
> 
> At the same time, in the case you are talking about, I really do
> suspect that we have a (non-constant) power of two, and that you
> should have just used "round_up()" which works fine regardless of
> size, and is always efficient.
I think you are correct in this.

       act_size = roundup_64(attr->length,
MLX5_SW_ICM_BLOCK_SIZE(dm_db->dev));

Where we have:

#define MLX5_SW_ICM_BLOCK_SIZE(dev) (1 << MLX5_LOG_SW_ICM_BLOCK_SIZE(dev))

Which pretty much guarantees that it is a power of two. Thus, the real
fix here is simply to s/roundup/round_up/ as you suggest.
> 
> On a slight tangent.. Maybe we should have something like this:
> 
> #define size_fn(x, prefix, ...) ({                      \
>         typeof(x) __ret;                                \
>         switch (sizeof(x)) {                            \
>         case 1: __ret = prefix##8(__VA_ARGS__); break;  \
>         case 2: __ret = prefix##16(__VA_ARGS__); break; \
>         case 4: __ret = prefix##32(__VA_ARGS__); break; \
>         case 8: __ret = prefix##64(__VA_ARGS__); break; \
>         default: __ret = prefix##bad(__VA_ARGS__);      \
>         } __ret; })
> 
> #define type_fn(x, prefix, ...) ({                              \
>         typeof(x) __ret;                                        \
>         if ((typeof(x))-1 > 1)                                  \
>                 __ret = size_fn(x, prefix##_u, __VA_ARGS__);    \
>         else                                                    \
>                 __ret = size_fn(x, prefix##_s, __VA_ARGS__);    \
>         __ret; })
> 
> which would allow typed integer functions like this. So you could do
> something like
> 
>      #define round_up(x, y) size_fn(x, round_up_size, x, y)
> 
> and then you define functions for round_up_size8/16/32/64 (and you
You mean define functions for round_up_size_{u|s}8/16/32/64
> have toi declare - but not define - round_up_sizebad()).
> 
> Of course, you probably want the usual "at least use
'int'" semantics,
> in which case the "type" should be "(x)+0":
> 
>      #define round_up(x, y) size_fn((x)+0, round_up_size, x, y)
> 
>  and the 8-bit and 16-bit cases will never be used.
I'm curious to what the advantage of that is?
> 
> We have a lot of cases where we end up using "type overloading"
by
> size. The most explicit case is perhaps "get_user()" and
"put_user()",
> but this whole round_up thing is another example.
> 
> Maybe we never really care about "char" and "short",
and always want
> just the "int-vs-long-vs-longlong"? That would make the cases
simpler
> (32 and 64). And maybe we never care about sign. But we could try to
> have some unified helper model like the above..
It may be simpler and perhaps more robust if we keep the char and short
cases.

I'm fine with adding something like this for round_up(), but do we want
to have a generic roundup_64() as well? I'm also thinking that we
perhaps should test for power of two on roundup():

#define roundup(x, y) (					\
{							\
	typeof(y) __y = y;				\
	typeof(x) __x;					\
							\
	if (__y & (__y - 1))				\
		__x = round_up(x, __y);			\
	else						\
		__x = (((x) + (__y - 1)) / __y) * __y;	\
	__x;						\
})


-- Steve

On 23/05/2019 16:27, Steven Rostedt wrote:
>
> I haven't yet tested this, but what about something like the following:
>
> ...perhaps forget about the constant check, and just force
> the power of two check:
>
> 							\
> 	if (!(__y & (__y >> 1))) {			\
> 		__x = round_up(x, y);			\
> 	} else {					\
You probably want

            if (!(__y & (__y - 1))

--

Roger

On Fri, 24 May 2019 16:11:14 +0100
Roger Willcocks <roger at filmlight.ltd.uk> wrote:
> On 23/05/2019 16:27, Steven Rostedt wrote:
> >
> > I haven't yet tested this, but what about something like the
following:
> >
> > ...perhaps forget about the constant check, and just force
> > the power of two check:
> >
> > 							\
> > 	if (!(__y & (__y >> 1))) {			\
> > 		__x = round_up(x, y);			\
> > 	} else {					\  
> 
> You probably want
> 
>             if (!(__y & (__y - 1))
> 
> --
Yes I do. I corrected it in my next email.

 http://lkml.kernel.org/r/20190523133648.591f9e78 at gandalf.local.home
> #define roundup(x, y) (					\
> {							\
> 	typeof(y) __y = y;				\
> 	typeof(x) __x;					\
> 							\
> 	if (__y & (__y - 1))				\
> 		__x = round_up(x, __y);			\
> 	else						\
> 		__x = (((x) + (__y - 1)) / __y) * __y;	\
> 	__x;						\
> })

-- Steve