Introduce a benchmark test for the fns(). It measures the total time
taken by fns() to process 1,000,000 test data generated using
get_random_long() for each n in the range [0, BITS_PER_LONG).

example:
test_bitops: fns:          5876762553 ns, 64000000 iterations

Signed-off-by: Kuan-Wei Chiu <visitorckw@gmail.com>
---

Changes in v3:
- Move the benchmark test for fns() to lib/test_bitops.c.
- Exclude the overhead of random number generation from the benchmark
  result.
- Change the output to print only a total gross instead of each n in
  the benchmark result.

 lib/test_bitops.c | 22 ++++++++++++++++++++++
 1 file changed, 22 insertions(+)

diff --git a/lib/test_bitops.c b/lib/test_bitops.c
index 3b7bcbee84db..ed939f124417 100644
--- a/lib/test_bitops.c
+++ b/lib/test_bitops.c
@@ -50,6 +50,26 @@ static unsigned long order_comb_long[][2] = {
 };
 #endif
 
+static unsigned long buf[1000000];
+
+static int __init test_fns(void)
+{
+	unsigned int i, n;
+	ktime_t time;
+
+	get_random_bytes(buf, sizeof(buf));
+	time = ktime_get();
+
+	for (n = 0; n < BITS_PER_LONG; n++)
+		for (i = 0; i < 1000000; i++)
+			fns(buf[i], n);
+
+	time = ktime_get() - time;
+	pr_err("fns:  %18llu ns, %6d iterations\n", time, BITS_PER_LONG * 1000000);
+
+	return 0;
+}
+
 static int __init test_bitops_startup(void)
 {
 	int i, bit_set;
@@ -94,6 +114,8 @@ static int __init test_bitops_startup(void)
 	if (bit_set != BITOPS_LAST)
 		pr_err("ERROR: FOUND SET BIT %d\n", bit_set);
 
+	test_fns();
+
 	pr_info("Completed bitops test\n");
 
 	return 0;
-- 
2.34.1

From: Kuan-Wei Chiu
> Sent: 01 May 2024 08:17
> 
> Introduce a benchmark test for the fns(). It measures the total time
> taken by fns() to process 1,000,000 test data generated using
> get_random_long() for each n in the range [0, BITS_PER_LONG).
> 
> example:
> test_bitops: fns:          5876762553 ns, 64000000 iterations

Great benchmark....

The compiler almost certainly optimises it all away.

Assigning the result of fns() to a file scope (global) volatile int
should stop that happening.

And a real test would actually check the result - just in case
someone does something silly.

	David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)

Hi David,

On Sun, May 05, 2024 at 01:03:23PM +0000, David Laight wrote:
> From: Kuan-Wei Chiu
> > Sent: 01 May 2024 08:17
> > 
> > Introduce a benchmark test for the fns(). It measures the total time
> > taken by fns() to process 1,000,000 test data generated using
> > get_random_long() for each n in the range [0, BITS_PER_LONG).
> > 
> > example:
> > test_bitops: fns:          5876762553 ns, 64000000 iterations
> 
> Great benchmark....
> 
> The compiler almost certainly optimises it all away.
> 
> Assigning the result of fns() to a file scope (global) volatile int
> should stop that happening.
> 
Thank you for your review. There is an updated v5 of this patch [1],
which has already been accepted and included in Yury's bitmap-for-next
branch of the bitmap tree. In the v5 patch, we have addressed the issue
you mentioned regarding the use of volatile variables to avoid compiler
optimizations.

> And a real test would actually check the result - just in case
> someone does something silly.
>
The fns() function is mainly a helper for find_nth_bit(), so its
accuracy should have been checked in find_nth_bit()'s tests. If you
want unit tests for fns() here too, that sounds good to me, but it
would likely be a separate patch. I'm happy to do it if you'd like.

Regards,
Kuan-Wei

> 	David
> 
> -
> Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
> Registration No: 1397386 (Wales)
>

On Mon, May 06, 2024 at 01:27:25AM +0800, Kuan-Wei Chiu wrote:
> Hi David,
> 
> On Sun, May 05, 2024 at 01:03:23PM +0000, David Laight wrote:
> > From: Kuan-Wei Chiu
> > > Sent: 01 May 2024 08:17
> > > 
> > > Introduce a benchmark test for the fns(). It measures the total time
> > > taken by fns() to process 1,000,000 test data generated using
> > > get_random_long() for each n in the range [0, BITS_PER_LONG).
> > > 
> > > example:
> > > test_bitops: fns:          5876762553 ns, 64000000 iterations
> > 
> > Great benchmark....
> > 
> > The compiler almost certainly optimises it all away.
> > 
> > Assigning the result of fns() to a file scope (global) volatile int
> > should stop that happening.
> > 
> Thank you for your review. There is an updated v5 of this patch [1],
> which has already been accepted and included in Yury's bitmap-for-next
> branch of the bitmap tree. In the v5 patch, we have addressed the issue
> you mentioned regarding the use of volatile variables to avoid compiler
> optimizations.
>
[1]: https://lore.kernel.org/lkml/20240502092443.6845-2-visitorckw@gmail.com/

> > And a real test would actually check the result - just in case
> > someone does something silly.
> >
> The fns() function is mainly a helper for find_nth_bit(), so its
> accuracy should have been checked in find_nth_bit()'s tests. If you
> want unit tests for fns() here too, that sounds good to me, but it
> would likely be a separate patch. I'm happy to do it if you'd like.
> 
> Regards,
> Kuan-Wei
> 
> > 	David
> > 
> > -
> > Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
> > Registration No: 1397386 (Wales)
> >