commit a37fbe666c upstream.
The number of times yet another open coded
`BITS_TO_LONGS(nbits) * sizeof(long)` can be spotted is huge.
Some generic helper is long overdue.
Add one, bitmap_size(), but with one detail.
BITS_TO_LONGS() uses DIV_ROUND_UP(). The latter works well when both
divident and divisor are compile-time constants or when the divisor
is not a pow-of-2. When it is however, the compilers sometimes tend
to generate suboptimal code (GCC 13):
48 83 c0 3f add $0x3f,%rax
48 c1 e8 06 shr $0x6,%rax
48 8d 14 c5 00 00 00 00 lea 0x0(,%rax,8),%rdx
%BITS_PER_LONG is always a pow-2 (either 32 or 64), but GCC still does
full division of `nbits + 63` by it and then multiplication by 8.
Instead of BITS_TO_LONGS(), use ALIGN() and then divide by 8. GCC:
8d 50 3f lea 0x3f(%rax),%edx
c1 ea 03 shr $0x3,%edx
81 e2 f8 ff ff 1f and $0x1ffffff8,%edx
Now it shifts `nbits + 63` by 3 positions (IOW performs fast division
by 8) and then masks bits[2:0]. bloat-o-meter:
add/remove: 0/0 grow/shrink: 20/133 up/down: 156/-773 (-617)
Clang does it better and generates the same code before/after starting
from -O1, except that with the ALIGN() approach it uses %edx and thus
still saves some bytes:
add/remove: 0/0 grow/shrink: 9/133 up/down: 18/-538 (-520)
Note that we can't expand DIV_ROUND_UP() by adding a check and using
this approach there, as it's used in array declarations where
expressions are not allowed.
Add this helper to tools/ as well.
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Acked-by: Yury Norov <yury.norov@gmail.com>
Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 65a0d3c146 ]
If the input is out of the range of the allowed values, either larger than
the largest value or closer to zero than the smallest non-zero allowed
value, then a division by zero would occur.
In the case of input too large, the division by zero will occur on the
first iteration. The best result (largest allowed value) will be found by
always choosing the semi-convergent and excluding the denominator based
limit when finding it.
In the case of the input too small, the division by zero will occur on the
second iteration. The numerator based semi-convergent should not be
calculated to avoid the division by zero. But the semi-convergent vs
previous convergent test is still needed, which effectively chooses
between 0 (the previous convergent) vs the smallest allowed fraction (best
semi-convergent) as the result.
Link: https://lkml.kernel.org/r/20210525144250.214670-1-tpiepho@gmail.com
Fixes: 323dd2c3ed ("lib/math/rational.c: fix possible incorrect result from rational fractions helper")
Signed-off-by: Trent Piepho <tpiepho@gmail.com>
Reported-by: Yiyuan Guo <yguoaz@gmail.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: Oskar Schirmer <oskar@scara.com>
Cc: Daniel Latypov <dlatypov@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
People report that utime and stime from /proc/<pid>/stat become very
wrong when the numbers are big enough, especially if you watch these
counters incrementally.
Specifically, the current implementation of: stime*rtime/total,
results in a saw-tooth function on top of the desired line, where the
teeth grow in size the larger the values become. IOW, it has a
relative error.
The result is that, when watching incrementally as time progresses
(for large values), we'll see periods of pure stime or utime increase,
irrespective of the actual ratio we're striving for.
Replace scale_stime() with a math64.h helper: mul_u64_u64_div_u64()
that is far more accurate. This also allows architectures to override
the implementation -- for instance they can opt for the old algorithm
if this new one turns out to be too expensive for them.
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200519172506.GA317395@hirez.programming.kicks-ass.net
Pull kselftest updates from Shuah Khan:
"This consists of:
- Several fixes from Masami Hiramatsu to improve coverage for lib and
sysctl tests.
- Clean up to vdso test and a new test for getcpu() from Mark Brown.
- Add new gen_tar selftests Makefile target generate selftest package
running "make gen_tar" in selftests directory from Veronika
Kabatova.
- Other miscellaneous fixes to timens, exec, tpm2 tests"
* tag 'linux-kselftest-5.8-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest:
selftests/sysctl: Make sysctl test driver as a module
selftests/sysctl: Fix to load test_sysctl module
lib: Make test_sysctl initialized as module
lib: Make prime number generator independently selectable
selftests/ftrace: Return unsupported if no error_log file
selftests/ftrace: Use printf for backslash included command
selftests/timens: handle a case when alarm clocks are not supported
Kernel selftests: Add check if TPM devices are supported
selftests: vdso: Add a selftest for vDSO getcpu()
selftests: vdso: Use a header file to prototype parse_vdso API
selftests: vdso: Rename vdso_test to vdso_test_gettimeofday
selftests/exec: Verify execve of non-regular files fail
selftests: introduce gen_tar Makefile target
Make prime number generator independently selectable from
kconfig. This allows us to enable CONFIG_PRIME_NUMBERS=m
and run the tools/testing/selftests/lib/prime_numbers.sh
without other DRM selftest modules.
Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
Reviewed-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Luis Chamberlain <mcgrof@kernel.org>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
In some cases the previous algorithm would not return the closest
approximation. This would happen when a semi-convergent was the
closest, as the previous algorithm would only consider convergents.
As an example, consider an initial value of 5/4, and trying to find the
closest approximation with a maximum of 4 for numerator and denominator.
The previous algorithm would return 1/1 as the closest approximation,
while this version will return the correct answer of 4/3.
To do this, the main loop performs effectively the same operations as it
did before. It must now keep track of the last three approximations,
n2/d2 .. n0/d0, while before it only needed the last two.
If an exact answer is not found, the algorithm will now calculate the
best semi-convergent term, t, which is a single expression with two
divisions:
min((max_numerator - n0) / n1, (max_denominator - d0) / d1)
This will be used if it is better than previous convergent. The test
for this is generally a simple comparison, 2*t > a. But in an edge
case, where the convergent's final term is even and the best allowable
semi-convergent has a final term of exactly half the convergent's final
term, the more complex comparison (d0*dp > d1*d) is used.
I also wrote some comments explaining the code. While one still needs
to look up the math elsewhere, they should help a lot to follow how the
code relates to that math.
This routine is used in two places in the video4linux code, but in those
cases it is only used to reduce a fraction to lowest terms, which the
existing code will do correctly. This could be done more efficiently
with a different library routine but it would still be the Euclidean
alogrithm at its heart. So no change.
The remain users are places where a fractional PLL divider is
programmed. What would happen is something asked for a clock of X MHz
but instead gets Y MHz, where Y is close to X but not exactly due to the
hardware limitations. After this change they might, in some cases, get
Y' MHz, where Y' is a little closer to X then Y was.
Users like this are: Three UARTs, in 8250_mid, 8250_lpss, and imx. One
GPU in vp4_hdmi. And three clock drivers, clk-cdce706, clk-si5351, and
clk-fractional-divider. The last is a generic clock driver and so would
have more users referenced via device tree entries.
I think there's a bug in that one, it's limiting an N bit field that is
offset-by-1 to the range 0 .. (1<<N)-2, when it should be (1<<N)-1 as
the upper limit.
I have an IMX system, one of the UARTs using this, so I can provide a
real example. If I request a custom baud rate of 1499978, the driver
will program the PLL to produce a baud rate of 1500000. After this
change, the fractional divider in the UART is programmed to a ratio of
65535/65536, which produces a baud rate of 1499977.0625. Closer to the
requested value.
Link: http://lkml.kernel.org/r/20190330205855.19396-1-tpiepho@gmail.com
Signed-off-by: Trent Piepho <tpiepho@gmail.com>
Cc: Oskar Schirmer <oskar@scara.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Add SPDX license identifiers to all Make/Kconfig files which:
- Have no license information of any form
These files fall under the project license, GPL v2 only. The resulting SPDX
license identifier is:
GPL-2.0-only
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add SPDX license identifiers to all files which:
- Have no license information of any form
- Have MODULE_LICENCE("GPL*") inside which was used in the initial
scan/conversion to ignore the file
These files fall under the project license, GPL v2 only. The resulting SPDX
license identifier is:
GPL-2.0-only
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add SPDX license identifiers to all files which:
- Have no license information of any form
- Have EXPORT_.*_SYMBOL_GPL inside which was used in the
initial scan/conversion to ignore the file
These files fall under the project license, GPL v2 only. The resulting SPDX
license identifier is:
GPL-2.0-only
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Alexander Lobakin
Trent Piepho
David S. Miller
Andy Shevchenko
Alexander A. Klimov
Oleg Nesterov
Linus Torvalds
Christophe JAILLET
Masami Hiramatsu
Thomas Gleixner