mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Andy Shevchenko <andriy.shevchenko@intel.com>
To: rodrigo.alencar@analog.com
Cc: linux-kernel@vger.kernel.org, linux-iio@vger.kernel.org,
	Andrew Morton <akpm@linux-foundation.org>,
	Jonathan Cameron <jic23@kernel.org>,
	David Lechner <dlechner@baylibre.com>,
	Andy Shevchenko <andy@kernel.org>
Subject: Re: [PATCH v4 4/6] iio: test: rescale: compare scales against the precision of the output
Date: Mon, 26 Oct 2026 16:40:29 +0200	[thread overview]
Message-ID: <at9mXdRhiM-9TVyg@ashevche-desk.local> (raw)
In-Reply-To: <20261026-e-notation-v4-4-a366a6d14d57@analog.com>

On Mon, Oct 26, 2026 at 08:54:08AM -0300, Rodrigo Alencar via B4 Relay wrote:
> 
> The scale tests format the rescaled value with iio_format_value(), which
> prints at most nano precision, and require the relative error to the
> expected value to be below 1 ppm. Both strings are parsed back with
> iio_str_to_fixpoint() at nano precision, so for the small scale cases the
> check only passes because the expected value, like the formatted one, is
> truncated to nano. For example, "0.000000033594" is parsed as 33 nano,
> equal to the formatted "0.000000033", although the actual error is about
> 1.8%. If the expected value were rounded instead, it would be parsed as
> 34 nano and the test would fail even though the output is as accurate as
> the format allows.
> 
> Parse both values with kstrtodec64() at a scale of 10, one digit more
> than the output precision, so the expected value keeps the digits lost
> when formatting. Scale 10 is the largest that still fits the overflow
> cases (about 2.1e8) in an s64. Compute the relative error with
> mul_u64_u64_div_u64(), as the product with 10^6 no longer fits in 64 bits
> for those cases.
> 
> Then allow a relative error up to the relative size of one nano of the
> expected value, the precision of the formatted output. This is 0 ppm for
> values of 0.01 and above, so those cases are checked as before, and it
> accepts the formatting error for the small scale cases.

...

> +/*
> + * iio_format_value() prints scales with up to nano precision. Parse values
> + * with one more digit, so the expected values keep the precision lost there.
> + * Scale 10 is the largest that fits the overflow test cases in an s64.
>   */
> +#define IIO_TEST_DEC_SCALE	10
> +#define IIO_TEST_NANO		10	/* one nano at IIO_TEST_DEC_SCALE */

I would put them as (9 + 1) to show that actually show that "with one more digit".
I also don't follow the comment near to the last definition. What does it mean?

-- 
With Best Regards,
Andy Shevchenko



  reply	other threads:[~2026-10-26 14:40 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-10-26 11:54 [PATCH v4 0/6] Add E notation support to kstrtodec64() and IIO sysfs writes Rodrigo Alencar via B4 Relay
2026-10-26 11:54 ` [PATCH v4 1/6] lib: kstrtox: support E notation in kstrtoudec64() and kstrtodec64() Rodrigo Alencar via B4 Relay
2026-10-27 10:23   ` Rodrigo Alencar
2026-10-26 11:54 ` [PATCH v4 2/6] lib: kstrtox: round to nearest, ties to even in kstrto[u]dec64() Rodrigo Alencar via B4 Relay
2026-10-26 11:54 ` [PATCH v4 3/6] lib: test-kstrtox: add E notation tests for kstrto[u]dec64() Rodrigo Alencar via B4 Relay
2026-10-26 11:54 ` [PATCH v4 4/6] iio: test: rescale: compare scales against the precision of the output Rodrigo Alencar via B4 Relay
2026-10-26 14:40   ` Andy Shevchenko [this message]
2026-10-26 15:29     ` Rodrigo Alencar
2026-10-26 11:54 ` [PATCH v4 5/6] iio: core: use kstrtodec64() to parse fixed-point values Rodrigo Alencar via B4 Relay
2026-10-26 14:54   ` Andy Shevchenko
2026-10-26 15:53     ` Rodrigo Alencar
2026-10-26 11:54 ` [PATCH v4 6/6] iio: test: add kunit tests for channel info writes Rodrigo Alencar via B4 Relay

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=at9mXdRhiM-9TVyg@ashevche-desk.local \
    --to=andriy.shevchenko@intel.com \
    --cc=akpm@linux-foundation.org \
    --cc=andy@kernel.org \
    --cc=dlechner@baylibre.com \
    --cc=jic23@kernel.org \
    --cc=linux-iio@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=rodrigo.alencar@analog.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®