Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.33 as permitted sender) client-ip=23.128.96.33;
MIME-Version: 1.0
References: <20231023102327.3074212-1-glider@google.com> <20231023102327.3074212-2-glider@google.com>
 <ZTbcXEe74q6jG4XZ@yury-ThinkPad>
In-Reply-To: <ZTbcXEe74q6jG4XZ@yury-ThinkPad>
From:   Alexander Potapenko <glider@google.com>
Date:   Tue, 24 Oct 2023 13:56:18 +0200
Message-ID: <CAG_fn=XZdT7k4fr7G5+-+DC9Oon-3y2PenyZomSiObbc4M=GQw@mail.gmail.com>
Subject: Re: [PATCH v8 2/2] lib/test_bitmap: add tests for bitmap_{read,write}()
To:     Yury Norov <yury.norov@gmail.com>
Cc:     catalin.marinas@arm.com, will@kernel.org, pcc@google.com,
        andreyknvl@gmail.com, andriy.shevchenko@linux.intel.com,
        aleksander.lobakin@intel.com, linux@rasmusvillemoes.dk,
        alexandru.elisei@arm.com, linux-kernel@vger.kernel.org,
        linux-arm-kernel@lists.infradead.org, eugenis@google.com,
        syednwaris@gmail.com, william.gray@linaro.org
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

> > +
> > +static void __init test_bitmap_write_perf(void)
> > +{
> > +     DECLARE_BITMAP(bitmap, TEST_BIT_LEN);
> > +     unsigned int cnt, nbits, i;
> > +     unsigned long val = 0xfeedface;
> > +     ktime_t time;
> > +
> > +     bitmap_zero(bitmap, TEST_BIT_LEN);
> > +     time = ktime_get();
> > +     for (cnt = 0; cnt < 5; cnt++) {
> > +             for (nbits = 1; nbits <= BITS_PER_LONG; nbits++) {
> > +                     for (i = 0; i < TEST_BIT_LEN; i++) {
> > +                             if (i + nbits > TEST_BIT_LEN)
> > +                                     break;
> > +                             bitmap_write(bitmap, val, i, nbits);
> > +                     }
> > +             }
> > +     }
> > +     time = ktime_get() - time;
> > +     pr_err("Time spent in %s:\t%llu\n", __func__, time);
>
> For the perf part, can you add the output example to the commit
> message, and compare numbers with non-optimized for-loop()?
>

I don't understand the purpose of this comparison.
It is evident that bitmap_write() is faster than the non-optimized
loop doing BITS_PER_LONG reads and writes of a single bit.
It is moot how much faster the current implementation is, because the
loop implementation is just a concept describing the behavior of
bitmap_write().

My understanding was that the performance tests will help if someone
decides to optimize bitmap_write() further - in that case they would
be able to compare the performance before and after their changes.
But I fail to see how it helps to compare the current implementation
to something that is a priori slower.