Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Wed, 4 Jul 2018 15:36:17 +0000
From:   Dmitry Torokhov <dmitry.torokhov@gmail.com>
To:     Yury Norov <ynorov@caviumnetworks.com>
Cc:     Alexander Shishkin <alexander.shishkin@linux.intel.com>,
        Peter Zijlstra <peterz@infradead.org>,
        Ingo Molnar <mingo@redhat.com>,
        Arnaldo Carvalho de Melo <acme@kernel.org>,
        Jiri Olsa <jolsa@redhat.com>,
        Namhyung Kim <namhyung@kernel.org>,
        Kate Stewart <kstewart@linuxfoundation.org>,
        Matthew Wilcox <willy@infradead.org>,
        Philippe Ombredanne <pombredanne@nexb.com>,
        David Ahern <dsahern@gmail.com>,
        David Carrillo-Cisneros <davidcc@google.com>,
        Andi Kleen <ak@linux.intel.com>,
        Jin Yao <yao.jin@linux.intel.com>,
        linux-kernel@vger.kernel.org,
        Andy Shevchenko <andriy.shevchenko@linux.intel.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        Mike Snitzer <snitzer@redhat.com>
Subject: Re: [PATCH 2/2] bitmap: sync tools with new bitmap allocation API
Message-ID: <20180704153617.tps2asygskoym46d@penguin>
References: <20180623073502.16321-1-ynorov@caviumnetworks.com>
 <20180623073502.16321-2-ynorov@caviumnetworks.com>
 <20180624213103.GA166241@dtor-ws>
 <20180624224522.GA23998@yury-thinkpad>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20180624224522.GA23998@yury-thinkpad>
User-Agent: NeoMutt/20170113 (1.7.2)
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Mon, Jun 25, 2018 at 01:45:22AM +0300, Yury Norov wrote:
> On Sun, Jun 24, 2018 at 02:31:03PM -0700, Dmitry Torokhov wrote:
> > External Email
> > 
> > On Sat, Jun 23, 2018 at 10:35:02AM +0300, Yury Norov wrote:
> > > On top of next-20180622 and Andy Shevchenko series:
> > > https://lkml.org/lkml/2018/6/18/841
> > >
> > > The series mentioned above introduces helpers for bitmap allocation.
> > > tools/ has its own bitmap_alloc() which differs from bitmap_alloc()
> > > proposed in new kernel API, and is equivalent to bitmap_zalloc().
> > > In this series tools is switched to new API.
> > >
> > > This is RFC because I didn't find counterpart free() call to some
> > > bitmap_zalloc()'s. So I didn't convert them to bitmap_free(). Could
> > > someone point me out? The functions are:
> > > setup_nodes();
> > > do_read_bitmap(); // Free is called, but only in fail path.
> > 
> > Yes, because if we succeed we effectively return allocated bitmap to the
> > caller. You'd need to trace upwards and see how it all gets cleaned up.
> > But given that this is userspace and is not expected to be long-lived,
> > maybe nobody bothered freeing memory and we instead rely on the kernel
> > to clean it all up when process terminates.
> > 
> > Thanks.
> > 
> > > memory_node__read();
> > >
> > > Signed-off-by: Yury Norov <ynorov@caviumnetworks.com>
> > > ---
> > >  tools/include/linux/bitmap.h | 19 +++++++++++++++----
> > >  tools/perf/builtin-c2c.c     | 10 +++++-----
> > >  tools/perf/tests/bitmap.c    |  4 ++--
> > >  tools/perf/tests/mem2node.c  |  4 ++--
> > >  tools/perf/util/header.c     |  6 +++---
> > >  5 files changed, 27 insertions(+), 16 deletions(-)
> > >
> > > diff --git a/tools/include/linux/bitmap.h b/tools/include/linux/bitmap.h
> > > index 48c208437bbd..b9b85b94c937 100644
> > > --- a/tools/include/linux/bitmap.h
> > > +++ b/tools/include/linux/bitmap.h
> > > @@ -98,12 +98,23 @@ static inline int test_and_set_bit(int nr, unsigned long *addr)
> > >  }
> > >
> > >  /**
> > > - * bitmap_alloc - Allocate bitmap
> > > - * @nbits: Number of bits
> > > + * Allocation and deallocation of bitmap.
> > >   */
> > > -static inline unsigned long *bitmap_alloc(int nbits)
> > > +static inline unsigned long *bitmap_alloc(unsigned int nbits, gfp_t flags)
> > 
> > This makes absolutely no sense for userspace API. What gfp_t even means
> > here?
> > 
> > If you want to introduce bitmap_zalloc and bitmap_free it is fine but
> > adding dummy parameters to match kernel API exactly is a folly.
> 
> Identical API makes easier porting the code from kernel to tools.
> Refer for example declaration of kmalloc in:
> tools/testing/radix-tree/linux.c
> tools/testing/scatterlist/linux/mm.h
> tools/virtio/linux/kernel.h
> tools/virtio/ringtest/ptr_ring.c

These are unittests for the APIs in question, of course they would have
to match exactly.

perf tool however is not a unittest, so there is no need to match kernel
API.

Thanks.

-- 
Dmitry