[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <1F3AC3675D538145B1661F571FE1805F19A79E7E@irsmsx105.ger.corp.intel.com>
Date: Fri, 5 Jun 2015 17:00:05 +0000
From: "Tirdea, Irina" <irina.tirdea@...el.com>
To: Dmitry Torokhov <dmitry.torokhov@...il.com>
CC: 'Antonio Ospite' <ao2@....it>, Bastien Nocera <hadess@...ess.net>,
"linux-input@...r.kernel.org" <linux-input@...r.kernel.org>,
"devicetree@...r.kernel.org" <devicetree@...r.kernel.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: RE: [PATCH 2/9] input: goodix: fix variable length array warning
> -----Original Message-----
> From: Dmitry Torokhov [mailto:dmitry.torokhov@...il.com]
> Sent: 05 June, 2015 19:41
> To: Tirdea, Irina
> Cc: 'Antonio Ospite'; Bastien Nocera; linux-input@...r.kernel.org; devicetree@...r.kernel.org; linux-kernel@...r.kernel.org
> Subject: Re: [PATCH 2/9] input: goodix: fix variable length array warning
>
> On Fri, Jun 05, 2015 at 04:34:38PM +0000, Tirdea, Irina wrote:
> >
> >
> > > -----Original Message-----
> > > From: linux-input-owner@...r.kernel.org [mailto:linux-input-owner@...r.kernel.org] On Behalf Of Antonio Ospite
> > > Sent: 03 June, 2015 23:50
> > > To: Tirdea, Irina
> > > Cc: Dmitry Torokhov; Bastien Nocera; linux-input@...r.kernel.org; devicetree@...r.kernel.org; linux-kernel@...r.kernel.org
> > > Subject: Re: [PATCH 2/9] input: goodix: fix variable length array warning
> > >
> > > On Wed, 3 Jun 2015 10:26:47 +0000
> > > "Tirdea, Irina" <irina.tirdea@...el.com> wrote:
> > >
> > > > > -----Original Message-----
> > > > > From: Antonio Ospite [mailto:ao2@....it]
> > > > > Sent: 28 May, 2015 18:58
> > > > > To: Tirdea, Irina
> > > > > Cc: Dmitry Torokhov; Bastien Nocera; linux-input@...r.kernel.org; devicetree@...r.kernel.org; linux-kernel@...r.kernel.org
> > > > > Subject: Re: [PATCH 2/9] input: goodix: fix variable length array warning
> > > > >
> > > > > On Thu, 28 May 2015 15:47:38 +0300
> > > > > Irina Tirdea <irina.tirdea@...el.com> wrote:
> > > > >
> > > > > > Fix sparse warning:
> > > > > > drivers/input/touchscreen/goodix.c:182:26: warning:
> > > > > > Variable length array is used.
> > > > > >
> > > > > > Replace the variable length array with fixed length.
> > > > > >
> > > > > > Signed-off-by: Irina Tirdea <irina.tirdea@...el.com>
> > > > > > ---
> > > > > > drivers/input/touchscreen/goodix.c | 2 +-
> > > > > > 1 file changed, 1 insertion(+), 1 deletion(-)
> > > > > >
> > > > > > diff --git a/drivers/input/touchscreen/goodix.c b/drivers/input/touchscreen/goodix.c
> > > > > > index c2e785c..dac1b3c 100644
> > > > > > --- a/drivers/input/touchscreen/goodix.c
> > > > > > +++ b/drivers/input/touchscreen/goodix.c
> > > > > > @@ -147,7 +147,7 @@ static void goodix_ts_report_touch(struct goodix_ts_data *ts, u8 *coor_data)
> > > > > > */
> > > > > > static void goodix_process_events(struct goodix_ts_data *ts)
> > > > > > {
> > > > > > - u8 point_data[1 + GOODIX_CONTACT_SIZE * ts->max_touch_num];
> > > > > > + u8 point_data[1 + GOODIX_CONTACT_SIZE * GOODIX_MAX_CONTACTS];
> > > > >
> > > > > Hi,
> > > > >
> > > >
> > > > Hi Antonio,
> > > >
> > > > > this fixes the warning from sparse, but also changes the semantics of
> > > > > the code: ts->max_touch_num is less that GOODIX_MAX_CONTACTS for 5
> > > > > touches devices and in this case we'll end up using more memory than is
> > > > > necessary.
> > > > >
> > > >
> > > > I wasn't sure if it is better to save the 5 bytes or fix the warning,
> > > > so I sent this to get some more input.
> > > > Thanks for the feedback, I will drop this patch.
> > > >
> > >
> > > Use kmalloc() or, alternatively, add at least a comment telling why you
> > > think that sacrificing a few bytes —only for some devices— has
> > > advantages over dynamic allocation.
> > >
> >
> > You are right, kmalloc will solve both problems - the sparse warning and allocating
> > more bytes than necessary. Don't know why I did not think of that.
> > Will use that in v2.
>
> Please leave the patch as is. We can spare 80 bytes on the stack given
> that we are running in threaded IRQ. kmallocing will use more code and
> runtime resources and won't give any benefits.
I was actually thinking of allocating it with devm_kzalloc just once at device init and
storing a pointer to it in the goodix_ts_data structure.
Thanks,
Irina
>
> Thanks.
>
> --
> Dmitry
Powered by blists - more mailing lists