From mboxrd@z Thu Jan 1 00:00:00 1970 From: Henrik Rydberg Subject: Re: [PATCH] Input: bcm5974.c initialize raw_w, raw_x and raw_y before it get used Date: Sun, 13 Sep 2009 13:39:21 +0200 Message-ID: <4AACD9E9.4000702@bitmath.org> References: <1252775770.3687.7.camel@ht.satnam> <4AAC2DBE.9040303@bitmath.org> <1252830690.3440.5.camel@ht.satnam> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: Received: from csmtp2.one.com ([195.47.247.206]:58119 "EHLO csmtp2.one.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751627AbZIMLj0 (ORCPT ); Sun, 13 Sep 2009 07:39:26 -0400 In-Reply-To: <1252830690.3440.5.camel@ht.satnam> Sender: linux-input-owner@vger.kernel.org List-Id: linux-input@vger.kernel.org To: Jaswinder Singh Rajput Cc: Dmitry Torokhov , linux-input@vger.kernel.org Jaswinder Singh Rajput wrote: > On Sun, 2009-09-13 at 01:24 +0200, Henrik Rydberg wrote: >> Jaswinder Singh Rajput wrote: >>> raw_w, raw_x and raw_y is used uninitialized for !raw_n >> Thanks for the heads up, but actually not, since !raw_n also implies >> !(ptest > PRESSURE_LOW). >> >=20 > Then can we move 'if (ptest > PRESSURE_LOW && origin)' stuff to 'if > (raw_n)'. If not then my patch is correct. Yes, that's it, thanks. So this patch ought to solve the warning cleanl= y: diff --git a/drivers/input/mouse/bcm5974.c b/drivers/input/mouse/bcm597= 4.c index 2d8fc0b..3b598de 100644 --- a/drivers/input/mouse/bcm5974.c +++ b/drivers/input/mouse/bcm5974.c @@ -345,21 +345,22 @@ static int report_tp_state(struct bcm5974 *dev, i= nt size) /* set the integrated button if applicable */ if (c->tp_type =3D=3D TYPE2) ibt =3D raw2int(dev->tp_data[BUTTON_TYPE2]); - } - /* while tracking finger still valid, count all fingers */ - if (ptest > PRESSURE_LOW && origin) { - abs_p =3D ptest; - abs_w =3D int2bound(&c->w, raw_w); - abs_x =3D int2bound(&c->x, raw_x - c->x.devmin); - abs_y =3D int2bound(&c->y, c->y.devmax - raw_y); - while (raw_n--) { - ptest =3D int2bound(&c->p, raw2int(f->force_major)); - if (ptest > PRESSURE_LOW) - nmax++; - if (ptest > PRESSURE_HIGH) - nmin++; - f++; + /* while tracking finger still valid, count all fingers */ + if (ptest > PRESSURE_LOW && origin) { + abs_p =3D ptest; + abs_w =3D int2bound(&c->w, raw_w); + abs_x =3D int2bound(&c->x, raw_x - c->x.devmin); + abs_y =3D int2bound(&c->y, c->y.devmax - raw_y); + while (raw_n--) { + ptest =3D int2bound(&c->p, + raw2int(f->force_major)); + if (ptest > PRESSURE_LOW) + nmax++; + if (ptest > PRESSURE_HIGH) + nmin++; + f++; + } } } >=20 >>> This also fixed these compilation warnings : >>> >>> CC [M] drivers/input/mouse/bcm5974.o >>> drivers/input/mouse/bcm5974.c: In function =E2=80=98report_tp_state= =E2=80=99: >>> drivers/input/mouse/bcm5974.c:319: warning: =E2=80=98raw_y=E2=80=99= may be used uninitialized in this function >>> drivers/input/mouse/bcm5974.c:319: warning: =E2=80=98raw_x=E2=80=99= may be used uninitialized in this function >>> drivers/input/mouse/bcm5974.c:319: warning: =E2=80=98raw_w=E2=80=99= may be used uninitialized in this function >>> >>> Signed-off-by: Jaswinder Singh Rajput >>> --- >>> drivers/input/mouse/bcm5974.c | 3 ++- >>> 1 files changed, 2 insertions(+), 1 deletions(-) >>> >>> diff --git a/drivers/input/mouse/bcm5974.c b/drivers/input/mouse/bc= m5974.c >>> index 2d8fc0b..171f345 100644 >>> --- a/drivers/input/mouse/bcm5974.c >>> +++ b/drivers/input/mouse/bcm5974.c >>> @@ -345,7 +345,8 @@ static int report_tp_state(struct bcm5974 *dev,= int size) >>> /* set the integrated button if applicable */ >>> if (c->tp_type =3D=3D TYPE2) >>> ibt =3D raw2int(dev->tp_data[BUTTON_TYPE2]); >>> - } >>> + } else >>> + raw_w =3D raw_x =3D raw_y =3D 0; >>> =20 >>> /* while tracking finger still valid, count all fingers */ >>> if (ptest > PRESSURE_LOW && origin) { >> I would prefer treating raw_p on the same footing here, completing t= he set of >> non-obviously initialized variables. It might also make sense to uti= lize the >> same initialization technique already used in the code, thus: >> >=20 > No, then you are wasting cpu cycles and doing double initialization f= or > some cases. >=20 >> diff --git a/drivers/input/mouse/bcm5974.c b/drivers/input/mouse/bcm= 5974.c >> index 2d8fc0b..2f85876 100644 >> --- a/drivers/input/mouse/bcm5974.c >> +++ b/drivers/input/mouse/bcm5974.c >> @@ -316,7 +316,7 @@ static int report_tp_state(struct bcm5974 *dev, = int size) >> const struct bcm5974_config *c =3D &dev->cfg; >> const struct tp_finger *f; >> struct input_dev *input =3D dev->input; >> - int raw_p, raw_w, raw_x, raw_y, raw_n; >> + int raw_p =3D 0, raw_w =3D 0, raw_x =3D 0, raw_y =3D 0, raw_n; >> int ptest =3D 0, origin =3D 0, ibt =3D 0, nmin =3D 0, nmax =3D 0; >> int abs_p =3D 0, abs_w =3D 0, abs_x =3D 0, abs_y =3D 0; >> >> >> I wonder how many cpu cycles in the world are spent making compilers= happy. >> >=20 > It is not compiler mistake it is programming/logic mistakes. We shoul= d > be thankful to compiler to pointing mistakes made by us. The uninitialized-variable warnings have certainly proved their value m= any times over. I was merely reflecting over the fact that the enclosing space of= all such errors could be made smaller by deduction at compile time. Cheers, Henrik -- To unsubscribe from this list: send the line "unsubscribe linux-input" = in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html