Em Tue, 19 May 2015 08:05:56 -0400 (EDT) Federico Simoncelli <fsimonce@xxxxxxxxxx> escreveu: > ----- Original Message ----- > > From: "Mauro Carvalho Chehab" <mchehab@xxxxxxxxxxxxxxx> > > To: "Linux Media Mailing List" <linux-media@xxxxxxxxxxxxxxx> > > Cc: "Mauro Carvalho Chehab" <mchehab@xxxxxxxxxxxxxxx>, "Mauro Carvalho Chehab" <mchehab@xxxxxxxxxxxxx>, "Lars-Peter > > Clausen" <lars@xxxxxxxxxx>, "Michael Buesch" <m@xxxxxxx>, "Antti Palosaari" <crope@xxxxxx>, "Hans Verkuil" > > <hverkuil@xxxxxxxxx>, "Sakari Ailus" <sakari.ailus@xxxxxxxxxxxxxxx>, "Ondrej Zary" <linux@xxxxxxxxxxxxxxxxxxxx>, > > "Ramakrishnan Muthukrishnan" <ramakrmu@xxxxxxxxx>, "Laurent Pinchart" <laurent.pinchart@xxxxxxxxxxxxxxxx>, "Takashi > > Iwai" <tiwai@xxxxxxx>, "Amber Thrall" <amber.rose.thrall@xxxxxxxxx>, "Federico Simoncelli" <fsimonce@xxxxxxxxxx>, > > "James Harper" <james.harper@xxxxxxxxxxxxxxxxx>, "Dan Carpenter" <dan.carpenter@xxxxxxxxxx>, "Konrad Rzeszutek Wilk" > > <konrad.wilk@xxxxxxxxxx> > > Sent: Tuesday, May 19, 2015 1:00:57 PM > > Subject: [PATCH 2/2] drivers: Simplify the return code > > > > If the last thing we do in a function is to call another > > function and then return its value, we don't need to store > > the returned code into some ancillary var. > > > > Signed-off-by: Mauro Carvalho Chehab <mchehab@xxxxxxxxxxxxxxx> > > > > diff --git a/drivers/media/dvb-frontends/lgs8gxx.c > > b/drivers/media/dvb-frontends/lgs8gxx.c > > index 3c92f36ea5c7..9b0166cdc7c2 100644 > > --- a/drivers/media/dvb-frontends/lgs8gxx.c > > +++ b/drivers/media/dvb-frontends/lgs8gxx.c > > @@ -544,11 +544,7 @@ static int lgs8gxx_set_mpeg_mode(struct lgs8gxx_state > > *priv, > > t |= clk_pol ? TS_CLK_INVERTED : TS_CLK_NORMAL; > > t |= clk_gated ? TS_CLK_GATED : TS_CLK_FREERUN; > > > > - ret = lgs8gxx_write_reg(priv, reg_addr, t); > > - if (ret != 0) > > - return ret; > > - > > - return 0; > > + return lgs8gxx_write_reg(priv, reg_addr, t); > > } > > Personally I prefer the current style because it's more consistent with all > the other calls in the same function (return ret when ret != 0). > > It also allows you to easily add/remove calls without having to deal with > the last special case return my_last_fun_call(...). > > Anyway it's not a big deal, I think it's your call. First of all, I wrote this patch to test coccinelle (with was broken in Fedora 21 due to a dependency issue[1]). Once done, why not submit? So, it is not that I'm really wanting to do that. [1] https://bugzilla.redhat.com/show_bug.cgi?id=1218987 Yet, as someone that review a lot of code, the above is a way better, as I can instantly check that the data returned by lgs8gxx_write_reg() will also be returned to the caller function. Ok, this optimizes a few milliseconds of my brain's processing, but it still means a faster review to me, with is a gain, as I spend big periods of time reviewing code. So, the simpler, the merrier. Also, I don't trust too much on compiler optimization. I mean: I don't expect that gcc will always do the right thing and remove the useless if. Writing software for so long time, I know that software has bugs. So, if I see something that can be easily optimized like the above code, I do. It is one less thing to trust that the compiler will do right. Regards, Mauro -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
