Return-Path: MIME-Version: 1.0 In-Reply-To: <1300888771-26437-1-git-send-email-par-gunnar.p.hjalmdahl@stericsson.com> References: <1300888771-26437-1-git-send-email-par-gunnar.p.hjalmdahl@stericsson.com> Date: Wed, 23 Mar 2011 15:53:22 +0100 Message-ID: Subject: Re: [PATCH 1/2] staging: Add ST-Ericsson CG2900 driver From: Vitaly Wool To: Par-Gunnar Hjalmdahl Cc: Greg Kroah-Hartman , devel@driverdev.osuosl.org, Linus Walleij , linux-kernel@vger.kernel.org, linux-bluetooth@vger.kernel.org, Pavan Savoy , Alan Cox , Arnd Bergmann , Marcel Holtmann , Lukasz Rymanowski , Linus Walleij , Par-Gunnar Hjalmdahl , Lee Jones Content-Type: text/plain; charset=ISO-8859-1 List-ID: On Wed, Mar 23, 2011 at 2:59 PM, Par-Gunnar Hjalmdahl wrote: > diff --git a/drivers/staging/cg2900/TODO b/drivers/staging/cg2900/TODO > new file mode 100644 > index 0000000..f6fb76a > --- /dev/null > +++ b/drivers/staging/cg2900/TODO > @@ -0,0 +1,21 @@ > +TODO > +---- > + > + - Decide upon architecture. Some people consider architecture in the cg= 2900 > + =A0 driver to be too complex. We consider it to be not more complex tha= n needed. > + > + - Currently the cg2900_uart registers as protocol driver against hci_ld= isc.c. > + =A0 There is however some common functionality with hci_h4.c and the cg= 2900 could > + =A0 therefore register it's vendor specific channels to hci_h4.c, but t= his would > + =A0 require adding a registration functionality in the hci_h4 file. > + > + - Some people demand that the cg2900 driver re-use the Bluetooth driver= to send > + =A0 and receive BT commands and events. That is however not possible wi= th current > + =A0 BT API and might not be feasible, for example when using FM only in > + =A0 the cg2900 chip. > + > + - TI has already delivered a driver for a multi-function chip called ti= -st. > + =A0 This driver is currently located in drivers/misc/ti-st/. There has = however > + =A0 been criticism raised against design/architecture of the driver. Th= ere > + =A0 currently also doesn't seem to be a way to add support for cg2900 i= n that > + =A0 driver even though some people has raised this as an alternative. As someone who is definitely the part of these demotivating and unreasonable "some people" I have to say that even though I'm okay with this driver going into staging, I see a huge gap between the desire of its creators to have something generic and the ability to do so. I am absolutely sure that given the widely advertised peculiarities of this chip, the authors should stop trying to come up with a generic solution and just present a simplistic MFD driver for their particular device. Thanks, Vitaly