From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jakub =?UTF-8?B?S2ljacWEc2tp?= Subject: Re: [PATCH 2/2] spi: add statistics gathering and reporting methods Date: Tue, 5 May 2015 17:48:31 +0200 Message-ID: <20150505174831.5d9acedb@north> References: <1430741564-2849-1-git-send-email-kernel@martin.sperl.org> <1430741564-2849-2-git-send-email-kernel@martin.sperl.org> <20150504134226.GY15510@sirena.org.uk> <9A0ADDBC-5D5A-428F-B434-9CCB97505DCB@martin.sperl.org> <20150505142008.2087baac@north> <20150505150246.GG22845@sirena.org.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Martin Sperl , linux-spi-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Mark Brown Return-path: In-Reply-To: <20150505150246.GG22845-GFdadSzt00ze9xe1eoZjHA@public.gmane.org> Sender: linux-spi-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-ID: On Tue, 5 May 2015 16:02:46 +0100, Mark Brown wrote: > On Tue, May 05, 2015 at 02:20:08PM +0200, Jakub Kici=C5=84ski wrote: > > On Mon, 4 May 2015 16:25:01 +0200, Martin Sperl wrote: >=20 > > > that makes it much more complicated to add things per spi_master > > > implementation (which was my original idea: have a framework that= exposes > > > the distinct transfer modes of the spi-bcm2835 driver (polling, i= nterrupt > > > and dma driven) and also to get some stats on how often we run in= to those > > > =E2=80=9Ccorner=E2=80=9D cases - that is also why I want to gathe= r the histogram > > > statistics, as it may make it easier to understand the use cases = of people > > > who complain about =E2=80=9Cperformance=E2=80=9D in case they sha= re those infos=E2=80=A6 >=20 > > You can also consider adding few strategically placed tracepoints. = =20 > > It saves the amount of kernel code necessary and is generally more > > flexible. >=20 > We already have a bunch of tracepoints in there (probably sufficientl= y > detailed for what's here) but they solve a different problem to > statistics - while you can extract statistics at least as far back as > the tracepoints go but it's slightly cumbersome and for longer term > tests the buffer will roll round at some point and it's a bit more wo= rk > for people to set them up. Sure, I definitely agree with complicated setup especially that Martin is targeting this at RPi users who are usually not the most seasoned of kernel developers. Perhaps one day someone will write a generic tracepoint-like statistic subsystem which can be disabled at runtime and easily compiled-out... -- To unsubscribe from this list: send the line "unsubscribe linux-spi" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html