From mboxrd@z Thu Jan 1 00:00:00 1970 From: Peter Hurley Subject: Re: [PATCH 2/4] of: DT quirks infrastructure Date: Fri, 20 Feb 2015 10:00:06 -0500 Message-ID: <54E74BF6.208@hurleysoftware.com> References: <1424271576-1952-3-git-send-email-pantelis.antoniou@konsulko.com> <20150218154106.GC29429@leverpostej> <20150218173115.GG29429@leverpostej> <76BD1B22-BAED-4205-9B34-186907CE0217@konsulko.com> <54E613E7.2020405@gmail.com> <670D0881-DBF0-45E8-A502-A6DB2B77A750@konsulko.com> <54E61DD2.3060002@gmail.com> <53F2F94C-0C43-4A54-B8CD-EEC454A0AC19@konsulko.com> <54E742F2.80506@hurleysoftware.com> <20150220143533.GA29908@odux.rfo.atmel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: In-Reply-To: <20150220143533.GA29908-FuRPzXQv2LUWBfJKYY8PcdBPR1lH4CV8@public.gmane.org> Sender: devicetree-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Pantelis Antoniou , frowand.list-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org, Mark Rutland , "devicetree-u79uwXL29TY76Z2rM5mHXA@public.gmane.org" , Tony Lindgren , Koen Kooi , Nicolas Ferre , "linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org" , Grant Likely , "linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org" , Matt Porter , Guenter Roeck List-Id: devicetree@vger.kernel.org On 02/20/2015 09:35 AM, Ludovic Desroches wrote: > On Fri, Feb 20, 2015 at 09:21:38AM -0500, Peter Hurley wrote: >> On 02/19/2015 12:38 PM, Pantelis Antoniou wrote: >>> >>>> On Feb 19, 2015, at 19:30 , Frank Rowand = wrote: >>>> >>>> On 2/19/2015 9:00 AM, Pantelis Antoniou wrote: >>>>> Hi Frank, >>>>> >>>>>> On Feb 19, 2015, at 18:48 , Frank Rowand wrote: >>>>>> >>>>>> On 2/19/2015 6:29 AM, Pantelis Antoniou wrote: >>>>>>> Hi Mark, >>>>>>> >>>>>>>> On Feb 18, 2015, at 19:31 , Mark Rutland wrote: >>>>>>>> >>>>>>>>>>> +While this may in theory work, in practice it is very cumb= ersome >>>>>>>>>>> +for the following reasons: >>>>>>>>>>> + >>>>>>>>>>> +1. The act of selecting a different boot device tree blob = requires >>>>>>>>>>> +a reasonably advanced bootloader with some kind of configu= ration or >>>>>>>>>>> +scripting capabilities. Sadly this is not the case many ti= mes, the >>>>>>>>>>> +bootloader is extremely dumb and can only use a single dt = blob. >>>>>>>>>> >>>>>>>>>> You can have several bootloader builds, or even a single bui= ld with >>>>>>>>>> something like appended DTB to get an appropriate DTB if the= same binary >>>>>>>>>> will otherwise work across all variants of a board. >>>>>>>>>> >>>>>>>>> >>>>>>>>> No, the same DTB will not work across all the variants of a b= oard. >>>>>>>> >>>>>>>> I wasn't on about the DTB. I was on about the loader binary, i= n the case >>>>>>>> the FW/bootloader could be common even if the DTB couldn't. >>>>>>>> >>>>>>>> To some extent there must be a DTB that will work across all v= ariants >>>>>>>> (albeit with limited utility) or the quirk approach wouldn't w= ork=E2=80=A6 >>>>>>>> >>>>>>> >>>>>>> That=E2=80=99s not correct; the only part of the DTB that needs= to be common >>>>>>> is the model property that would allow the quirk detection logi= c to fire. >>>>>>> >>>>>>> So, there is a base DTB that will work on all variants, but tha= t only means >>>>>>> that it will work only up to the point that the quirk detector = method >>>>>>> can work. So while in recommended practice there are common sub= sets >>>>>>> of the DTB that might work, they might be unsafe. >>>>>>> >>>>>>> For instance on the beaglebone the regulator configuration is d= ifferent >>>>>>> between white and black, it is imperative you get them right ot= herwise >>>>>>> you risk board damage. >>>>>>> >>>>>>>>>> So it's not necessarily true that you need a complex bootloa= der. >>>>>>>>>> >>>>>>>>> >>>>>>>>>>> +2. On many instances boot time is extremely critical; in s= ome cases >>>>>>>>>>> +there are hard requirements like having working video feed= s in under >>>>>>>>>>> +2 seconds from power-up. This leaves an extremely small ti= me budget for >>>>>>>>>>> +boot-up, as low as 500ms to kernel entry. The sanest way t= o get there >>>>>>>>>>> +is by removing the standard bootloader from the normal boo= t sequence >>>>>>>>>>> +altogether by having a very small boot shim that loads the= kernel and >>>>>>>>>>> +immediately jumps to kernel, like falcon-boot mode in u-bo= ot does. >>>>>>>>>> >>>>>>>>>> Given my previous comments above I don't see why this is rel= evant. >>>>>>>>>> You're already passing _some_ DTB here, so if you can organi= se for the >>>>>>>>>> board to statically provide a sane DTB that's fine, or you c= an resort to >>>>>>>>>> appended DTB if it's not possible to update the board config= uration. >>>>>>>>>> >>>>>>>>> >>>>>>>>> You=E2=80=99re missing the point. I can=E2=80=99t use the sam= e DTB for each revision of the >>>>>>>>> board. Each board is similar but it=E2=80=99s not identical. >>>>>>>> >>>>>>>> I think you've misunderstood my point. If you program the boar= d with the >>>>>>>> relevant DTB, or use appended DTB, then you will pass the corr= ect DTB to >>>>>>>> the kernel without need for quirks. >>>>>>>> >>>>>>>> I understand that each variant is somewhat incompatible (and h= ence needs >>>>>>>> its own DTB). >>>>>>> >>>>>>> In theory it might work, in practice this does not. Ludovic men= tioned that they >>>>>>> have 27 different DTBs in use at the moment. At a relatively co= mmon 60k per DTB >>>>>>> that=E2=80=99s 27x60k =3D 1.6MB of DTBs, that need to be instal= led. >>>>>> >>>>>> < snip > >>>>>> >>>>>> Or you can install the correct DTB on the board. You trust your= manufacturing line >>>>>> to install the correct resistors. You trust your manufacturing = line to install the >>>>>> correct kernel version (eg an updated version to resolve a secur= ity issue). >>>>>> >>>>>> I thought the DT blob was supposed to follow the same standard t= hat other OS's or >>>>>> bootloaders understood. Are you willing to break that? (This i= s one of those >>>>>> ripples I mentioned in my other emails.) >>>>>> >>>>> >>>>> Trust no-one. >>>>> >>>>> This is one of those things that the kernel community doesn=E2=80= =99t understand which makes people >>>>> who push product quite mad. >>>>> >>>>> Engineering a product is not only about meeting customer spec, in= order to turn a profit >>>>> the whole endeavor must be engineered as well for manufacturabili= ty. >>>>> >>>>> Yes, you can always manually install files in the bootloader. For= 1 board no problem. >>>>> For 10 doable. For 100 I guess you can hire an extra guy. For 1 m= illion? Guess what, >>>>> instead of turning a profit you=E2=80=99re losing money if you on= ly have a few cents of profit >>>>> per unit. >>>> >>>> I'm not installing physical components manually. Why would I be i= nstalling software >>>> manually? (rhetorical question) >>>> >>> >>> Because on high volume product runs the flash comes preprogrammed a= nd is soldered as is. >>> >>> Having a single binary to flash to every revision of the board make= s logistics considerably >>> easier. >>> >>> Having to boot and tweak the bootloader settings to select the corr= ect dtb (even if it=E2=80=99s present >>> on the flash medium) takes time and is error-prone. >>> >>> Factory time =3D=3D money, errors =3D=3D money. >>> >>>>> >>>>> No knobs to tweak means no knobs to break. And a broken knob can = have pretty bad consequences >>>>> for a few million units.=20 >>>> >>>> And you produce a few million units before testing that the first = one off the line works? >>>> >>> >>> The first one off the line works. The rest will get some burn in an= d functional testing if you=E2=80=99re >>> lucky. In many cases where the product is very cheap it might make = financial sense to just ship >>> as is and deal with recalls, if you=E2=80=99re reasonably happy aft= er a little bit of statistical sampling. >>> >>> Hardware is hard :) >> >> I'm failing to see how this series improves your manufacturing proce= ss at all. >> >> 1. Won't you have to provide the factory with different eeprom image= s for the >> White and Black? You _trust_ them to get that right, or more lik= ely, you >> have process control procedures in place so that you don't get 1 = million Blacks >> flashed with the White eeprom image. >> >> 2. The White and Black use different memory technology so it's not a= s if the >> eMMC from the Black will end up on the White SMT line (or vice ve= rsa). >> >> 3 For that matter, why wouldn't you worry that all the microSD card= s intended >> for the White were accidentally assembled with the first 50,000 B= lacks; at >> that point you're losing a lot more than a few cents of profit. A= nd that has >> nothing to do with what image you provided. >> >=20 > As you said, we can imagine many reasons to have a failure during the > production, having several DTB files will increase the risk. It's interesting that you don't see the added complexity of open-coding the i2c driver or mixing DTS fragments for different designs as increas= ed risk (for us all). >> 3. The factory is just as likely to use some other customer's image = by accident, >> so you're just as likely to have the same failure rate if you hav= e no test >> process at the factory. >> >> 4. If you're using offline programming, the image has to be tested a= fter >> reflow anyway. >> >> IOW, your QA process will not change at all =3D=3D same cost. >> >> Regards, >> Peter Hurley -- To unsubscribe from this list: send the line "unsubscribe devicetree" i= n the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html