From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754979AbbBTPY6 (ORCPT ); Fri, 20 Feb 2015 10:24:58 -0500 Received: from mail-qg0-f50.google.com ([209.85.192.50]:37541 "EHLO mail-qg0-f50.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754433AbbBTPY4 (ORCPT ); Fri, 20 Feb 2015 10:24:56 -0500 Message-ID: <54E751BF.3000604@hurleysoftware.com> Date: Fri, 20 Feb 2015 10:24:47 -0500 From: Peter Hurley User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.4.0 MIME-Version: 1.0 To: Pantelis Antoniou CC: frowand.list@gmail.com, Mark Rutland , "devicetree@vger.kernel.org" , Tony Lindgren , Koen Kooi , Nicolas Ferre , "linux-kernel@vger.kernel.org" , Grant Likely , "linux-arm-kernel@lists.infradead.org" , Matt Porter , Guenter Roeck Subject: Re: [PATCH 2/4] of: DT quirks infrastructure 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> <54E74BF6.208@hurleysoftware.com> In-Reply-To: Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 8bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 02/20/2015 10:02 AM, Pantelis Antoniou wrote: > Hi Peter, > >> On Feb 20, 2015, at 17:00 , Peter Hurley wrote: >> >> 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 cumbersome >>>>>>>>>>>>> +for the following reasons: >>>>>>>>>>>>> + >>>>>>>>>>>>> +1. The act of selecting a different boot device tree blob requires >>>>>>>>>>>>> +a reasonably advanced bootloader with some kind of configuration or >>>>>>>>>>>>> +scripting capabilities. Sadly this is not the case many times, the >>>>>>>>>>>>> +bootloader is extremely dumb and can only use a single dt blob. >>>>>>>>>>>> >>>>>>>>>>>> You can have several bootloader builds, or even a single build 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 board. >>>>>>>>>> >>>>>>>>>> I wasn't on about the DTB. I was on about the loader binary, in 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 variants >>>>>>>>>> (albeit with limited utility) or the quirk approach wouldn't work… >>>>>>>>>> >>>>>>>>> >>>>>>>>> That’s not correct; the only part of the DTB that needs to be common >>>>>>>>> is the model property that would allow the quirk detection logic to fire. >>>>>>>>> >>>>>>>>> So, there is a base DTB that will work on all variants, but that 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 subsets >>>>>>>>> of the DTB that might work, they might be unsafe. >>>>>>>>> >>>>>>>>> For instance on the beaglebone the regulator configuration is different >>>>>>>>> between white and black, it is imperative you get them right otherwise >>>>>>>>> you risk board damage. >>>>>>>>> >>>>>>>>>>>> So it's not necessarily true that you need a complex bootloader. >>>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>>>> +2. On many instances boot time is extremely critical; in some cases >>>>>>>>>>>>> +there are hard requirements like having working video feeds in under >>>>>>>>>>>>> +2 seconds from power-up. This leaves an extremely small time budget for >>>>>>>>>>>>> +boot-up, as low as 500ms to kernel entry. The sanest way to get there >>>>>>>>>>>>> +is by removing the standard bootloader from the normal boot sequence >>>>>>>>>>>>> +altogether by having a very small boot shim that loads the kernel and >>>>>>>>>>>>> +immediately jumps to kernel, like falcon-boot mode in u-boot does. >>>>>>>>>>>> >>>>>>>>>>>> Given my previous comments above I don't see why this is relevant. >>>>>>>>>>>> You're already passing _some_ DTB here, so if you can organise for the >>>>>>>>>>>> board to statically provide a sane DTB that's fine, or you can resort to >>>>>>>>>>>> appended DTB if it's not possible to update the board configuration. >>>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> You’re missing the point. I can’t use the same DTB for each revision of the >>>>>>>>>>> board. Each board is similar but it’s not identical. >>>>>>>>>> >>>>>>>>>> I think you've misunderstood my point. If you program the board with the >>>>>>>>>> relevant DTB, or use appended DTB, then you will pass the correct DTB to >>>>>>>>>> the kernel without need for quirks. >>>>>>>>>> >>>>>>>>>> I understand that each variant is somewhat incompatible (and hence needs >>>>>>>>>> its own DTB). >>>>>>>>> >>>>>>>>> In theory it might work, in practice this does not. Ludovic mentioned that they >>>>>>>>> have 27 different DTBs in use at the moment. At a relatively common 60k per DTB >>>>>>>>> that’s 27x60k = 1.6MB of DTBs, that need to be installed. >>>>>>>> >>>>>>>> < 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 security issue). >>>>>>>> >>>>>>>> I thought the DT blob was supposed to follow the same standard that other OS's or >>>>>>>> bootloaders understood. Are you willing to break that? (This is one of those >>>>>>>> ripples I mentioned in my other emails.) >>>>>>>> >>>>>>> >>>>>>> Trust no-one. >>>>>>> >>>>>>> This is one of those things that the kernel community doesn’t 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 manufacturability. >>>>>>> >>>>>>> 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 million? Guess what, >>>>>>> instead of turning a profit you’re losing money if you only have a few cents of profit >>>>>>> per unit. >>>>>> >>>>>> I'm not installing physical components manually. Why would I be installing software >>>>>> manually? (rhetorical question) >>>>>> >>>>> >>>>> Because on high volume product runs the flash comes preprogrammed and is soldered as is. >>>>> >>>>> Having a single binary to flash to every revision of the board makes logistics considerably >>>>> easier. >>>>> >>>>> Having to boot and tweak the bootloader settings to select the correct dtb (even if it’s present >>>>> on the flash medium) takes time and is error-prone. >>>>> >>>>> Factory time == money, errors == money. >>>>> >>>>>>> >>>>>>> No knobs to tweak means no knobs to break. And a broken knob can have pretty bad consequences >>>>>>> for a few million units. >>>>>> >>>>>> 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 and functional testing if you’re >>>>> 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’re reasonably happy after a little bit of statistical sampling. >>>>> >>>>> Hardware is hard :) >>>> >>>> I'm failing to see how this series improves your manufacturing process at all. >>>> >>>> 1. Won't you have to provide the factory with different eeprom images for the >>>> White and Black? You _trust_ them to get that right, or more likely, 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 as if the >>>> eMMC from the Black will end up on the White SMT line (or vice versa). >>>> >>>> 3 For that matter, why wouldn't you worry that all the microSD cards intended >>>> for the White were accidentally assembled with the first 50,000 Blacks; at >>>> that point you're losing a lot more than a few cents of profit. And that has >>>> nothing to do with what image you provided. >>>> >>> >>> 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 increased risk >> (for us all). >> >> > > You don’t have to use it. > diff --git a/arch/arm/mach-omap2/Makefile b/arch/arm/mach-omap2/Makefile > index 5d27dfd..02129e7 100644 > --- a/arch/arm/mach-omap2/Makefile > +++ b/arch/arm/mach-omap2/Makefile > @@ -259,6 +259,11 @@ obj-$(CONFIG_MACH_CRANEBOARD) += board-am3517crane.o > > obj-$(CONFIG_MACH_SBC3530) += board-omap3stalker.o > > +# DT quirks > +ifneq ($(CONFIG_OF_DYNAMIC),) > +obj-$(CONFIG_SOC_AM33XX) += am33xx-dt-quirks.o > +endif Won't this automatically be included on my Black that supports DT overlays? > Some people really do though. As for increased risk > I expect to see arguments instead of a statement. No one is wasting your time with random arguments. Please keep your tone civil. Regards, Peter Hurley >>>> 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 have no test >>>> process at the factory. >>>> >>>> 4. If you're using offline programming, the image has to be tested after >>>> reflow anyway. >>>> >>>> IOW, your QA process will not change at all == same cost. >>>> >>>> Regards, >>>> Peter Hurley >> > > Regards > > — Pantelis > 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 10:24:47 -0500 Message-ID: <54E751BF.3000604@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> <54E74BF6.208@hurleysoftware.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org To: Pantelis Antoniou Cc: frowand.list@gmail.com, Mark Rutland , "devicetree@vger.kernel.org" , Tony Lindgren , Koen Kooi , Nicolas Ferre , "linux-kernel@vger.kernel.org" , Grant Likely , "linux-arm-kernel@lists.infradead.org" , Matt Porter , Guenter Roeck List-Id: devicetree@vger.kernel.org On 02/20/2015 10:02 AM, Pantelis Antoniou wrote: > Hi Peter, >=20 >> On Feb 20, 2015, at 17:00 , Peter Hurley = wrote: >> >> 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 cu= mbersome >>>>>>>>>>>>> +for the following reasons: >>>>>>>>>>>>> + >>>>>>>>>>>>> +1. The act of selecting a different boot device tree blo= b requires >>>>>>>>>>>>> +a reasonably advanced bootloader with some kind of confi= guration or >>>>>>>>>>>>> +scripting capabilities. Sadly this is not the case many = times, the >>>>>>>>>>>>> +bootloader is extremely dumb and can only use a single d= t blob. >>>>>>>>>>>> >>>>>>>>>>>> You can have several bootloader builds, or even a single b= uild with >>>>>>>>>>>> something like appended DTB to get an appropriate DTB if t= he same binary >>>>>>>>>>>> will otherwise work across all variants of a board. >>>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> No, the same DTB will not work across all the variants of a= board. >>>>>>>>>> >>>>>>>>>> I wasn't on about the DTB. I was on about the loader binary,= in 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= variants >>>>>>>>>> (albeit with limited utility) or the quirk approach wouldn't= work=E2=80=A6 >>>>>>>>>> >>>>>>>>> >>>>>>>>> That=E2=80=99s not correct; the only part of the DTB that nee= ds to be common >>>>>>>>> is the model property that would allow the quirk detection lo= gic to fire. >>>>>>>>> >>>>>>>>> So, there is a base DTB that will work on all variants, but t= hat only means >>>>>>>>> that it will work only up to the point that the quirk detecto= r method >>>>>>>>> can work. So while in recommended practice there are common s= ubsets >>>>>>>>> of the DTB that might work, they might be unsafe. >>>>>>>>> >>>>>>>>> For instance on the beaglebone the regulator configuration is= different >>>>>>>>> between white and black, it is imperative you get them right = otherwise >>>>>>>>> you risk board damage. >>>>>>>>> >>>>>>>>>>>> So it's not necessarily true that you need a complex bootl= oader. >>>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>>>> +2. On many instances boot time is extremely critical; in= some cases >>>>>>>>>>>>> +there are hard requirements like having working video fe= eds in under >>>>>>>>>>>>> +2 seconds from power-up. This leaves an extremely small = time budget for >>>>>>>>>>>>> +boot-up, as low as 500ms to kernel entry. The sanest way= to get there >>>>>>>>>>>>> +is by removing the standard bootloader from the normal b= oot sequence >>>>>>>>>>>>> +altogether by having a very small boot shim that loads t= he kernel and >>>>>>>>>>>>> +immediately jumps to kernel, like falcon-boot mode in u-= boot does. >>>>>>>>>>>> >>>>>>>>>>>> Given my previous comments above I don't see why this is r= elevant. >>>>>>>>>>>> You're already passing _some_ DTB here, so if you can orga= nise for the >>>>>>>>>>>> board to statically provide a sane DTB that's fine, or you= can resort to >>>>>>>>>>>> appended DTB if it's not possible to update the board conf= iguration. >>>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> You=E2=80=99re missing the point. I can=E2=80=99t use the s= ame DTB for each revision of the >>>>>>>>>>> board. Each board is similar but it=E2=80=99s not identical= =2E >>>>>>>>>> >>>>>>>>>> I think you've misunderstood my point. If you program the bo= ard with the >>>>>>>>>> relevant DTB, or use appended DTB, then you will pass the co= rrect DTB to >>>>>>>>>> the kernel without need for quirks. >>>>>>>>>> >>>>>>>>>> I understand that each variant is somewhat incompatible (and= hence needs >>>>>>>>>> its own DTB). >>>>>>>>> >>>>>>>>> In theory it might work, in practice this does not. Ludovic m= entioned that they >>>>>>>>> have 27 different DTBs in use at the moment. At a relatively = common 60k per DTB >>>>>>>>> that=E2=80=99s 27x60k =3D 1.6MB of DTBs, that need to be inst= alled. >>>>>>>> >>>>>>>> < snip > >>>>>>>> >>>>>>>> Or you can install the correct DTB on the board. You trust yo= ur manufacturing line >>>>>>>> to install the correct resistors. You trust your manufacturin= g line to install the >>>>>>>> correct kernel version (eg an updated version to resolve a sec= urity issue). >>>>>>>> >>>>>>>> I thought the DT blob was supposed to follow the same standard= that other OS's or >>>>>>>> bootloaders understood. Are you willing to break that? (This= is 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 manufacturabi= lity. >>>>>>> >>>>>>> Yes, you can always manually install files in the bootloader. F= or 1 board no problem. >>>>>>> For 10 doable. For 100 I guess you can hire an extra guy. For 1= million? Guess what, >>>>>>> instead of turning a profit you=E2=80=99re losing money if you = only have a few cents of profit >>>>>>> per unit. >>>>>> >>>>>> I'm not installing physical components manually. Why would I be= installing software >>>>>> manually? (rhetorical question) >>>>>> >>>>> >>>>> Because on high volume product runs the flash comes preprogrammed= and is soldered as is. >>>>> >>>>> Having a single binary to flash to every revision of the board ma= kes logistics considerably >>>>> easier. >>>>> >>>>> Having to boot and tweak the bootloader settings to select the co= rrect 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 ca= n have pretty bad consequences >>>>>>> for a few million units.=20 >>>>>> >>>>>> And you produce a few million units before testing that the firs= t one off the line works? >>>>>> >>>>> >>>>> The first one off the line works. The rest will get some burn in = and functional testing if you=E2=80=99re >>>>> lucky. In many cases where the product is very cheap it might mak= e financial sense to just ship >>>>> as is and deal with recalls, if you=E2=80=99re reasonably happy a= fter a little bit of statistical sampling. >>>>> >>>>> Hardware is hard :) >>>> >>>> I'm failing to see how this series improves your manufacturing pro= cess at all. >>>> >>>> 1. Won't you have to provide the factory with different eeprom ima= ges for the >>>> White and Black? You _trust_ them to get that right, or more li= kely, 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= as if the >>>> eMMC from the Black will end up on the White SMT line (or vice v= ersa). >>>> >>>> 3 For that matter, why wouldn't you worry that all the microSD ca= rds intended >>>> for the White were accidentally assembled with the first 50,000 = Blacks; at >>>> that point you're losing a lot more than a few cents of profit. = And that has >>>> nothing to do with what image you provided. >>>> >>> >>> As you said, we can imagine many reasons to have a failure during t= he >>> production, having several DTB files will increase the risk. >> >> It's interesting that you don't see the added complexity of open-cod= ing >> the i2c driver or mixing DTS fragments for different designs as incr= eased risk >> (for us all). >> >> >=20 > You don=E2=80=99t have to use it. > diff --git a/arch/arm/mach-omap2/Makefile b/arch/arm/mach-omap2/Makef= ile > index 5d27dfd..02129e7 100644 > --- a/arch/arm/mach-omap2/Makefile > +++ b/arch/arm/mach-omap2/Makefile > @@ -259,6 +259,11 @@ obj-$(CONFIG_MACH_CRANEBOARD) +=3D board-am3517= crane.o > =20 > obj-$(CONFIG_MACH_SBC3530) +=3D board-omap3stalker.o > =20 > +# DT quirks > +ifneq ($(CONFIG_OF_DYNAMIC),) > +obj-$(CONFIG_SOC_AM33XX) +=3D am33xx-dt-quirks.o > +endif Won't this automatically be included on my Black that supports DT overl= ays? > Some people really do though. As for increased risk > I expect to see arguments instead of a statement. No one is wasting your time with random arguments. Please keep your ton= e civil. Regards, Peter Hurley >>>> 3. The factory is just as likely to use some other customer's imag= e by accident, >>>> so you're just as likely to have the same failure rate if you ha= ve no test >>>> process at the factory. >>>> >>>> 4. If you're using offline programming, the image has to be tested= after >>>> reflow anyway. >>>> >>>> IOW, your QA process will not change at all =3D=3D same cost. >>>> >>>> Regards, >>>> Peter Hurley >> >=20 > Regards >=20 > =E2=80=94 Pantelis >=20 From mboxrd@z Thu Jan 1 00:00:00 1970 From: peter@hurleysoftware.com (Peter Hurley) Date: Fri, 20 Feb 2015 10:24:47 -0500 Subject: [PATCH 2/4] of: DT quirks infrastructure In-Reply-To: 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> <54E74BF6.208@hurleysoftware.com> Message-ID: <54E751BF.3000604@hurleysoftware.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On 02/20/2015 10:02 AM, Pantelis Antoniou wrote: > Hi Peter, > >> On Feb 20, 2015, at 17:00 , Peter Hurley wrote: >> >> 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 cumbersome >>>>>>>>>>>>> +for the following reasons: >>>>>>>>>>>>> + >>>>>>>>>>>>> +1. The act of selecting a different boot device tree blob requires >>>>>>>>>>>>> +a reasonably advanced bootloader with some kind of configuration or >>>>>>>>>>>>> +scripting capabilities. Sadly this is not the case many times, the >>>>>>>>>>>>> +bootloader is extremely dumb and can only use a single dt blob. >>>>>>>>>>>> >>>>>>>>>>>> You can have several bootloader builds, or even a single build 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 board. >>>>>>>>>> >>>>>>>>>> I wasn't on about the DTB. I was on about the loader binary, in 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 variants >>>>>>>>>> (albeit with limited utility) or the quirk approach wouldn't work? >>>>>>>>>> >>>>>>>>> >>>>>>>>> That?s not correct; the only part of the DTB that needs to be common >>>>>>>>> is the model property that would allow the quirk detection logic to fire. >>>>>>>>> >>>>>>>>> So, there is a base DTB that will work on all variants, but that 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 subsets >>>>>>>>> of the DTB that might work, they might be unsafe. >>>>>>>>> >>>>>>>>> For instance on the beaglebone the regulator configuration is different >>>>>>>>> between white and black, it is imperative you get them right otherwise >>>>>>>>> you risk board damage. >>>>>>>>> >>>>>>>>>>>> So it's not necessarily true that you need a complex bootloader. >>>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>>>> +2. On many instances boot time is extremely critical; in some cases >>>>>>>>>>>>> +there are hard requirements like having working video feeds in under >>>>>>>>>>>>> +2 seconds from power-up. This leaves an extremely small time budget for >>>>>>>>>>>>> +boot-up, as low as 500ms to kernel entry. The sanest way to get there >>>>>>>>>>>>> +is by removing the standard bootloader from the normal boot sequence >>>>>>>>>>>>> +altogether by having a very small boot shim that loads the kernel and >>>>>>>>>>>>> +immediately jumps to kernel, like falcon-boot mode in u-boot does. >>>>>>>>>>>> >>>>>>>>>>>> Given my previous comments above I don't see why this is relevant. >>>>>>>>>>>> You're already passing _some_ DTB here, so if you can organise for the >>>>>>>>>>>> board to statically provide a sane DTB that's fine, or you can resort to >>>>>>>>>>>> appended DTB if it's not possible to update the board configuration. >>>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> You?re missing the point. I can?t use the same DTB for each revision of the >>>>>>>>>>> board. Each board is similar but it?s not identical. >>>>>>>>>> >>>>>>>>>> I think you've misunderstood my point. If you program the board with the >>>>>>>>>> relevant DTB, or use appended DTB, then you will pass the correct DTB to >>>>>>>>>> the kernel without need for quirks. >>>>>>>>>> >>>>>>>>>> I understand that each variant is somewhat incompatible (and hence needs >>>>>>>>>> its own DTB). >>>>>>>>> >>>>>>>>> In theory it might work, in practice this does not. Ludovic mentioned that they >>>>>>>>> have 27 different DTBs in use at the moment. At a relatively common 60k per DTB >>>>>>>>> that?s 27x60k = 1.6MB of DTBs, that need to be installed. >>>>>>>> >>>>>>>> < 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 security issue). >>>>>>>> >>>>>>>> I thought the DT blob was supposed to follow the same standard that other OS's or >>>>>>>> bootloaders understood. Are you willing to break that? (This is one of those >>>>>>>> ripples I mentioned in my other emails.) >>>>>>>> >>>>>>> >>>>>>> Trust no-one. >>>>>>> >>>>>>> This is one of those things that the kernel community doesn?t 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 manufacturability. >>>>>>> >>>>>>> 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 million? Guess what, >>>>>>> instead of turning a profit you?re losing money if you only have a few cents of profit >>>>>>> per unit. >>>>>> >>>>>> I'm not installing physical components manually. Why would I be installing software >>>>>> manually? (rhetorical question) >>>>>> >>>>> >>>>> Because on high volume product runs the flash comes preprogrammed and is soldered as is. >>>>> >>>>> Having a single binary to flash to every revision of the board makes logistics considerably >>>>> easier. >>>>> >>>>> Having to boot and tweak the bootloader settings to select the correct dtb (even if it?s present >>>>> on the flash medium) takes time and is error-prone. >>>>> >>>>> Factory time == money, errors == money. >>>>> >>>>>>> >>>>>>> No knobs to tweak means no knobs to break. And a broken knob can have pretty bad consequences >>>>>>> for a few million units. >>>>>> >>>>>> 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 and functional testing if you?re >>>>> 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?re reasonably happy after a little bit of statistical sampling. >>>>> >>>>> Hardware is hard :) >>>> >>>> I'm failing to see how this series improves your manufacturing process at all. >>>> >>>> 1. Won't you have to provide the factory with different eeprom images for the >>>> White and Black? You _trust_ them to get that right, or more likely, 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 as if the >>>> eMMC from the Black will end up on the White SMT line (or vice versa). >>>> >>>> 3 For that matter, why wouldn't you worry that all the microSD cards intended >>>> for the White were accidentally assembled with the first 50,000 Blacks; at >>>> that point you're losing a lot more than a few cents of profit. And that has >>>> nothing to do with what image you provided. >>>> >>> >>> 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 increased risk >> (for us all). >> >> > > You don?t have to use it. > diff --git a/arch/arm/mach-omap2/Makefile b/arch/arm/mach-omap2/Makefile > index 5d27dfd..02129e7 100644 > --- a/arch/arm/mach-omap2/Makefile > +++ b/arch/arm/mach-omap2/Makefile > @@ -259,6 +259,11 @@ obj-$(CONFIG_MACH_CRANEBOARD) += board-am3517crane.o > > obj-$(CONFIG_MACH_SBC3530) += board-omap3stalker.o > > +# DT quirks > +ifneq ($(CONFIG_OF_DYNAMIC),) > +obj-$(CONFIG_SOC_AM33XX) += am33xx-dt-quirks.o > +endif Won't this automatically be included on my Black that supports DT overlays? > Some people really do though. As for increased risk > I expect to see arguments instead of a statement. No one is wasting your time with random arguments. Please keep your tone civil. Regards, Peter Hurley >>>> 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 have no test >>>> process at the factory. >>>> >>>> 4. If you're using offline programming, the image has to be tested after >>>> reflow anyway. >>>> >>>> IOW, your QA process will not change at all == same cost. >>>> >>>> Regards, >>>> Peter Hurley >> > > Regards > > ? Pantelis >