| Message ID | 1495120094-44581-1-git-send-email-matthew.weber@rockwellcollins.com |
|---|---|
| State | Superseded |
| Headers | show |
>>>>> "Matt" == Matt Weber <matthew.weber@rockwellcollins.com> writes: > From: Sam Voss <samuel.voss@rockwellcollins.com> > Add the option to enable linux sysfs gpio framework > configure option in menuconfig. > Signed-off-by: Sam Voss <samuel.voss@rockwellcollins.com> > Signed-off-by: Matt Weber <matthew.weber@rockwellcollins.com> > --- > Change Log > v1 -> v2 > - upstream accepted linux gpio fix, removed patch Didn't you just agree with Baruch that we could just enable it unconditially and not have a Config.in option for it?
Peter / Baruch, On Fri, May 19, 2017 at 8:14 AM, Peter Korsgaard <peter@korsgaard.com> wrote: > > >>>>> "Matt" == Matt Weber <matthew.weber@rockwellcollins.com> writes: > > > From: Sam Voss <samuel.voss@rockwellcollins.com> > > Add the option to enable linux sysfs gpio framework > > configure option in menuconfig. > > > Signed-off-by: Sam Voss <samuel.voss@rockwellcollins.com> > > Signed-off-by: Matt Weber <matthew.weber@rockwellcollins.com> > > > --- > > > Change Log > > > v1 -> v2 > > - upstream accepted linux gpio fix, removed patch > > Didn't you just agree with Baruch that we could just enable it > unconditially and not have a Config.in option for it? > Not sure we reached a conclusion. I guess the question would be, right now in buildroot it doesn't enable this option. So if we by default go linux-gpio it disables the legacy options which were the previous default (kernel's older then sysfs gpio support). I'm good either way, but the fundamental package behavior changes. Matt
Hello, On Fri, 19 May 2017 08:58:25 -0500, Matthew Weber wrote: > Not sure we reached a conclusion. I guess the question would be, > right now in buildroot it doesn't enable this option. So if we by > default go linux-gpio it disables the legacy options which were the > previous default (kernel's older then sysfs gpio support). I'm good > either way, but the fundamental package behavior changes. Package behavior also changes every time we bump a package. So Baruch's question was whether the size impact was significant or not. If it's not, then no need to make an option for it. Thomas
>>>>> "Matthew" == Matthew Weber <matthew.weber@rockwellcollins.com> writes: Hi, >> Didn't you just agree with Baruch that we could just enable it >> unconditially and not have a Config.in option for it? >> > Not sure we reached a conclusion. I guess the question would be, > right now in buildroot it doesn't enable this option. So if we by > default go linux-gpio it disables the legacy options which were the > previous default (kernel's older then sysfs gpio support) Do you mean that enabling this driver disables anything else? I don't see that. The kernel's gpio sysfs interface is close to 10 years old by now, so I doubt that is an issue. From the looks of it, the linuxgpio driver seems to be very small compared to the total size of avrdude, so unconditionally enabling it sounds sensible to me. I notice that we are using a fork of avrdude which hasn't really been updated significantly since 2013 instead of the official one which had a release in 2016. Any reason we shouldn't be using the official one instead?
Peter On Fri, May 19, 2017 at 9:48 AM, Peter Korsgaard <peter@korsgaard.com> wrote: >>>>>> "Matthew" == Matthew Weber <matthew.weber@rockwellcollins.com> writes: > > Hi, > > >> Didn't you just agree with Baruch that we could just enable it > >> unconditially and not have a Config.in option for it? > >> > > > Not sure we reached a conclusion. I guess the question would be, > > right now in buildroot it doesn't enable this option. So if we by > > default go linux-gpio it disables the legacy options which were the > > previous default (kernel's older then sysfs gpio support) > > Do you mean that enabling this driver disables anything else? I don't > see that. The kernel's gpio sysfs interface is close to 10 years old by > now, so I doubt that is an issue. > > From the looks of it, the linuxgpio driver seems to be very small > compared to the total size of avrdude, so unconditionally enabling it > sounds sensible to me. Ok, I'm good with unconditionally enabling. > > I notice that we are using a fork of avrdude which hasn't really been > updated significantly since 2013 instead of the official one which had a > release in 2016. Any reason we shouldn't be using the official one > instead? Very good question, we are unsure. I've added a few people that have modified that package to this email to ask. Wojciech / Samuel / Gregory, any reason to stay on this fork of avrdude? Matt
>> I notice that we are using a fork of avrdude which hasn't really been >> updated significantly since 2013 instead of the official one which had a >> release in 2016. Any reason we shouldn't be using the official one >> instead? > > Very good question, we are unsure. I've added a few people that have > modified that package to this email to ask. > > Wojciech / Samuel / Gregory, any reason to stay on this fork of avrdude? The biggest reason is due to the fact that the spi functionality has not been merged back into mainline, and requests to do it have fallen onto deaf ears. I sent an email recently to the mainline avrdude asking about getting this functionality merged in and have yet to hear anything from them.
diff --git a/package/avrdude/Config.in b/package/avrdude/Config.in index 3757f17..ac982e1 100644 --- a/package/avrdude/Config.in +++ b/package/avrdude/Config.in @@ -15,6 +15,16 @@ config BR2_PACKAGE_AVRDUDE https://github.com/kcuzner/avrdude +if BR2_PACKAGE_AVRDUDE + +config BR2_PACKAGE_AVRDUDE_LINUXGPIO + bool "linux gpio" + depends on BR2_PACKAGE_AVRDUDE + help + Enables the tool to use the Linux Sysfs based GPIO framework for GPIO + access by configuring AVRDude with '--enable-linuxgpio'. +endif + comment "avrdude needs a uClibc or glibc toolchain w/ threads, wchar, dynamic library" depends on !BR2_TOOLCHAIN_HAS_THREADS || !BR2_USE_WCHAR || BR2_STATIC_LIBS \ || !(BR2_TOOLCHAIN_USES_UCLIBC || BR2_TOOLCHAIN_USES_GLIBC) diff --git a/package/avrdude/avrdude.mk b/package/avrdude/avrdude.mk index f914ed7..19ffdfa 100644 --- a/package/avrdude/avrdude.mk +++ b/package/avrdude/avrdude.mk @@ -23,6 +23,12 @@ else ifeq ($(BR2_PACKAGE_LIBFTDI),y) AVRDUDE_DEPENDENCIES += libftdi endif +ifeq ($(BR2_PACKAGE_AVRDUDE_LINUXGPIO),y) +AVRDUDE_CONF_OPTS = --enable-linuxgpio +else +AVRDUDE_CONF_OPTS = --disable-linuxgpio +endif + # if /etc/avrdude.conf exists, the installation process creates a # backup file, which we do not want in the context of Buildroot. define AVRDUDE_REMOVE_BACKUP_FILE