Re: [yocto] OE Summit report from SCaLE17x
On Fri, Mar 15, 2019 at 10:27 AM Trevor Woerner wrote: > I'd like to thank the OE Board of Directors who > gave me permission to create and run a mini-conference on their behalf. > ... and provided funding!! (without which this event would never have been able to happen) So thanks to the Board for the funding, and to everyone who has contributed financially to OE who provide this funding. -- ___ yocto mailing list yocto@yoctoproject.org https://lists.yoctoproject.org/listinfo/yocto
[yocto] OE Summit report from SCaLE17x
Hi, Earlier this week I got back from SCaLE17x during which I taught an E-ALE class on Buildroot, was a TA for the other E-ALE classes, organized and ran the first ever (hopefully of more to come) OE Summit, and gave a general intro talk on OE. I wasn't planning on giving an OE talk at SCaLE, but the person who was supposed to talk wasn't able to make it. I had travelled to SCaLE with a backup talk prepared as a "plan B", and it was good that I did. I'm quite happy with how the inaugural OpenEmbedded Summit went. Everything went off without a hitch, and there was a lot of interaction and feedback between the speakers and the audience. We had 4 speakers: myself (Togan Labs): Using OE Drew Moseley (mender.io): Mechanisms for Enabling and Configuring WiFi in OE Alistair Francis (Western Digital): RISC-V and OE Jon Mason (ARM): Kernel Development Workflows Using OE Counting attendance is a bit tricky because people sometimes wander in and out. Do we count someone who wanders in partway through but leaves before the end? My talk was the least attended. I forgot to count the audience members (since I was giving the talk), but I did notice that a couple people wandered in during the talk. Behan counted 12; like I said, I didn't think to get a specific number myself. In any case, my talk wasn't in the programme, since I was filling in for someone who didn't make it to the conference. Drew's and Jon's talks had roughly 20 audience members each, and Alistair's talk was about 30. It's too bad the OE Summit was pitted against SCaLE's "Embedded Track". I think we ended up "competing" for audience members between the two of us. The OE Summit was put in a rather large room. Between teaching and helping out with E-ALE and running the OE Summit, I didn't attend a single talk at the conference nor spend time in any other part of the conference except for these two rooms (which were beside each other). I have been told anecdotally, however, that the attendance numbers we saw for the OE Summit were in-line with the numbers of attendees for several of the other talks, but I can't comment personally. As we were still setting up for my OE talk, the very first talk of the OE Summit, one of the audience members put up his hand and asked what the difference is between "Yocto" and "OE". We weren't specifically prepared to comment off-the-cuff about the relationship between the two projects; it's not as if we had a prepared statement to read on the topic. We tried to fumble through a satisfactory answer while remaining as neutral as possible on the topic. But it sure would be great if the two projects could get together and put out an official statement on the matter to which we could point all such curious people. Interestingly enough, this person was sure that The Yocto Project pre-dated OE, which we had to correct. Then he assumed these were two projects that had simply forked from each other at some point in the past. It was probably actually more confusing to him that OE and YP are two separate projects, but with almost all the same developers, working on almost all the same code. In any case, The Yocto Project's Community Manager was on-hand, thankfully, and I don't think he had any objections to any of the things we said in answering this question. My talk was a general introduction talk about OE, but all the other talks assumed the audience knew something about OE. This turned out to not be the case. Pretty much all the speakers had to field questions along the lines of "so what is OE anyway?". Sadly, in many cases, saying "it's like Buildroot" is often what would help the person asking the question the most. I think it's noteworthy that more people seem to know what Buildroot is, but not OE/YP. Although I realize this is just one conference, and a very small sample set, so it wouldn't be fair to generalize from these numbers. But it is true that these are the sorts of questions we were frequently asked, so it's noteworthy in that respect. Oddly enough, when I was finished the slides portion of the Buildroot class, the very first statement a student made was to complain about how terribly difficult Yocto is to learn and use. Why this came up in a class about Buildroot is confusing to me too, maybe he looked me up online and found my affiliation with OE? I don't know. I was careful to not mention the words "OE" or "Yocto" at all in my Buildroot talk. In any case, in his opinion, using Yocto requires a masters degree. Obviously he's exaggerating for effect. I'd like to personally thank the speakers who agreed to give a talk at a conference that didn't exist prior to last weekend, who worked hard on their talks, and who put together some very interesting presentations. I think we all walked away having learned a bit more about OE. I'd like to thank Behan and Tom King who helped organize. Behan who setup his recording equipment to record the talks (just in case there's an issue with SCaL
[yocto] can't find libudev.h
Hello, I am trying to build a recipe that has a header dependency on
libudev.h. I have added a
DEPENDS += "systemd"
because it looks like systemd provides libudev, but I continue to get the
following error:
hid.c:44:21: fatal error: libudev.h: No such file or directory
This is what a find shows:
find /build -name libudev.h 2>/dev/null
/build/multiplexor-glibc/work/corei7-64-oe-linux/systemd/1_241-r0/git/src/libudev/libudev.h
/build/multiplexor-glibc/work/corei7-64-oe-linux/systemd/1_241-r0/sysroot-destdir/usr/include/libudev.h
/build/multiplexor-glibc/work/corei7-64-oe-linux/abc-app/0.1-r0/recipe-sysroot/usr/include/libudev.h
/build/multiplexor-glibc/work/intel_corei7_64-oe-linux/image-multiplexor-dev/1.0-r0/rootfs/usr/include/libudev.h
/build/multiplexor-glibc/sysroots-components/corei7-64/systemd/usr/include/libudev.h
Any thoughts on how to resolve or troubleshoot this are appreciated.
Recipe and full error message are below.
LICENSE = "CLOSED"
LIC_FILES_CHKSUM = ""
SRC_URI = "\
git://
g...@git.bitbucket.is.company.com/abc/abc.git;protocol=ssh;destsuffix=abc;name=abc
\
git://
g...@git.bitbucket.is.company.com/abc/sequencer.git;protocol=ssh;destsuffix=abc/sequencer;name=sequencer
\
git://
github.com/kerrydwong/MCP2210-Library.git;destsuffix=abc/MCP2210-Library;name=MCP2210
\
"
# Modify these as desired
#PV = "1.0+git${SRCPV}"
PV = "0.1"
SRCREV_default = "${AUTOREV}"
SRCREV_abc = "${AUTOREV}"
SRCREV_sequencer = "${AUTOREV}"
SRCREV_MCP2210 = "${AUTOREV}"
DEPENDS += "systemd"
RDEPENDS_${PN} += "systemd"
S = "${WORKDIR}/abc"
# NOTE: this is a Makefile-only piece of software, so we cannot generate
much of the
# recipe automatically - you will need to examine the Makefile yourself and
ensure
# that the appropriate arguments are passed in.
do_patch[noexec] = "1"
do_configure () {
# Specify any needed configure commands here
:
}
do_compile () {
# You will almost certainly need to add additional arguments here
oe_runmake
}
do_install () {
# NOTE: unable to determine what to put here - there is a Makefile
but no
# target named "install", so you will need to define this yourself
:
}
Log data follows:
| DEBUG: Executing shell function do_compile
| NOTE: make -j 72
| cd MCP2210-Library; make CXXFLAGS=-fPIC CFLAGS=-fPIC all
| make[1]: Entering directory
'/build/multiplexor-glibc/work/corei7-64-oe-linux/abc-app/0.1-r0/abc/MCP2210-Library'
| for CONF in Debug Release ; \
| do \
| "make" -f nbproject/Makefile-${CONF}.mk QMAKE= SUBPROJECTS=
.build-conf; \
| done
| make[2]: Entering directory
'/build/multiplexor-glibc/work/corei7-64-oe-linux/abc-app/0.1-r0/abc/MCP2210-Library'
| "make" -f nbproject/Makefile-Debug.mk dist/Debug/GNU-Linux-x86/hidusbtest
| make[3]: Entering directory
'/build/multiplexor-glibc/work/corei7-64-oe-linux/abc-app/0.1-r0/abc/MCP2210-Library'
| mkdir -p build/Debug/GNU-Linux-x86
| rm -f build/Debug/GNU-Linux-x86/hid.o.d
| gcc -fPIC -c -g -MMD -MP -MF build/Debug/GNU-Linux-x86/hid.o.d -o
build/Debug/GNU-Linux-x86/hid.o hid.c
| hid.c:44:21: fatal error: libudev.h: No such file or directory
| compilation terminated.
--
___
yocto mailing list
yocto@yoctoproject.org
https://lists.yoctoproject.org/listinfo/yocto
Re: [yocto] can't find libudev.h
On Fri, 15 Mar 2019 at 17:00, Derek Dresser wrote: > | gcc -fPIC -c -g -MMD -MP -MF build/Debug/GNU-Linux-x86/hid.o.d -o > build/Debug/GNU-Linux-x86/hid.o hid.c 'gcc' isn't a cross compiler, and it isn't using the sysroot flags. The problem with people writing their own makefiles is that they think it's easy, but it isn't. You'll have to figure out why the maze of Makefiles in that project don't respect CC/CPPFLAGS/CFLAGS/LDFLAGS that are exported into the environment. Ross -- ___ yocto mailing list yocto@yoctoproject.org https://lists.yoctoproject.org/listinfo/yocto
[yocto] [meta-security][PATCH] libseccomp: update to 2.4.0
Update the syscall table for Linux v5.0-rc5.
also a security releated issue;
https://github.com/seccomp/libseccomp/issues/139
Signed-off-by: Armin Kuster
---
.../libseccomp/{libseccomp_2.3.3.bb => libseccomp_2.4.0.bb} | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
rename recipes-security/libseccomp/{libseccomp_2.3.3.bb =>
libseccomp_2.4.0.bb} (95%)
diff --git a/recipes-security/libseccomp/libseccomp_2.3.3.bb
b/recipes-security/libseccomp/libseccomp_2.4.0.bb
similarity index 95%
rename from recipes-security/libseccomp/libseccomp_2.3.3.bb
rename to recipes-security/libseccomp/libseccomp_2.4.0.bb
index 9c66db6..41ffd62 100644
--- a/recipes-security/libseccomp/libseccomp_2.3.3.bb
+++ b/recipes-security/libseccomp/libseccomp_2.4.0.bb
@@ -4,9 +4,9 @@ SECTION = "security"
LICENSE = "LGPL-2.1"
LIC_FILES_CHKSUM =
"file://LICENSE;beginline=0;endline=1;md5=8eac08d22113880357ceb8e7c37f989f"
-SRCREV = "74b190e1aa05f07da0c61fb9a30dbc9c18ce2c9d"
+SRCREV = "4d64011741375bb1a4ba7d71905ca37b97885083"
-SRC_URI = "git://github.com/seccomp/libseccomp.git;branch=release-2.3 \
+SRC_URI = "git://github.com/seccomp/libseccomp.git;branch=release-2.4 \
file://run-ptest \
"
--
2.17.1
--
___
yocto mailing list
yocto@yoctoproject.org
https://lists.yoctoproject.org/listinfo/yocto
Re: [yocto] [OE-core] OE Summit report from SCaLE17x
Trevor, Thank you for leading this effort. Its much appreciated. On 3/15/19 7:27 AM, Trevor Woerner wrote: > Hi, > > Earlier this week I got back from SCaLE17x during which I taught an E-ALE > class on Buildroot, was a TA for the other E-ALE classes, organized and ran > the first ever (hopefully of more to come) OE Summit, and gave a general intro > talk on OE. I wasn't planning on giving an OE talk at SCaLE, but the person > who was supposed to talk wasn't able to make it. I had travelled to SCaLE with > a backup talk prepared as a "plan B", and it was good that I did. > > I'm quite happy with how the inaugural OpenEmbedded Summit went. Everything > went off without a hitch, and there was a lot of interaction and feedback > between the speakers and the audience. > > We had 4 speakers: > > myself (Togan Labs): Using OE > Drew Moseley (mender.io): Mechanisms for Enabling and Configuring WiFi > in OE > Alistair Francis (Western Digital): RISC-V and OE > Jon Mason (ARM): Kernel Development Workflows Using OE Thank you all for participating in the first ever OE Summit. Kind regards, Armin > > Counting attendance is a bit tricky because people sometimes wander in and > out. Do we count someone who wanders in partway through but leaves before the > end? My talk was the least attended. I forgot to count the audience members > (since I was giving the talk), but I did notice that a couple people wandered > in during the talk. Behan counted 12; like I said, I didn't think to get a > specific number myself. In any case, my talk wasn't in the programme, since I > was filling in for someone who didn't make it to the conference. Drew's and > Jon's talks had roughly 20 audience members each, and Alistair's talk was > about 30. > > It's too bad the OE Summit was pitted against SCaLE's "Embedded Track". I > think we ended up "competing" for audience members between the two of us. The > OE Summit was put in a rather large room. Between teaching and helping out > with E-ALE and running the OE Summit, I didn't attend a single talk at the > conference nor spend time in any other part of the conference except for > these two rooms (which were beside each other). I have been told anecdotally, > however, that the attendance numbers we saw for the OE Summit were in-line > with the numbers of attendees for several of the other talks, but I can't > comment personally. > > As we were still setting up for my OE talk, the very first talk of the OE > Summit, one of the audience members put up his hand and asked what the > difference is between "Yocto" and "OE". We weren't specifically prepared to > comment off-the-cuff about the relationship between the two projects; it's not > as if we had a prepared statement to read on the topic. We tried to fumble > through a satisfactory answer while remaining as neutral as possible on the > topic. But it sure would be great if the two projects could get together and > put out an official statement on the matter to which we could point all such > curious people. Interestingly enough, this person was sure that The Yocto > Project pre-dated OE, which we had to correct. Then he assumed these were > two projects that had simply forked from each other at some point in the > past. It was probably actually more confusing to him that OE and YP are > two separate projects, but with almost all the same developers, working on > almost all the same code. In any case, The Yocto Project's Community Manager > was on-hand, thankfully, and I don't think he had any objections to any of > the things we said in answering this question. > > My talk was a general introduction talk about OE, but all the other talks > assumed the audience knew something about OE. This turned out to not be the > case. Pretty much all the speakers had to field questions along the lines of > "so what is OE anyway?". Sadly, in many cases, saying "it's like Buildroot" is > often what would help the person asking the question the most. I think it's > noteworthy that more people seem to know what Buildroot is, but not OE/YP. > Although I realize this is just one conference, and a very small sample set, > so it wouldn't be fair to generalize from these numbers. But it is true that > these are the sorts of questions we were frequently asked, so it's noteworthy > in that respect. > > Oddly enough, when I was finished the slides portion of the Buildroot class, > the very first statement a student made was to complain about how terribly > difficult Yocto is to learn and use. Why this came up in a class about > Buildroot is confusing to me too, maybe he looked me up online and found my > affiliation with OE? I don't know. I was careful to not mention the words "OE" > or "Yocto" at all in my Buildroot talk. In any case, in his opinion, using > Yocto requires a masters degree. Obviously he's exaggerating for effect. > > I'd like to personally thank the speakers who agreed to give a talk at a > conference that didn't exist pr
[yocto] [meta-security][PATCH] keyutils: fix QA WARNING
WARNING: keyutils-1.6-r0 do_package: QA Issue: keyutils: Files/directories were
installed but not shipped in any package:
/lib/pkgconfig
/lib/pkgconfig/libkeyutils.pc
Signed-off-by: Armin Kuster
---
recipes-security/keyutils/keyutils_1.6.bb | 3 +++
1 file changed, 3 insertions(+)
diff --git a/recipes-security/keyutils/keyutils_1.6.bb
b/recipes-security/keyutils/keyutils_1.6.bb
index 730e174..054e0c1 100644
--- a/recipes-security/keyutils/keyutils_1.6.bb
+++ b/recipes-security/keyutils/keyutils_1.6.bb
@@ -35,6 +35,7 @@ EXTRA_OEMAKE = "'CFLAGS=${CFLAGS} -Wall' \
"
do_install () {
+install -d ${D}/${nonarch_base_libdir}/pkgconfig
oe_runmake DESTDIR=${D} install
}
@@ -43,4 +44,6 @@ do_install_ptest () {
sed -i -e 's/OSDIST=Unknown/OSDIST=${DISTRO}/'
${D}${PTEST_PATH}/tests/prepare.inc.sh
}
+FILES_${PN}-dev += "${nonarch_base_libdir}/pkgconfig/libkeyutils.pc"
+
RDEPENDS_${PN}-ptest += "glibc-utils lsb"
--
2.17.1
--
___
yocto mailing list
yocto@yoctoproject.org
https://lists.yoctoproject.org/listinfo/yocto
Re: [yocto] can't find libudev.h
On Fri, Mar 15, 2019 at 10:23 AM Burton, Ross wrote: > On Fri, 15 Mar 2019 at 17:00, Derek Dresser wrote: > > | gcc -fPIC -c -g -MMD -MP -MF build/Debug/GNU-Linux-x86/hid.o.d -o > build/Debug/GNU-Linux-x86/hid.o hid.c > > 'gcc' isn't a cross compiler, and it isn't using the sysroot flags. > > The problem with people writing their own makefiles is that they think > it's easy, but it isn't. You'll have to figure out why the maze of > Makefiles in that project don't respect CC/CPPFLAGS/CFLAGS/LDFLAGS > that are exported into the environment. > You might want to consider a different driver: https://archive.fosdem.org/2018/schedule/event/rapid_spi_over_usb/attachments/audio/2461/export/events/attachments/rapid_spi_over_usb/audio/2461/USB_SPI_FOSDEM_2018.pdf > Ross > -- > ___ > yocto mailing list > yocto@yoctoproject.org > https://lists.yoctoproject.org/listinfo/yocto > -- ___ yocto mailing list yocto@yoctoproject.org https://lists.yoctoproject.org/listinfo/yocto
