Issue: LIN10-5872
Add IRQ statistic code in kvm/i8259.c.
(LOCAL REV: NOT UPSTREAM) -- oe-core already included it.
Signed-off-by: Jiping Ma
---
...1-kvm-i8259-support-info-pic-and-info-irq.patch | 76 +
.../0002-i8259-use-DEBUG_IRQ_COUNT-always.patch| 89 +++
The action of "dhclient -x eth0" and "dhclient -r eth0" is
same when enable ENABLE_GENTLE_SHUTDOWN. Disable ENABLE_GENTLE_SHUTDOWN
that will use the default signal hander.
Signed-off-by: Jiping Ma
---
meta/recipes-connectivity/dhcp/dhcp_4.4.1.bb | 1 -
1 file changed, 1 deletion(-)
diff --git
hi,
not sure if it's been discussed reported before. but i ended up in a
situation where a package is "empty" (e.g. it's not created) even
though "files-in-packages.txt" in its buildhistory folder shows it
should contain a file.
I have a synthetic example:
$ cat
Changes in this release:
979a351 [common] Add A2RGB10 fourcc definition
22cba79 build: meson: enables va messaging and visibility
b89f6ec va/va_trace: add trace support for RIR(rolling intra refresh).
844b90c va/va_trace: add trace support for ROI(region of interest).
Signed-off-by: Anuj Mittal
Please ignore V1 of [OE-core][PATCH] dhcp:"dhclient -x eth0" action is not
correct.
The action of "dhclient -x eth0" and "dhclient -r eth0" is
same when enable ENABLE_GENTLE_SHUTDOWN. Disable ENABLE_GENTLE_SHUTDOWN
that will use the default signal hander.
Signed-off-by: Jiping Ma
---
On Wed, May 22, 2019 at 04:29:33PM -0700, Khem Raj wrote:
> On Tue, May 21, 2019 at 11:49 PM Adrian Bunk wrote:
> >
> > On Tue, May 21, 2019 at 06:32:28PM -0700, Khem Raj wrote:
> > >...
> > > PN and PN-binutils is not empty anymore on musl
> > >...
> > > ++#if !defined(FNM_EXTMATCH)
> > > ++#
Please ignore V1 of [OE-core][PATCH] dhcp:"dhclient -x eth0" action is not
correct.
The action of "dhclient -x eth0" and "dhclient -r eth0" is
same when enable ENABLE_GENTLE_SHUTDOWN. Disable ENABLE_GENTLE_SHUTDOWN
that will use the default signal hander.
Signed-off-by: Jiping Ma
---
musl > 2.21 changed the implementation of the freeaddrinfo() function
not allowing anymore to pass null pointers to it.
This was causing a segmentation fault in connman.
---
...-gweb-fix-segfault-with-musl-v1.1.21.patch | 34 +++
.../connman/connman_1.37.bb | 5
This is the one-line mentioned in [1].
[1]
http://lists.openembedded.org/pipermail/openembedded-devel/2019-May/199769.html
Signed-off-by: Andreas Müller
---
meta/lib/oe/package_manager.py | 1 +
1 file changed, 1 insertion(+)
diff --git a/meta/lib/oe/package_manager.py
On Wed, May 22, 2019 at 11:10 PM Adrian Bunk wrote:
> On Wed, May 22, 2019 at 04:29:33PM -0700, Khem Raj wrote:
> > On Tue, May 21, 2019 at 11:49 PM Adrian Bunk wrote:
> > >
> > > On Tue, May 21, 2019 at 06:32:28PM -0700, Khem Raj wrote:
> > > >...
> > > > PN and PN-binutils is not empty
Just to make it clean: this flow of cleanup patches is *much
appreciated*. Thanks!
Ross
On Wed, 22 May 2019 at 11:59, Adrian Bunk wrote:
>
> Signed-off-by: Adrian Bunk
> ---
> ...t-Bsymbolic-fixes-configure-on-macOS.patch | 35 +++
>
Why -help- and not -doc?
Ross
On Thu, 23 May 2019 at 13:30, Andreas Müller wrote:
>
> This is the one-line mentioned in [1].
>
> [1]
> http://lists.openembedded.org/pipermail/openembedded-devel/2019-May/199769.html
>
> Signed-off-by: Andreas Müller
> ---
> meta/lib/oe/package_manager.py | 1
On Thu, 23 May 2019 at 13:44, Andreas Müller wrote:
> > Why -help- and not -doc?
> >
> Because there are help files installed in 'help' directory and
> displayed by a help browser.
Sounds a lot like documentation to me...
--
___
Openembedded-core
* just to make sure it's expaned by bitbake before it gets executed in shell
* e.g. with cmake.bbclass and cmake recipe (any recipe without
PACKAGECONFIG options have this issue) it looks like this:
bitbake -e cmake | grep EXTRA_OECMAKE=
EXTRA_OECMAKE="
On Thu, 2019-05-23 at 12:06 +0100, nick83ola wrote:
> Hi Andre,
>
> in glibc freeaddrinfo is checking for null
>
> https://github.molgen.mpg.de/git-mirror/glibc/blob/20003c49884422da7ffbc459cdeee768a6fee07b/sysdeps/posix/getaddrinfo.c#L2663-L2674
>
> void
> freeaddrinfo (struct addrinfo *ai)
>
I think a neater solution might be to just aways set that when
handling PACKAGECONFIG in base.bbclass. Presumably the problem here
is that if PACKAGECONFIG isn't set then PACKAGECONFIG_CONFARGS is
never set, so lets ensure it is always set.
Ross
On Thu, 23 May 2019 at 14:55, Martin Jansa
This reverts commit 0c68c621d3b8690e3a346170006f75b032511a8a.
It should be upstreamed first
Secondly, musl and glibc are under different licenses and importing
code from one to other could taint it.
Signed-off-by: Khem Raj
---
...0001-unistd.h-Add-TEMP_FAILURE_RETRY.patch | 36
On Thu, May 23, 2019 at 2:39 PM Burton, Ross wrote:
>
> Why -help- and not -doc?
>
Because there are help files installed in 'help' directory and
displayed by a help browser.
Andreas
--
___
Openembedded-core mailing list
On Thu, 23 May 2019 at 11:34, Adrian Bunk wrote:
> systemd is fundamentally Linux-only and not portable to other kernels.
>
> systemd upstream is using glibc extensions not present in other
> C libraries.
>
> systemd upstream is accepting technically correct patches that help
> building with
On Thu, 23 May 2019 at 12:07, nick83ola wrote:
> that is why I put it into the musl specific patches.
> But ok I will repost it putting it in the generic SRC_URI.
Rule of thumb is to only apply patches in an override if they break
other cases outside of the override. In this case, it's just an
Devtool does not create commits; it's up to the developer and/or AUH.
I wouldn't necessarily want AUH to try to auto-explain the changes, as
it's worse than no explanation imo :)
Alex
On Thu, 23 May 2019 at 14:23, Burton, Ross wrote:
>
> Can you file a bug to get devtool to state that in the
musl > v1.1.21 changed the implementation of the freeaddrinfo() function
not allowing anymore to pass null pointers to it.
This was causing a segmentation fault in connman.
Signed-off-by: Nicola Lunghi
---
...-gweb-fix-segfault-with-musl-v1.1.21.patch | 34 +++
Hi Andre,
in glibc freeaddrinfo is checking for null
https://github.molgen.mpg.de/git-mirror/glibc/blob/20003c49884422da7ffbc459cdeee768a6fee07b/sysdeps/posix/getaddrinfo.c#L2663-L2674
void
freeaddrinfo (struct addrinfo *ai)
{
struct addrinfo *p;
while (ai != NULL)
{
p =
On Thu, May 23, 2019 at 02:56:11PM +0100, André Draszik wrote:
>...
> Also, in connman's gweb.c 'addr' is initialised to NULL. Again,
> NULL isn't anything 'returned by getaddrinfo()', so even in glibc
> it only works by pure luck.
>...
Not pure luck, free(NULL) is valid but passing random
Can you file a bug to get devtool to state that in the commit message?
Ross
On Thu, 23 May 2019 at 10:27, Alexander Kanavin wrote:
>
> Yes. If devtool drops them, then they are already in the code.
>
> Alex
>
> On Thu, 23 May 2019 at 03:30, Khem Raj wrote:
> >
> > On Wed, May 22, 2019 at 4:02
On Thu, 23 May 2019 at 07:48, Jiping Ma wrote:
> Issue: LIN10-5872
>
> Add IRQ statistic code in kvm/i8259.c.
>
> (LOCAL REV: NOT UPSTREAM) -- oe-core already included it.
Can you remove all these Wind River-specific lines and replace them
with a better commit message?
Ross
--
On 5/23/19 2:39 PM, Burton, Ross wrote:
> Why -help- and not -doc?
I was thinking it may make better sense to have a variable that DEFAULTS to
"*-locale-%s", but can be extended to have other locale based files.. Such as a
*-doc-%s, or *-man-%s, etc...
(We will -never- get all of the possible
On Thu, May 23, 2019 at 06:52:23AM -0700, Khem Raj wrote:
> On Wed, May 22, 2019 at 11:10 PM Adrian Bunk wrote:
>
> > On Wed, May 22, 2019 at 04:29:33PM -0700, Khem Raj wrote:
> > > On Tue, May 21, 2019 at 11:49 PM Adrian Bunk wrote:
> > > >
> > > > On Tue, May 21, 2019 at 06:32:28PM -0700,
On 5/23/19 11:12 AM, David Frey wrote:
On 5/22/2019 3:03 PM, Adrian Bunk wrote:
On Wed, May 22, 2019 at 02:55:41PM -0700, David Frey wrote:
udev is an optional dependency of bluez5, so use PACKAGECONFIG to allow
users to decide if they want udev support.
Signed-off-by: David Frey
---
On Thu, May 23, 2019 at 07:58:19AM -0700, Khem Raj wrote:
>...
> Secondly, musl and glibc are under different licenses and importing
> code from one to other could taint it.
>...
How many LOC are considered the legal limit in OE for that?
Many of the musl support patches (e.g. [1]) contain more
On 5/22/2019 3:03 PM, Adrian Bunk wrote:
> On Wed, May 22, 2019 at 02:55:41PM -0700, David Frey wrote:
>> udev is an optional dependency of bluez5, so use PACKAGECONFIG to allow
>> users to decide if they want udev support.
>>
>> Signed-off-by: David Frey
>> ---
>>
On 5/23/2019 9:30 AM, Joshua Watt wrote:
>
> On 5/23/19 11:12 AM, David Frey wrote:
>> On 5/22/2019 3:03 PM, Adrian Bunk wrote:
>>> On Wed, May 22, 2019 at 02:55:41PM -0700, David Frey wrote:
udev is an optional dependency of bluez5, so use PACKAGECONFIG to allow
users to decide if they
udev is an optional dependency of bluez5, so use PACKAGECONFIG to allow
users to decide if they want udev support.
Signed-off-by: David Frey
---
meta/recipes-connectivity/bluez5/bluez5.inc | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git
Recipe upgrade statistics:
* Failed (devtool error): 38
apt, 1.8.1, Aníbal Limón
openssh, 8.0p1, Armin Kuster
gettext, 0.20.1, Robert Yang
vulkan, 1.1.106.0, Ross Burton
sysklogd, 1.6, Chen Qi
ifupdown, 0.8.35, Anuj Mittal
On Thu, May 23, 2019 at 9:06 AM Adrian Bunk wrote:
>
> On Thu, May 23, 2019 at 07:58:19AM -0700, Khem Raj wrote:
> >...
> > Secondly, musl and glibc are under different licenses and importing
> > code from one to other could taint it.
> >...
>
> How many LOC are considered the legal limit in OE
This is mainly whitespace clean up, plus using the with statement when
writing files.
Signed-off-by: Peter Kjellerstedt
---
.../texinfo-dummy/template.py | 63 +--
1 file changed, 30 insertions(+), 33 deletions(-)
diff --git
The original version of template.py parses the arguments manually. This
fails when looking for the -E option if, e.g., an -I option is specified
without any space before its argument, and that argument contains the
letter 'E'.
A minor difference to the original version is that it parsed the
On 5/23/19 6:26 AM, Nicola Lunghi wrote:
musl > v1.1.21 changed the implementation of the freeaddrinfo() function
not allowing anymore to pass null pointers to it.
This was causing
Looks ok. We had similar issue with nfs-utils
https://patchwork.openembedded.org/patch/156772/
a
> -Original Message-
> From: openembedded-core-boun...@lists.openembedded.org core-boun...@lists.openembedded.org> On Behalf Of
> kai.k...@windriver.com
> Sent: den 23 maj 2019 10:26
> To: richard.pur...@linuxfoundation.org
> Cc: openembedded-core@lists.openembedded.org
> Subject:
On Thu, May 23, 2019 at 1:41 PM Peter Kjellerstedt <
peter.kjellerst...@axis.com> wrote:
> > -Original Message-
> > From: openembedded-core-boun...@lists.openembedded.org > core-boun...@lists.openembedded.org> On Behalf Of
> > kai.k...@windriver.com
> > Sent: den 23 maj 2019 10:26
> >
On 05/23/2019 08:22 PM, Burton, Ross wrote:
On Thu, 23 May 2019 at 11:34, Adrian Bunk wrote:
systemd is fundamentally Linux-only and not portable to other kernels.
systemd upstream is using glibc extensions not present in other
C libraries.
systemd upstream is accepting technically correct
Adds a PACKAGECONFIG option to enable the libdns_sd compatibility
library.
Signed-off-by: Joshua Watt
---
meta/recipes-connectivity/avahi/avahi.inc | 7 +++
1 file changed, 7 insertions(+)
diff --git a/meta/recipes-connectivity/avahi/avahi.inc
b/meta/recipes-connectivity/avahi/avahi.inc
On 5/22/19 11:53 PM, Nicolas Dechesne wrote:
hi,
not sure if it's been discussed reported before. but i ended up in a
situation where a package is "empty" (e.g. it's not created) even
though "files-in-packages.txt" in its buildhistory folder shows it
should contain a file.
I have a
On 5/23/19 7:17 PM, Chen Qi wrote:
systemd upstream only care about glibc. We made musl specific
patches so that systemd could work. But currently these patches
contain potential security issues.
So apply these patches only when the libc is musl.
LGTM
Signed-off-by: Chen Qi
---
On 5/23/19 6:45 PM, ChenQi wrote:
On 05/23/2019 08:22 PM, Burton, Ross wrote:
On Thu, 23 May 2019 at 11:34, Adrian Bunk wrote:
systemd is fundamentally Linux-only and not portable to other kernels.
systemd upstream is using glibc extensions not present in other
C libraries.
systemd
systemd upstream only care about glibc. We made musl specific
patches so that systemd could work. But currently these patches
contain potential security issues.
So apply these patches only when the libc is musl.
Signed-off-by: Chen Qi
---
meta/recipes-core/systemd/systemd_242.bb | 2 +-
1 file
*** BLURB HERE ***
The following changes since commit 06381a917e7e7648ace4746cbda55dcd138a64ec:
perl/modules: Add various missing ptest perl module dependencies (2019-05-22
07:24:52 +0100)
are available in the git repository at:
git://git.pokylinux.org/poky-contrib ChenQi/systemd-musl
Fixes build with musl where TEMP_FAILURE_RETRY is not provided
Signed-off-by: Khem Raj
---
...tional-TEMP_FAILURE_RETRY-macro-copy.patch | 36 +++
meta/recipes-connectivity/ofono/ofono_1.29.bb | 1 +
2 files changed, 37 insertions(+)
create mode 100644
On Thu, 2019-05-23 at 08:47 +0100, Nicola Lunghi wrote:
> musl > 2.21 changed the implementation of the freeaddrinfo() function
I think you mean 1.1.21 here and in the commit subject.
A.
> not allowing anymore to pass null pointers to it.
> This was causing a segmentation fault
systemd is fundamentally Linux-only and not portable to other kernels.
systemd upstream is using glibc extensions not present in other
C libraries.
systemd upstream is accepting technically correct patches that help
building with musl, but there is no interest upstream in keeping systemd
Signed-off-by: Alexander Kanavin
---
meta/recipes-extended/chkconfig/chkconfig_1.3.58.bb | 9 -
1 file changed, 4 insertions(+), 5 deletions(-)
diff --git a/meta/recipes-extended/chkconfig/chkconfig_1.3.58.bb
b/meta/recipes-extended/chkconfig/chkconfig_1.3.58.bb
index
Signed-off-by: Alexander Kanavin
---
meta/recipes-extended/chkconfig/chkconfig_1.3.58.bb | 1 +
1 file changed, 1 insertion(+)
diff --git a/meta/recipes-extended/chkconfig/chkconfig_1.3.58.bb
b/meta/recipes-extended/chkconfig/chkconfig_1.3.58.bb
index 788a37b6ed0..d21dd697042 100644
---
From: Kai Kang
v3:
* add the patch to make systemd as default init manager
* rename the .network with wired.network
* call to bb.utils.contains('OVERRIDES', ...) is not right. Install
wired.network unconditionally and remove in in do_install_append_qemuall()
BTW,
From: Kai Kang
Add a configure file for systemd.networkd to configure wired network
interfaces with dhcp. It works with common network interfaces such eth0
and eno1. And do not install it for qemu bsps.
Refer to
https://github.com/YoeDistro/meta-yoe/tree/master/recipes-core/systemd
[YOCTO
On Thu, 2019-05-23 at 08:47 +0100, Nicola Lunghi wrote:
> musl > 2.21 changed the implementation of the freeaddrinfo() function
> not allowing anymore to pass null pointers to it.
> This was causing a segmentation fault in connman.
> ---
> ...-gweb-fix-segfault-with-musl-v1.1.21.patch | 34
From: Kai Kang
Move configurations from local.conf.sample.extended to local.conf.sample
to make systemd as default init manager for poky.
[YOCTO #13031]
Signed-off-by: Kai Kang
---
meta-poky/conf/local.conf.sample | 9 +
meta-poky/conf/local.conf.sample.extended | 9
Yes. If devtool drops them, then they are already in the code.
Alex
On Thu, 23 May 2019 at 03:30, Khem Raj wrote:
>
> On Wed, May 22, 2019 at 4:02 AM Alexander Kanavin
> wrote:
> >
> > Signed-off-by: Alexander Kanavin
> > ---
> > ...UPWARD-only-if-its-defined-in-fenv.h.patch | 75
On Thu, 23 May 2019 at 00:24, Richard Purdie
wrote:
> Ultimately I'd say that all these checks should work with INSANE_SKIP.
> I was just wondering about this issue myself for the chkconfig
> autobuilder warnings...
chkconfig can be easily addressed by switching to git:// I think. Patch coming.
58 matches
Mail list logo
