[oe] [meta-qt5][PATCH 2/2] qtbase: Add packageconfigs for renameat2 and getentropy use

[Khem Raj]

On Fri, Feb 23, 2018 at 12:00 AM, Samuli Piippo <samuli.piippo at gmail.com> wrote:
> renameat2 is a system call without a libc wrapper, so the configure
> test will always fail. The system call will still be used if that is
> defined, so PACKAGECONFIG doesn't make sense for it.
>

thats what I was thinking that testing kernel headers for the syscall
may not be sufficient. We might have to pass minimum supported kernel
version to make this decision to enable/disable these calls.

> On 23 February 2018 at 04:17, Khem Raj <raj.khem at gmail.com> wrote:
>> On Thu, Feb 22, 2018 at 5:27 PM, Martin Jansa <martin.jansa at gmail.com> wrote:
>>> On Wed, Feb 21, 2018 at 08:29:40PM -0800, Khem Raj wrote:
>>>> These features depend on underlying syscall support in kernel
>>>> and if older kernels are in use, then we can have a knob to
>>>> turn them off.
>>>>
>>>> Signed-off-by: Khem Raj <raj.khem at gmail.com>
>>>> ---
>>>>  recipes-qt/qt5/qtbase_git.bb | 6 +++++-
>>>>  1 file changed, 5 insertions(+), 1 deletion(-)
>>>>
>>>> diff --git a/recipes-qt/qt5/qtbase_git.bb b/recipes-qt/qt5/qtbase_git.bb
>>>> index e495b8c..843648f 100644
>>>> --- a/recipes-qt/qt5/qtbase_git.bb
>>>> +++ b/recipes-qt/qt5/qtbase_git.bb
>>>> @@ -71,7 +71,7 @@ PACKAGECONFIG_DISTRO ?= ""
>>>>  PACKAGECONFIG_RELEASE ?= "release"
>>>>  # This is in qt5.inc, because qtwebkit-examples are using it to enable ca-certificates dependency
>>>>  # PACKAGECONFIG_OPENSSL ?= "openssl"
>>>> -PACKAGECONFIG_DEFAULT ?= "dbus udev evdev widgets tools libs freetype tests"
>>>> +PACKAGECONFIG_DEFAULT ?= "dbus udev evdev widgets tools libs freetype tests renameat2 getentropy"
>>>
>>> Should renameat2 be enabled by default?
>>>
>>> Either the test for it is broken in 5.11 or it's not available in
>>> default setup.
>>>
>>
>> Its ok to keep them enabled by default. but I think this is a bug that
>> should be reported
>> to upstream QT, if the feature is knobbale then it should have worked.
>> --
>> _______________________________________________
>> Openembedded-devel mailing list
>> Openembedded-devel at lists.openembedded.org
>> http://lists.openembedded.org/mailman/listinfo/openembedded-devel