Message ID | 20250923121422.24593-2-adam.blank.g@gmail.com |
---|---|
State | New |
Headers | show |
Series | Fix KBUILD_DEFCONFIG documentation | expand |
On Tue Sep 23, 2025 at 2:14 PM CEST, Adam Blank via lists.yoctoproject.org wrote: > The description of the relation between KBUILD_DEFCONFIG and > SRC_URI is reversed. In fact it is the SRC_URI provided > defconfig which will be dropped by the kernel-yocto class > if both are provided. > > Signed-off-by: Adam Blank <adam.blank.g@gmail.com> > --- > documentation/kernel-dev/common.rst | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/documentation/kernel-dev/common.rst b/documentation/kernel-dev/common.rst > index 654c4e0a0..c7b25765a 100644 > --- a/documentation/kernel-dev/common.rst > +++ b/documentation/kernel-dev/common.rst > @@ -655,7 +655,7 @@ Aside from modifying your kernel recipe and providing your own > :term:`SRC_URI` to use a ``defconfig`` other than your "in-tree" file (e.g. > a kernel's ``linux-``\ `machine`\ ``.inc`` file). In other words, if the > build system detects a statement that identifies an "out-of-tree" > -``defconfig`` file, that statement will override your > +``defconfig`` file, that statement will be overriden by your s/overriden/overridden/ Thanks, I can also verify that behavior so the logic is indeed reverted. In this case I think the previous sentence should also be removed: Aside from modifying your kernel recipe and providing your own ``defconfig`` file, you need to be sure no files or statements set :term:`SRC_URI` to use a ``defconfig`` other than your "in-tree" file (e.g. a kernel's ``linux-``\ `machine`\ ``.inc`` file). And instead only have one: If the build system detects a statement that identifies an "out-of-tree" ``defconfig`` file, that statement will be overriden by your :term:`KBUILD_DEFCONFIG` variable. What do you think? Antonin
Yes, that was also my first approach, but eventually I decided towards a minimal change. Shall I send a V2 of the patchset? Thanks, Adam On Thu, 25 Sept 2025 at 10:51, Antonin Godard <antonin.godard@bootlin.com> wrote: > On Tue Sep 23, 2025 at 2:14 PM CEST, Adam Blank via lists.yoctoproject.org > wrote: > > The description of the relation between KBUILD_DEFCONFIG and > > SRC_URI is reversed. In fact it is the SRC_URI provided > > defconfig which will be dropped by the kernel-yocto class > > if both are provided. > > > > Signed-off-by: Adam Blank <adam.blank.g@gmail.com> > > --- > > documentation/kernel-dev/common.rst | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/documentation/kernel-dev/common.rst > b/documentation/kernel-dev/common.rst > > index 654c4e0a0..c7b25765a 100644 > > --- a/documentation/kernel-dev/common.rst > > +++ b/documentation/kernel-dev/common.rst > > @@ -655,7 +655,7 @@ Aside from modifying your kernel recipe and > providing your own > > :term:`SRC_URI` to use a ``defconfig`` other than your "in-tree" file > (e.g. > > a kernel's ``linux-``\ `machine`\ ``.inc`` file). In other words, if the > > build system detects a statement that identifies an "out-of-tree" > > -``defconfig`` file, that statement will override your > > +``defconfig`` file, that statement will be overriden by your > > s/overriden/overridden/ > > Thanks, I can also verify that behavior so the logic is indeed reverted. > > In this case I think the previous sentence should also be removed: > > Aside from modifying your kernel recipe and providing your own > ``defconfig`` > file, you need to be sure no files or statements set :term:`SRC_URI` to > use a > ``defconfig`` other than your "in-tree" file (e.g. a kernel's ``linux-``\ > `machine`\ ``.inc`` file). > > And instead only have one: > > If the build system detects a statement that identifies an "out-of-tree" > ``defconfig`` file, that statement will be overriden by your > :term:`KBUILD_DEFCONFIG` variable. > > What do you think? > > Antonin > > -- > Antonin Godard, Bootlin > Embedded Linux and Kernel engineering > https://bootlin.com > >
On Thu, 25 Sep 2025, Antonin Godard via lists.yoctoproject.org wrote: ... snip ... > And instead only have one: > > If the build system detects a statement that identifies an > "out-of-tree" ``defconfig`` file, that statement will be overriden > by your :term:`KBUILD_DEFCONFIG` variable. I dislike the order of this information -- rather than saying A "will be overridden" by B, I prefer the more direct "B will take precedence over A". But that's just me. rday
On Thu Sep 25, 2025 at 11:09 AM CEST, Adam Blank wrote: > Yes, that was also my first approach, but eventually I decided towards a > minimal change. > Shall I send a V2 of the patchset? Yes, please. Feel free to rephrase to something better as suggested by Robert in this thread. Thanks! Antonin
diff --git a/documentation/kernel-dev/common.rst b/documentation/kernel-dev/common.rst index 654c4e0a0..c7b25765a 100644 --- a/documentation/kernel-dev/common.rst +++ b/documentation/kernel-dev/common.rst @@ -655,7 +655,7 @@ Aside from modifying your kernel recipe and providing your own :term:`SRC_URI` to use a ``defconfig`` other than your "in-tree" file (e.g. a kernel's ``linux-``\ `machine`\ ``.inc`` file). In other words, if the build system detects a statement that identifies an "out-of-tree" -``defconfig`` file, that statement will override your +``defconfig`` file, that statement will be overriden by your :term:`KBUILD_DEFCONFIG` variable. See the
The description of the relation between KBUILD_DEFCONFIG and SRC_URI is reversed. In fact it is the SRC_URI provided defconfig which will be dropped by the kernel-yocto class if both are provided. Signed-off-by: Adam Blank <adam.blank.g@gmail.com> --- documentation/kernel-dev/common.rst | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)