[v2,04/13] initramfs: rename gen_initramfs_list.sh to gen_initramfs.sh
diff mbox series

Message ID 20200104150238.19834-5-masahiroy@kernel.org
State New
Headers show
Series
  • initramfs: a lot of cleanups
Related show

Commit Message

Masahiro Yamada Jan. 4, 2020, 3:02 p.m. UTC
The comments in usr/Makefile wrongly refer to the script name (twice).

Line 37:
    # The dependency list is generated by gen_initramfs.sh -l

Line 54:
    # 4) Arguments to gen_initramfs.sh changes

There does not exist such a script.

I was going to fix the comments, but after some consideration, I thought
"gen_initramfs.sh" would be more suitable than "gen_initramfs_list.sh"
because it generates an initramfs image in the common usage.

The script generates a list that can be fed to gen_init_cpio only when
it is directly run without -o or -l option.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
---

Changes in v2: None

 usr/Makefile                                    | 2 +-
 usr/{gen_initramfs_list.sh => gen_initramfs.sh} | 0
 2 files changed, 1 insertion(+), 1 deletion(-)
 rename usr/{gen_initramfs_list.sh => gen_initramfs.sh} (100%)

Comments

Greg Thelen Jan. 6, 2020, 7:43 a.m. UTC | #1
Masahiro Yamada <masahiroy@kernel.org> wrote:

> The comments in usr/Makefile wrongly refer to the script name (twice).
>
> Line 37:
>     # The dependency list is generated by gen_initramfs.sh -l
>
> Line 54:
>     # 4) Arguments to gen_initramfs.sh changes
>
> There does not exist such a script.
>
> I was going to fix the comments, but after some consideration, I thought
> "gen_initramfs.sh" would be more suitable than "gen_initramfs_list.sh"
> because it generates an initramfs image in the common usage.
>
> The script generates a list that can be fed to gen_init_cpio only when
> it is directly run without -o or -l option.
>
> Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
> ---
>
> Changes in v2: None
>
>  usr/Makefile                                    | 2 +-
>  usr/{gen_initramfs_list.sh => gen_initramfs.sh} | 0
>  2 files changed, 1 insertion(+), 1 deletion(-)
>  rename usr/{gen_initramfs_list.sh => gen_initramfs.sh} (100%)

Will this break klibc?  It might have a ref to the old name.
https://git.kernel.org/pub/scm/libs/klibc/klibc.git/tree/usr/Kbuild#n55

> diff --git a/usr/Makefile b/usr/Makefile
> index 55c942da01cd..e44a66b8c051 100644
> --- a/usr/Makefile
> +++ b/usr/Makefile
> @@ -24,7 +24,7 @@ $(obj)/initramfs_data.o: $(obj)/$(datafile_y) FORCE
>  # Generate the initramfs cpio archive
>  
>  hostprogs-y := gen_init_cpio
> -initramfs   := $(CONFIG_SHELL) $(srctree)/$(src)/gen_initramfs_list.sh
> +initramfs   := $(CONFIG_SHELL) $(srctree)/$(src)/gen_initramfs.sh
>  ramfs-input := $(if $(filter-out "",$(CONFIG_INITRAMFS_SOURCE)), \
>  			$(shell echo $(CONFIG_INITRAMFS_SOURCE)),-d)
>  ramfs-args  := \
> diff --git a/usr/gen_initramfs_list.sh b/usr/gen_initramfs.sh
> similarity index 100%
> rename from usr/gen_initramfs_list.sh
> rename to usr/gen_initramfs.sh
Masahiro Yamada Jan. 6, 2020, 10:26 a.m. UTC | #2
(+CC Ben Hutchings, H. Peter Anvin)

In my understanding, the klibc build system is standalone.
So, the change in Linux kernel does not affect klibc at all.
Only the depending part is UAPI headers (make headers_install).

So, this patch
(https://lore.kernel.org/patchwork/patch/1175336/)
should be OK.

Please correct me if I am wrong.


On Mon, Jan 6, 2020 at 4:43 PM Greg Thelen <gthelen@google.com> wrote:
>
> Masahiro Yamada <masahiroy@kernel.org> wrote:
>
> > The comments in usr/Makefile wrongly refer to the script name (twice).
> >
> > Line 37:
> >     # The dependency list is generated by gen_initramfs.sh -l
> >
> > Line 54:
> >     # 4) Arguments to gen_initramfs.sh changes
> >
> > There does not exist such a script.
> >
> > I was going to fix the comments, but after some consideration, I thought
> > "gen_initramfs.sh" would be more suitable than "gen_initramfs_list.sh"
> > because it generates an initramfs image in the common usage.
> >
> > The script generates a list that can be fed to gen_init_cpio only when
> > it is directly run without -o or -l option.
> >
> > Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
> > ---
> >
> > Changes in v2: None
> >
> >  usr/Makefile                                    | 2 +-
> >  usr/{gen_initramfs_list.sh => gen_initramfs.sh} | 0
> >  2 files changed, 1 insertion(+), 1 deletion(-)
> >  rename usr/{gen_initramfs_list.sh => gen_initramfs.sh} (100%)
>
> Will this break klibc?  It might have a ref to the old name.
> https://git.kernel.org/pub/scm/libs/klibc/klibc.git/tree/usr/Kbuild#n55

I do not think so.

As I stated above, the klibc build system is independent of
any script in the Linux kernel.

The klibc Makefile refers to
scripts/gen_initramfs_list.sh, which does not exist.

My path is renaming
usr/gen_initramfs_list.sh to usr/gen_initramfs.sh


If the renaming had been problematic for klibc,
commit f6f57a46435d7253a52a1a07a58183678ad266a0
("initramfs: move gen_initramfs_list.sh from scripts/ to usr/")
would have already caused a problem.



Ben, Hans,
Is usr/Kbuild in klibc used?
If it is not used, is it better to delete it to avoid confusion?


Masahiro Yamada




> > diff --git a/usr/Makefile b/usr/Makefile
> > index 55c942da01cd..e44a66b8c051 100644
> > --- a/usr/Makefile
> > +++ b/usr/Makefile
> > @@ -24,7 +24,7 @@ $(obj)/initramfs_data.o: $(obj)/$(datafile_y) FORCE
> >  # Generate the initramfs cpio archive
> >
> >  hostprogs-y := gen_init_cpio
> > -initramfs   := $(CONFIG_SHELL) $(srctree)/$(src)/gen_initramfs_list.sh
> > +initramfs   := $(CONFIG_SHELL) $(srctree)/$(src)/gen_initramfs.sh
> >  ramfs-input := $(if $(filter-out "",$(CONFIG_INITRAMFS_SOURCE)), \
> >                       $(shell echo $(CONFIG_INITRAMFS_SOURCE)),-d)
> >  ramfs-args  := \
> > diff --git a/usr/gen_initramfs_list.sh b/usr/gen_initramfs.sh
> > similarity index 100%
> > rename from usr/gen_initramfs_list.sh
> > rename to usr/gen_initramfs.sh
Ben Hutchings Jan. 7, 2020, 3:02 a.m. UTC | #3
On Mon, 2020-01-06 at 19:26 +0900, Masahiro Yamada wrote:
> (+CC Ben Hutchings, H. Peter Anvin)
> 
> In my understanding, the klibc build system is standalone.
> So, the change in Linux kernel does not affect klibc at all.
> Only the depending part is UAPI headers (make headers_install).
> 
> So, this patch
> (https://lore.kernel.org/patchwork/patch/1175336/)
> should be OK.
> 
> Please correct me if I am wrong.

I think you're right.

[...]
> I do not think so.
> 
> As I stated above, the klibc build system is independent of
> any script in the Linux kernel.
> 
> The klibc Makefile refers to
> scripts/gen_initramfs_list.sh, which does not exist.
[...]

Right.  I believe the original intent was to include klibc in the
kernel tree, so this would have existed because of that.

There's no sign of that in the current klibc.git, but I think it might
have gone through "git filter-branch" at some point after it was
decided not to include it upstream.  In any case, this was already
broken for as long as klibc has been separate.

Ben.
H. Peter Anvin Jan. 7, 2020, 3:21 a.m. UTC | #4
On January 6, 2020 2:26:22 AM PST, Masahiro Yamada <masahiroy@kernel.org> wrote:
>(+CC Ben Hutchings, H. Peter Anvin)
>
>In my understanding, the klibc build system is standalone.
>So, the change in Linux kernel does not affect klibc at all.
>Only the depending part is UAPI headers (make headers_install).
>
>So, this patch
>(https://lore.kernel.org/patchwork/patch/1175336/)
>should be OK.
>
>Please correct me if I am wrong.
>
>
>On Mon, Jan 6, 2020 at 4:43 PM Greg Thelen <gthelen@google.com> wrote:
>>
>> Masahiro Yamada <masahiroy@kernel.org> wrote:
>>
>> > The comments in usr/Makefile wrongly refer to the script name
>(twice).
>> >
>> > Line 37:
>> >     # The dependency list is generated by gen_initramfs.sh -l
>> >
>> > Line 54:
>> >     # 4) Arguments to gen_initramfs.sh changes
>> >
>> > There does not exist such a script.
>> >
>> > I was going to fix the comments, but after some consideration, I
>thought
>> > "gen_initramfs.sh" would be more suitable than
>"gen_initramfs_list.sh"
>> > because it generates an initramfs image in the common usage.
>> >
>> > The script generates a list that can be fed to gen_init_cpio only
>when
>> > it is directly run without -o or -l option.
>> >
>> > Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
>> > ---
>> >
>> > Changes in v2: None
>> >
>> >  usr/Makefile                                    | 2 +-
>> >  usr/{gen_initramfs_list.sh => gen_initramfs.sh} | 0
>> >  2 files changed, 1 insertion(+), 1 deletion(-)
>> >  rename usr/{gen_initramfs_list.sh => gen_initramfs.sh} (100%)
>>
>> Will this break klibc?  It might have a ref to the old name.
>>
>https://git.kernel.org/pub/scm/libs/klibc/klibc.git/tree/usr/Kbuild#n55
>
>I do not think so.
>
>As I stated above, the klibc build system is independent of
>any script in the Linux kernel.
>
>The klibc Makefile refers to
>scripts/gen_initramfs_list.sh, which does not exist.
>
>My path is renaming
>usr/gen_initramfs_list.sh to usr/gen_initramfs.sh
>
>
>If the renaming had been problematic for klibc,
>commit f6f57a46435d7253a52a1a07a58183678ad266a0
>("initramfs: move gen_initramfs_list.sh from scripts/ to usr/")
>would have already caused a problem.
>
>
>
>Ben, Hans,
>Is usr/Kbuild in klibc used?
>If it is not used, is it better to delete it to avoid confusion?
>
>
>Masahiro Yamada
>
>
>
>
>> > diff --git a/usr/Makefile b/usr/Makefile
>> > index 55c942da01cd..e44a66b8c051 100644
>> > --- a/usr/Makefile
>> > +++ b/usr/Makefile
>> > @@ -24,7 +24,7 @@ $(obj)/initramfs_data.o: $(obj)/$(datafile_y)
>FORCE
>> >  # Generate the initramfs cpio archive
>> >
>> >  hostprogs-y := gen_init_cpio
>> > -initramfs   := $(CONFIG_SHELL)
>$(srctree)/$(src)/gen_initramfs_list.sh
>> > +initramfs   := $(CONFIG_SHELL) $(srctree)/$(src)/gen_initramfs.sh
>> >  ramfs-input := $(if $(filter-out "",$(CONFIG_INITRAMFS_SOURCE)), \
>> >                       $(shell echo $(CONFIG_INITRAMFS_SOURCE)),-d)
>> >  ramfs-args  := \
>> > diff --git a/usr/gen_initramfs_list.sh b/usr/gen_initramfs.sh
>> > similarity index 100%
>> > rename from usr/gen_initramfs_list.sh
>> > rename to usr/gen_initramfs.sh

That is correct.

The klibc integration work was mainly fine in an actual kernel tree, which pulled the klibc tree.

It has obviously bitrotted somewhat, but it wouldn't be all that hard to resurrect it if there is ever interest.

Patch
diff mbox series

diff --git a/usr/Makefile b/usr/Makefile
index 55c942da01cd..e44a66b8c051 100644
--- a/usr/Makefile
+++ b/usr/Makefile
@@ -24,7 +24,7 @@  $(obj)/initramfs_data.o: $(obj)/$(datafile_y) FORCE
 # Generate the initramfs cpio archive
 
 hostprogs-y := gen_init_cpio
-initramfs   := $(CONFIG_SHELL) $(srctree)/$(src)/gen_initramfs_list.sh
+initramfs   := $(CONFIG_SHELL) $(srctree)/$(src)/gen_initramfs.sh
 ramfs-input := $(if $(filter-out "",$(CONFIG_INITRAMFS_SOURCE)), \
 			$(shell echo $(CONFIG_INITRAMFS_SOURCE)),-d)
 ramfs-args  := \
diff --git a/usr/gen_initramfs_list.sh b/usr/gen_initramfs.sh
similarity index 100%
rename from usr/gen_initramfs_list.sh
rename to usr/gen_initramfs.sh