On 7/27/21 6:03 PM, Masahiro Yamada wrote:
> On Wed, Jul 28, 2021 at 12:56 AM Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
>>
>> On Wed, Jul 28, 2021 at 12:39:24AM +0900, Masahiro Yamada wrote:
>>> Set the x bit to some scripts to make them directly executable.
>>>
>>> Especially, scripts/checkdeclares.pl is not hooked by anyone.
>>> It should be executable since it is tedious to type
>>> 'perl scripts/checkdeclares.pl'.>>>
>>> The original patch [1] set the x bit properly, but it was lost when
>>> it was merged as commit 21917bded72c ("scripts: a new script for
>>> checking duplicate struct declaration").
>>>
>>> [1] https://lore.kernel.org/lkml/20210401110943.1010796-1-wanjiabing@xxxxxxxx/
>>>
>>> Signed-off-by: Masahiro Yamada <masahiroy@xxxxxxxxxx>
>>> ---
>>>
>>> scripts/checkdeclares.pl | 0
>>> scripts/gcc-plugins/gen-random-seed.sh | 0
>>> scripts/syscallnr.sh | 0
>>> scripts/xen-hypercalls.sh | 0
>>> 4 files changed, 0 insertions(+), 0 deletions(-)
>>> mode change 100644 => 100755 scripts/checkdeclares.pl
>>> mode change 100644 => 100755 scripts/gcc-plugins/gen-random-seed.sh
>>> mode change 100644 => 100755 scripts/syscallnr.sh
>>> mode change 100644 => 100755 scripts/xen-hypercalls.sh
>>
>> Please no, as other tools (i.e. patch), can not set mode bits, and some
>> people still rely on patch in places.
>>
>> If these need to be called by other parts of the build, we should
>> execute them properly, not rely on the mode settings.
>>
>> thanks,
>>
>> greg k-h
>
>
> I believe tools should be executable.
>
> If the x bit were missing in scripts/checkpatch.pl
> for example, we would need to run 'perl scripts/checkpatch.pl'
> instead of 'scripts/checkpatch.pl'. That is annoying.
>
>
> Most of the scripts under the scripts/ directory
> are already executable, and we rely on that fact.
> Some of them are run directly, and I do not hear
> from anyone who complains about that.
>
>
...
>
>
> Even if it did not work on somebody's tools,
> the diff files are provided for bug-fix
> releases (for example, 5.13.x), not the entire source.
>
> Developers (except Andrew Morton) use git
> to merge patches like this, so I see no issue
> on changing the mode.
Sure, once the changes are in a git tree, it's not an
issue, so I don't see a problem with it.
Someone may have to go a few weeks without such a change,
but that's not a big deal.
--
~Randy
