[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20241203102603.GA21636@noisy.programming.kicks-ass.net>
Date: Tue, 3 Dec 2024 11:26:03 +0100
From: Peter Zijlstra <peterz@...radead.org>
To: Masahiro Yamada <masahiroy@...nel.org>
Cc: Linus Torvalds <torvalds@...ux-foundation.org>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
Matthias Maennich <maennich@...gle.com>,
linux-kbuild@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 0/3] module: fixups after module namespace conversion
On Tue, Dec 03, 2024 at 07:21:04PM +0900, Masahiro Yamada wrote:
>
> [1/3]
> Commit cdd30ebb1b9f added double-quotes to enclose the variable in the
> semantic patch, scripts/coccinelle/misc/add_namespace.cocci.
> Since then, 'make nsdepds' always changes the missing import to silly code:
> MODULE_IMPORT_NS("ns");
>
> [2/3]
> MODULE_IMPORT_NS("") is misleading in the context of documentation
> about MODULE_IMPORT_NS() tags in general.
>
> [3/3]
> Convert DEFAULT_SYMBOL_NAMESPACE from a macro expansion to a string
> literal. There is no good reason to keep inconsistency in the default
> case.
>
>
>
> Masahiro Yamada (3):
> scripts/nsdeps: get 'make nsdeps' working again
> doc: module: revert misconversions for MODULE_IMPORT_NS()
> module: Convert default symbol namespace to string literal
Thanks, I totally missed the DEFAULT_SYMBOL_NAMESPACE thing.
And my coccinelle foo is non-existent, I would've never spotted that.
Powered by blists - more mailing lists