[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <87il4a9n37.fsf@meer.lwn.net>
Date: Wed, 03 Jan 2024 13:30:20 -0700
From: Jonathan Corbet <corbet@....net>
To: Randy Dunlap <rdunlap@...radead.org>, linux-kernel@...r.kernel.org
Cc: Randy Dunlap <rdunlap@...radead.org>, Thomas Gleixner
<tglx@...utronix.de>, x86@...nel.org, linux-doc@...r.kernel.org
Subject: Re: [RFC PATCH] kernel-doc: handle X86 DEFINE_IDTENTRY() variants
Randy Dunlap <rdunlap@...radead.org> writes:
> Teach scripts/kernel-doc to handle the various DEFINE_IDTENTRY*() flavors.
>
> This corrects 2 kernel-doc warnings:
>
> arch/x86/entry/common.c:211: warning: expecting prototype for int80_emulation(). Prototype was for DEFINE_IDTENTRY_RAW() instead
>
> arch/x86/kernel/apic/apic.c:2170: warning: expecting prototype for spurious_interrupt(). Prototype was for DEFINE_IDTENTRY_IRQ() instead
>
> The script uses 'uname -m' to determine if it is running on i386 or x86_64
> or something else. It also uses "ARCH=<arch>" in the environment variables
> to allow for overriding the processed ARCH.
>
> Alternatively, we could remove the "/**" kernel-doc markers from those
> 2 functions. There are 60 uses of DEFINE_IDTENTRY*() that I see and
> only 2 of them have kernel-doc comments.
So I feel like I'm missing something here; the docs build should be the
same regardless of the architecture it's running on, right? So why do
we need architecture checks in kernel-doc?
Honestly, it might be better to just remove the kerneldoc comments
rather than add this much more complexity.
Thanks,
jon
Powered by blists - more mailing lists