Re: [PATCH] drivers: pnp: proc.c: Handle errors while attaching devices

From: "Valdis Klētnieks" <valdis.kletnieks@vt.edu>
To: Leon Romanovsky <leon@kernel.org>
Cc: bkkarthik <bkkarthik@pesu.pes.edu>,
	Anupama K Patil <anupamakpatil123@gmail.com>,
	Jaroslav Kysela <perex@perex.cz>,
	"Rafael J. Wysocki" <rafael.j.wysocki@intel.com>,
	linux-acpi@vger.kernel.org, linux-kernel@vger.kernel.org,
	skhan@linuxfoundation.org, gregkh@linuxfoundation.org,
	kernelnewbies@kernelnewbies.org
Subject: Re: [PATCH] drivers: pnp: proc.c: Handle errors while attaching devices
Date: Thu, 29 Apr 2021 00:31:13 -0400	[thread overview]
Message-ID: <210271.1619670673@turing-police> (raw)
In-Reply-To: <YIeSc1qePhuQ1XRK@unreal>

[-- Attachment #1: Type: text/plain, Size: 904 bytes --]

On Tue, 27 Apr 2021 07:26:27 +0300, Leon Romanovsky said:
> On Mon, Apr 26, 2021 at 11:20:32PM +0530, bkkarthik wrote:
> > These were only intended for a clean-up job, the idea of this function came from how PCI handles procfs.
> > Maybe those should be changed?
>
> Probably, the CONFIG_PROC_FS around pci_proc_*() is not needed too.

Will that actually build correctly if it's an embedded system or something build with
CONFIG_PROC_FS=n?  I'd expect that to die a horrid death while linking vmlinx due
to stuff inside that #ifdef calling symbols only present with PROC_FS=m/y.

In general, inline ifdef's are frowned upon, so if you come across one in the kernel
source, that's probably a *big* hint that either (a) refactoring the code to eliminate
an inline ifdef was just too ugly to be allowed to live or (b) you *have* to put a guard
around it because you're guaranteed a build failure otherwise.

[-- Attachment #2: Type: application/pgp-signature, Size: 832 bytes --]