Re: [PATCHv4 1/9] zram: Preparation for multi-zcomp support

From: Sergey Senozhatsky <senozhatsky@chromium.org>
To: Minchan Kim <minchan@kernel.org>
Cc: Sergey Senozhatsky <senozhatsky@chromium.org>,
	Andrew Morton <akpm@linux-foundation.org>,
	Nitin Gupta <ngupta@vflare.org>,
	linux-kernel@vger.kernel.org, linux-mm@kvack.org
Subject: Re: [PATCHv4 1/9] zram: Preparation for multi-zcomp support
Date: Thu, 3 Nov 2022 11:40:31 +0900	[thread overview]
Message-ID: <Y2MqH+EdN7wX1vcl@google.com> (raw)
In-Reply-To: <Y2LPUlircv6a74mJ@google.com>

On (22/11/02 13:13), Minchan Kim wrote:
[..]
> >  
> > +static void zram_destroy_comps(struct zram *zram)
> > +{
> > +	u32 idx;
> > +
> > +	for (idx = 0; idx < ZRAM_MAX_ZCOMPS; idx++) {
> > +		struct zcomp *comp = zram->comps[idx];
> > +
> > +		zram->comps[idx] = NULL;
> > +		if (IS_ERR_OR_NULL(comp))
> 
> nit:
> 
> Why don't you use just NULL check? I don't see any error setting
> for zram->comps(Maybe later patch? Will keep check)?

A defense measure. zcomp_create() returns err pointer, so here
I check for err and nil just in case (if somehow someday we
accidentally have err values stored in comps). It's cheap and
safer than NULL.