[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAGM2reZsZVhhg2=dQZf6D-NmPTFRN-_95+s61pC7Axz5G5mkMQ@mail.gmail.com>
Date: Thu, 28 Jun 2018 08:12:04 -0400
From: Pavel Tatashin <pasha.tatashin@...cle.com>
To: osalvador@...hadventures.net
Cc: bhe@...hat.com, LKML <linux-kernel@...r.kernel.org>,
Andrew Morton <akpm@...ux-foundation.org>,
dave.hansen@...el.com, pagupta@...hat.com,
Linux Memory Management List <linux-mm@...ck.org>,
kirill.shutemov@...ux.intel.com
Subject: Re: [PATCH v6 4/5] mm/sparse: Optimize memmap allocation during sparse_init()
> > + if (nr_consumed_maps >= nr_present_sections) {
> > + pr_err("nr_consumed_maps goes beyond nr_present_sections\n");
> > + break;
> > + }
>
> Hi Baoquan,
>
> I am sure I am missing something here, but is this check really needed?
>
> I mean, for_each_present_section_nr() only returns the section nr if the section
> has been marked as SECTION_MARKED_PRESENT.
> That happens in memory_present(), where now we also increment nr_present_sections whenever
> we find a present section.
>
> So, for_each_present_section_nr() should return the same nr of section as nr_present_sections.
> Since we only increment nr_consumed_maps once in the loop, I am not so sure we can
> go beyond nr_present_sections.
>
> Did I overlook something?
You did not, this is basically a safety check. A BUG_ON() would be
better here. As, this something that should really not happening, and
would mean a bug in the current project.
Powered by blists - more mailing lists