From: Baoquan He <bhe@redhat.com>
To: Dave Hansen <dave.hansen@linux.intel.com>
Cc: "Kirill A. Shutemov" <kirill@shutemov.name>,
linux-kernel@vger.kernel.org, linux-mm@kvack.org,
akpm@linux-foundation.org, kirill.shutemov@linux.intel.com,
mhocko@suse.com, tglx@linutronix.de, douly.fnst@cn.fujitsu.com
Subject: Re: [PATCH 2/2] mm/sparse.c: Add nr_present_sections to change the mem_map allocation
Date: Thu, 1 Feb 2018 22:19:34 +0800 [thread overview]
Message-ID: <20180201141934.GC1770@localhost.localdomain> (raw)
In-Reply-To: <6def8374-2de2-a30c-69ff-2a49fb57dc9a@linux.intel.com>
On 02/01/18 at 05:49am, Dave Hansen wrote:
> On 02/01/2018 02:16 AM, Kirill A. Shutemov wrote:
> > On Thu, Feb 01, 2018 at 03:19:56PM +0800, Baoquan He wrote:
> >> In sparse_init(), we allocate usemap_map and map_map which are pointer
> >> array with the size of NR_MEM_SECTIONS. The memory consumption can be
> >> ignorable in 4-level paging mode. While in 5-level paging, this costs
> >> much memory, 512M. Kdump kernel even can't boot up with a normal
> >> 'crashkernel=' setting.
> >>
> >> Here add a new variable to record the number of present sections. Let's
> >> allocate the usemap_map and map_map with the size of nr_present_sections.
> >> We only need to make sure that for the ith present section, usemap_map[i]
> >> and map_map[i] store its usemap and mem_map separately.
> >>
> >> This change can save much memory on most of systems. Anytime, we should
> >> avoid to define array or allocate memory with the size of NR_MEM_SECTIONS.
> > That's very desirable outcome. But I don't know much about sparsemem.
>
> ... with the downside being that we can no longer hot-add memory that
> was not part of the original, present sections.
>
> Is that OK?
Thanks for looking into this, Dave!
I suppose these functions changed here are only called during system
bootup, namely in paging_init(). Hot-add memory goes in a different
path, __add_section() -> sparse_add_one_section(), different called
functions.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2018-02-01 14:19 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-02-01 7:19 [PATCH 0/2] Optimize the code of mem_map allocation in Baoquan He
2018-02-01 7:19 ` [PATCH 1/2] mm/sparsemem: Defer the ms->section_mem_map clearing a little later Baoquan He
2018-02-01 14:15 ` Dave Hansen
2018-02-01 14:38 ` Baoquan He
2018-02-01 7:19 ` [PATCH 2/2] mm/sparse.c: Add nr_present_sections to change the mem_map allocation Baoquan He
2018-02-01 10:16 ` Kirill A. Shutemov
2018-02-01 13:49 ` Dave Hansen
2018-02-01 14:19 ` Baoquan He [this message]
2018-02-01 14:23 ` Dave Hansen
2018-02-01 14:33 ` Baoquan He
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20180201141934.GC1770@localhost.localdomain \
--to=bhe@redhat.com \
--cc=akpm@linux-foundation.org \
--cc=dave.hansen@linux.intel.com \
--cc=douly.fnst@cn.fujitsu.com \
--cc=kirill.shutemov@linux.intel.com \
--cc=kirill@shutemov.name \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mhocko@suse.com \
--cc=tglx@linutronix.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).