From: Andrew Morton <akpm@digeo.com>
To: "Martin J. Bligh" <mbligh@aracnet.com>
Cc: linux-kernel <linux-kernel@vger.kernel.org>,
linux-mm mailing list <linux-mm@kvack.org>
Subject: Re: ZONE_NORMAL exhaustion (dcache slab)
Date: Mon, 21 Oct 2002 14:49:35 -0700 [thread overview]
Message-ID: <3DB4766F.D3AB15B9@digeo.com> (raw)
In-Reply-To: 309670000.1035236015@flay
"Martin J. Bligh" wrote:
>
> >> Nope, kept OOMing and killing everything .
> >
> > Something broke.
>
> Even I worked that out ;-)
Well I'm feeling especially helpful today.
> > Blockdevices only use ZONE_NORMAL for their pagecache. That cat will
> > selectively put pressure on the normal zone (and DMA zone, of course).
>
> Ah, I recall that now. That's fundamentally screwed.
When filesystems want to access metadata, they will typically read
a block into a buffer_head and access the memory directly.
mnm:/usr/src/25> grep -rI b_data fs | wc -l
844
That's a lot of kmaps need adding.
So we constrain blockdev->bd_inode->i_mapping->gfp_mask so that
the blockdev's pagecache memory is always in the direct-addressed
region.
It would be possible to fix on a per-fs basis - teach a filesystem
to kmap bh->b_page appropriately and then set __GFP_HIGHMEM in the
blockdev's gfp_mask.
But it doesn't seem to cause a lot of trouble in practice.
WARNING: multiple messages have this Message-ID (diff)
From: Andrew Morton <akpm@digeo.com>
To: "Martin J. Bligh" <mbligh@aracnet.com>
Cc: linux-kernel <linux-kernel@vger.kernel.org>,
linux-mm mailing list <linux-mm@kvack.org>
Subject: Re: ZONE_NORMAL exhaustion (dcache slab)
Date: Mon, 21 Oct 2002 14:49:35 -0700 [thread overview]
Message-ID: <3DB4766F.D3AB15B9@digeo.com> (raw)
In-Reply-To: 309670000.1035236015@flay
"Martin J. Bligh" wrote:
>
> >> Nope, kept OOMing and killing everything .
> >
> > Something broke.
>
> Even I worked that out ;-)
Well I'm feeling especially helpful today.
> > Blockdevices only use ZONE_NORMAL for their pagecache. That cat will
> > selectively put pressure on the normal zone (and DMA zone, of course).
>
> Ah, I recall that now. That's fundamentally screwed.
When filesystems want to access metadata, they will typically read
a block into a buffer_head and access the memory directly.
mnm:/usr/src/25> grep -rI b_data fs | wc -l
844
That's a lot of kmaps need adding.
So we constrain blockdev->bd_inode->i_mapping->gfp_mask so that
the blockdev's pagecache memory is always in the direct-addressed
region.
It would be possible to fix on a per-fs basis - teach a filesystem
to kmap bh->b_page appropriately and then set __GFP_HIGHMEM in the
blockdev's gfp_mask.
But it doesn't seem to cause a lot of trouble in practice.
--
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/
next prev parent reply other threads:[~2002-10-21 21:43 UTC|newest]
Thread overview: 41+ messages / expand[flat|nested] mbox.gz Atom feed top
2002-10-21 20:40 ZONE_NORMAL exhaustion (dcache slab) Martin J. Bligh
2002-10-21 20:40 ` Martin J. Bligh
2002-10-21 21:13 ` Andrew Morton
2002-10-21 21:13 ` Andrew Morton
2002-10-21 21:16 ` Martin J. Bligh
2002-10-21 21:16 ` Martin J. Bligh
2002-10-21 21:33 ` Andrew Morton
2002-10-21 21:33 ` Andrew Morton
2002-10-21 21:33 ` Martin J. Bligh
2002-10-21 21:33 ` Martin J. Bligh
2002-10-21 21:49 ` Andrew Morton [this message]
2002-10-21 21:49 ` Andrew Morton
2002-10-21 22:30 ` Rik van Riel
2002-10-21 22:30 ` Rik van Riel
2002-10-21 22:53 ` Andrew Morton
2002-10-21 22:53 ` Andrew Morton
2002-10-22 0:31 ` Martin J. Bligh
2002-10-22 0:31 ` Martin J. Bligh
2002-10-22 3:39 ` Andrew Morton
2002-10-22 3:39 ` Andrew Morton
2002-10-22 3:53 ` Martin J. Bligh
2002-10-22 3:53 ` Martin J. Bligh
2002-10-22 4:20 ` Andrew Morton
2002-10-22 4:20 ` Andrew Morton
2002-10-22 5:49 ` Martin J. Bligh
2002-10-22 5:49 ` Martin J. Bligh
2002-10-22 6:21 ` Andrew Morton
2002-10-22 6:21 ` Andrew Morton
2002-10-22 16:13 ` Martin J. Bligh
2002-10-22 16:13 ` Martin J. Bligh
2002-10-24 11:35 ` Ed Tomlinson
2002-10-24 11:35 ` Ed Tomlinson
2002-10-24 14:28 ` Martin J. Bligh
2002-10-24 14:28 ` Martin J. Bligh
2002-10-22 16:33 ` Rik van Riel
2002-10-22 16:33 ` Rik van Riel
2002-10-22 16:54 ` ARP address resolving in kernel Steffen Persvold
2002-10-22 17:05 ` ZONE_NORMAL exhaustion (dcache slab) Andrew Morton
2002-10-22 17:05 ` Andrew Morton
2002-10-22 16:21 ` Dipankar Sarma
2002-10-22 16:21 ` Dipankar Sarma
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=3DB4766F.D3AB15B9@digeo.com \
--to=akpm@digeo.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mbligh@aracnet.com \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.