mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: "Drokin, Oleg" <oleg.drokin@intel.com>
To: Dan Carpenter <dan.carpenter@oracle.com>
Cc: "Simmons, James A." <simmonsja@ornl.gov>,
	Julia Lawall <Julia.Lawall@lip6.fr>,
	"devel@driverdev.osuosl.org" <devel@driverdev.osuosl.org>,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	"kernel-janitors@vger.kernel.org"
	<kernel-janitors@vger.kernel.org>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"HPDD-discuss@lists.01.org" <HPDD-discuss@ml01.01.org>
Subject: Re: [HPDD-discuss] [PATCH 2/11] Staging: lustre: fld: Use kzalloc and kfree
Date: Fri, 1 May 2015 20:12:44 +0000	[thread overview]
Message-ID: <9C4D3E11-27BE-42F1-8AC7-9D8DC5D31F93@intel.com> (raw)
In-Reply-To: <20150501200221.GF14154@mwanda>


On May 1, 2015, at 4:02 PM, Dan Carpenter wrote:

> We are hopefully going to get rid of OBD_ALLOC_LARGE() as well, though.
> 
> It's simple enough to write a function:
> 
> void *obd_zalloc(size_t size)
> {
> 	if (size > 4 * PAGE_CACHE_SIZE)
> 		return vzalloc(size);
> 	else
> 		return kmalloc(size, GFP_NOFS);

kzalloc here too. Except e also want to have locality of allocations.

> }
> 
> Except, huh?  Shouldn't we be using GFP_NOFS for the vzalloc() side?
> There was some discussion of that GFP_NOFS was a bit buggy back in 2010
> (http://marc.info/?l=linux-mm&m=128942194520631&w=4) but the current
> lustre code doesn't try to pass GFP_NOFS.

The patch I submitted was rejected, or so I think to remember, because we use __vmalloc_node
or something and it's not an exported symbol.
http://www.spinics.net/lists/linux-mm/msg83997.html

> Then it's simple enough to change OBD_FREE_LARGE() to kvfree().
> 
> Also it's weird that only the lustre people have thought of this trick
> to allocate big chunks of RAM and no one else has.  What would happen if
> we just change vmalloc() so it worked this way for everyone?

We are certainly not alone.
I saw this in a few other pieces of code.

void *ext4_kvmalloc(size_t size, gfp_t flags)
{
        void *ret;

        ret = kmalloc(size, flags | __GFP_NOWARN);
        if (!ret)
                ret = __vmalloc(size, flags, PAGE_KERNEL);
        return ret;
}

or kmem_zalloc_large in xfs.

The difference at hand is that we pessimistically assume anything over certain threshold would
fail in kmalloc anyway and others actually do try kmalloc and only switch to vmalloc if kmaloc failed.
Considerign how expensive (and unsafe) vmalloc is, there might be some benefit to converting to their
way of doing things too.

Bye,
    Oleg

  reply	other threads:[~2015-05-01 20:12 UTC|newest]

Thread overview: 36+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-05-01 15:51 [PATCH 0/11] " Julia Lawall
2015-05-01 15:51 ` [PATCH 11/11] staging: lustre: ptlrpc: " Julia Lawall
2015-05-01 15:51 ` [PATCH 10/11] staging: lustre: osc: " Julia Lawall
2015-05-01 15:51 ` [PATCH 9/11] staging: lustre: obdecho: " Julia Lawall
2015-05-01 15:51 ` [PATCH 8/11] staging: lustre: obdclass: " Julia Lawall
2015-05-01 18:30   ` walter harms
2015-05-01 18:42     ` Julia Lawall
2015-05-01 15:51 ` [PATCH 7/11] staging: lustre: mgc: " Julia Lawall
2015-05-01 15:51 ` [PATCH 6/11] staging: lustre: mdc: " Julia Lawall
2015-05-01 15:51 ` [PATCH 5/11] staging: lustre: lmv: " Julia Lawall
2015-05-01 15:51 ` [PATCH 4/11] staging: lustre: ldlm: " Julia Lawall
2015-05-01 15:51 ` [PATCH 3/11] staging: lustre: lclient: " Julia Lawall
2015-05-01 15:51 ` [PATCH 2/11] Staging: lustre: fld: " Julia Lawall
2015-05-01 17:38   ` [HPDD-discuss] " Simmons, James A.
2015-05-01 17:48     ` Julia Lawall
2015-05-01 18:49       ` Drokin, Oleg
2015-05-01 20:18       ` Simmons, James A.
2015-05-01 20:47         ` Greg Kroah-Hartman
2015-05-01 22:57           ` Simmons, James A.
2015-05-01 20:49         ` Drokin, Oleg
2015-05-01 22:59           ` Simmons, James A.
2015-05-01 18:05     ` Greg Kroah-Hartman
2015-05-01 20:02     ` Dan Carpenter
2015-05-01 20:12       ` Drokin, Oleg [this message]
2015-05-01 20:36       ` Simmons, James A.
2015-05-01 20:49         ` Greg Kroah-Hartman
2015-05-01 20:52           ` Drokin, Oleg
2015-05-01 20:58             ` Greg Kroah-Hartman
2015-05-01 21:13               ` Drokin, Oleg
2015-05-02  6:02                 ` Julia Lawall
2015-05-02  8:14                 ` Dan Carpenter
2015-05-02  9:05                   ` Julia Lawall
2015-05-03  3:12                   ` Drokin, Oleg
2015-05-01 15:51 ` [PATCH 1/11] staging: lustre: fid: " Julia Lawall
2015-05-03 18:17 ` [PATCH 0/11] " Greg Kroah-Hartman
2015-05-03 18:39   ` Julia Lawall

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=9C4D3E11-27BE-42F1-8AC7-9D8DC5D31F93@intel.com \
    --to=oleg.drokin@intel.com \
    --cc=HPDD-discuss@ml01.01.org \
    --cc=Julia.Lawall@lip6.fr \
    --cc=dan.carpenter@oracle.com \
    --cc=devel@driverdev.osuosl.org \
    --cc=gregkh@linuxfoundation.org \
    --cc=kernel-janitors@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=simmonsja@ornl.gov \
    /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

all inboxes | Powered by JetHome®