All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Lu, Aaron" <aaron.lu@intel.com>
To: "bcodding@redhat.com" <bcodding@redhat.com>,
	"jlayton@redhat.com" <jlayton@redhat.com>
Cc: "torvalds@linux-foundation.org" <torvalds@linux-foundation.org>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"lkp@01.org" <lkp@01.org>, "Ye, Xiaolong" <xiaolong.ye@intel.com>
Subject: Re: [LKP] [lkp-robot] [fs/locks] 52306e882f: stress-ng.lockofd.ops_per_sec -11% regression
Date: Wed, 6 Dec 2017 01:43:52 +0000	[thread overview]
Message-ID: <1512524641.18934.2.camel@intel.com> (raw)
In-Reply-To: <1512471662.4125.2.camel@redhat.com>

On Tue, 2017-12-05 at 06:01 -0500, Jeff Layton wrote:
> On Tue, 2017-12-05 at 13:57 +0800, Aaron Lu wrote:
> > On Wed, Nov 08, 2017 at 03:22:33PM +0800, Aaron Lu wrote:
> > > On Thu, Sep 28, 2017 at 04:02:23PM +0800, kernel test robot wrote:
> > > > 
> > > > Greeting,
> > > > 
> > > > FYI, we noticed a -11% regression of stress-ng.lockofd.ops_per_sec due to commit:
> > > > 
> > > > 
> > > > commit: 52306e882f77d3fd73f91435c41373d634acc5d2 ("fs/locks: Use allocation rather than the stack in fcntl_getlk()")
> > > > https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master
> > > 
> > > It's been a while, I wonder what do you think of this regression?
> > > 
> > > The test stresses byte-range locks AFAICS and since the commit uses
> > > dynamic allocation instead of stack for the 'struct file_lock', it sounds
> > > natural the performance regressed for this test.
> > > 
> > > Now the question is, do we care about the performance regression here?
> > 
> > Appreciated it if you can share your opinion on this, thanks.
> > 
> > Regards,
> > Aaron
> >  
> 
> Sorry I missed your earlier mail about this. My feeling is to not worry

Never mind :)

> about it. struct file_lock is rather large, so putting it on the stack
> was always a bit dangerous, and F_GETLK is a rather uncommon operation
> anyway.
> 
> That said, if there are real-world workloads that have regressed because
> of this patch, I'm definitely open to backing it out.
> 
> Does anyone else have opinions on the matter?

Your comments makes sense to me, thanks for the reply.

WARNING: multiple messages have this Message-ID (diff)
From: Lu, Aaron <aaron.lu@intel.com>
To: lkp@lists.01.org
Subject: Re: [lkp-robot] [fs/locks] 52306e882f: stress-ng.lockofd.ops_per_sec -11% regression
Date: Wed, 06 Dec 2017 01:43:52 +0000	[thread overview]
Message-ID: <1512524641.18934.2.camel@intel.com> (raw)
In-Reply-To: <1512471662.4125.2.camel@redhat.com>

[-- Attachment #1: Type: text/plain, Size: 1603 bytes --]

On Tue, 2017-12-05 at 06:01 -0500, Jeff Layton wrote:
> On Tue, 2017-12-05 at 13:57 +0800, Aaron Lu wrote:
> > On Wed, Nov 08, 2017 at 03:22:33PM +0800, Aaron Lu wrote:
> > > On Thu, Sep 28, 2017 at 04:02:23PM +0800, kernel test robot wrote:
> > > > 
> > > > Greeting,
> > > > 
> > > > FYI, we noticed a -11% regression of stress-ng.lockofd.ops_per_sec due to commit:
> > > > 
> > > > 
> > > > commit: 52306e882f77d3fd73f91435c41373d634acc5d2 ("fs/locks: Use allocation rather than the stack in fcntl_getlk()")
> > > > https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master
> > > 
> > > It's been a while, I wonder what do you think of this regression?
> > > 
> > > The test stresses byte-range locks AFAICS and since the commit uses
> > > dynamic allocation instead of stack for the 'struct file_lock', it sounds
> > > natural the performance regressed for this test.
> > > 
> > > Now the question is, do we care about the performance regression here?
> > 
> > Appreciated it if you can share your opinion on this, thanks.
> > 
> > Regards,
> > Aaron
> >  
> 
> Sorry I missed your earlier mail about this. My feeling is to not worry

Never mind :)

> about it. struct file_lock is rather large, so putting it on the stack
> was always a bit dangerous, and F_GETLK is a rather uncommon operation
> anyway.
> 
> That said, if there are real-world workloads that have regressed because
> of this patch, I'm definitely open to backing it out.
> 
> Does anyone else have opinions on the matter?

Your comments makes sense to me, thanks for the reply.

  reply	other threads:[~2017-12-06  1:44 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-09-28  8:02 [lkp-robot] [fs/locks] 52306e882f: stress-ng.lockofd.ops_per_sec -11% regression kernel test robot
2017-09-28  8:02 ` kernel test robot
2017-11-08  7:22 ` [LKP] " Aaron Lu
2017-11-08  7:22   ` Aaron Lu
2017-12-05  5:57   ` [LKP] " Aaron Lu
2017-12-05  5:57     ` Aaron Lu
2017-12-05 11:01     ` [LKP] " Jeff Layton
2017-12-05 11:01       ` Jeff Layton
2017-12-06  1:43       ` Lu, Aaron [this message]
2017-12-06  1:43         ` Lu, Aaron

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=1512524641.18934.2.camel@intel.com \
    --to=aaron.lu@intel.com \
    --cc=bcodding@redhat.com \
    --cc=jlayton@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=lkp@01.org \
    --cc=torvalds@linux-foundation.org \
    --cc=xiaolong.ye@intel.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.