linux-nfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jeff Layton <jlayton@redhat.com>
To: Anna Schumaker <Anna.Schumaker@netapp.com>,
	trond.myklebust@primarydata.com
Cc: linux-nfs@vger.kernel.org
Subject: Re: [PATCH 4/9] nfs: add a freezable_schedule_timeout_unsafe() and use it when waiting to retry LOCK
Date: Thu, 08 Sep 2016 14:36:22 -0400	[thread overview]
Message-ID: <1473359782.23262.10.camel@redhat.com> (raw)
In-Reply-To: <3c94ea23-fd38-be69-2eef-e641c8bb1ff5@Netapp.com>

On Thu, 2016-09-08 at 14:20 -0400, Anna Schumaker wrote:
> Hi Jeff,
> 
> Can you CC the freezer folks on this patch during the second
> posting?  They'll probably need to at send an ACK for the changes to
> their include file.
> 
> Thanks,
> Anna
> 

Will do...

OTOH, now that I look, I'm not sure we really need a *_unsafe variant
here at all.

File locking is somewhat different in that we're expected to sleep in
the kernel, and I don't think we hold any kernel locks that are the
usual worry here (i_rwsem and the like). So we might can just use
freezable_schedule_timeout here.

I'll see if I can experiment with that before the next posting. It'd be
nice to eliminate one of the unsafe sleep callsites.

> On 09/06/2016 11:12 AM, Jeff Layton wrote:
> > 
> > We actually want to use TASK_INTERRUPTIBLE sleeps here. Once the
> > task
> > wakes up, if there is a signal pending then we'll be returning an
> > error
> > anyway. So, we might as well wake up immediately for non-fatal
> > signals
> > as well. That allows us to return to userland more quickly in that
> > case,
> > but won't change the error that userland sees.
> > 
> > Signed-off-by: Jeff Layton <jlayton@redhat.com>
> > ---
> >  fs/nfs/nfs4proc.c       |  3 ++-
> >  include/linux/freezer.h | 13 +++++++++++++
> >  2 files changed, 15 insertions(+), 1 deletion(-)
> > 
> > diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
> > index e3bf95369daf..e9232d71bc64 100644
> > --- a/fs/nfs/nfs4proc.c
> > +++ b/fs/nfs/nfs4proc.c
> > @@ -5537,7 +5537,8 @@ int nfs4_proc_delegreturn(struct inode
> > *inode, struct rpc_cred *cred, const nfs4
> >  static unsigned long
> >  nfs4_set_lock_task_retry(unsigned long timeout)
> >  {
> > -	freezable_schedule_timeout_killable_unsafe(timeout);
> > +	set_current_state(TASK_INTERRUPTIBLE);
> > +	freezable_schedule_timeout_unsafe(timeout);
> >  	timeout <<= 1;
> >  	if (timeout > NFS4_LOCK_MAXTIMEOUT)
> >  		return NFS4_LOCK_MAXTIMEOUT;
> > diff --git a/include/linux/freezer.h b/include/linux/freezer.h
> > index dd03e837ebb7..fe31601e7f55 100644
> > --- a/include/linux/freezer.h
> > +++ b/include/linux/freezer.h
> > @@ -197,6 +197,19 @@ static inline long
> > freezable_schedule_timeout(long timeout)
> >   * Like schedule_timeout_interruptible(), but should not block the
> > freezer.  Do not
> >   * call this with locks held.
> >   */
> > +static inline long freezable_schedule_timeout_unsafe(long timeout)
> > +{
> > +	long __retval;
> > +	freezer_do_not_count();
> > +	__retval = schedule_timeout(timeout);
> > +	freezer_count_unsafe();
> > +	return __retval;
> > +}
> > +
> > +/*
> > + * Like schedule_timeout_interruptible(), but should not block the
> > freezer.  Do not
> > + * call this with locks held.
> > + */
> >  static inline long freezable_schedule_timeout_interruptible(long
> > timeout)
> >  {
> >  	long __retval;
> > 
> 

-- 
Jeff Layton <jlayton@redhat.com>

  reply	other threads:[~2016-09-08 18:36 UTC|newest]

Thread overview: 24+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-09-06 15:12 [PATCH 0/9] nfs: add CB_NOTIFY_LOCK support to nfs client Jeff Layton
2016-09-06 15:12 ` [PATCH 1/9] nfs: the length argument to read_buf should be unsigned Jeff Layton
2016-09-08 17:39   ` Anna Schumaker
2016-09-08 18:19     ` Jeff Layton
2016-09-06 15:12 ` [PATCH 2/9] nfs: eliminate pointless and confusing do_vfs_lock wrappers Jeff Layton
2016-09-06 15:12 ` [PATCH 3/9] nfs: check for POSIX lock capability on server even for flock locks Jeff Layton
2016-09-06 15:12 ` [PATCH 4/9] nfs: add a freezable_schedule_timeout_unsafe() and use it when waiting to retry LOCK Jeff Layton
2016-09-06 16:39   ` Jeff Layton
2016-09-08 18:20   ` Anna Schumaker
2016-09-08 18:36     ` Jeff Layton [this message]
2016-09-06 15:12 ` [PATCH 5/9] nfs: add handling for CB_NOTIFY_LOCK in client Jeff Layton
2016-09-08 20:11   ` Anna Schumaker
2016-09-06 15:12 ` [PATCH 6/9] nfs: move nfs4_set_lock_state call into caller Jeff Layton
2016-09-08 19:47   ` Anna Schumaker
2016-09-08 21:41     ` Jeff Layton
2016-09-06 15:12 ` [PATCH 7/9] nfs: add code to allow client to wait on lock callbacks Jeff Layton
2016-09-08 19:59   ` Anna Schumaker
2016-09-08 21:42     ` Jeff Layton
2016-09-06 15:12 ` [PATCH 8/9] nfs: ensure that the filehandle in CB_NOTIFY_LOCK request matches the inode Jeff Layton
2016-09-08 20:07   ` Anna Schumaker
2016-09-08 21:43     ` Jeff Layton
2016-09-06 15:12 ` [PATCH 9/9] nfs: track whether server sets MAY_NOTIFY_LOCK flag Jeff Layton
2016-09-08 20:15   ` Anna Schumaker
2016-09-08 21:47     ` Jeff Layton

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=1473359782.23262.10.camel@redhat.com \
    --to=jlayton@redhat.com \
    --cc=Anna.Schumaker@netapp.com \
    --cc=linux-nfs@vger.kernel.org \
    --cc=trond.myklebust@primarydata.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 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).