netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jeff Layton <jlayton@kernel.org>
To: NeilBrown <neilb@suse.de>, Lorenzo Bianconi <lorenzo@kernel.org>
Cc: linux-nfs@vger.kernel.org, lorenzo.bianconi@redhat.com,
	 chuck.lever@oracle.com, netdev@vger.kernel.org, kuba@kernel.org
Subject: Re: [PATCH v8 2/6] NFSD: convert write_threads to netlink command
Date: Wed, 17 Apr 2024 09:04:01 -0400	[thread overview]
Message-ID: <4659cdbbe84d40e98cb5926885189d08c94cb2f4.camel@kernel.org> (raw)
In-Reply-To: <4571e67647036cbd5833637aca07b6c4cc0e5e67.camel@kernel.org>

On Tue, 2024-04-16 at 18:39 -0400, Jeff Layton wrote:
> On Wed, 2024-04-17 at 07:55 +1000, NeilBrown wrote:
> > On Tue, 16 Apr 2024, Lorenzo Bianconi wrote:
> > > Introduce write_threads netlink command similar to the one available
> > > through the procfs.
> > 
> > I think this should support write_pool_threads too.
> > i.e.  the number of threads should be an array.  If it is a singleton,
> > the it does write_threads.   If larger it does write_pool_threads.
> > I don't think we want to add a separate command later for pool_threads.
> > 
> > NeilBrown
> > 
> 
> That is a very good point. We'll take a closer look at the current
> pool_threads interface and see how we can extend this one to cover that
> use case. Making it an array sounds reasonable at first blush.
> 

Lorenzo and I had a look this morning:

Doing this properly is going to take some refactoring, I think. The
existing pool code has a weird restriction where you can't bring the
server up and down exclusively using the pool_threads interface. There
are no userland tools for it either, so it's out of scope since it's not
really part of the how rpc.nfsd works today.

I think the way forward here is to go ahead and make this interface use
an array of integers (like Neil suggests), but for now if someone tries
to send down a THREAD_SET operation and the pool_mode isn't global,
we'll just return an error of some sort (EINVAL, EOPNOTSUPP, ?).

One the new interfaces are in I'll take a look at adding support for
managing other pool_modes properly.

Does that sound ok?


> > 
> > > 
> > > Tested-by: Jeff Layton <jlayton@kernel.org>
> > > Reviewed-by: Jeff Layton <jlayton@kernel.org>
> > > Co-developed-by: Jeff Layton <jlayton@kernel.org>
> > > Signed-off-by: Jeff Layton <jlayton@kernel.org>
> > > Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
> > > ---
> > >  Documentation/netlink/specs/nfsd.yaml |  33 ++++++++
> > >  fs/nfsd/netlink.c                     |  19 +++++
> > >  fs/nfsd/netlink.h                     |   2 +
> > >  fs/nfsd/nfsctl.c                      | 104 ++++++++++++++++++++++++++
> > >  include/uapi/linux/nfsd_netlink.h     |  11 +++
> > >  5 files changed, 169 insertions(+)
> > > 
> > > diff --git a/Documentation/netlink/specs/nfsd.yaml b/Documentation/netlink/specs/nfsd.yaml
> > > index 05acc73e2e33..cbe6c5fd6c4d 100644
> > > --- a/Documentation/netlink/specs/nfsd.yaml
> > > +++ b/Documentation/netlink/specs/nfsd.yaml
> > > @@ -62,6 +62,18 @@ attribute-sets:
> > >          name: compound-ops
> > >          type: u32
> > >          multi-attr: true
> > > +  -
> > > +    name: server-worker
> > > +    attributes:
> > > +      -
> > > +        name: threads
> > > +        type: u32
> > > +      -
> > > +        name: gracetime
> > > +        type: u32
> > > +      -
> > > +        name: leasetime
> > > +        type: u32
> > >  
> > >  operations:
> > >    list:
> > > @@ -87,3 +99,24 @@ operations:
> > >              - sport
> > >              - dport
> > >              - compound-ops
> > > +    -
> > > +      name: threads-set
> > > +      doc: set the number of running threads
> > > +      attribute-set: server-worker
> > > +      flags: [ admin-perm ]
> > > +      do:
> > > +        request:
> > > +          attributes:
> > > +            - threads
> > > +            - gracetime
> > > +            - leasetime
> > > +    -
> > > +      name: threads-get
> > > +      doc: get the number of running threads
> > > +      attribute-set: server-worker
> > > +      do:
> > > +        reply:
> > > +          attributes:
> > > +            - threads
> > > +            - gracetime
> > > +            - leasetime
> > > diff --git a/fs/nfsd/netlink.c b/fs/nfsd/netlink.c
> > > index 0e1d635ec5f9..20a646af0324 100644
> > > --- a/fs/nfsd/netlink.c
> > > +++ b/fs/nfsd/netlink.c
> > > @@ -10,6 +10,13 @@
> > >  
> > >  #include <uapi/linux/nfsd_netlink.h>
> > >  
> > > +/* NFSD_CMD_THREADS_SET - do */
> > > +static const struct nla_policy nfsd_threads_set_nl_policy[NFSD_A_SERVER_WORKER_LEASETIME + 1] = {
> > > +	[NFSD_A_SERVER_WORKER_THREADS] = { .type = NLA_U32, },
> > > +	[NFSD_A_SERVER_WORKER_GRACETIME] = { .type = NLA_U32, },
> > > +	[NFSD_A_SERVER_WORKER_LEASETIME] = { .type = NLA_U32, },
> > > +};
> > > +
> > >  /* Ops table for nfsd */
> > >  static const struct genl_split_ops nfsd_nl_ops[] = {
> > >  	{
> > > @@ -19,6 +26,18 @@ static const struct genl_split_ops nfsd_nl_ops[] = {
> > >  		.done	= nfsd_nl_rpc_status_get_done,
> > >  		.flags	= GENL_CMD_CAP_DUMP,
> > >  	},
> > > +	{
> > > +		.cmd		= NFSD_CMD_THREADS_SET,
> > > +		.doit		= nfsd_nl_threads_set_doit,
> > > +		.policy		= nfsd_threads_set_nl_policy,
> > > +		.maxattr	= NFSD_A_SERVER_WORKER_LEASETIME,
> > > +		.flags		= GENL_ADMIN_PERM | GENL_CMD_CAP_DO,
> > > +	},
> > > +	{
> > > +		.cmd	= NFSD_CMD_THREADS_GET,
> > > +		.doit	= nfsd_nl_threads_get_doit,
> > > +		.flags	= GENL_CMD_CAP_DO,
> > > +	},
> > >  };
> > >  
> > >  struct genl_family nfsd_nl_family __ro_after_init = {
> > > diff --git a/fs/nfsd/netlink.h b/fs/nfsd/netlink.h
> > > index d83dd6bdee92..4137fac477e4 100644
> > > --- a/fs/nfsd/netlink.h
> > > +++ b/fs/nfsd/netlink.h
> > > @@ -16,6 +16,8 @@ int nfsd_nl_rpc_status_get_done(struct netlink_callback *cb);
> > >  
> > >  int nfsd_nl_rpc_status_get_dumpit(struct sk_buff *skb,
> > >  				  struct netlink_callback *cb);
> > > +int nfsd_nl_threads_set_doit(struct sk_buff *skb, struct genl_info *info);
> > > +int nfsd_nl_threads_get_doit(struct sk_buff *skb, struct genl_info *info);
> > >  
> > >  extern struct genl_family nfsd_nl_family;
> > >  
> > > diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c
> > > index f2e442d7fe16..38a5df03981b 100644
> > > --- a/fs/nfsd/nfsctl.c
> > > +++ b/fs/nfsd/nfsctl.c
> > > @@ -1653,6 +1653,110 @@ int nfsd_nl_rpc_status_get_done(struct netlink_callback *cb)
> > >  	return 0;
> > >  }
> > >  
> > > +/**
> > > + * nfsd_nl_threads_set_doit - set the number of running threads
> > > + * @skb: reply buffer
> > > + * @info: netlink metadata and command arguments
> > > + *
> > > + * Return 0 on success or a negative errno.
> > > + */
> > > +int nfsd_nl_threads_set_doit(struct sk_buff *skb, struct genl_info *info)
> > > +{
> > > +	struct net *net = genl_info_net(info);
> > > +	struct nfsd_net *nn = net_generic(net, nfsd_net_id);
> > > +	int ret = -EBUSY;
> > > +	u32 nthreads;
> > > +
> > > +	if (GENL_REQ_ATTR_CHECK(info, NFSD_A_SERVER_WORKER_THREADS))
> > > +		return -EINVAL;
> > > +
> > > +	nthreads = nla_get_u32(info->attrs[NFSD_A_SERVER_WORKER_THREADS]);
> > > +
> > > +	mutex_lock(&nfsd_mutex);
> > > +	if (info->attrs[NFSD_A_SERVER_WORKER_GRACETIME] ||
> > > +	    info->attrs[NFSD_A_SERVER_WORKER_LEASETIME]) {
> > > +		const struct nlattr *attr;
> > > +
> > > +		if (nn->nfsd_serv && nn->nfsd_serv->sv_nrthreads)
> > > +			goto out_unlock;
> > > +
> > > +		ret = -EINVAL;
> > > +		attr = info->attrs[NFSD_A_SERVER_WORKER_GRACETIME];
> > > +		if (attr) {
> > > +			u32 gracetime = nla_get_u32(attr);
> > > +
> > > +			if (gracetime < 10 || gracetime > 3600)
> > > +				goto out_unlock;
> > > +
> > > +			nn->nfsd4_grace = gracetime;
> > > +		}
> > > +
> > > +		attr = info->attrs[NFSD_A_SERVER_WORKER_LEASETIME];
> > > +		if (attr) {
> > > +			u32 leasetime = nla_get_u32(attr);
> > > +
> > > +			if (leasetime < 10 || leasetime > 3600)
> > > +				goto out_unlock;
> > > +
> > > +			nn->nfsd4_lease = leasetime;
> > > +		}
> > > +	}
> > > +
> > > +	ret = nfsd_svc(nthreads, net, get_current_cred());
> > > +out_unlock:
> > > +	mutex_unlock(&nfsd_mutex);
> > > +
> > > +	return ret == nthreads ? 0 : ret;
> > > +}
> > > +
> > > +/**
> > > + * nfsd_nl_threads_get_doit - get the number of running threads
> > > + * @skb: reply buffer
> > > + * @info: netlink metadata and command arguments
> > > + *
> > > + * Return 0 on success or a negative errno.
> > > + */
> > > +int nfsd_nl_threads_get_doit(struct sk_buff *skb, struct genl_info *info)
> > > +{
> > > +	struct net *net = genl_info_net(info);
> > > +	struct nfsd_net *nn = net_generic(net, nfsd_net_id);
> > > +	void *hdr;
> > > +	int err;
> > > +
> > > +	skb = genlmsg_new(GENLMSG_DEFAULT_SIZE, GFP_KERNEL);
> > > +	if (!skb)
> > > +		return -ENOMEM;
> > > +
> > > +	hdr = genlmsg_iput(skb, info);
> > > +	if (!hdr) {
> > > +		err = -EMSGSIZE;
> > > +		goto err_free_msg;
> > > +	}
> > > +
> > > +	mutex_lock(&nfsd_mutex);
> > > +	err = nla_put_u32(skb, NFSD_A_SERVER_WORKER_GRACETIME,
> > > +			  nn->nfsd4_grace) ||
> > > +	      nla_put_u32(skb, NFSD_A_SERVER_WORKER_LEASETIME,
> > > +			  nn->nfsd4_lease) ||
> > > +	      nla_put_u32(skb, NFSD_A_SERVER_WORKER_THREADS,
> > > +			  nn->nfsd_serv ? nn->nfsd_serv->sv_nrthreads : 0);
> > > +	mutex_unlock(&nfsd_mutex);
> > > +
> > > +	if (err) {
> > > +		err = -EINVAL;
> > > +		goto err_free_msg;
> > > +	}
> > > +
> > > +	genlmsg_end(skb, hdr);
> > > +
> > > +	return genlmsg_reply(skb, info);
> > > +
> > > +err_free_msg:
> > > +	nlmsg_free(skb);
> > > +
> > > +	return err;
> > > +}
> > > +
> > >  /**
> > >   * nfsd_net_init - Prepare the nfsd_net portion of a new net namespace
> > >   * @net: a freshly-created network namespace
> > > diff --git a/include/uapi/linux/nfsd_netlink.h b/include/uapi/linux/nfsd_netlink.h
> > > index 3cd044edee5d..ccc78a5ee650 100644
> > > --- a/include/uapi/linux/nfsd_netlink.h
> > > +++ b/include/uapi/linux/nfsd_netlink.h
> > > @@ -29,8 +29,19 @@ enum {
> > >  	NFSD_A_RPC_STATUS_MAX = (__NFSD_A_RPC_STATUS_MAX - 1)
> > >  };
> > >  
> > > +enum {
> > > +	NFSD_A_SERVER_WORKER_THREADS = 1,
> > > +	NFSD_A_SERVER_WORKER_GRACETIME,
> > > +	NFSD_A_SERVER_WORKER_LEASETIME,
> > > +
> > > +	__NFSD_A_SERVER_WORKER_MAX,
> > > +	NFSD_A_SERVER_WORKER_MAX = (__NFSD_A_SERVER_WORKER_MAX - 1)
> > > +};
> > > +
> > >  enum {
> > >  	NFSD_CMD_RPC_STATUS_GET = 1,
> > > +	NFSD_CMD_THREADS_SET,
> > > +	NFSD_CMD_THREADS_GET,
> > >  
> > >  	__NFSD_CMD_MAX,
> > >  	NFSD_CMD_MAX = (__NFSD_CMD_MAX - 1)
> > > -- 
> > > 2.44.0
> > > 
> > > 
> > 
> 

-- 
Jeff Layton <jlayton@kernel.org>

  reply	other threads:[~2024-04-17 13:04 UTC|newest]

Thread overview: 23+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-04-15 19:44 [PATCH v8 0/6] convert write_threads, write_version and write_ports to netlink commands Lorenzo Bianconi
2024-04-15 19:44 ` [PATCH v8 1/6] nfsd: move nfsd_mutex handling into nfsd_svc callers Lorenzo Bianconi
2024-04-15 19:44 ` [PATCH v8 2/6] NFSD: convert write_threads to netlink command Lorenzo Bianconi
2024-04-16 21:55   ` NeilBrown
2024-04-16 22:39     ` Jeff Layton
2024-04-17 13:04       ` Jeff Layton [this message]
2024-04-16 22:28   ` NeilBrown
2024-04-16 22:47     ` Jeff Layton
2024-04-15 19:44 ` [PATCH v8 3/6] NFSD: add write_version " Lorenzo Bianconi
2024-04-16  3:16   ` NeilBrown
2024-04-16 10:26     ` Jeff Layton
2024-04-16 21:48       ` NeilBrown
2024-04-16 22:33         ` Jeff Layton
2024-04-16 23:05           ` NeilBrown
2024-04-17  0:10             ` Jeff Layton
2024-04-17  0:43               ` NeilBrown
2024-04-17 11:25                 ` Jeff Layton
2024-04-22  3:36                   ` NeilBrown
2024-04-23 12:42                     ` Jeff Layton
2024-04-15 19:44 ` [PATCH v8 4/6] SUNRPC: introduce svc_xprt_create_from_sa utility routine Lorenzo Bianconi
2024-04-15 19:44 ` [PATCH v8 5/6] SUNRPC: add a new svc_find_listener helper Lorenzo Bianconi
2024-04-15 19:44 ` [PATCH v8 6/6] NFSD: add listener-{set,get} netlink command Lorenzo Bianconi
2024-04-16 19:05 ` [PATCH v8 0/6] convert write_threads, write_version and write_ports to netlink commands Chuck Lever

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=4659cdbbe84d40e98cb5926885189d08c94cb2f4.camel@kernel.org \
    --to=jlayton@kernel.org \
    --cc=chuck.lever@oracle.com \
    --cc=kuba@kernel.org \
    --cc=linux-nfs@vger.kernel.org \
    --cc=lorenzo.bianconi@redhat.com \
    --cc=lorenzo@kernel.org \
    --cc=neilb@suse.de \
    --cc=netdev@vger.kernel.org \
    /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).