From: bfields@fieldses.org (J. Bruce Fields)
To: Olga Kornievskaia <kolga@netapp.com>
Cc: bfields@redhat.com, linux-nfs@vger.kernel.org
Subject: Re: [PATCH v6 05/10] NFSD first draft of async copy
Date: Thu, 25 Jan 2018 17:04:40 -0500 [thread overview]
Message-ID: <20180125220440.GA21492@fieldses.org> (raw)
In-Reply-To: <20171024174752.74910-6-kolga@netapp.com>
Nit: this could use a better subject line.
On Tue, Oct 24, 2017 at 01:47:47PM -0400, Olga Kornievskaia wrote:
...
> + if (!copy->cp_synchronous) {
> + status = nfsd4_init_copy_res(copy, 0);
> + async_copy = kzalloc(sizeof(struct nfsd4_copy), GFP_KERNEL);
> + if (!async_copy) {
> + status = nfserrno(-ENOMEM);
> + goto out;
> + }
> + dup_copy_fields(copy, async_copy);
> + memcpy(©->cp_res.cb_stateid, ©->cp_dst_stateid,
> + sizeof(copy->cp_dst_stateid));
> + spin_lock(&async_copy->cp_clp->async_lock);
> + list_add(&async_copy->copies,
> + &async_copy->cp_clp->async_copies);
> + spin_unlock(&async_copy->cp_clp->async_lock);
At this point other threads could in theory look up this async_copy, but
its copy_task field is not yet initialized. I don't *think* that's a
problem for nfsd4_shutdown_copy, because I don't think the server could
be processing rpc's for this client any more at that point. But I think
a malicious client might be able to trigger a NULL dereference in
nfsd4_offload_cancel.
Is there any reason not to assign copy_task before adding it to this
list?
--b.
> + async_copy->copy_task = kthread_create(nfsd4_do_async_copy,
> + async_copy, "%s", "copy thread");
> + if (IS_ERR(async_copy->copy_task)) {
> + status = PTR_ERR(async_copy->copy_task);
> + goto out_err_dec;
> + }
> + wake_up_process(async_copy->copy_task);
> + } else {
> + status = nfsd4_do_copy(copy, 1);
> }
> -
> - fput(src);
> - fput(dst);
> out:
> return status;
> +out_err_dec:
> + cleanup_async_copy(async_copy);
> + goto out;
> }
>
> static __be32
> diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
> index 0c04f81..d7767a1 100644
> --- a/fs/nfsd/nfs4state.c
> +++ b/fs/nfsd/nfs4state.c
> @@ -1774,6 +1774,8 @@ static struct nfs4_client *alloc_client(struct xdr_netobj name)
> #ifdef CONFIG_NFSD_PNFS
> INIT_LIST_HEAD(&clp->cl_lo_states);
> #endif
> + INIT_LIST_HEAD(&clp->async_copies);
> + spin_lock_init(&clp->async_lock);
> spin_lock_init(&clp->cl_lock);
> rpc_init_wait_queue(&clp->cl_cb_waitq, "Backchannel slot table");
> return clp;
> diff --git a/fs/nfsd/state.h b/fs/nfsd/state.h
> index f8b0210..9189062 100644
> --- a/fs/nfsd/state.h
> +++ b/fs/nfsd/state.h
> @@ -352,6 +352,8 @@ struct nfs4_client {
> struct rpc_wait_queue cl_cb_waitq; /* backchannel callers may */
> /* wait here for slots */
> struct net *net;
> + struct list_head async_copies; /* list of async copies */
> + spinlock_t async_lock; /* lock for async copies */
> };
>
> /* struct nfs4_client_reset
> diff --git a/fs/nfsd/xdr4.h b/fs/nfsd/xdr4.h
> index 9b0c099..0a19954 100644
> --- a/fs/nfsd/xdr4.h
> +++ b/fs/nfsd/xdr4.h
> @@ -529,6 +529,15 @@ struct nfsd4_copy {
> struct nfsd4_callback cp_cb;
> __be32 nfserr;
> struct knfsd_fh fh;
> +
> + struct nfs4_client *cp_clp;
> +
> + struct file *fh_src;
> + struct file *fh_dst;
> + struct net *net;
> +
> + struct list_head copies;
> + struct task_struct *copy_task;
> };
>
> struct nfsd4_seek {
> --
> 1.8.3.1
next prev parent reply other threads:[~2018-01-25 22:04 UTC|newest]
Thread overview: 48+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-10-24 17:47 [PATCH v6 00/10] NFSD support for asynchronous COPY Olga Kornievskaia
2017-10-24 17:47 ` [PATCH v6 01/10] NFSD CB_OFFLOAD xdr Olga Kornievskaia
2018-01-25 16:43 ` J. Bruce Fields
2018-01-26 15:16 ` Olga Kornievskaia
2017-10-24 17:47 ` [PATCH v6 02/10] NFSD OFFLOAD_STATUS xdr Olga Kornievskaia
2017-10-24 17:47 ` [PATCH v6 03/10] NFSD OFFLOAD_CANCEL xdr Olga Kornievskaia
2017-10-24 17:47 ` [PATCH v6 04/10] NFSD xdr callback stateid in async COPY reply Olga Kornievskaia
2017-10-24 17:47 ` [PATCH v6 05/10] NFSD first draft of async copy Olga Kornievskaia
2018-01-25 22:04 ` J. Bruce Fields [this message]
2018-01-26 15:17 ` Olga Kornievskaia
2018-02-15 19:59 ` Olga Kornievskaia
2018-02-15 20:06 ` J. Bruce Fields
2018-01-25 22:29 ` J. Bruce Fields
2018-01-26 15:17 ` Olga Kornievskaia
2018-01-26 21:34 ` J. Bruce Fields
2018-02-02 19:50 ` Olga Kornievskaia
2018-02-02 19:55 ` J. Bruce Fields
2017-10-24 17:47 ` [PATCH v6 06/10] NFSD return nfs4_stid in nfs4_preprocess_stateid_op Olga Kornievskaia
2017-10-24 17:47 ` [PATCH v6 07/10] NFSD create new stateid for async copy Olga Kornievskaia
2018-01-26 21:37 ` J. Bruce Fields
2018-01-26 21:59 ` J. Bruce Fields
2018-02-02 20:45 ` Olga Kornievskaia
2018-02-02 21:45 ` J. Bruce Fields
2018-02-15 22:18 ` Olga Kornievskaia
2018-02-16 1:43 ` J. Bruce Fields
2018-02-16 16:06 ` Olga Kornievskaia
2018-02-16 18:12 ` J. Bruce Fields
2018-02-16 20:53 ` Olga Kornievskaia
2018-02-20 18:48 ` J. Bruce Fields
2018-03-06 17:15 ` Olga Kornievskaia
2018-03-06 19:33 ` J. Bruce Fields
2017-10-24 17:47 ` [PATCH v6 08/10] NFSD handle OFFLOAD_CANCEL op Olga Kornievskaia
2018-02-16 17:28 ` Olga Kornievskaia
2018-02-16 18:10 ` J. Bruce Fields
2017-10-24 17:47 ` [PATCH v6 09/10] NFSD support OFFLOAD_STATUS Olga Kornievskaia
2017-10-24 17:47 ` [PATCH v6 10/10] NFSD stop queued async copies on client shutdown Olga Kornievskaia
2018-01-25 22:22 ` J. Bruce Fields
2018-01-26 15:17 ` Olga Kornievskaia
2017-11-03 19:57 ` [PATCH v6 00/10] NFSD support for asynchronous COPY Olga Kornievskaia
2017-11-10 15:01 ` Olga Kornievskaia
2017-11-14 0:48 ` J. Bruce Fields
2017-11-28 20:28 ` Olga Kornievskaia
2017-11-30 20:18 ` J. Bruce Fields
2017-11-30 23:03 ` Olga Kornievskaia
2017-12-04 21:32 ` J. Bruce Fields
[not found] ` <CAN-5tyEVSwBmPMtUBJYDdLi7FK2MNMGuDQrrsvp776zD3Jcw0w@mail.gmail.com>
2018-01-22 16:51 ` Olga Kornievskaia
2018-01-25 22:33 ` J. Bruce Fields
2018-01-26 15:16 ` Olga Kornievskaia
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=20180125220440.GA21492@fieldses.org \
--to=bfields@fieldses.org \
--cc=bfields@redhat.com \
--cc=kolga@netapp.com \
--cc=linux-nfs@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).