From: Trond Myklebust <trondmy@hammerspace.com>
To: "linux-nfs@vger.kernel.org" <linux-nfs@vger.kernel.org>,
"chuck.lever@oracle.com" <chuck.lever@oracle.com>
Subject: Re: [PATCH RFC] NFSD: Hold rcu_read_lock while getting refs
Date: Mon, 3 Oct 2022 13:16:48 +0000 [thread overview]
Message-ID: <aaebdc3facb1466da6390937b38ee9a31d2f9344.camel@hammerspace.com> (raw)
In-Reply-To: <166463917715.10124.3789034969503323129.stgit@bazille.1015granger.net>
On Sat, 2022-10-01 at 11:48 -0400, Chuck Lever wrote:
> nfsd_file is RCU-freed, so it's possible that one could be found
> that's in the process of being freed and the memory recycled. Ensure
> we hold the rcu_read_lock while attempting to get a reference on the
> object.
>
> Suggested-by: Jeff Layton <jlayton@kernel.org>
> Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
> ---
> fs/nfsd/filecache.c | 34 +++++++++++-----------------------
> fs/nfsd/trace.h | 27 ---------------------------
> 2 files changed, 11 insertions(+), 50 deletions(-)
>
> This is what I was thinking... Compile-tested only.
>
>
> diff --git a/fs/nfsd/filecache.c b/fs/nfsd/filecache.c
> index be152e3e3a80..6e17f74fb29f 100644
> --- a/fs/nfsd/filecache.c
> +++ b/fs/nfsd/filecache.c
> @@ -1056,10 +1056,12 @@ nfsd_file_do_acquire(struct svc_rqst *rqstp,
> struct svc_fh *fhp,
>
> retry:
> /* Avoid allocation if the item is already in cache */
> - nf = rhashtable_lookup_fast(&nfsd_file_rhash_tbl, &key,
> - nfsd_file_rhash_params);
> + rcu_read_lock();
> + nf = rhashtable_lookup(&nfsd_file_rhash_tbl, &key,
> + nfsd_file_rhash_params);
> if (nf)
> nf = nfsd_file_get(nf);
> + rcu_read_unlock();
>
That definitely deserves a 'Fixes:' line so you can unbreak 6.0.
> if (nf)
> goto wait_for_construction;
>
> @@ -1069,21 +1071,14 @@ nfsd_file_do_acquire(struct svc_rqst *rqstp,
> struct svc_fh *fhp,
> goto out_status;
> }
>
> - nf = rhashtable_lookup_get_insert_key(&nfsd_file_rhash_tbl,
> - &key, &new->nf_rhash,
> -
> nfsd_file_rhash_params);
> - if (!nf) {
> - nf = new;
> - goto open_file;
> - }
> - if (IS_ERR(nf))
> - goto insert_err;
> - nf = nfsd_file_get(nf);
> - if (nf == NULL) {
> - nf = new;
> - goto open_file;
> + if (rhashtable_lookup_insert_key(&nfsd_file_rhash_tbl,
> + &key, &new->nf_rhash,
> + nfsd_file_rhash_params)) {
> + nfsd_file_slab_free(&new->nf_rcu);
> + goto retry;
> }
> - nfsd_file_slab_free(&new->nf_rcu);
> + nf = new;
> + goto open_file;
>
> wait_for_construction:
> wait_on_bit(&nf->nf_flags, NFSD_FILE_PENDING,
> TASK_UNINTERRUPTIBLE);
> @@ -1143,13 +1138,6 @@ nfsd_file_do_acquire(struct svc_rqst *rqstp,
> struct svc_fh *fhp,
> smp_mb__after_atomic();
> wake_up_bit(&nf->nf_flags, NFSD_FILE_PENDING);
> goto out;
> -
> -insert_err:
> - nfsd_file_slab_free(&new->nf_rcu);
> - trace_nfsd_file_insert_err(rqstp, key.inode, may_flags,
> PTR_ERR(nf));
> - nf = NULL;
> - status = nfserr_jukebox;
> - goto out_status;
> }
>
> /**
> diff --git a/fs/nfsd/trace.h b/fs/nfsd/trace.h
> index 06a96e955bd0..c15467b2e8d9 100644
> --- a/fs/nfsd/trace.h
> +++ b/fs/nfsd/trace.h
> @@ -954,33 +954,6 @@ TRACE_EVENT(nfsd_file_create,
> )
> );
>
> -TRACE_EVENT(nfsd_file_insert_err,
> - TP_PROTO(
> - const struct svc_rqst *rqstp,
> - const struct inode *inode,
> - unsigned int may_flags,
> - long error
> - ),
> - TP_ARGS(rqstp, inode, may_flags, error),
> - TP_STRUCT__entry(
> - __field(u32, xid)
> - __field(const void *, inode)
> - __field(unsigned long, may_flags)
> - __field(long, error)
> - ),
> - TP_fast_assign(
> - __entry->xid = be32_to_cpu(rqstp->rq_xid);
> - __entry->inode = inode;
> - __entry->may_flags = may_flags;
> - __entry->error = error;
> - ),
> - TP_printk("xid=0x%x inode=%p may_flags=%s error=%ld",
> - __entry->xid, __entry->inode,
> - show_nfsd_may_flags(__entry->may_flags),
> - __entry->error
> - )
> -);
> -
> TRACE_EVENT(nfsd_file_cons_err,
> TP_PROTO(
> const struct svc_rqst *rqstp,
>
>
--
Trond Myklebust
Linux NFS client maintainer, Hammerspace
trond.myklebust@hammerspace.com
prev parent reply other threads:[~2022-10-03 13:17 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-10-01 15:48 [PATCH RFC] NFSD: Hold rcu_read_lock while getting refs Chuck Lever
2022-10-03 11:34 ` Jeff Layton
2022-10-03 13:16 ` Trond Myklebust [this message]
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=aaebdc3facb1466da6390937b38ee9a31d2f9344.camel@hammerspace.com \
--to=trondmy@hammerspace.com \
--cc=chuck.lever@oracle.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).