public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	stable@vger.kernel.org,
	Trond Myklebust <trond.myklebust@hammerspace.com>,
	Sasha Levin <sashal@kernel.org>
Subject: [PATCH 5.10 45/63] NFSv4: Fix a pNFS layout related use-after-free race when freeing the inode
Date: Mon,  4 Jan 2021 16:57:38 +0100	[thread overview]
Message-ID: <20210104155711.007242412@linuxfoundation.org> (raw)
In-Reply-To: <20210104155708.800470590@linuxfoundation.org>

From: Trond Myklebust <trond.myklebust@hammerspace.com>

[ Upstream commit b6d49ecd1081740b6e632366428b960461f8158b ]

When returning the layout in nfs4_evict_inode(), we need to ensure that
the layout is actually done being freed before we can proceed to free the
inode itself.

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 fs/nfs/nfs4super.c |  2 +-
 fs/nfs/pnfs.c      | 33 +++++++++++++++++++++++++++++++--
 fs/nfs/pnfs.h      |  5 +++++
 3 files changed, 37 insertions(+), 3 deletions(-)

diff --git a/fs/nfs/nfs4super.c b/fs/nfs/nfs4super.c
index 93f5c1678ec29..984cc42ee54d8 100644
--- a/fs/nfs/nfs4super.c
+++ b/fs/nfs/nfs4super.c
@@ -67,7 +67,7 @@ static void nfs4_evict_inode(struct inode *inode)
 	nfs_inode_evict_delegation(inode);
 	/* Note that above delegreturn would trigger pnfs return-on-close */
 	pnfs_return_layout(inode);
-	pnfs_destroy_layout(NFS_I(inode));
+	pnfs_destroy_layout_final(NFS_I(inode));
 	/* First call standard NFS clear_inode() code */
 	nfs_clear_inode(inode);
 	nfs4_xattr_cache_zap(inode);
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index 0e50b9d45c320..07f59dc8cb2e7 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -294,6 +294,7 @@ void
 pnfs_put_layout_hdr(struct pnfs_layout_hdr *lo)
 {
 	struct inode *inode;
+	unsigned long i_state;
 
 	if (!lo)
 		return;
@@ -304,8 +305,12 @@ pnfs_put_layout_hdr(struct pnfs_layout_hdr *lo)
 		if (!list_empty(&lo->plh_segs))
 			WARN_ONCE(1, "NFS: BUG unfreed layout segments.\n");
 		pnfs_detach_layout_hdr(lo);
+		i_state = inode->i_state;
 		spin_unlock(&inode->i_lock);
 		pnfs_free_layout_hdr(lo);
+		/* Notify pnfs_destroy_layout_final() that we're done */
+		if (i_state & (I_FREEING | I_CLEAR))
+			wake_up_var(lo);
 	}
 }
 
@@ -734,8 +739,7 @@ pnfs_free_lseg_list(struct list_head *free_me)
 	}
 }
 
-void
-pnfs_destroy_layout(struct nfs_inode *nfsi)
+static struct pnfs_layout_hdr *__pnfs_destroy_layout(struct nfs_inode *nfsi)
 {
 	struct pnfs_layout_hdr *lo;
 	LIST_HEAD(tmp_list);
@@ -753,9 +757,34 @@ pnfs_destroy_layout(struct nfs_inode *nfsi)
 		pnfs_put_layout_hdr(lo);
 	} else
 		spin_unlock(&nfsi->vfs_inode.i_lock);
+	return lo;
+}
+
+void pnfs_destroy_layout(struct nfs_inode *nfsi)
+{
+	__pnfs_destroy_layout(nfsi);
 }
 EXPORT_SYMBOL_GPL(pnfs_destroy_layout);
 
+static bool pnfs_layout_removed(struct nfs_inode *nfsi,
+				struct pnfs_layout_hdr *lo)
+{
+	bool ret;
+
+	spin_lock(&nfsi->vfs_inode.i_lock);
+	ret = nfsi->layout != lo;
+	spin_unlock(&nfsi->vfs_inode.i_lock);
+	return ret;
+}
+
+void pnfs_destroy_layout_final(struct nfs_inode *nfsi)
+{
+	struct pnfs_layout_hdr *lo = __pnfs_destroy_layout(nfsi);
+
+	if (lo)
+		wait_var_event(lo, pnfs_layout_removed(nfsi, lo));
+}
+
 static bool
 pnfs_layout_add_bulk_destroy_list(struct inode *inode,
 		struct list_head *layout_list)
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index 2661c44c62db4..78c3893918486 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -266,6 +266,7 @@ struct pnfs_layout_segment *pnfs_layout_process(struct nfs4_layoutget *lgp);
 void pnfs_layoutget_free(struct nfs4_layoutget *lgp);
 void pnfs_free_lseg_list(struct list_head *tmp_list);
 void pnfs_destroy_layout(struct nfs_inode *);
+void pnfs_destroy_layout_final(struct nfs_inode *);
 void pnfs_destroy_all_layouts(struct nfs_client *);
 int pnfs_destroy_layouts_byfsid(struct nfs_client *clp,
 		struct nfs_fsid *fsid,
@@ -710,6 +711,10 @@ static inline void pnfs_destroy_layout(struct nfs_inode *nfsi)
 {
 }
 
+static inline void pnfs_destroy_layout_final(struct nfs_inode *nfsi)
+{
+}
+
 static inline struct pnfs_layout_segment *
 pnfs_get_lseg(struct pnfs_layout_segment *lseg)
 {
-- 
2.27.0




  parent reply	other threads:[~2021-01-04 16:06 UTC|newest]

Thread overview: 75+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-01-04 15:56 [PATCH 5.10 00/63] 5.10.5-rc1 review Greg Kroah-Hartman
2021-01-04 15:56 ` [PATCH 5.10 01/63] net/sched: sch_taprio: reset child qdiscs before freeing them Greg Kroah-Hartman
2021-01-04 22:58   ` Sasha Levin
2021-01-04 23:06     ` Jakub Kicinski
2021-01-04 15:56 ` [PATCH 5.10 02/63] mptcp: fix security context on server socket Greg Kroah-Hartman
2021-01-04 15:56 ` [PATCH 5.10 03/63] ethtool: fix error paths in ethnl_set_channels() Greg Kroah-Hartman
2021-01-04 15:56 ` [PATCH 5.10 04/63] ethtool: fix string set id check Greg Kroah-Hartman
2021-01-04 15:56 ` [PATCH 5.10 05/63] md/raid10: initialize r10_bio->read_slot before use Greg Kroah-Hartman
2021-01-04 15:56 ` [PATCH 5.10 06/63] drm/amd/display: Add get_dig_frontend implementation for DCEx Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 07/63] io_uring: close a small race gap for files cancel Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 08/63] jffs2: Allow setting rp_size to zero during remounting Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 09/63] jffs2: Fix NULL pointer dereference in rp_size fs option parsing Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 10/63] spi: dw-bt1: Fix undefined devm_mux_control_get symbol Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 11/63] opp: fix memory leak in _allocate_opp_table Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 12/63] opp: Call the missing clk_put() on error Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 13/63] scsi: block: Fix a race in the runtime power management code Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 14/63] mm/hugetlb: fix deadlock in hugetlb_cow error path Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 15/63] mm: memmap defer init doesnt work as expected Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 16/63] lib/zlib: fix inflating zlib streams on s390 Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 17/63] io_uring: dont assume mm is constant across submits Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 18/63] io_uring: use bottom half safe lock for fixed file data Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 19/63] io_uring: add a helper for setting a ref node Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 20/63] io_uring: fix io_sqe_files_unregister() hangs Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 21/63] kernel/io_uring: cancel io_uring before task works Greg Kroah-Hartman
2021-01-04 16:06   ` Pavel Begunkov
2021-01-04 17:43     ` Sasha Levin
2021-01-04 15:57 ` [PATCH 5.10 22/63] uapi: move constants from <linux/kernel.h> to <linux/const.h> Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 23/63] tools headers UAPI: Sync linux/const.h with the kernel headers Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 24/63] cgroup: Fix memory leak when parsing multiple source parameters Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 25/63] zlib: move EXPORT_SYMBOL() and MODULE_LICENSE() out of dfltcc_syms.c Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 26/63] scsi: cxgb4i: Fix TLS dependency Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 27/63] Bluetooth: hci_h5: close serdev device and free hu in h5_close Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 28/63] fbcon: Disable accelerated scrolling Greg Kroah-Hartman
2021-01-07  8:13   ` Geert Uytterhoeven
2021-01-04 15:57 ` [PATCH 5.10 29/63] reiserfs: add check for an invalid ih_entry_count Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 30/63] misc: vmw_vmci: fix kernel info-leak by initializing dbells in vmci_ctx_get_chkpt_doorbells() Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 31/63] media: gp8psk: initialize stats at power control logic Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 32/63] f2fs: fix shift-out-of-bounds in sanity_check_raw_super() Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 33/63] ALSA: seq: Use bool for snd_seq_queue internal flags Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 34/63] ALSA: rawmidi: Access runtime->avail always in spinlock Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 35/63] bfs: dont use WARNING: string when its just info Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 36/63] ext4: check for invalid block size early when mounting a file system Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 37/63] fcntl: Fix potential deadlock in send_sig{io, urg}() Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 38/63] io_uring: check kthread stopped flag when sq thread is unparked Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 39/63] rtc: sun6i: Fix memleak in sun6i_rtc_clk_init Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 40/63] module: set MODULE_STATE_GOING state when a module fails to load Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 41/63] quota: Dont overflow quota file offsets Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 42/63] rtc: pl031: fix resource leak in pl031_probe Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 43/63] powerpc: sysdev: add missing iounmap() on error in mpic_msgr_probe() Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 44/63] i3c master: fix missing destroy_workqueue() on error in i3c_master_register Greg Kroah-Hartman
2021-01-04 15:57 ` Greg Kroah-Hartman [this message]
2021-01-04 15:57 ` [PATCH 5.10 46/63] f2fs: avoid race condition for shrinker count Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 47/63] f2fs: fix race of pending_pages in decompression Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 48/63] module: delay kobject uevent until after module init call Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 49/63] powerpc/64: irq replay remove decrementer overflow check Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 50/63] fs/namespace.c: WARN if mnt_count has become negative Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 51/63] watchdog: rti-wdt: fix reference leak in rti_wdt_probe Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 52/63] um: random: Register random as hwrng-core device Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 53/63] um: ubd: Submit all data segments atomically Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 54/63] NFSv4.2: Dont error when exiting early on a READ_PLUS buffer overflow Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 55/63] ceph: fix inode refcount leak when ceph_fill_inode on non-I_NEW inode fails Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 56/63] drm/amd/display: updated wm table for Renoir Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 57/63] tick/sched: Remove bogus boot "safety" check Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 58/63] s390: always clear kernel stack backchain before calling functions Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 59/63] io_uring: remove racy overflow list fast checks Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 60/63] ALSA: pcm: Clear the full allocated memory at hw_params Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 61/63] dm verity: skip verity work if I/O error when system is shutting down Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 62/63] ext4: avoid s_mb_prefetch to be zero in individual scenarios Greg Kroah-Hartman
2021-01-04 15:57 ` [PATCH 5.10 63/63] device-dax: Fix range release Greg Kroah-Hartman
2021-01-05  6:06 ` [PATCH 5.10 00/63] 5.10.5-rc1 review Daniel Díaz
2021-01-05 12:55 ` Jeffrin Jose T
2021-01-05 13:05   ` Greg Kroah-Hartman
2021-01-06 18:43     ` Jeffrin Jose T
2021-01-05 16:38 ` Shuah Khan
2021-01-05 18:17 ` Guenter Roeck

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=20210104155711.007242412@linuxfoundation.org \
    --to=gregkh@linuxfoundation.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=sashal@kernel.org \
    --cc=stable@vger.kernel.org \
    --cc=trond.myklebust@hammerspace.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