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, kernel test robot <rong.a.chen@intel.com>,
	Florian Westphal <fw@strlen.de>,
	Pablo Neira Ayuso <pablo@netfilter.org>
Subject: [PATCH 5.4 76/76] netfilter: nft_compat: remove flush counter optimization
Date: Mon, 18 Jan 2021 12:35:16 +0100	[thread overview]
Message-ID: <20210118113344.604874964@linuxfoundation.org> (raw)
In-Reply-To: <20210118113340.984217512@linuxfoundation.org>

From: Florian Westphal <fw@strlen.de>

commit 2f941622fd88328ca75806c45c9e9709286a0609 upstream.

WARNING: CPU: 1 PID: 16059 at lib/refcount.c:31 refcount_warn_saturate+0xdf/0xf
[..]
 __nft_mt_tg_destroy+0x42/0x50 [nft_compat]
 nft_target_destroy+0x63/0x80 [nft_compat]
 nf_tables_expr_destroy+0x1b/0x30 [nf_tables]
 nf_tables_rule_destroy+0x3a/0x70 [nf_tables]
 nf_tables_exit_net+0x186/0x3d0 [nf_tables]

Happens when a compat expr is destoyed from abort path.
There is no functional impact; after this work queue is flushed
unconditionally if its pending.

This removes the waitcount optimization.  Test of repeated
iptables-restore of a ~60k kubernetes ruleset doesn't indicate
a slowdown.  In case the counter is needed after all for some workloads
we can revert this and increment the refcount for the
!= NFT_PREPARE_TRANS case to avoid the increment/decrement imbalance.

While at it, also flush for match case, this was an oversight
in the original patch.

Fixes: ffe8923f109b7e ("netfilter: nft_compat: make sure xtables destructors have run")
Reported-by: kernel test robot <rong.a.chen@intel.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 net/netfilter/nft_compat.c |   37 ++++++++++++++-----------------------
 1 file changed, 14 insertions(+), 23 deletions(-)

--- a/net/netfilter/nft_compat.c
+++ b/net/netfilter/nft_compat.c
@@ -27,8 +27,6 @@ struct nft_xt_match_priv {
 	void *info;
 };
 
-static refcount_t nft_compat_pending_destroy = REFCOUNT_INIT(1);
-
 static int nft_compat_chain_validate_dependency(const struct nft_ctx *ctx,
 						const char *tablename)
 {
@@ -215,6 +213,17 @@ static int nft_parse_compat(const struct
 	return 0;
 }
 
+static void nft_compat_wait_for_destructors(void)
+{
+	/* xtables matches or targets can have side effects, e.g.
+	 * creation/destruction of /proc files.
+	 * The xt ->destroy functions are run asynchronously from
+	 * work queue.  If we have pending invocations we thus
+	 * need to wait for those to finish.
+	 */
+	nf_tables_trans_destroy_flush_work();
+}
+
 static int
 nft_target_init(const struct nft_ctx *ctx, const struct nft_expr *expr,
 		const struct nlattr * const tb[])
@@ -238,14 +247,7 @@ nft_target_init(const struct nft_ctx *ct
 
 	nft_target_set_tgchk_param(&par, ctx, target, info, &e, proto, inv);
 
-	/* xtables matches or targets can have side effects, e.g.
-	 * creation/destruction of /proc files.
-	 * The xt ->destroy functions are run asynchronously from
-	 * work queue.  If we have pending invocations we thus
-	 * need to wait for those to finish.
-	 */
-	if (refcount_read(&nft_compat_pending_destroy) > 1)
-		nf_tables_trans_destroy_flush_work();
+	nft_compat_wait_for_destructors();
 
 	ret = xt_check_target(&par, size, proto, inv);
 	if (ret < 0)
@@ -260,7 +262,6 @@ nft_target_init(const struct nft_ctx *ct
 
 static void __nft_mt_tg_destroy(struct module *me, const struct nft_expr *expr)
 {
-	refcount_dec(&nft_compat_pending_destroy);
 	module_put(me);
 	kfree(expr->ops);
 }
@@ -468,6 +469,8 @@ __nft_match_init(const struct nft_ctx *c
 
 	nft_match_set_mtchk_param(&par, ctx, match, info, &e, proto, inv);
 
+	nft_compat_wait_for_destructors();
+
 	return xt_check_match(&par, size, proto, inv);
 }
 
@@ -716,14 +719,6 @@ static const struct nfnetlink_subsystem
 
 static struct nft_expr_type nft_match_type;
 
-static void nft_mt_tg_deactivate(const struct nft_ctx *ctx,
-				 const struct nft_expr *expr,
-				 enum nft_trans_phase phase)
-{
-	if (phase == NFT_TRANS_COMMIT)
-		refcount_inc(&nft_compat_pending_destroy);
-}
-
 static const struct nft_expr_ops *
 nft_match_select_ops(const struct nft_ctx *ctx,
 		     const struct nlattr * const tb[])
@@ -762,7 +757,6 @@ nft_match_select_ops(const struct nft_ct
 	ops->type = &nft_match_type;
 	ops->eval = nft_match_eval;
 	ops->init = nft_match_init;
-	ops->deactivate = nft_mt_tg_deactivate,
 	ops->destroy = nft_match_destroy;
 	ops->dump = nft_match_dump;
 	ops->validate = nft_match_validate;
@@ -853,7 +847,6 @@ nft_target_select_ops(const struct nft_c
 	ops->size = NFT_EXPR_SIZE(XT_ALIGN(target->targetsize));
 	ops->init = nft_target_init;
 	ops->destroy = nft_target_destroy;
-	ops->deactivate = nft_mt_tg_deactivate,
 	ops->dump = nft_target_dump;
 	ops->validate = nft_target_validate;
 	ops->data = target;
@@ -917,8 +910,6 @@ static void __exit nft_compat_module_exi
 	nfnetlink_subsys_unregister(&nfnl_compat_subsys);
 	nft_unregister_expr(&nft_target_type);
 	nft_unregister_expr(&nft_match_type);
-
-	WARN_ON_ONCE(refcount_read(&nft_compat_pending_destroy) != 1);
 }
 
 MODULE_ALIAS_NFNL_SUBSYS(NFNL_SUBSYS_NFT_COMPAT);



  parent reply	other threads:[~2021-01-18 18:58 UTC|newest]

Thread overview: 79+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-01-18 11:34 [PATCH 5.4 00/76] 5.4.91-rc1 review Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 01/76] kbuild: enforce -Werror=return-type Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 02/76] btrfs: prevent NULL pointer dereference in extent_io_tree_panic Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 03/76] ASoC: dapm: remove widget from dirty list on free Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 04/76] x86/hyperv: check cpu mask after interrupt has been disabled Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 05/76] tracing/kprobes: Do the notrace functions check without kprobes on ftrace Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 06/76] mips: fix Section mismatch in reference Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 07/76] mips: lib: uncached: fix non-standard usage of variable sp Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 08/76] MIPS: boot: Fix unaligned access with CONFIG_MIPS_RAW_APPENDED_DTB Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 09/76] MIPS: Fix malformed NT_FILE and NT_SIGINFO in 32bit coredumps Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 10/76] MIPS: relocatable: fix possible boot hangup with KASLR enabled Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 11/76] RDMA/ocrdma: Fix use after free in ocrdma_dealloc_ucontext_pd() Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 12/76] ACPI: scan: Harden acpi_device_add() against device ID overflows Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 13/76] mm/hugetlb: fix potential missing huge page size info Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 14/76] dm raid: fix discard limits for raid1 Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 15/76] dm snapshot: flush merged data before committing metadata Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 16/76] dm integrity: fix the maximum number of arguments Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 17/76] r8152: Add Lenovo Powered USB-C Travel Hub Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 18/76] btrfs: tree-checker: check if chunk item end overflows Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 19/76] drm/i915/backlight: fix CPU mode backlight takeover on LPT Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 20/76] ext4: fix bug for rename with RENAME_WHITEOUT Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 21/76] ext4: dont leak old mountpoint samples Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 22/76] smb3: remove unused flag passed into close functions Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 23/76] cifs: fix interrupted close commands Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 24/76] dm integrity: fix flush with external metadata device Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 25/76] ARC: build: remove non-existing bootpImage from KBUILD_IMAGE Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 26/76] ARC: build: add uImage.lzma to the top-level target Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 27/76] ARC: build: add boot_targets to PHONY Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 28/76] ARC: build: move symlink creation to arch/arc/Makefile to avoid race Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 29/76] netfilter: ipset: fixes possible oops in mtype_resize Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 30/76] btrfs: fix transaction leak and crash after RO remount caused by qgroup rescan Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 31/76] regulator: bd718x7: Add enable times Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 32/76] ethernet: ucc_geth: fix definition and size of ucc_geth_tx_global_pram Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 33/76] habanalabs: register to pci shutdown callback Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 34/76] habanalabs: Fix memleak in hl_device_reset Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 35/76] hwmon: (pwm-fan) Ensure that calculation doesnt discard big period values Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 36/76] lib/raid6: Let $(UNROLL) rules work with macOS userland Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 37/76] bfq: Fix computation of shallow depth Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 38/76] arch/arc: add copy_user_page() to <asm/page.h> to fix build error on ARC Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 39/76] misdn: dsp: select CONFIG_BITREVERSE Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 40/76] net: ethernet: fs_enet: Add missing MODULE_LICENSE Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 41/76] selftests: fix the return value for UDP GRO test Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 42/76] nvme-pci: mark Samsung PM1725a as IGNORE_DEV_SUBNQN Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 43/76] nvmet-rdma: Fix list_del corruption on queue establishment failure Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 44/76] drm/amdgpu: fix a GPU hang issue when remove device Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 45/76] usb: typec: Fix copy paste error for NVIDIA alt-mode description Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 46/76] ACPI: scan: add stub acpi_create_platform_device() for !CONFIG_ACPI Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 47/76] drm/msm: Call msm_init_vram before binding the gpu Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 48/76] ARM: picoxcell: fix missing interrupt-parent properties Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 49/76] perf intel-pt: Fix CPU too large error Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 50/76] dump_common_audit_data(): fix racy accesses to ->d_name Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 51/76] ASoC: meson: axg-tdm-interface: fix loopback Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 52/76] ASoC: meson: axg-tdmin: fix axg skew offset Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 53/76] ASoC: Intel: fix error code cnl_set_dsp_D0() Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 54/76] nvme-tcp: fix possible data corruption with bio merges Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 55/76] NFS4: Fix use-after-free in trace_event_raw_event_nfs4_set_lock Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 56/76] pNFS: We want return-on-close to complete when evicting the inode Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 57/76] pNFS: Mark layout for return if return-on-close was not sent Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 58/76] pNFS: Stricter ordering of layoutget and layoutreturn Greg Kroah-Hartman
2021-01-18 11:34 ` [PATCH 5.4 59/76] NFS/pNFS: Fix a leak of the layout plh_outstanding counter Greg Kroah-Hartman
2021-01-18 11:35 ` [PATCH 5.4 60/76] NFS: nfs_igrab_and_active must first reference the superblock Greg Kroah-Hartman
2021-01-18 11:35 ` [PATCH 5.4 61/76] ext4: fix superblock checksum failure when setting password salt Greg Kroah-Hartman
2021-01-18 11:35 ` [PATCH 5.4 62/76] RDMA/restrack: Dont treat as an error allocation ID wrapping Greg Kroah-Hartman
2021-01-18 11:35 ` [PATCH 5.4 63/76] RDMA/usnic: Fix memleak in find_free_vf_and_create_qp_grp Greg Kroah-Hartman
2021-01-18 11:35 ` [PATCH 5.4 64/76] bnxt_en: Improve stats context resource accounting with RDMA driver loaded Greg Kroah-Hartman
2021-01-18 11:35 ` [PATCH 5.4 65/76] RDMA/mlx5: Fix wrong free of blue flame register on error Greg Kroah-Hartman
2021-01-18 11:35 ` [PATCH 5.4 66/76] IB/mlx5: Fix error unwinding when set_has_smi_cap fails Greg Kroah-Hartman
2021-01-18 11:35 ` [PATCH 5.4 67/76] drm/i915/dsi: Use unconditional msleep for the panel_on_delay when there is no reset-deassert MIPI-sequence Greg Kroah-Hartman
2021-01-18 11:35 ` [PATCH 5.4 68/76] mm, slub: consider rest of partial list if acquire_slab() fails Greg Kroah-Hartman
2021-01-18 11:35 ` [PATCH 5.4 69/76] iommu/vt-d: Fix unaligned addresses for intel_flush_svm_range_dev() Greg Kroah-Hartman
2021-01-18 11:35 ` [PATCH 5.4 70/76] net: sunrpc: interpret the return value of kstrtou32 correctly Greg Kroah-Hartman
2021-01-18 11:35 ` [PATCH 5.4 71/76] dm: eliminate potential source of excessive kernel log noise Greg Kroah-Hartman
2021-01-18 11:35 ` [PATCH 5.4 72/76] ALSA: fireface: Fix integer overflow in transmit_midi_msg() Greg Kroah-Hartman
2021-01-18 11:35 ` [PATCH 5.4 73/76] ALSA: firewire-tascam: Fix integer overflow in midi_port_work() Greg Kroah-Hartman
2021-01-18 11:35 ` [PATCH 5.4 74/76] netfilter: conntrack: fix reading nf_conntrack_buckets Greg Kroah-Hartman
2021-01-18 11:35 ` [PATCH 5.4 75/76] netfilter: nf_nat: Fix memleak in nf_nat_init Greg Kroah-Hartman
2021-01-18 11:35 ` Greg Kroah-Hartman [this message]
2021-01-18 14:43 ` [PATCH 5.4 00/76] 5.4.91-rc1 review Naresh Kamboju
2021-01-18 15:53   ` 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=20210118113344.604874964@linuxfoundation.org \
    --to=gregkh@linuxfoundation.org \
    --cc=fw@strlen.de \
    --cc=linux-kernel@vger.kernel.org \
    --cc=pablo@netfilter.org \
    --cc=rong.a.chen@intel.com \
    --cc=stable@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