public inbox for stable@vger.kernel.org
 help / color / mirror / Atom feed
From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: stable@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	patches@lists.linux.dev,
	Florian Kauer <florian.kauer@linutronix.de>,
	Kurt Kanzenbach <kurt@linutronix.de>,
	Vinicius Costa Gomes <vinicius.gomes@intel.com>,
	Simon Horman <simon.horman@corigine.com>,
	Naama Meir <naamax.meir@linux.intel.com>,
	Tony Nguyen <anthony.l.nguyen@intel.com>,
	Jakub Kicinski <kuba@kernel.org>, Sasha Levin <sashal@kernel.org>
Subject: [PATCH 5.15 54/78] igc: Prevent garbled TX queue with XDP ZEROCOPY
Date: Tue, 25 Jul 2023 12:46:45 +0200	[thread overview]
Message-ID: <20230725104453.364066666@linuxfoundation.org> (raw)
In-Reply-To: <20230725104451.275227789@linuxfoundation.org>

From: Florian Kauer <florian.kauer@linutronix.de>

[ Upstream commit 78adb4bcf99effbb960c5f9091e2e062509d1030 ]

In normal operation, each populated queue item has
next_to_watch pointing to the last TX desc of the packet,
while each cleaned item has it set to 0. In particular,
next_to_use that points to the next (necessarily clean)
item to use has next_to_watch set to 0.

When the TX queue is used both by an application using
AF_XDP with ZEROCOPY as well as a second non-XDP application
generating high traffic, the queue pointers can get in
an invalid state where next_to_use points to an item
where next_to_watch is NOT set to 0.

However, the implementation assumes at several places
that this is never the case, so if it does hold,
bad things happen. In particular, within the loop inside
of igc_clean_tx_irq(), next_to_clean can overtake next_to_use.
Finally, this prevents any further transmission via
this queue and it never gets unblocked or signaled.
Secondly, if the queue is in this garbled state,
the inner loop of igc_clean_tx_ring() will never terminate,
completely hogging a CPU core.

The reason is that igc_xdp_xmit_zc() reads next_to_use
before acquiring the lock, and writing it back
(potentially unmodified) later. If it got modified
before locking, the outdated next_to_use is written
pointing to an item that was already used elsewhere
(and thus next_to_watch got written).

Fixes: 9acf59a752d4 ("igc: Enable TX via AF_XDP zero-copy")
Signed-off-by: Florian Kauer <florian.kauer@linutronix.de>
Reviewed-by: Kurt Kanzenbach <kurt@linutronix.de>
Tested-by: Kurt Kanzenbach <kurt@linutronix.de>
Acked-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Tested-by: Naama Meir <naamax.meir@linux.intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Link: https://lore.kernel.org/r/20230717175444.3217831-1-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/net/ethernet/intel/igc/igc_main.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

--- a/drivers/net/ethernet/intel/igc/igc_main.c
+++ b/drivers/net/ethernet/intel/igc/igc_main.c
@@ -2748,15 +2748,15 @@ static void igc_xdp_xmit_zc(struct igc_r
 	struct netdev_queue *nq = txring_txq(ring);
 	union igc_adv_tx_desc *tx_desc = NULL;
 	int cpu = smp_processor_id();
-	u16 ntu = ring->next_to_use;
 	struct xdp_desc xdp_desc;
-	u16 budget;
+	u16 budget, ntu;
 
 	if (!netif_carrier_ok(ring->netdev))
 		return;
 
 	__netif_tx_lock(nq, cpu);
 
+	ntu = ring->next_to_use;
 	budget = igc_desc_unused(ring);
 
 	while (xsk_tx_peek_desc(pool, &xdp_desc) && budget--) {



  parent reply	other threads:[~2023-07-25 11:11 UTC|newest]

Thread overview: 86+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-07-25 10:45 [PATCH 5.15 00/78] 5.15.123-rc1 review Greg Kroah-Hartman
2023-07-25 10:45 ` [PATCH 5.15 01/78] ALSA: hda/realtek - remove 3k pull low procedure Greg Kroah-Hartman
2023-07-25 10:45 ` [PATCH 5.15 02/78] ALSA: hda/realtek: Add quirk for Clevo NS70AU Greg Kroah-Hartman
2023-07-25 10:45 ` [PATCH 5.15 03/78] ALSA: hda/realtek: Enable Mute LED on HP Laptop 15s-eq2xxx Greg Kroah-Hartman
2023-07-25 10:45 ` [PATCH 5.15 04/78] keys: Fix linking a duplicate key to a keyrings assoc_array Greg Kroah-Hartman
2023-07-25 10:45 ` [PATCH 5.15 05/78] perf probe: Add test for regression introduced by switch to die_get_decl_file() Greg Kroah-Hartman
2023-07-25 10:45 ` [PATCH 5.15 06/78] btrfs: fix warning when putting transaction with qgroups enabled after abort Greg Kroah-Hartman
2023-07-25 10:45 ` [PATCH 5.15 07/78] fuse: revalidate: dont invalidate if interrupted Greg Kroah-Hartman
2023-07-25 10:45 ` [PATCH 5.15 08/78] btrfs: zoned: fix memory leak after finding block group with super blocks Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 09/78] fuse: ioctl: translate ENOSYS in outarg Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 10/78] selftests: tc: set timeout to 15 minutes Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 11/78] selftests: tc: add ct action kconfig dep Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 12/78] regmap: Drop initial version of maximum transfer length fixes Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 13/78] regmap: Account for register length in SMBus I/O limits Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 14/78] can: bcm: Fix UAF in bcm_proc_show() Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 15/78] selftests: tc: add ConnTrack procfs kconfig Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 16/78] drm/client: Fix memory leak in drm_client_target_cloned Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 17/78] drm/client: Fix memory leak in drm_client_modeset_probe Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 18/78] drm/amd/display: Disable MPC split by default on special asic Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 19/78] drm/amd/display: Keep PHY active for DP displays on DCN31 Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 20/78] ASoC: fsl_sai: Disable bit clock with transmitter Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 21/78] ASoC: codecs: wcd938x: fix missing clsh ctrl error handling Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 22/78] ASoC: codecs: wcd-mbhc-v2: fix resource leaks on component remove Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 23/78] ASoC: codecs: wcd938x: " Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 24/78] ASoC: codecs: wcd938x: fix missing mbhc init error handling Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 25/78] ASoC: codecs: wcd934x: fix resource leaks on component remove Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 26/78] ASoC: codecs: wcd938x: fix codec initialisation race Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 27/78] ASoC: codecs: wcd938x: fix soundwire " Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 28/78] ext4: correct inline offset when handling xattrs in inode body Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 29/78] [PATCH AUTOSEL 4.14 1/9] drm/radeon: Fix integer overflow in radeon_cs_parser_init Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 30/78] [PATCH AUTOSEL 4.14 2/9] ALSA: emu10k1: roll up loops in DSP setup code for Audigy Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 31/78] [PATCH AUTOSEL 4.14 3/9] quota: Properly disable quotas when add_dquot_ref() fails Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 32/78] [PATCH AUTOSEL 4.14 4/9] quota: fix warning in dqgrab() Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 33/78] [PATCH AUTOSEL 4.14 5/9] udf: Fix uninitialized array access for some pathnames Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 34/78] [PATCH AUTOSEL 4.14 6/9] fs: jfs: Fix UBSAN: array-index-out-of-bounds in dbAllocDmapLev Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 35/78] [PATCH AUTOSEL 4.14 7/9] MIPS: dec: prom: Address -Warray-bounds warning Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 36/78] [PATCH AUTOSEL 4.14 8/9] FS: JFS: Fix null-ptr-deref Read in txBegin Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 37/78] [PATCH AUTOSEL 4.14 9/9] FS: JFS: Check for read-only mounted filesystem " Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 38/78] spi: bcm63xx: fix max prepend length Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 39/78] fbdev: imxfb: warn about invalid left/right margin Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 40/78] perf build: Fix library not found error when using CSLIBS Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 41/78] pinctrl: amd: Use amd_pinconf_set() for all config options Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 42/78] net: ethernet: ti: cpsw_ale: Fix cpsw_ale_get_field()/cpsw_ale_set_field() Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 43/78] bridge: Add extack warning when enabling STP in netns Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 44/78] ethernet: use eth_hw_addr_set() instead of ether_addr_copy() Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 45/78] of: net: add a helper for loading netdev->dev_addr Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 46/78] ethernet: use of_get_ethdev_address() Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 47/78] net: ethernet: mtk_eth_soc: handle probe deferral Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 48/78] net: sched: cls_bpf: Undo tcf_bind_filter in case of an error Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 49/78] iavf: Fix use-after-free in free_netdev Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 50/78] iavf: Fix out-of-bounds when setting channels on remove Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 51/78] security: keys: Modify mismatched function name Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 52/78] octeontx2-pf: Dont allocate BPIDs for LBK interfaces Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 53/78] bpf: Fix subprog idx logic in check_max_stack_depth Greg Kroah-Hartman
2023-07-25 10:46 ` Greg Kroah-Hartman [this message]
2023-07-25 10:46 ` [PATCH 5.15 55/78] tcp: annotate data-races around tcp_rsk(req)->ts_recent Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 56/78] net: ipv4: Use kfree_sensitive instead of kfree Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 57/78] net:ipv6: check return value of pskb_trim() Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 58/78] Revert "tcp: avoid the lookup process failing to get sk in ehash table" Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 59/78] fbdev: au1200fb: Fix missing IRQ check in au1200fb_drv_probe Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 60/78] llc: Dont drop packet from non-root netns Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 61/78] netfilter: nf_tables: fix spurious set element insertion failure Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 62/78] netfilter: nft_set_pipapo: fix improper element removal Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 63/78] netfilter: nf_tables: skip bound chain in netns release path Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 64/78] netfilter: nf_tables: skip bound chain on rule flush Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 65/78] tcp: annotate data-races around tp->tcp_tx_delay Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 66/78] tcp: annotate data-races around tp->keepalive_time Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 67/78] tcp: annotate data-races around tp->keepalive_intvl Greg Kroah-Hartman
2023-07-25 10:46 ` [PATCH 5.15 68/78] tcp: annotate data-races around tp->keepalive_probes Greg Kroah-Hartman
2023-07-25 10:47 ` [PATCH 5.15 69/78] tcp: annotate data-races around icsk->icsk_syn_retries Greg Kroah-Hartman
2023-07-25 10:47 ` [PATCH 5.15 70/78] tcp: annotate data-races around tp->linger2 Greg Kroah-Hartman
2023-07-25 10:47 ` [PATCH 5.15 71/78] tcp: annotate data-races around rskq_defer_accept Greg Kroah-Hartman
2023-07-25 10:47 ` [PATCH 5.15 72/78] tcp: annotate data-races around tp->notsent_lowat Greg Kroah-Hartman
2023-07-25 10:47 ` [PATCH 5.15 73/78] tcp: annotate data-races around icsk->icsk_user_timeout Greg Kroah-Hartman
2023-07-25 10:47 ` [PATCH 5.15 74/78] tcp: annotate data-races around fastopenq.max_qlen Greg Kroah-Hartman
2023-07-25 10:47 ` [PATCH 5.15 75/78] net: phy: prevent stale pointer dereference in phy_init() Greg Kroah-Hartman
2023-07-25 10:47 ` [PATCH 5.15 76/78] jbd2: recheck chechpointing non-dirty buffer Greg Kroah-Hartman
2023-07-25 10:47 ` [PATCH 5.15 77/78] tracing/histograms: Return an error if we fail to add histogram to hist_vars list Greg Kroah-Hartman
2023-07-25 10:47 ` [PATCH 5.15 78/78] nixge: fix mac address error handling again Greg Kroah-Hartman
2023-07-25 16:27 ` [PATCH 5.15 00/78] 5.15.123-rc1 review Jon Hunter
2023-07-25 18:08 ` SeongJae Park
2023-07-25 20:20 ` Shuah Khan
2023-07-25 21:52 ` Florian Fainelli
2023-07-26 14:19 ` Naresh Kamboju
2023-07-27  0:02 ` Guenter Roeck
2023-07-27  0:10 ` Ron Economos

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=20230725104453.364066666@linuxfoundation.org \
    --to=gregkh@linuxfoundation.org \
    --cc=anthony.l.nguyen@intel.com \
    --cc=florian.kauer@linutronix.de \
    --cc=kuba@kernel.org \
    --cc=kurt@linutronix.de \
    --cc=naamax.meir@linux.intel.com \
    --cc=patches@lists.linux.dev \
    --cc=sashal@kernel.org \
    --cc=simon.horman@corigine.com \
    --cc=stable@vger.kernel.org \
    --cc=vinicius.gomes@intel.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