From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
stable@vger.kernel.org, Eric Dumazet <edumazet@google.com>,
Saeed Mahameed <saeedm@mellanox.com>,
Tariq Toukan <tariqt@mellanox.com>,
"David S. Miller" <davem@davemloft.net>
Subject: [PATCH 4.9 37/63] net/mlx4_en: Force CHECKSUM_NONE for short ethernet frames
Date: Mon, 25 Feb 2019 22:11:37 +0100 [thread overview]
Message-ID: <20190225195038.653707314@linuxfoundation.org> (raw)
In-Reply-To: <20190225195035.713274200@linuxfoundation.org>
4.9-stable review patch. If anyone has any objections, please let me know.
------------------
From: Saeed Mahameed <saeedm@mellanox.com>
[ Upstream commit 29dded89e80e3fff61efb34f07a8a3fba3ea146d ]
When an ethernet frame is padded to meet the minimum ethernet frame
size, the padding octets are not covered by the hardware checksum.
Fortunately the padding octets are usually zero's, which don't affect
checksum. However, it is not guaranteed. For example, switches might
choose to make other use of these octets.
This repeatedly causes kernel hardware checksum fault.
Prior to the cited commit below, skb checksum was forced to be
CHECKSUM_NONE when padding is detected. After it, we need to keep
skb->csum updated. However, fixing up CHECKSUM_COMPLETE requires to
verify and parse IP headers, it does not worth the effort as the packets
are so small that CHECKSUM_COMPLETE has no significant advantage.
Future work: when reporting checksum complete is not an option for
IP non-TCP/UDP packets, we can actually fallback to report checksum
unnecessary, by looking at cqe IPOK bit.
Fixes: 88078d98d1bb ("net: pskb_trim_rcsum() and CHECKSUM_COMPLETE are friends")
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
drivers/net/ethernet/mellanox/mlx4/en_rx.c | 21 ++++++++++++++++++++-
1 file changed, 20 insertions(+), 1 deletion(-)
--- a/drivers/net/ethernet/mellanox/mlx4/en_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_rx.c
@@ -777,13 +777,27 @@ static int get_fixed_ipv6_csum(__wsum hw
return 0;
}
#endif
+
+#define short_frame(size) ((size) <= ETH_ZLEN + ETH_FCS_LEN)
+
static int check_csum(struct mlx4_cqe *cqe, struct sk_buff *skb, void *va,
netdev_features_t dev_features)
{
__wsum hw_checksum = 0;
+ void *hdr;
- void *hdr = (u8 *)va + sizeof(struct ethhdr);
+ /* CQE csum doesn't cover padding octets in short ethernet
+ * frames. And the pad field is appended prior to calculating
+ * and appending the FCS field.
+ *
+ * Detecting these padded frames requires to verify and parse
+ * IP headers, so we simply force all those small frames to skip
+ * checksum complete.
+ */
+ if (short_frame(skb->len))
+ return -EINVAL;
+ hdr = (u8 *)va + sizeof(struct ethhdr);
hw_checksum = csum_unfold((__force __sum16)cqe->checksum);
if (cqe->vlan_my_qpn & cpu_to_be32(MLX4_CQE_CVLAN_PRESENT_MASK) &&
@@ -945,6 +959,11 @@ xdp_drop:
}
if (likely(dev->features & NETIF_F_RXCSUM)) {
+ /* TODO: For IP non TCP/UDP packets when csum complete is
+ * not an option (not supported or any other reason) we can
+ * actually check cqe IPOK status bit and report
+ * CHECKSUM_UNNECESSARY rather than CHECKSUM_NONE
+ */
if (cqe->status & cpu_to_be16(MLX4_CQE_STATUS_TCP |
MLX4_CQE_STATUS_UDP)) {
if ((cqe->status & cpu_to_be16(MLX4_CQE_STATUS_IPOK)) &&
next prev parent reply other threads:[~2019-02-25 22:03 UTC|newest]
Thread overview: 69+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-02-25 21:11 [PATCH 4.9 00/63] 4.9.161-stable review Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 01/63] mac80211: Free mpath object when rhashtable insertion fails Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 02/63] libceph: handle an empty authorize reply Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 03/63] ceph: avoid repeatedly adding inode to mdsc->snap_flush_list Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 04/63] numa: change get_mempolicy() to use nr_node_ids instead of MAX_NUMNODES Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 05/63] proc, oom: do not report alien mms when setting oom_score_adj Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 06/63] KEYS: allow reaching the keys quotas exactly Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 07/63] mfd: ti_am335x_tscadc: Use PLATFORM_DEVID_AUTO while registering mfd cells Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 08/63] mfd: twl-core: Fix section annotations on {,un}protect_pm_master Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 09/63] mfd: db8500-prcmu: Fix some section annotations Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 10/63] mfd: mt6397: Do not call irq_domain_remove if PMIC unsupported Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 11/63] mfd: ab8500-core: Return zero in get_register_interruptible() Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 12/63] mfd: qcom_rpm: write fw_version to CTRL_REG Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 13/63] mfd: wm5110: Add missing ASRC rate register Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 14/63] mfd: mc13xxx: Fix a missing check of a register-read failure Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 15/63] qed: Fix qed_ll2_post_rx_buffer_notify_fw() by adding a write memory barrier Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 16/63] net: hns: Fix use after free identified by SLUB debug Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 17/63] MIPS: ath79: Enable OF serial ports in the default config Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 18/63] scsi: qla4xxx: check return code of qla4xxx_copy_from_fwddb_param Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 19/63] scsi: isci: initialize shost fully before calling scsi_add_host() Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 20/63] MIPS: jazz: fix 64bit build Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 21/63] net: stmmac: Fix PCI module removal leak Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 22/63] isdn: i4l: isdn_tty: Fix some concurrency double-free bugs Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 23/63] atm: he: fix sign-extension overflow on large shift Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 24/63] leds: lp5523: fix a missing check of return value of lp55xx_read Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 25/63] mlxsw: spectrum_switchdev: Do not treat static FDB entries as sticky Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 26/63] net/mlx5e: Fix wrong (zero) TX drop counter indication for representor Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 27/63] isdn: avm: Fix string plus integer warning from Clang Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 28/63] net: ethernet: stmmac: change dma descriptors to __le32 Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 29/63] RDMA/srp: Rework SCSI device reset handling Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 30/63] KEYS: user: Align the payload buffer Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 31/63] KEYS: always initialize keyring_index_key::desc_len Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 32/63] batman-adv: fix uninit-value in batadv_interface_tx() Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 33/63] net/packet: fix 4gb buffer limit due to overflow check Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 34/63] team: avoid complex list operations in team_nl_cmd_options_set() Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 35/63] sit: check if IPv6 enabled before calling ip6_err_gen_icmpv6_unreach() Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 36/63] sctp: call gso_reset_checksum when computing checksum in sctp_gso_segment Greg Kroah-Hartman
2019-02-25 21:11 ` Greg Kroah-Hartman [this message]
2019-02-25 21:11 ` [PATCH 4.9 38/63] parisc: Fix ptrace syscall number modification Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 39/63] ARCv2: Enable unaligned access in early ASM code Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 40/63] ARC: U-boot: check arguments paranoidly Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 41/63] ARC: define ARCH_SLAB_MINALIGN = 8 Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 42/63] hpet: Make cmd parameter of hpet_ioctl_common() unsigned Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 43/63] clocksource: Use GENMASK_ULL in definition of CLOCKSOURCE_MASK Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 44/63] netpoll: Fix device name check in netpoll_setup() Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 45/63] tracing: Use cpumask_available() to check if cpumask variable may be used Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 46/63] x86/boot: Disable the address-of-packed-member compiler warning Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 47/63] drm/i915: Consistently use enum pipe for PCH transcoders Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 48/63] drm/i915: Fix enum pipe vs. enum transcoder for the PCH transcoder Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 49/63] kbuild: move cc-option and cc-disable-warning after incl. arch Makefile Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 50/63] kbuild: clang: fix build failures with sparse check Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 51/63] kbuild: clang: remove crufty HOSTCFLAGS Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 52/63] kbuild: clang: disable unused variable warnings only when constant Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 53/63] kbuild: set no-integrated-as before incl. arch Makefile Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 54/63] kbuild: add -no-integrated-as Clang option unconditionally Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 55/63] irqchip/gic-v3: Convert arm64 GIC accessors to {read,write}_sysreg_s Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 56/63] mm/zsmalloc.c: change stat type parameter to int Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 57/63] mm/zsmalloc.c: fix -Wunneeded-internal-declaration warning Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 58/63] Revert "bridge: do not add port to router list when receives query with source 0.0.0.0" Greg Kroah-Hartman
2019-02-25 21:11 ` [PATCH 4.9 59/63] netfilter: nf_tables: fix flush after rule deletion in the same batch Greg Kroah-Hartman
2019-02-25 21:12 ` [PATCH 4.9 60/63] pinctrl: max77620: Use define directive for max77620_pinconf_param values Greg Kroah-Hartman
2019-02-25 21:12 ` [PATCH 4.9 61/63] phy: tegra: remove redundant self assignment of map Greg Kroah-Hartman
2019-02-25 21:12 ` [PATCH 4.9 62/63] sched/sysctl: Fix attributes of some extern declarations Greg Kroah-Hartman
2019-02-25 21:12 ` [PATCH 4.9 63/63] kbuild: consolidate Clang compiler flags Greg Kroah-Hartman
2019-02-26 8:33 ` [PATCH 4.9 00/63] 4.9.161-stable review Naresh Kamboju
2019-02-26 12:23 ` Jon Hunter
2019-02-26 15:06 ` shuah
2019-02-26 17:41 ` Guenter Roeck
2019-02-26 18:18 ` 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=20190225195038.653707314@linuxfoundation.org \
--to=gregkh@linuxfoundation.org \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=linux-kernel@vger.kernel.org \
--cc=saeedm@mellanox.com \
--cc=stable@vger.kernel.org \
--cc=tariqt@mellanox.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;
as well as URLs for NNTP newsgroup(s).