Linux kernel -stable discussions
 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, Felix Fietkau <nbd@nbd.name>,
	Willem de Bruijn <willemb@google.com>,
	Jakub Kicinski <kuba@kernel.org>, Sasha Levin <sashal@kernel.org>
Subject: [PATCH 6.1 57/97] net: ipv6: fix UDPv6 GSO segmentation with NAT
Date: Wed,  7 May 2025 20:39:32 +0200	[thread overview]
Message-ID: <20250507183809.290394861@linuxfoundation.org> (raw)
In-Reply-To: <20250507183806.987408728@linuxfoundation.org>

6.1-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Felix Fietkau <nbd@nbd.name>

[ Upstream commit b936a9b8d4a585ccb6d454921c36286bfe63e01d ]

If any address or port is changed, update it in all packets and recalculate
checksum.

Fixes: 9fd1ff5d2ac7 ("udp: Support UDP fraglist GRO/GSO.")
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Link: https://patch.msgid.link/20250426153210.14044-1-nbd@nbd.name
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 net/ipv4/udp_offload.c | 61 +++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 60 insertions(+), 1 deletion(-)

diff --git a/net/ipv4/udp_offload.c b/net/ipv4/udp_offload.c
index 2f1f038b0dc1b..d415b4fb2f1f4 100644
--- a/net/ipv4/udp_offload.c
+++ b/net/ipv4/udp_offload.c
@@ -246,6 +246,62 @@ static struct sk_buff *__udpv4_gso_segment_list_csum(struct sk_buff *segs)
 	return segs;
 }
 
+static void __udpv6_gso_segment_csum(struct sk_buff *seg,
+				     struct in6_addr *oldip,
+				     const struct in6_addr *newip,
+				     __be16 *oldport, __be16 newport)
+{
+	struct udphdr *uh = udp_hdr(seg);
+
+	if (ipv6_addr_equal(oldip, newip) && *oldport == newport)
+		return;
+
+	if (uh->check) {
+		inet_proto_csum_replace16(&uh->check, seg, oldip->s6_addr32,
+					  newip->s6_addr32, true);
+
+		inet_proto_csum_replace2(&uh->check, seg, *oldport, newport,
+					 false);
+		if (!uh->check)
+			uh->check = CSUM_MANGLED_0;
+	}
+
+	*oldip = *newip;
+	*oldport = newport;
+}
+
+static struct sk_buff *__udpv6_gso_segment_list_csum(struct sk_buff *segs)
+{
+	const struct ipv6hdr *iph;
+	const struct udphdr *uh;
+	struct ipv6hdr *iph2;
+	struct sk_buff *seg;
+	struct udphdr *uh2;
+
+	seg = segs;
+	uh = udp_hdr(seg);
+	iph = ipv6_hdr(seg);
+	uh2 = udp_hdr(seg->next);
+	iph2 = ipv6_hdr(seg->next);
+
+	if (!(*(const u32 *)&uh->source ^ *(const u32 *)&uh2->source) &&
+	    ipv6_addr_equal(&iph->saddr, &iph2->saddr) &&
+	    ipv6_addr_equal(&iph->daddr, &iph2->daddr))
+		return segs;
+
+	while ((seg = seg->next)) {
+		uh2 = udp_hdr(seg);
+		iph2 = ipv6_hdr(seg);
+
+		__udpv6_gso_segment_csum(seg, &iph2->saddr, &iph->saddr,
+					 &uh2->source, uh->source);
+		__udpv6_gso_segment_csum(seg, &iph2->daddr, &iph->daddr,
+					 &uh2->dest, uh->dest);
+	}
+
+	return segs;
+}
+
 static struct sk_buff *__udp_gso_segment_list(struct sk_buff *skb,
 					      netdev_features_t features,
 					      bool is_ipv6)
@@ -258,7 +314,10 @@ static struct sk_buff *__udp_gso_segment_list(struct sk_buff *skb,
 
 	udp_hdr(skb)->len = htons(sizeof(struct udphdr) + mss);
 
-	return is_ipv6 ? skb : __udpv4_gso_segment_list_csum(skb);
+	if (is_ipv6)
+		return __udpv6_gso_segment_list_csum(skb);
+	else
+		return __udpv4_gso_segment_list_csum(skb);
 }
 
 struct sk_buff *__udp_gso_segment(struct sk_buff *gso_skb,
-- 
2.39.5




  parent reply	other threads:[~2025-05-07 18:48 UTC|newest]

Thread overview: 107+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-05-07 18:38 [PATCH 6.1 00/97] 6.1.138-rc1 review Greg Kroah-Hartman
2025-05-07 18:38 ` [PATCH 6.1 01/97] Revert "rndis_host: Flag RNDIS modems as WWAN devices" Greg Kroah-Hartman
2025-05-07 18:38 ` [PATCH 6.1 02/97] ALSA: usb-audio: Add second USB ID for Jabra Evolve 65 headset Greg Kroah-Hartman
2025-05-07 18:38 ` [PATCH 6.1 03/97] drm/nouveau: Fix WARN_ON in nouveau_fence_context_kill() Greg Kroah-Hartman
2025-05-07 18:38 ` [PATCH 6.1 04/97] EDAC/altera: Test the correct error reg offset Greg Kroah-Hartman
2025-05-07 18:38 ` [PATCH 6.1 05/97] EDAC/altera: Set DDR and SDMMC interrupt mask before registration Greg Kroah-Hartman
2025-05-07 18:38 ` [PATCH 6.1 06/97] i2c: imx-lpi2c: Fix clock count when probe defers Greg Kroah-Hartman
2025-05-07 18:38 ` [PATCH 6.1 07/97] arm64: errata: Add missing sentinels to Spectre-BHB MIDR arrays Greg Kroah-Hartman
2025-05-07 18:38 ` [PATCH 6.1 08/97] parisc: Fix double SIGFPE crash Greg Kroah-Hartman
2025-05-07 18:38 ` [PATCH 6.1 09/97] perf/x86/intel: KVM: Mask PEBS_ENABLE loaded for guest with vCPUs value Greg Kroah-Hartman
2025-05-07 18:38 ` [PATCH 6.1 10/97] amd-xgbe: Fix to ensure dependent features are toggled with RX checksum offload Greg Kroah-Hartman
2025-05-07 18:38 ` [PATCH 6.1 11/97] irqchip/qcom-mpm: Prevent crash when trying to handle non-wake GPIOs Greg Kroah-Hartman
2025-05-07 18:38 ` [PATCH 6.1 12/97] mmc: renesas_sdhi: Fix error handling in renesas_sdhi_probe Greg Kroah-Hartman
2025-05-07 18:38 ` [PATCH 6.1 13/97] wifi: brcm80211: fmac: Add error handling for brcmf_usb_dl_writeimage() Greg Kroah-Hartman
2025-05-07 18:38 ` [PATCH 6.1 14/97] dm-integrity: fix a warning on invalid table line Greg Kroah-Hartman
2025-05-07 18:38 ` [PATCH 6.1 15/97] dm: always update the array size in realloc_argv on success Greg Kroah-Hartman
2025-05-07 18:38 ` [PATCH 6.1 16/97] iommu/amd: Fix potential buffer overflow in parse_ivrs_acpihid Greg Kroah-Hartman
2025-05-07 18:38 ` [PATCH 6.1 17/97] iommu/vt-d: Apply quirk_iommu_igfx for 8086:0044 (QM57/QS57) Greg Kroah-Hartman
2025-05-07 18:38 ` [PATCH 6.1 18/97] platform/x86/intel-uncore-freq: Fix missing uncore sysfs during CPU hotplug Greg Kroah-Hartman
2025-05-07 18:38 ` [PATCH 6.1 19/97] ksmbd: fix use-after-free in kerberos authentication Greg Kroah-Hartman
2025-05-07 18:38 ` [PATCH 6.1 20/97] cpufreq: Avoid using inconsistent policy->min and policy->max Greg Kroah-Hartman
2025-05-07 18:38 ` [PATCH 6.1 21/97] cpufreq: Fix setting policy limits when frequency tables are used Greg Kroah-Hartman
2025-05-07 18:38 ` [PATCH 6.1 22/97] tracing: Fix oob write in trace_seq_to_buffer() Greg Kroah-Hartman
2025-05-07 18:38 ` [PATCH 6.1 23/97] xfs: fix error returns from xfs_bmapi_write Greg Kroah-Hartman
2025-05-07 18:38 ` [PATCH 6.1 24/97] xfs: fix xfs_bmap_add_extent_delay_real for partial conversions Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 25/97] xfs: remove a racy if_bytes check in xfs_reflink_end_cow_extent Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 26/97] xfs: require XFS_SB_FEAT_INCOMPAT_LOG_XATTRS for attr log intent item recovery Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 27/97] xfs: check opcode and iovec count match in xlog_recover_attri_commit_pass2 Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 28/97] xfs: validate recovered name buffers when recovering xattr items Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 29/97] xfs: revert commit 44af6c7e59b12 Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 30/97] xfs: match lock mode in xfs_buffered_write_iomap_begin() Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 31/97] xfs: make the seq argument to xfs_bmapi_convert_delalloc() optional Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 32/97] xfs: make xfs_bmapi_convert_delalloc() to allocate the target offset Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 33/97] xfs: convert delayed extents to unwritten when zeroing post eof blocks Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 34/97] xfs: allow symlinks with short remote targets Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 35/97] xfs: make sure sb_fdblocks is non-negative Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 36/97] xfs: fix freeing speculative preallocations for preallocated files Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 37/97] xfs: allow unlinked symlinks and dirs with zero size Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 38/97] xfs: restrict when we try to align cow fork delalloc to cowextsz hints Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 39/97] KVM: x86: Load DR6 with guest value only before entering .vcpu_run() loop Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 40/97] dm-bufio: dont schedule in atomic context Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 41/97] ASoC: soc-pcm: Fix hw_params() and DAPM widget sequence Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 42/97] wifi: plfxlc: Remove erroneous assert in plfxlc_mac_release Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 43/97] vxlan: vnifilter: Fix unlocked deletion of default FDB entry Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 44/97] net/mlx5: E-Switch, Initialize MAC Address for Default GID Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 45/97] net/mlx5: E-switch, Fix error handling for enabling roce Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 46/97] net: mscc: ocelot: treat 802.1ad tagged traffic as 802.1Q-untagged Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 47/97] net: mscc: ocelot: delete PVID VLAN when readding it as non-PVID Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 48/97] net: ethernet: mtk-star-emac: fix spinlock recursion issues on rx/tx poll Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 49/97] net: ethernet: mtk-star-emac: rearm interrupts in rx_poll only when advised Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 50/97] net_sched: drr: Fix double list add in class with netem as child qdisc Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 51/97] net_sched: hfsc: Fix a UAF vulnerability " Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 52/97] net_sched: ets: Fix double list add " Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 53/97] net_sched: qfq: " Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 54/97] ice: Check VF VSI Pointer Value in ice_vc_add_fdir_fltr() Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 55/97] net: dlink: Correct endianness handling of led_mode Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 56/97] net: dsa: felix: fix broken taprio gate states after clock jump Greg Kroah-Hartman
2025-05-07 18:39 ` Greg Kroah-Hartman [this message]
2025-05-07 18:39 ` [PATCH 6.1 58/97] bnxt_en: Fix coredump logic to free allocated buffer Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 59/97] bnxt_en: Fix out-of-bound memcpy() during ethtool -w Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 60/97] bnxt_en: Fix ethtool -d byte order for 32-bit values Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 61/97] nvme-tcp: fix premature queue removal and I/O failover Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 62/97] net: lan743x: Fix memleak issue when GSO enabled Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 63/97] net: fec: ERR007885 Workaround for conventional TX Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 64/97] net: hns3: store rx VLAN tag offload state for VF Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 65/97] net: hns3: fix an interrupt residual problem Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 66/97] net: hns3: fixed debugfs tm_qset size Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 67/97] net: hns3: defer calling ptp_clock_register() Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 68/97] net: vertexcom: mse102x: Fix possible stuck of SPI interrupt Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 69/97] net: vertexcom: mse102x: Fix LEN_MASK Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 70/97] net: vertexcom: mse102x: Add range check for CMD_RTS Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 71/97] net: vertexcom: mse102x: Fix RX error handling Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 72/97] md: move initialization and destruction of io_acct_set to md.c Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 73/97] PCI: imx6: Skip controller_id generation logic for i.MX7D Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 74/97] sch_htb: make htb_qlen_notify() idempotent Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 75/97] sch_drr: make drr_qlen_notify() idempotent Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 76/97] sch_hfsc: make hfsc_qlen_notify() idempotent Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 77/97] sch_qfq: make qfq_qlen_notify() idempotent Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 78/97] sch_ets: make est_qlen_notify() idempotent Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 79/97] Revert "x86/kexec: Allocate PGD for x86_64 transition page tables separately" Greg Kroah-Hartman
2025-05-07 18:49   ` David Woodhouse
2025-05-07 18:39 ` [PATCH 6.1 80/97] firmware: arm_scmi: Balance device refcount when destroying devices Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 81/97] firmware: arm_ffa: Skip Rx buffer ownership release if not acquired Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 82/97] ARM: dts: opos6ul: add ksz8081 phy properties Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 83/97] net: phy: microchip: force IRQ polling mode for lan88xx Greg Kroah-Hartman
2025-05-07 18:39 ` [PATCH 6.1 84/97] Revert "drm/meson: vclk: fix calculation of 59.94 fractional rates" Greg Kroah-Hartman
2025-05-07 18:40 ` [PATCH 6.1 85/97] irqchip/gic-v2m: Mark a few functions __init Greg Kroah-Hartman
2025-05-07 18:40 ` [PATCH 6.1 86/97] irqchip/gic-v2m: Prevent use after free of gicv2m_get_fwnode() Greg Kroah-Hartman
2025-05-07 18:40 ` [PATCH 6.1 87/97] memcg: drain obj stock on cpu hotplug teardown Greg Kroah-Hartman
2025-05-07 18:40 ` [PATCH 6.1 88/97] riscv: uprobes: Add missing fence.i after building the XOL buffer Greg Kroah-Hartman
2025-05-07 18:40 ` [PATCH 6.1 89/97] iommu/arm-smmu-v3: Use the new rb tree helpers Greg Kroah-Hartman
2025-05-07 18:40 ` [PATCH 6.1 90/97] iommu/arm-smmu-v3: Fix iommu_device_probe bug due to duplicated stream ids Greg Kroah-Hartman
2025-05-07 18:40 ` [PATCH 6.1 91/97] drm/amd/display: phase2 enable mst hdcp multiple displays Greg Kroah-Hartman
2025-05-07 18:40 ` [PATCH 6.1 92/97] drm/amd/display: Clean up style problems in amdgpu_dm_hdcp.c Greg Kroah-Hartman
2025-05-07 18:40 ` [PATCH 6.1 93/97] drm/amd/display: Change HDCP update sequence for DM Greg Kroah-Hartman
2025-05-07 18:40 ` [PATCH 6.1 94/97] drm/amd/display: Add scoped mutexes for amdgpu_dm_dhcp Greg Kroah-Hartman
2025-05-07 18:40 ` [PATCH 6.1 95/97] drm/amd/display: Fix slab-use-after-free in hdcp Greg Kroah-Hartman
2025-05-07 18:40 ` [PATCH 6.1 96/97] ASoC: Use of_property_read_bool() Greg Kroah-Hartman
2025-05-07 18:40 ` [PATCH 6.1 97/97] ASoC: soc-core: Stop using of_property_read_bool() for non-boolean properties Greg Kroah-Hartman
2025-05-08  7:21 ` [PATCH 6.1 00/97] 6.1.138-rc1 review Pavel Machek
2025-05-08  9:45 ` Jon Hunter
2025-05-08  9:48   ` Jon Hunter
2025-05-08  9:52     ` Jon Hunter
2025-05-08 11:24       ` Greg Kroah-Hartman
2025-05-08 12:21         ` Jon Hunter
2025-05-08 11:28 ` Florian Fainelli
2025-05-08 15:00 ` Shuah Khan

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=20250507183809.290394861@linuxfoundation.org \
    --to=gregkh@linuxfoundation.org \
    --cc=kuba@kernel.org \
    --cc=nbd@nbd.name \
    --cc=patches@lists.linux.dev \
    --cc=sashal@kernel.org \
    --cc=stable@vger.kernel.org \
    --cc=willemb@google.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