From: Sasha Levin Date: Sun, 23 Mar 2025 17:44:05 +0000 (-0400) Subject: Fixes for 5.15 X-Git-Tag: v6.1.132~38 X-Git-Url: http://git.ipfire.org/gitweb.cgi?a=commitdiff_plain;h=ccc1f469b00447c4f8c3ee6334ac1097e051f152;p=thirdparty%2Fkernel%2Fstable-queue.git Fixes for 5.15 Signed-off-by: Sasha Levin --- diff --git a/queue-5.15/arm-dts-bcm2711-don-t-mark-timer-regs-unconfigured.patch b/queue-5.15/arm-dts-bcm2711-don-t-mark-timer-regs-unconfigured.patch new file mode 100644 index 0000000000..f8f7a101e6 --- /dev/null +++ b/queue-5.15/arm-dts-bcm2711-don-t-mark-timer-regs-unconfigured.patch @@ -0,0 +1,44 @@ +From 4e767ac23de3b77a9204f76e37f303ee78da6189 Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Sat, 22 Feb 2025 10:41:13 +0100 +Subject: ARM: dts: bcm2711: Don't mark timer regs unconfigured + +From: Phil Elwell + +[ Upstream commit c24f272ae751a9f54f8816430e7f2d56031892cc ] + +During upstream process of Raspberry Pi 4 back in 2019 the ARMv7 stubs +didn't configured the ARM architectural timer. This firmware issue has +been fixed in 2020, which gave users enough time to update their system. + +So drop this property to allow the use of the vDSO version of +clock_gettime. + +Link: https://github.com/raspberrypi/tools/pull/113 +Fixes: 7dbe8c62ceeb ("ARM: dts: Add minimal Raspberry Pi 4 support") +Signed-off-by: Phil Elwell +Signed-off-by: Stefan Wahren +Reviewed-by: Krzysztof Kozlowski +Link: https://lore.kernel.org/r/20250222094113.48198-1-wahrenst@gmx.net +Signed-off-by: Florian Fainelli +Signed-off-by: Sasha Levin +--- + arch/arm/boot/dts/bcm2711.dtsi | 2 -- + 1 file changed, 2 deletions(-) + +diff --git a/arch/arm/boot/dts/bcm2711.dtsi b/arch/arm/boot/dts/bcm2711.dtsi +index 926bc467d4516..da0eb87026028 100644 +--- a/arch/arm/boot/dts/bcm2711.dtsi ++++ b/arch/arm/boot/dts/bcm2711.dtsi +@@ -450,8 +450,6 @@ IRQ_TYPE_LEVEL_LOW)>, + IRQ_TYPE_LEVEL_LOW)>, + ; +- /* This only applies to the ARMv7 stub */ +- arm,cpu-registers-not-fw-configured; + }; + + cpus: cpus { +-- +2.39.5 + diff --git a/queue-5.15/arm-dts-bcm2711-pl011-uarts-are-actually-r1p5.patch b/queue-5.15/arm-dts-bcm2711-pl011-uarts-are-actually-r1p5.patch new file mode 100644 index 0000000000..68457729ff --- /dev/null +++ b/queue-5.15/arm-dts-bcm2711-pl011-uarts-are-actually-r1p5.patch @@ -0,0 +1,74 @@ +From 0eefd4d24972d85a7eac77f96328bc5369f60377 Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Sun, 23 Feb 2025 13:56:13 +0100 +Subject: ARM: dts: bcm2711: PL011 UARTs are actually r1p5 + +From: Phil Elwell + +[ Upstream commit 0de09025f161f67c07978c4742e221243d070d41 ] + +The ARM PL011 UART instances in BCM2711 are r1p5 spec, which means they +have 32-entry FIFOs. The correct periphid value for this is 0x00341011. +Thanks to N Buchwitz for pointing this out. + +Signed-off-by: Phil Elwell +Signed-off-by: Stefan Wahren +Link: https://lore.kernel.org/r/20250223125614.3592-2-wahrenst@gmx.net +Fixes: 7dbe8c62ceeb ("ARM: dts: Add minimal Raspberry Pi 4 support") +Signed-off-by: Florian Fainelli +Signed-off-by: Sasha Levin +--- + arch/arm/boot/dts/bcm2711.dtsi | 9 +++++---- + 1 file changed, 5 insertions(+), 4 deletions(-) + +diff --git a/arch/arm/boot/dts/bcm2711.dtsi b/arch/arm/boot/dts/bcm2711.dtsi +index 89af57482bc8f..926bc467d4516 100644 +--- a/arch/arm/boot/dts/bcm2711.dtsi ++++ b/arch/arm/boot/dts/bcm2711.dtsi +@@ -133,7 +133,7 @@ uart2: serial@7e201400 { + clocks = <&clocks BCM2835_CLOCK_UART>, + <&clocks BCM2835_CLOCK_VPU>; + clock-names = "uartclk", "apb_pclk"; +- arm,primecell-periphid = <0x00241011>; ++ arm,primecell-periphid = <0x00341011>; + status = "disabled"; + }; + +@@ -144,7 +144,7 @@ uart3: serial@7e201600 { + clocks = <&clocks BCM2835_CLOCK_UART>, + <&clocks BCM2835_CLOCK_VPU>; + clock-names = "uartclk", "apb_pclk"; +- arm,primecell-periphid = <0x00241011>; ++ arm,primecell-periphid = <0x00341011>; + status = "disabled"; + }; + +@@ -155,7 +155,7 @@ uart4: serial@7e201800 { + clocks = <&clocks BCM2835_CLOCK_UART>, + <&clocks BCM2835_CLOCK_VPU>; + clock-names = "uartclk", "apb_pclk"; +- arm,primecell-periphid = <0x00241011>; ++ arm,primecell-periphid = <0x00341011>; + status = "disabled"; + }; + +@@ -166,7 +166,7 @@ uart5: serial@7e201a00 { + clocks = <&clocks BCM2835_CLOCK_UART>, + <&clocks BCM2835_CLOCK_VPU>; + clock-names = "uartclk", "apb_pclk"; +- arm,primecell-periphid = <0x00241011>; ++ arm,primecell-periphid = <0x00341011>; + status = "disabled"; + }; + +@@ -1142,6 +1142,7 @@ &txp { + }; + + &uart0 { ++ arm,primecell-periphid = <0x00341011>; + interrupts = ; + }; + +-- +2.39.5 + diff --git a/queue-5.15/bluetooth-fix-error-code-in-chan_alloc_skb_cb.patch b/queue-5.15/bluetooth-fix-error-code-in-chan_alloc_skb_cb.patch new file mode 100644 index 0000000000..863312e17c --- /dev/null +++ b/queue-5.15/bluetooth-fix-error-code-in-chan_alloc_skb_cb.patch @@ -0,0 +1,45 @@ +From 6c4dea43176be6b171bead1a5e9a0a9cfc3d10f1 Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Mon, 10 Mar 2025 22:46:56 +0300 +Subject: Bluetooth: Fix error code in chan_alloc_skb_cb() + +From: Dan Carpenter + +[ Upstream commit 72d061ee630d0dbb45c2920d8d19b3861c413e54 ] + +The chan_alloc_skb_cb() function is supposed to return error pointers on +error. Returning NULL will lead to a NULL dereference. + +Fixes: 6b8d4a6a0314 ("Bluetooth: 6LoWPAN: Use connected oriented channel instead of fixed one") +Signed-off-by: Dan Carpenter +Signed-off-by: Luiz Augusto von Dentz +Signed-off-by: Sasha Levin +--- + net/bluetooth/6lowpan.c | 7 ++++++- + 1 file changed, 6 insertions(+), 1 deletion(-) + +diff --git a/net/bluetooth/6lowpan.c b/net/bluetooth/6lowpan.c +index 580b0940f067a..c4a1b478cf3e9 100644 +--- a/net/bluetooth/6lowpan.c ++++ b/net/bluetooth/6lowpan.c +@@ -824,11 +824,16 @@ static struct sk_buff *chan_alloc_skb_cb(struct l2cap_chan *chan, + unsigned long hdr_len, + unsigned long len, int nb) + { ++ struct sk_buff *skb; ++ + /* Note that we must allocate using GFP_ATOMIC here as + * this function is called originally from netdev hard xmit + * function in atomic context. + */ +- return bt_skb_alloc(hdr_len + len, GFP_ATOMIC); ++ skb = bt_skb_alloc(hdr_len + len, GFP_ATOMIC); ++ if (!skb) ++ return ERR_PTR(-ENOMEM); ++ return skb; + } + + static void chan_suspend_cb(struct l2cap_chan *chan) +-- +2.39.5 + diff --git a/queue-5.15/firmware-imx-scu-fix-of-node-leak-in-.probe.patch b/queue-5.15/firmware-imx-scu-fix-of-node-leak-in-.probe.patch new file mode 100644 index 0000000000..7daa1b44a1 --- /dev/null +++ b/queue-5.15/firmware-imx-scu-fix-of-node-leak-in-.probe.patch @@ -0,0 +1,36 @@ +From 1cbea9ae502e346dc9a22644dcbbf7ffe25046ee Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Tue, 24 Dec 2024 12:34:56 +0900 +Subject: firmware: imx-scu: fix OF node leak in .probe() + +From: Joe Hattori + +[ Upstream commit fbf10b86f6057cf79300720da4ea4b77e6708b0d ] + +imx_scu_probe() calls of_parse_phandle_with_args(), but does not +release the OF node reference obtained by it. Add a of_node_put() call +after done with the node. + +Fixes: f25a066d1a07 ("firmware: imx-scu: Support one TX and one RX") +Signed-off-by: Joe Hattori +Signed-off-by: Shawn Guo +Signed-off-by: Sasha Levin +--- + drivers/firmware/imx/imx-scu.c | 1 + + 1 file changed, 1 insertion(+) + +diff --git a/drivers/firmware/imx/imx-scu.c b/drivers/firmware/imx/imx-scu.c +index dca79caccd01c..fa25c082109ac 100644 +--- a/drivers/firmware/imx/imx-scu.c ++++ b/drivers/firmware/imx/imx-scu.c +@@ -279,6 +279,7 @@ static int imx_scu_probe(struct platform_device *pdev) + return ret; + + sc_ipc->fast_ipc = of_device_is_compatible(args.np, "fsl,imx8-mu-scu"); ++ of_node_put(args.np); + + num_channel = sc_ipc->fast_ipc ? 2 : SCU_MU_CHAN_NUM; + for (i = 0; i < num_channel; i++) { +-- +2.39.5 + diff --git a/queue-5.15/ipv6-fix-memleak-of-nhc_pcpu_rth_output-in-fib_check.patch b/queue-5.15/ipv6-fix-memleak-of-nhc_pcpu_rth_output-in-fib_check.patch new file mode 100644 index 0000000000..afd0f2b3b0 --- /dev/null +++ b/queue-5.15/ipv6-fix-memleak-of-nhc_pcpu_rth_output-in-fib_check.patch @@ -0,0 +1,49 @@ +From 5573e0bf4efbddab8817136c8f2f6e136cfcecc8 Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Tue, 11 Mar 2025 18:03:25 -0700 +Subject: ipv6: Fix memleak of nhc_pcpu_rth_output in fib_check_nh_v6_gw(). + +From: Kuniyuki Iwashima + +[ Upstream commit 9740890ee20e01f99ff1dde84c63dcf089fabb98 ] + +fib_check_nh_v6_gw() expects that fib6_nh_init() cleans up everything +when it fails. + +Commit 7dd73168e273 ("ipv6: Always allocate pcpu memory in a fib6_nh") +moved fib_nh_common_init() before alloc_percpu_gfp() within fib6_nh_init() +but forgot to add cleanup for fib6_nh->nh_common.nhc_pcpu_rth_output in +case it fails to allocate fib6_nh->rt6i_pcpu, resulting in memleak. + +Let's call fib_nh_common_release() and clear nhc_pcpu_rth_output in the +error path. + +Note that we can remove the fib6_nh_release() call in nh_create_ipv6() +later in net-next.git. + +Fixes: 7dd73168e273 ("ipv6: Always allocate pcpu memory in a fib6_nh") +Signed-off-by: Kuniyuki Iwashima +Link: https://patch.msgid.link/20250312010333.56001-1-kuniyu@amazon.com +Signed-off-by: Paolo Abeni +Signed-off-by: Sasha Levin +--- + net/ipv6/route.c | 3 ++- + 1 file changed, 2 insertions(+), 1 deletion(-) + +diff --git a/net/ipv6/route.c b/net/ipv6/route.c +index 94526436b91e8..875cab88f3891 100644 +--- a/net/ipv6/route.c ++++ b/net/ipv6/route.c +@@ -3630,7 +3630,8 @@ int fib6_nh_init(struct net *net, struct fib6_nh *fib6_nh, + in6_dev_put(idev); + + if (err) { +- lwtstate_put(fib6_nh->fib_nh_lws); ++ fib_nh_common_release(&fib6_nh->nh_common); ++ fib6_nh->nh_common.nhc_pcpu_rth_output = NULL; + fib6_nh->fib_nh_lws = NULL; + dev_put(dev); + } +-- +2.39.5 + diff --git a/queue-5.15/ipv6-set-errno-after-ip_fib_metrics_init-in-ip6_rout.patch b/queue-5.15/ipv6-set-errno-after-ip_fib_metrics_init-in-ip6_rout.patch new file mode 100644 index 0000000000..b01024b42e --- /dev/null +++ b/queue-5.15/ipv6-set-errno-after-ip_fib_metrics_init-in-ip6_rout.patch @@ -0,0 +1,57 @@ +From 385ec29722594c6ea50c1c40f700a949dfc3306f Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Tue, 11 Mar 2025 18:38:48 -0700 +Subject: ipv6: Set errno after ip_fib_metrics_init() in + ip6_route_info_create(). + +From: Kuniyuki Iwashima + +[ Upstream commit 9a81fc3480bf5dbe2bf80e278c440770f6ba2692 ] + +While creating a new IPv6, we could get a weird -ENOMEM when +RTA_NH_ID is set and either of the conditions below is true: + + 1) CONFIG_IPV6_SUBTREES is enabled and rtm_src_len is specified + 2) nexthop_get() fails + +e.g.) + + # strace ip -6 route add fe80::dead:beef:dead:beef nhid 1 from :: + recvmsg(3, {msg_iov=[{iov_base=[...[ + {error=-ENOMEM, msg=[... [...]]}, + [{nla_len=49, nla_type=NLMSGERR_ATTR_MSG}, "Nexthops can not be used with so"...] + ]], iov_len=32768}], msg_iovlen=1, msg_controllen=0, msg_flags=0}, 0) = 148 + +Let's set err explicitly after ip_fib_metrics_init() in +ip6_route_info_create(). + +Fixes: f88d8ea67fbd ("ipv6: Plumb support for nexthop object in a fib6_info") +Signed-off-by: Kuniyuki Iwashima +Reviewed-by: David Ahern +Link: https://patch.msgid.link/20250312013854.61125-1-kuniyu@amazon.com +Signed-off-by: Paolo Abeni +Signed-off-by: Sasha Levin +--- + net/ipv6/route.c | 2 ++ + 1 file changed, 2 insertions(+) + +diff --git a/net/ipv6/route.c b/net/ipv6/route.c +index 875cab88f3891..f8b2fdaef67f4 100644 +--- a/net/ipv6/route.c ++++ b/net/ipv6/route.c +@@ -3807,10 +3807,12 @@ static struct fib6_info *ip6_route_info_create(struct fib6_config *cfg, + if (nh) { + if (rt->fib6_src.plen) { + NL_SET_ERR_MSG(extack, "Nexthops can not be used with source routing"); ++ err = -EINVAL; + goto out_free; + } + if (!nexthop_get(nh)) { + NL_SET_ERR_MSG(extack, "Nexthop has been deleted"); ++ err = -ENOENT; + goto out_free; + } + rt->nh = nh; +-- +2.39.5 + diff --git a/queue-5.15/net-atm-fix-use-after-free-in-lec_send.patch b/queue-5.15/net-atm-fix-use-after-free-in-lec_send.patch new file mode 100644 index 0000000000..4651d3594b --- /dev/null +++ b/queue-5.15/net-atm-fix-use-after-free-in-lec_send.patch @@ -0,0 +1,46 @@ +From b170d5d637177259d89a6ecf1faf2963ad9e8374 Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Fri, 14 Mar 2025 13:10:57 +0300 +Subject: net: atm: fix use after free in lec_send() + +From: Dan Carpenter + +[ Upstream commit f3009d0d6ab78053117f8857b921a8237f4d17b3 ] + +The ->send() operation frees skb so save the length before calling +->send() to avoid a use after free. + +Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") +Signed-off-by: Dan Carpenter +Reviewed-by: Simon Horman +Link: https://patch.msgid.link/c751531d-4af4-42fe-affe-6104b34b791d@stanley.mountain +Signed-off-by: Paolo Abeni +Signed-off-by: Sasha Levin +--- + net/atm/lec.c | 3 ++- + 1 file changed, 2 insertions(+), 1 deletion(-) + +diff --git a/net/atm/lec.c b/net/atm/lec.c +index 7226c784dbe0c..ca9952c52fb5c 100644 +--- a/net/atm/lec.c ++++ b/net/atm/lec.c +@@ -181,6 +181,7 @@ static void + lec_send(struct atm_vcc *vcc, struct sk_buff *skb) + { + struct net_device *dev = skb->dev; ++ unsigned int len = skb->len; + + ATM_SKB(skb)->vcc = vcc; + atm_account_tx(vcc, skb); +@@ -191,7 +192,7 @@ lec_send(struct atm_vcc *vcc, struct sk_buff *skb) + } + + dev->stats.tx_packets++; +- dev->stats.tx_bytes += skb->len; ++ dev->stats.tx_bytes += len; + } + + static void lec_tx_timeout(struct net_device *dev, unsigned int txqueue) +-- +2.39.5 + diff --git a/queue-5.15/net-neighbor-add-missing-policy-for-ndtpa_queue_lenb.patch b/queue-5.15/net-neighbor-add-missing-policy-for-ndtpa_queue_lenb.patch new file mode 100644 index 0000000000..d8ee02954c --- /dev/null +++ b/queue-5.15/net-neighbor-add-missing-policy-for-ndtpa_queue_lenb.patch @@ -0,0 +1,39 @@ +From 3b2acf8b0931f72ee0b2343dfb98fbf992438be8 Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Sun, 16 Mar 2025 00:51:13 +0800 +Subject: net/neighbor: add missing policy for NDTPA_QUEUE_LENBYTES + +From: Lin Ma + +[ Upstream commit 90a7138619a0c55e2aefaad27b12ffc2ddbeed78 ] + +Previous commit 8b5c171bb3dc ("neigh: new unresolved queue limits") +introduces new netlink attribute NDTPA_QUEUE_LENBYTES to represent +approximative value for deprecated QUEUE_LEN. However, it forgot to add +the associated nla_policy in nl_ntbl_parm_policy array. Fix it with one +simple NLA_U32 type policy. + +Fixes: 8b5c171bb3dc ("neigh: new unresolved queue limits") +Signed-off-by: Lin Ma +Link: https://patch.msgid.link/20250315165113.37600-1-linma@zju.edu.cn +Signed-off-by: Paolo Abeni +Signed-off-by: Sasha Levin +--- + net/core/neighbour.c | 1 + + 1 file changed, 1 insertion(+) + +diff --git a/net/core/neighbour.c b/net/core/neighbour.c +index 9549738b81842..b83878b5bf788 100644 +--- a/net/core/neighbour.c ++++ b/net/core/neighbour.c +@@ -2173,6 +2173,7 @@ static const struct nla_policy nl_neightbl_policy[NDTA_MAX+1] = { + static const struct nla_policy nl_ntbl_parm_policy[NDTPA_MAX+1] = { + [NDTPA_IFINDEX] = { .type = NLA_U32 }, + [NDTPA_QUEUE_LEN] = { .type = NLA_U32 }, ++ [NDTPA_QUEUE_LENBYTES] = { .type = NLA_U32 }, + [NDTPA_PROXY_QLEN] = { .type = NLA_U32 }, + [NDTPA_APP_PROBES] = { .type = NLA_U32 }, + [NDTPA_UCAST_PROBES] = { .type = NLA_U32 }, +-- +2.39.5 + diff --git a/queue-5.15/rdma-bnxt_re-add-missing-paranthesis-in-map_qp_id_to.patch b/queue-5.15/rdma-bnxt_re-add-missing-paranthesis-in-map_qp_id_to.patch new file mode 100644 index 0000000000..1960de3e15 --- /dev/null +++ b/queue-5.15/rdma-bnxt_re-add-missing-paranthesis-in-map_qp_id_to.patch @@ -0,0 +1,42 @@ +From 62c9eb5d17248de241ef0ac3fb9831447ab01853 Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Mon, 3 Mar 2025 08:59:37 -0800 +Subject: RDMA/bnxt_re: Add missing paranthesis in map_qp_id_to_tbl_indx + +From: Kashyap Desai + +[ Upstream commit 67ee8d496511ad8e1cb88f72944847e7b3e4e47c ] + +The modulo operation returns wrong result without the +paranthesis and that resulted in wrong QP table indexing. + +Fixes: 84cf229f4001 ("RDMA/bnxt_re: Fix the qp table indexing") +Reviewed-by: Kalesh AP +Signed-off-by: Kashyap Desai +Signed-off-by: Selvin Xavier +Link: https://patch.msgid.link/1741021178-2569-3-git-send-email-selvin.xavier@broadcom.com +Signed-off-by: Leon Romanovsky +Signed-off-by: Sasha Levin +--- + drivers/infiniband/hw/bnxt_re/qplib_rcfw.h | 3 ++- + 1 file changed, 2 insertions(+), 1 deletion(-) + +diff --git a/drivers/infiniband/hw/bnxt_re/qplib_rcfw.h b/drivers/infiniband/hw/bnxt_re/qplib_rcfw.h +index aaf06cd939e69..08ea292512797 100644 +--- a/drivers/infiniband/hw/bnxt_re/qplib_rcfw.h ++++ b/drivers/infiniband/hw/bnxt_re/qplib_rcfw.h +@@ -220,9 +220,10 @@ int bnxt_qplib_deinit_rcfw(struct bnxt_qplib_rcfw *rcfw); + int bnxt_qplib_init_rcfw(struct bnxt_qplib_rcfw *rcfw, + struct bnxt_qplib_ctx *ctx, int is_virtfn); + void bnxt_qplib_mark_qp_error(void *qp_handle); ++ + static inline u32 map_qp_id_to_tbl_indx(u32 qid, struct bnxt_qplib_rcfw *rcfw) + { + /* Last index of the qp_tbl is for QP1 ie. qp_tbl_size - 1*/ +- return (qid == 1) ? rcfw->qp_tbl_size - 1 : qid % rcfw->qp_tbl_size - 2; ++ return (qid == 1) ? rcfw->qp_tbl_size - 1 : (qid % (rcfw->qp_tbl_size - 2)); + } + #endif /* __BNXT_QPLIB_RCFW_H__ */ +-- +2.39.5 + diff --git a/queue-5.15/rdma-bnxt_re-avoid-clearing-vlan_id-mask-in-modify-q.patch b/queue-5.15/rdma-bnxt_re-avoid-clearing-vlan_id-mask-in-modify-q.patch new file mode 100644 index 0000000000..8499cdb28f --- /dev/null +++ b/queue-5.15/rdma-bnxt_re-avoid-clearing-vlan_id-mask-in-modify-q.patch @@ -0,0 +1,42 @@ +From 4bf93df296a63dc76a7acabe19dd8a6cad0a4954 Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Mon, 10 Mar 2025 22:16:36 -0700 +Subject: RDMA/bnxt_re: Avoid clearing VLAN_ID mask in modify qp path + +From: Saravanan Vajravel + +[ Upstream commit 81c0db302a674f8004ed805393d17fd76f552e83 ] + +Driver is always clearing the mask that sets the VLAN ID/Service Level +in the adapter. Recent change for supporting multiple traffic class +exposed this issue. + +Allow setting SL and VLAN_ID while QP is moved from INIT to RTR state. + +Fixes: 1ac5a4047975 ("RDMA/bnxt_re: Add bnxt_re RoCE driver") +Fixes: c64b16a37b6d ("RDMA/bnxt_re: Support different traffic class") +Signed-off-by: Saravanan Vajravel +Signed-off-by: Selvin Xavier +Link: https://patch.msgid.link/1741670196-2919-1-git-send-email-selvin.xavier@broadcom.com +Signed-off-by: Leon Romanovsky +Signed-off-by: Sasha Levin +--- + drivers/infiniband/hw/bnxt_re/qplib_fp.c | 2 -- + 1 file changed, 2 deletions(-) + +diff --git a/drivers/infiniband/hw/bnxt_re/qplib_fp.c b/drivers/infiniband/hw/bnxt_re/qplib_fp.c +index 27cf6e62422aa..3725f05ad297b 100644 +--- a/drivers/infiniband/hw/bnxt_re/qplib_fp.c ++++ b/drivers/infiniband/hw/bnxt_re/qplib_fp.c +@@ -1181,8 +1181,6 @@ static void __modify_flags_from_init_state(struct bnxt_qplib_qp *qp) + qp->path_mtu = + CMDQ_MODIFY_QP_PATH_MTU_MTU_2048; + } +- qp->modify_flags &= +- ~CMDQ_MODIFY_QP_MODIFY_MASK_VLAN_ID; + /* Bono FW require the max_dest_rd_atomic to be >= 1 */ + if (qp->max_dest_rd_atomic < 1) + qp->max_dest_rd_atomic = 1; +-- +2.39.5 + diff --git a/queue-5.15/rdma-hns-fix-a-missing-rollback-in-error-path-of-hns.patch b/queue-5.15/rdma-hns-fix-a-missing-rollback-in-error-path-of-hns.patch new file mode 100644 index 0000000000..1a41277c1d --- /dev/null +++ b/queue-5.15/rdma-hns-fix-a-missing-rollback-in-error-path-of-hns.patch @@ -0,0 +1,39 @@ +From 65a53338b19bc2466c8496534c4a276326a65844 Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Tue, 11 Mar 2025 16:48:55 +0800 +Subject: RDMA/hns: Fix a missing rollback in error path of + hns_roce_create_qp_common() + +From: Junxian Huang + +[ Upstream commit 444907dd45cbe62fd69398805b6e2c626fab5b3a ] + +When ib_copy_to_udata() fails in hns_roce_create_qp_common(), +hns_roce_qp_remove() should be called in the error path to +clean up resources in hns_roce_qp_store(). + +Fixes: 0f00571f9433 ("RDMA/hns: Use new SQ doorbell register for HIP09") +Signed-off-by: Junxian Huang +Link: https://patch.msgid.link/20250311084857.3803665-6-huangjunxian6@hisilicon.com +Signed-off-by: Leon Romanovsky +Signed-off-by: Sasha Levin +--- + drivers/infiniband/hw/hns/hns_roce_qp.c | 2 +- + 1 file changed, 1 insertion(+), 1 deletion(-) + +diff --git a/drivers/infiniband/hw/hns/hns_roce_qp.c b/drivers/infiniband/hw/hns/hns_roce_qp.c +index d7f620eb4f21d..3875563abf374 100644 +--- a/drivers/infiniband/hw/hns/hns_roce_qp.c ++++ b/drivers/infiniband/hw/hns/hns_roce_qp.c +@@ -1117,7 +1117,7 @@ static int hns_roce_create_qp_common(struct hns_roce_dev *hr_dev, + min(udata->outlen, sizeof(resp))); + if (ret) { + ibdev_err(ibdev, "copy qp resp failed!\n"); +- goto err_store; ++ goto err_flow_ctrl; + } + } + +-- +2.39.5 + diff --git a/queue-5.15/rdma-hns-fix-soft-lockup-during-bt-pages-loop.patch b/queue-5.15/rdma-hns-fix-soft-lockup-during-bt-pages-loop.patch new file mode 100644 index 0000000000..a5b0e0260f --- /dev/null +++ b/queue-5.15/rdma-hns-fix-soft-lockup-during-bt-pages-loop.patch @@ -0,0 +1,100 @@ +From 0ef240ffdccc3354a9e00b6395618cbb69dc633c Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Tue, 11 Mar 2025 16:48:52 +0800 +Subject: RDMA/hns: Fix soft lockup during bt pages loop + +From: Junxian Huang + +[ Upstream commit 25655580136de59ec89f09089dd28008ea440fc9 ] + +Driver runs a for-loop when allocating bt pages and mapping them with +buffer pages. When a large buffer (e.g. MR over 100GB) is being allocated, +it may require a considerable loop count. This will lead to soft lockup: + + watchdog: BUG: soft lockup - CPU#27 stuck for 22s! + ... + Call trace: + hem_list_alloc_mid_bt+0x124/0x394 [hns_roce_hw_v2] + hns_roce_hem_list_request+0xf8/0x160 [hns_roce_hw_v2] + hns_roce_mtr_create+0x2e4/0x360 [hns_roce_hw_v2] + alloc_mr_pbl+0xd4/0x17c [hns_roce_hw_v2] + hns_roce_reg_user_mr+0xf8/0x190 [hns_roce_hw_v2] + ib_uverbs_reg_mr+0x118/0x290 + + watchdog: BUG: soft lockup - CPU#35 stuck for 23s! + ... + Call trace: + hns_roce_hem_list_find_mtt+0x7c/0xb0 [hns_roce_hw_v2] + mtr_map_bufs+0xc4/0x204 [hns_roce_hw_v2] + hns_roce_mtr_create+0x31c/0x3c4 [hns_roce_hw_v2] + alloc_mr_pbl+0xb0/0x160 [hns_roce_hw_v2] + hns_roce_reg_user_mr+0x108/0x1c0 [hns_roce_hw_v2] + ib_uverbs_reg_mr+0x120/0x2bc + +Add a cond_resched() to fix soft lockup during these loops. In order not +to affect the allocation performance of normal-size buffer, set the loop +count of a 100GB MR as the threshold to call cond_resched(). + +Fixes: 38389eaa4db1 ("RDMA/hns: Add mtr support for mixed multihop addressing") +Signed-off-by: Junxian Huang +Link: https://patch.msgid.link/20250311084857.3803665-3-huangjunxian6@hisilicon.com +Signed-off-by: Leon Romanovsky +Signed-off-by: Sasha Levin +--- + drivers/infiniband/hw/hns/hns_roce_hem.c | 16 +++++++++++++++- + 1 file changed, 15 insertions(+), 1 deletion(-) + +diff --git a/drivers/infiniband/hw/hns/hns_roce_hem.c b/drivers/infiniband/hw/hns/hns_roce_hem.c +index 57b84f5dc0f16..64ee875cc6d1c 100644 +--- a/drivers/infiniband/hw/hns/hns_roce_hem.c ++++ b/drivers/infiniband/hw/hns/hns_roce_hem.c +@@ -1410,6 +1410,11 @@ static int hem_list_alloc_root_bt(struct hns_roce_dev *hr_dev, + return ret; + } + ++/* This is the bottom bt pages number of a 100G MR on 4K OS, assuming ++ * the bt page size is not expanded by cal_best_bt_pg_sz() ++ */ ++#define RESCHED_LOOP_CNT_THRESHOLD_ON_4K 12800 ++ + /* construct the base address table and link them by address hop config */ + int hns_roce_hem_list_request(struct hns_roce_dev *hr_dev, + struct hns_roce_hem_list *hem_list, +@@ -1418,6 +1423,7 @@ int hns_roce_hem_list_request(struct hns_roce_dev *hr_dev, + { + const struct hns_roce_buf_region *r; + int ofs, end; ++ int loop; + int unit; + int ret; + int i; +@@ -1435,7 +1441,10 @@ int hns_roce_hem_list_request(struct hns_roce_dev *hr_dev, + continue; + + end = r->offset + r->count; +- for (ofs = r->offset; ofs < end; ofs += unit) { ++ for (ofs = r->offset, loop = 1; ofs < end; ofs += unit, loop++) { ++ if (!(loop % RESCHED_LOOP_CNT_THRESHOLD_ON_4K)) ++ cond_resched(); ++ + ret = hem_list_alloc_mid_bt(hr_dev, r, unit, ofs, + hem_list->mid_bt[i], + &hem_list->btm_bt); +@@ -1492,9 +1501,14 @@ void *hns_roce_hem_list_find_mtt(struct hns_roce_dev *hr_dev, + struct list_head *head = &hem_list->btm_bt; + struct hns_roce_hem_item *hem, *temp_hem; + void *cpu_base = NULL; ++ int loop = 1; + int nr = 0; + + list_for_each_entry_safe(hem, temp_hem, head, sibling) { ++ if (!(loop % RESCHED_LOOP_CNT_THRESHOLD_ON_4K)) ++ cond_resched(); ++ loop++; ++ + if (hem_list_page_is_in_range(hem, offset)) { + nr = offset - hem->start; + cpu_base = hem->addr + nr * BA_BYTE_LEN; +-- +2.39.5 + diff --git a/queue-5.15/rdma-hns-fix-unmatched-condition-in-error-path-of-al.patch b/queue-5.15/rdma-hns-fix-unmatched-condition-in-error-path-of-al.patch new file mode 100644 index 0000000000..f4a13054a7 --- /dev/null +++ b/queue-5.15/rdma-hns-fix-unmatched-condition-in-error-path-of-al.patch @@ -0,0 +1,64 @@ +From f2a13e377606895ffb3cd07c921203a7ad019d1b Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Tue, 11 Mar 2025 16:48:53 +0800 +Subject: RDMA/hns: Fix unmatched condition in error path of alloc_user_qp_db() + +From: Junxian Huang + +[ Upstream commit b9f59a24ba35a7d955a9f8e148dd9f85b7b40a01 ] + +Currently the condition of unmapping sdb in error path is not exactly +the same as the condition of mapping in alloc_user_qp_db(). This may +cause a problem of unmapping an unmapped db in some case, such as +when the QP is XRC TGT. Unified the two conditions. + +Fixes: 90ae0b57e4a5 ("RDMA/hns: Combine enable flags of qp") +Signed-off-by: Junxian Huang +Link: https://patch.msgid.link/20250311084857.3803665-4-huangjunxian6@hisilicon.com +Signed-off-by: Leon Romanovsky +Signed-off-by: Sasha Levin +--- + drivers/infiniband/hw/hns/hns_roce_qp.c | 8 +++++--- + 1 file changed, 5 insertions(+), 3 deletions(-) + +diff --git a/drivers/infiniband/hw/hns/hns_roce_qp.c b/drivers/infiniband/hw/hns/hns_roce_qp.c +index ff019e32c4552..d7f620eb4f21d 100644 +--- a/drivers/infiniband/hw/hns/hns_roce_qp.c ++++ b/drivers/infiniband/hw/hns/hns_roce_qp.c +@@ -842,12 +842,14 @@ static int alloc_user_qp_db(struct hns_roce_dev *hr_dev, + struct hns_roce_ib_create_qp *ucmd, + struct hns_roce_ib_create_qp_resp *resp) + { ++ bool has_sdb = user_qp_has_sdb(hr_dev, init_attr, udata, resp, ucmd); + struct hns_roce_ucontext *uctx = rdma_udata_to_drv_context(udata, + struct hns_roce_ucontext, ibucontext); ++ bool has_rdb = user_qp_has_rdb(hr_dev, init_attr, udata, resp); + struct ib_device *ibdev = &hr_dev->ib_dev; + int ret; + +- if (user_qp_has_sdb(hr_dev, init_attr, udata, resp, ucmd)) { ++ if (has_sdb) { + ret = hns_roce_db_map_user(uctx, ucmd->sdb_addr, &hr_qp->sdb); + if (ret) { + ibdev_err(ibdev, +@@ -858,7 +860,7 @@ static int alloc_user_qp_db(struct hns_roce_dev *hr_dev, + hr_qp->en_flags |= HNS_ROCE_QP_CAP_SQ_RECORD_DB; + } + +- if (user_qp_has_rdb(hr_dev, init_attr, udata, resp)) { ++ if (has_rdb) { + ret = hns_roce_db_map_user(uctx, ucmd->db_addr, &hr_qp->rdb); + if (ret) { + ibdev_err(ibdev, +@@ -872,7 +874,7 @@ static int alloc_user_qp_db(struct hns_roce_dev *hr_dev, + return 0; + + err_sdb: +- if (hr_qp->en_flags & HNS_ROCE_QP_CAP_SQ_RECORD_DB) ++ if (has_sdb) + hns_roce_db_unmap_user(uctx, &hr_qp->sdb); + err_out: + return ret; +-- +2.39.5 + diff --git a/queue-5.15/rdma-hns-fix-wrong-value-of-max_sge_rd.patch b/queue-5.15/rdma-hns-fix-wrong-value-of-max_sge_rd.patch new file mode 100644 index 0000000000..b5d95ab510 --- /dev/null +++ b/queue-5.15/rdma-hns-fix-wrong-value-of-max_sge_rd.patch @@ -0,0 +1,38 @@ +From 0cff0cad6b65a3df8fca1590f99a95679306c927 Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Tue, 11 Mar 2025 16:48:57 +0800 +Subject: RDMA/hns: Fix wrong value of max_sge_rd + +From: Junxian Huang + +[ Upstream commit 6b5e41a8b51fce520bb09bd651a29ef495e990de ] + +There is no difference between the sge of READ and non-READ +operations in hns RoCE. Set max_sge_rd to the same value as +max_send_sge. + +Fixes: 9a4435375cd1 ("IB/hns: Add driver files for hns RoCE driver") +Signed-off-by: Junxian Huang +Link: https://patch.msgid.link/20250311084857.3803665-8-huangjunxian6@hisilicon.com +Signed-off-by: Leon Romanovsky +Signed-off-by: Sasha Levin +--- + drivers/infiniband/hw/hns/hns_roce_main.c | 2 +- + 1 file changed, 1 insertion(+), 1 deletion(-) + +diff --git a/drivers/infiniband/hw/hns/hns_roce_main.c b/drivers/infiniband/hw/hns/hns_roce_main.c +index e2d2f8f2bdbcf..83a6b8fbe10f0 100644 +--- a/drivers/infiniband/hw/hns/hns_roce_main.c ++++ b/drivers/infiniband/hw/hns/hns_roce_main.c +@@ -185,7 +185,7 @@ static int hns_roce_query_device(struct ib_device *ib_dev, + IB_DEVICE_RC_RNR_NAK_GEN; + props->max_send_sge = hr_dev->caps.max_sq_sg; + props->max_recv_sge = hr_dev->caps.max_rq_sg; +- props->max_sge_rd = 1; ++ props->max_sge_rd = hr_dev->caps.max_sq_sg; + props->max_cq = hr_dev->caps.num_cqs; + props->max_cqe = hr_dev->caps.max_cqes; + props->max_mr = hr_dev->caps.num_mtpts; +-- +2.39.5 + diff --git a/queue-5.15/rdma-hns-remove-redundant-phy_addr-in-hns_roce_hem_l.patch b/queue-5.15/rdma-hns-remove-redundant-phy_addr-in-hns_roce_hem_l.patch new file mode 100644 index 0000000000..97a8cc83d6 --- /dev/null +++ b/queue-5.15/rdma-hns-remove-redundant-phy_addr-in-hns_roce_hem_l.patch @@ -0,0 +1,96 @@ +From 1297ff3f3b955cf76f67d2072a99a72a4fafba7a Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Thu, 22 Sep 2022 20:33:10 +0800 +Subject: RDMA/hns: Remove redundant 'phy_addr' in hns_roce_hem_list_find_mtt() + +From: Chengchang Tang + +[ Upstream commit 5f652387c5423a82453c5cb446a88834bf41a94b ] + +This parameter has never been used. Remove it to simplify the function. + +Link: https://lore.kernel.org/r/20220922123315.3732205-8-xuhaoyue1@hisilicon.com +Signed-off-by: Chengchang Tang +Signed-off-by: Yunsheng Lin +Signed-off-by: Haoyue Xu +Signed-off-by: Jason Gunthorpe +Stable-dep-of: 25655580136d ("RDMA/hns: Fix soft lockup during bt pages loop") +Signed-off-by: Sasha Levin +--- + drivers/infiniband/hw/hns/hns_roce_hem.c | 7 +------ + drivers/infiniband/hw/hns/hns_roce_hem.h | 2 +- + drivers/infiniband/hw/hns/hns_roce_mr.c | 4 ++-- + 3 files changed, 4 insertions(+), 9 deletions(-) + +diff --git a/drivers/infiniband/hw/hns/hns_roce_hem.c b/drivers/infiniband/hw/hns/hns_roce_hem.c +index 99708a7bcda78..57b84f5dc0f16 100644 +--- a/drivers/infiniband/hw/hns/hns_roce_hem.c ++++ b/drivers/infiniband/hw/hns/hns_roce_hem.c +@@ -1487,19 +1487,17 @@ void hns_roce_hem_list_init(struct hns_roce_hem_list *hem_list) + + void *hns_roce_hem_list_find_mtt(struct hns_roce_dev *hr_dev, + struct hns_roce_hem_list *hem_list, +- int offset, int *mtt_cnt, u64 *phy_addr) ++ int offset, int *mtt_cnt) + { + struct list_head *head = &hem_list->btm_bt; + struct hns_roce_hem_item *hem, *temp_hem; + void *cpu_base = NULL; +- u64 phy_base = 0; + int nr = 0; + + list_for_each_entry_safe(hem, temp_hem, head, sibling) { + if (hem_list_page_is_in_range(hem, offset)) { + nr = offset - hem->start; + cpu_base = hem->addr + nr * BA_BYTE_LEN; +- phy_base = hem->dma_addr + nr * BA_BYTE_LEN; + nr = hem->end + 1 - offset; + break; + } +@@ -1508,8 +1506,5 @@ void *hns_roce_hem_list_find_mtt(struct hns_roce_dev *hr_dev, + if (mtt_cnt) + *mtt_cnt = nr; + +- if (phy_addr) +- *phy_addr = phy_base; +- + return cpu_base; + } +diff --git a/drivers/infiniband/hw/hns/hns_roce_hem.h b/drivers/infiniband/hw/hns/hns_roce_hem.h +index fa84ce33076ac..150922b22eaa6 100644 +--- a/drivers/infiniband/hw/hns/hns_roce_hem.h ++++ b/drivers/infiniband/hw/hns/hns_roce_hem.h +@@ -132,7 +132,7 @@ void hns_roce_hem_list_release(struct hns_roce_dev *hr_dev, + struct hns_roce_hem_list *hem_list); + void *hns_roce_hem_list_find_mtt(struct hns_roce_dev *hr_dev, + struct hns_roce_hem_list *hem_list, +- int offset, int *mtt_cnt, u64 *phy_addr); ++ int offset, int *mtt_cnt); + + static inline void hns_roce_hem_first(struct hns_roce_hem *hem, + struct hns_roce_hem_iter *iter) +diff --git a/drivers/infiniband/hw/hns/hns_roce_mr.c b/drivers/infiniband/hw/hns/hns_roce_mr.c +index 604dd38b5c8fd..791a45802d6b2 100644 +--- a/drivers/infiniband/hw/hns/hns_roce_mr.c ++++ b/drivers/infiniband/hw/hns/hns_roce_mr.c +@@ -614,7 +614,7 @@ static int mtr_map_region(struct hns_roce_dev *hr_dev, struct hns_roce_mtr *mtr, + while (offset < end && npage < max_count) { + count = 0; + mtts = hns_roce_hem_list_find_mtt(hr_dev, &mtr->hem_list, +- offset, &count, NULL); ++ offset, &count); + if (!mtts) + return -ENOBUFS; + +@@ -864,7 +864,7 @@ int hns_roce_mtr_find(struct hns_roce_dev *hr_dev, struct hns_roce_mtr *mtr, + mtt_count = 0; + mtts = hns_roce_hem_list_find_mtt(hr_dev, &mtr->hem_list, + start_index + total, +- &mtt_count, NULL); ++ &mtt_count); + if (!mtts || !mtt_count) + goto done; + +-- +2.39.5 + diff --git a/queue-5.15/revert-gre-fix-ipv6-link-local-address-generation.patch b/queue-5.15/revert-gre-fix-ipv6-link-local-address-generation.patch new file mode 100644 index 0000000000..d2afa840b8 --- /dev/null +++ b/queue-5.15/revert-gre-fix-ipv6-link-local-address-generation.patch @@ -0,0 +1,66 @@ +From d618f1a7671c39dc1ee61c5175439326bee07bb2 Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Wed, 19 Mar 2025 22:26:50 +0100 +Subject: Revert "gre: Fix IPv6 link-local address generation." + +From: Guillaume Nault + +[ Upstream commit fc486c2d060f67d672ddad81724f7c8a4d329570 ] + +This reverts commit 183185a18ff96751db52a46ccf93fff3a1f42815. + +This patch broke net/forwarding/ip6gre_custom_multipath_hash.sh in some +circumstances (https://lore.kernel.org/netdev/Z9RIyKZDNoka53EO@mini-arch/). +Let's revert it while the problem is being investigated. + +Fixes: 183185a18ff9 ("gre: Fix IPv6 link-local address generation.") +Signed-off-by: Guillaume Nault +Link: https://patch.msgid.link/8b1ce738eb15dd841aab9ef888640cab4f6ccfea.1742418408.git.gnault@redhat.com +Acked-by: Stanislav Fomichev +Signed-off-by: Paolo Abeni +Signed-off-by: Sasha Levin +--- + net/ipv6/addrconf.c | 15 ++++++--------- + 1 file changed, 6 insertions(+), 9 deletions(-) + +diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c +index f5cca40b71610..932a10f64adcb 100644 +--- a/net/ipv6/addrconf.c ++++ b/net/ipv6/addrconf.c +@@ -3145,13 +3145,16 @@ static void add_v4_addrs(struct inet6_dev *idev) + struct in6_addr addr; + struct net_device *dev; + struct net *net = dev_net(idev->dev); +- int scope, plen; ++ int scope, plen, offset = 0; + u32 pflags = 0; + + ASSERT_RTNL(); + + memset(&addr, 0, sizeof(struct in6_addr)); +- memcpy(&addr.s6_addr32[3], idev->dev->dev_addr, 4); ++ /* in case of IP6GRE the dev_addr is an IPv6 and therefore we use only the last 4 bytes */ ++ if (idev->dev->addr_len == sizeof(struct in6_addr)) ++ offset = sizeof(struct in6_addr) - 4; ++ memcpy(&addr.s6_addr32[3], idev->dev->dev_addr + offset, 4); + + if (!(idev->dev->flags & IFF_POINTOPOINT) && idev->dev->type == ARPHRD_SIT) { + scope = IPV6_ADDR_COMPATv4; +@@ -3459,13 +3462,7 @@ static void addrconf_gre_config(struct net_device *dev) + return; + } + +- /* Generate the IPv6 link-local address using addrconf_addr_gen(), +- * unless we have an IPv4 GRE device not bound to an IP address and +- * which is in EUI64 mode (as __ipv6_isatap_ifid() would fail in this +- * case). Such devices fall back to add_v4_addrs() instead. +- */ +- if (!(dev->type == ARPHRD_IPGRE && *(__be32 *)dev->dev_addr == 0 && +- idev->cnf.addr_gen_mode == IN6_ADDR_GEN_MODE_EUI64)) { ++ if (dev->type == ARPHRD_ETHER) { + addrconf_addr_gen(idev, true); + return; + } +-- +2.39.5 + diff --git a/queue-5.15/series b/queue-5.15/series index ab8d2bca05..825a78af3a 100644 --- a/queue-5.15/series +++ b/queue-5.15/series @@ -76,3 +76,20 @@ drm-amd-display-check-for-invalid-input-params-when-.patch drm-amd-display-fix-null-check-for-pipe_ctx-plane_st.patch smb-client-fix-match_session-bug-preventing-session-reuse.patch smb-client-fix-potential-uaf-in-cifs_debug_files_proc_show.patch +firmware-imx-scu-fix-of-node-leak-in-.probe.patch +xfrm_output-force-software-gso-only-in-tunnel-mode.patch +arm-dts-bcm2711-pl011-uarts-are-actually-r1p5.patch +rdma-bnxt_re-add-missing-paranthesis-in-map_qp_id_to.patch +arm-dts-bcm2711-don-t-mark-timer-regs-unconfigured.patch +rdma-bnxt_re-avoid-clearing-vlan_id-mask-in-modify-q.patch +rdma-hns-remove-redundant-phy_addr-in-hns_roce_hem_l.patch +rdma-hns-fix-soft-lockup-during-bt-pages-loop.patch +rdma-hns-fix-unmatched-condition-in-error-path-of-al.patch +rdma-hns-fix-a-missing-rollback-in-error-path-of-hns.patch +rdma-hns-fix-wrong-value-of-max_sge_rd.patch +bluetooth-fix-error-code-in-chan_alloc_skb_cb.patch +ipv6-fix-memleak-of-nhc_pcpu_rth_output-in-fib_check.patch +ipv6-set-errno-after-ip_fib_metrics_init-in-ip6_rout.patch +net-atm-fix-use-after-free-in-lec_send.patch +net-neighbor-add-missing-policy-for-ndtpa_queue_lenb.patch +revert-gre-fix-ipv6-link-local-address-generation.patch diff --git a/queue-5.15/xfrm_output-force-software-gso-only-in-tunnel-mode.patch b/queue-5.15/xfrm_output-force-software-gso-only-in-tunnel-mode.patch new file mode 100644 index 0000000000..57cde950c4 --- /dev/null +++ b/queue-5.15/xfrm_output-force-software-gso-only-in-tunnel-mode.patch @@ -0,0 +1,77 @@ +From 94722d46a84f0bb1260e8804d22bb96a667d2e6f Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Wed, 19 Feb 2025 12:52:48 +0200 +Subject: xfrm_output: Force software GSO only in tunnel mode + +From: Cosmin Ratiu + +[ Upstream commit 0aae2867aa6067f73d066bc98385e23c8454a1d7 ] + +The cited commit fixed a software GSO bug with VXLAN + IPSec in tunnel +mode. Unfortunately, it is slightly broader than necessary, as it also +severely affects performance for Geneve + IPSec transport mode over a +device capable of both HW GSO and IPSec crypto offload. In this case, +xfrm_output unnecessarily triggers software GSO instead of letting the +HW do it. In simple iperf3 tests over Geneve + IPSec transport mode over +a back-2-back pair of NICs with MTU 1500, the performance was observed +to be up to 6x worse when doing software GSO compared to leaving it to +the hardware. + +This commit makes xfrm_output only trigger software GSO in crypto +offload cases for already encapsulated packets in tunnel mode, as not +doing so would then cause the inner tunnel skb->inner_networking_header +to be overwritten and break software GSO for that packet later if the +device turns out to not be capable of HW GSO. + +Taking a closer look at the conditions for the original bug, to better +understand the reasons for this change: +- vxlan_build_skb -> iptunnel_handle_offloads sets inner_protocol and + inner network header. +- then, udp_tunnel_xmit_skb -> ip_tunnel_xmit adds outer transport and + network headers. +- later in the xmit path, xfrm_output -> xfrm_outer_mode_output -> + xfrm4_prepare_output -> xfrm4_tunnel_encap_add overwrites the inner + network header with the one set in ip_tunnel_xmit before adding the + second outer header. +- __dev_queue_xmit -> validate_xmit_skb checks whether GSO segmentation + needs to happen based on dev features. In the original bug, the hw + couldn't segment the packets, so skb_gso_segment was invoked. +- deep in the .gso_segment callback machinery, __skb_udp_tunnel_segment + tries to use the wrong inner network header, expecting the one set in + iptunnel_handle_offloads but getting the one set by xfrm instead. +- a bit later, ipv6_gso_segment accesses the wrong memory based on that + wrong inner network header. + +With the new change, the original bug (or similar ones) cannot happen +again, as xfrm will now trigger software GSO before applying a tunnel. +This concern doesn't exist in packet offload mode, when the HW adds +encapsulation headers. For the non-offloaded packets (crypto in SW), +software GSO is still done unconditionally in the else branch. + +Reviewed-by: Dragos Tatulea +Reviewed-by: Yael Chemla +Reviewed-by: Leon Romanovsky +Fixes: a204aef9fd77 ("xfrm: call xfrm_output_gso when inner_protocol is set in xfrm_output") +Signed-off-by: Cosmin Ratiu +Signed-off-by: Steffen Klassert +Signed-off-by: Sasha Levin +--- + net/xfrm/xfrm_output.c | 2 +- + 1 file changed, 1 insertion(+), 1 deletion(-) + +diff --git a/net/xfrm/xfrm_output.c b/net/xfrm/xfrm_output.c +index 4dc4a7bbe51cf..29ce7f6f16a09 100644 +--- a/net/xfrm/xfrm_output.c ++++ b/net/xfrm/xfrm_output.c +@@ -737,7 +737,7 @@ int xfrm_output(struct sock *sk, struct sk_buff *skb) + skb->encapsulation = 1; + + if (skb_is_gso(skb)) { +- if (skb->inner_protocol) ++ if (skb->inner_protocol && x->props.mode == XFRM_MODE_TUNNEL) + return xfrm_output_gso(net, sk, skb); + + skb_shinfo(skb)->gso_type |= SKB_GSO_ESP; +-- +2.39.5 +