kernel: bump 5.10 to 5.10.20
authorRui Salvaterra <rsalvaterra@gmail.com>
Thu, 4 Mar 2021 19:51:39 +0000 (19:51 +0000)
committerDavid Bauer <mail@david-bauer.net>
Fri, 5 Mar 2021 19:35:01 +0000 (20:35 +0100)
Also add a new kconfig symbol (CONFIG_KCMP) to the generic config,
disabling the SYS_kcmp syscall (it was split from
CONFIG_CHECKPOINT_RESTORE, which is disabled by default, so the
previous behaviour is kept).

Removed (upstreamed) patches:
  070-net-icmp-pass-zeroed-opts-from-icmp-v6-_ndo_send-bef.patch
  081-wireguard-device-do-not-generate-ICMP-for-non-IP-pac.patch
  082-wireguard-queueing-get-rid-of-per-peer-ring-buffers.patch
  083-wireguard-kconfig-use-arm-chacha-even-with-no-neon.patch
  830-v5.12-0002-usb-serial-option-update-interface-mapping-for-ZTE-P685M.patch

Manually rebased patches:
  313-helios4-dts-status-led-alias.patch
  104-powerpc-mpc85xx-change-P2020RDB-dts-file-for-OpenWRT.patch

Run tested:
  ath79 (TL-WDR3600)
  mvebu (Turris Omnia)

Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
27 files changed:
include/kernel-version.mk
target/linux/bcm63xx/patches-5.10/401-bcm963xx_real_rootfs_length.patch
target/linux/generic/backport-5.10/070-net-icmp-pass-zeroed-opts-from-icmp-v6-_ndo_send-bef.patch [deleted file]
target/linux/generic/backport-5.10/080-wireguard-peer-put-frequently-used-members-above-cac.patch
target/linux/generic/backport-5.10/081-wireguard-device-do-not-generate-ICMP-for-non-IP-pac.patch [deleted file]
target/linux/generic/backport-5.10/082-wireguard-queueing-get-rid-of-per-peer-ring-buffers.patch [deleted file]
target/linux/generic/backport-5.10/083-wireguard-kconfig-use-arm-chacha-even-with-no-neon.patch [deleted file]
target/linux/generic/backport-5.10/830-v5.12-0002-usb-serial-option-update-interface-mapping-for-ZTE-P685M.patch [deleted file]
target/linux/generic/config-5.10
target/linux/generic/hack-5.10/204-module_strip.patch
target/linux/generic/hack-5.10/221-module_exports.patch
target/linux/generic/pending-5.10/203-kallsyms_uncompressed.patch
target/linux/generic/pending-5.10/308-mips32r2_tune.patch
target/linux/generic/pending-5.10/470-mtd-spi-nor-support-limiting-4K-sectors-support-base.patch
target/linux/generic/pending-5.10/479-mtd-spi-nor-add-xtx-xt25f128b.patch
target/linux/generic/pending-5.10/481-mtd-spi-nor-rework-broken-flash-reset-support.patch
target/linux/generic/pending-5.10/703-phy-add-detach-callback-to-struct-phy_driver.patch
target/linux/generic/pending-5.10/920-mangle_bootargs.patch
target/linux/imx6/patches-5.10/300-ARM-dts-imx6q-apalis-ixora-add-status-LEDs-aliases.patch
target/linux/imx6/patches-5.10/301-ARM-dts-imx6q-apalis-ixora-make-switch3-reset-button.patch
target/linux/mediatek/patches-5.10/602-arm64-dts-mediatek-Split-PCIe-node-for-MT2712-MT7622.patch
target/linux/mediatek/patches-5.10/700-net-ethernet-mtk_eth_soc-add-support-for-coherent-DM.patch
target/linux/mediatek/patches-5.10/710-pci-pcie-mediatek-add-support-for-coherent-DMA.patch
target/linux/mpc85xx/patches-5.10/104-powerpc-mpc85xx-change-P2020RDB-dts-file-for-OpenWRT.patch
target/linux/mvebu/patches-5.10/313-helios4-dts-status-led-alias.patch
target/linux/mvebu/patches-5.10/700-mvneta-tx-queue-workaround.patch
target/linux/rockchip/patches-5.10/101-dts-rockchip-add-usb3-controller-node-for-RK3328-SoCs.patch

index 9fbd8614406eec97c4806abd715137fa2fb55bec..33fe6b2d0eeba1c18811e4ed8b340d74d9494ee3 100644 (file)
@@ -7,10 +7,10 @@ ifdef CONFIG_TESTING_KERNEL
 endif
 
 LINUX_VERSION-5.4 = .102
-LINUX_VERSION-5.10 = .18
+LINUX_VERSION-5.10 = .20
 
 LINUX_KERNEL_HASH-5.4.102 = fd697ce1c3f6024d4ae77d4eb5a1552199407b60cb8e90bc621e23cbce639aed
-LINUX_KERNEL_HASH-5.10.18 = 3bc1ee2b1bf73b5ba936721953f3f9599fd165cef906cd5163c68d23cb9bb611
+LINUX_KERNEL_HASH-5.10.20 = 9be37146feba42be05137cf900a7d9012990b5a1d5e59bc0c8da1f86952930a3
 
 remove_uri_prefix=$(subst git://,,$(subst http://,,$(subst https://,,$(1))))
 sanitize_uri=$(call qstrip,$(subst @,_,$(subst :,_,$(subst .,_,$(subst -,_,$(subst /,_,$(1)))))))
index e92ec2d94c1e664dd46437b8359bf4f012a8c2bf..eccdf0d7c5d57c65375e155445f42430ec679904 100644 (file)
@@ -15,7 +15,7 @@
        /* 240-255: Unused at present */
 --- a/drivers/mtd/parsers/parser_imagetag.c
 +++ b/drivers/mtd/parsers/parser_imagetag.c
-@@ -132,7 +132,8 @@ static int bcm963xx_parse_imagetag_parti
+@@ -136,7 +136,8 @@ static int bcm963xx_parse_imagetag_parti
                } else {
                        /* OpenWrt layout */
                        rootfsaddr = kerneladdr + kernellen;
diff --git a/target/linux/generic/backport-5.10/070-net-icmp-pass-zeroed-opts-from-icmp-v6-_ndo_send-bef.patch b/target/linux/generic/backport-5.10/070-net-icmp-pass-zeroed-opts-from-icmp-v6-_ndo_send-bef.patch
deleted file mode 100644 (file)
index d2c0844..0000000
+++ /dev/null
@@ -1,337 +0,0 @@
-From 4a25324891a32d080589a6e3a4dec2be2d9e3d60 Mon Sep 17 00:00:00 2001
-From: "Jason A. Donenfeld" <Jason@zx2c4.com>
-Date: Tue, 23 Feb 2021 14:18:58 +0100
-Subject: [PATCH] net: icmp: pass zeroed opts from icmp{,v6}_ndo_send before
- sending
-
-commit ee576c47db60432c37e54b1e2b43a8ca6d3a8dca upstream.
-
-The icmp{,v6}_send functions make all sorts of use of skb->cb, casting
-it with IPCB or IP6CB, assuming the skb to have come directly from the
-inet layer. But when the packet comes from the ndo layer, especially
-when forwarded, there's no telling what might be in skb->cb at that
-point. As a result, the icmp sending code risks reading bogus memory
-contents, which can result in nasty stack overflows such as this one
-reported by a user:
-
-    panic+0x108/0x2ea
-    __stack_chk_fail+0x14/0x20
-    __icmp_send+0x5bd/0x5c0
-    icmp_ndo_send+0x148/0x160
-
-In icmp_send, skb->cb is cast with IPCB and an ip_options struct is read
-from it. The optlen parameter there is of particular note, as it can
-induce writes beyond bounds. There are quite a few ways that can happen
-in __ip_options_echo. For example:
-
-    // sptr/skb are attacker-controlled skb bytes
-    sptr = skb_network_header(skb);
-    // dptr/dopt points to stack memory allocated by __icmp_send
-    dptr = dopt->__data;
-    // sopt is the corrupt skb->cb in question
-    if (sopt->rr) {
-        optlen  = sptr[sopt->rr+1]; // corrupt skb->cb + skb->data
-        soffset = sptr[sopt->rr+2]; // corrupt skb->cb + skb->data
-       // this now writes potentially attacker-controlled data, over
-       // flowing the stack:
-        memcpy(dptr, sptr+sopt->rr, optlen);
-    }
-
-In the icmpv6_send case, the story is similar, but not as dire, as only
-IP6CB(skb)->iif and IP6CB(skb)->dsthao are used. The dsthao case is
-worse than the iif case, but it is passed to ipv6_find_tlv, which does
-a bit of bounds checking on the value.
-
-This is easy to simulate by doing a `memset(skb->cb, 0x41,
-sizeof(skb->cb));` before calling icmp{,v6}_ndo_send, and it's only by
-good fortune and the rarity of icmp sending from that context that we've
-avoided reports like this until now. For example, in KASAN:
-
-    BUG: KASAN: stack-out-of-bounds in __ip_options_echo+0xa0e/0x12b0
-    Write of size 38 at addr ffff888006f1f80e by task ping/89
-    CPU: 2 PID: 89 Comm: ping Not tainted 5.10.0-rc7-debug+ #5
-    Call Trace:
-     dump_stack+0x9a/0xcc
-     print_address_description.constprop.0+0x1a/0x160
-     __kasan_report.cold+0x20/0x38
-     kasan_report+0x32/0x40
-     check_memory_region+0x145/0x1a0
-     memcpy+0x39/0x60
-     __ip_options_echo+0xa0e/0x12b0
-     __icmp_send+0x744/0x1700
-
-Actually, out of the 4 drivers that do this, only gtp zeroed the cb for
-the v4 case, while the rest did not. So this commit actually removes the
-gtp-specific zeroing, while putting the code where it belongs in the
-shared infrastructure of icmp{,v6}_ndo_send.
-
-This commit fixes the issue by passing an empty IPCB or IP6CB along to
-the functions that actually do the work. For the icmp_send, this was
-already trivial, thanks to __icmp_send providing the plumbing function.
-For icmpv6_send, this required a tiny bit of refactoring to make it
-behave like the v4 case, after which it was straight forward.
-
-Fixes: a2b78e9b2cac ("sunvnet: generate ICMP PTMUD messages for smaller port MTUs")
-Reported-by: SinYu <liuxyon@gmail.com>
-Reviewed-by: Willem de Bruijn <willemb@google.com>
-Link: https://lore.kernel.org/netdev/CAF=yD-LOF116aHub6RMe8vB8ZpnrrnoTdqhobEx+bvoA8AsP0w@mail.gmail.com/T/
-Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
-Link: https://lore.kernel.org/r/20210223131858.72082-1-Jason@zx2c4.com
-Signed-off-by: Jakub Kicinski <kuba@kernel.org>
-[Jason: backported to 5.10]
-Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
----
- include/linux/icmpv6.h | 17 ++++++++++++++---
- include/linux/ipv6.h   |  1 -
- include/net/icmp.h     |  6 +++++-
- net/ipv4/icmp.c        |  5 +++--
- net/ipv6/icmp.c        | 16 ++++++++--------
- net/ipv6/ip6_icmp.c    | 12 +++++++-----
- 6 files changed, 37 insertions(+), 20 deletions(-)
-
---- a/drivers/net/gtp.c
-+++ b/drivers/net/gtp.c
-@@ -539,7 +539,6 @@ static int gtp_build_skb_ip4(struct sk_b
-       if (!skb_is_gso(skb) && (iph->frag_off & htons(IP_DF)) &&
-           mtu < ntohs(iph->tot_len)) {
-               netdev_dbg(dev, "packet too big, fragmentation needed\n");
--              memset(IPCB(skb), 0, sizeof(*IPCB(skb)));
-               icmp_ndo_send(skb, ICMP_DEST_UNREACH, ICMP_FRAG_NEEDED,
-                             htonl(mtu));
-               goto err_rt;
---- a/include/linux/icmpv6.h
-+++ b/include/linux/icmpv6.h
-@@ -3,6 +3,7 @@
- #define _LINUX_ICMPV6_H
- #include <linux/skbuff.h>
-+#include <linux/ipv6.h>
- #include <uapi/linux/icmpv6.h>
- static inline struct icmp6hdr *icmp6_hdr(const struct sk_buff *skb)
-@@ -15,13 +16,16 @@ static inline struct icmp6hdr *icmp6_hdr
- #if IS_ENABLED(CONFIG_IPV6)
- typedef void ip6_icmp_send_t(struct sk_buff *skb, u8 type, u8 code, __u32 info,
--                           const struct in6_addr *force_saddr);
-+                           const struct in6_addr *force_saddr,
-+                           const struct inet6_skb_parm *parm);
- #if IS_BUILTIN(CONFIG_IPV6)
- void icmp6_send(struct sk_buff *skb, u8 type, u8 code, __u32 info,
--              const struct in6_addr *force_saddr);
--static inline void icmpv6_send(struct sk_buff *skb, u8 type, u8 code, __u32 info)
-+              const struct in6_addr *force_saddr,
-+              const struct inet6_skb_parm *parm);
-+static inline void __icmpv6_send(struct sk_buff *skb, u8 type, u8 code, __u32 info,
-+                               const struct inet6_skb_parm *parm)
- {
--      icmp6_send(skb, type, code, info, NULL);
-+      icmp6_send(skb, type, code, info, NULL, parm);
- }
- static inline int inet6_register_icmp_sender(ip6_icmp_send_t *fn)
- {
-@@ -34,18 +38,28 @@ static inline int inet6_unregister_icmp_
-       return 0;
- }
- #else
--extern void icmpv6_send(struct sk_buff *skb, u8 type, u8 code, __u32 info);
-+extern void __icmpv6_send(struct sk_buff *skb, u8 type, u8 code, __u32 info,
-+                        const struct inet6_skb_parm *parm);
- extern int inet6_register_icmp_sender(ip6_icmp_send_t *fn);
- extern int inet6_unregister_icmp_sender(ip6_icmp_send_t *fn);
- #endif
-+static inline void icmpv6_send(struct sk_buff *skb, u8 type, u8 code, __u32 info)
-+{
-+      __icmpv6_send(skb, type, code, info, IP6CB(skb));
-+}
-+
- int ip6_err_gen_icmpv6_unreach(struct sk_buff *skb, int nhs, int type,
-                              unsigned int data_len);
- #if IS_ENABLED(CONFIG_NF_NAT)
- void icmpv6_ndo_send(struct sk_buff *skb_in, u8 type, u8 code, __u32 info);
- #else
--#define icmpv6_ndo_send icmpv6_send
-+static inline void icmpv6_ndo_send(struct sk_buff *skb_in, u8 type, u8 code, __u32 info)
-+{
-+      struct inet6_skb_parm parm = { 0 };
-+      __icmpv6_send(skb_in, type, code, info, &parm);
-+}
- #endif
- #else
---- a/include/linux/ipv6.h
-+++ b/include/linux/ipv6.h
-@@ -84,7 +84,6 @@ struct ipv6_params {
-       __s32 autoconf;
- };
- extern struct ipv6_params ipv6_defaults;
--#include <linux/icmpv6.h>
- #include <linux/tcp.h>
- #include <linux/udp.h>
---- a/include/net/icmp.h
-+++ b/include/net/icmp.h
-@@ -46,7 +46,11 @@ static inline void icmp_send(struct sk_b
- #if IS_ENABLED(CONFIG_NF_NAT)
- void icmp_ndo_send(struct sk_buff *skb_in, int type, int code, __be32 info);
- #else
--#define icmp_ndo_send icmp_send
-+static inline void icmp_ndo_send(struct sk_buff *skb_in, int type, int code, __be32 info)
-+{
-+      struct ip_options opts = { 0 };
-+      __icmp_send(skb_in, type, code, info, &opts);
-+}
- #endif
- int icmp_rcv(struct sk_buff *skb);
---- a/net/ipv4/icmp.c
-+++ b/net/ipv4/icmp.c
-@@ -775,13 +775,14 @@ EXPORT_SYMBOL(__icmp_send);
- void icmp_ndo_send(struct sk_buff *skb_in, int type, int code, __be32 info)
- {
-       struct sk_buff *cloned_skb = NULL;
-+      struct ip_options opts = { 0 };
-       enum ip_conntrack_info ctinfo;
-       struct nf_conn *ct;
-       __be32 orig_ip;
-       ct = nf_ct_get(skb_in, &ctinfo);
-       if (!ct || !(ct->status & IPS_SRC_NAT)) {
--              icmp_send(skb_in, type, code, info);
-+              __icmp_send(skb_in, type, code, info, &opts);
-               return;
-       }
-@@ -796,7 +797,7 @@ void icmp_ndo_send(struct sk_buff *skb_i
-       orig_ip = ip_hdr(skb_in)->saddr;
-       ip_hdr(skb_in)->saddr = ct->tuplehash[0].tuple.src.u3.ip;
--      icmp_send(skb_in, type, code, info);
-+      __icmp_send(skb_in, type, code, info, &opts);
-       ip_hdr(skb_in)->saddr = orig_ip;
- out:
-       consume_skb(cloned_skb);
---- a/net/ipv6/icmp.c
-+++ b/net/ipv6/icmp.c
-@@ -331,10 +331,9 @@ static int icmpv6_getfrag(void *from, ch
- }
- #if IS_ENABLED(CONFIG_IPV6_MIP6)
--static void mip6_addr_swap(struct sk_buff *skb)
-+static void mip6_addr_swap(struct sk_buff *skb, const struct inet6_skb_parm *opt)
- {
-       struct ipv6hdr *iph = ipv6_hdr(skb);
--      struct inet6_skb_parm *opt = IP6CB(skb);
-       struct ipv6_destopt_hao *hao;
-       struct in6_addr tmp;
-       int off;
-@@ -351,7 +350,7 @@ static void mip6_addr_swap(struct sk_buf
-       }
- }
- #else
--static inline void mip6_addr_swap(struct sk_buff *skb) {}
-+static inline void mip6_addr_swap(struct sk_buff *skb, const struct inet6_skb_parm *opt) {}
- #endif
- static struct dst_entry *icmpv6_route_lookup(struct net *net,
-@@ -446,7 +445,8 @@ static int icmp6_iif(const struct sk_buf
-  *    Send an ICMP message in response to a packet in error
-  */
- void icmp6_send(struct sk_buff *skb, u8 type, u8 code, __u32 info,
--              const struct in6_addr *force_saddr)
-+              const struct in6_addr *force_saddr,
-+              const struct inet6_skb_parm *parm)
- {
-       struct inet6_dev *idev = NULL;
-       struct ipv6hdr *hdr = ipv6_hdr(skb);
-@@ -542,7 +542,7 @@ void icmp6_send(struct sk_buff *skb, u8
-       if (!(skb->dev->flags & IFF_LOOPBACK) && !icmpv6_global_allow(net, type))
-               goto out_bh_enable;
--      mip6_addr_swap(skb);
-+      mip6_addr_swap(skb, parm);
-       sk = icmpv6_xmit_lock(net);
-       if (!sk)
-@@ -559,7 +559,7 @@ void icmp6_send(struct sk_buff *skb, u8
-               /* select a more meaningful saddr from input if */
-               struct net_device *in_netdev;
--              in_netdev = dev_get_by_index(net, IP6CB(skb)->iif);
-+              in_netdev = dev_get_by_index(net, parm->iif);
-               if (in_netdev) {
-                       ipv6_dev_get_saddr(net, in_netdev, &fl6.daddr,
-                                          inet6_sk(sk)->srcprefs,
-@@ -640,7 +640,7 @@ EXPORT_SYMBOL(icmp6_send);
-  */
- void icmpv6_param_prob(struct sk_buff *skb, u8 code, int pos)
- {
--      icmp6_send(skb, ICMPV6_PARAMPROB, code, pos, NULL);
-+      icmp6_send(skb, ICMPV6_PARAMPROB, code, pos, NULL, IP6CB(skb));
-       kfree_skb(skb);
- }
-@@ -697,10 +697,10 @@ int ip6_err_gen_icmpv6_unreach(struct sk
-       }
-       if (type == ICMP_TIME_EXCEEDED)
-               icmp6_send(skb2, ICMPV6_TIME_EXCEED, ICMPV6_EXC_HOPLIMIT,
--                         info, &temp_saddr);
-+                         info, &temp_saddr, IP6CB(skb2));
-       else
-               icmp6_send(skb2, ICMPV6_DEST_UNREACH, ICMPV6_ADDR_UNREACH,
--                         info, &temp_saddr);
-+                         info, &temp_saddr, IP6CB(skb2));
-       if (rt)
-               ip6_rt_put(rt);
---- a/net/ipv6/ip6_icmp.c
-+++ b/net/ipv6/ip6_icmp.c
-@@ -33,23 +33,25 @@ int inet6_unregister_icmp_sender(ip6_icm
- }
- EXPORT_SYMBOL(inet6_unregister_icmp_sender);
--void icmpv6_send(struct sk_buff *skb, u8 type, u8 code, __u32 info)
-+void __icmpv6_send(struct sk_buff *skb, u8 type, u8 code, __u32 info,
-+                 const struct inet6_skb_parm *parm)
- {
-       ip6_icmp_send_t *send;
-       rcu_read_lock();
-       send = rcu_dereference(ip6_icmp_send);
-       if (send)
--              send(skb, type, code, info, NULL);
-+              send(skb, type, code, info, NULL, parm);
-       rcu_read_unlock();
- }
--EXPORT_SYMBOL(icmpv6_send);
-+EXPORT_SYMBOL(__icmpv6_send);
- #endif
- #if IS_ENABLED(CONFIG_NF_NAT)
- #include <net/netfilter/nf_conntrack.h>
- void icmpv6_ndo_send(struct sk_buff *skb_in, u8 type, u8 code, __u32 info)
- {
-+      struct inet6_skb_parm parm = { 0 };
-       struct sk_buff *cloned_skb = NULL;
-       enum ip_conntrack_info ctinfo;
-       struct in6_addr orig_ip;
-@@ -57,7 +59,7 @@ void icmpv6_ndo_send(struct sk_buff *skb
-       ct = nf_ct_get(skb_in, &ctinfo);
-       if (!ct || !(ct->status & IPS_SRC_NAT)) {
--              icmpv6_send(skb_in, type, code, info);
-+              __icmpv6_send(skb_in, type, code, info, &parm);
-               return;
-       }
-@@ -72,7 +74,7 @@ void icmpv6_ndo_send(struct sk_buff *skb
-       orig_ip = ipv6_hdr(skb_in)->saddr;
-       ipv6_hdr(skb_in)->saddr = ct->tuplehash[0].tuple.src.u3.in6;
--      icmpv6_send(skb_in, type, code, info);
-+      __icmpv6_send(skb_in, type, code, info, &parm);
-       ipv6_hdr(skb_in)->saddr = orig_ip;
- out:
-       consume_skb(cloned_skb);
index 1e0115d98226e77e544f299dacba195dd8f344a3..444fd677b465a7ca2aca6409362bf16f462ddbbe 100644 (file)
@@ -22,7 +22,7 @@ Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
 --- a/drivers/net/wireguard/peer.h
 +++ b/drivers/net/wireguard/peer.h
 @@ -39,6 +39,7 @@ struct wg_peer {
-       struct crypt_queue tx_queue, rx_queue;
+       struct prev_queue tx_queue, rx_queue;
        struct sk_buff_head staged_packet_queue;
        int serial_work_cpu;
 +      bool is_dead;
diff --git a/target/linux/generic/backport-5.10/081-wireguard-device-do-not-generate-ICMP-for-non-IP-pac.patch b/target/linux/generic/backport-5.10/081-wireguard-device-do-not-generate-ICMP-for-non-IP-pac.patch
deleted file mode 100644 (file)
index 9b1fee7..0000000
+++ /dev/null
@@ -1,47 +0,0 @@
-From 49da2a610d63cef849f0095e601821ad6edfbef7 Mon Sep 17 00:00:00 2001
-From: "Jason A. Donenfeld" <Jason@zx2c4.com>
-Date: Mon, 22 Feb 2021 17:25:47 +0100
-Subject: [PATCH] wireguard: device: do not generate ICMP for non-IP packets
-
-commit 99fff5264e7ab06f45b0ad60243475be0a8d0559 upstream.
-
-If skb->protocol doesn't match the actual skb->data header, it's
-probably not a good idea to pass it off to icmp{,v6}_ndo_send, which is
-expecting to reply to a valid IP packet. So this commit has that early
-mismatch case jump to a later error label.
-
-Fixes: e7096c131e51 ("net: WireGuard secure network tunnel")
-Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
-Signed-off-by: Jakub Kicinski <kuba@kernel.org>
-Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
----
- drivers/net/wireguard/device.c | 7 ++++---
- 1 file changed, 4 insertions(+), 3 deletions(-)
-
---- a/drivers/net/wireguard/device.c
-+++ b/drivers/net/wireguard/device.c
-@@ -138,7 +138,7 @@ static netdev_tx_t wg_xmit(struct sk_buf
-               else if (skb->protocol == htons(ETH_P_IPV6))
-                       net_dbg_ratelimited("%s: No peer has allowed IPs matching %pI6\n",
-                                           dev->name, &ipv6_hdr(skb)->daddr);
--              goto err;
-+              goto err_icmp;
-       }
-       family = READ_ONCE(peer->endpoint.addr.sa_family);
-@@ -201,12 +201,13 @@ static netdev_tx_t wg_xmit(struct sk_buf
- err_peer:
-       wg_peer_put(peer);
--err:
--      ++dev->stats.tx_errors;
-+err_icmp:
-       if (skb->protocol == htons(ETH_P_IP))
-               icmp_ndo_send(skb, ICMP_DEST_UNREACH, ICMP_HOST_UNREACH, 0);
-       else if (skb->protocol == htons(ETH_P_IPV6))
-               icmpv6_ndo_send(skb, ICMPV6_DEST_UNREACH, ICMPV6_ADDR_UNREACH, 0);
-+err:
-+      ++dev->stats.tx_errors;
-       kfree_skb(skb);
-       return ret;
- }
diff --git a/target/linux/generic/backport-5.10/082-wireguard-queueing-get-rid-of-per-peer-ring-buffers.patch b/target/linux/generic/backport-5.10/082-wireguard-queueing-get-rid-of-per-peer-ring-buffers.patch
deleted file mode 100644 (file)
index 1e7c5e8..0000000
+++ /dev/null
@@ -1,560 +0,0 @@
-From 1771bbcc5bc99f569dd82ec9e1b7c397a2fb50ac Mon Sep 17 00:00:00 2001
-From: "Jason A. Donenfeld" <Jason@zx2c4.com>
-Date: Mon, 22 Feb 2021 17:25:48 +0100
-Subject: [PATCH] wireguard: queueing: get rid of per-peer ring buffers
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-commit 8b5553ace83cced775eefd0f3f18b5c6214ccf7a upstream.
-
-Having two ring buffers per-peer means that every peer results in two
-massive ring allocations. On an 8-core x86_64 machine, this commit
-reduces the per-peer allocation from 18,688 bytes to 1,856 bytes, which
-is an 90% reduction. Ninety percent! With some single-machine
-deployments approaching 500,000 peers, we're talking about a reduction
-from 7 gigs of memory down to 700 megs of memory.
-
-In order to get rid of these per-peer allocations, this commit switches
-to using a list-based queueing approach. Currently GSO fragments are
-chained together using the skb->next pointer (the skb_list_* singly
-linked list approach), so we form the per-peer queue around the unused
-skb->prev pointer (which sort of makes sense because the links are
-pointing backwards). Use of skb_queue_* is not possible here, because
-that is based on doubly linked lists and spinlocks. Multiple cores can
-write into the queue at any given time, because its writes occur in the
-start_xmit path or in the udp_recv path. But reads happen in a single
-workqueue item per-peer, amounting to a multi-producer, single-consumer
-paradigm.
-
-The MPSC queue is implemented locklessly and never blocks. However, it
-is not linearizable (though it is serializable), with a very tight and
-unlikely race on writes, which, when hit (some tiny fraction of the
-0.15% of partial adds on a fully loaded 16-core x86_64 system), causes
-the queue reader to terminate early. However, because every packet sent
-queues up the same workqueue item after it is fully added, the worker
-resumes again, and stopping early isn't actually a problem, since at
-that point the packet wouldn't have yet been added to the encryption
-queue. These properties allow us to avoid disabling interrupts or
-spinning. The design is based on Dmitry Vyukov's algorithm [1].
-
-Performance-wise, ordinarily list-based queues aren't preferable to
-ringbuffers, because of cache misses when following pointers around.
-However, we *already* have to follow the adjacent pointers when working
-through fragments, so there shouldn't actually be any change there. A
-potential downside is that dequeueing is a bit more complicated, but the
-ptr_ring structure used prior had a spinlock when dequeueing, so all and
-all the difference appears to be a wash.
-
-Actually, from profiling, the biggest performance hit, by far, of this
-commit winds up being atomic_add_unless(count, 1, max) and atomic_
-dec(count), which account for the majority of CPU time, according to
-perf. In that sense, the previous ring buffer was superior in that it
-could check if it was full by head==tail, which the list-based approach
-cannot do.
-
-But all and all, this enables us to get massive memory savings, allowing
-WireGuard to scale for real world deployments, without taking much of a
-performance hit.
-
-[1] http://www.1024cores.net/home/lock-free-algorithms/queues/intrusive-mpsc-node-based-queue
-
-Reviewed-by: Dmitry Vyukov <dvyukov@google.com>
-Reviewed-by: Toke Høiland-Jørgensen <toke@redhat.com>
-Fixes: e7096c131e51 ("net: WireGuard secure network tunnel")
-Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
-Signed-off-by: Jakub Kicinski <kuba@kernel.org>
-Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
----
- drivers/net/wireguard/device.c   | 12 ++---
- drivers/net/wireguard/device.h   | 15 +++---
- drivers/net/wireguard/peer.c     | 28 ++++-------
- drivers/net/wireguard/peer.h     |  4 +-
- drivers/net/wireguard/queueing.c | 86 +++++++++++++++++++++++++-------
- drivers/net/wireguard/queueing.h | 45 ++++++++++++-----
- drivers/net/wireguard/receive.c  | 16 +++---
- drivers/net/wireguard/send.c     | 31 ++++--------
- 8 files changed, 144 insertions(+), 93 deletions(-)
-
---- a/drivers/net/wireguard/device.c
-+++ b/drivers/net/wireguard/device.c
-@@ -235,8 +235,8 @@ static void wg_destruct(struct net_devic
-       destroy_workqueue(wg->handshake_receive_wq);
-       destroy_workqueue(wg->handshake_send_wq);
-       destroy_workqueue(wg->packet_crypt_wq);
--      wg_packet_queue_free(&wg->decrypt_queue, true);
--      wg_packet_queue_free(&wg->encrypt_queue, true);
-+      wg_packet_queue_free(&wg->decrypt_queue);
-+      wg_packet_queue_free(&wg->encrypt_queue);
-       rcu_barrier(); /* Wait for all the peers to be actually freed. */
-       wg_ratelimiter_uninit();
-       memzero_explicit(&wg->static_identity, sizeof(wg->static_identity));
-@@ -338,12 +338,12 @@ static int wg_newlink(struct net *src_ne
-               goto err_destroy_handshake_send;
-       ret = wg_packet_queue_init(&wg->encrypt_queue, wg_packet_encrypt_worker,
--                                 true, MAX_QUEUED_PACKETS);
-+                                 MAX_QUEUED_PACKETS);
-       if (ret < 0)
-               goto err_destroy_packet_crypt;
-       ret = wg_packet_queue_init(&wg->decrypt_queue, wg_packet_decrypt_worker,
--                                 true, MAX_QUEUED_PACKETS);
-+                                 MAX_QUEUED_PACKETS);
-       if (ret < 0)
-               goto err_free_encrypt_queue;
-@@ -368,9 +368,9 @@ static int wg_newlink(struct net *src_ne
- err_uninit_ratelimiter:
-       wg_ratelimiter_uninit();
- err_free_decrypt_queue:
--      wg_packet_queue_free(&wg->decrypt_queue, true);
-+      wg_packet_queue_free(&wg->decrypt_queue);
- err_free_encrypt_queue:
--      wg_packet_queue_free(&wg->encrypt_queue, true);
-+      wg_packet_queue_free(&wg->encrypt_queue);
- err_destroy_packet_crypt:
-       destroy_workqueue(wg->packet_crypt_wq);
- err_destroy_handshake_send:
---- a/drivers/net/wireguard/device.h
-+++ b/drivers/net/wireguard/device.h
-@@ -27,13 +27,14 @@ struct multicore_worker {
- struct crypt_queue {
-       struct ptr_ring ring;
--      union {
--              struct {
--                      struct multicore_worker __percpu *worker;
--                      int last_cpu;
--              };
--              struct work_struct work;
--      };
-+      struct multicore_worker __percpu *worker;
-+      int last_cpu;
-+};
-+
-+struct prev_queue {
-+      struct sk_buff *head, *tail, *peeked;
-+      struct { struct sk_buff *next, *prev; } empty; // Match first 2 members of struct sk_buff.
-+      atomic_t count;
- };
- struct wg_device {
---- a/drivers/net/wireguard/peer.c
-+++ b/drivers/net/wireguard/peer.c
-@@ -32,27 +32,22 @@ struct wg_peer *wg_peer_create(struct wg
-       peer = kzalloc(sizeof(*peer), GFP_KERNEL);
-       if (unlikely(!peer))
-               return ERR_PTR(ret);
--      peer->device = wg;
-+      if (dst_cache_init(&peer->endpoint_cache, GFP_KERNEL))
-+              goto err;
-+      peer->device = wg;
-       wg_noise_handshake_init(&peer->handshake, &wg->static_identity,
-                               public_key, preshared_key, peer);
--      if (dst_cache_init(&peer->endpoint_cache, GFP_KERNEL))
--              goto err_1;
--      if (wg_packet_queue_init(&peer->tx_queue, wg_packet_tx_worker, false,
--                               MAX_QUEUED_PACKETS))
--              goto err_2;
--      if (wg_packet_queue_init(&peer->rx_queue, NULL, false,
--                               MAX_QUEUED_PACKETS))
--              goto err_3;
--
-       peer->internal_id = atomic64_inc_return(&peer_counter);
-       peer->serial_work_cpu = nr_cpumask_bits;
-       wg_cookie_init(&peer->latest_cookie);
-       wg_timers_init(peer);
-       wg_cookie_checker_precompute_peer_keys(peer);
-       spin_lock_init(&peer->keypairs.keypair_update_lock);
--      INIT_WORK(&peer->transmit_handshake_work,
--                wg_packet_handshake_send_worker);
-+      INIT_WORK(&peer->transmit_handshake_work, wg_packet_handshake_send_worker);
-+      INIT_WORK(&peer->transmit_packet_work, wg_packet_tx_worker);
-+      wg_prev_queue_init(&peer->tx_queue);
-+      wg_prev_queue_init(&peer->rx_queue);
-       rwlock_init(&peer->endpoint_lock);
-       kref_init(&peer->refcount);
-       skb_queue_head_init(&peer->staged_packet_queue);
-@@ -68,11 +63,7 @@ struct wg_peer *wg_peer_create(struct wg
-       pr_debug("%s: Peer %llu created\n", wg->dev->name, peer->internal_id);
-       return peer;
--err_3:
--      wg_packet_queue_free(&peer->tx_queue, false);
--err_2:
--      dst_cache_destroy(&peer->endpoint_cache);
--err_1:
-+err:
-       kfree(peer);
-       return ERR_PTR(ret);
- }
-@@ -197,8 +188,7 @@ static void rcu_release(struct rcu_head
-       struct wg_peer *peer = container_of(rcu, struct wg_peer, rcu);
-       dst_cache_destroy(&peer->endpoint_cache);
--      wg_packet_queue_free(&peer->rx_queue, false);
--      wg_packet_queue_free(&peer->tx_queue, false);
-+      WARN_ON(wg_prev_queue_peek(&peer->tx_queue) || wg_prev_queue_peek(&peer->rx_queue));
-       /* The final zeroing takes care of clearing any remaining handshake key
-        * material and other potentially sensitive information.
---- a/drivers/net/wireguard/peer.h
-+++ b/drivers/net/wireguard/peer.h
-@@ -36,7 +36,7 @@ struct endpoint {
- struct wg_peer {
-       struct wg_device *device;
--      struct crypt_queue tx_queue, rx_queue;
-+      struct prev_queue tx_queue, rx_queue;
-       struct sk_buff_head staged_packet_queue;
-       int serial_work_cpu;
-       bool is_dead;
-@@ -46,7 +46,7 @@ struct wg_peer {
-       rwlock_t endpoint_lock;
-       struct noise_handshake handshake;
-       atomic64_t last_sent_handshake;
--      struct work_struct transmit_handshake_work, clear_peer_work;
-+      struct work_struct transmit_handshake_work, clear_peer_work, transmit_packet_work;
-       struct cookie latest_cookie;
-       struct hlist_node pubkey_hash;
-       u64 rx_bytes, tx_bytes;
---- a/drivers/net/wireguard/queueing.c
-+++ b/drivers/net/wireguard/queueing.c
-@@ -9,8 +9,7 @@ struct multicore_worker __percpu *
- wg_packet_percpu_multicore_worker_alloc(work_func_t function, void *ptr)
- {
-       int cpu;
--      struct multicore_worker __percpu *worker =
--              alloc_percpu(struct multicore_worker);
-+      struct multicore_worker __percpu *worker = alloc_percpu(struct multicore_worker);
-       if (!worker)
-               return NULL;
-@@ -23,7 +22,7 @@ wg_packet_percpu_multicore_worker_alloc(
- }
- int wg_packet_queue_init(struct crypt_queue *queue, work_func_t function,
--                       bool multicore, unsigned int len)
-+                       unsigned int len)
- {
-       int ret;
-@@ -31,25 +30,78 @@ int wg_packet_queue_init(struct crypt_qu
-       ret = ptr_ring_init(&queue->ring, len, GFP_KERNEL);
-       if (ret)
-               return ret;
--      if (function) {
--              if (multicore) {
--                      queue->worker = wg_packet_percpu_multicore_worker_alloc(
--                              function, queue);
--                      if (!queue->worker) {
--                              ptr_ring_cleanup(&queue->ring, NULL);
--                              return -ENOMEM;
--                      }
--              } else {
--                      INIT_WORK(&queue->work, function);
--              }
-+      queue->worker = wg_packet_percpu_multicore_worker_alloc(function, queue);
-+      if (!queue->worker) {
-+              ptr_ring_cleanup(&queue->ring, NULL);
-+              return -ENOMEM;
-       }
-       return 0;
- }
--void wg_packet_queue_free(struct crypt_queue *queue, bool multicore)
-+void wg_packet_queue_free(struct crypt_queue *queue)
- {
--      if (multicore)
--              free_percpu(queue->worker);
-+      free_percpu(queue->worker);
-       WARN_ON(!__ptr_ring_empty(&queue->ring));
-       ptr_ring_cleanup(&queue->ring, NULL);
- }
-+
-+#define NEXT(skb) ((skb)->prev)
-+#define STUB(queue) ((struct sk_buff *)&queue->empty)
-+
-+void wg_prev_queue_init(struct prev_queue *queue)
-+{
-+      NEXT(STUB(queue)) = NULL;
-+      queue->head = queue->tail = STUB(queue);
-+      queue->peeked = NULL;
-+      atomic_set(&queue->count, 0);
-+      BUILD_BUG_ON(
-+              offsetof(struct sk_buff, next) != offsetof(struct prev_queue, empty.next) -
-+                                                      offsetof(struct prev_queue, empty) ||
-+              offsetof(struct sk_buff, prev) != offsetof(struct prev_queue, empty.prev) -
-+                                                       offsetof(struct prev_queue, empty));
-+}
-+
-+static void __wg_prev_queue_enqueue(struct prev_queue *queue, struct sk_buff *skb)
-+{
-+      WRITE_ONCE(NEXT(skb), NULL);
-+      WRITE_ONCE(NEXT(xchg_release(&queue->head, skb)), skb);
-+}
-+
-+bool wg_prev_queue_enqueue(struct prev_queue *queue, struct sk_buff *skb)
-+{
-+      if (!atomic_add_unless(&queue->count, 1, MAX_QUEUED_PACKETS))
-+              return false;
-+      __wg_prev_queue_enqueue(queue, skb);
-+      return true;
-+}
-+
-+struct sk_buff *wg_prev_queue_dequeue(struct prev_queue *queue)
-+{
-+      struct sk_buff *tail = queue->tail, *next = smp_load_acquire(&NEXT(tail));
-+
-+      if (tail == STUB(queue)) {
-+              if (!next)
-+                      return NULL;
-+              queue->tail = next;
-+              tail = next;
-+              next = smp_load_acquire(&NEXT(next));
-+      }
-+      if (next) {
-+              queue->tail = next;
-+              atomic_dec(&queue->count);
-+              return tail;
-+      }
-+      if (tail != READ_ONCE(queue->head))
-+              return NULL;
-+      __wg_prev_queue_enqueue(queue, STUB(queue));
-+      next = smp_load_acquire(&NEXT(tail));
-+      if (next) {
-+              queue->tail = next;
-+              atomic_dec(&queue->count);
-+              return tail;
-+      }
-+      return NULL;
-+}
-+
-+#undef NEXT
-+#undef STUB
---- a/drivers/net/wireguard/queueing.h
-+++ b/drivers/net/wireguard/queueing.h
-@@ -17,12 +17,13 @@ struct wg_device;
- struct wg_peer;
- struct multicore_worker;
- struct crypt_queue;
-+struct prev_queue;
- struct sk_buff;
- /* queueing.c APIs: */
- int wg_packet_queue_init(struct crypt_queue *queue, work_func_t function,
--                       bool multicore, unsigned int len);
--void wg_packet_queue_free(struct crypt_queue *queue, bool multicore);
-+                       unsigned int len);
-+void wg_packet_queue_free(struct crypt_queue *queue);
- struct multicore_worker __percpu *
- wg_packet_percpu_multicore_worker_alloc(work_func_t function, void *ptr);
-@@ -135,8 +136,31 @@ static inline int wg_cpumask_next_online
-       return cpu;
- }
-+void wg_prev_queue_init(struct prev_queue *queue);
-+
-+/* Multi producer */
-+bool wg_prev_queue_enqueue(struct prev_queue *queue, struct sk_buff *skb);
-+
-+/* Single consumer */
-+struct sk_buff *wg_prev_queue_dequeue(struct prev_queue *queue);
-+
-+/* Single consumer */
-+static inline struct sk_buff *wg_prev_queue_peek(struct prev_queue *queue)
-+{
-+      if (queue->peeked)
-+              return queue->peeked;
-+      queue->peeked = wg_prev_queue_dequeue(queue);
-+      return queue->peeked;
-+}
-+
-+/* Single consumer */
-+static inline void wg_prev_queue_drop_peeked(struct prev_queue *queue)
-+{
-+      queue->peeked = NULL;
-+}
-+
- static inline int wg_queue_enqueue_per_device_and_peer(
--      struct crypt_queue *device_queue, struct crypt_queue *peer_queue,
-+      struct crypt_queue *device_queue, struct prev_queue *peer_queue,
-       struct sk_buff *skb, struct workqueue_struct *wq, int *next_cpu)
- {
-       int cpu;
-@@ -145,8 +169,9 @@ static inline int wg_queue_enqueue_per_d
-       /* We first queue this up for the peer ingestion, but the consumer
-        * will wait for the state to change to CRYPTED or DEAD before.
-        */
--      if (unlikely(ptr_ring_produce_bh(&peer_queue->ring, skb)))
-+      if (unlikely(!wg_prev_queue_enqueue(peer_queue, skb)))
-               return -ENOSPC;
-+
-       /* Then we queue it up in the device queue, which consumes the
-        * packet as soon as it can.
-        */
-@@ -157,9 +182,7 @@ static inline int wg_queue_enqueue_per_d
-       return 0;
- }
--static inline void wg_queue_enqueue_per_peer(struct crypt_queue *queue,
--                                           struct sk_buff *skb,
--                                           enum packet_state state)
-+static inline void wg_queue_enqueue_per_peer_tx(struct sk_buff *skb, enum packet_state state)
- {
-       /* We take a reference, because as soon as we call atomic_set, the
-        * peer can be freed from below us.
-@@ -167,14 +190,12 @@ static inline void wg_queue_enqueue_per_
-       struct wg_peer *peer = wg_peer_get(PACKET_PEER(skb));
-       atomic_set_release(&PACKET_CB(skb)->state, state);
--      queue_work_on(wg_cpumask_choose_online(&peer->serial_work_cpu,
--                                             peer->internal_id),
--                    peer->device->packet_crypt_wq, &queue->work);
-+      queue_work_on(wg_cpumask_choose_online(&peer->serial_work_cpu, peer->internal_id),
-+                    peer->device->packet_crypt_wq, &peer->transmit_packet_work);
-       wg_peer_put(peer);
- }
--static inline void wg_queue_enqueue_per_peer_napi(struct sk_buff *skb,
--                                                enum packet_state state)
-+static inline void wg_queue_enqueue_per_peer_rx(struct sk_buff *skb, enum packet_state state)
- {
-       /* We take a reference, because as soon as we call atomic_set, the
-        * peer can be freed from below us.
---- a/drivers/net/wireguard/receive.c
-+++ b/drivers/net/wireguard/receive.c
-@@ -444,7 +444,6 @@ packet_processed:
- int wg_packet_rx_poll(struct napi_struct *napi, int budget)
- {
-       struct wg_peer *peer = container_of(napi, struct wg_peer, napi);
--      struct crypt_queue *queue = &peer->rx_queue;
-       struct noise_keypair *keypair;
-       struct endpoint endpoint;
-       enum packet_state state;
-@@ -455,11 +454,10 @@ int wg_packet_rx_poll(struct napi_struct
-       if (unlikely(budget <= 0))
-               return 0;
--      while ((skb = __ptr_ring_peek(&queue->ring)) != NULL &&
-+      while ((skb = wg_prev_queue_peek(&peer->rx_queue)) != NULL &&
-              (state = atomic_read_acquire(&PACKET_CB(skb)->state)) !=
-                      PACKET_STATE_UNCRYPTED) {
--              __ptr_ring_discard_one(&queue->ring);
--              peer = PACKET_PEER(skb);
-+              wg_prev_queue_drop_peeked(&peer->rx_queue);
-               keypair = PACKET_CB(skb)->keypair;
-               free = true;
-@@ -508,7 +506,7 @@ void wg_packet_decrypt_worker(struct wor
-               enum packet_state state =
-                       likely(decrypt_packet(skb, PACKET_CB(skb)->keypair)) ?
-                               PACKET_STATE_CRYPTED : PACKET_STATE_DEAD;
--              wg_queue_enqueue_per_peer_napi(skb, state);
-+              wg_queue_enqueue_per_peer_rx(skb, state);
-               if (need_resched())
-                       cond_resched();
-       }
-@@ -531,12 +529,10 @@ static void wg_packet_consume_data(struc
-       if (unlikely(READ_ONCE(peer->is_dead)))
-               goto err;
--      ret = wg_queue_enqueue_per_device_and_peer(&wg->decrypt_queue,
--                                                 &peer->rx_queue, skb,
--                                                 wg->packet_crypt_wq,
--                                                 &wg->decrypt_queue.last_cpu);
-+      ret = wg_queue_enqueue_per_device_and_peer(&wg->decrypt_queue, &peer->rx_queue, skb,
-+                                                 wg->packet_crypt_wq, &wg->decrypt_queue.last_cpu);
-       if (unlikely(ret == -EPIPE))
--              wg_queue_enqueue_per_peer_napi(skb, PACKET_STATE_DEAD);
-+              wg_queue_enqueue_per_peer_rx(skb, PACKET_STATE_DEAD);
-       if (likely(!ret || ret == -EPIPE)) {
-               rcu_read_unlock_bh();
-               return;
---- a/drivers/net/wireguard/send.c
-+++ b/drivers/net/wireguard/send.c
-@@ -239,8 +239,7 @@ void wg_packet_send_keepalive(struct wg_
-       wg_packet_send_staged_packets(peer);
- }
--static void wg_packet_create_data_done(struct sk_buff *first,
--                                     struct wg_peer *peer)
-+static void wg_packet_create_data_done(struct wg_peer *peer, struct sk_buff *first)
- {
-       struct sk_buff *skb, *next;
-       bool is_keepalive, data_sent = false;
-@@ -262,22 +261,19 @@ static void wg_packet_create_data_done(s
- void wg_packet_tx_worker(struct work_struct *work)
- {
--      struct crypt_queue *queue = container_of(work, struct crypt_queue,
--                                               work);
-+      struct wg_peer *peer = container_of(work, struct wg_peer, transmit_packet_work);
-       struct noise_keypair *keypair;
-       enum packet_state state;
-       struct sk_buff *first;
--      struct wg_peer *peer;
--      while ((first = __ptr_ring_peek(&queue->ring)) != NULL &&
-+      while ((first = wg_prev_queue_peek(&peer->tx_queue)) != NULL &&
-              (state = atomic_read_acquire(&PACKET_CB(first)->state)) !=
-                      PACKET_STATE_UNCRYPTED) {
--              __ptr_ring_discard_one(&queue->ring);
--              peer = PACKET_PEER(first);
-+              wg_prev_queue_drop_peeked(&peer->tx_queue);
-               keypair = PACKET_CB(first)->keypair;
-               if (likely(state == PACKET_STATE_CRYPTED))
--                      wg_packet_create_data_done(first, peer);
-+                      wg_packet_create_data_done(peer, first);
-               else
-                       kfree_skb_list(first);
-@@ -306,16 +302,14 @@ void wg_packet_encrypt_worker(struct wor
-                               break;
-                       }
-               }
--              wg_queue_enqueue_per_peer(&PACKET_PEER(first)->tx_queue, first,
--                                        state);
-+              wg_queue_enqueue_per_peer_tx(first, state);
-               if (need_resched())
-                       cond_resched();
-       }
- }
--static void wg_packet_create_data(struct sk_buff *first)
-+static void wg_packet_create_data(struct wg_peer *peer, struct sk_buff *first)
- {
--      struct wg_peer *peer = PACKET_PEER(first);
-       struct wg_device *wg = peer->device;
-       int ret = -EINVAL;
-@@ -323,13 +317,10 @@ static void wg_packet_create_data(struct
-       if (unlikely(READ_ONCE(peer->is_dead)))
-               goto err;
--      ret = wg_queue_enqueue_per_device_and_peer(&wg->encrypt_queue,
--                                                 &peer->tx_queue, first,
--                                                 wg->packet_crypt_wq,
--                                                 &wg->encrypt_queue.last_cpu);
-+      ret = wg_queue_enqueue_per_device_and_peer(&wg->encrypt_queue, &peer->tx_queue, first,
-+                                                 wg->packet_crypt_wq, &wg->encrypt_queue.last_cpu);
-       if (unlikely(ret == -EPIPE))
--              wg_queue_enqueue_per_peer(&peer->tx_queue, first,
--                                        PACKET_STATE_DEAD);
-+              wg_queue_enqueue_per_peer_tx(first, PACKET_STATE_DEAD);
- err:
-       rcu_read_unlock_bh();
-       if (likely(!ret || ret == -EPIPE))
-@@ -393,7 +384,7 @@ void wg_packet_send_staged_packets(struc
-       packets.prev->next = NULL;
-       wg_peer_get(keypair->entry.peer);
-       PACKET_CB(packets.next)->keypair = keypair;
--      wg_packet_create_data(packets.next);
-+      wg_packet_create_data(peer, packets.next);
-       return;
- out_invalid:
diff --git a/target/linux/generic/backport-5.10/083-wireguard-kconfig-use-arm-chacha-even-with-no-neon.patch b/target/linux/generic/backport-5.10/083-wireguard-kconfig-use-arm-chacha-even-with-no-neon.patch
deleted file mode 100644 (file)
index 8a8c04e..0000000
+++ /dev/null
@@ -1,30 +0,0 @@
-From 514091206bc055a159348ae8575276dc925aea24 Mon Sep 17 00:00:00 2001
-From: "Jason A. Donenfeld" <Jason@zx2c4.com>
-Date: Mon, 22 Feb 2021 17:25:49 +0100
-Subject: [PATCH] wireguard: kconfig: use arm chacha even with no neon
-
-commit bce2473927af8de12ad131a743f55d69d358c0b9 upstream.
-
-The condition here was incorrect: a non-neon fallback implementation is
-available on arm32 when NEON is not supported.
-
-Reported-by: Ilya Lipnitskiy <ilya.lipnitskiy@gmail.com>
-Fixes: e7096c131e51 ("net: WireGuard secure network tunnel")
-Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
-Signed-off-by: Jakub Kicinski <kuba@kernel.org>
-Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
----
- drivers/net/Kconfig | 2 +-
- 1 file changed, 1 insertion(+), 1 deletion(-)
-
---- a/drivers/net/Kconfig
-+++ b/drivers/net/Kconfig
-@@ -87,7 +87,7 @@ config WIREGUARD
-       select CRYPTO_CURVE25519_X86 if X86 && 64BIT
-       select ARM_CRYPTO if ARM
-       select ARM64_CRYPTO if ARM64
--      select CRYPTO_CHACHA20_NEON if (ARM || ARM64) && KERNEL_MODE_NEON
-+      select CRYPTO_CHACHA20_NEON if ARM || (ARM64 && KERNEL_MODE_NEON)
-       select CRYPTO_POLY1305_NEON if ARM64 && KERNEL_MODE_NEON
-       select CRYPTO_POLY1305_ARM if ARM
-       select CRYPTO_CURVE25519_NEON if ARM && KERNEL_MODE_NEON
diff --git a/target/linux/generic/backport-5.10/830-v5.12-0002-usb-serial-option-update-interface-mapping-for-ZTE-P685M.patch b/target/linux/generic/backport-5.10/830-v5.12-0002-usb-serial-option-update-interface-mapping-for-ZTE-P685M.patch
deleted file mode 100644 (file)
index 87eed64..0000000
+++ /dev/null
@@ -1,73 +0,0 @@
-From 6420a569504e212d618d4a4736e2c59ed80a8478 Mon Sep 17 00:00:00 2001
-From: Lech Perczak <lech.perczak@gmail.com>
-Date: Sun, 7 Feb 2021 01:54:43 +0100
-Subject: USB: serial: option: update interface mapping for ZTE P685M
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-This patch prepares for qmi_wwan driver support for the device.
-Previously "option" driver mapped itself to interfaces 0 and 3 (matching
-ff/ff/ff), while interface 3 is in fact a QMI port.
-Interfaces 1 and 2 (matching ff/00/00) expose AT commands,
-and weren't supported previously at all.
-Without this patch, a possible conflict would exist if device ID was
-added to qmi_wwan driver for interface 3.
-
-Update and simplify device ID to match interfaces 0-2 directly,
-to expose QCDM (0), PCUI (1), and modem (2) ports and avoid conflict
-with QMI (3), and ADB (4).
-
-The modem is used inside ZTE MF283+ router and carriers identify it as
-such.
-Interface mapping is:
-0: QCDM, 1: AT (PCUI), 2: AT (Modem), 3: QMI, 4: ADB
-
-T:  Bus=02 Lev=02 Prnt=02 Port=05 Cnt=01 Dev#=  3 Spd=480  MxCh= 0
-D:  Ver= 2.01 Cls=00(>ifc ) Sub=00 Prot=00 MxPS=64 #Cfgs=  1
-P:  Vendor=19d2 ProdID=1275 Rev=f0.00
-S:  Manufacturer=ZTE,Incorporated
-S:  Product=ZTE Technologies MSM
-S:  SerialNumber=P685M510ZTED0000CP&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&0
-C:* #Ifs= 5 Cfg#= 1 Atr=a0 MxPwr=500mA
-I:* If#= 0 Alt= 0 #EPs= 2 Cls=ff(vend.) Sub=ff Prot=ff Driver=option
-E:  Ad=81(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
-E:  Ad=01(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
-I:* If#= 1 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=00 Prot=00 Driver=option
-E:  Ad=83(I) Atr=03(Int.) MxPS=  10 Ivl=32ms
-E:  Ad=82(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
-E:  Ad=02(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
-I:* If#= 2 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=00 Prot=00 Driver=option
-E:  Ad=85(I) Atr=03(Int.) MxPS=  10 Ivl=32ms
-E:  Ad=84(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
-E:  Ad=03(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
-I:* If#= 3 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=ff Prot=ff Driver=qmi_wwan
-E:  Ad=87(I) Atr=03(Int.) MxPS=   8 Ivl=32ms
-E:  Ad=86(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
-E:  Ad=04(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
-I:* If#= 4 Alt= 0 #EPs= 2 Cls=ff(vend.) Sub=42 Prot=01 Driver=(none)
-E:  Ad=88(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
-E:  Ad=05(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
-
-Cc: Johan Hovold <johan@kernel.org>
-Cc: Bjørn Mork <bjorn@mork.no>
-Signed-off-by: Lech Perczak <lech.perczak@gmail.com>
-Link: https://lore.kernel.org/r/20210207005443.12936-1-lech.perczak@gmail.com
-Cc: stable@vger.kernel.org
-Signed-off-by: Johan Hovold <johan@kernel.org>
----
- drivers/usb/serial/option.c | 3 ++-
- 1 file changed, 2 insertions(+), 1 deletion(-)
-
---- a/drivers/usb/serial/option.c
-+++ b/drivers/usb/serial/option.c
-@@ -1569,7 +1569,8 @@ static const struct usb_device_id option
-       { USB_DEVICE_AND_INTERFACE_INFO(ZTE_VENDOR_ID, 0x1272, 0xff, 0xff, 0xff) },
-       { USB_DEVICE_AND_INTERFACE_INFO(ZTE_VENDOR_ID, 0x1273, 0xff, 0xff, 0xff) },
-       { USB_DEVICE_AND_INTERFACE_INFO(ZTE_VENDOR_ID, 0x1274, 0xff, 0xff, 0xff) },
--      { USB_DEVICE_AND_INTERFACE_INFO(ZTE_VENDOR_ID, 0x1275, 0xff, 0xff, 0xff) },
-+      { USB_DEVICE(ZTE_VENDOR_ID, 0x1275),    /* ZTE P685M */
-+        .driver_info = RSVD(3) | RSVD(4) },
-       { USB_DEVICE_AND_INTERFACE_INFO(ZTE_VENDOR_ID, 0x1276, 0xff, 0xff, 0xff) },
-       { USB_DEVICE_AND_INTERFACE_INFO(ZTE_VENDOR_ID, 0x1277, 0xff, 0xff, 0xff) },
-       { USB_DEVICE_AND_INTERFACE_INFO(ZTE_VENDOR_ID, 0x1278, 0xff, 0xff, 0xff) },
index e5b8ef9e89aa2cb76a9043ad4eebe58e2cc8c660..040d62a51d1e72f4f80c75a25da04055e593f71b 100644 (file)
@@ -2779,6 +2779,7 @@ CONFIG_KALLSYMS_BASE_RELATIVE=y
 # CONFIG_KARMA_PARTITION is not set
 # CONFIG_KASAN is not set
 CONFIG_KASAN_STACK=1
+# CONFIG_KCMP is not set
 # CONFIG_KCOV is not set
 # CONFIG_KERNEL_BZIP2 is not set
 # CONFIG_KERNEL_CAT is not set
index 79fc1057aba5eeac6c1f8eeac23dc3516d223ec2..2a162600fc0270ac1e8d57f28ec8254f8b682bc6 100644 (file)
@@ -88,7 +88,7 @@ Signed-off-by: Felix Fietkau <nbd@nbd.name>
  
 --- a/init/Kconfig
 +++ b/init/Kconfig
-@@ -2327,6 +2327,13 @@ config UNUSED_KSYMS_WHITELIST
+@@ -2338,6 +2338,13 @@ config UNUSED_KSYMS_WHITELIST
          one per line. The path can be absolute, or relative to the kernel
          source tree.
  
@@ -104,7 +104,7 @@ Signed-off-by: Felix Fietkau <nbd@nbd.name>
  config MODULES_TREE_LOOKUP
 --- a/kernel/module.c
 +++ b/kernel/module.c
-@@ -3144,9 +3144,11 @@ static int setup_load_info(struct load_i
+@@ -3161,9 +3161,11 @@ static int setup_load_info(struct load_i
  
  static int check_modinfo(struct module *mod, struct load_info *info, int flags)
  {
@@ -117,7 +117,7 @@ Signed-off-by: Felix Fietkau <nbd@nbd.name>
        if (flags & MODULE_INIT_IGNORE_VERMAGIC)
                modmagic = NULL;
  
-@@ -3167,6 +3169,7 @@ static int check_modinfo(struct module *
+@@ -3184,6 +3186,7 @@ static int check_modinfo(struct module *
                                mod->name);
                add_taint_module(mod, TAINT_OOT_MODULE, LOCKDEP_STILL_OK);
        }
index 9dd57ba9334380f63019b179e176bb7e5e9f041d..655cc61560863da21b35903a00e90333b408b478 100644 (file)
@@ -56,7 +56,7 @@ Signed-off-by: Felix Fietkau <nbd@nbd.name>
        }                                                               \
                                                                        \
        /* __*init sections */                                          \
-@@ -1008,6 +1018,8 @@
+@@ -1014,6 +1024,8 @@
  
  #define COMMON_DISCARDS                                                       \
        SANITIZER_DISCARDS                                              \
index 348a5afa3d1e29699c54923338ec92be55dc7c60..9f2014c8ec482a704e817ae845f0cd738daec2e0 100644 (file)
@@ -13,7 +13,7 @@ Signed-off-by: Felix Fietkau <nbd@nbd.name>
 
 --- a/init/Kconfig
 +++ b/init/Kconfig
-@@ -1384,6 +1384,17 @@ config SYSCTL_ARCH_UNALIGN_ALLOW
+@@ -1385,6 +1385,17 @@ config SYSCTL_ARCH_UNALIGN_ALLOW
          the unaligned access emulation.
          see arch/parisc/kernel/unaligned.c for reference
  
index 54c0e01023855867724937e88f238c1f47638213..bbea9473824b6a78c93326fd511788deaa657351 100644 (file)
@@ -11,7 +11,7 @@ Signed-off-by: Felix Fietkau <nbd@nbd.name>
 
 --- a/arch/mips/Makefile
 +++ b/arch/mips/Makefile
-@@ -155,7 +155,7 @@ cflags-$(CONFIG_CPU_VR41XX)        += -march=r4
+@@ -174,7 +174,7 @@ cflags-$(CONFIG_CPU_VR41XX)        += -march=r4
  cflags-$(CONFIG_CPU_R4X00)    += -march=r4600 -Wa,--trap
  cflags-$(CONFIG_CPU_TX49XX)   += -march=r4600 -Wa,--trap
  cflags-$(CONFIG_CPU_MIPS32_R1)        += -march=mips32 -Wa,--trap
index bbdab181029ea69eedbffdb3db5b08bf5a5941f9..d000094e1dc40346e838f101a56c5f380ea1d5d6 100644 (file)
@@ -39,7 +39,7 @@ Signed-off-by: Felix Fietkau <nbd@nbd.name>
  endif # MTD_SPI_NOR
 --- a/drivers/mtd/spi-nor/core.c
 +++ b/drivers/mtd/spi-nor/core.c
-@@ -2784,6 +2784,21 @@ static void spi_nor_info_init_params(str
+@@ -2786,6 +2786,21 @@ static void spi_nor_info_init_params(str
         */
        erase_mask = 0;
        i = 0;
@@ -61,7 +61,7 @@ Signed-off-by: Felix Fietkau <nbd@nbd.name>
        if (info->flags & SECT_4K_PMC) {
                erase_mask |= BIT(i);
                spi_nor_set_erase_type(&map->erase_type[i], 4096u,
-@@ -2795,6 +2810,7 @@ static void spi_nor_info_init_params(str
+@@ -2797,6 +2812,7 @@ static void spi_nor_info_init_params(str
                                       SPINOR_OP_BE_4K);
                i++;
        }
index 3e7cd036794bc2c60b85b0c5d356da1fb9976533..9ff6ffae5894c9b35f8a2edbf076bc781cff811d 100644 (file)
@@ -59,7 +59,7 @@ Signed-off-by: Felix Fietkau <nbd@nbd.name>
 +};
 --- a/drivers/mtd/spi-nor/core.c
 +++ b/drivers/mtd/spi-nor/core.c
-@@ -2024,6 +2024,7 @@ static const struct spi_nor_manufacturer
+@@ -2026,6 +2026,7 @@ static const struct spi_nor_manufacturer
        &spi_nor_winbond,
        &spi_nor_xilinx,
        &spi_nor_xmc,
index e8e737ffca955095fa9f1fa930ddc7cb4199742c..5cb85172fb85aec2abb695e5376d5101564ccf81 100644 (file)
@@ -19,7 +19,7 @@ Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
 
 --- a/drivers/mtd/spi-nor/core.c
 +++ b/drivers/mtd/spi-nor/core.c
-@@ -1445,6 +1445,23 @@ destroy_erase_cmd_list:
+@@ -1447,6 +1447,23 @@ destroy_erase_cmd_list:
        return ret;
  }
  
@@ -43,7 +43,7 @@ Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
  /*
   * Erase an address range on the nor chip.  The address range may extend
   * one or more erase sectors.  Return an error is there is a problem erasing.
-@@ -1472,6 +1489,10 @@ static int spi_nor_erase(struct mtd_info
+@@ -1474,6 +1491,10 @@ static int spi_nor_erase(struct mtd_info
        if (ret)
                return ret;
  
@@ -54,7 +54,7 @@ Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
        /* whole-chip erase? */
        if (len == mtd->size && !(nor->flags & SNOR_F_NO_OP_CHIP_ERASE)) {
                unsigned long timeout;
-@@ -1531,6 +1552,7 @@ static int spi_nor_erase(struct mtd_info
+@@ -1533,6 +1554,7 @@ static int spi_nor_erase(struct mtd_info
        ret = spi_nor_write_disable(nor);
  
  erase_err:
@@ -62,7 +62,7 @@ Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
        spi_nor_unlock_and_unprep(nor);
  
        return ret;
-@@ -1870,7 +1892,9 @@ static int spi_nor_lock(struct mtd_info
+@@ -1872,7 +1894,9 @@ static int spi_nor_lock(struct mtd_info
        if (ret)
                return ret;
  
@@ -72,7 +72,7 @@ Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
  
        spi_nor_unlock_and_unprep(nor);
        return ret;
-@@ -1885,7 +1909,9 @@ static int spi_nor_unlock(struct mtd_inf
+@@ -1887,7 +1911,9 @@ static int spi_nor_unlock(struct mtd_inf
        if (ret)
                return ret;
  
@@ -82,7 +82,7 @@ Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
  
        spi_nor_unlock_and_unprep(nor);
        return ret;
-@@ -1900,7 +1926,9 @@ static int spi_nor_is_locked(struct mtd_
+@@ -1902,7 +1928,9 @@ static int spi_nor_is_locked(struct mtd_
        if (ret)
                return ret;
  
@@ -92,7 +92,7 @@ Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
  
        spi_nor_unlock_and_unprep(nor);
        return ret;
-@@ -2093,6 +2121,10 @@ static int spi_nor_read(struct mtd_info
+@@ -2095,6 +2123,10 @@ static int spi_nor_read(struct mtd_info
        if (ret)
                return ret;
  
@@ -103,7 +103,7 @@ Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
        while (len) {
                loff_t addr = from;
  
-@@ -2116,6 +2148,7 @@ static int spi_nor_read(struct mtd_info
+@@ -2118,6 +2150,7 @@ static int spi_nor_read(struct mtd_info
        ret = 0;
  
  read_err:
@@ -111,7 +111,7 @@ Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
        spi_nor_unlock_and_unprep(nor);
        return ret;
  }
-@@ -2138,6 +2171,10 @@ static int spi_nor_write(struct mtd_info
+@@ -2140,6 +2173,10 @@ static int spi_nor_write(struct mtd_info
        if (ret)
                return ret;
  
@@ -122,7 +122,7 @@ Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
        for (i = 0; i < len; ) {
                ssize_t written;
                loff_t addr = to + i;
-@@ -2180,6 +2217,7 @@ static int spi_nor_write(struct mtd_info
+@@ -2182,6 +2219,7 @@ static int spi_nor_write(struct mtd_info
        }
  
  write_err:
@@ -130,7 +130,7 @@ Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
        spi_nor_unlock_and_unprep(nor);
        return ret;
  }
-@@ -2975,9 +3013,13 @@ static int spi_nor_init(struct spi_nor *
+@@ -2977,9 +3015,13 @@ static int spi_nor_init(struct spi_nor *
                 * reboots (e.g., crashes). Warn the user (or hopefully, system
                 * designer) that this is bad.
                 */
index 371cffc58b68dd4c87643b58dbcc3281cadb9b2b..ec77d59cf0e6733bf40710a767d3b9b5747624e5 100644 (file)
@@ -11,7 +11,7 @@ Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
 
 --- a/drivers/net/phy/phy_device.c
 +++ b/drivers/net/phy/phy_device.c
-@@ -1663,6 +1663,9 @@ void phy_detach(struct phy_device *phyde
+@@ -1644,6 +1644,9 @@ void phy_detach(struct phy_device *phyde
        struct module *ndev_owner = NULL;
        struct mii_bus *bus;
  
index 14fdef31fdcd5a2b96560ec6fc3d0973de9321ee..9008a88e2e241a6be9a53c97abbffa10c8ccccde 100644 (file)
@@ -13,7 +13,7 @@ Signed-off-by: Imre Kaloz <kaloz@openwrt.org>
 
 --- a/init/Kconfig
 +++ b/init/Kconfig
-@@ -1780,6 +1780,15 @@ config EMBEDDED
+@@ -1791,6 +1791,15 @@ config EMBEDDED
          an embedded system so certain expert options are available
          for configuration.
  
index 0a27e7854db231c4abf86010aef54419cbe085c5..e14dc484b08a74d9549c55734cfb018b016d9e12 100644 (file)
@@ -14,7 +14,7 @@ Signed-off-by: Petr Štetiar <ynezz@true.cz>
 
 --- a/arch/arm/boot/dts/imx6q-apalis-ixora-v1.1.dts
 +++ b/arch/arm/boot/dts/imx6q-apalis-ixora-v1.1.dts
-@@ -61,6 +61,10 @@
+@@ -25,6 +25,10 @@
                i2c2 = &i2c2;
                rtc0 = &rtc_i2c;
                rtc1 = &snvs_rtc;
@@ -25,7 +25,7 @@ Signed-off-by: Petr Štetiar <ynezz@true.cz>
        };
  
        chosen {
-@@ -128,22 +132,22 @@
+@@ -92,22 +96,22 @@
                pinctrl-names = "default";
                pinctrl-0 = <&pinctrl_leds_ixora>;
  
@@ -56,7 +56,7 @@ Signed-off-by: Petr Štetiar <ynezz@true.cz>
                };
 --- a/arch/arm/boot/dts/imx6q-apalis-ixora.dts
 +++ b/arch/arm/boot/dts/imx6q-apalis-ixora.dts
-@@ -60,6 +60,10 @@
+@@ -24,6 +24,10 @@
                i2c2 = &i2c2;
                rtc0 = &rtc_i2c;
                rtc1 = &snvs_rtc;
@@ -67,7 +67,7 @@ Signed-off-by: Petr Štetiar <ynezz@true.cz>
        };
  
        chosen {
-@@ -127,22 +131,22 @@
+@@ -91,22 +95,22 @@
                pinctrl-names = "default";
                pinctrl-0 = <&pinctrl_leds_ixora>;
  
index 2bb1a164681ee1ae5cdd394c1a8d51d764684ea0..8f07f0b1a5111efd2b21b19178c753eb4c10d7d2 100644 (file)
@@ -14,7 +14,7 @@ Signed-off-by: Petr Štetiar <ynezz@true.cz>
 
 --- a/arch/arm/boot/dts/imx6q-apalis-ixora-v1.1.dts
 +++ b/arch/arm/boot/dts/imx6q-apalis-ixora-v1.1.dts
-@@ -74,7 +74,7 @@
+@@ -38,7 +38,7 @@
        gpio-keys {
                compatible = "gpio-keys";
                pinctrl-names = "default";
@@ -23,7 +23,7 @@ Signed-off-by: Petr Štetiar <ynezz@true.cz>
  
                wakeup {
                        label = "Wake-Up";
-@@ -83,6 +83,13 @@
+@@ -47,6 +47,13 @@
                        debounce-interval = <10>;
                        wakeup-source;
                };
@@ -37,7 +37,7 @@ Signed-off-by: Petr Štetiar <ynezz@true.cz>
        };
  
        lcd_display: disp0 {
-@@ -298,4 +305,10 @@
+@@ -275,4 +282,10 @@
                        MX6QDL_PAD_NANDF_D2__GPIO2_IO02 0x1b0b0
                >;
        };
@@ -50,7 +50,7 @@ Signed-off-by: Petr Štetiar <ynezz@true.cz>
  };
 --- a/arch/arm/boot/dts/imx6q-apalis-ixora.dts
 +++ b/arch/arm/boot/dts/imx6q-apalis-ixora.dts
-@@ -73,7 +73,7 @@
+@@ -37,7 +37,7 @@
        gpio-keys {
                compatible = "gpio-keys";
                pinctrl-names = "default";
@@ -59,7 +59,7 @@ Signed-off-by: Petr Štetiar <ynezz@true.cz>
  
                wakeup {
                        label = "Wake-Up";
-@@ -82,6 +82,13 @@
+@@ -46,6 +46,13 @@
                        debounce-interval = <10>;
                        wakeup-source;
                };
@@ -73,7 +73,7 @@ Signed-off-by: Petr Štetiar <ynezz@true.cz>
        };
  
        lcd_display: disp0 {
-@@ -299,4 +306,10 @@
+@@ -276,4 +283,10 @@
                        MX6QDL_PAD_NANDF_D2__GPIO2_IO02 0x1b0b0
                >;
        };
index a680cf01492a88047539a5b9f823517cb14cb033..19468ae037f0705a37b14d5b3b519f61d396cbc6 100644 (file)
@@ -281,7 +281,7 @@ Signed-off-by: chuanjia.liu <Chuanjia.Liu@mediatek.com>
  &pio {
 --- a/arch/arm64/boot/dts/mediatek/mt7622.dtsi
 +++ b/arch/arm64/boot/dts/mediatek/mt7622.dtsi
-@@ -792,45 +792,41 @@
+@@ -794,45 +794,41 @@
                #reset-cells = <1>;
        };
  
@@ -344,7 +344,7 @@ Signed-off-by: chuanjia.liu <Chuanjia.Liu@mediatek.com>
                        interrupt-map-mask = <0 0 0 7>;
                        interrupt-map = <0 0 0 1 &pcie_intc0 0>,
                                        <0 0 0 2 &pcie_intc0 1>,
-@@ -842,15 +838,39 @@
+@@ -844,15 +840,39 @@
                                #interrupt-cells = <1>;
                        };
                };
index 8c9194e5f281d4a63db44bf3271fe28a18a46cea..873317eb0052e7d9fc4dd407facf13687c684f7f 100644 (file)
@@ -18,7 +18,7 @@ Signed-off-by: Felix Fietkau <nbd@nbd.name>
                        interface-type = "ace";
                        reg = <0x5000 0x1000>;
                };
-@@ -967,6 +967,8 @@
+@@ -969,6 +969,8 @@
                power-domains = <&scpsys MT7622_POWER_DOMAIN_ETHSYS>;
                mediatek,ethsys = <&ethsys>;
                mediatek,sgmiisys = <&sgmiisys>;
index 503cc8937dca1064d3324a6e44ff664af71ee786..0156d89bbaa7a19a46f22c69e9e32f35778ca9a3 100644 (file)
@@ -10,7 +10,7 @@ Signed-off-by: Felix Fietkau <nbd@nbd.name>
 
 --- a/arch/arm64/boot/dts/mediatek/mt7622.dtsi
 +++ b/arch/arm64/boot/dts/mediatek/mt7622.dtsi
-@@ -803,6 +803,8 @@
+@@ -805,6 +805,8 @@
                reg = <0 0x1a143000 0 0x1000>;
                reg-names = "port0";
                mediatek,pcie-cfg = <&pciecfg>;
@@ -19,7 +19,7 @@ Signed-off-by: Felix Fietkau <nbd@nbd.name>
                #address-cells = <3>;
                #size-cells = <2>;
                interrupts = <GIC_SPI 228 IRQ_TYPE_LEVEL_LOW>;
-@@ -820,6 +822,7 @@
+@@ -822,6 +824,7 @@
                bus-range = <0x00 0xff>;
                ranges = <0x82000000 0 0x20000000  0x0 0x20000000  0 0x8000000>;
                status = "disabled";
@@ -27,7 +27,7 @@ Signed-off-by: Felix Fietkau <nbd@nbd.name>
  
                slot0: pcie@0,0 {
                        reg = <0x0000 0 0 0 0>;
-@@ -846,6 +849,8 @@
+@@ -848,6 +851,8 @@
                reg = <0 0x1a145000 0 0x1000>;
                reg-names = "port1";
                mediatek,pcie-cfg = <&pciecfg>;
@@ -36,7 +36,7 @@ Signed-off-by: Felix Fietkau <nbd@nbd.name>
                #address-cells = <3>;
                #size-cells = <2>;
                interrupts = <GIC_SPI 229 IRQ_TYPE_LEVEL_LOW>;
-@@ -864,6 +869,7 @@
+@@ -866,6 +871,7 @@
                bus-range = <0x00 0xff>;
                ranges = <0x82000000 0 0x28000000  0x0 0x28000000  0 0x8000000>;
                status = "disabled";
@@ -44,7 +44,7 @@ Signed-off-by: Felix Fietkau <nbd@nbd.name>
  
                slot1: pcie@1,0 {
                        reg = <0x0800 0 0 0 0>;
-@@ -923,6 +929,11 @@
+@@ -925,6 +931,11 @@
                };
        };
  
index d96b1d6c3054191b5262de8c2aae9ccce262d3c6..94ed26c3df0f170b96e3320f433b39d07464f078 100644 (file)
@@ -104,13 +104,13 @@ Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
                };
  
 @@ -85,6 +80,7 @@
-                        compatible = "fsl,p2020-fcm-nand",
-                                     "fsl,elbc-fcm-nand";
-                        reg = <0x1 0x0 0x40000>;
-+                       nand-ecc-mode = "none";
-
-                        partition@0 {
-                                /* This location must not be altered  */
+                       compatible = "fsl,p2020-fcm-nand",
+                                    "fsl,elbc-fcm-nand";
+                       reg = <0x1 0x0 0x40000>;
++                      nand-ecc-mode = "none";
+                       partition@0 {
+                               /* This location must not be altered  */
 @@ -140,13 +136,43 @@
        soc: soc@ffe00000 {
                ranges = <0x0 0x0 0xffe00000 0x100000>;
@@ -155,7 +155,7 @@ Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
                spi@7000 {
                        flash@0 {
                                #address-cells = <1>;
-@@ -200,10 +225,12 @@
+@@ -200,10 +226,12 @@
                        phy0: ethernet-phy@0 {
                                interrupts = <3 1 0 0>;
                                reg = <0x0>;
index 4c4fbec76400c476ce020be06052ba268b69159f..607f436297126d9af9613ac42314742d157b4880 100644 (file)
  
        memory {
                device_type = "memory";
-@@ -70,10 +77,9 @@
+@@ -73,10 +80,9 @@
+               pinctrl-names = "default";
+               pinctrl-0 = <&helios_system_led_pins>;
  
-       system-leds {
-               compatible = "gpio-leds";
 -              status-led {
 +              led_status: status-led {
                        label = "helios4:green:status";
index 2e840fd7877f221859abb99462c0fdc30ce3e160..da4705d90d70423bc314679363b9b426c335d434 100644 (file)
@@ -9,7 +9,7 @@ Signed-off-by: Felix Fietkau <nbd@nbd.name>
 ---
 --- a/drivers/net/ethernet/marvell/mvneta.c
 +++ b/drivers/net/ethernet/marvell/mvneta.c
-@@ -4896,6 +4896,14 @@ static int mvneta_ethtool_set_eee(struct
+@@ -4903,6 +4903,14 @@ static int mvneta_ethtool_set_eee(struct
        return phylink_ethtool_set_eee(pp->phylink, eee);
  }
  
@@ -24,7 +24,7 @@ Signed-off-by: Felix Fietkau <nbd@nbd.name>
  static const struct net_device_ops mvneta_netdev_ops = {
        .ndo_open            = mvneta_open,
        .ndo_stop            = mvneta_stop,
-@@ -4906,6 +4914,7 @@ static const struct net_device_ops mvnet
+@@ -4913,6 +4921,7 @@ static const struct net_device_ops mvnet
        .ndo_fix_features    = mvneta_fix_features,
        .ndo_get_stats64     = mvneta_get_stats64,
        .ndo_do_ioctl        = mvneta_ioctl,
index eb2c9b9d83b62035c01bb4ba68ff878348dc29fa..e8123ee025949e2cdec4faa56ba30da9d3ebcee2 100644 (file)
@@ -26,7 +26,7 @@ use-case. You've been warned.
 
 --- a/arch/arm64/boot/dts/rockchip/rk3328.dtsi
 +++ b/arch/arm64/boot/dts/rockchip/rk3328.dtsi
-@@ -983,6 +983,33 @@
+@@ -984,6 +984,33 @@
                status = "disabled";
        };