)]}' { "log": [ { "commit": "f1359c240191e686614847905fc861cbda480b47", "tree": "88c249e4f8e088a8de7ae64c1a7ecbacf5e6fac1", "parents": [ "58e416e283284c15b8daf0b2b425f43242f62b0a" ], "author": { "name": "Nagamani PV", "email": "nagamani@linux.ibm.com", "time": "Mon Mar 30 13:44:36 2026 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Mar 31 20:14:56 2026 -0700" }, "message": "net/iucv: Add missing kernel-doc return value descriptions\n\nAdd missing return value descriptions for several functions in\nnet/iucv/af_iucv.c and net/iucv/iucv.c to address kernel-doc warnings.\n\nWarnings detected with:\nscripts/kernel-doc -none -Wall net/iucv/*\nWarning: net/iucv/af_iucv.c:131 No description found for return value of \u0027iucv_msg_length\u0027\nWarning: net/iucv/af_iucv.c:150 No description found for return value of \u0027iucv_sock_in_state\u0027\n...\n\nNo functional change.\n\nReviewed-by: Aswin Karuvally \u003caswin@linux.ibm.com\u003e\nReviewed-by: Alexandra Winter \u003cwintera@linux.ibm.com\u003e\nSigned-off-by: Nagamani PV \u003cnagamani@linux.ibm.com\u003e\nSigned-off-by: Alexandra Winter \u003cwintera@linux.ibm.com\u003e\nLink: https://patch.msgid.link/20260330114436.2010108-1-wintera@linux.ibm.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "58e416e283284c15b8daf0b2b425f43242f62b0a", "tree": "1935f1e11f15fcd9b6a95b76b9e7a7a45299fc93", "parents": [ "d7709812e13d06132ddae3d21540472ea5cb11c5" ], "author": { "name": "Fernando Fernandez Mancera", "email": "fmancera@suse.de", "time": "Mon Mar 30 14:10:33 2026 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Mar 31 20:10:48 2026 -0700" }, "message": "net: vxlan: check ipv6_mod_enabled() on neigh_reduce()\n\nIPv6 must be enabled or otherwise neigh_reduce() might cause a kernel\npanic. This was prevented by a check on in6_dev. Use ipv6_mod_enabled()\ninstead as it is cleaner and also consistent with the code at\nroute_shortcircuit().\n\nSigned-off-by: Fernando Fernandez Mancera \u003cfmancera@suse.de\u003e\nReviewed-by: Ido Schimmel \u003cidosch@nvidia.com\u003e\nLink: https://patch.msgid.link/20260330121033.4479-1-fmancera@suse.de\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "d7709812e13d06132ddae3d21540472ea5cb11c5", "tree": "df74ce4a38159a3460f18e289f0bef0f2fb2d6b0", "parents": [ "e5a3f10ed7c97decd61b3274982f048b58ce69bc" ], "author": { "name": "Erni Sri Satya Vennela", "email": "ernis@linux.microsoft.com", "time": "Thu Mar 26 10:30:56 2026 -0700" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Mar 31 19:42:06 2026 -0700" }, "message": "net: mana: hardening: Validate adapter_mtu from MANA_QUERY_DEV_CONFIG\n\nAs a part of MANA hardening for CVM, validate the adapter_mtu value\nreturned from the MANA_QUERY_DEV_CONFIG HWC command.\n\nThe adapter_mtu value is used to compute ndev-\u003emax_mtu via:\ngc-\u003eadapter_mtu - ETH_HLEN. If hardware returns a bogus adapter_mtu\nsmaller than ETH_HLEN (e.g. 0), the unsigned subtraction wraps to a\nhuge value, silently allowing oversized MTU settings.\n\nAdd a validation check to reject adapter_mtu values below\nETH_MIN_MTU + ETH_HLEN, returning -EPROTO to fail the device\nconfiguration early with a clear error message.\n\nSigned-off-by: Erni Sri Satya Vennela \u003cernis@linux.microsoft.com\u003e\nLink: https://patch.msgid.link/20260326173101.2010514-1-ernis@linux.microsoft.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "e5a3f10ed7c97decd61b3274982f048b58ce69bc", "tree": "fcb87acef183edd08f421c359906b0b16098f0ca", "parents": [ "a94ddc191f19579a7e0a5da2c012f1048ce10262", "aae5efaeb8aa4ada710d5b0cdbab77b9539c69eb" ], "author": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Mar 31 19:32:44 2026 -0700" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Mar 31 19:32:45 2026 -0700" }, "message": "Merge branch \u0027declance-improve-dma-error-reporting\u0027\n\nMaciej W. Rozycki says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\ndeclance: Improve DMA error reporting\n\nInspired by a recent discussion[1] I have come up with this pair of\nsmall improvements to DMA error reporting with declance.\n\n[1] Sebastian Andrzej Siewior, \"declance: Remove IRQF_ONESHOT\",\n \u003chttps://lore.kernel.org/r/20260127135334.qUEaYP9G@linutronix.de/\u003e\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nLink: https://patch.msgid.link/alpine.DEB.2.21.2603291835550.60268@angie.orcam.me.uk\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "aae5efaeb8aa4ada710d5b0cdbab77b9539c69eb", "tree": "fcb87acef183edd08f421c359906b0b16098f0ca", "parents": [ "ee769323b1bf60c0ec0338cc5ee6b1c725624ec6" ], "author": { "name": "Maciej W. Rozycki", "email": "macro@orcam.me.uk", "time": "Sun Mar 29 19:07:41 2026 +0100" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Mar 31 19:32:41 2026 -0700" }, "message": "declance: Include the offending address with DMA errors\n\nThe address latched in the I/O ASIC LANCE DMA Pointer Register uses the\nTURBOchannel bus address encoding and therefore bits 33:29 of location\nreferred occupy bits 4:0, bits 28:2 are left-shifted by 3, and bits 1:0\nare hardwired to zero. In reality no TURBOchannel system exceeds 1GiB\nof RAM though, so the address reported will always fit in 8 hex digits.\n\nSigned-off-by: Maciej W. Rozycki \u003cmacro@orcam.me.uk\u003e\nLink: https://patch.msgid.link/alpine.DEB.2.21.2603291839220.60268@angie.orcam.me.uk\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "ee769323b1bf60c0ec0338cc5ee6b1c725624ec6", "tree": "b22f19a4dd5b39a0ff9c499bd82c15239b14eda4", "parents": [ "a94ddc191f19579a7e0a5da2c012f1048ce10262" ], "author": { "name": "Maciej W. Rozycki", "email": "macro@orcam.me.uk", "time": "Sun Mar 29 19:07:24 2026 +0100" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Mar 31 19:32:41 2026 -0700" }, "message": "declance: Rate-limit DMA errors\n\nPrevent the system from becoming unusable due to a flood of DMA error\nmessages.\n\nSigned-off-by: Maciej W. Rozycki \u003cmacro@orcam.me.uk\u003e\nLink: https://patch.msgid.link/alpine.DEB.2.21.2603291838370.60268@angie.orcam.me.uk\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "a94ddc191f19579a7e0a5da2c012f1048ce10262", "tree": "c9443482b4498e7c642c170fc3e737eed635d7ab", "parents": [ "0cbd743aad9076c349334264774bcb9f9af1e40f" ], "author": { "name": "Lorenzo Bianconi", "email": "lorenzo@kernel.org", "time": "Mon Mar 30 00:03:49 2026 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Mar 31 19:30:10 2026 -0700" }, "message": "net: airoha: Fix typo in airoha_set_gdm2_loopback routine name\n\nRename airhoha_set_gdm2_loopback() in airoha_set_gdm2_loopback()\n\nSigned-off-by: Lorenzo Bianconi \u003clorenzo@kernel.org\u003e\nLink: https://patch.msgid.link/20260330-airoha_set_gdm2_loopback-fix-typo-v1-1-a1320ff6b6cc@kernel.org\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "0cbd743aad9076c349334264774bcb9f9af1e40f", "tree": "d7c7c22474681a93c7894637ea7abd2cd12cdf9b", "parents": [ "177750206f9d59463c45e95701e3e56c6f3577c6", "771a627ddd3b31fc5a27dc7af28a0e59ba13aea1" ], "author": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Mar 31 19:24:32 2026 -0700" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Mar 31 19:24:32 2026 -0700" }, "message": "Merge branch \u0027nfc-refactor-usb-endpoint-lookups\u0027\n\nJohan Hovold says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\nnfc: refactor USB endpoint lookups\n\nUse the common USB helper for looking up bulk and interrupt endpoints\ninstead of open coding.\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nLink: https://patch.msgid.link/20260330103655.1672331-1-johan@kernel.org\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "771a627ddd3b31fc5a27dc7af28a0e59ba13aea1", "tree": "d7c7c22474681a93c7894637ea7abd2cd12cdf9b", "parents": [ "13f2e141b944cf24171db83ae4ee07caf83d2123" ], "author": { "name": "Johan Hovold", "email": "johan@kernel.org", "time": "Mon Mar 30 12:36:55 2026 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Mar 31 19:24:30 2026 -0700" }, "message": "nfc: port100: refactor endpoint lookup\n\nUse the common USB helpers for looking up bulk and interrupt endpoints\n(and determining endpoint numbers) instead of open coding.\n\nSigned-off-by: Johan Hovold \u003cjohan@kernel.org\u003e\nLink: https://patch.msgid.link/20260330103655.1672331-4-johan@kernel.org\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "13f2e141b944cf24171db83ae4ee07caf83d2123", "tree": "89b8a63c172b2dc183b8dd195a62c0e6c594b380", "parents": [ "dba69cba4a5d827a0d05b9692bf8b4d9cbfe9206" ], "author": { "name": "Johan Hovold", "email": "johan@kernel.org", "time": "Mon Mar 30 12:36:54 2026 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Mar 31 19:24:30 2026 -0700" }, "message": "nfc: pn533: refactor endpoint lookup\n\nUse the common USB helpers for looking up bulk and interrupt endpoints\n(and determining endpoint numbers) instead of open coding.\n\nSigned-off-by: Johan Hovold \u003cjohan@kernel.org\u003e\nLink: https://patch.msgid.link/20260330103655.1672331-3-johan@kernel.org\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "dba69cba4a5d827a0d05b9692bf8b4d9cbfe9206", "tree": "3b39904806183d338d221224559779e8311f74fe", "parents": [ "177750206f9d59463c45e95701e3e56c6f3577c6" ], "author": { "name": "Johan Hovold", "email": "johan@kernel.org", "time": "Mon Mar 30 12:36:53 2026 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Mar 31 19:24:30 2026 -0700" }, "message": "nfc: nfcmrvl: refactor endpoint lookup\n\nUse the common USB helper for looking up bulk and interrupt endpoints\ninstead of open coding.\n\nSigned-off-by: Johan Hovold \u003cjohan@kernel.org\u003e\nLink: https://patch.msgid.link/20260330103655.1672331-2-johan@kernel.org\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "177750206f9d59463c45e95701e3e56c6f3577c6", "tree": "2762a950ce0c2c7d2c8162f7380d9372684895d2", "parents": [ "6730f184abdcf0cc9b6ee20f97c467ab497a900d", "abfca6b13bcfe4f2b70a88190d61f162656ed44e" ], "author": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Mar 31 19:23:42 2026 -0700" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Mar 31 19:23:43 2026 -0700" }, "message": "Merge branch \u0027net-refactor-usb-endpoint-lookups\u0027\n\nJohan Hovold says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\nnet: refactor USB endpoint lookups\n\nUse the common USB helpers for looking up bulk and interrupt endpoints\ninstead of open coding.\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nLink: https://patch.msgid.link/20260330102611.1671546-1-johan@kernel.org\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "abfca6b13bcfe4f2b70a88190d61f162656ed44e", "tree": "2762a950ce0c2c7d2c8162f7380d9372684895d2", "parents": [ "da45a55748f2cd89fb6100df46821af69bdcea99" ], "author": { "name": "Johan Hovold", "email": "johan@kernel.org", "time": "Mon Mar 30 12:26:11 2026 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Mar 31 19:23:41 2026 -0700" }, "message": "net: ipeth: refactor endpoint lookup\n\nUse the common USB helper for looking up bulk and interrupt endpoints\ninstead of open coding.\n\nSigned-off-by: Johan Hovold \u003cjohan@kernel.org\u003e\nLink: https://patch.msgid.link/20260330102611.1671546-3-johan@kernel.org\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "da45a55748f2cd89fb6100df46821af69bdcea99", "tree": "b2fa5a77910c5716d2b7ec8ad2f9b6d0eb8c2394", "parents": [ "6730f184abdcf0cc9b6ee20f97c467ab497a900d" ], "author": { "name": "Johan Hovold", "email": "johan@kernel.org", "time": "Mon Mar 30 12:26:10 2026 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Mar 31 19:23:41 2026 -0700" }, "message": "net: hso: refactor endpoint lookup\n\nUse the common USB helpers for looking up bulk and interrupt endpoints\ninstead of a custom implementation.\n\nSigned-off-by: Johan Hovold \u003cjohan@kernel.org\u003e\nLink: https://patch.msgid.link/20260330102611.1671546-2-johan@kernel.org\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "6730f184abdcf0cc9b6ee20f97c467ab497a900d", "tree": "91308fb67fded945f19cb16892a9fc59df13109d", "parents": [ "f843687c30272d55739ef153ace29c58db2575ee" ], "author": { "name": "Izabela Bakollari", "email": "ibakolla@redhat.com", "time": "Mon Mar 30 08:32:35 2026 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Mar 31 19:19:09 2026 -0700" }, "message": "sfc: add transmit timestamping support\n\nAdd ethtool op to advertise TX timestamping. Insert a\nskb_tx_timestamp call in __efx_enqueue_skb.\n\nSigned-off-by: Izabela Bakollari \u003cibakolla@redhat.com\u003e\nReviewed-by: Edward Cree \u003cecree.xilinx@gmail.com\u003e\nLink: https://patch.msgid.link/20260330063236.247057-1-ibakolla@redhat.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "f843687c30272d55739ef153ace29c58db2575ee", "tree": "7d93b2c553437233611443feb063199edcf3c777", "parents": [ "75171eeff3538c6859b50f6a516342c8f1ed6e00" ], "author": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon Mar 30 17:19:30 2026 -0700" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Mar 31 19:18:20 2026 -0700" }, "message": "selftests: drv-net: update the README with variants\n\nTest authors need to know about variants, existing tests don\u0027t use\nthem because variants are relatively recent.\n\nReviewed-by: Joe Damato \u003cjoe@dama.to\u003e\nLink: https://patch.msgid.link/20260331001930.3411279-1-kuba@kernel.org\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "75171eeff3538c6859b50f6a516342c8f1ed6e00", "tree": "8098dbb02b01e936a70a489ff77bfb1db3fd349c", "parents": [ "e9abf1da0af3f787a03b249945e5ca726c1b8013" ], "author": { "name": "Daniel Wagner", "email": "wagner.daniel.t@gmail.com", "time": "Mon Mar 30 23:53:10 2026 +0100" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Mar 31 19:17:27 2026 -0700" }, "message": "net: phy: bcm84881: add BCM84891/BCM84892 support\n\nThe BCM84891 and BCM84892 are 10GBASE-T PHYs in the same family as the\nBCM84881, sharing the register map and most callbacks. They add USXGMII\nas a host interface mode.\n\nbcm8489x_config_init() is separate from bcm84881_config_init(): it\nallows only USXGMII (the only host mode available on the tested\nhardware) and clears MDIO_CTRL1_LPOWER, which is set at boot on the\ntested platform. Does not recur on ifdown/ifup, cable events, or\nlink-partner advertisement changes, so config_init is sufficient.\n\nFor USXGMII, read_status() skips the 0x4011 host-mode register: it\nreturns the same value regardless of negotiated copper speed (USXGMII\nsymbol replication). Speed comes from phy_resolve_aneg_linkmode() via\nstandard C45 AN resolution.\n\nTested on TRENDnet TEG-S750 (RTL9303 + 1x BCM84891 + 4x BCM84892)\nrunning OpenWrt, where the MDIO controller driver is currently\nOpenWrt-specific. Link verified at 100M, 1G, 2.5G, 10G.\n\nSigned-off-by: Daniel Wagner \u003cwagner.daniel.t@gmail.com\u003e\nReviewed-by: Florian Fainelli \u003cflorian.fainelli@broadcom.com\u003e\nReviewed-by: Nicolai Buchwitz \u003cnb@tipi-net.de\u003e\nReviewed-by: Russell King (Oracle) \u003crmk+kernel@armlinux.org.uk\u003e\nLink: https://patch.msgid.link/20260330225310.2801264-1-wagner.daniel.t@gmail.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "e9abf1da0af3f787a03b249945e5ca726c1b8013", "tree": "cf6bbbfdd1b9dfa6229217e4d5cdf89ca1949a34", "parents": [ "9229cb5a941c981633a79c9867f339c4589a9da9" ], "author": { "name": "Daniel Golle", "email": "daniel@makrotopia.org", "time": "Mon Mar 30 23:52:09 2026 +0100" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Mar 31 19:16:31 2026 -0700" }, "message": "net: dsa: mxl862xx: cancel pending work on probe error\n\nCall mxl862xx_host_shutdown() in case dsa_register_switch() returns\nan error, so any still pending crc_err_work get canceled.\n\nFixes: a319d0c8c8ce (\"net: dsa: mxl862xx: add CRC for MDIO communication\")\nSigned-off-by: Daniel Golle \u003cdaniel@makrotopia.org\u003e\nReviewed-by: Andrew Lunn \u003candrew@lunn.ch\u003e\nLink: https://patch.msgid.link/3fd163f5bb88de426ca9847549f94b4296170ef0.1774911025.git.daniel@makrotopia.org\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "9229cb5a941c981633a79c9867f339c4589a9da9", "tree": "ed2947362ea65ed36e20df402c48d19b6f4335cb", "parents": [ "93d04e76bcf1e81f36f5ea7ad620a07747f1527c" ], "author": { "name": "Alexander Duyck", "email": "alexanderduyck@fb.com", "time": "Fri Mar 27 13:44:45 2026 -0700" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Tue Mar 31 16:05:19 2026 +0200" }, "message": "fbnic: Set Relaxed Ordering PCIe TLP attributes for DMA engines\n\nAdd ATTR CSR bit field definitions for the DMA engine TLP header\nconfiguration registers:\n AW_CFG: RDE_ATTR[17:15], RQM_ATTR[14:12], TQM_ATTR[11:9]\n AR_CFG: TDE_ATTR[17:15], RQM_ATTR[14:12], TQM_ATTR[11:9]\n\nThese fields control the PCIe TLP attribute bits for outbound\ntransactions from the TQM, RQM, RDE (write path), and TDE (read path)\nDMA engines. An enum is added with standard PCIe TLP attribute values:\nNS (No Snoop), RO (Relaxed Ordering), and IDO (ID-based Ordering).\n\nRead the PCIe Relaxed Ordering capability at probe time and store it in\nfbnic_dev. Configure Relaxed Ordering on the PCIe TLP attributes in\nfbnic_mbx_init_desc_ring when the capability is enabled. For the write\npath (AW_CFG), set RO on RDE and TQM attributes. For the read path\n(AR_CFG), set RO on all three attributes (TDE, RQM, TQM). This allows\nthe PCIe fabric to reorder these transactions for improved throughput.\n\nSigned-off-by: Alexander Duyck \u003calexanderduyck@fb.com\u003e\nSigned-off-by: Dimitri Daskalakis \u003cdaskald@meta.com\u003e\nLink: https://patch.msgid.link/20260327204445.3074446-1-dimitri.daskalakis1@gmail.com\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "93d04e76bcf1e81f36f5ea7ad620a07747f1527c", "tree": "4eb17f9a01c015b955a34a0e2d663fda1bff9cf0", "parents": [ "7fae6616704a17c64438ad4b73a6effa6c03ffda", "7db154aa58e18bb663cc317f0b62631f8b3d2b87" ], "author": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Tue Mar 31 10:28:57 2026 +0200" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Tue Mar 31 10:28:57 2026 +0200" }, "message": "Merge branch \u0027r8152-add-helper-functions-for-pla-usb-phy-ocp-registers\u0027\n\nChih Kai Hsu says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\nr8152: add helper functions for PLA/USB/PHY OCP registers\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nLink: https://patch.msgid.link/20260326073925.32976-453-nic_swsd@realtek.com\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "7db154aa58e18bb663cc317f0b62631f8b3d2b87", "tree": "4eb17f9a01c015b955a34a0e2d663fda1bff9cf0", "parents": [ "d16133f177fea620d6b413990683b6a8b2641be5" ], "author": { "name": "Chih Kai Hsu", "email": "hsu.chih.kai@realtek.com", "time": "Thu Mar 26 15:39:25 2026 +0800" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Tue Mar 31 10:28:26 2026 +0200" }, "message": "r8152: add helper functions for PHY OCP registers\n\nAdd the following bitwise operation functions for PHY OCP registers to\nsimplify the code.\n\n- ocp_reg_w0w1()\n- ocp_reg_clr_bits()\n- ocp_reg_set_bits()\n- sram_write_w0w1()\n- sram_clr_bits()\n- sram_set_bits()\n- r8152_mdio_clr_bit()\n- r8152_mdio_set_bit()\n- r8152_mdio_test_and_clr_bit()\n\nIn addition, remove variable set but not used from r8153_init(),\nr8153b_init() and r8153c_init().\n\nSigned-off-by: Chih Kai Hsu \u003chsu.chih.kai@realtek.com\u003e\nReviewed-by: Hayes Wang \u003chayeswang@realtek.com\u003e\nLink: https://patch.msgid.link/20260326073925.32976-456-nic_swsd@realtek.com\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "d16133f177fea620d6b413990683b6a8b2641be5", "tree": "c74bc5c5fd8b6bdd1e311fe0a7f48315992797b7", "parents": [ "48afd5124fd6129c46fd12cb06155384b1c4a0c4" ], "author": { "name": "Chih Kai Hsu", "email": "hsu.chih.kai@realtek.com", "time": "Thu Mar 26 15:39:24 2026 +0800" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Tue Mar 31 10:28:26 2026 +0200" }, "message": "r8152: add helper functions for PLA/USB OCP registers\n\nAdd the following bitwise operation functions for PLA/USB OCP registers\nto simplify the code.\n\n- ocp_dword_w0w1()\n- ocp_word_w0w1()\n- ocp_byte_w0w1()\n- ocp_dword_clr_bits()\n- ocp_dword_set_bits()\n- ocp_word_clr_bits()\n- ocp_word_set_bits()\n- ocp_word_test_and_clr_bits()\n- ocp_byte_clr_bits()\n- ocp_byte_set_bits()\n\nSigned-off-by: Chih Kai Hsu \u003chsu.chih.kai@realtek.com\u003e\nReviewed-by: Hayes Wang \u003chayeswang@realtek.com\u003e\nLink: https://patch.msgid.link/20260326073925.32976-455-nic_swsd@realtek.com\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "48afd5124fd6129c46fd12cb06155384b1c4a0c4", "tree": "7757e8eacb9b983cc1bbdefc1c6e30015770ee89", "parents": [ "7fae6616704a17c64438ad4b73a6effa6c03ffda" ], "author": { "name": "Chih Kai Hsu", "email": "hsu.chih.kai@realtek.com", "time": "Thu Mar 26 15:39:23 2026 +0800" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Tue Mar 31 10:28:26 2026 +0200" }, "message": "r8152: fix incorrect register write to USB_UPHY_XTAL\n\nThe old code used ocp_write_byte() to clear the OOBS_POLLING bit\n(BIT(8)) in the USB_UPHY_XTAL register, but this doesn\u0027t correctly\nclear a bit in the upper byte of the 16-bit register.\n\nFix this by using ocp_write_word() instead.\n\nFixes: 195aae321c82 (\"r8152: support new chips\")\nSigned-off-by: Chih Kai Hsu \u003chsu.chih.kai@realtek.com\u003e\nReviewed-by: Hayes Wang \u003chayeswang@realtek.com\u003e\nLink: https://patch.msgid.link/20260326073925.32976-454-nic_swsd@realtek.com\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "7fae6616704a17c64438ad4b73a6effa6c03ffda", "tree": "00bc2a423d54a9167e5ed3c1fb7a90d3a47a2777", "parents": [ "f4ef5b1c13316fa518b7a41e6bff82c33239823c" ], "author": { "name": "Maciej W. Rozycki", "email": "macro@orcam.me.uk", "time": "Sun Mar 29 13:32:25 2026 +0100" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon Mar 30 18:11:37 2026 -0700" }, "message": "FDDI: defxx: Rate-limit memory allocation errors\n\nPrevent the system from becoming unstable or unusable due to a flood of\nmemory allocation error messages under memory pressure, e.g.:\n\n[...]\nfddi0: Could not allocate receive buffer. Dropping packet.\nfddi0: Could not allocate receive buffer. Dropping packet.\nfddi0: Could not allocate receive buffer. Dropping packet.\nfddi0: Could not allocate receive buffer. Dropping packet.\nrcu: INFO: rcu_sched self-detected stall on CPU\nrcu: \t0-...!: (332 ticks this GP) idle\u003d255c/1/0x40000000 softirq\u003d16420123/16420123 fqs\u003d0\nrcu: \t(t\u003d2103 jiffies g\u003d35680089 q\u003d4 ncpus\u003d1)\nrcu: rcu_sched kthread timer wakeup didn\u0027t happen for 2102 jiffies! g35680089 f0x0 RCU_GP_WAIT_FQS(5) -\u003estate\u003d0x402\nrcu: \tPossible timer handling issue on cpu\u003d0 timer-softirq\u003d12779658\nrcu: rcu_sched kthread starved for 2103 jiffies! g35680089 f0x0 RCU_GP_WAIT_FQS(5) -\u003estate\u003d0x402 -\u003ecpu\u003d0\nrcu: \tUnless rcu_sched kthread gets sufficient CPU time, OOM is now expected behavior.\nrcu: RCU grace-period kthread stack dump:\ntask:rcu_sched state:I stack:0 pid:14 tgid:14 ppid:2 flags:0x00004000\nCall Trace:\n __schedule+0x258/0x580\n schedule+0x19/0xa0\n schedule_timeout+0x4a/0xb0\n ? hrtimers_cpu_dying+0x1b0/0x1b0\n rcu_gp_fqs_loop+0xb1/0x450\n rcu_gp_kthread+0x9d/0x130\n kthread+0xb2/0xe0\n ? rcu_gp_init+0x4a0/0x4a0\n ? kthread_park+0x90/0x90\n ret_from_fork+0x2d/0x50\n ? kthread_park+0x90/0x90\n ret_from_fork_asm+0x12/0x20\n entry_INT80_32+0x10d/0x10d\nCPU: 0 UID: 500 PID: 21895 Comm: 31370.exe Not tainted 6.13.0-dirty #2\n\n(here running the libstdc++-v3 testsuite).\n\nSigned-off-by: Maciej W. Rozycki \u003cmacro@orcam.me.uk\u003e\nReviewed-by: Andrew Lunn \u003candrew@lunn.ch\u003e\nLink: https://patch.msgid.link/alpine.DEB.2.21.2603291236590.60268@angie.orcam.me.uk\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "f4ef5b1c13316fa518b7a41e6bff82c33239823c", "tree": "b168c107a70becb6ed9d19eb7479f5463cab2271", "parents": [ "0fb03e7038e4f91d44776021ce301af8d4efe0ce" ], "author": { "name": "Maciej W. Rozycki", "email": "macro@orcam.me.uk", "time": "Sun Mar 29 13:32:34 2026 +0100" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon Mar 30 18:11:31 2026 -0700" }, "message": "FDDI: defza: Rate-limit memory allocation errors\n\nPrevent the system from becoming unstable or unusable due to a flood of\nmemory allocation error messages under memory pressure.\n\nSigned-off-by: Maciej W. Rozycki \u003cmacro@orcam.me.uk\u003e\nReviewed-by: Andrew Lunn \u003candrew@lunn.ch\u003e\nLink: https://patch.msgid.link/alpine.DEB.2.21.2603291252380.60268@angie.orcam.me.uk\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "0fb03e7038e4f91d44776021ce301af8d4efe0ce", "tree": "ca13c61b6846d8d57c74a218f21a9ec37034a2ee", "parents": [ "911e2c050963ccf239faec6ae9dee0f5e8f1cc5c", "7f9f301660051353282e579affcd40e7075ab223" ], "author": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon Mar 30 17:36:52 2026 -0700" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon Mar 30 17:36:52 2026 -0700" }, "message": "Merge branch \u0027net-stmmac-qcom-ethqos-more-cleanups\u0027\n\nRussell King says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\nnet: stmmac: qcom-ethqos: more cleanups\n\nFurther cleanups to qcom-ethqos, mainly concentrating on the RGMII\ncode, making it clearer what the differences are for each speed, thus\nmaking the code more readable.\n\nI\u0027m still not really happy with this. The speed specific configuration\nremains split between ethqos_fix_mac_speed_rgmii() and\nethqos_rgmii_macro_init(), where the latter is only ever called from\nthe former. So, I think further work is needed here - maybe it needs\nrestructuring into the various componenet parts of the RGMII block?\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nLink: https://patch.msgid.link/acZDEg9wdjhBTHlL@shell.armlinux.org.uk\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "7f9f301660051353282e579affcd40e7075ab223", "tree": "ca13c61b6846d8d57c74a218f21a9ec37034a2ee", "parents": [ "2d350a892aad3b3dfdcfc170e7681f4095808a0c" ], "author": { "name": "Russell King (Oracle)", "email": "rmk+kernel@armlinux.org.uk", "time": "Fri Mar 27 08:44:39 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon Mar 30 17:36:46 2026 -0700" }, "message": "net: stmmac: qcom-ethqos: move phase_shift to register update site\n\nMove the determination of the phase shift enable alongside the register\nupdate, and make \"phase_shift\" unsigned.\n\nSigned-off-by: Russell King (Oracle) \u003crmk+kernel@armlinux.org.uk\u003e\nTested-by: Mohd Ayaan Anwar \u003cmohd.anwar@oss.qualcomm.com\u003e\nLink: https://patch.msgid.link/E1w62o3-0000000E3DE-3Vf8@rmk-PC.armlinux.org.uk\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "2d350a892aad3b3dfdcfc170e7681f4095808a0c", "tree": "3959a6f2b18ed88bfe2cc5290e1b838a085645fa", "parents": [ "67343aa24e595ac2522d9f7e2daba7ec29f32926" ], "author": { "name": "Russell King (Oracle)", "email": "rmk+kernel@armlinux.org.uk", "time": "Fri Mar 27 08:44:34 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon Mar 30 17:36:46 2026 -0700" }, "message": "net: stmmac: qcom-ethqos: correct prg_rclk_dly comment\n\nThe comment for calculating the prg_rclk_dly value is incorrect as it\nomits the brackets around the divisor. Add the brackets to allow the\nreader to correctly evaluate the value. Validated with the values given\nin the driver.\n\nSigned-off-by: Russell King (Oracle) \u003crmk+kernel@armlinux.org.uk\u003e\nTested-by: Mohd Ayaan Anwar \u003cmohd.anwar@oss.qualcomm.com\u003e\nLink: https://patch.msgid.link/E1w62ny-0000000E3D8-38Yp@rmk-PC.armlinux.org.uk\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "67343aa24e595ac2522d9f7e2daba7ec29f32926", "tree": "1fb9ad4550b97f88e723f3996a4ac58090fae5ae", "parents": [ "3df0e86f8f8d708a820f7844f1a5830839f78efd" ], "author": { "name": "Russell King (Oracle)", "email": "rmk+kernel@armlinux.org.uk", "time": "Fri Mar 27 08:44:29 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon Mar 30 17:36:46 2026 -0700" }, "message": "net: stmmac: qcom-ethqos: move loopback decision next to reg update\n\nMove the loopback decision next to the register update, and make the\nlocal variable unsigned. As a result, there is now no need for the\ncomment referring to the programming being later.\n\nSigned-off-by: Russell King (Oracle) \u003crmk+kernel@armlinux.org.uk\u003e\nTested-by: Mohd Ayaan Anwar \u003cmohd.anwar@oss.qualcomm.com\u003e\nLink: https://patch.msgid.link/E1w62nt-0000000E3D2-2fWk@rmk-PC.armlinux.org.uk\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "3df0e86f8f8d708a820f7844f1a5830839f78efd", "tree": "9d5ad0126600e541c20c88007210c5c7358472df", "parents": [ "439a27f21ecc1c7c59d6e00c7c6b9be380f66a64" ], "author": { "name": "Russell King (Oracle)", "email": "rmk+kernel@armlinux.org.uk", "time": "Fri Mar 27 08:44:24 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon Mar 30 17:36:46 2026 -0700" }, "message": "net: stmmac: qcom-ethqos: simplify prg_rclk_dly programming\n\nRather than coding the entire register update twice with different\nvalues, use a local variable to specify the value and have one\nregister update statement that uses this local variable. This results\nin neater code.\n\nSigned-off-by: Russell King (Oracle) \u003crmk+kernel@armlinux.org.uk\u003e\nTested-by: Mohd Ayaan Anwar \u003cmohd.anwar@oss.qualcomm.com\u003e\nLink: https://patch.msgid.link/E1w62no-0000000E3Cw-2EmH@rmk-PC.armlinux.org.uk\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "439a27f21ecc1c7c59d6e00c7c6b9be380f66a64", "tree": "e3b9976acb420c7e19b4a0645658142c136ec491", "parents": [ "432c8a9f5528e362cc42de46901f25373a3a9928" ], "author": { "name": "Russell King (Oracle)", "email": "rmk+kernel@armlinux.org.uk", "time": "Fri Mar 27 08:44:19 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon Mar 30 17:36:46 2026 -0700" }, "message": "net: stmmac: qcom-ethqos: finally eliminate the switch\n\nMove the RCLK delay configuration out of the switch, which just leaves\nthe RGMII_CONFIG_LOOPBACK_EN setting in all three paths. This makes it\ntrivial to eliminate the switch.\n\nSigned-off-by: Russell King (Oracle) \u003crmk+kernel@armlinux.org.uk\u003e\nTested-by: Mohd Ayaan Anwar \u003cmohd.anwar@oss.qualcomm.com\u003e\nLink: https://patch.msgid.link/E1w62nj-0000000E3Cq-1lPL@rmk-PC.armlinux.org.uk\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "432c8a9f5528e362cc42de46901f25373a3a9928", "tree": "34291da9255c15515b2835de10e4a1dbead4706a", "parents": [ "dae1de3df3e17b98fe78c4b1ba26322b03cd9f46" ], "author": { "name": "Russell King (Oracle)", "email": "rmk+kernel@armlinux.org.uk", "time": "Fri Mar 27 08:44:14 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon Mar 30 17:36:45 2026 -0700" }, "message": "net: stmmac: qcom-ethqos: move RGMII_CONFIG2_RX_PROG_SWAP\n\nMove RGMII_CONFIG2_RX_PROG_SWAP out of the switch. 1G speed always\nsets this field. 100M and 10M sets it for has_emac_ge_3 devices,\notherwise it is cleared.\n\nSigned-off-by: Russell King (Oracle) \u003crmk+kernel@armlinux.org.uk\u003e\nTested-by: Mohd Ayaan Anwar \u003cmohd.anwar@oss.qualcomm.com\u003e\nLink: https://patch.msgid.link/E1w62ne-0000000E3Ck-1Haf@rmk-PC.armlinux.org.uk\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "dae1de3df3e17b98fe78c4b1ba26322b03cd9f46", "tree": "7188246025e92f624378acc6698c7079ecb6def9", "parents": [ "8b19a91844203db70584a2e56ad97d117e2c0277" ], "author": { "name": "Russell King (Oracle)", "email": "rmk+kernel@armlinux.org.uk", "time": "Fri Mar 27 08:44:09 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon Mar 30 17:36:45 2026 -0700" }, "message": "net: stmmac: qcom-ethqos: move RGMII_CONFIG2_RSVD_CONFIG15 out\n\nAll paths through the switch clear the RGMII_CONFIG2_RSVD_CONFIG15\nfield. move it out of the switch statement.\n\nSigned-off-by: Russell King (Oracle) \u003crmk+kernel@armlinux.org.uk\u003e\nTested-by: Mohd Ayaan Anwar \u003cmohd.anwar@oss.qualcomm.com\u003e\nLink: https://patch.msgid.link/E1w62nZ-0000000E3Ce-0lyP@rmk-PC.armlinux.org.uk\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "8b19a91844203db70584a2e56ad97d117e2c0277", "tree": "1c1179d939de3ab6f7e4025b62cde36e1f09271a", "parents": [ "dd07f2f9149a885ea2d8d19f95439227ec558106" ], "author": { "name": "Russell King (Oracle)", "email": "rmk+kernel@armlinux.org.uk", "time": "Fri Mar 27 08:44:04 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon Mar 30 17:36:45 2026 -0700" }, "message": "net: stmmac: qcom-ethqos: move 100M/10M speed programming\n\nMove the speed programming for 100M and 10M out of the switch. There\nis no programming done for 1G speed.\n\nIt looks like there are two fields, 7:6 which are programemd to \u00271\u0027\nto select a /2 divisor for 100M, and bits 16:8 which are programmed\nto \u002719\u0027 to select a /20 divisor.\n\nSigned-off-by: Russell King (Oracle) \u003crmk+kernel@armlinux.org.uk\u003e\nTested-by: Mohd Ayaan Anwar \u003cmohd.anwar@oss.qualcomm.com\u003e\nLink: https://patch.msgid.link/E1w62nU-0000000E3CX-0KF9@rmk-PC.armlinux.org.uk\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "dd07f2f9149a885ea2d8d19f95439227ec558106", "tree": "2fbf76c3fd63f4eb1c853891b57bf21dae262aff", "parents": [ "82d5fdc82a33bea93baec4795e835b016d2dc87e" ], "author": { "name": "Russell King (Oracle)", "email": "rmk+kernel@armlinux.org.uk", "time": "Fri Mar 27 08:43:58 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon Mar 30 17:36:45 2026 -0700" }, "message": "net: stmmac: qcom-ethqos: move two more RGMII_IO_MACRO_CONFIG2 out\n\nRGMII_CONFIG2_DATA_DIVIDE_CLK_SEL is always cleared, and\nRGMII_CONFIG2_TX_CLK_PHASE_SHIFT_EN is always updated with the phase\nshift in each path through the switch, so these are independent of\nthe speed. Move them out.\n\nSigned-off-by: Russell King (Oracle) \u003crmk+kernel@armlinux.org.uk\u003e\nTested-by: Mohd Ayaan Anwar \u003cmohd.anwar@oss.qualcomm.com\u003e\nLink: https://patch.msgid.link/E1w62nO-0000000E3CR-445p@rmk-PC.armlinux.org.uk\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "82d5fdc82a33bea93baec4795e835b016d2dc87e", "tree": "b29ef2f7d4269231c8e30dffa29998c744e29783", "parents": [ "6be23c4c636acc08c691e2b7371f341b13de23f7" ], "author": { "name": "Russell King (Oracle)", "email": "rmk+kernel@armlinux.org.uk", "time": "Fri Mar 27 08:43:53 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon Mar 30 17:36:45 2026 -0700" }, "message": "net: stmmac: qcom-ethqos: move 1G vs 100M/10M RGMII settings\n\nMove RGMII_CONFIG_BYPASS_TX_ID_EN, RGMII_CONFIG_POS_NEG_DATA_SEL and\nRGMII_CONFIG_PROG_SWAP. There are two states for these: one group for\n1G, and the logical inversion for 100M and 10M. Move this out of the\nswitch into an if-else clause.\n\nSigned-off-by: Russell King (Oracle) \u003crmk+kernel@armlinux.org.uk\u003e\nTested-by: Mohd Ayaan Anwar \u003cmohd.anwar@oss.qualcomm.com\u003e\nLink: https://patch.msgid.link/E1w62nJ-0000000E3CL-3YSr@rmk-PC.armlinux.org.uk\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "6be23c4c636acc08c691e2b7371f341b13de23f7", "tree": "d27afcaa7a4fb2e4d41b33b803b43e841b190a7e", "parents": [ "426ce4677e8101996d1ff230464ea696eba2d853" ], "author": { "name": "Russell King (Oracle)", "email": "rmk+kernel@armlinux.org.uk", "time": "Fri Mar 27 08:43:48 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon Mar 30 17:36:44 2026 -0700" }, "message": "net: stmmac: qcom-ethqos: move RGMII_CONFIG_DDR_MODE\n\nRGMII_CONFIG_DDR_MODE is always set irrespective of the speed. Move\nthis out of the switch.\n\nSigned-off-by: Russell King (Oracle) \u003crmk+kernel@armlinux.org.uk\u003e\nTested-by: Mohd Ayaan Anwar \u003cmohd.anwar@oss.qualcomm.com\u003e\nLink: https://patch.msgid.link/E1w62nE-0000000E3CF-331r@rmk-PC.armlinux.org.uk\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "426ce4677e8101996d1ff230464ea696eba2d853", "tree": "dda8ef95cd18eca20470c1bd6f7103a9799de9db", "parents": [ "e9ed46a0b129d743f6ca008101b055fdd2e5ea2e" ], "author": { "name": "Russell King (Oracle)", "email": "rmk+kernel@armlinux.org.uk", "time": "Fri Mar 27 08:43:43 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon Mar 30 17:36:44 2026 -0700" }, "message": "net: stmmac: qcom-ethqos: move detection of invalid RGMII speed\n\nMove detection of invalid RGMII speeds (which will never be triggered)\nbefore the switch() to allow register modifications that are common to\nall speeds to be moved out of the switch.\n\nSigned-off-by: Russell King (Oracle) \u003crmk+kernel@armlinux.org.uk\u003e\nTested-by: Mohd Ayaan Anwar \u003cmohd.anwar@oss.qualcomm.com\u003e\nLink: https://patch.msgid.link/E1w62n9-0000000E3C9-2Zkr@rmk-PC.armlinux.org.uk\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "e9ed46a0b129d743f6ca008101b055fdd2e5ea2e", "tree": "ef5a78b1ff874206de8be7a5066f2b55c42bca1e", "parents": [ "673416fb5b410b536f45e738ac34e482efd94e81" ], "author": { "name": "Russell King (Oracle)", "email": "rmk+kernel@armlinux.org.uk", "time": "Fri Mar 27 08:43:38 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon Mar 30 17:36:44 2026 -0700" }, "message": "net: stmmac: qcom-ethqos: eliminate configure_func\n\nSince ethqos_fix_mac_speed() is called via a function pointer, and only\nindirects via the configure_func function pointer, eliminate this\nunnecessary indirection.\n\nSigned-off-by: Russell King (Oracle) \u003crmk+kernel@armlinux.org.uk\u003e\nTested-by: Mohd Ayaan Anwar \u003cmohd.anwar@oss.qualcomm.com\u003e\nLink: https://patch.msgid.link/E1w62n4-0000000E3C3-251S@rmk-PC.armlinux.org.uk\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "673416fb5b410b536f45e738ac34e482efd94e81", "tree": "15e8dd3be9fb80deeeab62db5f79e0999c28bfcc", "parents": [ "c3dd3b1e76e0ff20ce97e99bc8d1f3152eb177b9" ], "author": { "name": "Russell King (Oracle)", "email": "rmk+kernel@armlinux.org.uk", "time": "Fri Mar 27 08:43:33 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon Mar 30 17:36:44 2026 -0700" }, "message": "net: stmmac: qcom-ethqos: pass ethqos to ethqos_pcs_set_inband()\n\nRather than getting the stmmac_priv pointer in\nethqos_configure_sgmii(), move it into ethqos_pcs_set_inband() and pass\nthe struct qcom_ethqos pointer instead.\n\nSigned-off-by: Russell King (Oracle) \u003crmk+kernel@armlinux.org.uk\u003e\nTested-by: Mohd Ayaan Anwar \u003cmohd.anwar@oss.qualcomm.com\u003e\nLink: https://patch.msgid.link/E1w62mz-0000000E3Bx-1Xd8@rmk-PC.armlinux.org.uk\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "c3dd3b1e76e0ff20ce97e99bc8d1f3152eb177b9", "tree": "b8c77dec25b6f0dd2fed209943d8adf532069edc", "parents": [ "911e2c050963ccf239faec6ae9dee0f5e8f1cc5c" ], "author": { "name": "Russell King (Oracle)", "email": "rmk+kernel@armlinux.org.uk", "time": "Fri Mar 27 08:43:28 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon Mar 30 17:36:44 2026 -0700" }, "message": "net: stmmac: qcom-ethqos: remove ethqos_configure()\n\nethqos_configure() does nothing more than indirect via\nethqos-\u003econfigure_func, and is only called from ethqos_fix_mac_speed()\njust below. Move the indirect call into ethqos_fix_mac_speed().\n\nSigned-off-by: Russell King (Oracle) \u003crmk+kernel@armlinux.org.uk\u003e\nTested-by: Mohd Ayaan Anwar \u003cmohd.anwar@oss.qualcomm.com\u003e\nLink: https://patch.msgid.link/E1w62mu-0000000E3Bq-15wa@rmk-PC.armlinux.org.uk\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "911e2c050963ccf239faec6ae9dee0f5e8f1cc5c", "tree": "c57728091c9cdd9702bd000cf647007d5efc2af5", "parents": [ "cf0d9080c6f795bc6be08babbffa29b62c06e9b0" ], "author": { "name": "Jan Hoffmann", "email": "jan@3e8.eu", "time": "Sun Mar 29 21:11:11 2026 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon Mar 30 16:27:44 2026 -0700" }, "message": "net: sfp: add quirk for ZOERAX SFP-2.5G-T\n\nThis is a 2.5G copper module which appears to be based on a Motorcomm\nYT8821 PHY. There doesn\u0027t seem to be a usable way to to access the PHY\n(I2C address 0x56 provides only read-only C22 access, and Rollball is\nalso not working).\n\nThe module does not report the correct extended compliance code for\n2.5GBase-T, and instead claims to support SONET OC-48 and Fibre Channel:\n\n Identifier : 0x03 (SFP)\n Extended identifier : 0x04 (GBIC/SFP defined by 2-wire interface ID)\n Connector : 0x07 (LC)\n Transceiver codes : 0x00 0x01 0x00 0x00 0x40 0x40 0x04 0x00 0x00\n Transceiver type : FC: Multimode, 50um (M5)\n Encoding : 0x05 (SONET Scrambled)\n BR Nominal : 2500MBd\n\nDespite this, the kernel still enables the correct 2500Base-X interface\nmode. However, for the module to actually work, it is also necessary to\ndisable inband auto-negotiation.\n\nEnable the existing \"sfp_quirk_oem_2_5g\" for this module, which handles\nthat and also sets the bit for 2500Base-T link mode.\n\nSigned-off-by: Jan Hoffmann \u003cjan@3e8.eu\u003e\nReviewed-by: Russell King (Oracle) \u003crmk+kernel@armlinux.org.uk\u003e\nLink: https://patch.msgid.link/20260329191304.720160-1-jan@3e8.eu\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "cf0d9080c6f795bc6be08babbffa29b62c06e9b0", "tree": "5d2f46b48109b2779f70e68b55fbbf60d3325c9a", "parents": [ "686aad8353a5a4d3358deaba08b748a6b6e6d6c2", "137ac69c15fd2a9acf5d4d95456fe75f23c9ffc5" ], "author": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:37:52 2026 -0700" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:37:53 2026 -0700" }, "message": "Merge branch \u0027net-hsr-subsystem-cleanups-and-modernization\u0027\n\nLuka Gejak says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\nnet: hsr: subsystem cleanups and modernization\n\nThis series contains two focused HSR cleanups with practical benefit.\nIt constifies protocol operation tables and replaces a hardcoded\nfunction name with __func__ to keep diagnostics correct across\nrefactoring.\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nLink: https://patch.msgid.link/20260326174600.136232-1-luka.gejak@linux.dev\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "137ac69c15fd2a9acf5d4d95456fe75f23c9ffc5", "tree": "5d2f46b48109b2779f70e68b55fbbf60d3325c9a", "parents": [ "3e09b370f830d6f00336390a5f622f7202dcd7cc" ], "author": { "name": "Luka Gejak", "email": "luka.gejak@linux.dev", "time": "Thu Mar 26 18:46:00 2026 +0100" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:37:51 2026 -0700" }, "message": "net: hsr: use __func__ instead of hardcoded function name\n\nReplace the hardcoded string \"hsr_get_untagged_frame\" with the\nstandard __func__ macro in netdev_warn_once() call to make the code\nmore robust to refactoring.\n\nSigned-off-by: Luka Gejak \u003cluka.gejak@linux.dev\u003e\nReviewed-by: Felix Maurer \u003cfmaurer@redhat.com\u003e\nLink: https://patch.msgid.link/20260326174600.136232-3-luka.gejak@linux.dev\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "3e09b370f830d6f00336390a5f622f7202dcd7cc", "tree": "9f9ccdbf291cf4421c777747afea3a7b8f7943ea", "parents": [ "686aad8353a5a4d3358deaba08b748a6b6e6d6c2" ], "author": { "name": "Luka Gejak", "email": "luka.gejak@linux.dev", "time": "Thu Mar 26 18:45:59 2026 +0100" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:37:50 2026 -0700" }, "message": "net: hsr: constify hsr_ops and prp_ops protocol operation structures\n\nThe hsr_ops and prp_ops structures are assigned to hsr-\u003eproto_ops during\ndevice initialization and are never modified at runtime. Declaring them\nas const allows the compiler to place these structures in read-only\nmemory, which improves security by preventing accidental or malicious\nmodification of the function pointers they contain.\n\nThe proto_ops field in struct hsr_priv is also updated to a const\npointer to maintain type consistency.\n\nSigned-off-by: Luka Gejak \u003cluka.gejak@linux.dev\u003e\nReviewed-by: Felix Maurer \u003cfmaurer@redhat.com\u003e\nLink: https://patch.msgid.link/20260326174600.136232-2-luka.gejak@linux.dev\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "686aad8353a5a4d3358deaba08b748a6b6e6d6c2", "tree": "37cccea77b9c35536c10e42cbc8fd8caf9982d44", "parents": [ "57ec1622b62ae8dcd88a0084c49c5b72b1850e68", "cd1082a96f9a1ffe053e48ac0a1b4fffb189c30e" ], "author": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:35:29 2026 -0700" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:35:30 2026 -0700" }, "message": "Merge branch \u0027macb-usrio-tsu-patches\u0027\n\nConor Dooley says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\nmacb usrio/tsu patches\n\nAt the very least, it\u0027d be good of the soc vendor folks could check\ntheir platforms and see if their usrio stuff actually lines up with what\nthe driver currently calls \"macb_default_usrio\". Ours didn\u0027t and it was\na nasty surprise.\n\nRyan and I figured out that the sama7g5 stuff is not actually using the\nsame usrio bits as earlier devices, so there\u0027s now more patches in this\nseries to split them apart. I\u0027ve not tested the split or the new\nproperty due to lack of hardware, but Ryan has.\n\nMarking this stuff net-next, because although they\u0027re fixes I don\u0027t see\nany particular urgency, and it avoids creating some dependencies between\ncleanup items and the fixes.\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nLink: https://patch.msgid.link/20260325-unsterile-flail-4c7729750dc4@spud\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "cd1082a96f9a1ffe053e48ac0a1b4fffb189c30e", "tree": "37cccea77b9c35536c10e42cbc8fd8caf9982d44", "parents": [ "32fc6a9f6e7585b2b39999a22635c06aa810d1bc" ], "author": { "name": "Théo Lebrun", "email": "theo.lebrun@bootlin.com", "time": "Wed Mar 25 16:28:18 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:34:29 2026 -0700" }, "message": "net: macb: drop usrio pointer on EyeQ5 config\n\nUSRIO is disabled on this platform, drop its inherited usrio config.\nWe will end up with MACB_CAPS_USRIO_DISABLED on this platform:\n - We have no config-\u003eusrio so macb_configure_caps() deduces that the\n feature is disabled.\n - Anecdotally, we would also land in the runtime detection codepath\n that reads DCFG1.\n\nSigned-off-by: Théo Lebrun \u003ctheo.lebrun@bootlin.com\u003e\nSigned-off-by: Conor Dooley \u003cconor.dooley@microchip.com\u003e\nLink: https://patch.msgid.link/20260325-stillness-undertake-d83054057b8d@spud\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "32fc6a9f6e7585b2b39999a22635c06aa810d1bc", "tree": "1ac6c0b7dacead18a7e678b33f1630bcbe0cb181", "parents": [ "47c86c463612edc3107eb7c0fd25b463f1756adf" ], "author": { "name": "Théo Lebrun", "email": "theo.lebrun@bootlin.com", "time": "Wed Mar 25 16:28:17 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:34:29 2026 -0700" }, "message": "net: macb: set MACB_CAPS_USRIO_DISABLED if no usrio config is provided\n\nbp-\u003eusrio is copied directly from dt_conf-\u003eusrio in macb_probe().\n\nIf dt_conf-\u003eusrio is NULL, we do not want to land in USRIO write\ncodepaths which dereference bp-\u003eusrio. Inherit automatically\nMACB_CAPS_USRIO_DISABLED to avoid those.\n\nThis means a macb_config that wants to disable usrio can simply drop\nits .usrio field, rather than add the disabled capability explicitly.\n\nNit: drop the dt_conf NULL check because the pointer is always valid.\n\nSigned-off-by: Théo Lebrun \u003ctheo.lebrun@bootlin.com\u003e\nSigned-off-by: Conor Dooley \u003cconor.dooley@microchip.com\u003e\nLink: https://patch.msgid.link/20260325-husband-cape-ec4945b9184c@spud\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "47c86c463612edc3107eb7c0fd25b463f1756adf", "tree": "f9a22824ea422caf09f6eb1de7901ec401648aa7", "parents": [ "41adda8764fd4d5e0f4e40cdcf47d60344dc4c4d" ], "author": { "name": "Théo Lebrun", "email": "theo.lebrun@bootlin.com", "time": "Wed Mar 25 16:28:16 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:34:29 2026 -0700" }, "message": "net: macb: runtime detect MACB_CAPS_USRIO_DISABLED\n\nDCFG1 (design config 1 register) carries a bit indicating whether User\nI/O feature has been enabled or not. The MACB/GEM driver has a cap flag\nindicating that HW has the feature disabled (default is enabled). Add\nthe missing connection between DCFG1 bit and MACB_CAPS_USRIO_DISABLED.\n\nIndirect impact: avoid useless writel() on USERIO register; this is not\nan important fix because USERIO is anyway read-only when feature is\ndisabled.\n\nIf for some reason a compatible sets USRIO_DISABLED but DCFG1 indicates\nit is enabled, we still keep the disabled capability flag. This ensures\nwe don\u0027t break \"cdns,np4-macb\" that sets the flag from compatible match\ndata.\n\nSigned-off-by: Théo Lebrun \u003ctheo.lebrun@bootlin.com\u003e\nSigned-off-by: Conor Dooley \u003cconor.dooley@microchip.com\u003e\nLink: https://patch.msgid.link/20260325-compactly-glue-f426a2e68904@spud\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "41adda8764fd4d5e0f4e40cdcf47d60344dc4c4d", "tree": "0fa9ffb05908618578830121d160674dacd55ef3", "parents": [ "09a6164a4f1d1b5364069bd290dba1d808718016" ], "author": { "name": "Conor Dooley", "email": "conor.dooley@microchip.com", "time": "Wed Mar 25 16:28:15 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:34:28 2026 -0700" }, "message": "net: macb: timer adjust mode is not supported\n\nThe ptp portion of this driver controls the tsu\u0027s timer using the\ncontrols for \"increment mode\", which is not compatible with the hardware\ntrying to control it via the gem_tsu_inc_ctrl and gem_tsu_ms inputs in\n\"timer adjust mode\". Abort probe if the property signalling that the\nrelevant signals have been wired up is present.\n\nSigned-off-by: Conor Dooley \u003cconor.dooley@microchip.com\u003e\nLink: https://patch.msgid.link/20260325-facebook-chop-cf792c53f1da@spud\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "09a6164a4f1d1b5364069bd290dba1d808718016", "tree": "42fa7d4b18d2814a5acd31c4c87414cb8e4f1d67", "parents": [ "b698a1e397ab9371911a130605908d26510e83c6" ], "author": { "name": "Conor Dooley", "email": "conor.dooley@microchip.com", "time": "Wed Mar 25 16:28:14 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:34:28 2026 -0700" }, "message": "dt-bindings: net: macb: add property indicating timer adjust mode\n\nThe GEM IP has two methods for modifying the ptp timer. The first of\nthese, named \"increment mode\", relies on software controlling the timer\nby setting tsu_timer_incr and tsu_timer_incr_sub_nsec and performing\nonce-off adjustments via the tsu_timer_adjust register. This is what the\nmacb driver uses. The second mechanism, \"timer adjust mode\" uses the\ngem_tsu_inc_ctrl and gem_tsu_ms signals to control the timer. These\nmodes are not intended to be used in parallel, but both can be possible\non the same device and which mode is used cannot be determined from the\ncompatible on all devices, because some users of the GEM IP are SoC\nFPGAs that permit configuring how the IP is wired up.\n\nAdd a property to indicate that gem_tsu_inc_ctrl and gem_tsu_ms are wired\nup for timer adjust mode.\n\nReviewed-by: Krzysztof Kozlowski \u003ckrzysztof.kozlowski@oss.qualcomm.com\u003e\nSigned-off-by: Conor Dooley \u003cconor.dooley@microchip.com\u003e\nLink: https://patch.msgid.link/20260325-daily-entitle-3640f7254da4@spud\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "b698a1e397ab9371911a130605908d26510e83c6", "tree": "149db380c52a84953f86a4eaed1f864ce18addb6", "parents": [ "3fe13d858f8391e38c75698719198d15a6a962dc" ], "author": { "name": "Conor Dooley", "email": "conor.dooley@microchip.com", "time": "Wed Mar 25 16:28:13 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:34:28 2026 -0700" }, "message": "net: macb: clean up tsu clk rate acquisition\n\ntsu_clk is grabbed during probe, so doesn\u0027t need to be re-grabbed here.\npclk is mandatory, probe will fail if it is err/NULL, so there\u0027s no need\nto check it here or have a !pclk 3rd arm. Simplify gem_get_tsu_rate() to\naccount for these facts.\n\nSigned-off-by: Conor Dooley \u003cconor.dooley@microchip.com\u003e\nLink: https://patch.msgid.link/20260325-hazing-penniless-14ba803efbb6@spud\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "3fe13d858f8391e38c75698719198d15a6a962dc", "tree": "2a2ad1844f7fe66ee326086e08f670b47facc8da", "parents": [ "c711311d6ba359ece75c77cca973f4c3b78e25b8" ], "author": { "name": "Conor Dooley", "email": "conor.dooley@microchip.com", "time": "Wed Mar 25 16:28:12 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:34:28 2026 -0700" }, "message": "net: macb: warn on pclk use as a tsu_clk fallback\n\nThe Candence GEM IP has a configuration parameter which determines the\nsource of the clock used for the timestamp unit (if it is enabled),\nswitching it between using the pclk and a dedicated input.\n\nWhen ptp support was added to the macb driver, a new tsu_clk was added\nto represent the dedicated input. While this is understandable, I think\nit is bug prone and that the tsu_clk should represent whatever clock is\nused for the timestamper and not just that specific input.\n\n\u003eFrom a driver point of view, the benefit of taking the conceptual\napproach is avoiding misconfiguring the driver when the hardware\nsupports ptp (and it is set as a capability in the relevant per-device\nstructure) but no tsu_clk is provided in devicetree. At the moment, the\ntimestamper will be registered and programmed with an increment that\nreflects the pclk in these cases, but will malfunction if the pclk and\ntsu_clk frequencies do not match. Obviously, this means the devicetree\nincorrectly represents the hardware, but this change in approach would\nmake the driver more resilient without meaningfully impacting correctly\ndescribed users.\n\nOut of the devices that claim MACB_CAPS_GEM_HAS_PTP the fu540, mpfs,\nsama5d2 and sama7g5-emac (but not sama7g5-gem) are at risk of having\nthis problem with the in-kernel devicetrees. mpfs and sama7g5-emac\nhave been confirmed to be incorrect, and sama5d2 is correct. It may be\nthat the other platforms actually do use the pclk for the timestamper\n(either by supplying pclk to the tsu_clk input of the IP, or by having\nthe IP block configured to use pclk instead of the tsu_clk input), but\nat least two are wrong, as they do not use pclk for the tsu_clk, so the\ndriver is registering the ptp clock incorrectly.\n\nAdd a warning if no tsu_clk is provided on a platform that uses the\ntimerstamper, to encourage people to specifically provide a tsu_clk and\navoid silently registering the timerstamper with the wrong clock. If the\npclk is actually used, it can be provided as a tsu_clk for improved\nclarity in devicetrees.\n\nWhile this changes the meaning of the devicetree property, it is\nbackwards compatible as there\u0027s no functional change for platforms that\ndidn\u0027t provide a tsu_clk and the changed meaning of providing a tsu_clk\nin the devicetree does not impact platforms that already provided one as\nthe decision about the tsu clock source is at IP instantiation time\nrather than at runtime, so there\u0027s no driver behaviour that needs to\nchange based on the input to the IP used for the timestamping unit.\n\nSigned-off-by: Conor Dooley \u003cconor.dooley@microchip.com\u003e\nLink: https://patch.msgid.link/20260325-dust-revision-368053e82d0e@spud\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "c711311d6ba359ece75c77cca973f4c3b78e25b8", "tree": "48373ae8725f76ae111b64938afbd15ede4234f2", "parents": [ "826cbe636e10581ea6a78541d756bce6049e5087" ], "author": { "name": "Conor Dooley", "email": "conor.dooley@microchip.com", "time": "Wed Mar 25 16:28:11 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:34:28 2026 -0700" }, "message": "net: macb: add mpfs specific usrio configuration\n\nOn mpfs the driver needs to make sure the tsu clock source is not the\nfabric, as this requires that the hardware is in Timer Adjust mode,\nwhich is not compatible with the linux driver trying to control the\nhardware. It is unlikely that this will be set, as the peripheral is\nreset during probe, but if the resets are not provided in devicetree\nit\u0027s probable that this bit is set incorrectly, as U-Boot\u0027s macb driver\nhas the same issue with using usrio settings for at91 platforms as the\ndefault.\n\nFixes: 8aad66aa59be5 (\"net: macb: add polarfire soc reset support\")\nSigned-off-by: Conor Dooley \u003cconor.dooley@microchip.com\u003e\nLink: https://patch.msgid.link/20260325-excavate-jester-798e7cfe02b5@spud\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "826cbe636e10581ea6a78541d756bce6049e5087", "tree": "5c8987d6924c2d3f13a897eba0eac94963ff392f", "parents": [ "6c5b565d7d41561709f3fe969e4eca1ba9e7f6f7" ], "author": { "name": "Conor Dooley", "email": "conor.dooley@microchip.com", "time": "Wed Mar 25 16:28:10 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:34:28 2026 -0700" }, "message": "net: macb: np4 doesn\u0027t need a usrio pointer\n\nUSRIO is disabled on this platform, having a pointer to a usrio config\nstructure doesn\u0027t actually do anything other than look weird.\n\nSigned-off-by: Conor Dooley \u003cconor.dooley@microchip.com\u003e\nLink: https://patch.msgid.link/20260325-passover-rimless-73c19c67d94b@spud\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "6c5b565d7d41561709f3fe969e4eca1ba9e7f6f7", "tree": "284a459c179ad915b4055714bdba3ce21ba56a32", "parents": [ "dfa36d7e860c2f008d92d91903f2d41ed0c88f3d" ], "author": { "name": "Conor Dooley", "email": "conor.dooley@microchip.com", "time": "Wed Mar 25 16:28:09 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:34:28 2026 -0700" }, "message": "net: macb: rework usrio refclk selection code\n\nThe USRIO based refclk selection code abuses a capability flag to set\nthe refclk to an external source based on match data/compatible on\nsama7g5-emac and use an internal source for the gmac.\n\nRyan previously added a property in an attempt to decouple the refclk\nsource from the compatible, because this is not fixed by compatible\nand there\u0027s variance based on the choices made by board designers.\n\nOriginally when Ryan added it, he removed the capability flag entirely\nfrom match data, but this changed the default for the sama7g5-emac and\nthe removal had to be reverted for these devices. Because these devices\ndefault to an external refclk, and the current property is only capable\nof communicating external refclks, there\u0027s no way to make the\nsama7g5-emac use an internal refclk.\n\nAdditionally, this property has no limiting based on compatible, and\nif used on a platform with an external refclk that is not controlled\nby USRIO the capability would be erroneously set. Because of the reuse\nof the at91_default_usrio struct by non-at91 devices, this could cause\nthe refclk bit to be set in error, on a system where the refclk is\nexternally provided without usrio settings being required.\n\nChange the new capability flag so that it actually represents the\nhardware being capable of controlling the refclk source via USRIO,\nand move the selection of default behaviour into the macb_usrio_config\nstruct provided as part of match data.\nModify the devicetree code to support a new property,\n\"cdns,refclk-source\" which will support devices with either default,\nretaining support for \"cdns,refclk-external\" for compatibility reasons.\n\nSigned-off-by: Conor Dooley \u003cconor.dooley@microchip.com\u003e\nLink: https://patch.msgid.link/20260325-tarantula-bullring-6ac44b39dd52@spud\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "dfa36d7e860c2f008d92d91903f2d41ed0c88f3d", "tree": "52a3f3af8797b058b6b6fb7159353516b578b62e", "parents": [ "039f185a00602fe4e760216ffb50dda6d1ed863a" ], "author": { "name": "Conor Dooley", "email": "conor.dooley@microchip.com", "time": "Wed Mar 25 16:28:08 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:34:27 2026 -0700" }, "message": "dt-bindings: net: cdns,macb: replace cdns,refclk-ext with cdns,refclk-source\n\nRyan added cdns,refclk-ext with the intent of decoupling the source of\nthe reference clock on sama7g5 (and related platforms) from the\ncompatible. Unfortunately, the default for sama7g5-emac is an external\nreference clock, so this property had no effect there, so that\ncompatibility with older devicetrees is preserved.\nReplace cdns,refclk-ext with one that supports both default states and\ntherefore is usable for sama7g5-emac.\n\nFor now, limit it to only the platforms that have USRIO controlled\nreference clock selection, but this could be generalised in the future.\n\nThe existing property only works on devices that are compatible with\nsama7g5-gem, so mark it deprecated, and limit its use to that specific\nscenario.\n\nSigned-off-by: Conor Dooley \u003cconor.dooley@microchip.com\u003e\nLink: https://patch.msgid.link/20260325-savior-untainted-03057ee0a917@spud\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "039f185a00602fe4e760216ffb50dda6d1ed863a", "tree": "bdb206c7fa93f8a0b6b4a109d17699f09d31b543", "parents": [ "a17871778ee28e4df054521e966e9f37c61f541b" ], "author": { "name": "Conor Dooley", "email": "conor.dooley@microchip.com", "time": "Wed Mar 25 16:28:07 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:34:27 2026 -0700" }, "message": "net: macb: split USRIO_HAS_CLKEN capability in two\n\nWhile trying to rework the internal/external refclk selection on\nsama7g5, Ryan and I noticed that the sama7g5 was \"overloading\" the\nmeaning of MACB_CAPS_USRIO_HAS_CLKEN, using it differently to how it was\noriginally intended.\nOriginally, on the macb hardware on sam9620 et al,\nMACB_CAPS_USRIO_HAS_CLKEN represented the hardware having a bit that\nneeded to be set to turn on the input clock to the transceivers. The\nsama7g5 doesn\u0027t have this bit, so for some reason the decision was made\nto reuse this capability flag to control selection of internal/external\nreferences.\n\nSplit the caps in two, so that capabilities do what they say on the tin,\nand allow reworking the refclk selection handling without impacting the\nolder devices that use MACB_CAPS_USRIO_CLKEN for its original purpose.\n\nSigned-off-by: Conor Dooley \u003cconor.dooley@microchip.com\u003e\nLink: https://patch.msgid.link/20260325-gradient-grading-b23b9e6ef9ff@spud\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "a17871778ee28e4df054521e966e9f37c61f541b", "tree": "f9074f7add5464c01180bff65ea816c405a99b0e", "parents": [ "8ccf062c6770618cb19f72ec0202463e01e2c7b8" ], "author": { "name": "Conor Dooley", "email": "conor.dooley@microchip.com", "time": "Wed Mar 25 16:28:06 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:34:27 2026 -0700" }, "message": "net: macb: rename macb_default_usrio to at91_default_usrio as not all platforms have mii mode control in usrio\n\nCalling this structure macb_default_usrio is misleading, I believe, as\nit implies that it should be used if your platform has nothing special\nto do in usrio. Since usrio is platform dependent, the default here is\nprobably for each usrio to do nothing, with the macb documentation I\nhave access to prescribing no standard behaviour here. We noticed that\nthis was problematic because on mpfs, a bit that macb_default_usrio\nsets to deal with the MII mode actually changes the source for the\ntsu_clk to something with how the majority of mpfs devices are actually\nconfigured!\n\nRename it to at91_default_usrio, since that\u0027s where the values actually\ncome from for these. I have no idea if any of the other platforms that\nuse the default actually copied at91\u0027s usrio configuration or if they\nhave usrio configurations where what the driver does has no impact.\n\nGate touching these bits behind a capability, like the clken refclock\nusrio knob, so that platforms without the MII mode stuff can avoid\nrunning this code.\n\nSigned-off-by: Conor Dooley \u003cconor.dooley@microchip.com\u003e\nLink: https://patch.msgid.link/20260325-landowner-preformed-2922ce736337@spud\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "8ccf062c6770618cb19f72ec0202463e01e2c7b8", "tree": "5064ac918f69731288414c7380eb67850a820dc8", "parents": [ "57ec1622b62ae8dcd88a0084c49c5b72b1850e68" ], "author": { "name": "Conor Dooley", "email": "conor.dooley@microchip.com", "time": "Wed Mar 25 16:28:05 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:34:27 2026 -0700" }, "message": "Revert \"net: macb: Clean up the .usrio settings in macb_config instances\"\n\nCommit 0ae998c4efd69 (\"net: macb: Clean up the .usrio settings in\nmacb_config instances\") was a misguided attempt to clean up the driver\nthat actually just propagated problematic code. The default for usrio is\nactually no usrio, and already there are issues with people using the\nproblematically named \"macb_default_usrio\" on platforms where the usrio\ndoes not have this so-called default behaviour. usrio is platform\nspecific and using the default at91 usrio settings should be opt-in\nonly. Revert the \"cleanup\" patch.\n\nSigned-off-by: Conor Dooley \u003cconor.dooley@microchip.com\u003e\nLink: https://patch.msgid.link/20260325-rundown-unrushed-dd82c9f5c56c@spud\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "57ec1622b62ae8dcd88a0084c49c5b72b1850e68", "tree": "09d9fcdf734ea6b9a42eb693b872e339d9f38ebf", "parents": [ "22e49419d4fab568c74e8e9b1aeb69c187d74d1e" ], "author": { "name": "Tomas Alvarez Vanoli", "email": "tomas.alvarez-vanoli@hitachienergy.com", "time": "Thu Mar 26 11:02:32 2026 +0100" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:21:35 2026 -0700" }, "message": "net: wan/fsl_ucc_hdlc: cleanup ucc_hdlc_poll\n\nImmediately after setting to 0 we are adding to it, and subtracting 0\nfrom budget. Replace with just assignment and no subtraction.\n\nSigned-off-by: Tomas Alvarez Vanoli \u003ctomas.alvarez-vanoli@hitachienergy.com\u003e\nReviewed-by: Christophe Leroy (CS GROUP) \u003cchleroy@kernel.org\u003e\nLink: https://patch.msgid.link/20260326100232.904289-1-tomas.alvarez-vanoli@hitachienergy.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "22e49419d4fab568c74e8e9b1aeb69c187d74d1e", "tree": "ea41c65c0f80df73d1b412e890dec902da86efb8", "parents": [ "d4383c7c78635ed96b89646eeb000d66022f06f4", "e3cdf6cf5fc6db0643723083e2c70fffe098e249" ], "author": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:09:10 2026 -0700" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:09:10 2026 -0700" }, "message": "Merge branch \u0027bnxt_en-add-xdp-rss-hash-metadata-support\u0027\n\nChris J Arges says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\nbnxt_en: Add XDP RSS hash metadata support\n\nThis series adds XDP RSS hash metadata extraction support for the bnxt_en\ndriver and includes selftests to validate the functionality. I was able\nto test this on a BCM57414 NIC.\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nLink: https://patch.msgid.link/20260325201139.2501937-1-carges@cloudflare.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "e3cdf6cf5fc6db0643723083e2c70fffe098e249", "tree": "ea41c65c0f80df73d1b412e890dec902da86efb8", "parents": [ "4ce06406958b67fdddcc2e6948237dd6ff6ba112" ], "author": { "name": "Chris J Arges", "email": "carges@cloudflare.com", "time": "Wed Mar 25 15:09:52 2026 -0500" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:09:08 2026 -0700" }, "message": "selftests: drv-net: xdp: Add rss_hash metadata tests\n\nThis test loads xdp_metadata.bpf which calls bpf_xdp_metadata_rx_hash() on\nincoming packets. The metadata from that packet is then sent to a BPF\nmap for validation. It borrows structure from xdp.py, reusing common\nfunctions.\n\nThe test checks the device\u0027s xdp-rx-metadata-features via netlink\nbefore running and skips on devices that do not advertise hash support.\nThis can be run on veth devices as well as real hardware.\n\nThe test is fairly simple and just verifies that a TCP or UDP packet can be\nidentified as an L4 flow. This minimal test also passes if run on a veth\ndevice.\n\nSigned-off-by: Chris J Arges \u003ccarges@cloudflare.com\u003e\nLink: https://patch.msgid.link/20260325201139.2501937-7-carges@cloudflare.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "4ce06406958b67fdddcc2e6948237dd6ff6ba112", "tree": "5537be3601f8393ca361af5db1c00d5b09c2a338", "parents": [ "5ec22d5f2afcf2f9d7d631dcc2f848dda64a4933" ], "author": { "name": "Chris J Arges", "email": "carges@cloudflare.com", "time": "Wed Mar 25 15:09:51 2026 -0500" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:09:08 2026 -0700" }, "message": "selftests: net: move common xdp.py functions into lib\n\nThis moves a few functions which can be useful to other python programs\nthat manipulate XDP programs. This also refactors xdp.py to use the\nrefactored functions.\n\nSigned-off-by: Chris J Arges \u003ccarges@cloudflare.com\u003e\nLink: https://patch.msgid.link/20260325201139.2501937-6-carges@cloudflare.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "5ec22d5f2afcf2f9d7d631dcc2f848dda64a4933", "tree": "760e67f168ad6d0a6532a7ff1810ef95bcbdba94", "parents": [ "4a037aeb6bc8e1fbf4d131d4d06544a9857ad604" ], "author": { "name": "Chris J Arges", "email": "carges@cloudflare.com", "time": "Wed Mar 25 15:09:50 2026 -0500" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:09:08 2026 -0700" }, "message": "bnxt_en: Implement XDP RSS hash metadata extraction for V3_CMP\n\nThis adds another conditional when cmp_type is CMP_TYPE_RX_L2_V3_CMP for\ndrivers that support this completion format.\n\nThis re-uses bnxt_rss_ext_op to provide similar functionality. One\nlimitation is for L4 hash-types, protocol-specific bits can\u0027t be\ndetermined.\n\nReviewed-by: Joe Damato \u003cjoe@dama.to\u003e\nSigned-off-by: Chris J Arges \u003ccarges@cloudflare.com\u003e\nLink: https://patch.msgid.link/20260325201139.2501937-5-carges@cloudflare.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "4a037aeb6bc8e1fbf4d131d4d06544a9857ad604", "tree": "395bf635dcd968cbd3cee1e24a3f622f604919da", "parents": [ "e5648b08cbfc9e7921138d9ace830b48824608f2" ], "author": { "name": "Chris J Arges", "email": "carges@cloudflare.com", "time": "Wed Mar 25 15:09:49 2026 -0500" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:09:07 2026 -0700" }, "message": "bnxt_en: Move bnxt_rss_ext_op into header\n\nThis allows bnxt_rss_ext_op to be used by other functions. In addition this\nmodifies the rxcmp argument to be const since the function only reads from\nthis structure.\n\nReviewed-by: Joe Damato \u003cjoe@dama.to\u003e\nSigned-off-by: Chris J Arges \u003ccarges@cloudflare.com\u003e\nLink: https://patch.msgid.link/20260325201139.2501937-4-carges@cloudflare.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "e5648b08cbfc9e7921138d9ace830b48824608f2", "tree": "4400900348ea9cedf5d3448cb1739de98f22b075", "parents": [ "542d3ec4508f02c9d5d9cdf92f7f3ae6958ca67c" ], "author": { "name": "Chris J Arges", "email": "carges@cloudflare.com", "time": "Wed Mar 25 15:09:48 2026 -0500" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:09:07 2026 -0700" }, "message": "bnxt_en: Implement XDP RSS hash metadata extraction\n\nAdd support for extracting RSS hash values and hash types from hardware\ncompletion descriptors in XDP programs for bnxt_en.\n\nAdd IP_TYPE definition for determining if completion is ipv4 or ipv6. In\naddition add ITYPE_ICMP flag for identifying ICMP completions.\n\nSigned-off-by: Chris J Arges \u003ccarges@cloudflare.com\u003e\nReviewed-by: Joe Damato \u003cjoe@dama.to\u003e\nReviewed-by: Andy Gospodarek \u003cgospo@broadcom.com\u003e\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "542d3ec4508f02c9d5d9cdf92f7f3ae6958ca67c", "tree": "e77bf8cb276748237cf90b29a02a68b0ed41e072", "parents": [ "d4383c7c78635ed96b89646eeb000d66022f06f4" ], "author": { "name": "Chris J Arges", "email": "carges@cloudflare.com", "time": "Wed Mar 25 15:09:47 2026 -0500" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 14:09:07 2026 -0700" }, "message": "bnxt_en: use bnxt_xdp_buff for xdp context\n\nThis adds bnxt_xdp_buff which embeds the xdp_buff struct and stores\npointers to hardware RX completion descriptors (rx_cmp and rx_cmp_ext)\nalong with the completion type.\n\nSigned-off-by: Chris J Arges \u003ccarges@cloudflare.com\u003e\nLink: https://patch.msgid.link/20260325201139.2501937-2-carges@cloudflare.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "d4383c7c78635ed96b89646eeb000d66022f06f4", "tree": "73f0cdf93727cd64851ee909b52c6040f4368bfd", "parents": [ "02d0e59e36e06fb728eb4dea8479f502c67b9fbc" ], "author": { "name": "Linus Walleij", "email": "linusw@kernel.org", "time": "Fri Mar 27 13:23:45 2026 +0100" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 11:49:21 2026 -0700" }, "message": "net: dsa: qca8k: Use the right GPIO header\n\nThe driver header for qca8k includes the legacy GPIO header\n\u003clinux/gpio.h\u003e but does not use any symbols from it and actually\nwants \u003clinux/gpio/consumer.h\u003e so fix this up.\n\nSigned-off-by: Linus Walleij \u003clinusw@kernel.org\u003e\nReviewed-by: Andrew Lunn \u003candrew@lunn.ch\u003e\nLink: https://patch.msgid.link/20260327-net-dsa-qca8k-v1-1-94e613a5c369@kernel.org\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "02d0e59e36e06fb728eb4dea8479f502c67b9fbc", "tree": "f0b79435b9274910a7231c398499a751dbf4b1bd", "parents": [ "30fcf28d83ee30494454b3f62f89a86e9e67bc25" ], "author": { "name": "Eric Dumazet", "email": "edumazet@google.com", "time": "Fri Mar 27 04:06:46 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 11:45:48 2026 -0700" }, "message": "tcp: use __jhash_final() in inet6_ehashfn()\n\nI misread jhash2() implementation.\n\nLast round should use __jhash_final() instead of __jhash_mix().\n\nUsing __jhash_mix() here leaves entropy distributed across a, b, and c,\nwhich might lead to incomplete diffusion of the faddr and fport bits\ninto the bucket index. Replacing this last __jhash_mix() with\n__jhash_final() provides the correct avalanche properties\nfor the returned value in c.\n\n$ scripts/bloat-o-meter -t vmlinux.0 vmlinux\nadd/remove: 0/0 grow/shrink: 0/1 up/down: 0/-4 (-4)\nFunction old new delta\ninet6_ehashfn 306 302 -4\nTotal: Before\u003d25155089, After\u003d25155085, chg -0.00%\n\nFixes: 854587e69ef3 (\"tcp: improve inet6_ehashfn() entropy\")\nSigned-off-by: Eric Dumazet \u003cedumazet@google.com\u003e\nReviewed-by: Kuniyuki Iwashima \u003ckuniyu@google.com\u003e\nLink: https://patch.msgid.link/20260327040646.3849503-1-edumazet@google.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "30fcf28d83ee30494454b3f62f89a86e9e67bc25", "tree": "1fa39fdd36dafb1320ebd5c6f5a337a79b769bea", "parents": [ "e531a081065d274a14f54441a38e1849453d06ec" ], "author": { "name": "谢致邦 (XIE Zhibang)", "email": "Yeking@Red54.com", "time": "Sat Mar 28 13:43:31 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 11:44:39 2026 -0700" }, "message": "net: stmmac: dwmac-rk: Fix typo in comment\n\nCorrect the typo \"rk3520\" to \"rk3528\" in comment.\n\nSigned-off-by: 谢致邦 (XIE Zhibang) \u003cYeking@Red54.com\u003e\nReviewed-by: Russell King (Oracle) \u003crmk+kernel@armlinux.org.uk\u003e\nLink: https://patch.msgid.link/tencent_833D2AD6577F21CF38ED1C3FE8814EB4B308@qq.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "e531a081065d274a14f54441a38e1849453d06ec", "tree": "20fb8c04d45f66a6fd293256787916a560bb2607", "parents": [ "ced629dc8e5c51ff2b5d847adeeb1035cd655d58", "b2c981e7c4653e3c276d5f3a0e012711d3596418" ], "author": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 11:21:29 2026 -0700" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 11:21:29 2026 -0700" }, "message": "Merge branch \u0027convert-config_ipv6-to-built-in-and-remove-stubs\u0027\n\nFernando Fernandez Mancera says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\nConvert CONFIG_IPV6 to built-in and remove stubs\n\nHistorically, the Linux kernel has supported compiling the IPv6 stack as\na loadable module. While this made sense in the early days of IPv6\nadoption, modern deployments and distributions overwhelmingly either\nbuild IPv6 directly into the kernel (CONFIG_IPV6\u003dy) or disable it\nentirely (CONFIG_IPV6\u003dn). The modular IPv6 use-case offers image size\nand memory savings for specific setups, this benefit is outweighed by\nthe architectural burden it imposes on the subsystems on implementation\nand maintenance.\n\nIn addition, most of the distributions are already using CONFIG_IPV6\u003dy\nby default [1], including openWRT [2] and Android gki_defconfig [3]. So\nthis won\u0027t have an impact on them. The most impacted architecture would\nprobably be arm64 as their default config is still using CONFIG_IPV6\u003dm.\n\nTo allow core networking, BPF, Netfilter, and various device drivers to\nsafely interact with a potentially unloaded IPv6 module, the kernel\nrelies on indirect call structures like ipv6_stub, ipv6_bpf_stub, and\nnf_ipv6_ops, along with dynamic RCU registrations for things like ICMPv6\nsenders.\n\nThis patch series addresses this by changing CONFIG_IPV6 from a tristate\nto a boolean, enforcing that IPv6 is either built-in or disabled. This\nallows us to completely rip out the stub infrastructures and safely\nreplace them with direct function calls.\n\nThe bloat-o-meter report the following results for m68k, arm64, x86_64\ndefconfig.\n\nm68k (keep on mind that CONFIG_IPV6 is disabled now):\nadd/remove: 65/938 grow/shrink: 36/254 up/down: 3022/-49692 (-46670)\n\narm64:\nadd/remove: 1251/265 grow/shrink: 81/46 up/down: 448740/-71519 (377221)\n\nx86_64:\nadd/remove: 62/98 grow/shrink: 10/39 up/down: 2497/-4357 (-1860)\n\nConsidering that each new kernel release increases sizes by 30-40KiB on\naverage, this size increase isn\u0027t a huge jump for the distributions that\nare still using CONFIG_IPV6\u003dm. For the ones that are already using\nCONFIG_IPV6\u003dy, the size is reduced actually.\n\nAll the patches has been independently build tested. With allmodconfig\nand allmodconfig + CONFIG_IPV6\u003dn. In addition, net selftest has been run\nagainst them on virtme-ng.\n\nThe series applied as a whole as been tested with allyesconfig and also\nallyesconfig + CONFIG_IPV6\u003dn but not all patches has been independently\ntested this way.\n\n[1] https://github.com/nyrahul/linux-kernel-configs\n\n[2] https://github.com/openwrt/openwrt/commit/832e7b817221d288df76b763ca12c585365db5d8\n\n[3] https://android.googlesource.com/kernel/common/+/refs/heads/android-mainline/arch/arm64/configs/gki_defconfig\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nLink: https://patch.msgid.link/20260325120928.15848-1-fmancera@suse.de\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "b2c981e7c4653e3c276d5f3a0e012711d3596418", "tree": "20fb8c04d45f66a6fd293256787916a560bb2607", "parents": [ "964870b4b9012bd42f29336f32b5c8be8ef3c00d" ], "author": { "name": "Fernando Fernandez Mancera", "email": "fmancera@suse.de", "time": "Wed Mar 25 13:08:52 2026 +0100" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 11:21:24 2026 -0700" }, "message": "netfilter: remove nf_ipv6_ops and use direct function calls\n\nAs IPv6 is built-in only, nf_ipv6_ops can be removed completely as it is\nnot longer necessary.\n\nConvert all nf_ipv6_ops usage to direct function calls instead. In\naddition, remove the ipv6_netfilter_init/fini() functions as they are\nnot necessary any longer.\n\nSigned-off-by: Fernando Fernandez Mancera \u003cfmancera@suse.de\u003e\nTested-by: Ricardo B. Marlière \u003crbm@suse.com\u003e\nLink: https://patch.msgid.link/20260325120928.15848-12-fmancera@suse.de\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "964870b4b9012bd42f29336f32b5c8be8ef3c00d", "tree": "36e3ada2886e9b81d897d83dcc61881f3be6324e", "parents": [ "ad84b1eefe28cee96c572b84bfa4f0fbfd425b68" ], "author": { "name": "Fernando Fernandez Mancera", "email": "fmancera@suse.de", "time": "Wed Mar 25 13:08:51 2026 +0100" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 11:21:24 2026 -0700" }, "message": "ipv6: remove ipv6_stub infrastructure completely\n\nAs IPv6 is built-in only and there are no more users of ipv6_stub, the\nipv6_stub is now entirely obsolete.\n\nRemove all the code related to the definition, initialization and usage.\n\nSigned-off-by: Fernando Fernandez Mancera \u003cfmancera@suse.de\u003e\nTested-by: Ricardo B. Marlière \u003crbm@suse.com\u003e\nLink: https://patch.msgid.link/20260325120928.15848-11-fmancera@suse.de\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "ad84b1eefe28cee96c572b84bfa4f0fbfd425b68", "tree": "10ee2d975e66b9a437d4ce3e6d481c8381a02779", "parents": [ "d76f6b170a10414a9c674b6db4bd0473b1b30050" ], "author": { "name": "Fernando Fernandez Mancera", "email": "fmancera@suse.de", "time": "Wed Mar 25 13:08:50 2026 +0100" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 11:21:24 2026 -0700" }, "message": "bpf: remove ipv6_bpf_stub completely and use direct function calls\n\nAs IPv6 is built-in only, the ipv6_bpf_stub can be removed completely.\n\nConvert all ipv6_bpf_stub usage to direct function calls instead. The\nfallback functions introduced previously will prevent linkage errors\nwhen CONFIG_IPV6 is disabled.\n\nSigned-off-by: Fernando Fernandez Mancera \u003cfmancera@suse.de\u003e\nTested-by: Ricardo B. Marlière \u003crbm@suse.com\u003e\nAcked-by: Daniel Borkmann \u003cdaniel@iogearbox.net\u003e\nReviewed-by: Martin KaFai Lau \u003cmartin.lau@kernel.org\u003e\nLink: https://patch.msgid.link/20260325120928.15848-10-fmancera@suse.de\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "d76f6b170a10414a9c674b6db4bd0473b1b30050", "tree": "456dc0e9966c93f419f993717dc3362b0715df4f", "parents": [ "d98adfbdd5c014f73d5f5e6b43735cc1b2d9450a" ], "author": { "name": "Fernando Fernandez Mancera", "email": "fmancera@suse.de", "time": "Wed Mar 25 13:08:49 2026 +0100" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 11:21:23 2026 -0700" }, "message": "net: convert remaining ipv6_stub users to direct function calls\n\nAs IPv6 is built-in only, the ipv6_stub infrastructure is no longer\nnecessary.\n\nConvert remaining ipv6_stub users to make direct function calls. The\nfallback functions introduced previously will prevent linkage errors\nwhen CONFIG_IPV6 is disabled.\n\nSigned-off-by: Fernando Fernandez Mancera \u003cfmancera@suse.de\u003e\nTested-by: Ricardo B. Marlière \u003crbm@suse.com\u003e\nLink: https://patch.msgid.link/20260325120928.15848-9-fmancera@suse.de\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "d98adfbdd5c014f73d5f5e6b43735cc1b2d9450a", "tree": "8758f221e4168850bdf82fca30010ab851ba7484", "parents": [ "29ae61b2fe7eeafc921a1481230aac4431520418" ], "author": { "name": "Fernando Fernandez Mancera", "email": "fmancera@suse.de", "time": "Wed Mar 25 13:08:48 2026 +0100" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 11:21:23 2026 -0700" }, "message": "ipv4: drop ipv6_stub usage and use direct function calls\n\nAs IPv6 is built-in only, the ipv6_stub infrastructure is no longer\nnecessary.\n\nThe IPv4 stack interacts with IPv6 mainly to support IPv4 routes with\nIPv6 next-hops (RFC 8950). Convert all these cross-family calls from\nipv6_stub to direct function calls. The fallback functions introduced\npreviously will prevent linkage errors when CONFIG_IPV6 is disabled.\n\nSigned-off-by: Fernando Fernandez Mancera \u003cfmancera@suse.de\u003e\nTested-by: Ricardo B. Marlière \u003crbm@suse.com\u003e\nLink: https://patch.msgid.link/20260325120928.15848-8-fmancera@suse.de\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "29ae61b2fe7eeafc921a1481230aac4431520418", "tree": "2d45d24955591d260b61da4dc3be69f430241221", "parents": [ "4b70b20215049472f200ed563a7a0d44a7188fd3" ], "author": { "name": "Fernando Fernandez Mancera", "email": "fmancera@suse.de", "time": "Wed Mar 25 13:08:47 2026 +0100" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 11:21:23 2026 -0700" }, "message": "drivers: net: drop ipv6_stub usage and use direct function calls\n\nAs IPv6 is built-in only, the ipv6_stub infrastructure is no longer\nnecessary.\n\nConvert all drivers currently utilizing ipv6_stub to make direct\nfunction calls. The fallback functions introduced previously will\nprevent linkage errors when CONFIG_IPV6 is disabled.\n\nSigned-off-by: Fernando Fernandez Mancera \u003cfmancera@suse.de\u003e\nTested-by: Ricardo B. Marlière \u003crbm@suse.com\u003e\nReviewed-by: Jason A. Donenfeld \u003cJason@zx2c4.com\u003e\nReviewed-by: Antonio Quartulli \u003cantonio@openvpn.net\u003e\nReviewed-by: Edward Cree \u003cecree.xilinx@gmail.com\u003e\nLink: https://patch.msgid.link/20260325120928.15848-7-fmancera@suse.de\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "4b70b20215049472f200ed563a7a0d44a7188fd3", "tree": "ee87f6bf7f8445eac346014ecb5ecac50aaee1d2", "parents": [ "d2042d35f413b7131cc571655bbcb2c049489fe7" ], "author": { "name": "Fernando Fernandez Mancera", "email": "fmancera@suse.de", "time": "Wed Mar 25 13:08:46 2026 +0100" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 11:21:23 2026 -0700" }, "message": "ipv6: prepare headers for ipv6_stub removal\n\nIn preparation for dropping ipv6_stub and converting its users to direct\nfunction calls, introduce static inline dummy functions and fallback\nmacros in the IPv6 networking headers. In addition, introduce checks on\nfib6_nh_init(), ip6_dst_lookup_flow() and ip6_fragment() to avoid a\ncrash due to ipv6.disable\u003d1 set during booting. The other functions are\nsafe as they cannot be called with ipv6.disable\u003d1 set.\n\nThese fallbacks ensure that when CONFIG_IPV6 is completely disabled,\nthere are no compiling or linking errors due to code paths not guarded\nby preprocessor macro IS_ENABLED(CONFIG_IPV6).\n\nIn addition, export ndisc_send_na(), ip6_route_input() and\nip6_fragment().\n\nSigned-off-by: Fernando Fernandez Mancera \u003cfmancera@suse.de\u003e\nTested-by: Ricardo B. Marlière \u003crbm@suse.com\u003e\nLink: https://patch.msgid.link/20260325120928.15848-6-fmancera@suse.de\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "d2042d35f413b7131cc571655bbcb2c049489fe7", "tree": "8b4730743ac246b7d60f3cdcf62a3090eec82b11", "parents": [ "fde39f7df10b3dc150abb87c4718efba93cbc755" ], "author": { "name": "Fernando Fernandez Mancera", "email": "fmancera@suse.de", "time": "Wed Mar 25 13:08:45 2026 +0100" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 11:21:23 2026 -0700" }, "message": "ipv6: remove dynamic ICMPv6 sender registration infrastructure\n\nAs IPv6 is built-in only, there is no need to maintain the sender\nregistration infrastructure used to allow built-in subsystems to send\nICMPv6 messages when IPv6 was compiled as a module.\n\nDrop the registration mechanism and the __icmpv6_send() sender\nimplementation. While icmpv6_send() users could be converted to\nicmp6_send() that doesn\u0027t seems necessary as none of them are using the\nforce_saddr parameter.\n\nSigned-off-by: Fernando Fernandez Mancera \u003cfmancera@suse.de\u003e\nTested-by: Ricardo B. Marlière \u003crbm@suse.com\u003e\nLink: https://patch.msgid.link/20260325120928.15848-5-fmancera@suse.de\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "fde39f7df10b3dc150abb87c4718efba93cbc755", "tree": "83c61242083946b5b7e2b60572e6df68493b2b53", "parents": [ "0557a34487b10f3e600a3a20b98ea164f36a3ad2" ], "author": { "name": "Fernando Fernandez Mancera", "email": "fmancera@suse.de", "time": "Wed Mar 25 13:08:44 2026 +0100" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 11:21:23 2026 -0700" }, "message": "ipv6: replace IS_BUILTIN(CONFIG_IPV6) with IS_ENABLED(CONFIG_IPV6)\n\nAs IPv6 is built-in only, it does not make sense to continue using\nIS_BUILTIN(CONFIG_IPV6). Therefore, replace it with IS_ENABLED() when\nnecessary and drop it if it isn\u0027t valid anymore.\n\nNotice that there is still one instance related to ICMPv6, as it\nrequires more changes it will be handle separately.\n\nSigned-off-by: Fernando Fernandez Mancera \u003cfmancera@suse.de\u003e\nTested-by: Ricardo B. Marlière \u003crbm@suse.com\u003e\nAcked-by: Martin KaFai Lau \u003cmartin.lau@kernel.org\u003e\nLink: https://patch.msgid.link/20260325120928.15848-4-fmancera@suse.de\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "0557a34487b10f3e600a3a20b98ea164f36a3ad2", "tree": "9d65ed04a6c14d6f6587bd93a1e7417b5f7d4d88", "parents": [ "309b905deee595619cc38719f48d63d57b8bff3d" ], "author": { "name": "Fernando Fernandez Mancera", "email": "fmancera@suse.de", "time": "Wed Mar 25 13:08:43 2026 +0100" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 11:21:22 2026 -0700" }, "message": "net: remove EXPORT_IPV6_MOD() and EXPORT_IPV6_MOD_GPL() macros\n\nAs IPv6 is built-in only, the macro is always evaluating to an empty\none. Remove it completely from the code.\n\nSigned-off-by: Fernando Fernandez Mancera \u003cfmancera@suse.de\u003e\nLink: https://patch.msgid.link/20260325120928.15848-3-fmancera@suse.de\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "309b905deee595619cc38719f48d63d57b8bff3d", "tree": "877158428ba63e826349eff2a21faa18682b2dc8", "parents": [ "ced629dc8e5c51ff2b5d847adeeb1035cd655d58" ], "author": { "name": "Fernando Fernandez Mancera", "email": "fmancera@suse.de", "time": "Wed Mar 25 13:08:42 2026 +0100" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Sun Mar 29 11:21:22 2026 -0700" }, "message": "ipv6: convert CONFIG_IPV6 to built-in only and clean up Kconfigs\n\nMaintaining a modular IPv6 stack offers image size savings for specific\nsetups, this benefit is outweighed by the architectural burden it\nimposes on the subsystems on implementation and maintenance. Therefore,\ndrop it.\n\nChange CONFIG_IPV6 from tristate to bool. Remove all Kconfig\ndependencies across the tree that explicitly checked for IPV6\u003dm. In\naddition, remove MODULE_DESCRIPTION(), MODULE_ALIAS(), MODULE_AUTHOR()\nand MODULE_LICENSE().\n\nThis is also replacing module_init() by device_initcall(). It is not\npossible to use fs_initcall() as IPv4 does because that creates a race\ncondition on IPv6 addrconf.\n\nFinally, modify the default configs from CONFIG_IPV6\u003dm to CONFIG_IPV6\u003dy\nexcept for m68k as according to the bloat-o-meter the image is\nincreasing by 330KB~ and that isn\u0027t acceptable. Instead, disable IPv6 on\nthis architecture by default. This is aligned with m68k RAM requirements\nand recommendations [1].\n\n[1] http://www.linux-m68k.org/faq/ram.html\n\nSigned-off-by: Fernando Fernandez Mancera \u003cfmancera@suse.de\u003e\nTested-by: Ricardo B. Marlière \u003crbm@suse.com\u003e\nAcked-by: Krzysztof Kozlowski \u003ckrzk@kernel.org\u003e # arm64\nLink: https://patch.msgid.link/20260325120928.15848-2-fmancera@suse.de\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "ced629dc8e5c51ff2b5d847adeeb1035cd655d58", "tree": "1ec252d9bc451cf100e188461ca4341b1baf509e", "parents": [ "77a347f82a8043c2ee77edf50cb585fb7b604325", "075196489a3797c2a931de68c438e06f8303dd5c" ], "author": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Fri Mar 27 21:07:26 2026 -0700" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Fri Mar 27 21:07:27 2026 -0700" }, "message": "Merge branch \u0027vrf-a-few-cleanups\u0027\n\nIdo Schimmel says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\nvrf: A few cleanups\n\nPerform a few cleanups in the VRF driver. Noticed these while reviewing\na recent patch [1]. See individual patches for more details.\n\n[1] https://lore.kernel.org/netdev/20260310105331.2371-1-lirongqing@baidu.com/\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nLink: https://patch.msgid.link/20260326203233.1128554-1-idosch@nvidia.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "075196489a3797c2a931de68c438e06f8303dd5c", "tree": "1ec252d9bc451cf100e188461ca4341b1baf509e", "parents": [ "50504e2579c1e0379bc0ffeaec67884e1d6c9212" ], "author": { "name": "Ido Schimmel", "email": "idosch@nvidia.com", "time": "Thu Mar 26 22:32:33 2026 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Fri Mar 27 21:07:24 2026 -0700" }, "message": "vrf: Remove unnecessary RCU protection around dst entries\n\nDuring initialization of a VRF device, the VRF driver creates two dst\nentries (for IPv4 and IPv6). They are attached to locally generated\npackets that are transmitted out of the VRF ports (via the\nl3mdev_l3_out() hook). Their purpose is to redirect packets towards the\nVRF device instead of having the packets egress directly out of the VRF\nports. This is useful, for example, when a queuing discipline is\nconfigured on the VRF device.\n\nIn order to avoid a NULL pointer dereference, commit b0e95ccdd775 (\"net:\nvrf: protect changes to private data with rcu\") made the pointers to the\ndst entries RCU protected. As far as I can tell, this was needed because\nback then the dst entries were released (and the pointers reset to NULL)\nbefore removing the VRF ports.\n\nLater on, commit f630c38ef0d7 (\"vrf: fix bug_on triggered by rx when\ndestroying a vrf\") moved the removal of the VRF ports to the VRF\ndevice\u0027s dellink() callback. As such, the tear down sequence of a VRF\ndevice looks as follows:\n\n1. VRF ports are removed.\n2. VRF device is unregistered.\n a. Device is closed.\n b. An RCU grace period passes.\n c. ndo_uninit() is called.\n i. dst entries are released.\n\nGiven the above, the Tx path will always see the same fully initialized\ndst entries and will never race with the ndo_uninit() callback.\n\nTherefore, there is no need to make the pointers to the dst entries RCU\nprotected. Remove it as well as the unnecessary NULL checks in the Tx\npath.\n\nSigned-off-by: Ido Schimmel \u003cidosch@nvidia.com\u003e\nReviewed-by: David Ahern \u003cdsahern@kernel.org\u003e\nLink: https://patch.msgid.link/20260326203233.1128554-4-idosch@nvidia.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "50504e2579c1e0379bc0ffeaec67884e1d6c9212", "tree": "77e17fb39f288a77a2e1eae89769949f1704a455", "parents": [ "ae3cdfd4e0b86e5d23b46b80e8b010f5392c9635" ], "author": { "name": "Ido Schimmel", "email": "idosch@nvidia.com", "time": "Thu Mar 26 22:32:32 2026 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Fri Mar 27 21:07:24 2026 -0700" }, "message": "vrf: Use dst_dev_put() instead of using loopback device\n\nUse dst_dev_put() to clean up the device referenced by the dst entry\ninstead of partially open coding it. Internally, the helper uses the\nblackhole device instead of the loopback device.\n\nReviewed-by: Petr Machata \u003cpetrm@nvidia.com\u003e\nReviewed-by: David Ahern \u003cdsahern@kernel.org\u003e\nSigned-off-by: Ido Schimmel \u003cidosch@nvidia.com\u003e\nLink: https://patch.msgid.link/20260326203233.1128554-3-idosch@nvidia.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "ae3cdfd4e0b86e5d23b46b80e8b010f5392c9635", "tree": "bc1f4a68fded6f40cf6468635896e4bc7a93458b", "parents": [ "77a347f82a8043c2ee77edf50cb585fb7b604325" ], "author": { "name": "Ido Schimmel", "email": "idosch@nvidia.com", "time": "Thu Mar 26 22:32:31 2026 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Fri Mar 27 21:07:24 2026 -0700" }, "message": "vrf: Remove unnecessary NULL check\n\nThe VRF driver always allocates an IPv4 dst entry for a VRF device and\nprevents the device from being registered if the allocation fails.\n\nTherefore, there is no need to check if the entry exists when tearing\ndown a VRF device. Remove the check.\n\nNote that the same is not true for the IPv6 dst entry. Its creation can\nbe skipped if IPv6 is administratively disabled (i.e.,\n\u0027ipv6.disable\u003d1\u0027).\n\nReviewed-by: Petr Machata \u003cpetrm@nvidia.com\u003e\nReviewed-by: David Ahern \u003cdsahern@kernel.org\u003e\nSigned-off-by: Ido Schimmel \u003cidosch@nvidia.com\u003e\nLink: https://patch.msgid.link/20260326203233.1128554-2-idosch@nvidia.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "77a347f82a8043c2ee77edf50cb585fb7b604325", "tree": "4289008f17f0a0b0899c47cb1203bc1a9653623d", "parents": [ "fb4b4a05aeeb8b0f253c5ddce21f4635dadc9550", "394863097e3603eafe819ab4085cbd0ddf371dd9" ], "author": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Fri Mar 27 20:57:40 2026 -0700" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Fri Mar 27 20:57:41 2026 -0700" }, "message": "Merge branch \u0027net-stmmac-disable-eee-on-i-mx\u0027\n\nLaurent Pinchart says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\nnet: stmmac: Disable EEE on i.MX\n\nThis small patch series fixes a long-standing interrupt storm issue with\nstmmac on NXP i.MX platforms.\n\nThe initial attempt to fix^Wwork around the problem in DT ([1]) was\npainfully but rightfully rejected by Russell, who helped me investigate\nthe issue in depth. It turned out that the root cause is a mistake in\nhow interrupts are wired in the SoC, a hardware bug that has been\nreplicated in all i.MX SoCs that integrate an stmmac. The only viable\nsolution is to disable EEE on those devices.\n\nIndividual patches explain the issue in more details. Patch 1/2,\nauthored by Russell, adds a new STMMAC_FLAG to disable EEE, and patch\n2/2 sets the flag for i.MX platforms.\n\n[1] https://lore.kernel.org/r/20251026122905.29028-1-laurent.pinchart@ideasonboard.com\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nLink: https://patch.msgid.link/20260325210003.2752013-1-laurent.pinchart@ideasonboard.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "394863097e3603eafe819ab4085cbd0ddf371dd9", "tree": "4289008f17f0a0b0899c47cb1203bc1a9653623d", "parents": [ "187b00a26679ae58a79f56c0024df1e3dbd7dff0" ], "author": { "name": "Laurent Pinchart", "email": "laurent.pinchart@ideasonboard.com", "time": "Wed Mar 25 23:00:03 2026 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Fri Mar 27 20:57:39 2026 -0700" }, "message": "net: stmmac: imx: Disable EEE\n\nThe i.MX8MP suffers from an interrupt storm related to the stmmac and\nEEE. A long and tedious analysis ([1]) concluded that the SoC wires the\nstmmac lpi_intr_o signal to an OR gate along with the main dwmac\ninterrupts, which causes an interrupt storm for two reasons.\n\nFirst, there\u0027s a race condition due to the interrupt deassertion being\nsynchronous to the RX clock domain:\n\n- When the PHY exits LPI mode, it restarts generating the RX clock\n (clk_rx_i input signal to the GMAC).\n- The MAC detects exit from LPI, and asserts lpi_intr_o. This triggers\n the ENET_EQOS interrupt.\n- Before the CPU has time to process the interrupt, the PHY enters LPI\n mode again, and stops generating the RX clock.\n- The CPU processes the interrupt and reads the GMAC4_LPI_CTRL_STATUS\n registers. This does not clear lpi_intr_o as there\u0027s no clk_rx_i.\n\nAn attempt was made to fixing the issue by not stopping RX_CLK in Rx LPI\nstate ([2]). This alleviates the symptoms but doesn\u0027t fix the issue.\nSince lpi_intr_o takes four RX_CLK cycles to clear, an interrupt storm\ncan still occur during that window. In 1000T mode this is harder to\nnotice, but slower receive clocks cause hundreds to thousands of\nspurious interrupts.\n\nFix the issue by disabling EEE completely on i.MX8MP.\n\n[1] https://lore.kernel.org/all/20251026122905.29028-1-laurent.pinchart@ideasonboard.com/\n[2] https://lore.kernel.org/all/20251123053518.8478-1-laurent.pinchart@ideasonboard.com/\n\nSigned-off-by: Laurent Pinchart \u003claurent.pinchart@ideasonboard.com\u003e\nReviewed-by: Russell King (Oracle) \u003crmk+kernel@armlinux.org.uk\u003e\nLink: https://patch.msgid.link/20260325210003.2752013-3-laurent.pinchart@ideasonboard.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "187b00a26679ae58a79f56c0024df1e3dbd7dff0", "tree": "d8e4bfa99b1b595fa046610b45782e5a9a66da7d", "parents": [ "fb4b4a05aeeb8b0f253c5ddce21f4635dadc9550" ], "author": { "name": "Russell King (Oracle)", "email": "rmk+kernel@armlinux.org.uk", "time": "Wed Mar 25 23:00:02 2026 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Fri Mar 27 20:57:38 2026 -0700" }, "message": "net: stmmac: provide flag to disable EEE\n\nSome platforms have problems when EEE is enabled, and thus need a way\nto disable stmmac EEE support. Add a flag before the other LPI related\nflags which tells stmmac to avoid populating the phylink LPI\ncapabilities, which causes phylink to call phy_disable_eee() for any\nPHY that is attached to the affected phylink instance.\n\niMX8MP is an example - the lpi_intr_o signal is wired to an OR gate\nalong with the main dwmac interrupts. Since lpi_intr_o is synchronous\nto the receive clock domain, and takes four clock cycles to clear, this\nleads to interrupt storms as the interrupt remains asserted for some\ntime after the LPI control and status register is read.\n\nThis problem becomes worse when the receive clock from the PHY stops\nwhen the receive path enters LPI state - which means that lpi_intr_o\ncan not deassert until the clock restarts. Since the LPI state of the\nreceive path depends on the link partner, this is out of our control.\nWe could disable RX clock stop at the PHY, but that doesn\u0027t get around\nthe slow-to-deassert lpi_intr_o mentioned in the above paragraph.\n\nPreviously, iMX8MP worked around this by disabling gigabit EEE, but\nthis is insufficient - the problem is also visible at 100M speeds,\nwhere the receive clock is slower.\n\nThere is extensive discussion and investigation in the thread linked\nbelow, the result of which is summarised in this commit message.\n\nReported-by: Laurent Pinchart \u003claurent.pinchart@ideasonboard.com\u003e\nCloses: https://lore.kernel.org/r/20251026122905.29028-1-laurent.pinchart@ideasonboard.com\nSigned-off-by: Russell King (Oracle) \u003crmk+kernel@armlinux.org.uk\u003e\nTested-by: Ovidiu Panait \u003covidiu.panait.rb@renesas.com\u003e\nSigned-off-by: Laurent Pinchart \u003claurent.pinchart@ideasonboard.com\u003e\nReviewed-by: Laurent Pinchart \u003claurent.pinchart@ideasonboard.com\u003e\nReviewed-by: Kieran Bingham \u003ckieran.bingham@ideasonboard.com\u003e\nLink: https://patch.msgid.link/20260325210003.2752013-2-laurent.pinchart@ideasonboard.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "fb4b4a05aeeb8b0f253c5ddce21f4635dadc9550", "tree": "d4b87ea3cdf52378b261a55ee5fe97f24f738bde", "parents": [ "b8f9410aefc15e55304788b958a796d22bcabb7e" ], "author": { "name": "Erni Sri Satya Vennela", "email": "ernis@linux.microsoft.com", "time": "Wed Mar 25 11:04:17 2026 -0700" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Fri Mar 27 20:45:02 2026 -0700" }, "message": "net: mana: Use at least SZ_4K in doorbell ID range check\n\nmana_gd_ring_doorbell() accesses offsets up to DOORBELL_OFFSET_EQ\n(0xFF8) + 8 bytes \u003d 4KB within each doorbell page. A db_page_size\nsmaller than SZ_4K is fundamentally incompatible with the driver:\ndoorbell pages would overlap and the device cannot function correctly.\n\nValidate db_page_size at the source and fail the\nprobe early if the value is below SZ_4K. This ensures the doorbell ID\nrange check in mana_gd_register_device() can rely on db_page_size\nbeing valid.\n\nFixes: 89fe91c65992 (\"net: mana: hardening: Validate doorbell ID from GDMA_REGISTER_DEVICE response\")\nSigned-off-by: Erni Sri Satya Vennela \u003cernis@linux.microsoft.com\u003e\nReviewed-by: Simon Horman \u003chorms@kernel.org\u003e\nLink: https://patch.msgid.link/20260325180423.1923060-1-ernis@linux.microsoft.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "b8f9410aefc15e55304788b958a796d22bcabb7e", "tree": "7583eee81df2a7a1bf22fb76004f1e1e58bca2fe", "parents": [ "d748047af1355df044faf8df0eedf0ef75f87de8" ], "author": { "name": "Jiri Pirko", "email": "jiri@nvidia.com", "time": "Wed Mar 25 16:28:01 2026 +0100" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Fri Mar 27 20:42:39 2026 -0700" }, "message": "mlx5: shd: Gracefully avoid shared devlink creation when no usable SN is found\n\nOn some HW, not even fall-back \"SERIALNO\" is found in VPD data. Unify\nthe behavior with the case there are no VPD data at all and avoid\ncreation of shared devlink instance.\n\nFixes: 2a8c8a03f306 (\"net/mlx5: Add a shared devlink instance for PFs on same chip\")\nReported-by: Adam Young \u003cadmiyo@amperemail.onmicrosoft.com\u003e\nCloses: https://lore.kernel.org/all/bab5b6bc-aa42-4af1-80d1-e56bcef06bc2@amperemail.onmicrosoft.com/\nReported-by: Ben Copeland \u003cben.copeland@linaro.org\u003e\nCloses: https://lore.kernel.org/all/20260324151014.860376-1-ben.copeland@linaro.org/\nSigned-off-by: Jiri Pirko \u003cjiri@nvidia.com\u003e\nTested-by: Ben Copeland \u003cben.copeland@linaro.org\u003e\nReviewed-by: Simon Horman \u003chorms@kernel.org\u003e\nLink: https://patch.msgid.link/20260325152801.236343-1-jiri@resnulli.us\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "d748047af1355df044faf8df0eedf0ef75f87de8", "tree": "6763a0860407de28ad74936929cb25e528a4c0ab", "parents": [ "e4cf6087cab382c7031e6b436ec55202fa9f2d7b" ], "author": { "name": "Michael S. Tsirkin", "email": "mst@redhat.com", "time": "Wed Mar 25 13:36:00 2026 -0400" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Fri Mar 27 17:14:31 2026 -0700" }, "message": "ptr_ring: disable KCSAN warnings\n\nEric Dumazet reported KCSAN warnings:\n\nBUG: KCSAN: data-race in pfifo_fast_dequeue / pfifo_fast_enqueue\n\nwrite to 0xffff88811d5ccc00 of 8 bytes by interrupt on cpu 0:\n__ptr_ring_zero_tail include/linux/ptr_ring.h:259 [inline]\n__ptr_ring_discard_one include/linux/ptr_ring.h:291 [inline]\n__ptr_ring_consume include/linux/ptr_ring.h:311 [inline]\n__skb_array_consume include/linux/skb_array.h:98 [inline]\npfifo_fast_dequeue+0x770/0x8f0 net/sched/sch_generic.c:770\ndequeue_skb net/sched/sch_generic.c:297 [inline]\nqdisc_restart net/sched/sch_generic.c:402 [inline]\n__qdisc_run+0x189/0xc80 net/sched/sch_generic.c:420\nqdisc_run include/net/pkt_sched.h:120 [inline]\nnet_tx_action+0x379/0x590 net/core/dev.c:5793\nhandle_softirqs+0xb9/0x280 kernel/softirq.c:622\ndo_softirq+0x45/0x60 kernel/softirq.c:523\n__local_bh_enable_ip+0x70/0x80 kernel/softirq.c:450\nlocal_bh_enable include/linux/bottom_half.h:33 [inline]\nbpf_test_run+0x2db/0x620 net/bpf/test_run.c:426\nbpf_prog_test_run_skb+0x9a4/0xef0 net/bpf/test_run.c:1159\nbpf_prog_test_run+0x204/0x340 kernel/bpf/syscall.c:4721\n__sys_bpf+0x52e/0x7e0 kernel/bpf/syscall.c:6246\n__do_sys_bpf kernel/bpf/syscall.c:6341 [inline]\n__se_sys_bpf kernel/bpf/syscall.c:6339 [inline]\n__x64_sys_bpf+0x41/0x50 kernel/bpf/syscall.c:6339\nx64_sys_call+0x10cb/0x3020 arch/x86/include/generated/asm/syscalls_64.h:322\ndo_syscall_x64 arch/x86/entry/syscall_64.c:63 [inline]\ndo_syscall_64+0x12c/0x370 arch/x86/entry/syscall_64.c:94\nentry_SYSCALL_64_after_hwframe+0x77/0x7f\n\nread to 0xffff88811d5ccc00 of 8 bytes by task 22632 on cpu 1:\n__ptr_ring_produce include/linux/ptr_ring.h:106 [inline]\nptr_ring_produce include/linux/ptr_ring.h:129 [inline]\nskb_array_produce include/linux/skb_array.h:44 [inline]\npfifo_fast_enqueue+0xd5/0x2c0 net/sched/sch_generic.c:741\ndev_qdisc_enqueue net/core/dev.c:4144 [inline]\n__dev_xmit_skb net/core/dev.c:4188 [inline]\n__dev_queue_xmit+0x6a4/0x1f20 net/core/dev.c:4795\ndev_queue_xmit include/linux/netdevice.h:3384 [inline]\n__bpf_tx_skb net/core/filter.c:2153 [inline]\n__bpf_redirect_common net/core/filter.c:2197 [inline]\n__bpf_redirect+0x862/0x990 net/core/filter.c:2204\n____bpf_clone_redirect net/core/filter.c:2487 [inline]\nbpf_clone_redirect+0x20c/0x290 net/core/filter.c:2450\nbpf_prog_53f18857bc887b09+0x22/0x2a\nbpf_dispatcher_nop_func include/linux/bpf.h:1402 [inline]\n__bpf_prog_run include/linux/filter.h:723 [inline]\nbpf_prog_run include/linux/filter.h:730 [inline]\nbpf_test_run+0x29d/0x620 net/bpf/test_run.c:423\nbpf_prog_test_run_skb+0x9a4/0xef0 net/bpf/test_run.c:1159\nbpf_prog_test_run+0x204/0x340 kernel/bpf/syscall.c:4721\n__sys_bpf+0x52e/0x7e0 kernel/bpf/syscall.c:6246\n__do_sys_bpf kernel/bpf/syscall.c:6341 [inline]\n__se_sys_bpf kernel/bpf/syscall.c:6339 [inline]\n__x64_sys_bpf+0x41/0x50 kernel/bpf/syscall.c:6339\nx64_sys_call+0x10cb/0x3020 arch/x86/include/generated/asm/syscalls_64.h:322\ndo_syscall_x64 arch/x86/entry/syscall_64.c:63 [inline]\ndo_syscall_64+0x12c/0x370 arch/x86/entry/syscall_64.c:94\nentry_SYSCALL_64_after_hwframe+0x77/0x7f\n\nvalue changed: 0xffff888104a93a00 -\u003e 0x0000000000000000\n\nReported by Kernel Concurrency Sanitizer on:\nCPU: 1 UID: 0 PID: 22632 Comm: syz.0.4135 Tainted: G W syzkaller #0\nPREEMPT(full)\nTainted: [W]\u003dWARN\nHardware name: Google Google Compute Engine/Google Compute Engine,\nBIOS Google 01/24/2026\n\nThere is no race on ring accesses: reading/writing a partial pointer\nwould be fine, because the reading is done by the producer\nwhich merely cares about NULL/non NULL.\nDocument and disable the warnings using data_race().\n\nSigned-off-by: Michael S. Tsirkin \u003cmst@redhat.com\u003e\nAcked-by: Jason Wang \u003cjasowang@redhat.com\u003e\nReviewed-by: Eric Dumazet \u003cedumazet@google.com\u003e\nLink: https://patch.msgid.link/dd3984b3bce9df3591927f927668cb31cc7ecf34.1774460059.git.mst@redhat.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "e4cf6087cab382c7031e6b436ec55202fa9f2d7b", "tree": "25596267090817b922c0b05c3f3cd11250d80fbb", "parents": [ "552994294fe27b42a6a735b0388029b45d776b38" ], "author": { "name": "Alexander Wilhelm", "email": "alexander.wilhelm@westermo.com", "time": "Thu Mar 26 08:17:52 2026 +0100" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Fri Mar 27 12:10:15 2026 +0000" }, "message": "net: qrtr: fix endian handling of confirm_rx field\n\nConvert confirm_rx to little endian when enqueueing and convert it back on\nreceive. This fixes control flow on big endian hosts, little endian is\nunaffected.\n\nOn transmit, store confirm_rx as __le32 using cpu_to_le32(). On receive,\napply le32_to_cpu() before using the value. !! ensures the value is 0 or 1\nin native endianness, so the conversion isn’t strictly required here, but\nit is kept for consistency and clarity.\n\nReviewed-by: Manivannan Sadhasivam \u003cmani@kernel.org\u003e\nSigned-off-by: Alexander Wilhelm \u003calexander.wilhelm@westermo.com\u003e\nReviewed-by: Simon Horman \u003chorms@kernel.org\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "552994294fe27b42a6a735b0388029b45d776b38", "tree": "93f42c8aa1c84ba9819c5c103d6dda4093acbb98", "parents": [ "5fffa6fe40f07e831dc387170405b7f3486d14d0" ], "author": { "name": "Jiayuan Chen", "email": "jiayuan.chen@linux.dev", "time": "Wed Mar 25 15:18:54 2026 +0800" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Thu Mar 26 20:44:45 2026 -0700" }, "message": "tcp: Fix inconsistent indenting warning\n\nSuppress such warning reported by test robot:\n include/net/tcp.h:1449 tcp_ca_event() warn: inconsistent indenting\n\nReported-by: kernel test robot \u003clkp@intel.com\u003e\nCloses: https://lore.kernel.org/oe-kbuild-all/202603251430.gQ3VuiKV-lkp@intel.com/\nSigned-off-by: Jiayuan Chen \u003cjiayuan.chen@linux.dev\u003e\nReviewed-by: Eric Dumazet \u003cedumazet@google.com\u003e\nLink: https://patch.msgid.link/20260325071854.805-1-jiayuan.chen@linux.dev\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "5fffa6fe40f07e831dc387170405b7f3486d14d0", "tree": "b1af7cfcb1aba4d286f5f76490da7b45495eeb34", "parents": [ "f8844dfeeae88db074146e3f8a865b5565f61b40", "af0331e1ac51f48d0cc6bb09547b8c8ae467a019" ], "author": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Thu Mar 26 20:39:47 2026 -0700" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Thu Mar 26 20:39:47 2026 -0700" }, "message": "Merge branch \u0027net-stmmac-remove-unused-and-unimplemented-axi-properties\u0027\n\nRussell King says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\nnet: stmmac: remove unused and unimplemented AXI properties\n\ncommit afea03656add (\"stmmac: rework DMA bus setting and introduce new\nplatform AXI structure\") added support for parsing all the stmmac AXI\nattributes, and added code to set most of the appropriate register bits\nwith three exceptions:\n\n\tsnps,kbbe\n\tsnps,mb\n\tsnps,rb\n\nThese were parsed by the driver, but the result of parsing was never\nused by any of the cores.\n\nMoreover, no DTS in the kernel makes use of these properties.\n\nThus, it doesn\u0027t make sense for the driver to parse these, so let\u0027s\nremove them. Also remove them from the DT binding document.\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nLink: https://patch.msgid.link/acJh4z3pRKkeaFbR@shell.armlinux.org.uk\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "af0331e1ac51f48d0cc6bb09547b8c8ae467a019", "tree": "b1af7cfcb1aba4d286f5f76490da7b45495eeb34", "parents": [ "a800398e746f8c9010c626a71d92a05b708f7622" ], "author": { "name": "Russell King (Oracle)", "email": "rmk+kernel@armlinux.org.uk", "time": "Tue Mar 24 10:05:45 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Thu Mar 26 20:39:43 2026 -0700" }, "message": "dt-bindings: remove unimplemented AXI snps,kbbe snps,mb and snps,rb\n\nRemove the AXI snps,kbbe snps,mb and snps,rb properties as they have\nnot been used, and although the driver parses these, the code hasn\u0027t\never used the parsed result. This parsing has now been removed.\n\nThese were introduced by commit afea03656add (\"stmmac: rework DMA bus\nsetting and introduce new platform AXI structure\").\n\nSigned-off-by: Russell King (Oracle) \u003crmk+kernel@armlinux.org.uk\u003e\nAcked-by: Conor Dooley \u003cconor.dooley@microchip.com\u003e\nLink: https://patch.msgid.link/E1w4ydt-0000000Dlph-3WvI@rmk-PC.armlinux.org.uk\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "a800398e746f8c9010c626a71d92a05b708f7622", "tree": "95fefbbe957a18d091e99857701c16788462578a", "parents": [ "f8844dfeeae88db074146e3f8a865b5565f61b40" ], "author": { "name": "Russell King (Oracle)", "email": "rmk+kernel@armlinux.org.uk", "time": "Tue Mar 24 10:05:40 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Thu Mar 26 20:39:43 2026 -0700" }, "message": "net: stmmac: remove axi_kbbe, axi_mb and axi_rb members\n\naxi_kbbe, axi_mb and axi_rb are all written, but nothing ever reads\ntheir values. Remove the code that sets these and the struct members.\n\nSigned-off-by: Russell King (Oracle) \u003crmk+kernel@armlinux.org.uk\u003e\nLink: https://patch.msgid.link/E1w4ydo-0000000Dlpb-34jd@rmk-PC.armlinux.org.uk\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "f8844dfeeae88db074146e3f8a865b5565f61b40", "tree": "9a3b1d9d85b8db0440f2164919ded41eed4be65a", "parents": [ "6a539eee855cbfe9c32507c70003b7710604fcfb" ], "author": { "name": "Di Zhu", "email": "zhud@hygon.cn", "time": "Mon Mar 23 12:17:30 2026 +0800" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Thu Mar 26 20:28:49 2026 -0700" }, "message": "virtio-net: enable NETIF_F_GRO_HW only if GRO-related offloads are supported\n\nNegotiating VIRTIO_NET_F_CTRL_GUEST_OFFLOADS indicates the device\nallows control over offload support, but the offloads that can be\ncontrolled may have nothing to do with GRO (e.g., if neither GUEST_TSO4\nnor GUEST_TSO6 is supported).\n\nIn such a setup, reporting NETIF_F_GRO_HW as available for the device\nis too optimistic and misleading to the user.\n\nImprove the situation by masking off NETIF_F_GRO_HW unless the device\npossesses actual GRO-related offload capabilities. Out of an abundance\nof caution, this does not change the current behaviour for hardware with\njust v6 or just v4 GRO: current interfaces do not allow distinguishing\nbetween v6/v4 GRO, so we can\u0027t expose them to userspace precisely.\n\nSigned-off-by: Di Zhu \u003czhud@hygon.cn\u003e\nAcked-by: Jason Wang \u003cjasowang@redhat.com\u003e\nLink: https://patch.msgid.link/20260323041730.986351-1-zhud@hygon.cn\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "6a539eee855cbfe9c32507c70003b7710604fcfb", "tree": "8b7b60d7e8a60a11d5798955773429853cd92145", "parents": [ "fe3e54253f0b04ec9e85d46e10aadbdbb31d29b2" ], "author": { "name": "Eric Dumazet", "email": "edumazet@google.com", "time": "Wed Mar 25 21:24:39 2026 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Thu Mar 26 20:11:53 2026 -0700" }, "message": "tcp: tcp_vegas: use tcp_vegas_cwnd_event_tx_start()\n\nWhile net/ipv4/tcp_yeah.c is correctly setting .cwnd_event_tx_start\nto tcp_vegas_cwnd_event_tx_start(), I forgot to do the same in tcp_vegas.c\n\nFixes: d1e59a469737 (\"tcp: add cwnd_event_tx_start to tcp_congestion_ops\")\nSigned-off-by: Eric Dumazet \u003cedumazet@google.com\u003e\nReviewed-by: Kuniyuki Iwashima \u003ckuniyu@google.com\u003e\nLink: https://patch.msgid.link/20260325212440.4146579-1-edumazet@google.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" } ], "next": "fe3e54253f0b04ec9e85d46e10aadbdbb31d29b2" }