)]}' { "log": [ { "commit": "c519cf9b7434183bb56ed1e200ac577a5fd34d9b", "tree": "1ff5878aa9a887a0de6070c68133a9542c062b8b", "parents": [ "6e15774d92056f7e4d226b2fc758124a6c803f99" ], "author": { "name": "Thorsten Blum", "email": "thorsten.blum@toblux.com", "time": "Mon May 27 12:36:19 2024 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon May 27 17:15:22 2024 -0700" }, "message": "docs: netdev: Fix typo in Signed-off-by tag\n\ns/of/off/\n\nSigned-off-by: Thorsten Blum \u003cthorsten.blum@toblux.com\u003e\nFixes: e110ba659271 (\"docs: netdev: add note about Changes Requested and revising commit messages\")\nLink: https://lore.kernel.org/r/20240527103618.265801-2-thorsten.blum@toblux.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "6e15774d92056f7e4d226b2fc758124a6c803f99", "tree": "abaac1bd13f422a913f394bdfe3489a5b5386ddc", "parents": [ "7a8cc96ebefaab3db0477a54b217214ad70990d1", "38af56e6668b455f7dd0a8e2d9afe74100068e17" ], "author": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon May 27 17:13:01 2024 -0700" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon May 27 17:13:01 2024 -0700" }, "message": "Merge branch \u0027selftests-mptcp-mark-unstable-subtests-as-flaky\u0027\n\nMatthieu Baerts says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\nselftests: mptcp: mark unstable subtests as flaky\n\nSome subtests can be unstable, failing once every X runs. Fixing them\ncan take time: there could be an issue in the kernel or in the subtest,\nand it is then important to do a proper analysis, not to hide real bugs.\n\nTo avoid creating noises on the different CIs where tests are more\nunstable than on our side, some subtests have been marked as flaky. As a\nresult, errors with these subtests (if any) are ignored.\n\nNote that the MPTCP CI will continue to track these flaky subtests. All\nthese unstable subtests are also tracked by our bug tracker.\n\nThese are fixes for the -net tree, because the instabilities are visible\nthere. The first patch introducing the flake support has no \u0027Fixes\u0027\ntags, mainly because it requires recent and important refactoring done\nin all MPTCP selftests. Backporting that to old versions where the flaky\ntests have been introduced would be too difficult, and probably not\nworth it. The other patches, adding MPTCP_LIB_SUBTEST_FLAKY\u003d1, have a\nFixes tag, simply to ease the backport of the future fixes removing them\nalong with the proper fix.\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nLink: https://lore.kernel.org/r/20240524-upstream-net-20240524-selftests-mptcp-flaky-v1-0-a352362f3f8e@kernel.org\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "38af56e6668b455f7dd0a8e2d9afe74100068e17", "tree": "abaac1bd13f422a913f394bdfe3489a5b5386ddc", "parents": [ "8c06ac2178a9dee887929232226e35a5cdda1793" ], "author": { "name": "Matthieu Baerts (NGI0)", "email": "matttbe@kernel.org", "time": "Fri May 24 18:30:59 2024 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon May 27 17:12:51 2024 -0700" }, "message": "selftests: mptcp: join: mark \u0027fail\u0027 tests as flaky\n\nThese tests are rarely unstable. It depends on the CI running the tests,\nespecially if it is also busy doing other tasks in parallel, and if a\ndebug kernel config is being used.\n\nIt looks like this issue is sometimes present with the NetDev CI. While\nthis is being investigated, the tests are marked as flaky not to create\nnoises on such CIs.\n\nFixes: b6e074e171bc (\"selftests: mptcp: add infinite map testcase\")\nLink: https://github.com/multipath-tcp/mptcp_net-next/issues/491\nReviewed-by: Mat Martineau \u003cmartineau@kernel.org\u003e\nSigned-off-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\nLink: https://lore.kernel.org/r/20240524-upstream-net-20240524-selftests-mptcp-flaky-v1-4-a352362f3f8e@kernel.org\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "8c06ac2178a9dee887929232226e35a5cdda1793", "tree": "4b50080cb9bb52e5138874d5da4e9892cd1662ed", "parents": [ "cc73a6577ae64247898269d138dee6b73ff710cc" ], "author": { "name": "Matthieu Baerts (NGI0)", "email": "matttbe@kernel.org", "time": "Fri May 24 18:30:58 2024 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon May 27 17:12:50 2024 -0700" }, "message": "selftests: mptcp: join: mark \u0027fastclose\u0027 tests as flaky\n\nThese tests are flaky since their introduction. This might be less or\nnot visible depending on the CI running the tests, especially if it is\nalso busy doing other tasks in parallel, and if a debug kernel config is\nbeing used.\n\nIt looks like this issue is often present with the NetDev CI. While this\nis being investigated, the tests are marked as flaky not to create\nnoises on such CIs.\n\nFixes: 01542c9bf9ab (\"selftests: mptcp: add fastclose testcase\")\nLink: https://github.com/multipath-tcp/mptcp_net-next/issues/324\nReviewed-by: Mat Martineau \u003cmartineau@kernel.org\u003e\nSigned-off-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\nLink: https://lore.kernel.org/r/20240524-upstream-net-20240524-selftests-mptcp-flaky-v1-3-a352362f3f8e@kernel.org\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "cc73a6577ae64247898269d138dee6b73ff710cc", "tree": "b2014d1f3f20ab8b7edc631ab56f830f9aa17c9a", "parents": [ "5597613fb3cf0e36d26cfd8fb2a63196da249333" ], "author": { "name": "Matthieu Baerts (NGI0)", "email": "matttbe@kernel.org", "time": "Fri May 24 18:30:57 2024 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon May 27 17:12:50 2024 -0700" }, "message": "selftests: mptcp: simult flows: mark \u0027unbalanced\u0027 tests as flaky\n\nThese tests are flaky since their introduction. This might be less or\nnot visible depending on the CI running the tests, especially if it is\nalso busy doing other tasks in parallel.\n\nA first analysis shown that the transfer can be slowed down when there\nare some re-injections at the MPTCP level. Such re-injections can of\ncourse happen, and disturb the transfer, but it looks strange to have\nthem in this lab. That could be caused by the kernel having access to\nless CPU cycles -- e.g. when other activities are executed in parallel\n-- or by a misinterpretation on the MPTCP packet scheduler side.\n\nWhile this is being investigated, the tests are marked as flaky not to\ncreate noises in other CIs.\n\nFixes: 219d04992b68 (\"mptcp: push pending frames when subflow has free space\")\nLink: https://github.com/multipath-tcp/mptcp_net-next/issues/475\nReviewed-by: Mat Martineau \u003cmartineau@kernel.org\u003e\nSigned-off-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\nLink: https://lore.kernel.org/r/20240524-upstream-net-20240524-selftests-mptcp-flaky-v1-2-a352362f3f8e@kernel.org\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "5597613fb3cf0e36d26cfd8fb2a63196da249333", "tree": "43461ced70c74caddafcefe6fe6a69b685fdb672", "parents": [ "7a8cc96ebefaab3db0477a54b217214ad70990d1" ], "author": { "name": "Matthieu Baerts (NGI0)", "email": "matttbe@kernel.org", "time": "Fri May 24 18:30:56 2024 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon May 27 17:12:50 2024 -0700" }, "message": "selftests: mptcp: lib: support flaky subtests\n\nSome subtests can be unstable, failing once every X runs. Fixing them\ncan take time: there could be an issue in the kernel or in the subtest,\nand it is then important to do a proper analysis, not to hide real bugs.\n\nTo avoid creating noises on the different CIs, it is important to have a\nsimple way to mark subtests as flaky, and ignore the errors. This is\nwhat this patch introduces: subtests can be marked as flaky by setting\nMPTCP_LIB_SUBTEST_FLAKY env var to 1, e.g.\n\n MPTCP_LIB_SUBTEST_FLAKY\u003d1 \u003crun flaky subtest\u003e\n\nThe subtest will be executed, and errors (if any) will be ignored. It is\nstill good to run these subtests, as it exercises code, and the results\ncan still be useful for the on-going investigations.\n\nNote that the MPTCP CI will continue to track these flaky subtests by\nsetting SELFTESTS_MPTCP_LIB_OVERRIDE_FLAKY env var to 1, and a ticket\nhas to be created before marking subtests as flaky.\n\nReviewed-by: Mat Martineau \u003cmartineau@kernel.org\u003e\nSigned-off-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\nLink: https://lore.kernel.org/r/20240524-upstream-net-20240524-selftests-mptcp-flaky-v1-1-a352362f3f8e@kernel.org\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "7a8cc96ebefaab3db0477a54b217214ad70990d1", "tree": "7194c1c7003853d949030f70033adb0016cd3dfd", "parents": [ "266aa3b4812e97942a8ce5c7aafa7da059f7b5b8", "82617b9a04649e83ee8731918aeadbb6e6d7cbc7" ], "author": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon May 27 17:11:45 2024 -0700" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon May 27 17:11:46 2024 -0700" }, "message": "Merge branch \u0027intel-wired-lan-driver-updates-2024-05-23-ice-idpf\u0027\n\nJacob Keller says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\nIntel Wired LAN Driver Updates 2024-05-23 (ice, idpf)\n\nThis series contains two fixes which finished up testing.\n\nFirst, Alexander fixes an issue in idpf caused by enabling NAPI and\ninterrupts prior to actually allocating the Rx buffers.\n\nSecond, Jacob fixes the ice driver VSI VLAN counting logic to ensure that\naddition and deletion of VLANs properly manages the total VSI count.\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nLink: https://lore.kernel.org/r/20240523-net-2024-05-23-intel-net-fixes-v1-0-17a923e0bb5f@intel.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "82617b9a04649e83ee8731918aeadbb6e6d7cbc7", "tree": "7194c1c7003853d949030f70033adb0016cd3dfd", "parents": [ "d514c8b54209de7a95ab37259fe32c7406976bd9" ], "author": { "name": "Jacob Keller", "email": "jacob.e.keller@intel.com", "time": "Thu May 23 10:45:30 2024 -0700" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon May 27 17:11:43 2024 -0700" }, "message": "ice: fix accounting if a VLAN already exists\n\nThe ice_vsi_add_vlan() function is used to add a VLAN filter for the target\nVSI. This function prepares a filter in the switch table for the given VSI.\nIf it succeeds, the vsi-\u003enum_vlan counter is incremented.\n\nIt is not considered an error to add a VLAN which already exists in the\nswitch table, so the function explicitly checks and ignores -EEXIST. The\nvsi-\u003enum_vlan counter is still incremented.\n\nThis seems incorrect, as it means we can double-count in the case where the\nsame VLAN is added twice by the caller. The actual table will have one less\nfilter than the count.\n\nThe ice_vsi_del_vlan() function similarly checks and handles the -ENOENT\ncondition for when deleting a filter that doesn\u0027t exist. This flow only\ndecrements the vsi-\u003enum_vlan if it actually deleted a filter.\n\nThe vsi-\u003enum_vlan counter is used only in a few places, primarily related\nto tracking the number of non-zero VLANs. If the vsi-\u003enum_vlans gets out of\nsync, then ice_vsi_num_non_zero_vlans() will incorrectly report more VLANs\nthan are present, and ice_vsi_has_non_zero_vlans() could return true\npotentially in cases where there are only VLAN 0 filters left.\n\nFix this by only incrementing the vsi-\u003enum_vlan in the case where we\nactually added an entry, and not in the case where the entry already\nexisted.\n\nFixes: a1ffafb0b4a4 (\"ice: Support configuring the device to Double VLAN Mode\")\nSigned-off-by: Jacob Keller \u003cjacob.e.keller@intel.com\u003e\nTested-by: Pucha Himasekhar Reddy \u003chimasekharx.reddy.pucha@intel.com\u003e\nReviewed-by: Simon Horman \u003chorms@kernel.org\u003e\nLink: https://lore.kernel.org/r/20240523-net-2024-05-23-intel-net-fixes-v1-2-17a923e0bb5f@intel.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "d514c8b54209de7a95ab37259fe32c7406976bd9", "tree": "aed85382b94ba3cac3cf253ba4d51fbc5defda8e", "parents": [ "266aa3b4812e97942a8ce5c7aafa7da059f7b5b8" ], "author": { "name": "Alexander Lobakin", "email": "aleksander.lobakin@intel.com", "time": "Thu May 23 10:45:29 2024 -0700" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon May 27 17:11:43 2024 -0700" }, "message": "idpf: don\u0027t enable NAPI and interrupts prior to allocating Rx buffers\n\nCurrently, idpf enables NAPI and interrupts prior to allocating Rx\nbuffers.\nThis may lead to frame loss (there are no buffers to place incoming\nframes) and even crashes on quick ifup-ifdown. Interrupts must be\nenabled only after all the resources are here and available.\nSplit interrupt init into two phases: initialization and enabling,\nand perform the second only after the queues are fully initialized.\nNote that we can\u0027t just move interrupt initialization down the init\nprocess, as the queues must have correct a ::q_vector pointer set\nand NAPI already added in order to allocate buffers correctly.\nAlso, during the deinit process, disable HW interrupts first and\nonly then disable NAPI. Otherwise, there can be a HW event leading\nto napi_schedule(), but the NAPI will already be unavailable.\n\nFixes: d4d558718266 (\"idpf: initialize interrupts and enable vport\")\nReported-by: Michal Kubiak \u003cmichal.kubiak@intel.com\u003e\nReviewed-by: Wojciech Drewek \u003cwojciech.drewek@intel.com\u003e\nSigned-off-by: Alexander Lobakin \u003caleksander.lobakin@intel.com\u003e\nReviewed-by: Simon Horman \u003chorms@kernel.org\u003e\nTested-by: Krishneil Singh \u003ckrishneil.k.singh@intel.com\u003e\nSigned-off-by: Jacob Keller \u003cjacob.e.keller@intel.com\u003e\nLink: https://lore.kernel.org/r/20240523-net-2024-05-23-intel-net-fixes-v1-1-17a923e0bb5f@intel.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "266aa3b4812e97942a8ce5c7aafa7da059f7b5b8", "tree": "e30ef008ae5c8702f807f344df7cfbb92cfec0a6", "parents": [ "4fb679040d9f758eeb3b4d01bbde6405bf20e64e" ], "author": { "name": "Alexander Lobakin", "email": "aleksander.lobakin@intel.com", "time": "Fri May 24 13:28:59 2024 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon May 27 17:00:22 2024 -0700" }, "message": "page_pool: fix \u0026page_pool_params kdoc issues\n\nAfter the tagged commit, @netdev got documented twice and the kdoc\nscript didn\u0027t notice that. Remove the second description added later\nand move the initial one according to the field position.\n\nAfter merging commit 5f8e4007c10d (\"kernel-doc: fix\nstruct_group_tagged() parsing\"), kdoc requires to describe struct\ngroups as well. \u0026page_pool_params has 2 struct groups which\ngenerated new warnings, describe them to resolve this.\n\nFixes: 403f11ac9ab7 (\"page_pool: don\u0027t use driver-set flags field directly\")\nSigned-off-by: Alexander Lobakin \u003caleksander.lobakin@intel.com\u003e\nLink: https://lore.kernel.org/r/20240524112859.2757403-1-aleksander.lobakin@intel.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "4fb679040d9f758eeb3b4d01bbde6405bf20e64e", "tree": "6a38ceba007df721804998664882a9aebd2c5865", "parents": [ "bf0497f53c8535f99b72041529d3f7708a6e2c0d" ], "author": { "name": "Horatiu Vultur", "email": "horatiu.vultur@microchip.com", "time": "Fri May 24 10:53:50 2024 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon May 27 16:58:39 2024 -0700" }, "message": "net: micrel: Fix lan8841_config_intr after getting out of sleep mode\n\nWhen the interrupt is enabled, the function lan8841_config_intr tries to\nclear any pending interrupts by reading the interrupt status, then\nchecks the return value for errors and then continue to enable the\ninterrupt. It has been seen that once the system gets out of sleep mode,\nthe interrupt status has the value 0x400 meaning that the PHY detected\nthat the link was in low power. That is correct value but the problem is\nthat the check is wrong. We try to check for errors but we return an\nerror also in this case which is not an error. Therefore fix this by\nreturning only when there is an error.\n\nFixes: a8f1a19d27ef (\"net: micrel: Add support for lan8841 PHY\")\nSigned-off-by: Horatiu Vultur \u003choratiu.vultur@microchip.com\u003e\nReviewed-by: Suman Ghosh \u003csumang@marvell.com\u003e\nReviewed-by: Andrew Lunn \u003candrew@lunn.ch\u003e\nReviewed-by: Russell King (Oracle) \u003crmk+kernel@armlinux.org.uk\u003e\nLink: https://lore.kernel.org/r/20240524085350.359812-1-horatiu.vultur@microchip.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "bf0497f53c8535f99b72041529d3f7708a6e2c0d", "tree": "21620285c420dd6c943f887730e3bdd372d4bbd8", "parents": [ "12f86b9af96a8b09969e4392311602f787b40834" ], "author": { "name": "Xiaolei Wang", "email": "xiaolei.wang@windriver.com", "time": "Fri May 24 13:05:28 2024 +0800" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon May 27 16:55:32 2024 -0700" }, "message": "net:fec: Add fec_enet_deinit()\n\nWhen fec_probe() fails or fec_drv_remove() needs to release the\nfec queue and remove a NAPI context, therefore add a function\ncorresponding to fec_enet_init() and call fec_enet_deinit() which\ndoes the opposite to release memory and remove a NAPI context.\n\nFixes: 59d0f7465644 (\"net: fec: init multi queue date structure\")\nSigned-off-by: Xiaolei Wang \u003cxiaolei.wang@windriver.com\u003e\nReviewed-by: Wei Fang \u003cwei.fang@nxp.com\u003e\nReviewed-by: Andrew Lunn \u003candrew@lunn.ch\u003e\nLink: https://lore.kernel.org/r/20240524050528.4115581-1-xiaolei.wang@windriver.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "12f86b9af96a8b09969e4392311602f787b40834", "tree": "218a973c7255a074f1fcc139e9927f5e4dbd26f7", "parents": [ "0fe53c0ab018b3399b8d4be95f32fd017c9719e1" ], "author": { "name": "Rob Herring (Arm)", "email": "robh@kernel.org", "time": "Thu May 23 12:17:50 2024 -0500" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon May 27 16:49:16 2024 -0700" }, "message": "dt-bindings: net: pse-pd: ti,tps23881: Fix missing \"additionalProperties\" constraints\n\nThe child nodes are missing \"additionalProperties\" constraints which\nmeans any undocumented properties or child nodes are allowed. Add the\nconstraints and all the undocumented properties exposed by the fix.\n\nFixes: f562202fedad (\"dt-bindings: net: pse-pd: Add bindings for TPS23881 PSE controller\")\nSigned-off-by: Rob Herring (Arm) \u003crobh@kernel.org\u003e\nAcked-by: Kory Maincent \u003ckory.maincent@bootlin.com\u003e\nLink: https://lore.kernel.org/r/20240523171750.2837331-1-robh@kernel.org\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "0fe53c0ab018b3399b8d4be95f32fd017c9719e1", "tree": "63637b51937821edf8568b3156c50d1ed0521984", "parents": [ "f4dca95fc0f6350918f2e6727e35b41f7f86fcce" ], "author": { "name": "Rob Herring (Arm)", "email": "robh@kernel.org", "time": "Thu May 23 12:17:31 2024 -0500" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon May 27 16:48:55 2024 -0700" }, "message": "dt-bindings: net: pse-pd: microchip,pd692x0: Fix missing \"additionalProperties\" constraints\n\nThe child nodes are missing \"additionalProperties\" constraints which\nmeans any undocumented properties or child nodes are allowed. Add the\nconstraints, and fix the fallout of wrong manager node regex and\nmissing properties.\n\nFixes: 9c1de033afad (\"dt-bindings: net: pse-pd: Add bindings for PD692x0 PSE controller\")\nSigned-off-by: Rob Herring (Arm) \u003crobh@kernel.org\u003e\nAcked-by: Kory Maincent \u003ckory.maincent@bootlin.com\u003e\nLink: https://lore.kernel.org/r/20240523171732.2836880-1-robh@kernel.org\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "f4dca95fc0f6350918f2e6727e35b41f7f86fcce", "tree": "9ce47f24f372bf30a1f3649952b26eb889d6fa89", "parents": [ "be008726d0ac338a6bb19c2da2853e3e2112b055" ], "author": { "name": "Eric Dumazet", "email": "edumazet@google.com", "time": "Thu May 23 13:05:27 2024 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon May 27 16:47:23 2024 -0700" }, "message": "tcp: reduce accepted window in NEW_SYN_RECV state\n\nJason commit made checks against ACK sequence less strict\nand can be exploited by attackers to establish spoofed flows\nwith less probes.\n\nInnocent users might use tcp_rmem[1] \u003d\u003d 1,000,000,000,\nor something more reasonable.\n\nAn attacker can use a regular TCP connection to learn the server\ninitial tp-\u003ercv_wnd, and use it to optimize the attack.\n\nIf we make sure that only the announced window (smaller than 65535)\nis used for ACK validation, we force an attacker to use\n65537 packets to complete the 3WHS (assuming server ISN is unknown)\n\nFixes: 378979e94e95 (\"tcp: remove 64 KByte limit for initial tp-\u003ercv_wnd value\")\nLink: https://datatracker.ietf.org/meeting/119/materials/slides-119-tcpm-ghost-acks-00\nSigned-off-by: Eric Dumazet \u003cedumazet@google.com\u003e\nAcked-by: Neal Cardwell \u003cncardwell@google.com\u003e\nReviewed-by: Jason Xing \u003ckerneljasonxing@gmail.com\u003e\nLink: https://lore.kernel.org/r/20240523130528.60376-1-edumazet@google.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "be008726d0ac338a6bb19c2da2853e3e2112b055", "tree": "5bd9fb7e267f7363f01e17937ee43aa6e3ea3c2e", "parents": [ "7b05ab85e28f615e70520d24c075249b4512044e" ], "author": { "name": "Willem de Bruijn", "email": "willemb@google.com", "time": "Thu May 23 10:13:45 2024 -0400" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon May 27 16:46:59 2024 -0700" }, "message": "net: gro: initialize network_offset in network layer\n\nSyzkaller was able to trigger\n\n kernel BUG at net/core/gro.c:424 !\n RIP: 0010:gro_pull_from_frag0 net/core/gro.c:424 [inline]\n RIP: 0010:gro_try_pull_from_frag0 net/core/gro.c:446 [inline]\n RIP: 0010:dev_gro_receive+0x242f/0x24b0 net/core/gro.c:571\n\nDue to using an incorrect NAPI_GRO_CB(skb)-\u003enetwork_offset.\n\nThe referenced commit sets this offset to 0 in skb_gro_reset_offset.\nThat matches the expected case in dev_gro_receive:\n\n pp \u003d INDIRECT_CALL_INET(ptype-\u003ecallbacks.gro_receive,\n ipv6_gro_receive, inet_gro_receive,\n \u0026gro_list-\u003elist, skb);\n\nBut syzkaller injected an skb with protocol ETH_P_TEB into an ip6gre\ndevice (by writing the IP6GRE encapsulated version to a TAP device).\nThe result was a first call to eth_gro_receive, and thus an extra\nETH_HLEN in network_offset that should not be there. First issue hit\nis when computing offset from network header in ipv6_gro_pull_exthdrs.\n\nInitialize both offsets in the network layer gro_receive.\n\nThis pairs with all reads in gro_receive, which use\nskb_gro_receive_network_offset().\n\nFixes: 186b1ea73ad8 (\"net: gro: use cb instead of skb-\u003enetwork_header\")\nReported-by: syzkaller \u003csyzkaller@googlegroups.com\u003e\nSigned-off-by: Willem de Bruijn \u003cwillemb@google.com\u003e\nCC: Richard Gobert \u003crichardbgobert@gmail.com\u003e\nReviewed-by: Eric Dumazet \u003cedumazet@google.com\u003e\nLink: https://lore.kernel.org/r/20240523141434.1752483-1-willemdebruijn.kernel@gmail.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "7b05ab85e28f615e70520d24c075249b4512044e", "tree": "6e3c4046ba5ec10c024b9c68bfdc150fd750b773", "parents": [ "2786ae339ef504f480753b54219b65471aaf98e8" ], "author": { "name": "Ido Schimmel", "email": "idosch@nvidia.com", "time": "Thu May 23 14:02:57 2024 +0300" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon May 27 16:46:05 2024 -0700" }, "message": "ipv4: Fix address dump when IPv4 is disabled on an interface\n\nCited commit started returning an error when user space requests to dump\nthe interface\u0027s IPv4 addresses and IPv4 is disabled on the interface.\nRestore the previous behavior and do not return an error.\n\nBefore cited commit:\n\n # ip address show dev dummy1\n 10: dummy1: \u003cBROADCAST,NOARP,UP,LOWER_UP\u003e mtu 1500 qdisc noqueue state UNKNOWN group default qlen 1000\n link/ether e2:40:68:98:d0:18 brd ff:ff:ff:ff:ff:ff\n inet6 fe80::e040:68ff:fe98:d018/64 scope link proto kernel_ll\n valid_lft forever preferred_lft forever\n # ip link set dev dummy1 mtu 67\n # ip address show dev dummy1\n 10: dummy1: \u003cBROADCAST,NOARP,UP,LOWER_UP\u003e mtu 67 qdisc noqueue state UNKNOWN group default qlen 1000\n link/ether e2:40:68:98:d0:18 brd ff:ff:ff:ff:ff:ff\n\nAfter cited commit:\n\n # ip address show dev dummy1\n 10: dummy1: \u003cBROADCAST,NOARP,UP,LOWER_UP\u003e mtu 1500 qdisc noqueue state UNKNOWN group default qlen 1000\n link/ether 32:2d:69:f2:9c:99 brd ff:ff:ff:ff:ff:ff\n inet6 fe80::302d:69ff:fef2:9c99/64 scope link proto kernel_ll\n valid_lft forever preferred_lft forever\n # ip link set dev dummy1 mtu 67\n # ip address show dev dummy1\n RTNETLINK answers: No such device\n Dump terminated\n\nWith this patch:\n\n # ip address show dev dummy1\n 10: dummy1: \u003cBROADCAST,NOARP,UP,LOWER_UP\u003e mtu 1500 qdisc noqueue state UNKNOWN group default qlen 1000\n link/ether de:17:56:bb:57:c0 brd ff:ff:ff:ff:ff:ff\n inet6 fe80::dc17:56ff:febb:57c0/64 scope link proto kernel_ll\n valid_lft forever preferred_lft forever\n # ip link set dev dummy1 mtu 67\n # ip address show dev dummy1\n 10: dummy1: \u003cBROADCAST,NOARP,UP,LOWER_UP\u003e mtu 67 qdisc noqueue state UNKNOWN group default qlen 1000\n link/ether de:17:56:bb:57:c0 brd ff:ff:ff:ff:ff:ff\n\nI fixed the exact same issue for IPv6 in commit c04f7dfe6ec2 (\"ipv6: Fix\naddress dump when IPv6 is disabled on an interface\"), but noted [1] that\nI am not doing the change for IPv4 because I am not aware of a way to\ndisable IPv4 on an interface other than unregistering it. I clearly\nmissed the above case.\n\n[1] https://lore.kernel.org/netdev/20240321173042.2151756-1-idosch@nvidia.com/\n\nFixes: cdb2f80f1c10 (\"inet: use xa_array iterator to implement inet_dump_ifaddr()\")\nReported-by: Carolina Jubran \u003ccjubran@nvidia.com\u003e\nReported-by: Yamen Safadi \u003cysafadi@nvidia.com\u003e\nTested-by: Carolina Jubran \u003ccjubran@nvidia.com\u003e\nReviewed-by: Petr Machata \u003cpetrm@nvidia.com\u003e\nSigned-off-by: Ido Schimmel \u003cidosch@nvidia.com\u003e\nReviewed-by: Eric Dumazet \u003cedumazet@google.com\u003e\nReviewed-by: David Ahern \u003cdsahern@kernel.org\u003e\nLink: https://lore.kernel.org/r/20240523110257.334315-1-idosch@nvidia.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "2786ae339ef504f480753b54219b65471aaf98e8", "tree": "97e30a59f15ef156645b0c8f7a389e848c0ede7c", "parents": [ "52a2f0608366a629d43dacd3191039c95fef74ba", "a63bf556160fb19591183383da6757f52119981d" ], "author": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon May 27 16:26:30 2024 -0700" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Mon May 27 16:26:30 2024 -0700" }, "message": "Merge tag \u0027for-netdev\u0027 of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf\n\nDaniel Borkmann says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\npull-request: bpf 2024-05-27\n\nWe\u0027ve added 15 non-merge commits during the last 7 day(s) which contain\na total of 18 files changed, 583 insertions(+), 55 deletions(-).\n\nThe main changes are:\n\n1) Fix broken BPF multi-uprobe PID filtering logic which filtered by thread\n while the promise was to filter by process, from Andrii Nakryiko.\n\n2) Fix the recent influx of syzkaller reports to sockmap which triggered\n a locking rule violation by performing a map_delete, from Jakub Sitnicki.\n\n3) Fixes to netkit driver in particular on skb-\u003epkt_type override upon pass\n verdict, from Daniel Borkmann.\n\n4) Fix an integer overflow in resolve_btfids which can wrongly trigger build\n failures, from Friedrich Vock.\n\n5) Follow-up fixes for ARC JIT reported by static analyzers,\n from Shahab Vahedi.\n\n* tag \u0027for-netdev\u0027 of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf:\n selftests/bpf: Cover verifier checks for mutating sockmap/sockhash\n Revert \"bpf, sockmap: Prevent lock inversion deadlock in map delete elem\"\n bpf: Allow delete from sockmap/sockhash only if update is allowed\n selftests/bpf: Add netkit test for pkt_type\n selftests/bpf: Add netkit tests for mac address\n netkit: Fix pkt_type override upon netkit pass verdict\n netkit: Fix setting mac address in l2 mode\n ARC, bpf: Fix issues reported by the static analyzers\n selftests/bpf: extend multi-uprobe tests with USDTs\n selftests/bpf: extend multi-uprobe tests with child thread case\n libbpf: detect broken PID filtering logic for multi-uprobe\n bpf: remove unnecessary rcu_read_{lock,unlock}() in multi-uprobe attach logic\n bpf: fix multi-uprobe PID filtering logic\n bpf: Fix potential integer overflow in resolve_btfids\n MAINTAINERS: Add myself as reviewer of ARM64 BPF JIT\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nLink: https://lore.kernel.org/r/20240527203551.29712-1-daniel@iogearbox.net\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "a63bf556160fb19591183383da6757f52119981d", "tree": "e6102ba20bfdef0c8312cf878c6e129995581b1c", "parents": [ "3b9ce0491a43e9af7f108b2f1bced7cd35931660" ], "author": { "name": "Jakub Sitnicki", "email": "jakub@cloudflare.com", "time": "Mon May 27 13:20:09 2024 +0200" }, "committer": { "name": "Daniel Borkmann", "email": "daniel@iogearbox.net", "time": "Mon May 27 19:34:26 2024 +0200" }, "message": "selftests/bpf: Cover verifier checks for mutating sockmap/sockhash\n\nVerifier enforces that only certain program types can mutate sock{map,hash}\nmaps, that is update it or delete from it. Add test coverage for these\nchecks so we don\u0027t regress.\n\nSigned-off-by: Jakub Sitnicki \u003cjakub@cloudflare.com\u003e\nSigned-off-by: Daniel Borkmann \u003cdaniel@iogearbox.net\u003e\nAcked-by: John Fastabend \u003cjohn.fastabend@gmail.com\u003e\nLink: https://lore.kernel.org/bpf/20240527-sockmap-verify-deletes-v1-3-944b372f2101@cloudflare.com\n" }, { "commit": "3b9ce0491a43e9af7f108b2f1bced7cd35931660", "tree": "3ba3ed3599c9ad372609e9e7ffb6bf1cba40e24d", "parents": [ "98e948fb60d41447fd8d2d0c3b8637fc6b6dc26d" ], "author": { "name": "Jakub Sitnicki", "email": "jakub@cloudflare.com", "time": "Mon May 27 13:20:08 2024 +0200" }, "committer": { "name": "Daniel Borkmann", "email": "daniel@iogearbox.net", "time": "Mon May 27 19:34:25 2024 +0200" }, "message": "Revert \"bpf, sockmap: Prevent lock inversion deadlock in map delete elem\"\n\nThis reverts commit ff91059932401894e6c86341915615c5eb0eca48.\n\nThis check is no longer needed. BPF programs attached to tracepoints are\nnow rejected by the verifier when they attempt to delete from a\nsockmap/sockhash maps.\n\nSigned-off-by: Jakub Sitnicki \u003cjakub@cloudflare.com\u003e\nSigned-off-by: Daniel Borkmann \u003cdaniel@iogearbox.net\u003e\nAcked-by: John Fastabend \u003cjohn.fastabend@gmail.com\u003e\nLink: https://lore.kernel.org/bpf/20240527-sockmap-verify-deletes-v1-2-944b372f2101@cloudflare.com\n" }, { "commit": "98e948fb60d41447fd8d2d0c3b8637fc6b6dc26d", "tree": "a96449f76eff9a59365db5570758d363176cea30", "parents": [ "95348e463eabc803341c67d562f9e0a5f0a48fe6" ], "author": { "name": "Jakub Sitnicki", "email": "jakub@cloudflare.com", "time": "Mon May 27 13:20:07 2024 +0200" }, "committer": { "name": "Daniel Borkmann", "email": "daniel@iogearbox.net", "time": "Mon May 27 19:33:40 2024 +0200" }, "message": "bpf: Allow delete from sockmap/sockhash only if update is allowed\n\nWe have seen an influx of syzkaller reports where a BPF program attached to\na tracepoint triggers a locking rule violation by performing a map_delete\non a sockmap/sockhash.\n\nWe don\u0027t intend to support this artificial use scenario. Extend the\nexisting verifier allowed-program-type check for updating sockmap/sockhash\nto also cover deleting from a map.\n\nFrom now on only BPF programs which were previously allowed to update\nsockmap/sockhash can delete from these map types.\n\nFixes: ff9105993240 (\"bpf, sockmap: Prevent lock inversion deadlock in map delete elem\")\nReported-by: Tetsuo Handa \u003cpenguin-kernel@i-love.sakura.ne.jp\u003e\nReported-by: syzbot+ec941d6e24f633a59172@syzkaller.appspotmail.com\nSigned-off-by: Jakub Sitnicki \u003cjakub@cloudflare.com\u003e\nSigned-off-by: Daniel Borkmann \u003cdaniel@iogearbox.net\u003e\nTested-by: syzbot+ec941d6e24f633a59172@syzkaller.appspotmail.com\nAcked-by: John Fastabend \u003cjohn.fastabend@gmail.com\u003e\nCloses: https://syzkaller.appspot.com/bug?extid\u003dec941d6e24f633a59172\nLink: https://lore.kernel.org/bpf/20240527-sockmap-verify-deletes-v1-1-944b372f2101@cloudflare.com\n" }, { "commit": "52a2f0608366a629d43dacd3191039c95fef74ba", "tree": "c19cc41466a77d157f5842e3f9ef6d6e299397f8", "parents": [ "1684842147677a1279bcff95f8adb6de9a656e30" ], "author": { "name": "Parthiban Veerasooran", "email": "Parthiban.Veerasooran@microchip.com", "time": "Thu May 23 14:23:14 2024 +0530" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Mon May 27 12:48:23 2024 +0200" }, "message": "net: usb: smsc95xx: fix changing LED_SEL bit value updated from EEPROM\n\nLED Select (LED_SEL) bit in the LED General Purpose IO Configuration\nregister is used to determine the functionality of external LED pins\n(Speed Indicator, Link and Activity Indicator, Full Duplex Link\nIndicator). The default value for this bit is 0 when no EEPROM is\npresent. If a EEPROM is present, the default value is the value of the\nLED Select bit in the Configuration Flags of the EEPROM. A USB Reset or\nLite Reset (LRST) will cause this bit to be restored to the image value\nlast loaded from EEPROM, or to be set to 0 if no EEPROM is present.\n\nWhile configuring the dual purpose GPIO/LED pins to LED outputs in the\nLED General Purpose IO Configuration register, the LED_SEL bit is changed\nas 0 and resulting the configured value from the EEPROM is cleared. The\nissue is fixed by using read-modify-write approach.\n\nFixes: f293501c61c5 (\"smsc95xx: configure LED outputs\")\nSigned-off-by: Parthiban Veerasooran \u003cParthiban.Veerasooran@microchip.com\u003e\nReviewed-by: Simon Horman \u003chorms@kernel.org\u003e\nReviewed-by: Woojung Huh \u003cwoojung.huh@microchip.com\u003e\nLink: https://lore.kernel.org/r/20240523085314.167650-1-Parthiban.Veerasooran@microchip.com\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "1684842147677a1279bcff95f8adb6de9a656e30", "tree": "e04723417993ccbbaf6bdb08e59e86aab8118094", "parents": [ "51d1b25a720982324871338b1a36b197ec9bd6f0" ], "author": { "name": "Hariprasad Kelam", "email": "hkelam@marvell.com", "time": "Thu May 23 13:06:26 2024 +0530" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Mon May 27 11:55:47 2024 +0200" }, "message": "Octeontx2-pf: Free send queue buffers incase of leaf to inner\n\nThere are two type of classes. \"Leaf classes\" that are the\nbottom of the class hierarchy. \"Inner classes\" that are neither\nthe root class nor leaf classes. QoS rules can only specify leaf\nclasses as targets for traffic.\n\n\t\t\t Root\n\t\t / \\\n\t\t / \\\n 1 2\n /\\\n / \\\n 4 5\n classes 1,4 and 5 are leaf classes.\n class 2 is a inner class.\n\nWhen a leaf class made as inner, or vice versa, resources associated\nwith send queue (send queue buffers and transmit schedulers) are not\ngetting freed.\n\nFixes: 5e6808b4c68d (\"octeontx2-pf: Add support for HTB offload\")\nSigned-off-by: Hariprasad Kelam \u003chkelam@marvell.com\u003e\nLink: https://lore.kernel.org/r/20240523073626.4114-1-hkelam@marvell.com\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "51d1b25a720982324871338b1a36b197ec9bd6f0", "tree": "9902f7dd18ec9436bfe906aafedb0e1cb89309f7", "parents": [ "97e1db06c7bb948da10ba85acad8030b56886593" ], "author": { "name": "Kuniyuki Iwashima", "email": "kuniyu@amazon.com", "time": "Thu May 23 00:42:18 2024 +0900" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Mon May 27 11:46:56 2024 +0200" }, "message": "af_unix: Read sk-\u003esk_hash under bindlock during bind().\n\nsyzkaller reported data-race of sk-\u003esk_hash in unix_autobind() [0],\nand the same ones exist in unix_bind_bsd() and unix_bind_abstract().\n\nThe three bind() functions prefetch sk-\u003esk_hash locklessly and\nuse it later after validating that unix_sk(sk)-\u003eaddr is NULL under\nunix_sk(sk)-\u003ebindlock.\n\nThe prefetched sk-\u003esk_hash is the hash value of unbound socket set\nin unix_create1() and does not change until bind() completes.\n\nThere could be a chance that sk-\u003esk_hash changes after the lockless\nread. However, in such a case, non-NULL unix_sk(sk)-\u003eaddr is visible\nunder unix_sk(sk)-\u003ebindlock, and bind() returns -EINVAL without using\nthe prefetched value.\n\nThe KCSAN splat is false-positive, but let\u0027s silence it by reading\nsk-\u003esk_hash under unix_sk(sk)-\u003ebindlock.\n\n[0]:\nBUG: KCSAN: data-race in unix_autobind / unix_autobind\n\nwrite to 0xffff888034a9fb88 of 4 bytes by task 4468 on cpu 0:\n __unix_set_addr_hash net/unix/af_unix.c:331 [inline]\n unix_autobind+0x47a/0x7d0 net/unix/af_unix.c:1185\n unix_dgram_connect+0x7e3/0x890 net/unix/af_unix.c:1373\n __sys_connect_file+0xd7/0xe0 net/socket.c:2048\n __sys_connect+0x114/0x140 net/socket.c:2065\n __do_sys_connect net/socket.c:2075 [inline]\n __se_sys_connect net/socket.c:2072 [inline]\n __x64_sys_connect+0x40/0x50 net/socket.c:2072\n do_syscall_x64 arch/x86/entry/common.c:52 [inline]\n do_syscall_64+0x4f/0x110 arch/x86/entry/common.c:83\n entry_SYSCALL_64_after_hwframe+0x46/0x4e\n\nread to 0xffff888034a9fb88 of 4 bytes by task 4465 on cpu 1:\n unix_autobind+0x28/0x7d0 net/unix/af_unix.c:1134\n unix_dgram_connect+0x7e3/0x890 net/unix/af_unix.c:1373\n __sys_connect_file+0xd7/0xe0 net/socket.c:2048\n __sys_connect+0x114/0x140 net/socket.c:2065\n __do_sys_connect net/socket.c:2075 [inline]\n __se_sys_connect net/socket.c:2072 [inline]\n __x64_sys_connect+0x40/0x50 net/socket.c:2072\n do_syscall_x64 arch/x86/entry/common.c:52 [inline]\n do_syscall_64+0x4f/0x110 arch/x86/entry/common.c:83\n entry_SYSCALL_64_after_hwframe+0x46/0x4e\n\nvalue changed: 0x000000e4 -\u003e 0x000001e3\n\nReported by Kernel Concurrency Sanitizer on:\nCPU: 1 PID: 4465 Comm: syz-executor.0 Not tainted 6.8.0-12822-gcd51db110a7e #12\nHardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.16.0-0-gd239552ce722-prebuilt.qemu.org 04/01/2014\n\nFixes: afd20b9290e1 (\"af_unix: Replace the big lock with small locks.\")\nReported-by: syzkaller \u003csyzkaller@googlegroups.com\u003e\nSigned-off-by: Kuniyuki Iwashima \u003ckuniyu@amazon.com\u003e\nLink: https://lore.kernel.org/r/20240522154218.78088-1-kuniyu@amazon.com\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "97e1db06c7bb948da10ba85acad8030b56886593", "tree": "59630c2207b9a6977654338a0c1071ad37babffa", "parents": [ "21a22ed618d072a47597e63ee591973c18524880" ], "author": { "name": "Kuniyuki Iwashima", "email": "kuniyu@amazon.com", "time": "Thu May 23 00:40:02 2024 +0900" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Mon May 27 11:46:31 2024 +0200" }, "message": "af_unix: Annotate data-race around unix_sk(sk)-\u003eaddr.\n\nOnce unix_sk(sk)-\u003eaddr is assigned under net-\u003eunx.table.locks and\nunix_sk(sk)-\u003ebindlock, *(unix_sk(sk)-\u003eaddr) and unix_sk(sk)-\u003epath are\nfully set up, and unix_sk(sk)-\u003eaddr is never changed.\n\nunix_getname() and unix_copy_addr() access the two fields locklessly,\nand commit ae3b564179bf (\"missing barriers in some of unix_sock -\u003eaddr\nand -\u003epath accesses\") added smp_store_release() and smp_load_acquire()\npairs.\n\nIn other functions, we still read unix_sk(sk)-\u003eaddr locklessly to check\nif the socket is bound, and KCSAN complains about it. [0]\n\nGiven these functions have no dependency for *(unix_sk(sk)-\u003eaddr) and\nunix_sk(sk)-\u003epath, READ_ONCE() is enough to annotate the data-race.\n\nNote that it is safe to access unix_sk(sk)-\u003eaddr locklessly if the socket\nis found in the hash table. For example, the lockless read of otheru-\u003eaddr\nin unix_stream_connect() is safe.\n\nNote also that newu-\u003eaddr there is of the child socket that is still not\naccessible from userspace, and smp_store_release() publishes the address\nin case the socket is accept()ed and unix_getname() / unix_copy_addr()\nis called.\n\n[0]:\nBUG: KCSAN: data-race in unix_bind / unix_listen\n\nwrite (marked) to 0xffff88805f8d1840 of 8 bytes by task 13723 on cpu 0:\n __unix_set_addr_hash net/unix/af_unix.c:329 [inline]\n unix_bind_bsd net/unix/af_unix.c:1241 [inline]\n unix_bind+0x881/0x1000 net/unix/af_unix.c:1319\n __sys_bind+0x194/0x1e0 net/socket.c:1847\n __do_sys_bind net/socket.c:1858 [inline]\n __se_sys_bind net/socket.c:1856 [inline]\n __x64_sys_bind+0x40/0x50 net/socket.c:1856\n do_syscall_x64 arch/x86/entry/common.c:52 [inline]\n do_syscall_64+0x4f/0x110 arch/x86/entry/common.c:83\n entry_SYSCALL_64_after_hwframe+0x46/0x4e\n\nread to 0xffff88805f8d1840 of 8 bytes by task 13724 on cpu 1:\n unix_listen+0x72/0x180 net/unix/af_unix.c:734\n __sys_listen+0xdc/0x160 net/socket.c:1881\n __do_sys_listen net/socket.c:1890 [inline]\n __se_sys_listen net/socket.c:1888 [inline]\n __x64_sys_listen+0x2e/0x40 net/socket.c:1888\n do_syscall_x64 arch/x86/entry/common.c:52 [inline]\n do_syscall_64+0x4f/0x110 arch/x86/entry/common.c:83\n entry_SYSCALL_64_after_hwframe+0x46/0x4e\n\nvalue changed: 0x0000000000000000 -\u003e 0xffff88807b5b1b40\n\nReported by Kernel Concurrency Sanitizer on:\nCPU: 1 PID: 13724 Comm: syz-executor.4 Not tainted 6.8.0-12822-gcd51db110a7e #12\nHardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.16.0-0-gd239552ce722-prebuilt.qemu.org 04/01/2014\n\nFixes: 1da177e4c3f4 (\"Linux-2.6.12-rc2\")\nReported-by: syzkaller \u003csyzkaller@googlegroups.com\u003e\nSigned-off-by: Kuniyuki Iwashima \u003ckuniyu@amazon.com\u003e\nLink: https://lore.kernel.org/r/20240522154002.77857-1-kuniyu@amazon.com\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "21a22ed618d072a47597e63ee591973c18524880", "tree": "4578349ff1febf8bceeba188e43947434db09db3", "parents": [ "e8021b94b0412c37bcc79027c2e382086b6ce449" ], "author": { "name": "Geliang Tang", "email": "tanggeliang@kylinos.cn", "time": "Wed May 22 18:45:04 2024 +0800" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Mon May 27 11:44:42 2024 +0200" }, "message": "selftests: hsr: Fix \"File exists\" errors for hsr_ping\n\nThe hsr_ping test reports the following errors:\n\n INFO: preparing interfaces for HSRv0.\n INFO: Initial validation ping.\n INFO: Longer ping test.\n INFO: Cutting one link.\n INFO: Delay the link and drop a few packages.\n INFO: All good.\n INFO: preparing interfaces for HSRv1.\n RTNETLINK answers: File exists\n RTNETLINK answers: File exists\n RTNETLINK answers: File exists\n RTNETLINK answers: File exists\n RTNETLINK answers: File exists\n RTNETLINK answers: File exists\n Error: ipv4: Address already assigned.\n Error: ipv6: address already assigned.\n Error: ipv4: Address already assigned.\n Error: ipv6: address already assigned.\n Error: ipv4: Address already assigned.\n Error: ipv6: address already assigned.\n INFO: Initial validation ping.\n\nThat is because the cleanup code for the 2nd round test before\n\"setup_hsr_interfaces 1\" is removed incorrectly in commit 680fda4f6714\n(\"test: hsr: Remove script code already implemented in lib.sh\").\n\nThis patch fixes it by re-setup the namespaces using\n\n\tsetup_ns ns1 ns2 ns3\n\ncommand before \"setup_hsr_interfaces 1\". It deletes previous namespaces\nand create new ones.\n\nFixes: 680fda4f6714 (\"test: hsr: Remove script code already implemented in lib.sh\")\nReviewed-by: Hangbin Liu \u003cliuhangbin@gmail.com\u003e\nSigned-off-by: Geliang Tang \u003ctanggeliang@kylinos.cn\u003e\nLink: https://lore.kernel.org/r/6485d3005f467758d49f0f313c8c009759ba6b05.1716374462.git.tanggeliang@kylinos.cn\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "e8021b94b0412c37bcc79027c2e382086b6ce449", "tree": "5e2720872cdc9e6558dae1623298f8fa339fe1da", "parents": [ "0b4f5add9fa59bfd42c1030f572db2e4c395181b" ], "author": { "name": "Roded Zats", "email": "rzats@paloaltonetworks.com", "time": "Wed May 22 10:30:44 2024 +0300" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Mon May 27 11:18:01 2024 +0200" }, "message": "enic: Validate length of nl attributes in enic_set_vf_port\n\nenic_set_vf_port assumes that the nl attribute IFLA_PORT_PROFILE\nis of length PORT_PROFILE_MAX and that the nl attributes\nIFLA_PORT_INSTANCE_UUID, IFLA_PORT_HOST_UUID are of length PORT_UUID_MAX.\nThese attributes are validated (in the function do_setlink in rtnetlink.c)\nusing the nla_policy ifla_port_policy. The policy defines IFLA_PORT_PROFILE\nas NLA_STRING, IFLA_PORT_INSTANCE_UUID as NLA_BINARY and\nIFLA_PORT_HOST_UUID as NLA_STRING. That means that the length validation\nusing the policy is for the max size of the attributes and not on exact\nsize so the length of these attributes might be less than the sizes that\nenic_set_vf_port expects. This might cause an out of bands\nread access in the memcpys of the data of these\nattributes in enic_set_vf_port.\n\nFixes: f8bd909183ac (\"net: Add ndo_{set|get}_vf_port support for enic dynamic vnics\")\nSigned-off-by: Roded Zats \u003crzats@paloaltonetworks.com\u003e\nLink: https://lore.kernel.org/r/20240522073044.33519-1-rzats@paloaltonetworks.com\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "95348e463eabc803341c67d562f9e0a5f0a48fe6", "tree": "e8062791b197694f238e725e2bcb2cadfb5910f7", "parents": [ "998ffeb2738e26f134dc8e63b5dcaece22573957" ], "author": { "name": "Daniel Borkmann", "email": "daniel@iogearbox.net", "time": "Fri May 24 18:36:19 2024 +0200" }, "committer": { "name": "Alexei Starovoitov", "email": "ast@kernel.org", "time": "Sat May 25 10:53:11 2024 -0700" }, "message": "selftests/bpf: Add netkit test for pkt_type\n\nAdd a test case to assert that the skb-\u003epkt_type which was set from the BPF\nprogram is retained from the netkit xmit side to the peer\u0027s device at tcx\ningress location.\n\n # ./vmtest.sh -- ./test_progs -t netkit\n [...]\n ./test_progs -t netkit\n [ 1.140780] bpf_testmod: loading out-of-tree module taints kernel.\n [ 1.141127] bpf_testmod: module verification failed: signature and/or required key missing - tainting kernel\n [ 1.284601] tsc: Refined TSC clocksource calibration: 3408.006 MHz\n [ 1.286672] clocksource: tsc: mask: 0xffffffffffffffff max_cycles: 0x311fd9b189d, max_idle_ns: 440795225691 ns\n [ 1.290384] clocksource: Switched to clocksource tsc\n #345 tc_netkit_basic:OK\n #346 tc_netkit_device:OK\n #347 tc_netkit_multi_links:OK\n #348 tc_netkit_multi_opts:OK\n #349 tc_netkit_neigh_links:OK\n #350 tc_netkit_pkt_type:OK\n Summary: 6/0 PASSED, 0 SKIPPED, 0 FAILED\n\nSigned-off-by: Daniel Borkmann \u003cdaniel@iogearbox.net\u003e\nLink: https://lore.kernel.org/r/20240524163619.26001-4-daniel@iogearbox.net\nSigned-off-by: Alexei Starovoitov \u003cast@kernel.org\u003e\n" }, { "commit": "998ffeb2738e26f134dc8e63b5dcaece22573957", "tree": "32d71431910f2beba683c40b443997986f08ba81", "parents": [ "3998d184267dfcff858aaa84d3de17429253629d" ], "author": { "name": "Daniel Borkmann", "email": "daniel@iogearbox.net", "time": "Fri May 24 18:36:18 2024 +0200" }, "committer": { "name": "Alexei Starovoitov", "email": "ast@kernel.org", "time": "Sat May 25 10:48:57 2024 -0700" }, "message": "selftests/bpf: Add netkit tests for mac address\n\nThis adds simple tests around setting MAC addresses in the different\nnetkit modes.\n\nSigned-off-by: Daniel Borkmann \u003cdaniel@iogearbox.net\u003e\nLink: https://lore.kernel.org/r/20240524163619.26001-3-daniel@iogearbox.net\nSigned-off-by: Alexei Starovoitov \u003cast@kernel.org\u003e\n" }, { "commit": "3998d184267dfcff858aaa84d3de17429253629d", "tree": "9862377f9644922177186ccfce9296e527fd3e03", "parents": [ "d6fe532b7499e4575f9647879b7a34625817fe7f" ], "author": { "name": "Daniel Borkmann", "email": "daniel@iogearbox.net", "time": "Fri May 24 18:36:17 2024 +0200" }, "committer": { "name": "Alexei Starovoitov", "email": "ast@kernel.org", "time": "Sat May 25 10:48:57 2024 -0700" }, "message": "netkit: Fix pkt_type override upon netkit pass verdict\n\nWhen running Cilium connectivity test suite with netkit in L2 mode, we\nfound that compared to tcx a few tests were failing which pushed traffic\ninto an L7 proxy sitting in host namespace. The problem in particular is\naround the invocation of eth_type_trans() in netkit.\n\nIn case of tcx, this is run before the tcx ingress is triggered inside\nhost namespace and thus if the BPF program uses the bpf_skb_change_type()\nhelper the newly set type is retained. However, in case of netkit, the\nlate eth_type_trans() invocation overrides the earlier decision from the\nBPF program which eventually leads to the test failure.\n\nInstead of eth_type_trans(), split out the relevant parts, meaning, reset\nof mac header and call to eth_skb_pkt_type() before the BPF program is run\nin order to have the same behavior as with tcx, and refactor a small helper\ncalled eth_skb_pull_mac() which is run in case it\u0027s passed up the stack\nwhere the mac header must be pulled. With this all connectivity tests pass.\n\nFixes: 35dfaad7188c (\"netkit, bpf: Add bpf programmable net device\")\nSigned-off-by: Daniel Borkmann \u003cdaniel@iogearbox.net\u003e\nAcked-by: Nikolay Aleksandrov \u003crazor@blackwall.org\u003e\nLink: https://lore.kernel.org/r/20240524163619.26001-2-daniel@iogearbox.net\nSigned-off-by: Alexei Starovoitov \u003cast@kernel.org\u003e\n" }, { "commit": "d6fe532b7499e4575f9647879b7a34625817fe7f", "tree": "303db3b2d23ec12cf01998f603ec97948b4f414b", "parents": [ "dd6a403795f0c7b5c566f86f2ee6b687278d3c1c" ], "author": { "name": "Daniel Borkmann", "email": "daniel@iogearbox.net", "time": "Fri May 24 18:36:16 2024 +0200" }, "committer": { "name": "Alexei Starovoitov", "email": "ast@kernel.org", "time": "Sat May 25 10:48:57 2024 -0700" }, "message": "netkit: Fix setting mac address in l2 mode\n\nWhen running Cilium connectivity test suite with netkit in L2 mode, we\nfound that it is expected to be able to specify a custom MAC address for\nthe devices, in particular, cilium-cni obtains the specified MAC address\nby querying the endpoint and sets the MAC address of the interface inside\nthe Pod. Thus, fix the missing support in netkit for L2 mode.\n\nFixes: 35dfaad7188c (\"netkit, bpf: Add bpf programmable net device\")\nSigned-off-by: Daniel Borkmann \u003cdaniel@iogearbox.net\u003e\nAcked-by: Nikolay Aleksandrov \u003crazor@blackwall.org\u003e\nAcked-by: Stanislav Fomichev \u003csdf@google.com\u003e\nLink: https://lore.kernel.org/r/20240524163619.26001-1-daniel@iogearbox.net\nSigned-off-by: Alexei Starovoitov \u003cast@kernel.org\u003e\n" }, { "commit": "dd6a403795f0c7b5c566f86f2ee6b687278d3c1c", "tree": "365a2e0e7266d9be0554f260d4bd58bf688a8f07", "parents": [ "590016ad83de770153a09151336d95544d6bd7ad" ], "author": { "name": "Shahab Vahedi", "email": "shahab@synopsys.com", "time": "Sat May 25 05:56:28 2024 +0200" }, "committer": { "name": "Alexei Starovoitov", "email": "ast@kernel.org", "time": "Sat May 25 10:47:21 2024 -0700" }, "message": "ARC, bpf: Fix issues reported by the static analyzers\n\nAlso updated couple of comments along the way.\n\nOne of the issues reported was indeed a bug in the code:\n\n memset(ctx, 0, sizeof(ctx)) // original line\n memset(ctx, 0, sizeof(*ctx)) // fixed line\n\nThat was a nice catch.\n\nReported-by: kernel test robot \u003clkp@intel.com\u003e\nCloses: https://lore.kernel.org/oe-kbuild-all/202405222314.UG5F2NHn-lkp@intel.com/\nCloses: https://lore.kernel.org/oe-kbuild-all/202405232036.Xqoc3b0J-lkp@intel.com/\nSigned-off-by: Shahab Vahedi \u003cshahab@synopsys.com\u003e\nLink: https://lore.kernel.org/r/20240525035628.1026-1-list+bpf@vahedi.org\nSigned-off-by: Alexei Starovoitov \u003cast@kernel.org\u003e\n" }, { "commit": "590016ad83de770153a09151336d95544d6bd7ad", "tree": "2f016edc5b9d6493df7368f06884af7bac6e574d", "parents": [ "44382b3ed6b2787710c8ade06c0e97f5970a47c8", "198034a87dfeb64d5a8359a5089022c6b923646e" ], "author": { "name": "Alexei Starovoitov", "email": "ast@kernel.org", "time": "Sat May 25 10:46:03 2024 -0700" }, "committer": { "name": "Alexei Starovoitov", "email": "ast@kernel.org", "time": "Sat May 25 10:46:03 2024 -0700" }, "message": "Merge branch \u0027fix-bpf-multi-uprobe-pid-filtering-logic\u0027\n\nAndrii Nakryiko says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\nFix BPF multi-uprobe PID filtering logic\n\nIt turns out that current implementation of multi-uprobe PID filtering logic\nis broken. It filters by thread, while the promise is filtering by process.\nPatch #1 fixes the logic trivially. The rest is testing and mitigations that\nare necessary for libbpf to not break users of USDT programs.\n\nv1-\u003ev2:\n - fix selftest in last patch (CI);\n - use semicolon in patch #3 (Jiri).\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nLink: https://lore.kernel.org/r/20240521163401.3005045-1-andrii@kernel.org\nSigned-off-by: Alexei Starovoitov \u003cast@kernel.org\u003e\n" }, { "commit": "198034a87dfeb64d5a8359a5089022c6b923646e", "tree": "2f016edc5b9d6493df7368f06884af7bac6e574d", "parents": [ "70342420a1cf1173bdec456e5fa574a804e422db" ], "author": { "name": "Andrii Nakryiko", "email": "andrii@kernel.org", "time": "Tue May 21 09:34:01 2024 -0700" }, "committer": { "name": "Alexei Starovoitov", "email": "ast@kernel.org", "time": "Sat May 25 10:46:02 2024 -0700" }, "message": "selftests/bpf: extend multi-uprobe tests with USDTs\n\nValidate libbpf\u0027s USDT-over-multi-uprobe logic by adding USDTs to\nexisting multi-uprobe tests. This checks correct libbpf fallback to\nsingular uprobes (when run on older kernels with buggy PID filtering).\nWe reuse already established child process and child thread testing\ninfrastructure, so additions are minimal. These test fail on either\nolder kernels or older version of libbpf that doesn\u0027t detect PID\nfiltering problems.\n\nAcked-by: Jiri Olsa \u003cjolsa@kernel.org\u003e\nSigned-off-by: Andrii Nakryiko \u003candrii@kernel.org\u003e\nLink: https://lore.kernel.org/r/20240521163401.3005045-6-andrii@kernel.org\nSigned-off-by: Alexei Starovoitov \u003cast@kernel.org\u003e\n" }, { "commit": "70342420a1cf1173bdec456e5fa574a804e422db", "tree": "4636693b5cc3ec019e1c1e69b420f136b41b0ee1", "parents": [ "04d939a2ab229a3821f04fc81f7c027842f501f1" ], "author": { "name": "Andrii Nakryiko", "email": "andrii@kernel.org", "time": "Tue May 21 09:34:00 2024 -0700" }, "committer": { "name": "Alexei Starovoitov", "email": "ast@kernel.org", "time": "Sat May 25 10:46:02 2024 -0700" }, "message": "selftests/bpf: extend multi-uprobe tests with child thread case\n\nExtend existing multi-uprobe tests to test that PID filtering works\ncorrectly. We already have child *process* tests, but we need also child\n*thread* tests. This patch adds spawn_thread() helper to start child\nthread, wait for it to be ready, and then instruct it to trigger desired\nuprobes.\n\nAdditionally, we extend BPF-side code to track thread ID, not just\nprocess ID. Also we detect whether extraneous triggerings with\nunexpected process IDs happened, and validate that none of that happened\nin practice.\n\nThese changes prove that fixed PID filtering logic for multi-uprobe\nworks as expected. These tests fail on old kernels.\n\nSigned-off-by: Andrii Nakryiko \u003candrii@kernel.org\u003e\nAcked-by: Jiri Olsa \u003cjolsa@kernel.org\u003e\nLink: https://lore.kernel.org/r/20240521163401.3005045-5-andrii@kernel.org\nSigned-off-by: Alexei Starovoitov \u003cast@kernel.org\u003e\n" }, { "commit": "04d939a2ab229a3821f04fc81f7c027842f501f1", "tree": "9d3f1d6d2832d365a5f87a1ff8b7dc9f2328b487", "parents": [ "4a8f635a60540888dab3804992e86410360339c8" ], "author": { "name": "Andrii Nakryiko", "email": "andrii@kernel.org", "time": "Tue May 21 09:33:59 2024 -0700" }, "committer": { "name": "Alexei Starovoitov", "email": "ast@kernel.org", "time": "Sat May 25 10:46:02 2024 -0700" }, "message": "libbpf: detect broken PID filtering logic for multi-uprobe\n\nLibbpf is automatically (and transparently to user) detecting\nmulti-uprobe support in the kernel, and, if supported, uses\nmulti-uprobes to improve USDT attachment speed.\n\nUSDTs can be attached system-wide or for the specific process by PID. In\nthe latter case, we rely on correct kernel logic of not triggering USDT\nfor unrelated processes.\n\nAs such, on older kernels that do support multi-uprobes, but still have\nbroken PID filtering logic, we need to fall back to singular uprobes.\n\nUnfortunately, whether user is using PID filtering or not is known at\nthe attachment time, which happens after relevant BPF programs were\nloaded into the kernel. Also unfortunately, we need to make a call\nwhether to use multi-uprobes or singular uprobe for SEC(\"usdt\") programs\nduring BPF object load time, at which point we have no information about\npossible PID filtering.\n\nThe distinction between single and multi-uprobes is small, but important\nfor the kernel. Multi-uprobes get BPF_TRACE_UPROBE_MULTI attach type,\nand kernel internally substitiute different implementation of some of\nBPF helpers (e.g., bpf_get_attach_cookie()) depending on whether uprobe\nis multi or singular. So, multi-uprobes and singular uprobes cannot be\nintermixed.\n\nAll the above implies that we have to make an early and conservative\ncall about the use of multi-uprobes. And so this patch modifies libbpf\u0027s\nexisting feature detector for multi-uprobe support to also check correct\nPID filtering. If PID filtering is not yet fixed, we fall back to\nsingular uprobes for USDTs.\n\nThis extension to feature detection is simple thanks to kernel\u0027s -EINVAL\naddition for pid \u003c 0.\n\nAcked-by: Jiri Olsa \u003cjolsa@kernel.org\u003e\nSigned-off-by: Andrii Nakryiko \u003candrii@kernel.org\u003e\nLink: https://lore.kernel.org/r/20240521163401.3005045-4-andrii@kernel.org\nSigned-off-by: Alexei Starovoitov \u003cast@kernel.org\u003e\n" }, { "commit": "4a8f635a60540888dab3804992e86410360339c8", "tree": "d5616c72bde13807d4ff75fab5acfa3c5371143b", "parents": [ "46ba0e49b64232adac35a2bc892f1710c5b0fb7f" ], "author": { "name": "Andrii Nakryiko", "email": "andrii@kernel.org", "time": "Tue May 21 09:33:58 2024 -0700" }, "committer": { "name": "Alexei Starovoitov", "email": "ast@kernel.org", "time": "Sat May 25 10:46:02 2024 -0700" }, "message": "bpf: remove unnecessary rcu_read_{lock,unlock}() in multi-uprobe attach logic\n\nget_pid_task() internally already calls rcu_read_lock() and\nrcu_read_unlock(), so there is no point to do this one extra time.\n\nThis is a drive-by improvement and has no correctness implications.\n\nAcked-by: Jiri Olsa \u003cjolsa@kernel.org\u003e\nSigned-off-by: Andrii Nakryiko \u003candrii@kernel.org\u003e\nLink: https://lore.kernel.org/r/20240521163401.3005045-3-andrii@kernel.org\nSigned-off-by: Alexei Starovoitov \u003cast@kernel.org\u003e\n" }, { "commit": "46ba0e49b64232adac35a2bc892f1710c5b0fb7f", "tree": "6fd4199126856396394680fd9553ac4c14fd81c9", "parents": [ "44382b3ed6b2787710c8ade06c0e97f5970a47c8" ], "author": { "name": "Andrii Nakryiko", "email": "andrii@kernel.org", "time": "Tue May 21 09:33:57 2024 -0700" }, "committer": { "name": "Alexei Starovoitov", "email": "ast@kernel.org", "time": "Sat May 25 10:46:02 2024 -0700" }, "message": "bpf: fix multi-uprobe PID filtering logic\n\nCurrent implementation of PID filtering logic for multi-uprobes in\nuprobe_prog_run() is filtering down to exact *thread*, while the intent\nfor PID filtering it to filter by *process* instead. The check in\nuprobe_prog_run() also differs from the analogous one in\nuprobe_multi_link_filter() for some reason. The latter is correct,\nchecking task-\u003emm, not the task itself.\n\nFix the check in uprobe_prog_run() to perform the same task-\u003emm check.\n\nWhile doing this, we also update get_pid_task() use to use PIDTYPE_TGID\ntype of lookup, given the intent is to get a representative task of an\nentire process. This doesn\u0027t change behavior, but seems more logical. It\nwould hold task group leader task now, not any random thread task.\n\nLast but not least, given multi-uprobe support is half-broken due to\nthis PID filtering logic (depending on whether PID filtering is\nimportant or not), we need to make it easy for user space consumers\n(including libbpf) to easily detect whether PID filtering logic was\nalready fixed.\n\nWe do it here by adding an early check on passed pid parameter. If it\u0027s\nnegative (and so has no chance of being a valid PID), we return -EINVAL.\nPrevious behavior would eventually return -ESRCH (\"No process found\"),\ngiven there can\u0027t be any process with negative PID. This subtle change\nwon\u0027t make any practical change in behavior, but will allow applications\nto detect PID filtering fixes easily. Libbpf fixes take advantage of\nthis in the next patch.\n\nCc: stable@vger.kernel.org\nAcked-by: Jiri Olsa \u003cjolsa@kernel.org\u003e\nFixes: b733eeade420 (\"bpf: Add pid filter support for uprobe_multi link\")\nSigned-off-by: Andrii Nakryiko \u003candrii@kernel.org\u003e\nLink: https://lore.kernel.org/r/20240521163401.3005045-2-andrii@kernel.org\nSigned-off-by: Alexei Starovoitov \u003cast@kernel.org\u003e\n" }, { "commit": "44382b3ed6b2787710c8ade06c0e97f5970a47c8", "tree": "ffd66a523cd4888ffba95883d30df99f22a9fb7d", "parents": [ "8d00547ea8754afdc4a550af2fb7af2e3ba93cf8" ], "author": { "name": "Friedrich Vock", "email": "friedrich.vock@gmx.de", "time": "Tue May 14 09:09:31 2024 +0200" }, "committer": { "name": "Daniel Borkmann", "email": "daniel@iogearbox.net", "time": "Fri May 24 17:12:12 2024 +0200" }, "message": "bpf: Fix potential integer overflow in resolve_btfids\n\nerr is a 32-bit integer, but elf_update returns an off_t, which is 64-bit\nat least on 64-bit platforms. If symbols_patch is called on a binary between\n2-4GB in size, the result will be negative when cast to a 32-bit integer,\nwhich the code assumes means an error occurred. This can wrongly trigger\nbuild failures when building very large kernel images.\n\nFixes: fbbb68de80a4 (\"bpf: Add resolve_btfids tool to resolve BTF IDs in ELF object\")\nSigned-off-by: Friedrich Vock \u003cfriedrich.vock@gmx.de\u003e\nSigned-off-by: Daniel Borkmann \u003cdaniel@iogearbox.net\u003e\nAcked-by: Daniel Borkmann \u003cdaniel@iogearbox.net\u003e\nLink: https://lore.kernel.org/bpf/20240514070931.199694-1-friedrich.vock@gmx.de\n" }, { "commit": "0b4f5add9fa59bfd42c1030f572db2e4c395181b", "tree": "feb6fd37a5bc3cb2183623e69247cc9b7c57755c", "parents": [ "128d54fbcb14b8717ecf596d3dbded327b9980b3", "83fea49f2711fc90c0d115b0ed04046b45155b65" ], "author": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Fri May 24 13:27:08 2024 +0100" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Fri May 24 13:27:08 2024 +0100" }, "message": "Merge branch \u0027mlx5-fixes\u0027\n\nTariq Toukan says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\nmlx5 fixes 24-05-22\n\nThis patchset provides bug fixes to mlx5 core and Eth drivers.\n\nSeries generated against:\ncommit 9c91c7fadb17 (\"net: mana: Fix the extra HZ in mana_hwc_send_request\")\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "83fea49f2711fc90c0d115b0ed04046b45155b65", "tree": "feb6fd37a5bc3cb2183623e69247cc9b7c57755c", "parents": [ "5c74195d5dd977e97556e6fa76909b831c241230" ], "author": { "name": "Gal Pressman", "email": "gal@nvidia.com", "time": "Wed May 22 22:26:59 2024 +0300" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Fri May 24 13:27:08 2024 +0100" }, "message": "net/mlx5e: Fix UDP GSO for encapsulated packets\n\nWhen the skb is encapsulated, adjust the inner UDP header instead of the\nouter one, and account for UDP header (instead of TCP) in the inline\nheader size calculation.\n\nFixes: 689adf0d4892 (\"net/mlx5e: Add UDP GSO support\")\nReported-by: Jason Baron \u003cjbaron@akamai.com\u003e\nCloses: https://lore.kernel.org/netdev/c42961cb-50b9-4a9a-bd43-87fe48d88d29@akamai.com/\nSigned-off-by: Gal Pressman \u003cgal@nvidia.com\u003e\nReviewed-by: Dragos Tatulea \u003cdtatulea@nvidia.com\u003e\nReviewed-by: Boris Pismenny \u003cborisp@nvidia.com\u003e\nSigned-off-by: Tariq Toukan \u003ctariqt@nvidia.com\u003e\nReviewed-by: Simon Horman \u003chorms@kernel.org\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "5c74195d5dd977e97556e6fa76909b831c241230", "tree": "a497c7221baaf996613dfd0a0f22c2bb4855f138", "parents": [ "f55cd31287e5f77f226c91d2f7756bafa0d583ed" ], "author": { "name": "Carolina Jubran", "email": "cjubran@nvidia.com", "time": "Wed May 22 22:26:58 2024 +0300" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Fri May 24 13:27:08 2024 +0100" }, "message": "net/mlx5e: Use rx_missed_errors instead of rx_dropped for reporting buffer exhaustion\n\nPreviously, the driver incorrectly used rx_dropped to report device\nbuffer exhaustion.\n\nAccording to the documentation, rx_dropped should not be used to count\npackets dropped due to buffer exhaustion, which is the purpose of\nrx_missed_errors.\n\nUse rx_missed_errors as intended for counting packets dropped due to\nbuffer exhaustion.\n\nFixes: 269e6b3af3bf (\"net/mlx5e: Report additional error statistics in get stats ndo\")\nSigned-off-by: Carolina Jubran \u003ccjubran@nvidia.com\u003e\nSigned-off-by: Tariq Toukan \u003ctariqt@nvidia.com\u003e\nReviewed-by: Simon Horman \u003chorms@kernel.org\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "f55cd31287e5f77f226c91d2f7756bafa0d583ed", "tree": "fdd32bb75dbbec885e5c572b0188d7d96ab1490b", "parents": [ "9a52f6d44f4521773b4699b4ed34b8e21d5a175c" ], "author": { "name": "Rahul Rameshbabu", "email": "rrameshbabu@nvidia.com", "time": "Wed May 22 22:26:57 2024 +0300" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Fri May 24 13:27:08 2024 +0100" }, "message": "net/mlx5e: Do not use ptp structure for tx ts stats when not initialized\n\nThe ptp channel instance is only initialized when ptp traffic is first\nprocessed by the driver. This means that there is a window in between when\nport timestamping is enabled and ptp traffic is sent where the ptp channel\ninstance is not initialized. Accessing statistics during this window will\nlead to an access violation (NULL + member offset). Check the validity of\nthe instance before attempting to query statistics.\n\n BUG: unable to handle page fault for address: 0000000000003524\n #PF: supervisor read access in kernel mode\n #PF: error_code(0x0000) - not-present page\n PGD 109dfc067 P4D 109dfc067 PUD 1064ef067 PMD 0\n Oops: 0000 [#1] SMP\n CPU: 0 PID: 420 Comm: ethtool Not tainted 6.9.0-rc2-rrameshbabu+ #245\n Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS Arch Linux 1.16.3-1-1 04/01/204\n RIP: 0010:mlx5e_stats_ts_get+0x4c/0x130\n \u003csnip\u003e\n Call Trace:\n \u003cTASK\u003e\n ? show_regs+0x60/0x70\n ? __die+0x24/0x70\n ? page_fault_oops+0x15f/0x430\n ? do_user_addr_fault+0x2c9/0x5c0\n ? exc_page_fault+0x63/0x110\n ? asm_exc_page_fault+0x27/0x30\n ? mlx5e_stats_ts_get+0x4c/0x130\n ? mlx5e_stats_ts_get+0x20/0x130\n mlx5e_get_ts_stats+0x15/0x20\n \u003csnip\u003e\n\nFixes: 3579032c08c1 (\"net/mlx5e: Implement ethtool hardware timestamping statistics\")\nSigned-off-by: Rahul Rameshbabu \u003crrameshbabu@nvidia.com\u003e\nSigned-off-by: Tariq Toukan \u003ctariqt@nvidia.com\u003e\nReviewed-by: Simon Horman \u003chorms@kernel.org\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "9a52f6d44f4521773b4699b4ed34b8e21d5a175c", "tree": "992aaaa30bb91781c5af475daad3421937cd09a4", "parents": [ "16d66a4fa81da07bc4ed19f4e53b87263c2f8d38" ], "author": { "name": "Rahul Rameshbabu", "email": "rrameshbabu@nvidia.com", "time": "Wed May 22 22:26:56 2024 +0300" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Fri May 24 13:27:08 2024 +0100" }, "message": "net/mlx5e: Fix IPsec tunnel mode offload feature check\n\nRemove faulty check disabling checksum offload and GSO for offload of\nsimple IPsec tunnel L4 traffic. Comment previously describing the deleted\ncode incorrectly claimed the check prevented double tunnel (or three layers\nof ip headers).\n\nFixes: f1267798c980 (\"net/mlx5: Fix checksum issue of VXLAN and IPsec crypto offload\")\nSigned-off-by: Rahul Rameshbabu \u003crrameshbabu@nvidia.com\u003e\nSigned-off-by: Tariq Toukan \u003ctariqt@nvidia.com\u003e\nReviewed-by: Simon Horman \u003chorms@kernel.org\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "16d66a4fa81da07bc4ed19f4e53b87263c2f8d38", "tree": "c2b0b099b5e2b9834321a668982485c83268ef15", "parents": [ "1b9f86c6d53245dab087f1b2c05727b5982142ff" ], "author": { "name": "Rahul Rameshbabu", "email": "rrameshbabu@nvidia.com", "time": "Wed May 22 22:26:55 2024 +0300" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Fri May 24 13:27:07 2024 +0100" }, "message": "net/mlx5: Use mlx5_ipsec_rx_status_destroy to correctly delete status rules\n\nrx_create no longer allocates a modify_hdr instance that needs to be\ncleaned up. The mlx5_modify_header_dealloc call will lead to a NULL pointer\ndereference. A leak in the rules also previously occurred since there are\nnow two rules populated related to status.\n\n BUG: kernel NULL pointer dereference, address: 0000000000000000\n #PF: supervisor read access in kernel mode\n #PF: error_code(0x0000) - not-present page\n PGD 109907067 P4D 109907067 PUD 116890067 PMD 0\n Oops: 0000 [#1] SMP\n CPU: 1 PID: 484 Comm: ip Not tainted 6.9.0-rc2-rrameshbabu+ #254\n Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS Arch Linux 1.16.3-1-1 04/01/2014\n RIP: 0010:mlx5_modify_header_dealloc+0xd/0x70\n \u003csnip\u003e\n Call Trace:\n \u003cTASK\u003e\n ? show_regs+0x60/0x70\n ? __die+0x24/0x70\n ? page_fault_oops+0x15f/0x430\n ? free_to_partial_list.constprop.0+0x79/0x150\n ? do_user_addr_fault+0x2c9/0x5c0\n ? exc_page_fault+0x63/0x110\n ? asm_exc_page_fault+0x27/0x30\n ? mlx5_modify_header_dealloc+0xd/0x70\n rx_create+0x374/0x590\n rx_add_rule+0x3ad/0x500\n ? rx_add_rule+0x3ad/0x500\n ? mlx5_cmd_exec+0x2c/0x40\n ? mlx5_create_ipsec_obj+0xd6/0x200\n mlx5e_accel_ipsec_fs_add_rule+0x31/0xf0\n mlx5e_xfrm_add_state+0x426/0xc00\n \u003csnip\u003e\n\nFixes: 94af50c0a9bb (\"net/mlx5e: Unify esw and normal IPsec status table creation/destruction\")\nSigned-off-by: Rahul Rameshbabu \u003crrameshbabu@nvidia.com\u003e\nSigned-off-by: Tariq Toukan \u003ctariqt@nvidia.com\u003e\nReviewed-by: Simon Horman \u003chorms@kernel.org\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "1b9f86c6d53245dab087f1b2c05727b5982142ff", "tree": "3b144c57e5285d4a22ef8cb287651da3fcb04d51", "parents": [ "fca3b4791850b7e2181f0b3195b66d53df83151b" ], "author": { "name": "Gal Pressman", "email": "gal@nvidia.com", "time": "Wed May 22 22:26:54 2024 +0300" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Fri May 24 13:27:07 2024 +0100" }, "message": "net/mlx5: Fix MTMP register capability offset in MCAM register\n\nThe MTMP register (0x900a) capability offset is off-by-one, move it to\nthe right place.\n\nFixes: 1f507e80c700 (\"net/mlx5: Expose NIC temperature via hardware monitoring kernel API\")\nSigned-off-by: Gal Pressman \u003cgal@nvidia.com\u003e\nReviewed-by: Cosmin Ratiu \u003ccratiu@nvidia.com\u003e\nSigned-off-by: Tariq Toukan \u003ctariqt@nvidia.com\u003e\nReviewed-by: Simon Horman \u003chorms@kernel.org\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "fca3b4791850b7e2181f0b3195b66d53df83151b", "tree": "7f7640722053e7da3f3e9d7984c4476fcba828ed", "parents": [ "51ef9305b8f40946d65c40368ffb4c14636d369a" ], "author": { "name": "Tariq Toukan", "email": "tariqt@nvidia.com", "time": "Wed May 22 22:26:53 2024 +0300" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Fri May 24 13:27:07 2024 +0100" }, "message": "net/mlx5: Do not query MPIR on embedded CPU function\n\nA proper query to MPIR needs to set the correct value in the depth field.\nOn embedded CPU this value is not necessarily zero. As there is no real\nuse case for multi-PF netdev on the embedded CPU of the smart NIC, block\nthis option.\n\nThis fixes the following failure:\nACCESS_REG(0x805) op_mod(0x1) failed, status bad system state(0x4), syndrome (0x685f19), err(-5)\n\nFixes: 678eb448055a (\"net/mlx5: SD, Implement basic query and instantiation\")\nSigned-off-by: Tariq Toukan \u003ctariqt@nvidia.com\u003e\nReviewed-by: Simon Horman \u003chorms@kernel.org\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "51ef9305b8f40946d65c40368ffb4c14636d369a", "tree": "31ebc6ce530805e764f40990e4178599da65d0d8", "parents": [ "128d54fbcb14b8717ecf596d3dbded327b9980b3" ], "author": { "name": "Maher Sanalla", "email": "msanalla@nvidia.com", "time": "Wed May 22 22:26:52 2024 +0300" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Fri May 24 13:27:07 2024 +0100" }, "message": "net/mlx5: Lag, do bond only if slaves agree on roce state\n\nCurrently, the driver does not enforce that lag bond slaves must have\nmatching roce capabilities. Yet, in mlx5_do_bond(), the driver attempts\nto enable roce on all vports of the bond slaves, causing the following\nsyndrome when one slave has no roce fw support:\n\nmlx5_cmd_out_err:809:(pid 25427): MODIFY_NIC_VPORT_CONTEXT(0×755) op_mod(0×0)\nfailed, status bad parameter(0×3), syndrome (0xc1f678), err(-22)\n\nThus, create HW lag only if bond\u0027s slaves agree on roce state,\neither all slaves have roce support resulting in a roce lag bond,\nor none do, resulting in a raw eth bond.\n\nFixes: 7907f23adc18 (\"net/mlx5: Implement RoCE LAG feature\")\nSigned-off-by: Maher Sanalla \u003cmsanalla@nvidia.com\u003e\nSigned-off-by: Tariq Toukan \u003ctariqt@nvidia.com\u003e\nReviewed-by: Simon Horman \u003chorms@kernel.org\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "128d54fbcb14b8717ecf596d3dbded327b9980b3", "tree": "c06ce26072664fb47fd2ea9fa7e38637dabb68ac", "parents": [ "06e785aeb9ea8a43d0a3967c1ba6e69d758e82d4" ], "author": { "name": "Mathieu Othacehe", "email": "othacehe@gnu.org", "time": "Tue May 21 08:54:06 2024 +0200" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Fri May 24 12:30:38 2024 +0100" }, "message": "net: phy: micrel: set soft_reset callback to genphy_soft_reset for KSZ8061\n\nFollowing a similar reinstate for the KSZ8081 and KSZ9031.\n\nOlder kernels would use the genphy_soft_reset if the PHY did not implement\na .soft_reset.\n\nThe KSZ8061 errata described here:\nhttps://ww1.microchip.com/downloads/en/DeviceDoc/KSZ8061-Errata-DS80000688B.pdf\nand worked around with 232ba3a51c (\"net: phy: Micrel KSZ8061: link failure after cable connect\")\nis back again without this soft reset.\n\nFixes: 6e2d85ec0559 (\"net: phy: Stop with excessive soft reset\")\nTested-by: Karim Ben Houcine \u003ckarim.benhoucine@landisgyr.com\u003e\nSigned-off-by: Mathieu Othacehe \u003cothacehe@gnu.org\u003e\nReviewed-by: Andrew Lunn \u003candrew@lunn.ch\u003e\nReviewed-by: Florian Fainelli \u003cflorian.fainelli@broadcom.com\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "06e785aeb9ea8a43d0a3967c1ba6e69d758e82d4", "tree": "7335dc1b362161ad0bd0438488ed56b8fabc5032", "parents": [ "66ad4829ddd0b5540dc0b076ef2818e89c8f720e" ], "author": { "name": "Matt Jan", "email": "zoo868e@gmail.com", "time": "Tue May 14 12:10:46 2024 +0800" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Fri May 24 10:36:55 2024 +0100" }, "message": "connector: Fix invalid conversion in cn_proc.h\n\nThe implicit conversion from unsigned int to enum\nproc_cn_event is invalid, so explicitly cast it\nfor compilation in a C++ compiler.\n/usr/include/linux/cn_proc.h: In function \u0027proc_cn_event valid_event(proc_cn_event)\u0027:\n/usr/include/linux/cn_proc.h:72:17: error: invalid conversion from \u0027unsigned int\u0027 to \u0027proc_cn_event\u0027 [-fpermissive]\n 72 | ev_type \u0026\u003d PROC_EVENT_ALL;\n | ^\n | |\n | unsigned int\n\nSigned-off-by: Matt Jan \u003czoo868e@gmail.com\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "66ad4829ddd0b5540dc0b076ef2818e89c8f720e", "tree": "628e2f73202a3e9e22865e8cd8f69db8692edfd3", "parents": [ "404001ddf3e40369510a08b8b7e895c2da692b79", "c71e3a5cffd5309d7f84444df03d5b72600cc417" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 23 12:49:37 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 23 12:49:37 2024 -0700" }, "message": "Merge tag \u0027net-6.10-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net\n\nPull networking fixes from Paolo Abeni:\n \"Quite smaller than usual. Notably it includes the fix for the unix\n regression from the past weeks. The TCP window fix will require some\n follow-up, already queued.\n\n Current release - regressions:\n\n - af_unix: fix garbage collection of embryos\n\n Previous releases - regressions:\n\n - af_unix: fix race between GC and receive path\n\n - ipv6: sr: fix missing sk_buff release in seg6_input_core\n\n - tcp: remove 64 KByte limit for initial tp-\u003ercv_wnd value\n\n - eth: r8169: fix rx hangup\n\n - eth: lan966x: remove ptp traps in case the ptp is not enabled\n\n - eth: ixgbe: fix link breakage vs cisco switches\n\n - eth: ice: prevent ethtool from corrupting the channels\n\n Previous releases - always broken:\n\n - openvswitch: set the skbuff pkt_type for proper pmtud support\n\n - tcp: Fix shift-out-of-bounds in dctcp_update_alpha()\n\n Misc:\n\n - a bunch of selftests stabilization patches\"\n\n* tag \u0027net-6.10-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (25 commits)\n r8169: Fix possible ring buffer corruption on fragmented Tx packets.\n idpf: Interpret .set_channels() input differently\n ice: Interpret .set_channels() input differently\n nfc: nci: Fix handling of zero-length payload packets in nci_rx_work()\n net: relax socket state check at accept time.\n tcp: remove 64 KByte limit for initial tp-\u003ercv_wnd value\n net: ti: icssg_prueth: Fix NULL pointer dereference in prueth_probe()\n tls: fix missing memory barrier in tls_init\n net: fec: avoid lock evasion when reading pps_enable\n Revert \"ixgbe: Manual AN-37 for troublesome link partners for X550 SFI\"\n testing: net-drv: use stats64 for testing\n net: mana: Fix the extra HZ in mana_hwc_send_request\n net: lan966x: Remove ptp traps in case the ptp is not enabled.\n openvswitch: Set the skbuff pkt_type for proper pmtud support.\n selftest: af_unix: Make SCM_RIGHTS into OOB data.\n af_unix: Fix garbage collection of embryos carrying OOB with SCM_RIGHTS\n tcp: Fix shift-out-of-bounds in dctcp_update_alpha().\n selftests/net: use tc rule to filter the na packet\n ipv6: sr: fix memleak in seg6_hmac_init_algo\n af_unix: Update unix_sk(sk)-\u003eoob_skb under sk_receive_queue lock.\n ...\n" }, { "commit": "404001ddf3e40369510a08b8b7e895c2da692b79", "tree": "48803b50a62640da7ffb046ca0e32ac378574289", "parents": [ "e82d2af501072930983bfa7a0d200fc8aeb4f63c", "1e8b7b3dbb3103d577a586ca72bc329f7b67120b" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 23 12:36:38 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 23 12:36:38 2024 -0700" }, "message": "Merge tag \u0027trace-fixes-v6.10\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace\n\nPull tracing fixes from Steven Rostedt:\n \"Minor last minute fixes:\n\n - Fix a very tight race between the ring buffer readers and resizing\n the ring buffer\n\n - Correct some stale comments in the ring buffer code\n\n - Fix kernel-doc in the rv code\n\n - Add a MODULE_DESCRIPTION to preemptirq_delay_test\"\n\n* tag \u0027trace-fixes-v6.10\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:\n rv: Update rv_en(dis)able_monitor doc to match kernel-doc\n tracing: Add MODULE_DESCRIPTION() to preemptirq_delay_test\n ring-buffer: Fix a race between readers and resize checks\n ring-buffer: Correct stale comments related to non-consuming readers\n" }, { "commit": "e82d2af501072930983bfa7a0d200fc8aeb4f63c", "tree": "c46d7c872de1288cc50a3a67743ed58767797be7", "parents": [ "d6a326d694e77fd1d90ae40ad3c73a6af0eb1b8d", "df73757cf8f66fa54c4721c53b0916af3c4d9818" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 23 12:32:15 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 23 12:32:15 2024 -0700" }, "message": "Merge tag \u0027trace-tools-v6.10-2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace\n\nPull tracing tool fix from Steven Rostedt:\n \"Fix printf format warnings in latency-collector.\n\n Use the printf format string with %s to take a string instead of\n taking in a string directly\"\n\n* tag \u0027trace-tools-v6.10-2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:\n tools/latency-collector: Fix -Wformat-security compile warns\n" }, { "commit": "d6a326d694e77fd1d90ae40ad3c73a6af0eb1b8d", "tree": "65620f42b46c745e16298f31b45d54b33ec0ebea", "parents": [ "bca2a25d3be779fbdf912de8e1ded6dcd483845d", "2c92ca849fcc6ee7d0c358e9959abc9f58661aea" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 23 12:28:01 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 23 12:28:01 2024 -0700" }, "message": "Merge tag \u0027trace-assign-str-v6.10\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace\n\nPull tracing cleanup from Steven Rostedt:\n \"Remove second argument of __assign_str()\n\n The __assign_str() macro logic of the TRACE_EVENT() macro was\n optimized so that it no longer needs the second argument. The\n __assign_str() is always matched with __string() field that takes a\n field name and the source for that field:\n\n __string(field, source)\n\n The TRACE_EVENT() macro logic will save off the source value and then\n use that value to copy into the ring buffer via the __assign_str().\n\n Before commit c1fa617caeb0 (\"tracing: Rework __assign_str() and\n __string() to not duplicate getting the string\"), the __assign_str()\n needed the second argument which would perform the same logic as the\n __string() source parameter did. Not only would this add overhead, but\n it was error prone as if the __assign_str() source produced something\n different, it may not have allocated enough for the string in the ring\n buffer (as the __string() source was used to determine how much to\n allocate)\n\n Now that the __assign_str() just uses the same string that was used in\n __string() it no longer needs the source parameter. It can now be\n removed\"\n\n* tag \u0027trace-assign-str-v6.10\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:\n tracing/treewide: Remove second parameter of __assign_str()\n" }, { "commit": "bca2a25d3be779fbdf912de8e1ded6dcd483845d", "tree": "426c75e626ace0c5e47d40be48c4be9daa380a1f", "parents": [ "2b7ced108e93b837f152841ac1f0bf45ed2a6b21", "1c9e709cde80fb612e07d9503ad04457e8a58da2" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 23 12:22:20 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 23 12:22:20 2024 -0700" }, "message": "Merge tag \u0027sparc-for-6.10-tag1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/alarsson/linux-sparc\n\nPull sparc updates from Andreas Larsson:\n\n - Avoid on-stack cpumask variables in a number of places\n\n - Move struct termio to asm/termios.h, matching other architectures and\n allowing certain user space applications to build also for sparc\n\n - Fix missing prototype warnings for sparc64\n\n - Fix version generation warnings for sparc32\n\n - Fix bug where non-consecutive CPU IDs lead to some CPUs not starting\n\n - Simplification using swap and cleanup using NULL for pointer\n\n - Convert sparc parport and chmc drivers to use remove callbacks\n returning void\n\n* tag \u0027sparc-for-6.10-tag1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/alarsson/linux-sparc:\n sparc/leon: Remove on-stack cpumask var\n sparc/pci_msi: Remove on-stack cpumask var\n sparc/of: Remove on-stack cpumask var\n sparc/irq: Remove on-stack cpumask var\n sparc/srmmu: Remove on-stack cpumask var\n sparc: chmc: Convert to platform remove callback returning void\n sparc: parport: Convert to platform remove callback returning void\n sparc: Compare pointers to NULL instead of 0\n sparc: Use swap() to fix Coccinelle warning\n sparc32: Fix version generation failed warnings\n sparc64: Fix number of online CPUs\n sparc64: Fix prototype warning for sched_clock\n sparc64: Fix prototype warnings in adi_64.c\n sparc64: Fix prototype warning for dma_4v_iotsb_bind\n sparc64: Fix prototype warning for uprobe_trap\n sparc64: Fix prototype warning for alloc_irqstack_bootmem\n sparc64: Fix prototype warning for vmemmap_free\n sparc64: Fix prototype warnings in traps_64.c\n sparc64: Fix prototype warning for init_vdso_image\n sparc: move struct termio to asm/termios.h\n" }, { "commit": "2b7ced108e93b837f152841ac1f0bf45ed2a6b21", "tree": "84ab89bc9f4b9aeea5aeb4cdda438333333b37d5", "parents": [ "2ef32ad2241340565c35baf77fc95053c84eeeb0", "e92bee9f861b466c676f0200be3e46af7bc4ac6b" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 23 12:09:22 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 23 12:09:22 2024 -0700" }, "message": "Merge tag \u0027arm64-fixes\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux\n\nPull arm64 fixes from Will Deacon:\n \"The major fix here is for a filesystem corruption issue reported on\n Apple M1 as a result of buggy management of the floating point\n register state introduced in 6.8. I initially reverted one of the\n offending patches, but in the end Ard cooked a proper fix so there\u0027s a\n revert+reapply in the series.\n\n Aside from that, we\u0027ve got some CPU errata workarounds and misc other\n fixes.\n\n - Fix broken FP register state tracking which resulted in filesystem\n corruption when dm-crypt is used\n\n - Workarounds for Arm CPU errata affecting the SSBS Spectre\n mitigation\n\n - Fix lockdep assertion in DMC620 memory controller PMU driver\n\n - Fix alignment of BUG table when CONFIG_DEBUG_BUGVERBOSE is\n disabled\"\n\n* tag \u0027arm64-fixes\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:\n arm64/fpsimd: Avoid erroneous elide of user state reload\n Reapply \"arm64: fpsimd: Implement lazy restore for kernel mode FPSIMD\"\n arm64: asm-bug: Add .align 2 to the end of __BUG_ENTRY\n perf/arm-dmc620: Fix lockdep assert in -\u003eevent_init()\n Revert \"arm64: fpsimd: Implement lazy restore for kernel mode FPSIMD\"\n arm64: errata: Add workaround for Arm errata 3194386 and 3312417\n arm64: cputype: Add Neoverse-V3 definitions\n arm64: cputype: Add Cortex-X4 definitions\n arm64: barrier: Restore spec_bar() macro\n" }, { "commit": "2ef32ad2241340565c35baf77fc95053c84eeeb0", "tree": "672ed232affb09035e73e1ba3411be7f71d3289a", "parents": [ "c760b3725e52403dc1b28644fb09c47a83cacea6", "c8fae27d141a32a1624d0d0d5419d94252824498" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 23 12:04:36 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 23 12:04:36 2024 -0700" }, "message": "Merge tag \u0027for_linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost\n\nPull virtio updates from Michael Tsirkin:\n \"Several new features here:\n\n - virtio-net is finally supported in vduse\n\n - virtio (balloon and mem) interaction with suspend is improved\n\n - vhost-scsi now handles signals better/faster\n\n And fixes, cleanups all over the place\"\n\n* tag \u0027for_linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost: (48 commits)\n virtio-pci: Check if is_avq is NULL\n virtio: delete vq in vp_find_vqs_msix() when request_irq() fails\n MAINTAINERS: add Eugenio Pérez as reviewer\n vhost-vdpa: Remove usage of the deprecated ida_simple_xx() API\n vp_vdpa: don\u0027t allocate unused msix vectors\n sound: virtio: drop owner assignment\n fuse: virtio: drop owner assignment\n scsi: virtio: drop owner assignment\n rpmsg: virtio: drop owner assignment\n nvdimm: virtio_pmem: drop owner assignment\n wifi: mac80211_hwsim: drop owner assignment\n vsock/virtio: drop owner assignment\n net: 9p: virtio: drop owner assignment\n net: virtio: drop owner assignment\n net: caif: virtio: drop owner assignment\n misc: nsm: drop owner assignment\n iommu: virtio: drop owner assignment\n drm/virtio: drop owner assignment\n gpio: virtio: drop owner assignment\n firmware: arm_scmi: virtio: drop owner assignment\n ...\n" }, { "commit": "df73757cf8f66fa54c4721c53b0916af3c4d9818", "tree": "0ba29096853e2965c02e229b30472f06ba80708b", "parents": [ "59c22f70b2951d81de410d477ae536ba951b4f37" ], "author": { "name": "Shuah Khan", "email": "skhan@linuxfoundation.org", "time": "Wed Apr 03 19:10:09 2024 -0600" }, "committer": { "name": "Steven Rostedt (Google)", "email": "rostedt@goodmis.org", "time": "Thu May 23 10:46:01 2024 -0400" }, "message": "tools/latency-collector: Fix -Wformat-security compile warns\n\nFix the following -Wformat-security compile warnings adding missing\nformat arguments:\n\nlatency-collector.c: In function ‘show_available’:\nlatency-collector.c:938:17: warning: format not a string literal and\nno format arguments [-Wformat-security]\n 938 | warnx(no_tracer_msg);\n | ^~~~~\n\nlatency-collector.c:943:17: warning: format not a string literal and\nno format arguments [-Wformat-security]\n 943 | warnx(no_latency_tr_msg);\n | ^~~~~\n\nlatency-collector.c: In function ‘find_default_tracer’:\nlatency-collector.c:986:25: warning: format not a string literal and\nno format arguments [-Wformat-security]\n 986 | errx(EXIT_FAILURE, no_tracer_msg);\n |\n ^~~~\nlatency-collector.c: In function ‘scan_arguments’:\nlatency-collector.c:1881:33: warning: format not a string literal and\nno format arguments [-Wformat-security]\n 1881 | errx(EXIT_FAILURE, no_tracer_msg);\n | ^~~~\n\nLink: https://lore.kernel.org/linux-trace-kernel/20240404011009.32945-1-skhan@linuxfoundation.org\n\nCc: stable@vger.kernel.org\nFixes: e23db805da2df (\"tracing/tools: Add the latency-collector to tools directory\")\nSigned-off-by: Shuah Khan \u003cskhan@linuxfoundation.org\u003e\nSigned-off-by: Steven Rostedt (Google) \u003crostedt@goodmis.org\u003e\n" }, { "commit": "c71e3a5cffd5309d7f84444df03d5b72600cc417", "tree": "db28cb32550f6e6f436ebb215cc495079400ff48", "parents": [ "3d8597d8d7d9b3faffe0f2361032123ee6c09c02" ], "author": { "name": "Ken Milmore", "email": "ken.milmore@gmail.com", "time": "Tue May 21 23:45:50 2024 +0100" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Thu May 23 15:45:32 2024 +0200" }, "message": "r8169: Fix possible ring buffer corruption on fragmented Tx packets.\n\nAn issue was found on the RTL8125b when transmitting small fragmented\npackets, whereby invalid entries were inserted into the transmit ring\nbuffer, subsequently leading to calls to dma_unmap_single() with a null\naddress.\n\nThis was caused by rtl8169_start_xmit() not noticing changes to nr_frags\nwhich may occur when small packets are padded (to work around hardware\nquirks) in rtl8169_tso_csum_v2().\n\nTo fix this, postpone inspecting nr_frags until after any padding has been\napplied.\n\nFixes: 9020845fb5d6 (\"r8169: improve rtl8169_start_xmit\")\nCc: stable@vger.kernel.org\nSigned-off-by: Ken Milmore \u003cken.milmore@gmail.com\u003e\nReviewed-by: Heiner Kallweit \u003chkallweit1@gmail.com\u003e\nLink: https://lore.kernel.org/r/27ead18b-c23d-4f49-a020-1fc482c5ac95@gmail.com\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "3d8597d8d7d9b3faffe0f2361032123ee6c09c02", "tree": "ca0f475835937dc836ff51b6e774b3113de991e6", "parents": [ "6671e352497ca4bb07a96c48e03907065ff77d8a", "5e7695e0219bf6acb96081af3ba0ca08b1829656" ], "author": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Thu May 23 13:02:28 2024 +0200" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Thu May 23 13:02:28 2024 +0200" }, "message": "Merge branch \u0027intel-interpret-set_channels-input-differently\u0027\n\nJacob Keller says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\nintel: Interpret .set_channels() input differently\n\nThe ice and idpf drivers can trigger a crash with AF_XDP due to incorrect\ninterpretation of the asymmetric Tx and Rx parameters in their\n.set_channels() implementations:\n\n1. ethtool -l \u003cIFNAME\u003e -\u003e combined: 40\n2. Attach AF_XDP to queue 30\n3. ethtool -L \u003cIFNAME\u003e rx 15 tx 15\n combined number is not specified, so command becomes {rx_count \u003d 15,\n tx_count \u003d 15, combined_count \u003d 40}.\n4. ethnl_set_channels checks, if there are any AF_XDP of queues from the\n new (combined_count + rx_count) to the old one, so from 55 to 40, check\n does not trigger.\n5. the driver interprets `rx 15 tx 15` as 15 combined channels and deletes\n the queue that AF_XDP is attached to.\n\nThis is fundamentally a problem with interpreting a request for asymmetric\nqueues as symmetric combined queues.\n\nFix the ice and idpf drivers to stop interpreting such requests as a\nrequest for combined queues. Due to current driver design for both ice and\nidpf, it is not possible to support requests of the same count of Tx and Rx\nqueues with independent interrupts, (i.e. ethtool -L \u003cIFNAME\u003e rx 15 tx 15)\nso such requests are now rejected.\n\nSigned-off-by: Jacob Keller \u003cjacob.e.keller@intel.com\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://lore.kernel.org/r/20240521-iwl-net-2024-05-14-set-channels-fixes-v2-0-7aa39e2e99f1@intel.com\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "5e7695e0219bf6acb96081af3ba0ca08b1829656", "tree": "ca0f475835937dc836ff51b6e774b3113de991e6", "parents": [ "05d6f442f31f901d27dbc64fd504a8ec7d5013de" ], "author": { "name": "Larysa Zaremba", "email": "larysa.zaremba@intel.com", "time": "Tue May 21 12:39:54 2024 -0700" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Thu May 23 13:02:26 2024 +0200" }, "message": "idpf: Interpret .set_channels() input differently\n\nUnlike ice, idpf does not check, if user has requested at least 1 combined\nchannel. Instead, it relies on a check in the core code. Unfortunately, the\ncheck does not trigger for us because of the hacky .set_channels()\ninterpretation logic that is not consistent with the core code.\n\nThis naturally leads to user being able to trigger a crash with an invalid\ninput. This is how:\n\n1. ethtool -l \u003cIFNAME\u003e -\u003e combined: 40\n2. ethtool -L \u003cIFNAME\u003e rx 0 tx 0\n combined number is not specified, so command becomes {rx_count \u003d 0,\n tx_count \u003d 0, combined_count \u003d 40}.\n3. ethnl_set_channels checks, if there is at least 1 RX and 1 TX channel,\n comparing (combined_count + rx_count) and (combined_count + tx_count)\n to zero. Obviously, (40 + 0) is greater than zero, so the core code\n deems the input OK.\n4. idpf interprets `rx 0 tx 0` as 0 channels and tries to proceed with such\n configuration.\n\nThe issue has to be solved fundamentally, as current logic is also known to\ncause AF_XDP problems in ice [0].\n\nInterpret the command in a way that is more consistent with ethtool\nmanual [1] (--show-channels and --set-channels) and new ice logic.\n\nConsidering that in the idpf driver only the difference between RX and TX\nqueues forms dedicated channels, change the correct way to set number of\nchannels to:\n\nethtool -L \u003cIFNAME\u003e combined 10 /* For symmetric queues */\nethtool -L \u003cIFNAME\u003e combined 8 tx 2 rx 0 /* For asymmetric queues */\n\n[0] https://lore.kernel.org/netdev/20240418095857.2827-1-larysa.zaremba@intel.com/\n[1] https://man7.org/linux/man-pages/man8/ethtool.8.html\n\nFixes: 02cbfba1add5 (\"idpf: add ethtool callbacks\")\nReviewed-by: Przemek Kitszel \u003cprzemyslaw.kitszel@intel.com\u003e\nReviewed-by: Igor Bagnucki \u003cigor.bagnucki@intel.com\u003e\nSigned-off-by: Larysa Zaremba \u003clarysa.zaremba@intel.com\u003e\nTested-by: Krishneil Singh \u003ckrishneil.k.singh@intel.com\u003e\nReviewed-by: Simon Horman \u003chorms@kernel.org\u003e\nSigned-off-by: Jacob Keller \u003cjacob.e.keller@intel.com\u003e\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "05d6f442f31f901d27dbc64fd504a8ec7d5013de", "tree": "3f39aa5555dfc14f555ce06d1c64725f8dc453c6", "parents": [ "6671e352497ca4bb07a96c48e03907065ff77d8a" ], "author": { "name": "Larysa Zaremba", "email": "larysa.zaremba@intel.com", "time": "Tue May 21 12:39:53 2024 -0700" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Thu May 23 13:02:26 2024 +0200" }, "message": "ice: Interpret .set_channels() input differently\n\nA bug occurs because a safety check guarding AF_XDP-related queues in\nethnl_set_channels(), does not trigger. This happens, because kernel and\nice driver interpret the ethtool command differently.\n\nHow the bug occurs:\n1. ethtool -l \u003cIFNAME\u003e -\u003e combined: 40\n2. Attach AF_XDP to queue 30\n3. ethtool -L \u003cIFNAME\u003e rx 15 tx 15\n combined number is not specified, so command becomes {rx_count \u003d 15,\n tx_count \u003d 15, combined_count \u003d 40}.\n4. ethnl_set_channels checks, if there are any AF_XDP of queues from the\n new (combined_count + rx_count) to the old one, so from 55 to 40, check\n does not trigger.\n5. ice interprets `rx 15 tx 15` as 15 combined channels and deletes the\n queue that AF_XDP is attached to.\n\nInterpret the command in a way that is more consistent with ethtool\nmanual [0] (--show-channels and --set-channels).\n\nConsidering that in the ice driver only the difference between RX and TX\nqueues forms dedicated channels, change the correct way to set number of\nchannels to:\n\nethtool -L \u003cIFNAME\u003e combined 10 /* For symmetric queues */\nethtool -L \u003cIFNAME\u003e combined 8 tx 2 rx 0 /* For asymmetric queues */\n\n[0] https://man7.org/linux/man-pages/man8/ethtool.8.html\n\nFixes: 87324e747fde (\"ice: Implement ethtool ops for channels\")\nReviewed-by: Michal Swiatkowski \u003cmichal.swiatkowski@linux.intel.com\u003e\nSigned-off-by: Larysa Zaremba \u003clarysa.zaremba@intel.com\u003e\nTested-by: Chandan Kumar Rout \u003cchandanx.rout@intel.com\u003e\nTested-by: Pucha Himasekhar Reddy \u003chimasekharx.reddy.pucha@intel.com\u003e\nAcked-by: Maciej Fijalkowski \u003cmaciej.fijalkowski@intel.com\u003e\nSigned-off-by: Jacob Keller \u003cjacob.e.keller@intel.com\u003e\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "6671e352497ca4bb07a96c48e03907065ff77d8a", "tree": "0cead9afb23c2acdd83ae94df018889fe58c96e4", "parents": [ "26afda78cda3da974fd4c287962c169e9462c495" ], "author": { "name": "Ryosuke Yasuoka", "email": "ryasuoka@redhat.com", "time": "Wed May 22 00:34:42 2024 +0900" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Thu May 23 12:39:44 2024 +0200" }, "message": "nfc: nci: Fix handling of zero-length payload packets in nci_rx_work()\n\nWhen nci_rx_work() receives a zero-length payload packet, it should not\ndiscard the packet and exit the loop. Instead, it should continue\nprocessing subsequent packets.\n\nFixes: d24b03535e5e (\"nfc: nci: Fix uninit-value in nci_dev_up and nci_ntf_packet\")\nSigned-off-by: Ryosuke Yasuoka \u003cryasuoka@redhat.com\u003e\nReviewed-by: Simon Horman \u003chorms@kernel.org\u003e\nReviewed-by: Krzysztof Kozlowski \u003ckrzysztof.kozlowski@linaro.org\u003e\nLink: https://lore.kernel.org/r/20240521153444.535399-1-ryasuoka@redhat.com\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "26afda78cda3da974fd4c287962c169e9462c495", "tree": "0fe6e9adfab09fc680b3228143d7b1b0a55aa6f2", "parents": [ "378979e94e953c2070acb4f0e0c98d29260bd09d" ], "author": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Tue May 21 16:01:00 2024 +0200" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Thu May 23 12:33:35 2024 +0200" }, "message": "net: relax socket state check at accept time.\n\nChristoph reported the following splat:\n\nWARNING: CPU: 1 PID: 772 at net/ipv4/af_inet.c:761 __inet_accept+0x1f4/0x4a0\nModules linked in:\nCPU: 1 PID: 772 Comm: syz-executor510 Not tainted 6.9.0-rc7-g7da7119fe22b #56\nHardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.11.0-2.el7 04/01/2014\nRIP: 0010:__inet_accept+0x1f4/0x4a0 net/ipv4/af_inet.c:759\nCode: 04 38 84 c0 0f 85 87 00 00 00 41 c7 04 24 03 00 00 00 48 83 c4 10 5b 41 5c 41 5d 41 5e 41 5f 5d c3 cc cc cc cc e8 ec b7 da fd \u003c0f\u003e 0b e9 7f fe ff ff e8 e0 b7 da fd 0f 0b e9 fe fe ff ff 89 d9 80\nRSP: 0018:ffffc90000c2fc58 EFLAGS: 00010293\nRAX: ffffffff836bdd14 RBX: 0000000000000000 RCX: ffff888104668000\nRDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000000\nRBP: dffffc0000000000 R08: ffffffff836bdb89 R09: fffff52000185f64\nR10: dffffc0000000000 R11: fffff52000185f64 R12: dffffc0000000000\nR13: 1ffff92000185f98 R14: ffff88810754d880 R15: ffff8881007b7800\nFS: 000000001c772880(0000) GS:ffff88811b280000(0000) knlGS:0000000000000000\nCS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033\nCR2: 00007fb9fcf2e178 CR3: 00000001045d2002 CR4: 0000000000770ef0\nDR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000\nDR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400\nPKRU: 55555554\nCall Trace:\n \u003cTASK\u003e\n inet_accept+0x138/0x1d0 net/ipv4/af_inet.c:786\n do_accept+0x435/0x620 net/socket.c:1929\n __sys_accept4_file net/socket.c:1969 [inline]\n __sys_accept4+0x9b/0x110 net/socket.c:1999\n __do_sys_accept net/socket.c:2016 [inline]\n __se_sys_accept net/socket.c:2013 [inline]\n __x64_sys_accept+0x7d/0x90 net/socket.c:2013\n do_syscall_x64 arch/x86/entry/common.c:52 [inline]\n do_syscall_64+0x58/0x100 arch/x86/entry/common.c:83\n entry_SYSCALL_64_after_hwframe+0x76/0x7e\nRIP: 0033:0x4315f9\nCode: fd ff 48 81 c4 80 00 00 00 e9 f1 fe ff ff 0f 1f 00 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 \u003c48\u003e 3d 01 f0 ff ff 0f 83 ab b4 fd ff c3 66 2e 0f 1f 84 00 00 00 00\nRSP: 002b:00007ffdb26d9c78 EFLAGS: 00000246 ORIG_RAX: 000000000000002b\nRAX: ffffffffffffffda RBX: 0000000000400300 RCX: 00000000004315f9\nRDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000004\nRBP: 00000000006e1018 R08: 0000000000400300 R09: 0000000000400300\nR10: 0000000000400300 R11: 0000000000000246 R12: 0000000000000000\nR13: 000000000040cdf0 R14: 000000000040ce80 R15: 0000000000000055\n \u003c/TASK\u003e\n\nThe reproducer invokes shutdown() before entering the listener status.\nAfter commit 94062790aedb (\"tcp: defer shutdown(SEND_SHUTDOWN) for\nTCP_SYN_RECV sockets\"), the above causes the child to reach the accept\nsyscall in FIN_WAIT1 status.\n\nEric noted we can relax the existing assertion in __inet_accept()\n\nReported-by: Christoph Paasch \u003ccpaasch@apple.com\u003e\nCloses: https://github.com/multipath-tcp/mptcp_net-next/issues/490\nSuggested-by: Eric Dumazet \u003cedumazet@google.com\u003e\nFixes: 94062790aedb (\"tcp: defer shutdown(SEND_SHUTDOWN) for TCP_SYN_RECV sockets\")\nReviewed-by: Eric Dumazet \u003cedumazet@google.com\u003e\nLink: https://lore.kernel.org/r/23ab880a44d8cfd967e84de8b93dbf48848e3d8c.1716299669.git.pabeni@redhat.com\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "378979e94e953c2070acb4f0e0c98d29260bd09d", "tree": "72053c2990a0e15b9a534333f83f3fbbfba679a4", "parents": [ "b31c7e78086127a7fcaa761e8d336ee855a920c6" ], "author": { "name": "Jason Xing", "email": "kernelxing@tencent.com", "time": "Tue May 21 21:42:20 2024 +0800" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Thu May 23 12:21:17 2024 +0200" }, "message": "tcp: remove 64 KByte limit for initial tp-\u003ercv_wnd value\n\nRecently, we had some servers upgraded to the latest kernel and noticed\nthe indicator from the user side showed worse results than before. It is\ncaused by the limitation of tp-\u003ercv_wnd.\n\nIn 2018 commit a337531b942b (\"tcp: up initial rmem to 128KB and SYN rwin\nto around 64KB\") limited the initial value of tp-\u003ercv_wnd to 65535, most\nCDN teams would not benefit from this change because they cannot have a\nlarge window to receive a big packet, which will be slowed down especially\nin long RTT. Small rcv_wnd means slow transfer speed, to some extent. It\u0027s\nthe side effect for the latency/time-sensitive users.\n\nTo avoid future confusion, current change doesn\u0027t affect the initial\nreceive window on the wire in a SYN or SYN+ACK packet which are set within\n65535 bytes according to RFC 7323 also due to the limit in\n__tcp_transmit_skb():\n\n th-\u003ewindow \u003d htons(min(tp-\u003ercv_wnd, 65535U));\n\nIn one word, __tcp_transmit_skb() already ensures that constraint is\nrespected, no matter how large tp-\u003ercv_wnd is. The change doesn\u0027t violate\nRFC.\n\nLet me provide one example if with or without the patch:\nBefore:\nclient --- SYN: rwindow\u003d65535 ---\u003e server\nclient \u003c--- SYN+ACK: rwindow\u003d65535 ---- server\nclient --- ACK: rwindow\u003d65536 ---\u003e server\nNote: for the last ACK, the calculation is 512 \u003c\u003c 7.\n\nAfter:\nclient --- SYN: rwindow\u003d65535 ---\u003e server\nclient \u003c--- SYN+ACK: rwindow\u003d65535 ---- server\nclient --- ACK: rwindow\u003d175232 ---\u003e server\nNote: I use the following command to make it work:\nip route change default via [ip] dev eth0 metric 100 initrwnd 120\nFor the last ACK, the calculation is 1369 \u003c\u003c 7.\n\nWhen we apply such a patch, having a large rcv_wnd if the user tweak this\nknob can help transfer data more rapidly and save some rtts.\n\nFixes: a337531b942b (\"tcp: up initial rmem to 128KB and SYN rwin to around 64KB\")\nSigned-off-by: Jason Xing \u003ckernelxing@tencent.com\u003e\nReviewed-by: Eric Dumazet \u003cedumazet@google.com\u003e\nAcked-by: Neal Cardwell \u003cncardwell@google.com\u003e\nLink: https://lore.kernel.org/r/20240521134220.12510-1-kerneljasonxing@gmail.com\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "b31c7e78086127a7fcaa761e8d336ee855a920c6", "tree": "85d0ff5213c596a6ab68ea75f3aa8341018d89cf", "parents": [ "91e61dd7a0af660408e87372d8330ceb218be302" ], "author": { "name": "Romain Gantois", "email": "romain.gantois@bootlin.com", "time": "Tue May 21 14:44:11 2024 +0200" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Thu May 23 12:11:09 2024 +0200" }, "message": "net: ti: icssg_prueth: Fix NULL pointer dereference in prueth_probe()\n\nIn the prueth_probe() function, if one of the calls to emac_phy_connect()\nfails due to of_phy_connect() returning NULL, then the subsequent call to\nphy_attached_info() will dereference a NULL pointer.\n\nCheck the return code of emac_phy_connect and fail cleanly if there is an\nerror.\n\nFixes: 128d5874c082 (\"net: ti: icssg-prueth: Add ICSSG ethernet driver\")\nCc: stable@vger.kernel.org\nSigned-off-by: Romain Gantois \u003cromain.gantois@bootlin.com\u003e\nReviewed-by: Simon Horman \u003chorms@kernel.org\u003e\nReviewed-by: MD Danish Anwar \u003cdanishanwar@ti.com\u003e\nLink: https://lore.kernel.org/r/20240521-icssg-prueth-fix-v1-1-b4b17b1433e9@bootlin.com\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "91e61dd7a0af660408e87372d8330ceb218be302", "tree": "69bc4696c019ef04ef22c34400827fd86e7ffbbe", "parents": [ "3b1c92f8e5371700fada307cc8fd2c51fa7bc8c1" ], "author": { "name": "Dae R. Jeong", "email": "threeearcat@gmail.com", "time": "Tue May 21 19:34:38 2024 +0900" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Thu May 23 12:03:26 2024 +0200" }, "message": "tls: fix missing memory barrier in tls_init\n\nIn tls_init(), a write memory barrier is missing, and store-store\nreordering may cause NULL dereference in tls_{setsockopt,getsockopt}.\n\nCPU0 CPU1\n----- -----\n// In tls_init()\n// In tls_ctx_create()\nctx \u003d kzalloc()\nctx-\u003esk_proto \u003d READ_ONCE(sk-\u003esk_prot) -(1)\n\n// In update_sk_prot()\nWRITE_ONCE(sk-\u003esk_prot, tls_prots) -(2)\n\n // In sock_common_setsockopt()\n READ_ONCE(sk-\u003esk_prot)-\u003esetsockopt()\n\n // In tls_{setsockopt,getsockopt}()\n ctx-\u003esk_proto-\u003esetsockopt() -(3)\n\nIn the above scenario, when (1) and (2) are reordered, (3) can observe\nthe NULL value of ctx-\u003esk_proto, causing NULL dereference.\n\nTo fix it, we rely on rcu_assign_pointer() which implies the release\nbarrier semantic. By moving rcu_assign_pointer() after ctx-\u003esk_proto is\ninitialized, we can ensure that ctx-\u003esk_proto are visible when\nchanging sk-\u003esk_prot.\n\nFixes: d5bee7374b68 (\"net/tls: Annotate access to sk_prot with READ_ONCE/WRITE_ONCE\")\nSigned-off-by: Yewon Choi \u003cwoni9911@gmail.com\u003e\nSigned-off-by: Dae R. Jeong \u003cthreeearcat@gmail.com\u003e\nLink: https://lore.kernel.org/netdev/ZU4OJG56g2V9z_H7@dragonet/T/\nLink: https://lore.kernel.org/r/Zkx4vjSFp0mfpjQ2@libra05\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "3b1c92f8e5371700fada307cc8fd2c51fa7bc8c1", "tree": "30d41b515ce3c8ab78a3b9a29dfa2c5dd1d2807a", "parents": [ "b35b1c0b4e166a427395deaf61e3140495dfcb89" ], "author": { "name": "Wei Fang", "email": "wei.fang@nxp.com", "time": "Tue May 21 10:38:00 2024 +0800" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Thu May 23 11:15:54 2024 +0200" }, "message": "net: fec: avoid lock evasion when reading pps_enable\n\nThe assignment of pps_enable is protected by tmreg_lock, but the read\noperation of pps_enable is not. So the Coverity tool reports a lock\nevasion warning which may cause data race to occur when running in a\nmultithread environment. Although this issue is almost impossible to\noccur, we\u0027d better fix it, at least it seems more logically reasonable,\nand it also prevents Coverity from continuing to issue warnings.\n\nFixes: 278d24047891 (\"net: fec: ptp: Enable PPS output based on ptp clock\")\nSigned-off-by: Wei Fang \u003cwei.fang@nxp.com\u003e\nLink: https://lore.kernel.org/r/20240521023800.17102-1-wei.fang@nxp.com\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "b35b1c0b4e166a427395deaf61e3140495dfcb89", "tree": "be0bab5271ffe08292e21fe84ab7766cf29bf860", "parents": [ "a61a459f58221f09810d6f60c657dda7add739fa" ], "author": { "name": "Jacob Keller", "email": "jacob.e.keller@intel.com", "time": "Mon May 20 17:21:27 2024 -0700" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Thu May 23 10:48:35 2024 +0200" }, "message": "Revert \"ixgbe: Manual AN-37 for troublesome link partners for X550 SFI\"\n\nThis reverts commit 565736048bd5f9888990569993c6b6bfdf6dcb6d.\n\nAccording to the commit, it implements a manual AN-37 for some\n\"troublesome\" Juniper MX5 switches. This appears to be a workaround for a\nparticular switch.\n\nIt has been reported that this causes a severe breakage for other switches,\nincluding a Cisco 3560CX-12PD-S.\n\nThe code appears to be a workaround for a specific switch which fails to\nlink in SFI mode. It expects to see AN-37 auto negotiation in order to\nlink. The Cisco switch is not expecting AN-37 auto negotiation. When the\ndevice starts the manual AN-37, the Cisco switch decides that the port is\nconfused and stops attempting to link with it. This persists until a power\ncycle. A simple driver unload and reload does not resolve the issue, even\nif loading with a version of the driver which lacks this workaround.\n\nThe authors of the workaround commit have not responded with\nclarifications, and the result of the workaround is complete failure to\nconnect with other switches.\n\nThis appears to be a case where the driver can either \"correctly\" link with\nthe Juniper MX5 switch, at the cost of bricking the link with the Cisco\nswitch, or it can behave properly for the Cisco switch, but fail to link\nwith the Junipir MX5 switch. I do not know enough about the standards\ninvolved to clearly determine whether either switch is at fault or behaving\nincorrectly. Nor do I know whether there exists some alternative fix which\ncorrects behavior with both switches.\n\nRevert the workaround for the Juniper switch.\n\nFixes: 565736048bd5 (\"ixgbe: Manual AN-37 for troublesome link partners for X550 SFI\")\nLink: https://lore.kernel.org/netdev/cbe874db-9ac9-42b8-afa0-88ea910e1e99@intel.com/T/\nLink: https://forum.proxmox.com/threads/intel-x553-sfp-ixgbe-no-go-on-pve8.135129/#post-612291\nSigned-off-by: Jacob Keller \u003cjacob.e.keller@intel.com\u003e\nCc: Jeff Daly \u003cjeffd@silicom-usa.com\u003e\nCc: kernel.org-fo5k2w@ycharbi.fr\nReviewed-by: Simon Horman \u003chorms@kernel.org\u003e\nLink: https://lore.kernel.org/r/20240520-net-2024-05-20-revert-silicom-switch-workaround-v1-1-50f80f261c94@intel.com\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "a61a459f58221f09810d6f60c657dda7add739fa", "tree": "3a9e5a52503bce750dc44e3362ca0c3d8c592efa", "parents": [ "9c91c7fadb1771dcc2815c5271d14566366d05c5" ], "author": { "name": "Joe Damato", "email": "jdamato@fastly.com", "time": "Mon May 20 23:58:43 2024 +0000" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Thu May 23 10:18:29 2024 +0200" }, "message": "testing: net-drv: use stats64 for testing\n\nTesting a network device that has large numbers of bytes/packets may\noverflow. Using stats64 when comparing fixes this problem.\n\nI tripped on this while iterating on a qstats patch for mlx5. See below\nfor confirmation without my added code that this is a bug.\n\nBefore this patch (with added debugging output):\n\n$ NETIF\u003deth0 tools/testing/selftests/drivers/net/stats.py\nKTAP version 1\n1..4\nok 1 stats.check_pause\nok 2 stats.check_fec\nrstat: 481708634 qstat: 666201639514 key: tx-bytes\nnot ok 3 stats.pkt_byte_sum\nok 4 stats.qstat_by_ifindex\n\nNote the huge delta above ^^^ in the rtnl vs qstats.\n\nAfter this patch:\n\n$ NETIF\u003deth0 tools/testing/selftests/drivers/net/stats.py\nKTAP version 1\n1..4\nok 1 stats.check_pause\nok 2 stats.check_fec\nok 3 stats.pkt_byte_sum\nok 4 stats.qstat_by_ifindex\n\nIt looks like rtnl_fill_stats in net/core/rtnetlink.c will attempt to\ncopy the 64bit stats into a 32bit structure which is probably why this\nbehavior is occurring.\n\nTo show this is happening, you can get the underlying stats that the\nstats.py test uses like this:\n\n$ ./cli.py --spec ../../../Documentation/netlink/specs/rt_link.yaml \\\n --do getlink --json \u0027{\"ifi-index\": 7}\u0027\n\nAnd examine the output (heavily snipped to show relevant fields):\n\n \u0027stats\u0027: {\n \u0027multicast\u0027: 3739197,\n \u0027rx-bytes\u0027: 1201525399,\n \u0027rx-packets\u0027: 56807158,\n \u0027tx-bytes\u0027: 492404458,\n \u0027tx-packets\u0027: 1200285371,\n\n \u0027stats64\u0027: {\n \u0027multicast\u0027: 3739197,\n \u0027rx-bytes\u0027: 35561263767,\n \u0027rx-packets\u0027: 56807158,\n \u0027tx-bytes\u0027: 666212335338,\n \u0027tx-packets\u0027: 1200285371,\n\nThe stats.py test prior to this patch was using the \u0027stats\u0027 structure\nabove, which matches the failure output on my system.\n\nComparing side by side, rx-bytes and tx-bytes, and getting ethtool -S\noutput:\n\nrx-bytes stats: 1201525399\nrx-bytes stats64: 35561263767\nrx-bytes ethtool: 36203402638\n\ntx-bytes stats: 492404458\ntx-bytes stats64: 666212335338\ntx-bytes ethtool: 666215360113\n\nNote that the above was taken from a system with an mlx5 NIC, which only\nexposes ndo_get_stats64.\n\nBased on the ethtool output and qstat output, it appears that stats.py\nshould be updated to use the \u0027stats64\u0027 structure for accurate\ncomparisons when packet/byte counters get very large.\n\nTo confirm that this was not related to the qstats code I was iterating\non, I booted a kernel without my driver changes and re-ran the test\nwhich shows the qstats are skipped (as they don\u0027t exist for mlx5):\n\nNETIF\u003deth0 tools/testing/selftests/drivers/net/stats.py\nKTAP version 1\n1..4\nok 1 stats.check_pause\nok 2 stats.check_fec\nok 3 stats.pkt_byte_sum # SKIP qstats not supported by the device\nok 4 stats.qstat_by_ifindex # SKIP No ifindex supports qstats\n\nBut, fetching the stats using the CLI\n\n$ ./cli.py --spec ../../../Documentation/netlink/specs/rt_link.yaml \\\n --do getlink --json \u0027{\"ifi-index\": 7}\u0027\n\nShows the same issue (heavily snipped for relevant fields only):\n\n \u0027stats\u0027: {\n \u0027multicast\u0027: 105489,\n \u0027rx-bytes\u0027: 530879526,\n \u0027rx-packets\u0027: 751415,\n \u0027tx-bytes\u0027: 2510191396,\n \u0027tx-packets\u0027: 27700323,\n \u0027stats64\u0027: {\n \u0027multicast\u0027: 105489,\n \u0027rx-bytes\u0027: 530879526,\n \u0027rx-packets\u0027: 751415,\n \u0027tx-bytes\u0027: 15395093284,\n \u0027tx-packets\u0027: 27700323,\n\nComparing side by side with ethtool -S on the unmodified mlx5 driver:\n\ntx-bytes stats: 2510191396\ntx-bytes stats64: 15395093284\ntx-bytes ethtool: 17718435810\n\nFixes: f0e6c86e4bab (\"testing: net-drv: add a driver test for stats reporting\")\nSigned-off-by: Joe Damato \u003cjdamato@fastly.com\u003e\nLink: https://lore.kernel.org/r/20240520235850.190041-1-jdamato@fastly.com\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "c760b3725e52403dc1b28644fb09c47a83cacea6", "tree": "652d83ee1ccf1ea723ba68dde69c03d64bd49fa3", "parents": [ "5c6f4d68e2aca67e425b7227369ec9fde8adfb6d", "db3e24a02e29b507c24c0adb4d22914c65dab763" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 18:59:29 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 18:59:29 2024 -0700" }, "message": "Merge tag \u0027mm-nonmm-stable-2024-05-22-17-30\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm\n\nPull more non-mm updates from Andrew Morton:\n\n - A series (\"kbuild: enable more warnings by default\") from Arnd\n Bergmann which enables a number of additional build-time warnings. We\n fixed all the fallout which we could find, there may still be a few\n stragglers.\n\n - Samuel Holland has developed the series \"Unified cross-architecture\n kernel-mode FPU API\". This does a lot of consolidation of\n per-architecture kernel-mode FPU usage and enables the use of newer\n AMD GPUs on RISC-V.\n\n - Tao Su has fixed some selftests build warnings in the series\n \"Selftests: Fix compilation warnings due to missing _GNU_SOURCE\n definition\".\n\n - This pull also includes a nilfs2 fixup from Ryusuke Konishi.\n\n* tag \u0027mm-nonmm-stable-2024-05-22-17-30\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm: (23 commits)\n nilfs2: make block erasure safe in nilfs_finish_roll_forward()\n selftests/harness: use 1024 in place of LINE_MAX\n Revert \"selftests/harness: remove use of LINE_MAX\"\n selftests/fpu: allow building on other architectures\n selftests/fpu: move FP code to a separate translation unit\n drm/amd/display: use ARCH_HAS_KERNEL_FPU_SUPPORT\n drm/amd/display: only use hard-float, not altivec on powerpc\n riscv: add support for kernel-mode FPU\n x86: implement ARCH_HAS_KERNEL_FPU_SUPPORT\n powerpc: implement ARCH_HAS_KERNEL_FPU_SUPPORT\n LoongArch: implement ARCH_HAS_KERNEL_FPU_SUPPORT\n lib/raid6: use CC_FLAGS_FPU for NEON CFLAGS\n arm64: crypto: use CC_FLAGS_FPU for NEON CFLAGS\n arm64: implement ARCH_HAS_KERNEL_FPU_SUPPORT\n ARM: crypto: use CC_FLAGS_FPU for NEON CFLAGS\n ARM: implement ARCH_HAS_KERNEL_FPU_SUPPORT\n arch: add ARCH_HAS_KERNEL_FPU_SUPPORT\n x86/fpu: fix asm/fpu/types.h include guard\n kbuild: enable -Wcast-function-type-strict unconditionally\n kbuild: enable -Wformat-truncation on clang\n ...\n" }, { "commit": "5c6f4d68e2aca67e425b7227369ec9fde8adfb6d", "tree": "b60f38675b4572047bcb840a89cb07329a5f6c22", "parents": [ "de7e71ef8bed222dd144d8878091ecb6d5dfd208", "99b80ac45f7ec351c2d1c9fbfec702213dcae566" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 17:32:04 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 17:32:04 2024 -0700" }, "message": "Merge tag \u0027mm-stable-2024-05-22-17-22\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm\n\nPull more mm updates from Andrew Morton:\n \"A series from Dave Chinner which cleans up and fixes the handling of\n nested allocations within stackdepot and page-owner\"\n\n* tag \u0027mm-stable-2024-05-22-17-22\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm:\n mm/page-owner: use gfp_nested_mask() instead of open coded masking\n stackdepot: use gfp_nested_mask() instead of open coded masking\n mm: lift gfp_kmemleak_mask() to gfp.h\n" }, { "commit": "2c92ca849fcc6ee7d0c358e9959abc9f58661aea", "tree": "0db4e762ad0b20f05d41bcfa53df57a46fa46ec3", "parents": [ "de7e71ef8bed222dd144d8878091ecb6d5dfd208" ], "author": { "name": "Steven Rostedt (Google)", "email": "rostedt@goodmis.org", "time": "Thu May 16 13:34:54 2024 -0400" }, "committer": { "name": "Steven Rostedt (Google)", "email": "rostedt@goodmis.org", "time": "Wed May 22 20:14:47 2024 -0400" }, "message": "tracing/treewide: Remove second parameter of __assign_str()\n\nWith the rework of how the __string() handles dynamic strings where it\nsaves off the source string in field in the helper structure[1], the\nassignment of that value to the trace event field is stored in the helper\nvalue and does not need to be passed in again.\n\nThis means that with:\n\n __string(field, mystring)\n\nWhich use to be assigned with __assign_str(field, mystring), no longer\nneeds the second parameter and it is unused. With this, __assign_str()\nwill now only get a single parameter.\n\nThere\u0027s over 700 users of __assign_str() and because coccinelle does not\nhandle the TRACE_EVENT() macro I ended up using the following sed script:\n\n git grep -l __assign_str | while read a ; do\n sed -e \u0027s/\\(__assign_str([^,]*[^ ,]\\) *,[^;]*/\\1)/\u0027 $a \u003e /tmp/test-file;\n mv /tmp/test-file $a;\n done\n\nI then searched for __assign_str() that did not end with \u0027;\u0027 as those\nwere multi line assignments that the sed script above would fail to catch.\n\nNote, the same updates will need to be done for:\n\n __assign_str_len()\n __assign_rel_str()\n __assign_rel_str_len()\n\nI tested this with both an allmodconfig and an allyesconfig (build only for both).\n\n[1] https://lore.kernel.org/linux-trace-kernel/20240222211442.634192653@goodmis.org/\n\nLink: https://lore.kernel.org/linux-trace-kernel/20240516133454.681ba6a0@rorschach.local.home\n\nCc: Masami Hiramatsu \u003cmhiramat@kernel.org\u003e\nCc: Mathieu Desnoyers \u003cmathieu.desnoyers@efficios.com\u003e\nCc: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\nCc: Julia Lawall \u003cJulia.Lawall@inria.fr\u003e\nSigned-off-by: Steven Rostedt (Google) \u003crostedt@goodmis.org\u003e\nAcked-by: Jani Nikula \u003cjani.nikula@intel.com\u003e\nAcked-by: Christian König \u003cchristian.koenig@amd.com\u003e for the amdgpu parts.\nAcked-by: Thomas Hellström \u003cthomas.hellstrom@linux.intel.com\u003e #for\nAcked-by: Rafael J. Wysocki \u003crafael@kernel.org\u003e # for thermal\nAcked-by: Takashi Iwai \u003ctiwai@suse.de\u003e\nAcked-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\t# xfs\nTested-by: Guenter Roeck \u003clinux@roeck-us.net\u003e\n" }, { "commit": "de7e71ef8bed222dd144d8878091ecb6d5dfd208", "tree": "329a4c2ee5a7b7defe6201e28381349b7e3e7cc1", "parents": [ "f8a6e48c6c6dc30dbd423a3f4b082df625664730" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Apr 07 13:18:39 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 14:37:23 2024 -0700" }, "message": "mm: simplify and improve print_vma_addr() output\n\nUse \u0027%pD\u0027 to print out the filename, and print out the actual offset\nwithin the file too, rather than just what the virtual address of the\nmapping is (which doesn\u0027t tell you anything about any mapping offsets).\n\nAlso, use the exact vma_lookup() instead of find_vma() - the latter\nlooks up any vma _after_ the address, which is of questionable value\n(yes, maybe you fell off the beginning, but you\u0027d be more likely to fall\noff the end).\n\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "f8a6e48c6c6dc30dbd423a3f4b082df625664730", "tree": "19b68bfffb5483634ec52f0c39a657f3cb54b929", "parents": [ "5f16eb0549ab502906fb2a10147dad4b9dc185c4", "b9b60b3199b70fe3ce74ff493b1870ccd7554134" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 14:13:22 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 14:13:22 2024 -0700" }, "message": "Merge local branch \u0027x86-codegen\u0027\n\nMerge trivial x86 code generation annoyances\n\n - Introduce helper macros for clang asm input problems\n\n - use said macros to improve trivially stupid code generation issues in\n bitops and array_index_mask_nospec\n\n - also improve codegen with 32-bit array index comparisons\n\nNone of these really matter, but I look at code generation and profiles\nfairly regularly, and these misfeatures caused the generated code to\nlook really odd and distract from the real issues.\n\n* branch \u0027x86-codegen\u0027 of local tree:\n x86: improve bitop code generation with clang\n x86: improve array_index_mask_nospec() code generation\n clang: work around asm input constraint problems\n" }, { "commit": "b9b60b3199b70fe3ce74ff493b1870ccd7554134", "tree": "232e0543c7015119bf752331a1642fa4b8f659de", "parents": [ "7453b9485114f7ffec4a99bccee469a4d4809894" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Apr 09 11:55:07 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 14:12:11 2024 -0700" }, "message": "x86: improve bitop code generation with clang\n\nThis uses the new ASM_INPUT_RM macro to avoid the bad code generation\nissue that clang has with more generic asm inputs.\n\nThis ends up avoiding generating code like this:\n\n \tmov %r10,(%rsp)\n \ttzcnt (%rsp),%rcx\n\nwhich now becomes just\n\n \ttzcnt %r10,%rcx\n\nand in the process ends up also removing a few unnecessary stack frames\nwhen the only use was that pointless \"asm uses memory location off stack\".\n\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "7453b9485114f7ffec4a99bccee469a4d4809894", "tree": "37e21820e4a0dc2fdbb04cb42d23e08eb9323266", "parents": [ "dbaaabd60e1662d2659eaeab0a4fc521667737ed" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Apr 08 11:38:30 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 14:12:11 2024 -0700" }, "message": "x86: improve array_index_mask_nospec() code generation\n\nDon\u0027t force the inputs to be \u0027unsigned long\u0027, when the comparison can\neasily be done in 32-bit if that\u0027s more appropriate.\n\nNote that while we can look at the inputs to choose an appropriate size\nfor the compare instruction, the output is fixed at \u0027unsigned long\u0027.\nThat\u0027s not technically optimal either, since a 32-bit \u0027sbbl\u0027 would often\nbe sufficient.\n\nBut for the outgoing mask we don\u0027t know how the mask ends up being used\n(ie we have uses that have an incoming 32-bit array index, but end up\nusing the mask for other things). That said, it only costs the extra\nREX prefix to always generate the 64-bit mask.\n\n[ A \u0027sbbl\u0027 also always technically generates a 64-bit mask, but with the\n upper 32 bits clear: that\u0027s fine for when the incoming index that will\n be masked is already 32-bit, but not if you use the mask to mask a\n pointer afterwards, like the file table lookup does ]\n\nCc: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nCc: H. Peter Anvin \u003chpa@zytor.com\u003e\nCc: Ingo Molnar \u003cmingo@kernel.org\u003e\nCc: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "dbaaabd60e1662d2659eaeab0a4fc521667737ed", "tree": "fe79efbe1efa9c9bb88b96a7735b49664b82d99a", "parents": [ "a38297e3fb012ddfa7ce0321a7e5a8daeb1872b6" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Apr 08 11:38:30 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 14:12:11 2024 -0700" }, "message": "clang: work around asm input constraint problems\n\nWork around clang problems with asm constraints that have multiple\npossibilities, particularly \"g\" and \"rm\".\n\nClang seems to turn inputs like that into the most generic form, which\nis the memory input - but to make matters worse, clang won\u0027t even use a\npossible original memory location, but will spill the value to stack,\nand use the stack for the asm input.\n\nSee\n\n https://github.com/llvm/llvm-project/issues/20571#issuecomment-980933442\n\nfor some explanation of why clang has this strange behavior, but the end\nresult is that \"g\" and \"rm\" really end up generating horrid code.\n\nLink: https://github.com/llvm/llvm-project/issues/20571\nCc: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nCc: H. Peter Anvin \u003chpa@zytor.com\u003e\nCc: Ingo Molnar \u003cmingo@kernel.org\u003e\nCc: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "5f16eb0549ab502906fb2a10147dad4b9dc185c4", "tree": "285bcf9df768d58f9aa5dbda8418f1ec961bd22b", "parents": [ "d90be6e4aaf23cd4a2c202891399cbafe669aaab", "f5b335dc025cfee90957efa90dc72fada0d5abb4" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 12:26:46 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 12:26:46 2024 -0700" }, "message": "Merge tag \u0027char-misc-6.10-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc\n\nPull char/misc and other driver subsystem updates from Greg KH:\n \"Here is the big set of char/misc and other driver subsystem updates\n for 6.10-rc1. Nothing major here, just lots of new drivers and updates\n for apis and new hardware types. Included in here are:\n\n - big IIO driver updates with more devices and drivers added\n\n - fpga driver updates\n\n - hyper-v driver updates\n\n - uio_pruss driver removal, no one uses it, other drivers control the\n same hardware now\n\n - binder minor updates\n\n - mhi driver updates\n\n - excon driver updates\n\n - counter driver updates\n\n - accessability driver updates\n\n - coresight driver updates\n\n - other hwtracing driver updates\n\n - nvmem driver updates\n\n - slimbus driver updates\n\n - spmi driver updates\n\n - other smaller misc and char driver updates\n\n All of these have been in linux-next for a while with no reported\n issues\"\n\n* tag \u0027char-misc-6.10-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc: (319 commits)\n misc: ntsync: mark driver as \"broken\" to prevent from building\n spmi: pmic-arb: Add multi bus support\n spmi: pmic-arb: Register controller for bus instead of arbiter\n spmi: pmic-arb: Make core resources acquiring a version operation\n spmi: pmic-arb: Make the APID init a version operation\n spmi: pmic-arb: Fix some compile warnings about members not being described\n dt-bindings: spmi: Deprecate qcom,bus-id\n dt-bindings: spmi: Add X1E80100 SPMI PMIC ARB schema\n spmi: pmic-arb: Replace three IS_ERR() calls by null pointer checks in spmi_pmic_arb_probe()\n spmi: hisi-spmi-controller: Do not override device identifier\n dt-bindings: spmi: hisilicon,hisi-spmi-controller: clean up example\n dt-bindings: spmi: hisilicon,hisi-spmi-controller: fix binding references\n spmi: make spmi_bus_type const\n extcon: adc-jack: Document missing struct members\n extcon: realtek: Remove unused of_gpio.h\n extcon: usbc-cros-ec: Convert to platform remove callback returning void\n extcon: usb-gpio: Convert to platform remove callback returning void\n extcon: max77843: Convert to platform remove callback returning void\n extcon: max3355: Convert to platform remove callback returning void\n extcon: intel-mrfld: Convert to platform remove callback returning void\n ...\n" }, { "commit": "d90be6e4aaf23cd4a2c202891399cbafe669aaab", "tree": "e30557c1d4cb9dc47f2b24c25668c19d6df9fbe5", "parents": [ "be81389c82e2c1ed0997629cb3d910f584666e33", "880a746fa3ea5916a012fa320fdfbcd3f331bea3" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 12:13:40 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 12:13:40 2024 -0700" }, "message": "Merge tag \u0027driver-core-6.10-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core\n\nPull driver core updates from Greg KH:\n \"Here is the small set of driver core and kernfs changes for 6.10-rc1.\n\n Nothing major here at all, just a small set of changes for some driver\n core apis, and minor fixups. Included in here are:\n\n - sysfs_bin_attr_simple_read() helper added and used\n\n - device_show_string() helper added and used\n\n All usages of these were acked by the various maintainers. Also in\n here are:\n\n - kernfs minor cleanup\n\n - removed unused functions\n\n - typo fix in documentation\n\n - pay attention to sysfs_create_link() failures in module.c finally\n\n All of these have been in linux-next for a very long time with no\n reported problems\"\n\n* tag \u0027driver-core-6.10-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core:\n device property: Fix a typo in the description of device_get_child_node_count()\n kernfs: mount: Remove unnecessary ‘NULL’ values from knparent\n scsi: Use device_show_string() helper for sysfs attributes\n platform/x86: Use device_show_string() helper for sysfs attributes\n perf: Use device_show_string() helper for sysfs attributes\n IB/qib: Use device_show_string() helper for sysfs attributes\n hwmon: Use device_show_string() helper for sysfs attributes\n driver core: Add device_show_string() helper for sysfs attributes\n treewide: Use sysfs_bin_attr_simple_read() helper\n sysfs: Add sysfs_bin_attr_simple_read() helper\n module: don\u0027t ignore sysfs_create_link() failures\n driver core: Remove unused platform_notify, platform_notify_remove\n" }, { "commit": "be81389c82e2c1ed0997629cb3d910f584666e33", "tree": "f5e7e34d3d13ae9b1c70c562bfb1f2eba7bd1684", "parents": [ "f6b8e86b7a65495d3947a1d1fc22183c52f786f6", "eb563dc752d33b0a5d4952964af15ca892f59524" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 12:11:48 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 12:11:48 2024 -0700" }, "message": "Merge tag \u0027staging-6.10-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging\n\nPull staging driver updates from Greg KH:\n \"Here is the big set of staging driver changes for 6.10-rc1. Not a lot\n of cleanups happening this kernel release, intern applications must be\n out of sync at the moment. But we did delete two drivers, wlan-ng and\n pi433, as they are no longer in use and the developers involved wanted\n them just gone entirely, allowing us to drop 19k lines from the tree.\n\n Other than the normal coding style cleanups here, there has been a lot\n of work on the vc04_services code, with the intent to finally get that\n out of staging hopefully soon. It\u0027s getting closer, which is nice to\n see.\n\n All of these have been in linux-next for a while with no reported\n issues\"\n\n* tag \u0027staging-6.10-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging: (98 commits)\n staging: pi433: Remove unused driver\n staging: vchiq_core: Add missing blank lines\n staging: vchiq_core: Drop unnecessary blank lines\n staging: vchiq_core: Add parentheses to VCHIQ_MSG_SRCPORT\n staging: vchiq_core: Use printk messages for devices\n staging: vchiq_arm: Drop unnecessary NULL check\n staging: vc04_services: Delete unnecessary NULL check\n staging: vc04_services: vchiq_arm: Fix NULL ptr dereferences\n Staging: rtl8192e: Rename variable DssCCk\n Staging: rtl8192e: Rename variable ExtHTCapInfo\n Staging: rtl8192e: Rename variable MPDUDensity\n Staging: rtl8192e: Rename variable MaxRxAMPDUFactor\n Staging: rtl8192e: Rename variable MaxAMSDUSize\n Staging: rtl8192e: Rename variable DelayBA\n Staging: rtl8192e: Rename variable RxSTBC\n Staging: rtl8192e: Rename variable TxSTBC\n Staging: rtl8192e: Rename variable GreenField\n Staging: rtl8192e: Rename variable ShortGI20Mhz\n Staging: rtl8192e: Rename variable ShortGI40Mhz\n Staging: rtl8192e: Rename variable MimoPwrSave\n ...\n" }, { "commit": "f6b8e86b7a65495d3947a1d1fc22183c52f786f6", "tree": "f0b15efd2dc851079a55a7b4e8a198fb6faa97ec", "parents": [ "89601f675b008ed0fd66c060fb23354a106436bb", "e21de1455a721a0cb4217b18589ede846f5b0686" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 11:53:02 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 11:53:02 2024 -0700" }, "message": "Merge tag \u0027tty-6.10-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty\n\nPull tty / serial updates from Greg KH:\n \"Here is the big set of tty/serial driver changes for 6.10-rc1.\n Included in here are:\n\n - Usual good set of api cleanups and evolution by Jiri Slaby to make\n the serial interfaces move out of the 1990\u0027s by using kfifos\n instead of hand-rolling their own logic.\n\n - 8250_exar driver updates\n\n - max3100 driver updates\n\n - sc16is7xx driver updates\n\n - exar driver updates\n\n - sh-sci driver updates\n\n - tty ldisc api addition to help refuse bindings\n\n - other smaller serial driver updates\n\n All of these have been in linux-next for a while with no reported\n issues\"\n\n* tag \u0027tty-6.10-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty: (113 commits)\n serial: Clear UPF_DEAD before calling tty_port_register_device_attr_serdev()\n serial: imx: Raise TX trigger level to 8\n serial: 8250_pnp: Simplify \"line\" related code\n serial: sh-sci: simplify locking when re-issuing RXDMA fails\n serial: sh-sci: let timeout timer only run when DMA is scheduled\n serial: sh-sci: describe locking requirements for invalidating RXDMA\n serial: sh-sci: protect invalidating RXDMA on shutdown\n tty: add the option to have a tty reject a new ldisc\n serial: core: Call device_set_awake_path() for console port\n dt-bindings: serial: brcm,bcm2835-aux-uart: convert to dtschema\n tty: serial: uartps: Add support for uartps controller reset\n arm64: zynqmp: Add resets property for UART nodes\n dt-bindings: serial: cdns,uart: Add optional reset property\n serial: 8250_pnp: Switch to DEFINE_SIMPLE_DEV_PM_OPS()\n serial: 8250_exar: Keep the includes sorted\n serial: 8250_exar: Make type of bit the same in exar_ee_*_bit()\n serial: 8250_exar: Use BIT() in exar_ee_read()\n serial: 8250_exar: Switch to use dev_err_probe()\n serial: 8250_exar: Return directly from switch-cases\n serial: 8250_exar: Decrease indentation level\n ...\n" }, { "commit": "89601f675b008ed0fd66c060fb23354a106436bb", "tree": "58520de5e06676ffb4fbb5d94a09fe97e378e244", "parents": [ "f3033eb79136dd27b17e7a192fac0155ceab5eb8", "51474ab44abf907023a8a875e799b07de461e466" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 11:40:09 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 11:40:09 2024 -0700" }, "message": "Merge tag \u0027usb-6.10-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb\n\nPull USB / Thunderbolt updates from Greg KH:\n \"Here is the big set of USB and Thunderbolt changes for 6.10-rc1.\n Nothing hugely earth-shattering, just constant forward progress for\n hardware support of new devices and cleanups over the drivers.\n\n Included in here are:\n\n - Thunderbolt / USB 4 driver updates\n\n - typec driver updates\n\n - dwc3 driver updates\n\n - gadget driver updates\n\n - uss720 driver id additions and fixes (people use USB-\u003earallel port\n devices still!)\n\n - onboard-hub driver rename and additions for new hardware\n\n - xhci driver updates\n\n - other small USB driver updates and additions for quirks and api\n changes\n\n All of these have been in linux-next for a while with no reported\n problems\"\n\n* tag \u0027usb-6.10-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb: (154 commits)\n drm/bridge: aux-hpd-bridge: correct devm_drm_dp_hpd_bridge_add() stub\n usb: fotg210: Add missing kernel doc description\n usb: dwc3: core: Fix unused variable warning in core driver\n usb: typec: tipd: rely on i2c_get_match_data()\n usb: typec: tipd: fix event checking for tps6598x\n usb: typec: tipd: fix event checking for tps25750\n dt-bindings: usb: qcom,dwc3: fix interrupt max items\n usb: fotg210: Use *-y instead of *-objs in Makefile\n usb: phy: tegra: Replace of_gpio.h by proper one\n usb: typec: ucsi: displayport: Fix potential deadlock\n usb: typec: qcom-pmic-typec: split HPD bridge alloc and registration\n usb: musc: Remove unused list \u0027buffers\u0027\n usb: dwc3: Wait unconditionally after issuing EndXfer command\n usb: gadget: u_audio: Clear uac pointer when freed.\n usb: gadget: u_audio: Fix race condition use of controls after free during gadget unbind.\n dt-bindings: usb: dwc3: Add QDU1000 compatible\n usb: core: Remove the useless struct usb_devmap which is just a bitmap\n MAINTAINERS: Remove {ehci,uhci}-platform.c from ARM/VT8500 entry\n USB: usb_parse_endpoint: ignore reserved bits\n usb: xhci: compact \u0027trb_in_td()\u0027 arguments\n ...\n" }, { "commit": "f3033eb79136dd27b17e7a192fac0155ceab5eb8", "tree": "444aecf3588c6b9d007e8dbb4a3c6baacec62322", "parents": [ "7eae27cd12a2d305ffad41a8e10cff3bb8c0dcb0", "f2994f5341e03b8680a88abc5f1dee950033c3a9" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 10:49:54 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 10:49:54 2024 -0700" }, "message": "Merge tag \u0027leds-next-6.10\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/lee/leds\n\nPull LED updates from Lee Jones:\n \"Core Frameworks:\n - Ensure seldom updated triggers have a brightness value before first\n update\n\n New Device Support:\n - Add support for Simatic IPC Device BX_59A to IPC LEDs Core\n - Add support for Qualcomm PMI8950 PWM to LPG Core\n\n New Functionality:\n - Add a bunch of new LED function identifiers\n - Add support for High Resolution Timers in LED Trigger Patten\n\n Fix-ups:\n - Shift out Audio Trigger to the Sound subsystem\n - Convert suitable calls to devm_* managed resources\n - Device Tree binding adaptions/conversions/creation\n - Remove superfluous code/variables/attributes and simplify overall\n - Use/convert to new/better APIs/helpers/MACROs instead of\n hand-rolling implementations\n\n Bug Fixes:\n - Repair enabling Torch Mode from V4L2 on the second LED\n - Ensure PWM is disabled when suspending\"\n\n* tag \u0027leds-next-6.10\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/lee/leds: (28 commits)\n leds: mt6370: Remove unused field \u0027reg_cfgs\u0027 from \u0027struct mt6370_priv\u0027\n leds: lp50xx: Remove unused field \u0027num_of_banked_leds\u0027 from \u0027struct lp50xx\u0027\n leds: lp50xx: Remove unused field \u0027bank_modules\u0027 from \u0027struct lp50xx_led\u0027\n leds: aat1290: Remove unused field \u0027torch_brightness\u0027 from \u0027struct aat1290_led\u0027\n leds: sun50i-a100: Use match_string() helper to simplify the code\n leds: pwm: Disable PWM when going to suspend\n leds: trigger: pattern: Add support for hrtimer\n leds: mt6360: Fix the second LED can not enable torch mode by V4L2\n dt-bindings: leds: leds-qcom-lpg: Add support for PMI8950 PWM\n leds: qcom-lpg: Add support for PMI8950 PWM\n leds: apu: Remove duplicate DMI lookup data\n leds: trigger: netdev: Remove not needed call to led_set_brightness in deactivate\n dt-bindings: leds: Add LED_FUNCTION_SPEED_* for link speed on LAN/WAN\n dt-bindings: leds: Add LED_FUNCTION_MOBILE for mobile network\n leds: simatic-ipc-leds-gpio: Add support for module BX-59A\n dt-bindings: leds: qcom-lpg: Document PM6150L compatible\n dt-bindings: leds: pca963x: Convert text bindings to YAML\n leds: an30259a: Use devm_mutex_init() for mutex initialization\n leds: mlxreg: Use devm_mutex_init() for mutex initialization\n leds: nic78bx: Use devm API to cleanup module\u0027s resources\n ...\n" }, { "commit": "7eae27cd12a2d305ffad41a8e10cff3bb8c0dcb0", "tree": "0f973c127f4ed014dbafa8755db6d0e87d1791c7", "parents": [ "a85629f435a4e724c414a6ae3e2f327272ab11af", "1fd949f653ee1a3c1776ef8a5295ae072c9b67f2" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 10:45:12 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 10:45:12 2024 -0700" }, "message": "Merge tag \u0027backlight-next-6.10\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/lee/backlight\n\nPull backlight updates from Lee Jones:\n \"Fix-ups:\n - FB Backlight interaction overhaul\n - Remove superfluous code and simplify overall\n - Constify various structs and struct attributes\n\n Bug Fixes:\n - Repair LED flickering\n - Fix signedness bugs\"\n\n* tag \u0027backlight-next-6.10\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/lee/backlight: (42 commits)\n backlight: sky81452-backlight: Remove unnecessary call to of_node_get()\n backlight: mp3309c: Fix LEDs flickering in PWM mode\n backlight: otm3225a: Drop driver owner assignment\n backlight: lp8788: Drop support for platform data\n backlight: lcd: Make lcd_class constant\n backlight: Make backlight_class constant\n backlight: mp3309c: Fix signedness bug in mp3309c_parse_fwnode()\n const_structs.checkpatch: add lcd_ops\n fbdev: omap: lcd_ams_delta: Constify lcd_ops\n fbdev: imx: Constify lcd_ops\n fbdev: clps711x: Constify lcd_ops\n HID: picoLCD: Constify lcd_ops\n backlight: tdo24m: Constify lcd_ops\n backlight: platform_lcd: Constify lcd_ops\n backlight: otm3225a: Constify lcd_ops\n backlight: ltv350qv: Constify lcd_ops\n backlight: lms501kf03: Constify lcd_ops\n backlight: lms283gf05: Constify lcd_ops\n backlight: l4f00242t03: Constify lcd_ops\n backlight: jornada720_lcd: Constify lcd_ops\n ...\n" }, { "commit": "a85629f435a4e724c414a6ae3e2f327272ab11af", "tree": "cc6279ca271f27e55b11a12860cd14cbfc9f15a5", "parents": [ "0bfbc914d9433d8ac2763a9ce99ce7721ee5c8e0", "1482489b5196f4203576ae1dc2ba4ce3ada381c7" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 10:41:14 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 10:41:14 2024 -0700" }, "message": "Merge tag \u0027mfd-next-6.10\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/lee/mfd\n\nPull MFD updates from Lee Jones:\n \"New Device Support:\n - Add support for X-Powers AXP717 PMIC to AXP22X\n - Add support for Rockchip RK816 PMIC to RK8XX\n - Add support for TI TPS65224 PMIC to TPS6594\n\n New Functionality:\n - Add Power Off functionality to Rohm BD71828\n - Allow I2C SMBus access in Renesas RSMU\n\n Fix-ups:\n - Device Tree binding adaptions/conversions/creation\n - Shift Intel support over to MSI interrupts\n - Generify adding platform data away from being ACPI specific\n - Use device core supplied attribute to register sysfs entries\n - Replace hand-rolled functionality with generic APIs\n - Utilise centrally provided helpers and macros\n - Clean-up error handling\n - Remove superfluous/duplicated/unused sections\n - Trivial; spelling, whitespace, coding-style adaptions\n - More Maple Tree conversions\"\n\n* tag \u0027mfd-next-6.10\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/lee/mfd: (44 commits)\n dt-bindings: mfd: Use full path to other schemas\n mfd: rsmu: support I2C SMBus access\n dt-bindings: mfd: Convert lp873x.txt to json-schema\n dt-bindings: mfd: aspeed: Drop \u0027oneOf\u0027 for pinctrl node\n dt-bindings: mfd: allwinner,sun6i-a31-prcm: Use hyphens in node names\n mfd: ssbi: Remove unused field \u0027slave\u0027 from \u0027struct ssbi\u0027\n mfd: kempld: Remove custom DMI matching code\n mfd: cs42l43: Update patching revision check\n dt-bindings: mfd: qcom: pm8xxx: Add pm8901 compatible\n mfd: timberdale: Remove redundant assignment to variable err\n dt-bindings: mfd: qcom,spmi-pmic: Add pbs to SPMI device types\n dt-bindings: mfd: syscon: Add ti,am62p-cpsw-mac-efuse compatible\n dt-bindings: mfd: qcom,tcsr: Add compatible for SDX75\n mfd: axp20x: Convert to use Maple Tree register cache\n mfd: bd71828: Remove commented code lines\n mfd: intel-m10-bmc: Change staging size to a variable\n dt-bindings: mfd: Add ROHM BD71879\n mfd: Tidy Kconfig dependency\u0027s parentheses\n mfd: ocelot-spi: Use spi_sync_transfer()\n dt-bindings: mfd: syscon: Add missing simple syscon compatibles\n ...\n" }, { "commit": "0bfbc914d9433d8ac2763a9ce99ce7721ee5c8e0", "tree": "ffcb63ac983a005728f7d4fd244bc32ec66d9b16", "parents": [ "4f05e82003d1c20da29fa593420b8d92e2c8d4e6", "92cce91949a497a8a4615f9ba5813b03f7a1f1d5" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 09:56:00 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 09:56:00 2024 -0700" }, "message": "Merge tag \u0027riscv-for-linus-6.10-mw1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux\n\nPull RISC-V updates from Palmer Dabbelt:\n\n - Add byte/half-word compare-and-exchange, emulated via LR/SC loops\n\n - Support for Rust\n\n - Support for Zihintpause in hwprobe\n\n - Add PR_RISCV_SET_ICACHE_FLUSH_CTX prctl()\n\n - Support lockless lockrefs\n\n* tag \u0027riscv-for-linus-6.10-mw1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux: (42 commits)\n riscv: defconfig: Enable CONFIG_CLK_SOPHGO_CV1800\n riscv: select ARCH_HAS_FAST_MULTIPLIER\n riscv: mm: still create swiotlb buffer for kmalloc() bouncing if required\n riscv: Annotate pgtable_l{4,5}_enabled with __ro_after_init\n riscv: Remove redundant CONFIG_64BIT from pgtable_l{4,5}_enabled\n riscv: mm: Always use an ASID to flush mm contexts\n riscv: mm: Preserve global TLB entries when switching contexts\n riscv: mm: Make asid_bits a local variable\n riscv: mm: Use a fixed layout for the MM context ID\n riscv: mm: Introduce cntx2asid/cntx2version helper macros\n riscv: Avoid TLB flush loops when affected by SiFive CIP-1200\n riscv: Apply SiFive CIP-1200 workaround to single-ASID sfence.vma\n riscv: mm: Combine the SMP and UP TLB flush code\n riscv: Only send remote fences when some other CPU is online\n riscv: mm: Broadcast kernel TLB flushes only when needed\n riscv: Use IPIs for remote cache/TLB flushes by default\n riscv: Factor out page table TLB synchronization\n riscv: Flush the instruction cache during SMP bringup\n riscv: hwprobe: export Zihintpause ISA extension\n riscv: misaligned: remove CONFIG_RISCV_M_MODE specific code\n ...\n" }, { "commit": "4f05e82003d1c20da29fa593420b8d92e2c8d4e6", "tree": "db3e2ebe9e14c55795fe00434356cc233c75d26d", "parents": [ "f33fda22a775d7c6d9b14757f94a535c0698c73a", "9cc1df421f00453afdcaf78b105d8e7fd03cce78" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 09:43:07 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 09:43:07 2024 -0700" }, "message": "Merge tag \u0027loongarch-6.10\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/chenhuacai/linux-loongson\n\nPull LoongArch updates from Huacai Chen:\n\n - Select some options in Kconfig\n\n - Give a chance to build with !CONFIG_SMP\n\n - Switch to use built-in rustc target\n\n - Add new supported device nodes to dts\n\n - Some bug fixes and other small changes\n\n - Update the default config file\n\n* tag \u0027loongarch-6.10\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/chenhuacai/linux-loongson:\n LoongArch: Update Loongson-3 default config file\n LoongArch: dts: Add new supported device nodes to Loongson-2K2000\n LoongArch: dts: Add new supported device nodes to Loongson-2K0500\n LoongArch: dts: Remove \"disabled\" state of clock controller node\n LoongArch: rust: Switch to use built-in rustc target\n LoongArch: Fix callchain parse error with kernel tracepoint events again\n LoongArch: Give a chance to build with !CONFIG_SMP\n LoongArch: Select THP_SWAP if HAVE_ARCH_TRANSPARENT_HUGEPAGE\n LoongArch: Select ARCH_WANT_DEFAULT_BPF_JIT\n LoongArch: Select ARCH_SUPPORTS_INT128 if CC_HAS_INT128\n LoongArch: Select ARCH_HAS_FAST_MULTIPLIER\n" }, { "commit": "f33fda22a775d7c6d9b14757f94a535c0698c73a", "tree": "5d76c55114b49a6affa39bf7132acc779f8fd3f1", "parents": [ "0e22bedd758643bc2cc161d54aa181e329da0ab3", "58d647506c92ccd3cfa0c453c68ddd14f40bf06f" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 09:31:01 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 09:31:01 2024 -0700" }, "message": "Merge tag \u0027microblaze-v6.10\u0027 of git://git.monstr.eu/linux-2.6-microblaze\n\nPull microblaze updates from Michal Simek:\n\n - Cleanup code around removed early_printk\n\n* tag \u0027microblaze-v6.10\u0027 of git://git.monstr.eu/linux-2.6-microblaze:\n microblaze: Remove early printk call from cpuinfo-static.c\n microblaze: Remove gcc flag for non existing early_printk.c file\n" }, { "commit": "0e22bedd758643bc2cc161d54aa181e329da0ab3", "tree": "73fd34e6567f36c423a1ff37bf34bff1b082ade8", "parents": [ "4f2d34b65b40937b43c38ba34ece5aa3bc210e0d", "e9229c18dae3b3c2556cea8413edd1f76c78d767" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 09:23:18 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 09:23:18 2024 -0700" }, "message": "Merge tag \u0027ovl-update-6.10\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/overlayfs/vfs\n\nPull overlayfs updates from Miklos Szeredi:\n\n - Add tmpfile support\n\n - Clean up include\n\n* tag \u0027ovl-update-6.10\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/overlayfs/vfs:\n ovl: remove duplicate included header\n ovl: remove upper umask handling from ovl_create_upper()\n ovl: implement tmpfile\n" }, { "commit": "4f2d34b65b40937b43c38ba34ece5aa3bc210e0d", "tree": "259d1ac5841ef3fb969537d07e87d937c88b2134", "parents": [ "681ce8623567ba7e7333908e9826b77145312dda", "529395d2ae6456c556405016ea0c43081fe607f3" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 09:18:51 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 09:18:51 2024 -0700" }, "message": "Merge tag \u0027fuse-update-6.10\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse\n\nPull fuse updates from Miklos Szeredi:\n\n - Add fs-verity support (Richard Fung)\n\n - Add multi-queue support to virtio-fs (Peter-Jan Gootzen)\n\n - Fix a bug in NOTIFY_RESEND handling (Hou Tao)\n\n - page -\u003e folio cleanup (Matthew Wilcox)\n\n* tag \u0027fuse-update-6.10\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse:\n virtio-fs: add multi-queue support\n virtio-fs: limit number of request queues\n fuse: clear FR_SENT when re-adding requests into pending list\n fuse: set FR_PENDING atomically in fuse_resend()\n fuse: Add initial support for fs-verity\n fuse: Convert fuse_readpages_end() to use folio_end_read()\n" }, { "commit": "681ce8623567ba7e7333908e9826b77145312dda", "tree": "545213689ec5dcfd291f0a50018639d631088afc", "parents": [ "29c73fc794c83505066ee6db893b2a83ac5fac63" ], "author": { "name": "Yafang Shao", "email": "laoar.shao@gmail.com", "time": "Wed May 15 17:17:27 2024 +0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 22 08:49:13 2024 -0700" }, "message": "vfs: Delete the associated dentry when deleting a file\n\nOur applications, built on Elasticsearch[0], frequently create and\ndelete files. These applications operate within containers, some with a\nmemory limit exceeding 100GB. Over prolonged periods, the accumulation\nof negative dentries within these containers can amount to tens of\ngigabytes.\n\nUpon container exit, directories are deleted. However, due to the\nnumerous associated dentries, this process can be time-consuming. Our\nusers have expressed frustration with this prolonged exit duration,\nwhich constitutes our first issue.\n\nSimultaneously, other processes may attempt to access the parent\ndirectory of the Elasticsearch directories. Since the task responsible\nfor deleting the dentries holds the inode lock, processes attempting\ndirectory lookup experience significant delays. This issue, our second\nproblem, is easily demonstrated:\n\n - Task 1 generates negative dentries:\n $ pwd\n ~/test\n $ mkdir es \u0026\u0026 cd es/ \u0026\u0026 ./create_and_delete_files.sh\n\n [ After generating tens of GB dentries ]\n\n $ cd ~/test \u0026\u0026 rm -rf es\n\n [ It will take a long duration to finish ]\n\n - Task 2 attempts to lookup the \u0027test/\u0027 directory\n $ pwd\n ~/test\n $ ls\n\n The \u0027ls\u0027 command in Task 2 experiences prolonged execution as Task 1\n is deleting the dentries.\n\nWe\u0027ve devised a solution to address both issues by deleting associated\ndentry when removing a file. Interestingly, we\u0027ve noted that a similar\npatch was proposed years ago[1], although it was rejected citing the\nabsence of tangible issues caused by negative dentries. Given our\ncurrent challenges, we\u0027re resubmitting the proposal. All relevant\nstakeholders from previous discussions have been included for reference.\n\nSome alternative solutions are also under discussion[2][3], such as\nshrinking child dentries outside of the parent inode lock or even\nasynchronously shrinking child dentries. However, given the\nstraightforward nature of the current solution, I believe this approach\nis still necessary.\n\n[ NOTE! This is a pretty fundamental change in how we deal with\n unlinking dentries, and it doesn\u0027t change the fact that you can have\n lots of negative dentries from just doing negative lookups.\n\n But the kernel test robot is at least initially happy with this from a\n performance angle, so I\u0027m applying this ASAP just to get more testing\n and as a \"known fix for an issue people hit in real life\".\n\n Put another way: we should still look at the alternatives, and this\n patch may get reverted if somebody finds a performance regression on\n some other load. - Linus ]\n\nSuggested-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\nSigned-off-by: Yafang Shao \u003claoar.shao@gmail.com\u003e\nLink: https://github.com/elastic/elasticsearch [0]\nLink: https://patchwork.kernel.org/project/linux-fsdevel/patch/1502099673-31620-1-git-send-email-wangkai86@huawei.com [1]\nLink: https://lore.kernel.org/linux-fsdevel/20240511200240.6354-2-torvalds@linux-foundation.org/ [2]\nLink: https://lore.kernel.org/linux-fsdevel/CAHk-\u003dwjEMf8Du4UFzxuToGDnF3yLaMcrYeyNAaH1NJWa6fwcNQ@mail.gmail.com/ [3]\nCc: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\nCc: Christian Brauner \u003cbrauner@kernel.org\u003e\nCc: Jan Kara \u003cjack@suse.cz\u003e\nCc: Waiman Long \u003clongman@redhat.com\u003e\nCc: Matthew Wilcox \u003cwilly@infradead.org\u003e\nCc: Wangkai \u003cwangkai86@huawei.com\u003e\nCc: Colin Walters \u003cwalters@verbum.org\u003e\nTested-by: kernel test robot \u003coliver.sang@intel.com\u003e\nLink: https://lore.kernel.org/all/202405221518.ecea2810-oliver.sang@intel.com/\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "c8fae27d141a32a1624d0d0d5419d94252824498", "tree": "de050c966953c6a0ee689055db7e60e516bb1eb1", "parents": [ "c0a7233f583f9aa53b5066785201746cdcd8fd24" ], "author": { "name": "Li Zhang", "email": "zhanglikernel@gmail.com", "time": "Sat Mar 16 13:25:54 2024 +0800" }, "committer": { "name": "Michael S. Tsirkin", "email": "mst@redhat.com", "time": "Wed May 22 08:39:41 2024 -0400" }, "message": "virtio-pci: Check if is_avq is NULL\n\n[bug]\nIn the virtio_pci_common.c function vp_del_vqs, vp_dev-\u003eis_avq is involved\nto determine whether it is admin virtqueue, but this function vp_dev-\u003eis_avq\n may be empty. For installations, virtio_pci_legacy does not assign a value\n to vp_dev-\u003eis_avq.\n\n[fix]\nCheck whether it is vp_dev-\u003eis_avq before use.\n\n[test]\nTest with virsh Attach device\nBefore this patch, the following command would crash the guest system\n\nAfter applying the patch, everything seems to be working fine.\n\nSigned-off-by: Li Zhang \u003czhanglikernel@gmail.com\u003e\nMessage-Id: \u003c1710566754-3532-1-git-send-email-zhanglikernel@gmail.com\u003e\nSigned-off-by: Michael S. Tsirkin \u003cmst@redhat.com\u003e\n" }, { "commit": "c0a7233f583f9aa53b5066785201746cdcd8fd24", "tree": "7cd0c7d1058a01e475af5bba1a5391c29594ddcc", "parents": [ "89875151fccdd024d571aa884ea97a0128b968b6", "894452180d732413fd29fa95a4820560fa44ca4a" ], "author": { "name": "Michael S. Tsirkin", "email": "mst@redhat.com", "time": "Sun May 12 08:13:45 2024 -0400" }, "committer": { "name": "Michael S. Tsirkin", "email": "mst@redhat.com", "time": "Wed May 22 08:32:48 2024 -0400" }, "message": "Merge tag \u0027stable/vduse-virtio-net\u0027 into vhost\n\nThis adds support for virtio-net to vduse.\n\nSigned-off-by: Michael S. Tsirkin \u003cmst@redhat.com\u003e\n" }, { "commit": "89875151fccdd024d571aa884ea97a0128b968b6", "tree": "8133b94d1b4a95bcbb2fba9f9e2bf1a8af231d34", "parents": [ "10e49da815e3e3dda7a86af2124edd687ea29a54" ], "author": { "name": "Jiri Pirko", "email": "jiri@nvidia.com", "time": "Fri Apr 26 17:08:45 2024 +0200" }, "committer": { "name": "Michael S. Tsirkin", "email": "mst@redhat.com", "time": "Wed May 22 08:31:38 2024 -0400" }, "message": "virtio: delete vq in vp_find_vqs_msix() when request_irq() fails\n\nWhen request_irq() fails, error path calls vp_del_vqs(). There, as vq is\npresent in the list, free_irq() is called for the same vector. That\ncauses following splat:\n\n[ 0.414355] Trying to free already-free IRQ 27\n[ 0.414403] WARNING: CPU: 1 PID: 1 at kernel/irq/manage.c:1899 free_irq+0x1a1/0x2d0\n[ 0.414510] Modules linked in:\n[ 0.414540] CPU: 1 PID: 1 Comm: swapper/0 Not tainted 6.9.0-rc4+ #27\n[ 0.414540] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.3-1.fc39 04/01/2014\n[ 0.414540] RIP: 0010:free_irq+0x1a1/0x2d0\n[ 0.414540] Code: 1e 00 48 83 c4 08 48 89 e8 5b 5d 41 5c 41 5d 41 5e 41 5f c3 cc cc cc cc 90 8b 74 24 04 48 c7 c7 98 80 6c b1 e8 00 c9 f7 ff 90 \u003c0f\u003e 0b 90 90 48 89 ee 4c 89 ef e8 e0 20 b8 00 49 8b 47 40 48 8b 40\n[ 0.414540] RSP: 0000:ffffb71480013ae0 EFLAGS: 00010086\n[ 0.414540] RAX: 0000000000000000 RBX: ffffa099c2722000 RCX: 0000000000000000\n[ 0.414540] RDX: 0000000000000000 RSI: ffffb71480013998 RDI: 0000000000000001\n[ 0.414540] RBP: 0000000000000246 R08: 00000000ffffdfff R09: 0000000000000001\n[ 0.414540] R10: 00000000ffffdfff R11: ffffffffb18729c0 R12: ffffa099c1c91760\n[ 0.414540] R13: ffffa099c1c916a4 R14: ffffa099c1d2f200 R15: ffffa099c1c91600\n[ 0.414540] FS: 0000000000000000(0000) GS:ffffa099fec40000(0000) knlGS:0000000000000000\n[ 0.414540] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033\n[ 0.414540] CR2: 0000000000000000 CR3: 0000000008e3e001 CR4: 0000000000370ef0\n[ 0.414540] Call Trace:\n[ 0.414540] \u003cTASK\u003e\n[ 0.414540] ? __warn+0x80/0x120\n[ 0.414540] ? free_irq+0x1a1/0x2d0\n[ 0.414540] ? report_bug+0x164/0x190\n[ 0.414540] ? handle_bug+0x3b/0x70\n[ 0.414540] ? exc_invalid_op+0x17/0x70\n[ 0.414540] ? asm_exc_invalid_op+0x1a/0x20\n[ 0.414540] ? free_irq+0x1a1/0x2d0\n[ 0.414540] vp_del_vqs+0xc1/0x220\n[ 0.414540] vp_find_vqs_msix+0x305/0x470\n[ 0.414540] vp_find_vqs+0x3e/0x1a0\n[ 0.414540] vp_modern_find_vqs+0x1b/0x70\n[ 0.414540] init_vqs+0x387/0x600\n[ 0.414540] virtnet_probe+0x50a/0xc80\n[ 0.414540] virtio_dev_probe+0x1e0/0x2b0\n[ 0.414540] really_probe+0xc0/0x2c0\n[ 0.414540] ? __pfx___driver_attach+0x10/0x10\n[ 0.414540] __driver_probe_device+0x73/0x120\n[ 0.414540] driver_probe_device+0x1f/0xe0\n[ 0.414540] __driver_attach+0x88/0x180\n[ 0.414540] bus_for_each_dev+0x85/0xd0\n[ 0.414540] bus_add_driver+0xec/0x1f0\n[ 0.414540] driver_register+0x59/0x100\n[ 0.414540] ? __pfx_virtio_net_driver_init+0x10/0x10\n[ 0.414540] virtio_net_driver_init+0x90/0xb0\n[ 0.414540] do_one_initcall+0x58/0x230\n[ 0.414540] kernel_init_freeable+0x1a3/0x2d0\n[ 0.414540] ? __pfx_kernel_init+0x10/0x10\n[ 0.414540] kernel_init+0x1a/0x1c0\n[ 0.414540] ret_from_fork+0x31/0x50\n[ 0.414540] ? __pfx_kernel_init+0x10/0x10\n[ 0.414540] ret_from_fork_asm+0x1a/0x30\n[ 0.414540] \u003c/TASK\u003e\n\nFix this by calling deleting the current vq when request_irq() fails.\n\nFixes: 0b0f9dc52ed0 (\"Revert \"virtio_pci: use shared interrupts for virtqueues\"\")\nSigned-off-by: Jiri Pirko \u003cjiri@nvidia.com\u003e\nMessage-Id: \u003c20240426150845.3999481-1-jiri@resnulli.us\u003e\nSigned-off-by: Michael S. Tsirkin \u003cmst@redhat.com\u003e\n" }, { "commit": "10e49da815e3e3dda7a86af2124edd687ea29a54", "tree": "5cfb93b13c1e45888212844d317080ecfc71c684", "parents": [ "f452001dca301d05e509f3e39998c442900e9937" ], "author": { "name": "Eugenio Pérez", "email": "eperezma@redhat.com", "time": "Tue Feb 13 19:24:50 2024 +0100" }, "committer": { "name": "Michael S. Tsirkin", "email": "mst@redhat.com", "time": "Wed May 22 08:31:18 2024 -0400" }, "message": "MAINTAINERS: add Eugenio Pérez as reviewer\n\nAdd myself as a reviewer of some VirtIO areas I\u0027m interested.\n\nUntil this point I\u0027ve been scanning manually the list looking for\nseries that touches this area. Adding myself to make this task easier.\n\nSigned-off-by: Eugenio Pérez \u003ceperezma@redhat.com\u003e\nMessage-Id: \u003c20240213182450.106796-1-eperezma@redhat.com\u003e\nSigned-off-by: Michael S. Tsirkin \u003cmst@redhat.com\u003e\n" }, { "commit": "f452001dca301d05e509f3e39998c442900e9937", "tree": "75d91b5025fb6b4cba62e91dfbcde3aa0233c2ea", "parents": [ "4d685629b727cf17be9446865076d52dd3fa0933" ], "author": { "name": "Christophe JAILLET", "email": "christophe.jaillet@wanadoo.fr", "time": "Sun Apr 14 12:04:26 2024 +0200" }, "committer": { "name": "Michael S. Tsirkin", "email": "mst@redhat.com", "time": "Wed May 22 08:31:18 2024 -0400" }, "message": "vhost-vdpa: Remove usage of the deprecated ida_simple_xx() API\n\nida_alloc() and ida_free() should be preferred to the deprecated\nida_simple_get() and ida_simple_remove().\n\nNote that the upper limit of ida_simple_get() is exclusive, but the one of\nida_alloc_max() is inclusive. So a -1 has been added when needed.\n\nSigned-off-by: Christophe JAILLET \u003cchristophe.jaillet@wanadoo.fr\u003e\nReviewed-by: Simon Horman \u003chorms@kernel.org\u003e\nMessage-Id: \u003c67c2edf49788c27d5f7a49fc701520b9fcf739b5.1713088999.git.christophe.jaillet@wanadoo.fr\u003e\nSigned-off-by: Michael S. Tsirkin \u003cmst@redhat.com\u003e\nAcked-by: Jason Wang \u003cjasowang@redhat.com\u003e\n" }, { "commit": "4d685629b727cf17be9446865076d52dd3fa0933", "tree": "2db0f47d8f10e05d82a0b81d7af950fdce9151d7", "parents": [ "1fa74f2449001de80bf2548cb42c04dd5b848d43" ], "author": { "name": "Yuxue Liu", "email": "yuxue.liu@jaguarmicro.com", "time": "Wed Apr 10 11:30:20 2024 +0800" }, "committer": { "name": "Michael S. Tsirkin", "email": "mst@redhat.com", "time": "Wed May 22 08:31:18 2024 -0400" }, "message": "vp_vdpa: don\u0027t allocate unused msix vectors\n\nWhen there is a ctlq and it doesn\u0027t require interrupt\ncallbacks,the original method of calculating vectors\nwastes hardware msi or msix resources as well as system\nIRQ resources.\n\nWhen conducting performance testing using testpmd in the\nguest os, it was found that the performance was lower compared\nto directly using vfio-pci to passthrough the device\n\nIn scenarios where the virtio device in the guest os does\nnot utilize interrupts, the vdpa driver still configures\nthe hardware\u0027s msix vector. Therefore, the hardware still\nsends interrupts to the host os. Because of this unnecessary\naction by the hardware, hardware performance decreases, and\nit also affects the performance of the host os.\n\nBefore modification:(interrupt mode)\n 32: 0 0 0 0 PCI-MSI 32768-edge vp-vdpa[0000:00:02.0]-0\n 33: 0 0 0 0 PCI-MSI 32769-edge vp-vdpa[0000:00:02.0]-1\n 34: 0 0 0 0 PCI-MSI 32770-edge vp-vdpa[0000:00:02.0]-2\n 35: 0 0 0 0 PCI-MSI 32771-edge vp-vdpa[0000:00:02.0]-config\n\nAfter modification:(interrupt mode)\n 32: 0 0 1 7 PCI-MSI 32768-edge vp-vdpa[0000:00:02.0]-0\n 33: 36 0 3 0 PCI-MSI 32769-edge vp-vdpa[0000:00:02.0]-1\n 34: 0 0 0 0 PCI-MSI 32770-edge vp-vdpa[0000:00:02.0]-config\n\nBefore modification:(virtio pmd mode for guest os)\n 32: 0 0 0 0 PCI-MSI 32768-edge vp-vdpa[0000:00:02.0]-0\n 33: 0 0 0 0 PCI-MSI 32769-edge vp-vdpa[0000:00:02.0]-1\n 34: 0 0 0 0 PCI-MSI 32770-edge vp-vdpa[0000:00:02.0]-2\n 35: 0 0 0 0 PCI-MSI 32771-edge vp-vdpa[0000:00:02.0]-config\n\nAfter modification:(virtio pmd mode for guest os)\n 32: 0 0 0 0 PCI-MSI 32768-edge vp-vdpa[0000:00:02.0]-config\n\nTo verify the use of the virtio PMD mode in the guest operating\nsystem, the following patch needs to be applied to QEMU:\nhttps://lore.kernel.org/all/20240408073311.2049-1-yuxue.liu@jaguarmicro.com\n\nSigned-off-by: Yuxue Liu \u003cyuxue.liu@jaguarmicro.com\u003e\nAcked-by: Jason Wang \u003cjasowang@redhat.com\u003e\nReviewed-by: Heng Qi \u003chengqi@linux.alibaba.com\u003e\nMessage-Id: \u003c20240410033020.1310-1-yuxue.liu@jaguarmicro.com\u003e\nSigned-off-by: Michael S. Tsirkin \u003cmst@redhat.com\u003e\n" }, { "commit": "1fa74f2449001de80bf2548cb42c04dd5b848d43", "tree": "6b09f5e631b1c46c6599b4552976408e13a483ff", "parents": [ "bc21020f61f5f93f39eeedc7ae56797e59b98816" ], "author": { "name": "Krzysztof Kozlowski", "email": "krzysztof.kozlowski@linaro.org", "time": "Sun Mar 31 10:44:12 2024 +0200" }, "committer": { "name": "Michael S. Tsirkin", "email": "mst@redhat.com", "time": "Wed May 22 08:31:18 2024 -0400" }, "message": "sound: virtio: drop owner assignment\n\nvirtio core already sets the .owner, so driver does not need to.\n\nSigned-off-by: Krzysztof Kozlowski \u003ckrzysztof.kozlowski@linaro.org\u003e\n\nMessage-Id: \u003c20240331-module-owner-virtio-v2-25-98f04bfaf46a@linaro.org\u003e\nSigned-off-by: Michael S. Tsirkin \u003cmst@redhat.com\u003e\nAcked-by: Anton Yakovlev \u003canton.yakovlev@opensynergy.com\u003e\n" }, { "commit": "bc21020f61f5f93f39eeedc7ae56797e59b98816", "tree": "e652096153fb13420a8e5a117ccbd6650d9d80ae", "parents": [ "e1e4d376836a134465f8f265281e4db5d3bcd9db" ], "author": { "name": "Krzysztof Kozlowski", "email": "krzysztof.kozlowski@linaro.org", "time": "Sun Mar 31 10:44:11 2024 +0200" }, "committer": { "name": "Michael S. Tsirkin", "email": "mst@redhat.com", "time": "Wed May 22 08:31:18 2024 -0400" }, "message": "fuse: virtio: drop owner assignment\n\nvirtio core already sets the .owner, so driver does not need to.\n\nSigned-off-by: Krzysztof Kozlowski \u003ckrzysztof.kozlowski@linaro.org\u003e\n\nMessage-Id: \u003c20240331-module-owner-virtio-v2-24-98f04bfaf46a@linaro.org\u003e\nSigned-off-by: Michael S. Tsirkin \u003cmst@redhat.com\u003e\nReviewed-by: Stefan Hajnoczi \u003cstefanha@redhat.com\u003e\n" } ], "next": "e1e4d376836a134465f8f265281e4db5d3bcd9db" }