{ "_class" : "hudson.model.FreeStyleBuild", "actions" : [ { "_class" : "hudson.model.CauseAction", "causes" : [ { "_class" : "org.jenkinsci.plugins.parameterizedscheduler.ParameterizedTimerTriggerCause", "shortDescription" : "Started by timer with parameters: {BUILD_CFG_TARGET_DEV=WR8750N/WR9500N/WG600HP (AR9344)}" } ] }, { "_class" : "hudson.model.ParametersAction", "parameters" : [ { "_class" : "hudson.model.BooleanParameterValue", "name" : "BUILD_CFG_DISTCLEAN", "value" : True }, { "_class" : "hudson.model.BooleanParameterValue", "name" : "BUILD_CFG_DIFFCONFIG", "value" : True }, { "_class" : "hudson.model.StringParameterValue", "name" : "BUILD_CFG_TARGET_DEV", "value" : "WR8750N/WR9500N/WG600HP (AR9344)" } ] }, { "_class" : "hudson.plugins.git.util.BuildData", "buildsByBranchName" : { "refs/remotes/origin/main" : { "_class" : "hudson.plugins.git.util.Build", "buildNumber" : 519, "buildResult" : None, "marked" : { "SHA1" : "795f369b181cd128962fa8f60a03eaf0ac8afac6", "branch" : [ { "SHA1" : "795f369b181cd128962fa8f60a03eaf0ac8afac6", "name" : "refs/remotes/origin/main" } ] }, "revision" : { "SHA1" : "795f369b181cd128962fa8f60a03eaf0ac8afac6", "branch" : [ { "SHA1" : "795f369b181cd128962fa8f60a03eaf0ac8afac6", "name" : "refs/remotes/origin/main" } ] } } }, "lastBuiltRevision" : { "SHA1" : "795f369b181cd128962fa8f60a03eaf0ac8afac6", "branch" : [ { "SHA1" : "795f369b181cd128962fa8f60a03eaf0ac8afac6", "name" : "refs/remotes/origin/main" } ] }, "remoteUrls" : [ "https://github.com/openwrt/openwrt" ], "scmName" : "" }, { }, { }, { }, { }, { "_class" : "org.jenkinsci.plugins.displayurlapi.actions.RunDisplayAction" } ], "artifacts" : [ { "displayPath" : "config.buildinfo", "fileName" : "config.buildinfo", "relativePath" : "bin/targets/ath79/tiny/config.buildinfo" }, { "displayPath" : "feeds.buildinfo", "fileName" : "feeds.buildinfo", "relativePath" : "bin/targets/ath79/tiny/feeds.buildinfo" }, { "displayPath" : "openwrt-ath79-tiny-nec_wg600hp-initramfs-factory.bin", "fileName" : "openwrt-ath79-tiny-nec_wg600hp-initramfs-factory.bin", "relativePath" : "bin/targets/ath79/tiny/openwrt-ath79-tiny-nec_wg600hp-initramfs-factory.bin" }, { "displayPath" : "openwrt-ath79-tiny-nec_wg600hp-initramfs-kernel.bin", "fileName" : "openwrt-ath79-tiny-nec_wg600hp-initramfs-kernel.bin", "relativePath" : "bin/targets/ath79/tiny/openwrt-ath79-tiny-nec_wg600hp-initramfs-kernel.bin" }, { "displayPath" : "openwrt-ath79-tiny-nec_wg600hp-squashfs-sysupgrade.bin", "fileName" : "openwrt-ath79-tiny-nec_wg600hp-squashfs-sysupgrade.bin", "relativePath" : "bin/targets/ath79/tiny/openwrt-ath79-tiny-nec_wg600hp-squashfs-sysupgrade.bin" }, { "displayPath" : "openwrt-ath79-tiny-nec_wg600hp-uboot.bin", "fileName" : "openwrt-ath79-tiny-nec_wg600hp-uboot.bin", "relativePath" : "bin/targets/ath79/tiny/openwrt-ath79-tiny-nec_wg600hp-uboot.bin" }, { "displayPath" : "openwrt-ath79-tiny-nec_wr8750n-initramfs-factory.bin", "fileName" : "openwrt-ath79-tiny-nec_wr8750n-initramfs-factory.bin", "relativePath" : "bin/targets/ath79/tiny/openwrt-ath79-tiny-nec_wr8750n-initramfs-factory.bin" }, { "displayPath" : "openwrt-ath79-tiny-nec_wr8750n-initramfs-kernel.bin", "fileName" : "openwrt-ath79-tiny-nec_wr8750n-initramfs-kernel.bin", "relativePath" : "bin/targets/ath79/tiny/openwrt-ath79-tiny-nec_wr8750n-initramfs-kernel.bin" }, { "displayPath" : "openwrt-ath79-tiny-nec_wr8750n-squashfs-sysupgrade.bin", "fileName" : "openwrt-ath79-tiny-nec_wr8750n-squashfs-sysupgrade.bin", "relativePath" : "bin/targets/ath79/tiny/openwrt-ath79-tiny-nec_wr8750n-squashfs-sysupgrade.bin" }, { "displayPath" : "openwrt-ath79-tiny-nec_wr8750n-uboot.bin", "fileName" : "openwrt-ath79-tiny-nec_wr8750n-uboot.bin", "relativePath" : "bin/targets/ath79/tiny/openwrt-ath79-tiny-nec_wr8750n-uboot.bin" }, { "displayPath" : "openwrt-ath79-tiny-nec_wr9500n-initramfs-factory.bin", "fileName" : "openwrt-ath79-tiny-nec_wr9500n-initramfs-factory.bin", "relativePath" : "bin/targets/ath79/tiny/openwrt-ath79-tiny-nec_wr9500n-initramfs-factory.bin" }, { "displayPath" : "openwrt-ath79-tiny-nec_wr9500n-initramfs-kernel.bin", "fileName" : "openwrt-ath79-tiny-nec_wr9500n-initramfs-kernel.bin", "relativePath" : "bin/targets/ath79/tiny/openwrt-ath79-tiny-nec_wr9500n-initramfs-kernel.bin" }, { "displayPath" : "openwrt-ath79-tiny-nec_wr9500n-squashfs-sysupgrade.bin", "fileName" : "openwrt-ath79-tiny-nec_wr9500n-squashfs-sysupgrade.bin", "relativePath" : "bin/targets/ath79/tiny/openwrt-ath79-tiny-nec_wr9500n-squashfs-sysupgrade.bin" }, { "displayPath" : "openwrt-ath79-tiny-nec_wr9500n-uboot.bin", "fileName" : "openwrt-ath79-tiny-nec_wr9500n-uboot.bin", "relativePath" : "bin/targets/ath79/tiny/openwrt-ath79-tiny-nec_wr9500n-uboot.bin" }, { "displayPath" : "openwrt-ath79-tiny.manifest", "fileName" : "openwrt-ath79-tiny.manifest", "relativePath" : "bin/targets/ath79/tiny/openwrt-ath79-tiny.manifest" }, { "displayPath" : "profiles.json", "fileName" : "profiles.json", "relativePath" : "bin/targets/ath79/tiny/profiles.json" }, { "displayPath" : "sha256sums", "fileName" : "sha256sums", "relativePath" : "bin/targets/ath79/tiny/sha256sums" }, { "displayPath" : "version.buildinfo", "fileName" : "version.buildinfo", "relativePath" : "bin/targets/ath79/tiny/version.buildinfo" } ], "building" : False, "description" : "diffconfig: true, device: WR8750N/WR9500N/WG600HP (AR9344)", "displayName" : "#519", "duration" : 2074465, "estimatedDuration" : 2120772, "executor" : None, "fullDisplayName" : "OpenWrt (master) for NEC Aterm (NetBSD based) #519", "id" : "519", "inProgress" : False, "keepLog" : False, "number" : 519, "queueId" : 84, "result" : "SUCCESS", "timestamp" : 1793230800229, "url" : "https://taiha.net/jenkins/job/OpenWrt-master-NEC-BSD-Aterm/519/", "builtOn" : "home-slave02_taihasv", "changeSet" : { "_class" : "hudson.plugins.git.GitChangeSetList", "items" : [ { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/realtek/files-6.18/drivers/net/dsa/rtl83xx/pie.c" ], "commitId" : "9d696f9ec1c41e74a4dafd846ecec5ca0052fe84", "timestamp" : 1793001377000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/markus.stockhausen", "fullName" : "markus.stockhausen" }, "authorEmail" : "markus.stockhausen@gmx.de", "comment" : "realtek: pie: serialize rule deletion with allocation\u000a\u000aPIE rule additions hold pie_mutex, but range deletion uses reg_mutex and\u000arule removal clears the allocation bitmap after dropping that lock. The\u000atwo paths can therefore access the engine concurrently. On RTL838x,\u000adeletion can also overwrite lookup changes made by a concurrent addition.\u000a\u000aUse pie_mutex for rule removal on RTL838x, RTL839x, RTL930x and RTL931x,\u000acovering both the hardware deletion and release of the allocated ID.\u000aRequire callers of the range-deletion helpers to hold that mutex and\u000aassert this with lockdep. Hold the same lock during PIE initialization,\u000aincluding the initial range clear and block configuration.\u000a\u000aThis serializes deletion with existing rule-add and RTL839x L3 write\u000apaths without nesting pie_mutex inside itself or acquiring reg_mutex.\u000a\u000aSigned-off-by: João Silva \u000aLink: https://github.com/openwrt/openwrt/pull/26190\u000aSigned-off-by: Markus Stockhausen \u000a", "date" : "2026-10-26 08:56:17 +0100", "id" : "9d696f9ec1c41e74a4dafd846ecec5ca0052fe84", "msg" : "realtek: pie: serialize rule deletion with allocation", "paths" : [ { "editType" : "edit", "file" : "target/linux/realtek/files-6.18/drivers/net/dsa/rtl83xx/pie.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/realtek/files-6.18/drivers/net/dsa/rtl83xx/pie.c" ], "commitId" : "14ba4bb0aa8928c98fbf5f3024bdfd2038348dd5", "timestamp" : 1793001377000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/markus.stockhausen", "fullName" : "markus.stockhausen" }, "authorEmail" : "markus.stockhausen@gmx.de", "comment" : "realtek: pie: bound rule deletion polling\u000a\u000aReplace the unbounded deletion-command loops on RTL838x, RTL839x,\u000aRTL930x and RTL931x with regmap_read_poll_timeout(). Return register\u000aaccess errors and stop waiting after 100 ms if the command stays busy.\u000a\u000aKeep affected rule IDs reserved after an error, since a timed-out command\u000amay still be pending in hardware. Stop PIE initialization if its initial\u000aclear fails, and only release individual IDs after successful deletion.\u000aOn RTL838x, restore block lookup only after the clear has completed.\u000a\u000aSigned-off-by: João Silva \u000aLink: https://github.com/openwrt/openwrt/pull/26190\u000aSigned-off-by: Markus Stockhausen \u000a", "date" : "2026-10-26 08:56:17 +0100", "id" : "14ba4bb0aa8928c98fbf5f3024bdfd2038348dd5", "msg" : "realtek: pie: bound rule deletion polling", "paths" : [ { "editType" : "edit", "file" : "target/linux/realtek/files-6.18/drivers/net/dsa/rtl83xx/pie.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/realtek/files-6.18/drivers/net/dsa/rtl83xx/pie.c" ], "commitId" : "3c5ccaaf58d9334d0a150dd5ffbd24da5e41f2d4", "timestamp" : 1793001377000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/markus.stockhausen", "fullName" : "markus.stockhausen" }, "authorEmail" : "markus.stockhausen@gmx.de", "comment" : "realtek: pie: convert IPv6 address and mask words consistently\u000a\u000aConvert IPv6 source and destination address words and their masks from\u000anetwork byte order before packing the RTL838x, RTL839x, RTL930x and\u000aRTL931x PIE entries. Use the inverse conversion when unpacking IPv6\u000adestination words in the RTL839x readback path.\u000a\u000aThis makes the register representation independent of CPU byte order\u000aand keeps address and mask handling consistent. Use a local word index\u000ato keep the conversions readable.\u000a\u000aSigned-off-by: João Silva \u000aLink: https://github.com/openwrt/openwrt/pull/26190\u000aSigned-off-by: Markus Stockhausen \u000a", "date" : "2026-10-26 08:56:17 +0100", "id" : "3c5ccaaf58d9334d0a150dd5ffbd24da5e41f2d4", "msg" : "realtek: pie: convert IPv6 address and mask words consistently", "paths" : [ { "editType" : "edit", "file" : "target/linux/realtek/files-6.18/drivers/net/dsa/rtl83xx/pie.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "rules.mk" ], "commitId" : "7375c575c3ea9871d41a1eaf63a2b9e96c63fc61", "timestamp" : 1793002233000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/ansuelsmth", "fullName" : "ansuelsmth" }, "authorEmail" : "ansuelsmth@gmail.com", "comment" : "rules: derive GCC_VERSION from the configuration\u000a\u000aThe toolchain writes its version into staging_dir/toolchain-*/info.mk,\u000awhich rules.mk includes. toolchain/Makefile creates that file from a\u000aplaceholder whose GCC_VERSION is the literal \"unknown\", and the only\u000athing that ever replaces it is Host/SetToolchainInfo, called from the\u000agcc prepare step.\u000a\u000aThe two live in different trees: the stamp that suppresses the prepare\u000astep sits in build_dir/toolchain-*, while info.mk sits in staging_dir.\u000aRemove staging_dir alone and the placeholder is reinstalled while the\u000aprepare stamp survives, so nothing rewrites it and every package is\u000abuilt with GCC_VERSION=unknown. apk rejects that outright:\u000a\u000a ERROR: info field 'version' has invalid value: package version is invalid\u000a ERROR: failed to create package: package version is invalid\u000a\u000aFor an internal toolchain the version is already known from\u000aCONFIG_GCC_VERSION, which rules.mk reads a few lines earlier as GCCV, so\u000afall back to it when info.mk offers nothing usable. TARGET_CROSS on the\u000aline above is defaulted the same way for the same reason.\u000a\u000aSigned-off-by: Daniel Golle \u000aLink: https://github.com/openwrt/openwrt/pull/26181\u000aSigned-off-by: Christian Marangi \u000a", "date" : "2026-10-26 09:10:33 +0100", "id" : "7375c575c3ea9871d41a1eaf63a2b9e96c63fc61", "msg" : "rules: derive GCC_VERSION from the configuration", "paths" : [ { "editType" : "edit", "file" : "rules.mk" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "package/firmware/linux-firmware/Makefile" ], "commitId" : "4dbc40c0630fccabc67c5ac7f613e6793a1d690e", "timestamp" : 1793002296000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/ansuelsmth", "fullName" : "ansuelsmth" }, "authorEmail" : "ansuelsmth@gmail.com", "comment" : "linux-firmware: update to 20261019\u000a\u000a% git log --no-merges --pretty=oneline --abbrev-commit 20260916...20261019\u000a158cd9ed6ebd rockchip: add PCIe3 PHY firmware for snps-pcie3\u000a733c58c2a8a9 Update ADSP firmware for Qualcomm qcs615 platform.\u000a4d3ecb245971 Revert \"amdgpu: update GC 11.0.0 firmware\"\u000ade4fccc398dc amdgpu: update vangogh firmware\u000a15f6f42c2991 amdgpu: update VPE 6.1.1 firmware\u000a347dcaed1ac0 amdgpu: update VCN 4.0.6 firmware\u000a2012998affe0 amdgpu: update PSP 14.0.1 firmware\u000a7a4a30b05a99 amdgpu: update GC 11.5.1 firmware\u000a9558ec3a8b69 amdgpu: update VCN 4.0.5 firmware\u000acf4b4addc1c4 amdgpu: update PSP 14.0.0 firmware\u000acab2fc931be4 amdgpu: update GC 11.5.0 firmware\u000a6deb84db6373 amdgpu: update renoir firmware\u000aa59975400aad amdgpu: update yellow carp firmware\u000a811ee4511577 amdgpu: update raven2 firmware\u000aa519eab5dca2 amdgpu: update raven firmware\u000a2d2cb01899a5 amdgpu: update VCN 3.1.2 firmware\u000a93b290fd8e35 amdgpu: update PSP 13.0.5 firmware\u000a9a3d7b4963aa amdgpu: update picasso firmware\u000aaf63ec014ba7 amdgpu: update PSP 13.0.11 firmware\u000afb52d937c35b amdgpu: update GC 11.0.4 firmware\u000aa4748ebfe8e9 amdgpu: update VCN 4.0.2 firmware\u000ad612c2461324 amdgpu: update PSP 13.0.4 firmware\u000a94daa4fe9a20 amdgpu: update GC 11.0.1 firmware\u000a7f7e805237cd amdgpu: update VCN 5.0.0 firmware\u000a85232b1cf902 amdgpu: update PSP 14.0.3 firmware\u000a58e210c52e72 amdgpu: update GC 12.0.1 firmware\u000aff4209304cec amdgpu: update PSP 14.0.2 firmware\u000a25cd19c34f83 amdgpu: update GC 12.0.0 firmware\u000a381bbd194f46 amdgpu: update VCN 4.0.4 firmware\u000abc9f817fb104 amdgpu: update PSP 13.0.7 firmware\u000a48d5ac9b57dc amdgpu: update GC 11.0.2 firmware\u000af547110145f6 amdgpu: update PSP 13.0.10 firmware\u000acc6d98c1dc6d amdgpu: update GC 11.0.3 firmware\u000a8dbdfc13ba7c amdgpu: update VCN 4.0.0 firmware\u000af8e341502af0 amdgpu: update PSP 13.0.0 firmware\u000a972f521e428f amdgpu: update GC 11.0.0 firmware\u000a9281ebe07688 amdgpu: update beige goby firmware\u000a921bdaa170f5 amdgpu: update dimgrey cavefish firmware\u000a77887010b3d8 amdgpu: update navy flounder firmware\u000a295a404bfc6b amdgpu: update sienna cichlid firmware\u000a9c2bed6472a6 amdgpu: update navi14 firmware\u000abb48b903db22 amdgpu: update navi12 firmware\u000aa97e3b5db17e amdgpu: update navi10 firmware\u000ada22110c7203 amdgpu: update VCN 5.0.1 firmware\u000a1cb1bc502d7a amdgpu: update GC 9.4.4 firmware\u000af102a868d206 amdgpu: update PSP 15.0.9 firmware\u000a5478478d2339 amdgpu: update GC 11.7.1 firmware\u000a36e99d99aa38 amdgpu: update VCN 5.3.0 firmware\u000ab1ca0bb6caa8 amdgpu: update PSP 15.0.0 firmware\u000a362b09d83a69 amdgpu: update PSP 14.0.5 firmware\u000a94cf07dfc413 amdgpu: update GC 11.5.3 firmware\u000ab8dadf247edd amdgpu: update PSP 14.0.4 firmware\u000a19b7ec9a3169 amdgpu: update GC 11.5.2 firmware\u000ae301e556e6bf amdgpu: update green sardine firmware\u000ac9944b924b18 amdgpu: update VCN 4.0.3 firmware\u000a6f6bfe6fa1e5 amdgpu: update PSP 13.0.6 firmware\u000aac2c59644fb2 amdgpu: update GC 9.4.3 firmware\u000a9ff0f3abf72d amdgpu: DMCUB updates for various ASICs\u000a50dd1bcf2f4b rtw89: init board_elm.bin with stuffs for board ID 1\u000a9f5eaf1e5fa1 rtw89: 8922d: update fw 0.35.119.3 with format suffix -1\u000a3dcdaa69640c rtw89: 8922d: update fw 0.35.115.2\u000a9936de9f9fdd wfx: update to firmware 3.18\u000abdeb9f0f93f1 nvidia: Update and add missing firmware files for Tegra SoCs\u000aee632be9e54c qcom:Update qdsp6sw firmware for shikra platform\u000a8d279b943f1f aic: Add AIC8800D80 Bluetooth and Wi-Fi firmware\u000a7609749670be qcom: update HPASS firmware for nord platform\u000af45ea109af75 amdgpu: DMCUB updates for various ASICs\u000a9ec083c3aba9 amdgpu: update gc 11.0.1 and 11.0.4 firmware\u000a3eadf1895677 qcom: update LPAICP firmware for shikra platform\u000a575bbe3f8356 ath12k: WCN7850 hw2.0: update board-2.bin\u000a5cdef65aaa63 panthor: Add initial firmware for Architecture 14.8 Arm Mali GPU\u000a6b5ab2d6f060 qcom: update ADSP firmware for glymur platform\u000a9c7ee0c8ef56 cirrus: cs35l41: Add firmware and tuning for Lenovo Thinkbook laptops\u000ac3bbdd9de36a cirrus: cs35l63: Add firmware for Cirrus CS35L63 for various Dell laptops\u000aae4b9e72fc5e cirrus: cs42l45: Add CS42L49 SDCA codec firmware for Dell laptops\u000a2180bb08bab1 qcom: fix qcs8300 gpdspr.jsn symlink target\u000a360b6a1c56da qcom: update ADSP firmware for x1e80100 platform\u000a6163f5ba0c33 linux-firmware: Update firmware file for Intel Solar core\u000ab95a9e931958 linux-firmware: Update firmware file for Intel Magnetar core\u000a2ce099f2c913 linux-firmware: Update firmware file for Intel BlazaruHrpGfp core\u000ab593e397a7c8 linux-firmware: Update firmware file for Intel BlazariGfp core\u000a458831d07217 linux-firmware: Update firmware file for Intel BlazarUFmPArl core\u000aa385fbdf1678 linux-firmware: Update firmware file for Intel BlazarI core\u000a8956e3383e94 linux-firmware: Update firmware file for Intel BlazariFmp core\u000a138a87a2591c linux-firmware: Update firmware file for Intel BlazarU core\u000a5c1ebeca629f linux-firmware: Update firmware file for Intel ScorpiusGfp core\u000a8a2941fae0d0 linux-firmware: Update firmware file for Intel Scorpius core\u000ac6e051da7ca9 cirrus: cs35l57 cs35l63: Add firmware for some Dell laptops\u000a185df4e8a45a WHENCE: add SPDX license references for non-standard license texts\u000a37d557ab3cc2 WHENCE: add initial SPDX identifiers\u000a8dcbe88cf1c0 cirrus: cs42l45: Add CS42L45 SDCA codec firmware for a Dell laptop\u000a675924146dab cirrus: cs35l41: Add support for Lenovo Yoga Slim 7 Carbon 14ACN6\u000a01d600e69036 linux-firmware: Add ISH firmware files for Intel Nova Lake-H/S platforms\u000a970505040744 qcom: add ADSP firmware for eliza platform\u000ae13358126f69 qcom: update ADSP firmware for hawi platform\u000af830e56c110c QCA: Add QCA2066 Bluetooth firmware hpnv21g.30c\u000a11184fe575c7 QCA: Update Bluetooth QCA2066 firmware to 2.1.0-00653\u000a5e7b5bc6a938 Revert \"qcom: update ADSP firmware for qcs615 platform\"q\u000a28b21de36022 qcom: Add gpu firmwares for Hawi and Maili chipsets\u000aa4c4b10b7ed6 qcom: update Rubik Pi 3 ADSP firmware\u000a4203b9738f72 cirrus: cs35l56: Correct firmware instances for 103c8c52 and 103c8c53\u000ad9dd66ff33ed contrib: post the MR comment to the project of the merge request\u000a83f41d84ccdc amdgpu: update VPE 2.0.0 firmware\u000abc5252f2c0de amdgpu: update VCN 5.3.0 firmware\u000ac3e8dd75984b amdgpu: update GC 11.7.0 MES firmware\u000a9ab11f1c5767 amdgpu: update PSP 15.0.9 firmware\u000a09c29d394433 amdgpu: update PSP 15.0.0 firmware\u000a39490329b234 qcom: vpu: add Gen2 firmware binary for Hawi\u000af1ea87d9cc84 qcom: vpu: add Gen2 firmware binary for Maili\u000ade45f8bef2ce linux-firmware: Update AMD cpu microcode\u000ab72356140866 iwlwifi: add Bz/Sc FW for core24.80-41 release\u000a092f5591c585 iwlwifi: Update Bz/Fm firmware for core24.80-41 release\u000a723c52333487 iwlwifi: Add Hr/Gf firmware for core24.80-41 release\u000a72caed25761a iwlwifi: update ty/So/Ma firmwares for core24.80-41 release\u000a0029ab7d3138 iwlwifi: Add cc/Qu/QuZ firmwares for core24.80-41 release\u000aca5fdc3e674b linux-firmware: mwifiex: add IW416 firmware\u000ae88807ffbf0f ath12k: WCN7850 hw2.0: update board-2.bin\u000a781bd88c531d ath12k: QCC2072 hw1.0: update board-2.bin\u000a4de2491cbe13 ci: add Qualcomm firmware version diff MR comment job\u000a4e1bb35244b7 qcom: sdx61: add missing firmware version\u000abd59c643be3c qcom: sdx35: add missing firmware version\u000ad04b22b2f7b1 qcom: aic100: add missing firmware versions\u000a77b75e7f7c09 qcom: qdu100: add missing firmware versions\u000a3cbed670aaa5 qcom: sdm845: add missing modem firmware version\u000ab259e2f9460c qcom: x1e80100: add missing Dell XPS 13 9345 DSP versions\u000a4d9d409cddcb qcom: add missing Adreno ZAP shader versions\u000a630e8b53cc2a contrib: report the version of the Qualcomm Bluetooth firmware\u000a28b24eb3bacf contrib: detect the version of Adreno GMU firmware\u000a0e45740a8632 contrib: detect the version of Adreno RGMU firmware\u000a968bab9bb0a8 contrib: detect the version of Adreno ZAP shaders\u000a4aeb44e73c5f contrib: detect the version of Adreno SQE and AQE microcode\u000a019260ca4ee3 contrib: report Iris video firmware versions in the WHENCE format\u000a30a63ce57b85 contrib: add Qualcomm firmware version scripts\u000a5d6ac99fb005 contrib: move common firmware script helpers to fw_helpers.py\u000a6a3381ca8301 qcom: use the WHENCE version format for the x1e80100 video firmware\u000a5dd792fbcbbc qca: separate the WCN685x uart firmware revisions\u000a16726a813864 qca: group the QCA61x4 USB firmware files\u000a8e96c7fb4935 qca: drop duplicate version lines\u000affce0cc01642 qca: fix the version of the QCA6698 Bluetooth firmware\u000aeedcf262a66a qca: fix the version of the WCN785x USB firmware\u000a828db6bf0dc9 qcom: fix the version of the a660 GMU firmware\u000ad77a7ff09172 qcom: fix the version of the Eliza SQE microcode\u000af35ac0f38f83 qcom: fix the version of the QCS615 ZAP shader\u000a837ae0e5becc qcom/shikra: link WiFi firmware from the ath10k subdir\u000a9a7c283fa003 Revert \"amdgpu: update GC 11.0.0 firmware\"\u000a0a5565a4a888 Revert \"amdgpu: update GC 11.0.0 firmware\"\u000af5e8adfa1e0c amdgpu: DMCUB updates for various ASICs\u000aa98a6daa9f4d intel: Update ibt-00a0-01a1-pci.ddc for BE211 - Whale Peak2 (WhP2)\u000a8fe920c980c3 ath11k: WCN6855 hw2.0: update board-2.bin\u000a30f83ae033e8 cirrus: cs42l45: Add CS42L45 SDCA codec firmware for Dell laptops\u000ac664be6ce9be qcom: add CDSP firmware for hawi platform\u000a10f09227a370 cirrus: cs42l45: Add CS42L45 SDCA codec firmware for Dell laptops\u000a2ded07e3da99 cirrus: cs35l63: Add Cirrus CS35L63 firmware mappings for some Dell laptops\u000ac50e97d36b10 QCA: Update Bluetooth firmware for QCC2072 UART interface: 1.1.0-00340 to 1.1.0-00355\u000a\u000aSigned-off-by: John Audia \u000aLink: https://github.com/openwrt/openwrt/pull/26125\u000aSigned-off-by: Christian Marangi \u000a", "date" : "2026-10-26 09:11:36 +0100", "id" : "4dbc40c0630fccabc67c5ac7f613e6793a1d690e", "msg" : "linux-firmware: update to 20261019", "paths" : [ { "editType" : "edit", "file" : "package/firmware/linux-firmware/Makefile" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "tools/mold/Makefile" ], "commitId" : "85398d134e0cce59efbd3c6f384a2da3df8dbfce", "timestamp" : 1793003639000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "mold: bump to 2.42.1\u000a\u000aUpdate mold to 2.42.1\u000a\u000aChangelog:\u000ahttps://github.com/rui314/mold/releases/tag/v2.42.1\u000a\u000aThis is the last version of mold based on C++, newer versions are based\u000aon Rust which needs different build systems.\u000a\u000aSigned-off-by: Carsten Schuette \u000aLink: https://github.com/openwrt/openwrt/pull/26188\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:33:59 +0100", "id" : "85398d134e0cce59efbd3c6f384a2da3df8dbfce", "msg" : "mold: bump to 2.42.1", "paths" : [ { "editType" : "edit", "file" : "tools/mold/Makefile" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_scheduler.c" ], "commitId" : "b771d55b76a8606146eb577e9f06f9647a33ea9a", "timestamp" : 1793004286000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: give a shaped mqprio class ten milliseconds of burst\u000a\u000aAn mqprio class was programmed with a bucket of one frame. A sender's\u000abursts leave the class's queues empty between them, and the tokens that\u000aaccrue meanwhile are lost, so the class runs well below its rate. Size\u000athe bucket to ten milliseconds of the class's rate, as cpu_port_rate\u000aalready does.\u000a\u000aOn an AX3600 with SQM's hw_ppe.qos at 305/165 Mbit/s, a single TCP\u000adownload over IPv4 or IPv6 carries 278-284 Mbit/s instead of 226-268\u000aMbit/s (3 runs each), and loaded RTT to the server stays within 0.3 ms\u000aof idle.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:46 +0100", "id" : "b771d55b76a8606146eb577e9f06f9647a33ea9a", "msg" : "qualcommax: qca_ppe: give a shaped mqprio class ten milliseconds of burst", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_scheduler.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_main.c" ], "commitId" : "1152a41775cabdbae6bce089d3cc10a40ccb4c2d", "timestamp" : 1793004286000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: record VSI 0 for a port outside a bridge\u000a\u000aqca_ppe_setup() and qca_ppe_port_bridge_leave() bind a port outside a\u000abridge to VSI 0 but record PPE_VSI_INVALID for it in port_vsi[]. Record\u000athe VSI the port is bound to. VSI 0 is reserved at setup, so no bridge\u000acompares equal to it.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:46 +0100", "id" : "1152a41775cabdbae6bce089d3cc10a40ccb4c2d", "msg" : "qualcommax: qca_ppe: record VSI 0 for a port outside a bridge", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_main.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_scheduler.c" ], "commitId" : "4a2fc6517d6a18d749d53e0e1907fcf69cd76487", "timestamp" : 1793004286000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: size a shaped port's queue from the tbf limit\u000a\u000aA shaped port's queue limit is 1 ms of the rate, too shallow to hold one\u000aflow at that rate, and a static limit per hash queue gives one flow a\u000aquarter of the depth at most. Take the depth from the tbf limit, read\u000afrom the bfifo tbf keeps it in, falling back to 1 ms, and set it before\u000athe rate is programmed so a refused rate restores the previous limit.\u000aEach shaped port's band queues share a dynamic admission group of that\u000adepth, and the queues of its rate-limited mqprio classes share another\u000aof 10 ms of each queue's rate. The classes of several ports at the same\u000adepth share one group, as each carries a share of the same download, such\u000aas every LAN port's band below one WAN rate. Three groups are free; a\u000afurther bottleneck splits its depth into static per-queue limits. Every\u000asuch queue has FORCE_AC set and the largest dynamic weight.\u000a\u000aMeasured on IPQ8074 at 50 Mbit/s with a 312 KB limit, one TCP flow to a\u000aserver 171 ms away, n=3: 38-40 Mbit/s with the depth shared, 14-17 Mbit/s\u000awith it split per queue. With SQM shaping the WAN and the download bands\u000aof lan1, lan2 and lan3, the three bands share one group and the WAN keeps\u000aanother, where the third LAN port fell back to static limits before.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:46 +0100", "id" : "4a2fc6517d6a18d749d53e0e1907fcf69cd76487", "msg" : "qualcommax: qca_ppe: size a shaped port's queue from the tbf limit", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_scheduler.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/patches-6.18/0953-net-ethernet-qca-add-EDMA.patch", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_main.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h", "target/linux/qualcommax/patches-6.18/0952-net-ethernet-qca-add-ppe.patch" ], "commitId" : "3844ee698bd81d362a2830985fe81603a3544828", "timestamp" : 1793004286000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: add PPE flow table access and debugfs dump\u000a\u000aThe PPE flow, host, nexthop, egress L3 interface and public address tables\u000aare unused. Add their register definitions and op engine accessors, enable\u000athe lookup with forward on miss so nothing changes yet, set the bucket\u000afunction of both hash blocks explicitly, and dump the flow table through\u000adebugfs. An entry is read back through the op engine because an IPv6 entry\u000aspans two flat slots, and a valid host entry is removed through the host op\u000aengine, not by zeroing it. Table sizes are per-SoC data, and the age period\u000afollows the board's PPE clock rate.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:46 +0100", "id" : "3844ee698bd81d362a2830985fe81603a3544828", "msg" : "qualcommax: qca_ppe: add PPE flow table access and debugfs dump", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/patches-6.18/0953-net-ethernet-qca-add-EDMA.patch" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_main.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h" }, { "editType" : "add", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/patches-6.18/0952-net-ethernet-qca-add-ppe.patch" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/patches-6.18/0953-net-ethernet-qca-add-EDMA.patch", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_scheduler.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_main.c", "target/linux/qualcommax/patches-6.18/0952-net-ethernet-qca-add-ppe.patch", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow_offload.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_vlan.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h" ], "commitId" : "882844eb2319d584df9970d07ab383cb7d73c45d", "timestamp" : 1793004286000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: add netfilter flowtable hardware offload\u000a\u000aRouted and NAT connections are forwarded by the CPU although the PPE has a\u000aflow table. Implement flowtable offload as mtk_ppe does: each rule becomes\u000aone route, SNAT or DNAT entry with shared, refcounted host, nexthop,\u000aegress L3 interface and public address entries, and the stats callback\u000areads the entry's age and counters and credits the bridge and VLAN uppers.\u000aEach egress L3 interface owns its index and carries the egress MTU,\u000abecause a nexthop naming interface 0 sent every LAN-to-WAN PPPoE packet to\u000athe CPU as L3_EXP_MTU_DF_FAIL, and a PPPoE-over-802.1Q uplink gets a\u000adedicated VSI with VLAN translation and a PPPoE session entry so its\u000adownload reaches the flow lookup. Rules rewriting both addresses, IPv6\u000aNAT, non-switch ingress and untranslated ingress tags stay in software,\u000afragments bypass the lookup, and flows are purged when their VSI is\u000areleased, a port's MTU changes or the flowtable block is released.\u000aDSA hands a flowtable bound to a user port to the conduit, as for\u000amtk_eth, and qca_edma passes it on to the switch.\u000a\u000aOn IPQ8074 with a PPPoE-over-VID7 uplink, both directions of an offloaded\u000aconnection stay off the conduit. A routed flow that leaves by its ingress\u000aport, a VLAN on a stick for one, is forwarded by its entry as well:\u000arouted TCP between two subnets on one port ran at 866-882 Mbit/s with the\u000areceive softirq rate at its idle level (3 runs).\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:46 +0100", "id" : "882844eb2319d584df9970d07ab383cb7d73c45d", "msg" : "qualcommax: qca_ppe: add netfilter flowtable hardware offload", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/patches-6.18/0952-net-ethernet-qca-add-ppe.patch" }, { "editType" : "add", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow_offload.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_main.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_vlan.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_scheduler.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/patches-6.18/0953-net-ethernet-qca-add-EDMA.patch" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow_offload.c" ], "commitId" : "46e3391b905f696d4d932c62effceaa78275595f", "timestamp" : 1793004286000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: offload GRE flowtable flows\u000a\u000aA GRE flow has no ports key, and the driver refuses it. Encode it as a\u000athree-tuple entry whose protocol field names no L4 protocol and which\u000acarries the IP protocol where the source port would be. The flowtable never\u000aoffloads NATed GRE, so these entries only route. fw4 offers only tcp and\u000audp, so a stock config needs a flow add rule that names gre.\u000a\u000aOn IPQ8074 IPv4 and IPv6 GRE packets go through one hardware entry.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:46 +0100", "id" : "46e3391b905f696d4d932c62effceaa78275595f", "msg" : "qualcommax: qca_ppe: offload GRE flowtable flows", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow_offload.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_scheduler.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow_offload.c" ], "commitId" : "dbdb7aa6d76574c2c3a7c0f129d52ee88e152e64", "timestamp" : 1793004286000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: prioritise sparse flows on a shaped port\u000a\u000aOn a shaped port a low-rate flow shares the bulk FIFO, takes its loss rate\u000aand recovers by RTO, so the latency probe of a speed test stalls. Use each\u000auser port's four unused unicast queues as a list one strict priority above\u000athe download band, inside its shaper, reached from flow priority profile\u000a12. A work every 100 ms re-adds a flow with that profile after ten periods\u000aunder eight packets, and back to its own profile once it exceeds eight in a\u000aperiod; the same pass credits counter deltas to the upper interfaces. The\u000asparse_flows module parameter turns it off.\u000a\u000aMeasured on IPQ8074 with a PPPoE uplink shaped to 305/165 Mbit/s, wired:\u000aLibreQoS +1/+1/+1 ms, against +1 ms in nine of fourteen runs before.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:46 +0100", "id" : "dbdb7aa6d76574c2c3a7c0f129d52ee88e152e64", "msg" : "qualcommax: qca_ppe: prioritise sparse flows on a shaped port", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow_offload.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_scheduler.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/airoha/patches-6.18/161-01-v7.2-net-airoha-Introduce-airoha_gdm_dev-struct.patch", "target/linux/airoha/patches-6.18/915-01-net-netfilter-flowtable-Add-the-capability-to-offloa.patch", "target/linux/generic/pending-6.18/708-netfilter-flowtable-tear-down-direct-xmit-flows-when.patch", "target/linux/airoha/patches-6.18/099-08-v6.19-net-airoha-ppe-Flush-PPE-SRAM-table-during-PPE-setup.patch", "target/linux/generic/pending-6.18/708-00-net-mlx5e-ignore-FLOW_ACTION_PRIORITY-on-flowtable-offload.patch", "target/linux/generic/pending-6.18/736-04-net-ethernet-mediatek-fix-ppe-flow-accounting-for-L2.patch", "target/linux/airoha/patches-6.18/165-01-v7.2-net-airoha-Remove-private-net_device-pointer-in-airo.patch", "target/linux/airoha/patches-6.18/191-v7.2-net-airoha-add-preliminary-support-to-configure-tx-h.patch", "target/linux/generic/pending-6.18/708-netfilter-flowtable-carry-a-priority-into-the-offload.patch", "target/linux/airoha/patches-6.18/134-v7.1-net-airoha-Delay-offloading-until-all-net_devices-ar.patch", "target/linux/airoha/patches-6.18/165-05-v7.2-net-airoha-Support-multiple-LAN-WAN-interfaces-for-h.patch", "target/linux/generic/hack-6.18/650-netfilter-add-xt_FLOWOFFLOAD-target.patch", "target/linux/airoha/patches-6.18/161-02-v7.2-net-airoha-Move-airoha_qdma-pointer-in-airoha_gdm_de.patch", "target/linux/generic/pending-6.18/795-01-net-ethernet-mtk_ppe_offload-use-rhashtable_lookup_f.patch", "target/linux/generic/pending-6.18/795-03-net-ethernet-mtk_ppe_offload-set-output-device-befor.patch", "target/linux/airoha/patches-6.18/185-v7.2-net-airoha-Fix-error-handling-in-airoha_ppe_flush_sr.patch", "target/linux/airoha/patches-6.18/915-02-net-airoha-Set-hw-QoS-parameter-according-to-the-pac.patch" ], "commitId" : "9e692c0e7b52334245a65e2009464238a3b112ab", "timestamp" : 1793004286000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "generic: add FLOW_ACTION_PRIORITY to the flowtable offload\u000a\u000aAdd 708, the netfilter-devel posting prepared for nf-next: the flowtable\u000astores the priority a rule set before \"flow add\", applies it in the\u000asoftware fast path and passes it to drivers as FLOW_ACTION_PRIORITY. An\u000aIPv4 priority that ip_forward() derived from the TOS is not stored,\u000asince it belongs to each packet; the fast path derives it per packet\u000ainstead. 708-00 makes mlx5 ignore the action on flowtable flows so they\u000astay offloaded. qca_ppe uses it in the next patch. The airoha, mediatek\u000aand xt_FLOWOFFLOAD patches are refreshed.\u000a\u000aLink: https://lore.kernel.org/netfilter-devel/20261004171636.3545085-1-julius@bairaktaris.de/\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:46 +0100", "id" : "9e692c0e7b52334245a65e2009464238a3b112ab", "msg" : "generic: add FLOW_ACTION_PRIORITY to the flowtable offload", "paths" : [ { "editType" : "edit", "file" : "target/linux/airoha/patches-6.18/191-v7.2-net-airoha-add-preliminary-support-to-configure-tx-h.patch" }, { "editType" : "add", "file" : "target/linux/generic/pending-6.18/708-netfilter-flowtable-carry-a-priority-into-the-offload.patch" }, { "editType" : "edit", "file" : "target/linux/airoha/patches-6.18/165-01-v7.2-net-airoha-Remove-private-net_device-pointer-in-airo.patch" }, { "editType" : "add", "file" : "target/linux/generic/pending-6.18/708-00-net-mlx5e-ignore-FLOW_ACTION_PRIORITY-on-flowtable-offload.patch" }, { "editType" : "edit", "file" : "target/linux/airoha/patches-6.18/161-01-v7.2-net-airoha-Introduce-airoha_gdm_dev-struct.patch" }, { "editType" : "edit", "file" : "target/linux/airoha/patches-6.18/915-01-net-netfilter-flowtable-Add-the-capability-to-offloa.patch" }, { "editType" : "edit", "file" : "target/linux/generic/pending-6.18/795-01-net-ethernet-mtk_ppe_offload-use-rhashtable_lookup_f.patch" }, { "editType" : "edit", "file" : "target/linux/generic/hack-6.18/650-netfilter-add-xt_FLOWOFFLOAD-target.patch" }, { "editType" : "edit", "file" : "target/linux/airoha/patches-6.18/134-v7.1-net-airoha-Delay-offloading-until-all-net_devices-ar.patch" }, { "editType" : "edit", "file" : "target/linux/airoha/patches-6.18/161-02-v7.2-net-airoha-Move-airoha_qdma-pointer-in-airoha_gdm_de.patch" }, { "editType" : "edit", "file" : "target/linux/generic/pending-6.18/708-netfilter-flowtable-tear-down-direct-xmit-flows-when.patch" }, { "editType" : "edit", "file" : "target/linux/airoha/patches-6.18/915-02-net-airoha-Set-hw-QoS-parameter-according-to-the-pac.patch" }, { "editType" : "edit", "file" : "target/linux/generic/pending-6.18/795-03-net-ethernet-mtk_ppe_offload-set-output-device-befor.patch" }, { "editType" : "edit", "file" : "target/linux/airoha/patches-6.18/165-05-v7.2-net-airoha-Support-multiple-LAN-WAN-interfaces-for-h.patch" }, { "editType" : "edit", "file" : "target/linux/airoha/patches-6.18/185-v7.2-net-airoha-Fix-error-handling-in-airoha_ppe_flush_sr.patch" }, { "editType" : "edit", "file" : "target/linux/generic/pending-6.18/736-04-net-ethernet-mediatek-fix-ppe-flow-accounting-for-L2.patch" }, { "editType" : "edit", "file" : "target/linux/airoha/patches-6.18/099-08-v6.19-net-airoha-ppe-Flush-PPE-SRAM-table-during-PPE-setup.patch" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow_offload.c" ], "commitId" : "3a26e63cc6fd5db4285b55e7e813b329894a0250", "timestamp" : 1793004286000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: offload FLOW_ACTION_PRIORITY to the flow entry\u000a\u000aAn offloaded flow never returns to the CPU, so its queue cannot follow the\u000aruleset's classification. Write the flowtable's FLOW_ACTION_PRIORITY into\u000athe entry's priority profile, which outranks DSCP and whose table maps each\u000aprofile to the same priority; profile 0 leaves DSCP to decide. A priority\u000aabove PPE_QOS_MAX_PRI, such as the 802.1d tags of 256 and up that Wi-Fi\u000astamps take, is ignored and leaves the queue to DSCP, rather than keeping\u000athe whole flow in software.\u000a\u000aOn IPQ8074 with meta priority set 0:5, IPv6 and IPv4 NAT packets of the\u000aflow leave on queue 5, and an unmarked flow is still classified by DSCP.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:46 +0100", "id" : "3a26e63cc6fd5db4285b55e7e813b329894a0250", "msg" : "qualcommax: qca_ppe: offload FLOW_ACTION_PRIORITY to the flow entry", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow_offload.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/airoha/patches-6.18/915-01-net-netfilter-flowtable-Add-the-capability-to-offloa.patch", "target/linux/generic/pending-6.18/709-netfilter-flowtable-advertise-the-vlan-match-in-used_keys.patch" ], "commitId" : "1b90dab6216e8b8d254bf6edd80d04b6673cfad1", "timestamp" : 1793004286000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "generic: advertise the flowtable vlan match\u000a\u000anf_flow_rule_match() builds a vlan match but does not set it in\u000aused_keys, so drivers never see the tag. 709 sets the bits; it is v3 as\u000asent to netfilter-devel. The airoha 915-01 patch is refreshed.\u000a\u000aLink: https://lore.kernel.org/netfilter-devel/20260913104251.648421-1-julius@bairaktaris.de/\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:46 +0100", "id" : "1b90dab6216e8b8d254bf6edd80d04b6673cfad1", "msg" : "generic: advertise the flowtable vlan match", "paths" : [ { "editType" : "edit", "file" : "target/linux/airoha/patches-6.18/915-01-net-netfilter-flowtable-Add-the-capability-to-offloa.patch" }, { "editType" : "add", "file" : "target/linux/generic/pending-6.18/709-netfilter-flowtable-advertise-the-vlan-match-in-used_keys.patch" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow_offload.c" ], "commitId" : "59c1ffad8b370bd334478d37753b19f5210f8b5e", "timestamp" : 1793004286000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: map the flowtable's ingress VLAN to a VSI\u000a\u000aA rule whose ingress frame keeps an 802.1Q tag now carries the tag as a\u000amatch key, but a flow entry has no tag field, only the VSI. Map the tag to\u000athe uplink's VSI on a PPPoE port or the VLAN's VSI on a bridge member; a\u000atag with no such classification, or a second tag, stays in software. Before\u000athe previous patch the key never arrived and a tagged ingress was matched\u000aon the PVID's VSI.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:46 +0100", "id" : "59c1ffad8b370bd334478d37753b19f5210f8b5e", "msg" : "qualcommax: qca_ppe: map the flowtable's ingress VLAN to a VSI", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow_offload.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow_offload.c" ], "commitId" : "114c4091f14dba9e2aca1a299e1d819d25cb29d7", "timestamp" : 1793004286000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: send slow-path packets of offloaded flows to the CPU\u000a\u000aThe flowtable matches every TCP flow with FIN and RST clear, but the driver\u000aignored FLOW_DISSECTOR_KEY_TCP and the hardware forwarded the teardown, so\u000aan offloaded connection only ended on the idle timeout. The software fast\u000apath also leaves IPv4 options, IPv6 extension headers and short TCP or UDP\u000aheaders to the stack, and a SYN on a closing flow tears it down. Program\u000aTCP flag exceptions for FIN, SYN and RST and enable them for routed flow\u000ahits with the header exceptions the software fast path leaves to the\u000astack, as qca-nss-ppe does: IPv4 options, IPv6 extension headers, a TTL\u000aor hop limit about to expire, fragments, and truncated or malformed L2,\u000aPPPoE, IP and L4 headers. Each redirects the packet to the CPU and leaves the flow entry\u000ain place, and every other exception is cleared. Accept the TCP key only\u000awhen every flag it excludes is one of the trapped ones.\u000a\u000aOn IPQ8074, over IPv4 and IPv6, a flow leaves hardware as the connection\u000acloses, and without the trap it stays offloaded until the idle timeout.\u000aPackets with IPv4 options, an IPv6 hop-by-hop header or a short TCP or\u000aUDP header on an offloaded flow reach the CPU, 3 of 3 each, and the flow\u000aentry stays (one run). 200 SYNs on an offloaded flow reach the CPU (two\u000aruns).\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:46 +0100", "id" : "114c4091f14dba9e2aca1a299e1d819d25cb29d7", "msg" : "qualcommax: qca_ppe: send slow-path packets of offloaded flows to the CPU", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow_offload.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow_offload.c" ], "commitId" : "aaa6875d7361b319242bb7769dc39cc84593f592", "timestamp" : 1793004286000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: use all sixteen PPPoE session entries\u000a\u000aThe uplink state was one per port and the PPPoE session table was indexed\u000aby port, so a port carried one session on one VLAN and a re-dial overwrote\u000athe session that older flows still matched on. Keep uplinks per port and\u000aVLAN, each with its own VSI and L3 interface, and give every session its\u000aown refcounted entry pointing at its uplink. Like the kernel's session\u000alookup, an entry matches the server's address too, so a frame with the\u000asession id from another source on the uplink stays in software. The same\u000aid from one server on two VLANs of one port is refused, as the hardware\u000acannot tell them apart.\u000a\u000aOn IPQ8074, with two sessions from a PPPoE server on one port, each session\u000aholds its own entry and the connections over both are offloaded. On a\u000aTelekom PPPoE uplink the entry reads back the server's address, and\u000adownloads stay offloaded at 260-284 Mbit/s behind a 305 Mbit/s shaper\u000awith the CPU idle (two servers, 2 runs each).\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:46 +0100", "id" : "aaa6875d7361b319242bb7769dc39cc84593f592", "msg" : "qualcommax: qca_ppe: use all sixteen PPPoE session entries", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow_offload.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow_offload.c" ], "commitId" : "3945d5a8ed3b4f2140c42c0e7ca9a4f029fa811d", "timestamp" : 1793004286000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: route a VLAN that no bridge VLAN classifies\u000a\u000aA VLAN device on a port outside any bridge, such as a DHCP WAN on\u000awan. or lan1.20, or on a port of a bridge that does not filter\u000aVLANs, has no VSI its tagged frames are classified into, so a flow\u000aarriving on it was refused as ingress_vlan_domain. Give such a port and\u000aVLAN the uplink a tagged PPPoE WAN uses, without a session, whichever\u000adirection of the flow is installed first. The L3 interface terminates\u000aPPPoE only once a session is attached to it. The offload debugfs file\u000alists the uplinks in use with the number of flows on each.\u000a\u000aOn IPQ8074, TCP routed both ways between VLANs 20 and 30 on one port of\u000aa bridge that does not filter VLANs runs at 933-936 Mbit/s combined with\u000athe softirq share of the CPU at 0% and no flow refused (3 runs), against\u000a907 Mbit/s at 21% with 53 flows refused before (1 run).\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:46 +0100", "id" : "3945d5a8ed3b4f2140c42c0e7ca9a4f029fa811d", "msg" : "qualcommax: qca_ppe: route a VLAN that no bridge VLAN classifies", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow_offload.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_main.c" ], "commitId" : "f7abd7bcaa5191804fc44e96f5a6a28ffb276c77", "timestamp" : 1793004286000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: add devlink resources for flow, PPPoE and meters\u000a\u000aThe flow table, the PPPoE session table and the ACL meters are the\u000atables the offload runs out of first, but devlink only reports the L2\u000atables. Register them as devlink resources with occupancy, so that\u000a\"devlink resource show\" says how full each one is.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:46 +0100", "id" : "f7abd7bcaa5191804fc44e96f5a6a28ffb276c77", "msg" : "qualcommax: qca_ppe: add devlink resources for flow, PPPoE and meters", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_main.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_acl.c" ], "commitId" : "713941a1ba8a60ca881aacd2b2ad0eb807200a6f", "timestamp" : 1793004286000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: default small_pkt_len to the qca-ppe-qos value\u000a\u000aqca-ppe-qos sets small_pkt_len to 1000, while the driver defaults to\u000a256. A module loaded without the package then leaves latency probes,\u000alarger DNS answers and game state in the bulk queue. Use 1000 as the\u000adriver default, which stays below a full-size segment on any path whose\u000aMSS is above 1000.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:46 +0100", "id" : "713941a1ba8a60ca881aacd2b2ad0eb807200a6f", "msg" : "qualcommax: qca_ppe: default small_pkt_len to the qca-ppe-qos value", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_acl.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.h", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c" ], "commitId" : "80884df9a23f024a574421e260868deee403a154", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_edma: keep the ring indices the driver owns in software\u000a\u000aThe data path read every ring index back from the engine through the\u000aregister map, including the producer index of the transmit and fill rings\u000aand the consumer index of the completion and receive rings, which only the\u000adriver moves. Each frame sent cost two register reads and a doorbell\u000awrite, and each poll about a dozen reads.\u000a\u000aKeep those indices in the queue, read the engine only for the indices it\u000amoves, and read the transmit consumer index only when the cached one\u000areports too little room. Write the transmit doorbell once per batch when\u000athe stack sets xmit_more, and always before the queue stops or a frame is\u000adropped. Access the index registers with readl() and writel(), which\u000aorder against the descriptor memory like the register map did.\u000a\u000aRouted Wi-Fi upload through a 1x1 station takes 4% fewer cycles per\u000apacket and 64-byte UDP upload 5% fewer (2 runs each).\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "80884df9a23f024a574421e260868deee403a154", "msg" : "qualcommax: qca_edma: keep the ring indices the driver owns in software", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.h" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/ipq807x/base-files/etc/init.d/smp_affinity", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.h", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c" ], "commitId" : "62e4e6ac5d2b2066c523dea20fed4c1a19b4531d", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_edma: poll in softirq context\u000a\u000aThreaded NAPI costs a context switch per poll, and the driver already\u000abinds each queue's interrupts to its own CPU, so softirq polling spreads\u000athe load the same way. Routed TCP from a wired host to a 2x2 HE160 station\u000atakes 11% fewer CPU cycles per packet upstream and 3% fewer downstream\u000awith one flow, 20% and 7% fewer with four flows (2 runs each). Threaded\u000aNAPI stays available through /sys/class/net//threaded.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "62e4e6ac5d2b2066c523dea20fed4c1a19b4531d", "msg" : "qualcommax: qca_edma: poll in softirq context", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/ipq807x/base-files/etc/init.d/smp_affinity" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.h" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.h", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c" ], "commitId" : "8d4ec787822d48b749b4957aa799b147a2821c33", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_edma: sync only the longest frame for the device\u000a\u000aThe page pool synced the whole receive buffer for the device each time a\u000apage was recycled, although no frame is longer than the MTU allows. Size\u000athe sync to the longest frame the ports admit and drop a longer one that\u000athe PPE still passes to the CPU, counting it as a length error. Cache\u000amaintenance on a routed 900 Mbit/s Wi-Fi downlink falls from 4.6% to\u000a3.6% of the CPU time (3 runs).\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "8d4ec787822d48b749b4957aa799b147a2821c33", "msg" : "qualcommax: qca_edma: sync only the longest frame for the device", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.h" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c" ], "commitId" : "3ce1d9eb181f9f2a851d84f9d6a6a9efdcbc0219", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_edma: read the next frame's headers ahead\u000a\u000aEvery received frame stalled on its first read of the preheader and the\u000aheaders behind it, which the receive sync had just evicted from the\u000acache. Read the next descriptor while the current frame is processed,\u000async its frame for the CPU and prefetch its first two cache lines.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "3ce1d9eb181f9f2a851d84f9d6a6a9efdcbc0219", "msg" : "qualcommax: qca_edma: read the next frame's headers ahead", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.h", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c" ], "commitId" : "6d5c2446833e24b6e8dd1883b9b46805b18c99e2", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_edma: mark frames the switch forwarded itself\u000a\u000aThe out-of-band tagger never sets offload_fwd_mark, so the bridge floods\u000aevery broadcast, multicast and unknown unicast frame from a switch port\u000aagain in software, after the PPE has flooded it in hardware. Frames the\u000aswitch forwarded or flooded to the CPU carry CPU code 0 in the receive\u000apreheader; hand those up with a METADATA_HW_PORT_MUX dst naming their\u000aport, as mtk_eth does, so DSA marks them and the tagger's skb extension\u000ais not allocated. A frame the switch trapped keeps its CPU code, goes\u000athrough the tagger and stays unmarked, so the bridge still forwards it.\u000a\u000aOn an AX3600, 1000 broadcasts sent on lan2 left lan1 and lan3 2000 times\u000aeach before and 1000 times after (3 runs), and 256 IGMP reports, which\u000athe switch traps, left each of them 256 times (3 runs).\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "6d5c2446833e24b6e8dd1883b9b46805b18c99e2", "msg" : "qualcommax: qca_edma: mark frames the switch forwarded itself", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.h" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c" ], "commitId" : "6914ff44db97010e5ae0d5da497f1b3f752f9312", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_edma: copy a transmitted head only when it is shared\u000a\u000aTransmit reallocated the head of every cloned skb before writing the\u000apreheader into its headroom, although a clone whose header was released,\u000aas TCP's are, may have its headroom written. Use skb_cow_head(), which\u000acopies the head only when the header is shared or the headroom is short.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "6914ff44db97010e5ae0d5da497f1b3f752f9312", "msg" : "qualcommax: qca_edma: copy a transmitted head only when it is shared", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/patches-6.18/0953-net-ethernet-qca-add-EDMA.patch", "target/linux/qualcommax/patches-6.18/0952-net-ethernet-qca-add-ppe.patch" ], "commitId" : "f3aac46ea7bc193dd06685f8294754abb6a1834e", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: declare what the PPE and EDMA drivers build on\u000a\u000aQCOM_80211AX_PPE registers a DSA switch on the out-of-band tag, takes its\u000aPCS from the UNIPHY driver and its registers through regmap, and QCOM_EDMA\u000auses the out-of-band tag's skb extension and page_pool, but neither says\u000aso: they only build because the target config enables all of it. Add the\u000adependencies and selects.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "f3aac46ea7bc193dd06685f8294754abb6a1834e", "msg" : "qualcommax: declare what the PPE and EDMA drivers build on", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/patches-6.18/0953-net-ethernet-qca-add-EDMA.patch" }, { "editType" : "edit", "file" : "target/linux/qualcommax/patches-6.18/0952-net-ethernet-qca-add-ppe.patch" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/pcs/pcs-qca-uniphy.c", "target/linux/qualcommax/files/include/linux/pcs/pcs-qca-uniphy.h" ], "commitId" : "62792fff96994a03981098b80d62c6c6cdeec06f", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: pcs-qca-uniphy: fix macro and clock callback slips\u000a\u000ato_qca_uniphy_pcs() ignored its argument and ended in a semicolon, the\u000aport clock index macros lacked outer parentheses, the reference clock's\u000ainit fields were chained with commas, and recalc_rate returned -EINVAL as\u000aa rate for an unknown UNIPHY. Use the argument, parenthesise the macros,\u000aend the statements and report a rate of 0. Drop the unused rst_ahb.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "62792fff96994a03981098b80d62c6c6cdeec06f", "msg" : "qualcommax: pcs-qca-uniphy: fix macro and clock callback slips", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/include/linux/pcs/pcs-qca-uniphy.h" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/pcs/pcs-qca-uniphy.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c" ], "commitId" : "c6a86d76b99c560f227c6886624bd237185f362d", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_edma: correct two comments\u000a\u000aThe transmit completion walk carried two comments saying the same thing,\u000aand the checksum comment read as if the driver accepted a padded frame's\u000asum, where it leaves the frame to software. Merge the first pair and\u000arestate the second.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "c6a86d76b99c560f227c6886624bd237185f362d", "msg" : "qualcommax: qca_edma: correct two comments", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_main.c" ], "commitId" : "bbb0d915e054c5482f23a06741b7494d565e9ace", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: keep a group's members when its lookup times out\u000a\u000aport_mdb_add starts a group from the CPU port alone on any lookup\u000aerror, so an operation that timed out rewrote an existing group with\u000aonly the CPU and the new port, dropping every other member. Start afresh\u000aonly when the group does not exist and pass any other error up.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "bbb0d915e054c5482f23a06741b7494d565e9ace", "msg" : "qualcommax: qca_ppe: keep a group's members when its lookup times out", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_main.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_main.c" ], "commitId" : "91ccd0713d44f687979f65305266c4881a704fdc", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: fold the CPU port's queue counters periodically\u000a\u000aThe fold work visits user ports only, so the CPU port's queue counters,\u000a32 bits of packets and 40 of bytes, were folded only when ethtool read\u000athem. At 1 Gbit/s the byte counter wraps in about 2.4 hours, and two\u000awraps between reads lose a whole period. Fold every available port.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "91ccd0713d44f687979f65305266c4881a704fdc", "msg" : "qualcommax: qca_ppe: fold the CPU port's queue counters periodically", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_main.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_main.c" ], "commitId" : "2ebcefb7526b24da2cd70221a6dbbb42b11a00f0", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: share the FDB operation code\u000a\u000aThe unicast and multicast FDB writes repeated the op engine sequence, the\u000aMAC packing was written out three times, and the LAG FDB ops copied the\u000aport ones apart from the destination. The LAG copies also took flow_lock,\u000awhich guards none of the FDB state; fdb_lock already serialises the op\u000aengine. Share the op engine write, the MAC packing and one FDB set helper\u000athat the port and LAG ops call.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "2ebcefb7526b24da2cd70221a6dbbb42b11a00f0", "msg" : "qualcommax: qca_ppe: share the FDB operation code", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_main.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_main.c" ], "commitId" : "fd5484ac4924d7dd21bc0df7329e5d149ac6c317", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: drop dead checks and accept an ethernet-ports node\u000a\u000appe_vlan_find() was declared twice, and four ethtool callbacks checked\u000athe port against num_ports, which DSA never exceeds. Probe and the\u000aIPQ6018 mux setup looked for a \"ports\" node only, while the binding and\u000athe DSA core also accept \"ethernet-ports\". Drop the duplicate and the\u000adead checks, and look for both names.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "fd5484ac4924d7dd21bc0df7329e5d149ac6c317", "msg" : "qualcommax: qca_ppe: drop dead checks and accept an ethernet-ports node", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_main.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_scheduler.c" ], "commitId" : "680b2858345d19bcea96d211fe482a4ad349956d", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: count only switched traffic in offloaded qdisc stats\u000a\u000aThe root tbf, ets and prio stats add the MAC's transmit counters to the\u000aqdisc's, but the MAC also sends every frame the host put through that\u000aqdisc, which the qdisc has counted itself, so host traffic was reported\u000atwice. Subtract what the host sent on the port over the same interval.\u000a\u000aOn a WAN port with an offloaded tbf, the qdisc reported the MAC's bytes\u000aplus its fq_codel child's before, and reports 177955309 bytes against the\u000aMAC's 177957143 after.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "680b2858345d19bcea96d211fe482a4ad349956d", "msg" : "qualcommax: qca_ppe: count only switched traffic in offloaded qdisc stats", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_scheduler.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_scheduler.c" ], "commitId" : "fcd37e330bed77dcb73c9d717852aaab58988322", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: refuse an mqprio class that splits a hashed band\u000a\u000aA priority selects a band of four queues and the flow hash picks the\u000aqueue in it, but mqprio's map was checked against priority i leaving on\u000aqueue i, a layout from before the hash spread. A class of one queue then\u000apassed the check and shaped a quarter of its priority's flows, chosen\u000aby hash. Require each class to own the whole band its priorities select.\u000ahw_ppe's map, 8@0 4@8, still passes.\u000a\u000aFixes: 3881df344eb (\"qualcommax: qca_ppe: spread flows across queues by RSS hash\")\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "fcd37e330bed77dcb73c9d717852aaab58988322", "msg" : "qualcommax: qca_ppe: refuse an mqprio class that splits a hashed band", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_scheduler.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_scheduler.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h" ], "commitId" : "8948da32427e35357867e61a0dd530e5350bccbb", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: keep a replacing mqprio's offload\u000a\u000atc qdisc replace of an offloaded mqprio installs the new qdisc first and\u000athen destroys the old one, whose offload removal carries no qdisc and\u000acleared the shapers and the transmit queue mapping the new qdisc had\u000ajust set. Count the port's offloaded mqprio qdiscs and ignore a removal\u000awhile another one is installed.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "8948da32427e35357867e61a0dd530e5350bccbb", "msg" : "qualcommax: qca_ppe: keep a replacing mqprio's offload", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_scheduler.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_scheduler.c" ], "commitId" : "6b9c6c5b6cb03d2cccd7fe42b446008e14414e0b", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: look a port's scheduler layout up in one place\u000a\u000aTen functions searched port_l0[] for the port with the same loop. Give\u000athe search a helper.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "6b9c6c5b6cb03d2cccd7fe42b446008e14414e0b", "msg" : "qualcommax: qca_ppe: look a port's scheduler layout up in one place", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_scheduler.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_scheduler.c" ], "commitId" : "94b80939378466333f918750b8fc40157038d85d", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: refuse mqprio frame preemption\u000a\u000amqprio's preemptible traffic classes were accepted and ignored. The MACs\u000ahave no frame preemption, so refuse them.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "94b80939378466333f918750b8fc40157038d85d", "msg" : "qualcommax: qca_ppe: refuse mqprio frame preemption", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_scheduler.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_scheduler.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_main.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_acl.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_debugfs.c" ], "commitId" : "fee126ace8a35af26aff3d965253a902a6836f3d", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: correct stale and wrong comments\u000a\u000aSeveral comments no longer matched the code: the qdisc band map claimed\u000anothing is programmed, the highest classifier priority was explained by a\u000anode sharing that was removed, the shaper state was described as tbf only,\u000athe root tbf comment sat above the gred code, the CPU port's admission\u000acomment said its scheduler is not built, and the ACL comment had ethtool\u000ainserting without rtnl, which it holds. Others narrated history or said\u000aFIXME. Correct or trim them.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "fee126ace8a35af26aff3d965253a902a6836f3d", "msg" : "qualcommax: qca_ppe: correct stale and wrong comments", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_scheduler.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_acl.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_debugfs.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_main.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c" ], "commitId" : "dd1227705a03ba479753267362823bbfca39b55a", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_edma: read a descriptor in one access\u000a\u000aThe descriptor rings are uncached, and the receive peek and the transmit\u000acompletion walk read a descriptor's address and status as separate 32-bit\u000aloads, the completion walk its status up to three times. Read both words\u000awith one 64-bit load.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "dd1227705a03ba479753267362823bbfca39b55a", "msg" : "qualcommax: qca_edma: read a descriptor in one access", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c" ], "commitId" : "2ca23dee6693fe78d5432d8e9fc3aa390d33d445", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_edma: prefetch the next segment of a frag_list frame\u000a\u000aCounting the descriptors of a frag_list frame walks its segments, and\u000aeach one misses on the skb and then on its shared info. Prefetch the next\u000asegment while the current one is counted.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "2ca23dee6693fe78d5432d8e9fc3aa390d33d445", "msg" : "qualcommax: qca_edma: prefetch the next segment of a frag_list frame", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_vlan.c" ], "commitId" : "33f270e580676520e2d8b4b0a4c9cf8b5a67c989", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: give a PVID-classified frame the VLAN's VID\u000a\u000aOn a VLAN-aware bridge, a frame that arrives untagged on a port and is\u000aclassified into its PVID VLAN carries no VID, so a port that egresses the\u000aVLAN tagged sends it with VID 0. Unicast between an untagged and a tagged\u000aport of the same VLAN therefore never arrives. Let the PVID rule add the\u000aVLAN's VID to the frame.\u000a\u000aOn an AX3600 with lan1 untagged in VLAN 1 and lan2 carrying VLAN 1\u000atagged, the PC's frames left lan2 tagged with VID 0 and its ping to a host\u000abehind lan2 failed; with the VID added, ping and TCP pass both ways.\u000a\u000aFixes: 142104bb90b (\"qualcommax: add PPE driver\")\u000aLink: https://github.com/openwrt/openwrt/issues/25904\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "33f270e580676520e2d8b4b0a4c9cf8b5a67c989", "msg" : "qualcommax: qca_ppe: give a PVID-classified frame the VLAN's VID", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_vlan.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/patches-6.18/0966-net-dsa-tag_oob-let-the-switch-send-frames-it-routed-to-the-CPU.patch", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_main.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.h", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow_offload.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h" ], "commitId" : "00ad5b53846b3a0ee386804a75f797c667167289", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: route flows to Wi-Fi through the CPU port\u000a\u000aA flow out of a device the switch cannot reach, such as a Wi-Fi\u000ainterface, was left to the software flowtable. Offload it with the CPU\u000aport as the egress port: the PPE routes and translates it and rewrites\u000athe Ethernet header, and the frame reaches the CPU finished, with the\u000aindex of the flow it matched in the receive preheader. qca_edma hands\u000asuch frames to the switch driver ahead of GRO, which queues them on the\u000aflow's egress device, so its qdisc, AQL and airtime fairness see them as\u000abefore. Routed TCP to a 2x2 HE160 station costs 27% fewer CPU cycles per\u000apacket over IPv4 (2 runs) and 23% fewer over IPv6 (1 run). The cpu_xmit\u000amodule parameter turns it off.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "00ad5b53846b3a0ee386804a75f797c667167289", "msg" : "qualcommax: qca_ppe: route flows to Wi-Fi through the CPU port", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.h" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow_offload.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_main.c" }, { "editType" : "add", "file" : "target/linux/qualcommax/patches-6.18/0966-net-dsa-tag_oob-let-the-switch-send-frames-it-routed-to-the-CPU.patch" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_main.c" ], "commitId" : "d6c75fb80965aa1b68d5333549ada9efc6cf151e", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: read an FDB lookup's entry before retiring its result\u000a\u000aThe FDB read results queue up: the data registers show the oldest, and\u000areading the result register retires it. The driver polled the result\u000aregister and then read the data, which by then was zero, so every\u000alookup missed. A port_mdb_add therefore always started the group from\u000athe CPU port alone and added another entry for the same group, and a\u000agroup with members on two ports was split over entries that each named\u000aone of them.\u000a\u000aRead the data before the result register in each poll, so the data read\u000ajust before the result that names the operation is its entry.\u000a\u000aOn an AX3600 with IGMP snooping, a group joined from lan1 and lan2 had\u000asix hardware entries, naming CPU and lan1 or CPU and lan2, and lan1\u000areceived none of the group's traffic; a lookup by hand returns the entry\u000awhen its data is read first.\u000a\u000aFixes: 0f9baab7f64 (\"qualcommax: qca_ppe: finish an FDB operation before reading its result\")\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "d6c75fb80965aa1b68d5333549ada9efc6cf151e", "msg" : "qualcommax: qca_ppe: read an FDB lookup's entry before retiring its result", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_main.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_acl.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow_offload.c" ], "commitId" : "68e25f82b10811cdbb99572bb9c4c35ecf54cbb0", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: keep PSH segments out of the small-packet priority\u000a\u000aA short TCP segment that carries PSH ends a write of a flow whose other\u000asegments are full size. Promoted by the small-packet rule, it overtakes\u000athem and reorders the flow on every write: a 16 Mbit/s download through\u000athe switch saw 35-42 reordered segments per 30 s, and none once the rule\u000amatches only frames with PSH clear (2-3 runs each). Acks, handshakes and\u000aevery protocol other than TCP, whose frames carry no TCP flags, still\u000amatch.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "68e25f82b10811cdbb99572bb9c4c35ecf54cbb0", "msg" : "qualcommax: qca_ppe: keep PSH segments out of the small-packet priority", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_acl.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow_offload.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "package/network/config/qca-ppe-qos/files/qca-ppe-qos.init" ], "commitId" : "fadbcb65d5699317d938a1ed46e23345c359e544", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qca-ppe-qos: keep PSH segments out of the Wi-Fi small-packet stamp\u000a\u000aThe u32 stamp moves short frames into a higher access category, so a\u000ashort TCP segment with PSH, which ends a write of a bulk flow, overtakes\u000athe flow's full-size segments. A download over Wi-Fi saw 240-460\u000areordered segments per 25 s, and none with two pass filters ahead of the\u000astamp for TCP segments with PSH set (3 runs each). The filters use fixed\u000aoffsets, so they cover IPv4 without options and IPv6 without extension\u000aheaders.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "fadbcb65d5699317d938a1ed46e23345c359e544", "msg" : "qca-ppe-qos: keep PSH segments out of the Wi-Fi small-packet stamp", "paths" : [ { "editType" : "edit", "file" : "package/network/config/qca-ppe-qos/files/qca-ppe-qos.init" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "package/network/config/qca-ppe-qos/files/qca-ppe-qos.init" ], "commitId" : "be48899f48a6c92ec9b2bd0551f674164ec09e33", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qca-ppe-qos: let 'enabled' switch only the marking rules\u000a\u000aWith enabled set to 0 the script tore everything down, the DSCP maps and\u000athe small-packet priority included, and skipped the Wi-Fi stamp, while\u000athe LuCI page offers the option as the switch for the marking rules.\u000aKeep the maps and the small-packet priority, and only flush the marking\u000afilters.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "be48899f48a6c92ec9b2bd0551f674164ec09e33", "msg" : "qca-ppe-qos: let 'enabled' switch only the marking rules", "paths" : [ { "editType" : "edit", "file" : "package/network/config/qca-ppe-qos/files/qca-ppe-qos.init" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "package/network/config/qca-ppe-qos/files/qca-ppe-qos.init" ], "commitId" : "bddefbe965f978b22e6dc26cbb2e95a43e53ea7f", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qca-ppe-qos: end the rule range at preference 509\u000a\u000aThe driver ranks a lower preference higher, so a filter at 511, where\u000ahw_ppe.qos marks its download band, ranks below every rule of this\u000ascript, not above as the comment said. Stop the range at 509, below the\u000asmall-packet rule's rank, and still flush up to 510, which earlier\u000aversions installed rules at.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "bddefbe965f978b22e6dc26cbb2e95a43e53ea7f", "msg" : "qca-ppe-qos: end the rule range at preference 509", "paths" : [ { "editType" : "edit", "file" : "package/network/config/qca-ppe-qos/files/qca-ppe-qos.init" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_scheduler.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h", "target/linux/qualcommax/patches-6.18/0952-net-ethernet-qca-add-ppe.patch", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow_offload.c" ], "commitId" : "8099efef1880f379fde94f90f51c058ddfef4b0a", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: hand shaped flows to a qdisc after the hardware routed them\u000a\u000aThe PPE has no queue that can run CAKE or mark ECN. A flow that crosses\u000athe shaped WAN link now keeps its lookup, NAT and encapsulation in\u000ahardware, and its finished frames reach the CPU only to pass a qdisc on\u000aa shaper device, one per direction, which the driver creates when the\u000adevlink parameter dl_shaper or ul_shaper names a device that does not\u000aexist yet. The PPE builds an upload frame whole, PPPoE and VLAN headers\u000aincluded. Once the qdisc releases a frame for a switch port, the shaper\u000adevice sends it straight into the conduit's ring; a download to Wi-Fi\u000agoes on to that interface. What the host forwards itself reaches the\u000asame devices through act_mirred, as it would reach an ifb: from the\u000auplink's egress, where the upload device builds the uplink's PPPoE and\u000aVLAN headers itself and sends the frame to the conduit too, and from the\u000aWAN port's ingress, from where the download device returns it to the\u000areceive path. Upload frames carry the conntrack the hardware path would\u000aotherwise lose, so CAKE's nat mode still tells the LAN hosts apart; a\u000adownload to a tagged LAN VLAN gets its tag after the shaper. An ifb\u000anamed instead works as well, at a higher cost. The devlink parameters\u000adl_shaper, ul_shaper, uplink and ct_dscp configure it, and sqm-scripts'\u000appe_cake.qos sets them.\u000a\u000aFlent from a wired client over a 300/150 Mbit/s PPPoE line shaped to\u000a330/168, 8 TCP streams each way: through the shaper devices the download\u000acarries 307.4 Mbit/s and the upload 157.2, as through IFBs, and the\u000arouter spends 13.4k CPU cycles per forwarded download packet instead of\u000a15.8k and 18.2k per upload packet instead of 22.4k; the four cores' load\u000ataken from idle-priority spinners falls from 26.5 to 21.6 % on the\u000adownload and from 19.4 to 15.8 % on the upload. Latency under load\u000astays at idle at the median through both; its 99th percentile rose by\u000a0.05-0.07 ms in six of seven runs and by 4.3 ms in one (3 runs each\u000aarm).\u000a\u000aLink: https://github.com/tohojo/sqm-scripts/pull/196\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "8099efef1880f379fde94f90f51c058ddfef4b0a", "msg" : "qualcommax: qca_ppe: hand shaped flows to a qdisc after the hardware routed them", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/patches-6.18/0952-net-ethernet-qca-add-ppe.patch" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow_offload.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_scheduler.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "package/network/config/qca-ppe-qos/Makefile" ], "commitId" : "be8b18e401fed78948828e66dcc281edf49da122", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qca-ppe-qos: name ppe_cake.qos in the package description\u000a\u000aThe classifiers pair with sqm-scripts' ppe_cake.qos as they do with\u000ahw_ppe.qos.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "be8b18e401fed78948828e66dcc281edf49da122", "msg" : "qca-ppe-qos: name ppe_cake.qos in the package description", "paths" : [ { "editType" : "edit", "file" : "package/network/config/qca-ppe-qos/Makefile" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c" ], "commitId" : "c31a509a4d0d787278b65b74e9a7be0333bb488d", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_edma: pick a multiqueue shaper's queue by the receive ring\u000a\u000aedma_rx_flow() hands a frame of an offloaded flow to its egress before\u000athe receive queue was recorded, so with cake_mq on the IFBs every such\u000aframe took the same queue. Record the ring first, so that each ring's\u000aCPU feeds a CAKE instance of its own.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "c31a509a4d0d787278b65b74e9a7be0333bb488d", "msg" : "qualcommax: qca_edma: pick a multiqueue shaper's queue by the receive ring", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_scheduler.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_debugfs.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h" ], "commitId" : "6cadbf46d30ae9294eb2ca9046497ba47edb2000", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_ppe: run CoDel's control law on the shaped queues\u000a\u000aThe queue manager drops a frame only on arrival, against a limit in\u000abuffers, so it bounds a queue but cannot tell a standing queue from a\u000aburst. A kthread reads the depth of every hashed queue of a shaped port\u000aand of the CPU port's download band, takes each bottleneck's delay from\u000aits backlog and rate, and when CoDel's law calls for a drop lowers the\u000alimit of each busy queue of that bottleneck to one frame under its depth\u000afor one tick. It ticks every 2 ms while a queue holds a full frame and\u000aevery 20 ms otherwise; aqm_tick_us=0 turns it off, and debugfs aqm shows\u000aits state.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "6cadbf46d30ae9294eb2ca9046497ba47edb2000", "msg" : "qualcommax: qca_ppe: run CoDel's control law on the shaped queues", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_scheduler.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_debugfs.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe.h" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.h" ], "commitId" : "698653058fe52a70d543a056a97e4a97eb047971", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: qca_edma: moderate receive interrupts over 100 us\u000a\u000aAt a shaped 330 Mbit/s download each receive interrupt found 2.4 frames,\u000athe 25 us moderation timer expiring long before the 16-frame threshold.\u000aWith 100 us it finds 6.7 (3.5 for the upload), and the router spends\u000a4.4 % fewer CPU cycles per forwarded packet on the download, with the\u000asame throughput and latency under load (Flent tcp_ndown 8 through CAKE,\u000a3 runs each; the upload's 1.8 % is within the 3 % that two identical\u000aconfigurations differ by). ethtool -C still sets it.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "698653058fe52a70d543a056a97e4a97eb047971", "msg" : "qualcommax: qca_edma: moderate receive interrupts over 100 us", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.h" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/patches-6.18/0966-net-dsa-tag_oob-let-the-switch-send-frames-it-routed-to-the-CPU.patch", "target/linux/qualcommax/files/include/linux/qca_edma.h", "target/linux/qualcommax/patches-6.18/0953-net-ethernet-qca-add-EDMA.patch", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_main.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.h", "target/linux/qualcommax/patches-6.18/0955-net-dsa-tag_qca-add-a-variant-behind-an-ethertype.patch", "target/linux/qualcommax/config-6.18", "target/linux/qualcommax/patches-6.18/0952-net-ethernet-qca-add-ppe.patch", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow_offload.c", "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c", "target/linux/qualcommax/patches-6.18/0950-net-dsa-add-out-of-band-tagging-protocol.patch" ], "commitId" : "b50530932b5a8dc7b086476ac930190ea6dc2d3b", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: tag with the QCA header instead of out of band\u000a\u000aThe out-of-band tagger was never merged upstream: it carries the port in\u000aan skb extension that DSA allocates for every frame sent, and that\u000aqca_edma allocated for every frame the switch trapped to the CPU. Use the\u000aQCA tagger instead, as airoha and mtk_eth_soc use their switch's in-band\u000atagger. On transmit, qca_edma reads the port out of the two-byte tag and\u000aremoves it once the frame is sure to go out, then names the port in the\u000apreheader as before. On receive, a forwarded frame keeps the metadata dst\u000athat DSA resolves without a tagger, and only a trapped frame, which the\u000abridge has to forward itself, gets the header the tagger reads its port\u000afrom. The switch's frames for a flow reach qca_ppe through a hook it\u000aregisters with qca_edma, and its shaper devices pass the port to\u000aqca_edma_port_xmit() directly, so the shaped path allocates nothing per\u000apacket. The tagger's needed headroom raises the conduit's MTU by 2.\u000a\u000aWired LAN, a VLAN-aware bridge, IGMP snooping with offloaded MDB\u000aentries, PPPoE over a VLAN uplink and Wi-Fi through the CPU port work\u000aunchanged on a Xiaomi AX3600.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "b50530932b5a8dc7b086476ac930190ea6dc2d3b", "msg" : "qualcommax: tag with the QCA header instead of out of band", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/config-6.18" }, { "editType" : "edit", "file" : "target/linux/qualcommax/patches-6.18/0952-net-ethernet-qca-add-ppe.patch" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_main.c" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.c" }, { "editType" : "delete", "file" : "target/linux/qualcommax/patches-6.18/0966-net-dsa-tag_oob-let-the-switch-send-frames-it-routed-to-the-CPU.patch" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_edma.h" }, { "editType" : "add", "file" : "target/linux/qualcommax/files/include/linux/qca_edma.h" }, { "editType" : "edit", "file" : "target/linux/qualcommax/patches-6.18/0953-net-ethernet-qca-add-EDMA.patch" }, { "editType" : "edit", "file" : "target/linux/qualcommax/patches-6.18/0955-net-dsa-tag_qca-add-a-variant-behind-an-ethertype.patch" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/drivers/net/ethernet/qualcomm/qca_ppe_flow_offload.c" }, { "editType" : "delete", "file" : "target/linux/qualcommax/patches-6.18/0950-net-dsa-add-out-of-band-tagging-protocol.patch" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/generic/backport-6.18/706-04-v7.3-netfilter-flowtable-use-the-vlan-id-not-the-full-tci.patch" ], "commitId" : "be75fbfa9b26eba3b60c291cff766c5fca8e929d", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "kernel: backport the flowtable VLAN ID fix\u000a\u000aBackport c203ecf90574, which keys the flowtable tuple on the 12-bit VLAN\u000aID instead of the whole TCI, so that frames whose tag carries a PCP\u000astill match their flow.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "be75fbfa9b26eba3b60c291cff766c5fca8e929d", "msg" : "kernel: backport the flowtable VLAN ID fix", "paths" : [ { "editType" : "add", "file" : "target/linux/generic/backport-6.18/706-04-v7.3-netfilter-flowtable-use-the-vlan-id-not-the-full-tci.patch" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/generic/pending-6.18/708-netfilter-flowtable-tear-down-direct-xmit-flows-when.patch" ], "commitId" : "09fe0f814008b287734a0582d985a9f99c6b3cdc", "timestamp" : 1793004287000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "generic: flowtable: check the bridge a roaming flow transmits on\u000a\u000aUpdate the imported fdb roaming teardown to v9. With a vlan device of one\u000abridge enslaved to another bridge, v8 counted the vlans below both\u000abridges, looked the bridge port up from the wrong device and read past\u000aencap[]; v9 resets the count at each bridge, so the gc checks the bridge\u000athe flow transmits on. v9 also carries Pablo's authorship and sign-off\u000afrom his posting with a note of the changes, instead of Co-developed-by.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:47 +0100", "id" : "09fe0f814008b287734a0582d985a9f99c6b3cdc", "msg" : "generic: flowtable: check the bridge a roaming flow transmits on", "paths" : [ { "editType" : "edit", "file" : "target/linux/generic/pending-6.18/708-netfilter-flowtable-tear-down-direct-xmit-flows-when.patch" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/generic/pending-6.18/643-net-dsa-fix-host-fdb-refcount-when-a-bridge-fdb-entry-roams.patch", "target/linux/generic/pending-6.18/642-net-bridge-switchdev-add-br_switchdev_fdb_find_port.patch" ], "commitId" : "c7924dcd74a6bd30abd5f935209528c071ddd287", "timestamp" : 1793004288000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "generic: dsa: track host fdb entries per port\u000a\u000aWhen a bridge fdb entry roams between ports, br_fdb_update() notifies\u000aswitchdev of the deletion and the addition without a lock, and DSA's\u000ahost address count on the CPU port goes wrong when two CPUs learn one\u000aaddress on different ports: \"qca-ppe 3a000000.ppe: port 5 failed to\u000adelete ... from fdb: -2\", and an address left on the CPU port.\u000a\u000aBackport the series prepared for net-next that makes each DSA user port\u000arecord the host addresses it holds and look up where the bridge has the\u000aentry when it handles each event, without a lock in the bridge's\u000alearning path.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/24806\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:44:48 +0100", "id" : "c7924dcd74a6bd30abd5f935209528c071ddd287", "msg" : "generic: dsa: track host fdb entries per port", "paths" : [ { "editType" : "add", "file" : "target/linux/generic/pending-6.18/642-net-bridge-switchdev-add-br_switchdev_fdb_find_port.patch" }, { "editType" : "add", "file" : "target/linux/generic/pending-6.18/643-net-dsa-fix-host-fdb-refcount-when-a-bridge-fdb-entry-roams.patch" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "package/network/config/firewall4/Makefile" ], "commitId" : "e2e50c9dc4b2dd57f4815ea9ff66c1c094bd21d8", "timestamp" : 1793004695000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "firewall4: update to Git HEAD (2026-10-26)\u000a\u000adf0c9021bbe7 ruleset: fix for missing counters handling code for ipsets\u000abac1d656e09b Do not emit helper jump without helper modules\u000a\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:51:35 +0100", "id" : "e2e50c9dc4b2dd57f4815ea9ff66c1c094bd21d8", "msg" : "firewall4: update to Git HEAD (2026-10-26)", "paths" : [ { "editType" : "edit", "file" : "package/network/config/firewall4/Makefile" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/patches-6.18/0953-net-ethernet-qca-add-EDMA.patch" ], "commitId" : "189012aed9ace46ca0f16650ce9b7136a631f05d", "timestamp" : 1793005059000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: refresh patches\u000a\u000aCI told us we need to do so, so lets do it.\u000a\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 09:57:39 +0100", "id" : "189012aed9ace46ca0f16650ce9b7136a631f05d", "msg" : "qualcommax: refresh patches", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/patches-6.18/0953-net-ethernet-qca-add-EDMA.patch" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/octeon/patches-6.18/900-cn70xx-fix-pem-bar-register-map.patch" ], "commitId" : "76d6b7fdbf0bc84bc4c89158faed8fe33beaf1e3", "timestamp" : 1793005782000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "octeon: fix PEM BAR register map on CN70XX\u000a\u000aOn CN70XX, the PEM register map differs from the other Octoen III models.\u000aPEM_BAR_CTL is at block offset 0xa0 and the BAR1 index table at 0x100, while\u000athe in-tree code uses the CN68XX offsets (BAR_CTL at 0x128, BAR1 index at 0xa8)\u000afor all Octeon III parts. The Cavium OCTEON-SDK cvmx-permx-defs.h special cases\u000aCN70XX for both registers.\u000a\u000aWith the generic offsets, __cvmx_pcie_rc_initialize_gen2() programs the wrong\u000aregisters on CN70XX. The intended BAR_CTL value lands on a BAR1 index entry,\u000aand the BAR1 index writes land on the real BAR_CTL (0xa8) and BAR2_MASK (0xb0).\u000aBAR2_MASK is ANDed with inbound (device DMA) addresses, so a bogus value there\u000asilently corrups all PCIe DMA to host memory. Observed on a CN7030 board while\u000adebugging a failing b43 firmware intialization followed by DMA errors.\u000a\u000aAdd model awareness helpers and use them for the three affected addresses.\u000a\u000aSigned-off-by: Andrew LaMarche \u000aLink: https://github.com/openwrt/openwrt/pull/26198\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 10:09:42 +0100", "id" : "76d6b7fdbf0bc84bc4c89158faed8fe33beaf1e3", "msg" : "octeon: fix PEM BAR register map on CN70XX", "paths" : [ { "editType" : "add", "file" : "target/linux/octeon/patches-6.18/900-cn70xx-fix-pem-bar-register-map.patch" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/ath79/dts/ar9344_wd_mynet-nxxx.dtsi" ], "commitId" : "c5e88157e9d33dca599a2e2f752b7b7bd4ee18d1", "timestamp" : 1793010935000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/ansuelsmth", "fullName" : "ansuelsmth" }, "authorEmail" : "ansuelsmth@gmail.com", "comment" : "ath79: use old SPI driver on WD MyNet to avoid filesystem errors\u000a\u000acomplaints about JFFS2 corruption on WD Mynet N750 devices started at 21.02.\u000ahttps://forum.openwrt.org/t/upgrade-from-barrier-breaker-14-07-to-latest-21-02-wd-my-net-n750/119833/\u000a\u000ait's still happening on 25.12, from my device:\u000ajffs2: notice: (594) jffs2_get_inode_nodes: Node header CRC failed at 0x006e9c.\u000ajffs2: notice: (723) jffs2_get_inode_nodes: Node header CRC failed at 0x006e58.\u000ajffs2: warning: (723) jffs2_do_read_inode_internal: no data nodes found for ino #37\u000affs2: notice: (723) jffs2_do_read_inode_internal: but it has children so we fake some modes for it\u000a\u000acommit ebf0d8dade (between 19.07 and 21.02) started using the new\u000aspi-ar934x driver for the ath79 target instead of spi-ath79.\u000a\u000ai don't know exactly what's wrong with the new driver with this device.\u000abut putting it back to the old one seems to fix it.\u000a\u000aI've been using my N750 successfully now for a week, changing configs,\u000ainstalling/removing packages, rebooting, yanking power.\u000aI don't get any more JFFS2 complaints, the device stays responsive and\u000aLUCI is usable as expected\u000a\u000aSigned-off-by: Ronald Ouyang \u000aLink: https://github.com/openwrt/openwrt/pull/26154\u000aSigned-off-by: Christian Marangi \u000a", "date" : "2026-10-26 11:35:35 +0100", "id" : "c5e88157e9d33dca599a2e2f752b7b7bd4ee18d1", "msg" : "ath79: use old SPI driver on WD MyNet to avoid filesystem errors", "paths" : [ { "editType" : "edit", "file" : "target/linux/ath79/dts/ar9344_wd_mynet-nxxx.dtsi" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/files/include/dt-bindings/net/pcs/qualcomm,ipq-uniphy.yaml", "target/linux/qualcommax/files/arch/arm64/boot/dts/qcom/ipq6018-ess.dtsi", "target/linux/qualcommax/files/arch/arm64/boot/dts/qcom/ipq8074-ess.dtsi" ], "commitId" : "a41558d2d3654938cbc22c1b3a9c6e68a1833089", "timestamp" : 1793015431000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: list UNIPHY clock output names RX first\u000a\u000aqca_uniphy_register_clks() registers each clock under the\u000aclock-output-names entry at its UNIPHY_CLK_* index, and the binding\u000aheader defines UNIPHY_CLK_RX as 0. The ipq6018 and ipq8074 nodes list\u000athe TX name first, so the RX clock carries the TX name and vice versa.\u000aSince GCC takes these clocks by phandle, debugfs reports every port TX\u000aRCG on uniphyN_gcc_rx_clk while the hardware selects uniphyN TX, and a\u000aparent lookup by name picks the wrong one. List the names RX first, as\u000aipq5018 already does, and make the binding match.\u000a\u000aTested on a Xiaomi AX3600 (ipq8074): nss_port1..5_tx_clk_src now report\u000auniphy0_gcc_tx_clk and the RX sources uniphy0_gcc_rx_clk, matching CFG\u000asrc 1 in their parent maps. The CFG registers and all rates are\u000aunchanged, and lan1 to lan2 iperf3 runs at 941 Mbit/s in each direction\u000a(one 5 s run each).\u000a\u000aFixes: 5df1a6ab80a (\"qualcommax: ipq60xx: pass UNIPHY clocks directly to GCC\")\u000aFixes: 49193000bf4 (\"qualcommax: ipq807x: pass UNIPHY clocks directly to GCC\")\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/26205\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 12:50:31 +0100", "id" : "a41558d2d3654938cbc22c1b3a9c6e68a1833089", "msg" : "qualcommax: list UNIPHY clock output names RX first", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/files/arch/arm64/boot/dts/qcom/ipq6018-ess.dtsi" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/include/dt-bindings/net/pcs/qualcomm,ipq-uniphy.yaml" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/arch/arm64/boot/dts/qcom/ipq8074-ess.dtsi" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "package/kernel/mac80211/patches/rtl/003-wifi-rtw88-usb-Download-the-reserved-page-synchronously.patch", "package/kernel/mac80211/Makefile" ], "commitId" : "a40fd94a3fa9bb787e631349032cf7878e3e6aad", "timestamp" : 1793018263000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/ansuelsmth", "fullName" : "ansuelsmth" }, "authorEmail" : "ansuelsmth@gmail.com", "comment" : "mac80211: realtek: download rsvd page synchronously on USB\u000a\u000aIn AP mode, rtw88 downloads the reserved page, which holds the beacon,\u000aand then polls BCN_VALID. Over USB the download only queues the URB, so\u000athe poll can run while the beacon is still on the bus and time out, in\u000abursts when the beacon changes often, such as when stations associate\u000aand disassociate (lwfinger/rtw88#451):\u000a\u000a rtw_8822bu: error beacon valid\u000a rtw_8822bu: failed to download drv rsvd page\u000a\u000aAdd v5 of the rtw88 patch. It waits for that URB in a new\u000artw_usb_write_port_sync() and kills it if it does not complete within\u000a5 s, as usb_bulk_msg() does, but keeps URB_ZERO_PACKET. This fixes the\u000ause-after-free that Ping-Ke Shih found in v3, where the completion\u000ahandler could run after the waiter had freed its context on timeout.\u000aOnly the line numbers and one context line differ from rtw-next.\u000a\u000av4 was tested with backports 7.2 on a Turris 1.x with a TP-Link Archer\u000aT4U v3 (RTL8812BU, handled by rtw_8822bu) as a 2.4 GHz AP over USB 2 and\u000aUSB 3 and as a 5 GHz AP over USB 3. The AP came up after each of 15\u000arestarts, and no \"error beacon valid\", \"failed to download drv rsvd\u000apage\" or USB write error was logged during them or 7 beacon changes. No\u000astation was associated, and the builds without the patch did not log\u000athese errors either, so this shows no regression; lwfinger/rtw88#451\u000aitself was not reproduced. v5 only reorders declarations, drops an\u000ainitializer and rewords a comment.\u000a\u000aPending upstream, submitted to rtw-next and acked by Ping-Ke Shih:\u000ahttps://lore.kernel.org/linux-wireless/20261013033737.6761-1-pepe.schlehofer@gmail.com/\u000a\u000aSigned-off-by: Josef Schlehofer \u000aLink: https://github.com/openwrt/openwrt/pull/25837\u000aSigned-off-by: Christian Marangi \u000a", "date" : "2026-10-26 13:37:43 +0100", "id" : "a40fd94a3fa9bb787e631349032cf7878e3e6aad", "msg" : "mac80211: realtek: download rsvd page synchronously on USB", "paths" : [ { "editType" : "edit", "file" : "package/kernel/mac80211/Makefile" }, { "editType" : "add", "file" : "package/kernel/mac80211/patches/rtl/003-wifi-rtw88-usb-Download-the-reserved-page-synchronously.patch" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "package/kernel/mac80211/patches/rtl/002-wifi-rtw88-8822b-Relax-ETSI-EDCCA-threshold-on-2.4-GHz.patch" ], "commitId" : "2462519a7f3811fe0535be440273e978cf66d080", "timestamp" : 1793018327000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/ansuelsmth", "fullName" : "ansuelsmth" }, "authorEmail" : "ansuelsmth@gmail.com", "comment" : "mac80211: realtek: relax 8822B 2.4 GHz ETSI EDCCA threshold\u000a\u000aUnder ETSI, rtw88 sets the RTL8822B EDCCA threshold to min(IGI, 36),\u000aat most about -74 dBm. On a Turris 1.x with a TP-Link Archer T4U v3\u000a(RTL8812BU) in USB 3 mode as a 2.4 GHz AP under CZ, in-band energy,\u000amostly from the adapter's own USB 3 link, kept IGI at 0x2a and the\u000athreshold at 36. The AP was not seen in any of 5 scans, and the\u000adriver reported \"failed to get tx report from firmware\".\u000a\u000aAdd the pending upstream patch that uses the normal threshold,\u000amax(IGI + 8, 48), on 2.4 GHz under ETSI, capped at 50 (nominally\u000a-60 dBm), and keeps the existing rule on 5 GHz and in JP. With it,\u000athe AP was seen in 4 of 5 scans on the same port and boot (16 of 19\u000aacross repeated runs), with no TX report failures. Only the line\u000anumbers are adapted to backports 7.2.\u000a\u000aPending upstream, submitted to rtw-next and acked by Ping-Ke Shih:\u000ahttps://lore.kernel.org/linux-wireless/20261011204509.46062-1-pepe.schlehofer@gmail.com/\u000a\u000aSigned-off-by: Josef Schlehofer \u000aLink: https://github.com/openwrt/openwrt/pull/25836\u000aSigned-off-by: Christian Marangi \u000a", "date" : "2026-10-26 13:38:47 +0100", "id" : "2462519a7f3811fe0535be440273e978cf66d080", "msg" : "mac80211: realtek: relax 8822B 2.4 GHz ETSI EDCCA threshold", "paths" : [ { "editType" : "add", "file" : "package/kernel/mac80211/patches/rtl/002-wifi-rtw88-8822b-Relax-ETSI-EDCCA-threshold-on-2.4-GHz.patch" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/ramips/dts/mt7621_mqmaker_witi.dts" ], "commitId" : "b5ff052caf4b7e90eb4b84941dec77bb5e4b5e08", "timestamp" : 1793022258000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "ramips: mt7621: fix SATA detection on MQMaker WiTi\u000a\u000aThe WiTi's SATA controller worked under OpenWrt 19.07, but\u000astopped being detected in the 21.02 release series following\u000achanges to the MT7621 PCIe driver. PCIe port 2 fails to\u000ainitialise, leaving the onboard ASMedia ASM1062 inaccessible.\u000a\u000aThis adds the missing board-specific reset GPIOs (19, 8 and 7).\u000aThe mapping was originally suggested by Sergio Paracuellos\u000aand tested by Florian Hofmann in the earlier issue discussion.\u000a\u000aI've tested the fix on a WiTi v2.0 (256 MB) with both OpenWrt main\u000a(Linux 6.18.55) and the 25.12 branch (Linux 6.12.112).\u000aBoth SATA ports negotiated 6 Gbps links, and I was able to\u000aread from two connected SSDs without errors.\u000a\u000aHope this helps anyone who is still using one of these boards\u000afor anything.\u000a\u000aLink: https://github.com/openwrt/openwrt/issues/9156\u000aLink: https://github.com/openwrt/openwrt/issues/9869\u000aSigned-off-by: Alex Sunekants \u000aLink: https://github.com/openwrt/openwrt/pull/26196\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 14:44:18 +0100", "id" : "b5ff052caf4b7e90eb4b84941dec77bb5e4b5e08", "msg" : "ramips: mt7621: fix SATA detection on MQMaker WiTi", "paths" : [ { "editType" : "edit", "file" : "target/linux/ramips/dts/mt7621_mqmaker_witi.dts" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "package/firmware/intel-microcode/Makefile" ], "commitId" : "6079d00acde51dc5a01cfd282d6647667e031f9e", "timestamp" : 1793022729000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "intel-microcode: switch source to Intel upstream, update to 20260925\u000a\u000aDebian tarball lags/currently 5 commits and 9 months old\u000arelative to upstream. Worse is that upstream is asleep at\u000athe switch with 3 outstanding bug reports since Feb/2026\u000ashowing them this fact[1-3].\u000a\u000aTherefore use intel/Intel-Linux-Processor-Microcode-Data-Files\u000adirectly and drop the Debian-supplied make step.\u000a\u000a1. https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1149038\u000a2. https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1136460\u000a3. https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1144154\u000a\u000aSigned-off-by: John Audia \u000aLink: https://github.com/openwrt/openwrt/pull/26184\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 14:52:09 +0100", "id" : "6079d00acde51dc5a01cfd282d6647667e031f9e", "msg" : "intel-microcode: switch source to Intel upstream, update to 20260925", "paths" : [ { "editType" : "edit", "file" : "package/firmware/intel-microcode/Makefile" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/kirkwood/base-files/etc/init.d/hwmon_fancontrol" ], "commitId" : "3c3a4150694ca9ea4b8a44c12273bd4780bbf7d0", "timestamp" : 1793024040000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "kirkwood: fix ReadyNAS Duo v2 fan controller detection\u000a\u000aThe hwmon index is not stable across boots. On the\u000aReadyNAS Duo v2, hwmon0 is the kirkwood_thermal sensor\u000awhile the g762 fan controller is registered as hwmon1.\u000a\u000aFind the fan controller by its hwmon name instead of\u000aassuming a fixed index.\u000a\u000aSigned-off-by: Jakub Skoczen \u000aLink: https://github.com/openwrt/openwrt/pull/26206\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 15:14:00 +0100", "id" : "3c3a4150694ca9ea4b8a44c12273bd4780bbf7d0", "msg" : "kirkwood: fix ReadyNAS Duo v2 fan controller detection", "paths" : [ { "editType" : "edit", "file" : "target/linux/kirkwood/base-files/etc/init.d/hwmon_fancontrol" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/realtek/files-6.18/drivers/net/dsa/rtl83xx/rtl-otto.h" ], "commitId" : "0074cd9b1d617bdb77a38f70d16653f5e31902ab", "timestamp" : 1793043963000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/markus.stockhausen", "fullName" : "markus.stockhausen" }, "authorEmail" : "markus.stockhausen@gmx.de", "comment" : "realtek: pie: preserve two-bit L2 frame type fields\u000a\u000aThe PIE L2 frame type and its mask occupy two bits in hardware, but their\u000apie_rule members are bool. Assigning LLC_Other (2) or a full mask (3)\u000atherefore truncates the value to 1 before the encoders can pack it.\u000a\u000aUse u8 for frame_type_l2 and frame_type_l2_m so all four switch families\u000apreserve the complete match and mask values during encoding and readback.\u000a\u000aSigned-off-by: João Silva \u000aLink: https://github.com/openwrt/openwrt/pull/26225\u000aSigned-off-by: Markus Stockhausen \u000a", "date" : "2026-10-26 20:46:03 +0100", "id" : "0074cd9b1d617bdb77a38f70d16653f5e31902ab", "msg" : "realtek: pie: preserve two-bit L2 frame type fields", "paths" : [ { "editType" : "edit", "file" : "target/linux/realtek/files-6.18/drivers/net/dsa/rtl83xx/rtl-otto.h" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/realtek/files-6.18/drivers/net/dsa/rtl83xx/pie.c" ], "commitId" : "6ca8a7f1d5b7051e365ade017d4fe481a5a622a7", "timestamp" : 1793044139000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/markus.stockhausen", "fullName" : "markus.stockhausen" }, "authorEmail" : "markus.stockhausen@gmx.de", "comment" : "realtek: pie: rtl930x: fix template 4 source port fields\u000a\u000aThe software description of RTL930x fixed template 4 lists SPM1 twice,\u000abut the hardware template contains SPM0 followed by SPM1. This encodes\u000athe upper source-port bitmap word into both fields, losing the lower\u000a16 source-port match bits and their mask.\u000a\u000aDescribe the two fields as SPM0, SPM1 so IPv6 source-address rules using\u000athis template match the intended ingress ports.\u000a\u000aSigned-off-by: João Silva \u000aLink: https://github.com/openwrt/openwrt/pull/26226\u000aSigned-off-by: Markus Stockhausen \u000a", "date" : "2026-10-26 20:48:59 +0100", "id" : "6ca8a7f1d5b7051e365ade017d4fe481a5a622a7", "msg" : "realtek: pie: rtl930x: fix template 4 source port fields", "paths" : [ { "editType" : "edit", "file" : "target/linux/realtek/files-6.18/drivers/net/dsa/rtl83xx/pie.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/generic/config-6.18" ], "commitId" : "5df2651e6a7002eec305fd1bef0c7f06050a949c", "timestamp" : 1793045293000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "generic: 6.18: add CONFIG_DETECT_HUNG_TASK_BLOCKER\u000a\u000aLinux 6.15 added DETECT_HUNG_TASK_BLOCKER, which depends on\u000aDETECT_HUNG_TASK. The generic config leaves DETECT_HUNG_TASK unset, so\u000athe symbol stays hidden by default, but selecting\u000aKERNEL_DETECT_HUNG_TASK makes it visible and the kernel build stops at\u000athe syncconfig prompt for it. Add it to the generic config.\u000a\u000aAssisted-by: Claude:claude-opus-5-5\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/26012\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 21:08:13 +0100", "id" : "5df2651e6a7002eec305fd1bef0c7f06050a949c", "msg" : "generic: 6.18: add CONFIG_DETECT_HUNG_TASK_BLOCKER", "paths" : [ { "editType" : "edit", "file" : "target/linux/generic/config-6.18" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/mediatek/filogic/config-6.18", "target/linux/mediatek/mt7622/config-6.18" ], "commitId" : "3d0cafad1cc516d3ed1e5db058a2686d23c5a8b2", "timestamp" : 1793045430000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "mediatek: disable perf events by default\u000a\u000aThe perf utility is not shipped by default. To use perf on a custom\u000abuild, enable KERNEL_PERF_EVENTS.\u000a\u000aSigned-off-by: Qingfang Deng \u000aLink: https://github.com/openwrt/openwrt/pull/26200\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 21:10:30 +0100", "id" : "3d0cafad1cc516d3ed1e5db058a2686d23c5a8b2", "msg" : "mediatek: disable perf events by default", "paths" : [ { "editType" : "edit", "file" : "target/linux/mediatek/mt7622/config-6.18" }, { "editType" : "edit", "file" : "target/linux/mediatek/filogic/config-6.18" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/mediatek/filogic/config-6.18", "target/linux/mediatek/mt7622/config-6.18" ], "commitId" : "4d3ed016a2d710ac5719863e7870fa91e39fc27e", "timestamp" : 1793045430000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "mediatek: disable unrelated errata\u000a\u000aMT7622 and MT798x (except MT7988) have a Cortex-A53 r0p4, as indicated\u000aby /proc/cpuinfo. Disable errata that only apply to earlier revisions.\u000a\u000aSigned-off-by: Qingfang Deng \u000aLink: https://github.com/openwrt/openwrt/pull/26200\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 21:10:30 +0100", "id" : "4d3ed016a2d710ac5719863e7870fa91e39fc27e", "msg" : "mediatek: disable unrelated errata", "paths" : [ { "editType" : "edit", "file" : "target/linux/mediatek/filogic/config-6.18" }, { "editType" : "edit", "file" : "target/linux/mediatek/mt7622/config-6.18" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/mediatek/filogic/config-6.18", "target/linux/mediatek/mt7622/config-6.18" ], "commitId" : "7d0bf92beefbb62e306f2318dfe5b47538bd85a6", "timestamp" : 1793045430000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "mediatek: disable unsupported arch features\u000a\u000aBoth A53 and A73 are baseline-ARMv8.0 and do not support those features.\u000a\u000aSigned-off-by: Qingfang Deng \u000aLink: https://github.com/openwrt/openwrt/pull/26200\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-26 21:10:30 +0100", "id" : "7d0bf92beefbb62e306f2318dfe5b47538bd85a6", "msg" : "mediatek: disable unsupported arch features", "paths" : [ { "editType" : "edit", "file" : "target/linux/mediatek/filogic/config-6.18" }, { "editType" : "edit", "file" : "target/linux/mediatek/mt7622/config-6.18" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "scripts/mkits-qsdk-ipq-image.sh" ], "commitId" : "9ec7903e371233ef913f1dee4e2aeb8872efff85", "timestamp" : 1793081966000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/daniel", "fullName" : "daniel" }, "authorEmail" : "daniel@makrotopia.org", "comment" : "scripts: mkits-qsdk-ipq-image: strip path from description\u000a\u000aThe FIT description of the boot script image is derived from the\u000afile name given on the command line. Take only the basename so that\u000athe script can be passed by path while the description stays the\u000asame.\u000a\u000aSigned-off-by: Daniel Golle \u000a", "date" : "2026-10-27 06:19:26 +0000", "id" : "9ec7903e371233ef913f1dee4e2aeb8872efff85", "msg" : "scripts: mkits-qsdk-ipq-image: strip path from description", "paths" : [ { "editType" : "edit", "file" : "scripts/mkits-qsdk-ipq-image.sh" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/image/ipq807x.mk" ], "commitId" : "454e516148eed22217dc878d486e28dd5cac5890", "timestamp" : 1793081967000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/daniel", "fullName" : "daniel" }, "authorEmail" : "daniel@makrotopia.org", "comment" : "qualcommax: ipq807x: fix Netgear factory image race\u000a\u000aRBR750, RBS750, RBR850 and RBS850 share one bootscript, and the\u000afactory image recipe copies it into the common kernel tmp directory\u000awith cp -f. When several images are built in parallel, one cp can\u000acreate the destination between the unlink and the create of another\u000aand fail with \"File exists\", aborting the target install.\u000a\u000aThe copy only serves to let dtc find the script next to the .its\u000afile. Pass the script by its absolute path instead, so that no shared\u000afile is written at all.\u000a\u000aFixes: d0bd7ad1a2 (\"qualcommax: ipq807x: add Netgear RBx750 support\")\u000aSigned-off-by: Daniel Golle \u000a", "date" : "2026-10-27 06:19:27 +0000", "id" : "454e516148eed22217dc878d486e28dd5cac5890", "msg" : "qualcommax: ipq807x: fix Netgear factory image race", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/image/ipq807x.mk" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/image/ipq60xx.mk" ], "commitId" : "936801c4489b02ad39eadd1a9b26d38dd3a40e0f", "timestamp" : 1793081967000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/daniel", "fullName" : "daniel" }, "authorEmail" : "daniel@makrotopia.org", "comment" : "qualcommax: ipq60xx: fix Netgear factory image race\u000a\u000aRBR350 and RBS350 share one bootscript, and the factory image recipe\u000acopies it into the common kernel tmp directory with cp -f. When both\u000aimages are built in parallel, one cp can create the destination\u000abetween the unlink and the create of the other and fail with \"File\u000aexists\", aborting the target install.\u000a\u000aThe copy only serves to let dtc find the script next to the .its\u000afile. Pass the script by its absolute path instead, so that no shared\u000afile is written at all.\u000a\u000aFixes: 67e6baf05e (\"qualcommax: ipq60xx: add Netgear RBx350 support\")\u000aSigned-off-by: Daniel Golle \u000a", "date" : "2026-10-27 06:19:27 +0000", "id" : "936801c4489b02ad39eadd1a9b26d38dd3a40e0f", "msg" : "qualcommax: ipq60xx: fix Netgear factory image race", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/image/ipq60xx.mk" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/realtek/patches-6.18/705-v7.1-net-sfp-initialize-i2c_block_size-at-adapter-configu.patch", "target/linux/realtek/patches-6.18/706-01-v7.2-net-sfp-apply-I2C-adapter-quirks-to-limit-block-size.patch", "target/linux/realtek/patches-6.18/706-02-v7.2-net-sfp-extend-SMBus-support.patch", "target/linux/realtek/patches-6.18/714-net-phy-sfp-add-support-for-SMBus.patch", "target/linux/generic/pending-6.18/753-net-sfp-add-quirk-for-ZYXEL-PMG3000-D20B.patch" ], "commitId" : "8ef3195f377420401969d31ecfd87c5e4179d166", "timestamp" : 1793091151000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "generic: 6.18: add SFP quirk for ZYXEL PMG3000-D20B\u000a\u000aThe ZYXEL PMG3000-D20B GPON SFP module reports LOS and TX_FAULT in a\u000away that prevents the SFP state machine from initialising it. Its\u000aEEPROM also advertises autonegotiation, which the module does not\u000asupport, and it needs a long startup delay before it comes up reliably.\u000a\u000aAdd a quirk which reuses the existing fixups to ignore LOS and\u000aTX_FAULT, drops the bogus autoneg capability and sets module_t_wait\u000ato T_START_UP_BAD_GPON (60s).\u000a\u000aSigned-off-by: Martin Ostertag \u000aLink: https://github.com/openwrt/openwrt/pull/26187\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-27 09:52:31 +0100", "id" : "8ef3195f377420401969d31ecfd87c5e4179d166", "msg" : "generic: 6.18: add SFP quirk for ZYXEL PMG3000-D20B", "paths" : [ { "editType" : "edit", "file" : "target/linux/realtek/patches-6.18/705-v7.1-net-sfp-initialize-i2c_block_size-at-adapter-configu.patch" }, { "editType" : "edit", "file" : "target/linux/realtek/patches-6.18/706-01-v7.2-net-sfp-apply-I2C-adapter-quirks-to-limit-block-size.patch" }, { "editType" : "edit", "file" : "target/linux/realtek/patches-6.18/706-02-v7.2-net-sfp-extend-SMBus-support.patch" }, { "editType" : "edit", "file" : "target/linux/realtek/patches-6.18/714-net-phy-sfp-add-support-for-SMBus.patch" }, { "editType" : "add", "file" : "target/linux/generic/pending-6.18/753-net-sfp-add-quirk-for-ZYXEL-PMG3000-D20B.patch" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/realtek/patches-6.18/824-net-dsa-pass-bridge-VLAN-protocol-changes-to-drivers.patch" ], "commitId" : "f3aa65fff90f2be10efedff798f8ba6bf6b90f0f", "timestamp" : 1793092466000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "realtek: refresh patches\u000a\u000aCI told us to refresh, so lets do it.\u000a\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-27 10:14:26 +0100", "id" : "f3aa65fff90f2be10efedff798f8ba6bf6b90f0f", "msg" : "realtek: refresh patches", "paths" : [ { "editType" : "edit", "file" : "target/linux/realtek/patches-6.18/824-net-dsa-pass-bridge-VLAN-protocol-changes-to-drivers.patch" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "package/firmware/ipq-wifi/Makefile" ], "commitId" : "f5a8129af3b8be8c9e629cefe603a7873ead21b8", "timestamp" : 1793093707000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "ipq-wifi: update to Git HEAD (2026-10-25)\u000a\u000aa4bb9873649c ipq9574/qcn9274: add Linksys LN6001 board files\u000a49ce14c099b1 ipq6018: add BDF for Wallys DR6018S\u000a\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-27 10:35:07 +0100", "id" : "f5a8129af3b8be8c9e629cefe603a7873ead21b8", "msg" : "ipq-wifi: update to Git HEAD (2026-10-25)", "paths" : [ { "editType" : "edit", "file" : "package/firmware/ipq-wifi/Makefile" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/ipq60xx/base-files/lib/upgrade/platform.sh", "target/linux/qualcommax/image/ipq60xx.mk", "target/linux/qualcommax/dts/ipq6010-dr6018s.dts", "target/linux/qualcommax/ipq60xx/base-files/etc/board.d/02_network", "package/firmware/ipq-wifi/Makefile" ], "commitId" : "01573cdc37e86b5e5308fffc81244b805bfb97f1", "timestamp" : 1793093783000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: ipq60xx: add support for Wallys DR6018S\u000a\u000aThe Wallys DR6018S is a dual-band Wi-Fi 6 board based on the\u000aQualcomm-Atheros IPQ6010 platform.\u000a\u000aProduct specification:\u000ahttps://www.wallystech.com/upload/DR6018S.pdf\u000a\u000aHardware:\u000a- Qualcomm-Atheros IPQ6010 quad-core ARM 64-bit A53 processor at 1.8 GHz\u000a- 1 GB DDR3L memory (2 x 512 MB)\u000a- 8 MB NOR flash and 256 MB NAND flash\u000a- Two 1 Gbps Ethernet ports with a QCA8075 PHY package\u000a- One Ethernet port supports passive or active PoE\u000a- Integrated 2x2 2.4 GHz 802.11ax radio, up to 573 Mbps\u000a- Integrated 2x2 5 GHz 802.11ax radio, up to 1201 Mbps\u000a- M.2 NGFF E-key socket with PCIe 3.0 support for an optional QCN9074\u000a Wi-Fi module\u000a- Optional SD card slot\u000a- Optional USB 3.0 port\u000a- Four-pin serial connector using GPIO44/GPIO45 at 115200 8N1\u000a- Four status LEDs and one reset button\u000a- Four MMCX antenna connectors\u000a- 24-48 V DC input with passive and active PoE support\u000a\u000aThis adds the device tree, FIT/UBI image definitions, network defaults,\u000aNVMEM-backed ART calibration and MAC address extraction, NAND sysupgrade\u000asupport and optional QCN9074 PCIe Wi-Fi support.\u000a\u000aThe SD card slot and USB 3.0 port are listed as board capabilities from\u000athe vendor specification. They are not required for the tested onboard\u000aWi-Fi configuration.\u000a\u000aMAC address layout:\u000a- WAN: MAC stored in 0:ART at offset 0x0000\u000a- LAN: MAC stored in 0:ART at offset 0x0006\u000a- 2.4 GHz and 5 GHz Wi-Fi: calibration and radio MACs stored in the\u000a calibration block at 0:ART offset 0x1000\u000a\u000aThe MAC addresses are unit-specific and are not hard-coded.\u000a\u000aBoard data is submitted separately through firmware_qca-wireless.\u000a\u000aInstallation:\u000a1. Boot the initramfs image through the vendor U-Boot:\u000a\u000a setenv serverip 192.168.1.4\u000a setenv ipaddr 192.168.1.1\u000a tftpboot 0x44000000 openwrt-qualcommax-ipq60xx-wallys_dr6018s-initramfs-uImage.itb\u000a bootm 0x44000000\u000a\u000a2. Transfer the sysupgrade image to /tmp and install it:\u000a\u000a sysupgrade -n /tmp/openwrt-qualcommax-ipq60xx-wallys_dr6018s-squashfs-sysupgrade.bin\u000a\u000aTested on a Wallys DR6018S:\u000a- Booting the initramfs image from the vendor U-Boot\u000a- Installing OpenWrt to SPI NAND\u000a- Rebooting and cold-booting from the installed system\u000a- 2.4 GHz and 5 GHz AP operation\u000a- QCN9074 PCIe module enumeration and Wi-Fi operation\u000a- Ethernet connectivity\u000a- Correct Ethernet and Wi-Fi MAC addresses\u000a\u000aSigned-off-by: Alan Gao <1226010136@qq.com>\u000aLink: https://github.com/openwrt/openwrt/pull/25955\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-27 10:36:23 +0100", "id" : "01573cdc37e86b5e5308fffc81244b805bfb97f1", "msg" : "qualcommax: ipq60xx: add support for Wallys DR6018S", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/image/ipq60xx.mk" }, { "editType" : "add", "file" : "target/linux/qualcommax/dts/ipq6010-dr6018s.dts" }, { "editType" : "edit", "file" : "target/linux/qualcommax/ipq60xx/base-files/lib/upgrade/platform.sh" }, { "editType" : "edit", "file" : "target/linux/qualcommax/ipq60xx/base-files/etc/board.d/02_network" }, { "editType" : "edit", "file" : "package/firmware/ipq-wifi/Makefile" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "package/network/config/wifi-scripts/files-ucode/usr/share/ucode/wifi/ap.uc", "package/network/config/wifi-scripts/files-ucode/usr/share/schema/wireless.wifi-iface.json" ], "commitId" : "c9ea07732f7561a346d5d95e525c53fe0b00cf14", "timestamp" : 1793094200000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/nbd", "fullName" : "nbd" }, "authorEmail" : "nbd@nbd.name", "comment" : "wifi-scripts: ignore sae_pwe 0 on an AP MLD with a 6 GHz link\u000a\u000ahostapd_config_check_bss_6g() turns sae_pwe 0 into 2 on 6 GHz only.\u000aWith an explicit sae_pwe 0 on an AP MLD, the 6 GHz link advertises SAE\u000aH2E in its RSNXE and the other links advertise no RSNXE. IEEE\u000a802.11be-2024 12.6.2 requires the same RSNXE on all links.\u000a\u000aFor sae and sae-mixed, use the band defaults in place of 0: 1 on 6 GHz,\u000awhere WPA3 v3.5 does not allow hunting-and-pecking, and 2 elsewhere.\u000aBoth values set the same RSNXE. psk2 and sae-compat keep the value,\u000asince their RSNE has no SAE AKM outside 6 GHz.\u000a\u000aSigned-off-by: Felix Fietkau \u000a", "date" : "2026-10-27 10:43:20 +0100", "id" : "c9ea07732f7561a346d5d95e525c53fe0b00cf14", "msg" : "wifi-scripts: ignore sae_pwe 0 on an AP MLD with a 6 GHz link", "paths" : [ { "editType" : "edit", "file" : "package/network/config/wifi-scripts/files-ucode/usr/share/schema/wireless.wifi-iface.json" }, { "editType" : "edit", "file" : "package/network/config/wifi-scripts/files-ucode/usr/share/ucode/wifi/ap.uc" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "package/network/config/wifi-scripts/files-ucode/usr/share/ucode/wifi/ap.uc", "package/network/config/wifi-scripts/files-ucode/usr/share/schema/wireless.wifi-iface.json" ], "commitId" : "035a94fdb5fe5cc23cb369d0ce220601d2727ea4", "timestamp" : 1793094200000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/nbd", "fullName" : "nbd" }, "authorEmail" : "nbd@nbd.name", "comment" : "wifi-scripts: default sae_pwe also with ppsk\u000a\u000aWith ppsk, iface_auth_type() leaves sae_pwe unset and hostapd uses 0.\u000aOn an AP MLD with a 6 GHz link, the 6 GHz link then advertises SAE H2E\u000ain its RSNXE and the other links advertise no RSNXE, as with an\u000aexplicit sae_pwe 0.\u000a\u000appsk skipped the defaults because sae_get_password() derived no PT for\u000aa per-station password, so H2E commits failed. Commit 8db58e80ab30\u000a(\"hostapd: derive the SAE PT of a per-station password\") added that, so\u000ause the defaults with ppsk too: 1 on 6 GHz and 2 on other bands.\u000a\u000aSigned-off-by: Felix Fietkau \u000a", "date" : "2026-10-27 10:43:20 +0100", "id" : "035a94fdb5fe5cc23cb369d0ce220601d2727ea4", "msg" : "wifi-scripts: default sae_pwe also with ppsk", "paths" : [ { "editType" : "edit", "file" : "package/network/config/wifi-scripts/files-ucode/usr/share/ucode/wifi/ap.uc" }, { "editType" : "edit", "file" : "package/network/config/wifi-scripts/files-ucode/usr/share/schema/wireless.wifi-iface.json" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "package/network/services/bridger/Makefile" ], "commitId" : "fbbcf2f78ba001471016f8b5ab0c265b532b22ea", "timestamp" : 1793094221000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/nbd", "fullName" : "nbd" }, "authorEmail" : "nbd@nbd.name", "comment" : "bridger: update to Git HEAD (2026-09-18)\u000a\u000a5084b9c68b0d flow: free flow struct on delete and skip idle path after free\u000ab2b73e247c3b device: prevent member_list double-add on repeated update\u000a4b008863ec30 device: clean up fdb entries on master change and device free\u000ac38234ea982b nl: guard bridger_nl_udebug_cb behind NL_UDEBUG\u000a7328c75e0683 nl: fix bridge vlan state notification handling\u000a335bb7a93793 nl: dump vlan state after link info during resync\u000a3855a8d2fe55 nl: fix vlan table leak and state loss on rebuild\u000a72293c5adf2e nl: clear offloaded flows when a vlan stops forwarding\u000a8a282a294061 nl: keep port state across link messages without protinfo\u000a385d82f1024b device: clean up references to freed devices and bridges\u000aeffe9fb0f351 flow: fix offload ifindex truncation\u000aa429ae138b17 nl: dump all offload filters when updating flow stats\u000a056524b9f39c flow: evict the least active flow when the offload map is full\u000ae8d03f033c5d ubus: apply bridge_local_tx regardless of blacklist contents\u000a04b40c1234c4 flow: skip offload for blacklisted or detached devices\u000a582e26c56fc8 nl: fix message leak in bridger_refresh_linkinfo\u000abb2b1f40ba4f nl: remove stale devices and fdb entries during resync\u000acefdf80bcf81 flow: reset the fdb refresh flag on all entries\u000ae0dbb73fa96c bridger-bpf: count packets for flows without a target port\u000a43c44960b8ea bridger-bpf: fix off-by-one in skb_ptr bounds check\u000a871ee2e5fd10 bridger-bpf: preserve vlan priority when retagging flows\u000a6b5071cf63a2 device: drop remaining kernel and config state on free\u000a240ce3e86f1b flow: remove the offload filter when a re-added flow stops qualifying\u000a6ef1f9d9c6dd nl: process fdb removal for aged out entries\u000ac355096362a2 nl: fix handling of bridge vlan change notifications\u000a1bb2be83ea48 main: fix link failure with ubus support disabled\u000a8b29b5d03c4b flow: fix pending flow handling for existing flows\u000a0bcfb0043645 flow: only evict flows when the offload map is full\u000a7a0d64fcab83 nl: report netlink errors back to the caller\u000a2a2d1261d12e nl: offload on the outermost device of the chain first\u000a\u000aSigned-off-by: Felix Fietkau \u000a", "date" : "2026-10-27 10:43:41 +0100", "id" : "fbbcf2f78ba001471016f8b5ab0c265b532b22ea", "msg" : "bridger: update to Git HEAD (2026-09-18)", "paths" : [ { "editType" : "edit", "file" : "package/network/services/bridger/Makefile" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "package/network/utils/ethtool/Makefile" ], "commitId" : "eeea495767bd378f65093294f7ddb492b3cab378", "timestamp" : 1793094272000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "ethtool: update to 7.2\u000a\u000aVersion 7.2 - Oct 19, 2026\u000a\u0009* Feature: support for KSZ87xx PHY tunables (--{get,set}-phy-tunable)\u000a\u0009* Fix: cmis: fix printing of Tx bias current (-m)\u000a\u0009* Fix: cmis: fix printing of channel-level flags (-m)\u000a\u0009* Fix: cmis: fix printing of CDB EPL pages (-m)\u000a\u0009* Fix: printing of transmitter technology for CMIS transceivers (-m)\u000a\u0009* Fix: avoid undefined behavior with CMIS transceivers (-m)\u000a\u0009* Fix: add missing CMIS transmitter technologies (-m)\u000a\u0009* Fix: printing of attenuation and wavelength (-m)\u000a\u0009* Fix: printing of link length (-m)\u000a\u0009* Fix: sfpid: print all implemented options (-m)\u000a\u0009* Fix: sfpid: print all compliance codes (-m)\u000a\u0009* Fix: qsfp: print all compliance codes (-m)\u000a\u0009* Fix: module-eeprom: clamp data length to attribute payload size (-m)\u000a\u0009* Fix: validate module EEPROM offset and length (-m)\u000a\u0009* Fix: rss: bound hash function masks in text and JSON output (-x)\u000a\u0009* Fix: fbnic: fix off-by-one shift in reg dump parser (-d)\u000a\u0009* Fix: fbnic: Fix various TX and RX CSR fields decoding (-d)\u000a\u0009* Fix: fbnic: Fix register dump names and bit labels (-d)\u000a\u000aSigned-off-by: Josef Schlehofer \u000aLink: https://github.com/openwrt/openwrt/pull/26216\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-27 10:44:32 +0100", "id" : "eeea495767bd378f65093294f7ddb492b3cab378", "msg" : "ethtool: update to 7.2", "paths" : [ { "editType" : "edit", "file" : "package/network/utils/ethtool/Makefile" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "package/network/utils/iw/Makefile" ], "commitId" : "0104d9e500aebd0344c0a6a24801d80dde1ce441", "timestamp" : 1793094420000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "iw: fix the documentation URL\u000a\u000aThe old wireless.kernel.org URL is redirected to a malformed address on\u000awireless.docs.kernel.org, which returns 404. Point it to the current\u000adocumentation page.\u000a\u000aSigned-off-by: Josef Schlehofer \u000aLink: https://github.com/openwrt/openwrt/pull/26218\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-27 10:47:00 +0100", "id" : "0104d9e500aebd0344c0a6a24801d80dde1ce441", "msg" : "iw: fix the documentation URL", "paths" : [ { "editType" : "edit", "file" : "package/network/utils/iw/Makefile" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "package/network/utils/iproute2/patches/200-drop_libbsd_dependency.patch", "package/network/utils/iproute2/patches/150-keep_libcap_optional.patch", "package/network/utils/iproute2/patches/001-1-tc-cake-add-cake_mq-support.patch", "package/network/utils/iproute2/patches/301-fix-UINT_MAX-undeclared-with-build-with-musl-libc.patch", "package/network/utils/iproute2/patches/140-keep_libmnl_optional.patch", "package/network/utils/iproute2/patches/130-no_netem_tipc_man_vdpa.patch", "package/network/utils/iproute2/Makefile", "package/network/utils/iproute2/patches/001-0-Update-kernel-headers.patch" ], "commitId" : "6b7ea6896a3a83c9ae80d31e8d35d911e032268f", "timestamp" : 1793094610000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "iproute2: update to 7.3.0\u000a\u000aDrop the patches that are included upstream: the kernel headers update,\u000athe cake_mq support and the musl UINT_MAX fix. Do not build the new\u000adpll tool, like the other tools that are not packaged, and refresh the\u000aremaining patches.\u000a\u000aSigned-off-by: Josef Schlehofer \u000aLink: https://github.com/openwrt/openwrt/pull/26217\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-27 10:50:10 +0100", "id" : "6b7ea6896a3a83c9ae80d31e8d35d911e032268f", "msg" : "iproute2: update to 7.3.0", "paths" : [ { "editType" : "edit", "file" : "package/network/utils/iproute2/Makefile" }, { "editType" : "edit", "file" : "package/network/utils/iproute2/patches/130-no_netem_tipc_man_vdpa.patch" }, { "editType" : "edit", "file" : "package/network/utils/iproute2/patches/150-keep_libcap_optional.patch" }, { "editType" : "edit", "file" : "package/network/utils/iproute2/patches/200-drop_libbsd_dependency.patch" }, { "editType" : "delete", "file" : "package/network/utils/iproute2/patches/001-0-Update-kernel-headers.patch" }, { "editType" : "delete", "file" : "package/network/utils/iproute2/patches/001-1-tc-cake-add-cake_mq-support.patch" }, { "editType" : "delete", "file" : "package/network/utils/iproute2/patches/301-fix-UINT_MAX-undeclared-with-build-with-musl-libc.patch" }, { "editType" : "edit", "file" : "package/network/utils/iproute2/patches/140-keep_libmnl_optional.patch" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/realtek/files-6.18/drivers/net/dsa/rtl83xx/common.c", "target/linux/realtek/files-6.18/drivers/net/dsa/rtl83xx/rtl-otto.h" ], "commitId" : "86f43bcb32558f0f33cac03ae9c5b7851ba79334", "timestamp" : 1793106513000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/markus.stockhausen", "fullName" : "markus.stockhausen" }, "authorEmail" : "markus.stockhausen@gmx.de", "comment" : "realtek: dsa: unregister the switch during remove and probe unwind\u000a\u000aThe Realtek platform driver calls dsa_switch_shutdown() on removal and\u000awhen L3 setup fails after DSA registration. Shutdown only disconnects\u000aports for system shutdown; it does not undo dsa_register_switch(). The\u000alink-state interrupt requested with request_irq() is also left installed.\u000a\u000aTrack successful IRQ registration and release the interrupt before\u000aunregistering the switch, while its port devices still exist. Cancel\u000acounter polling on the L3 probe failure path as well as on removal, then\u000ause dsa_unregister_switch() before releasing TC state and the workqueue.\u000a\u000aSigned-off-by: João Silva \u000aLink: https://github.com/openwrt/openwrt/pull/26228\u000aSigned-off-by: Markus Stockhausen \u000a", "date" : "2026-10-27 14:08:33 +0100", "id" : "86f43bcb32558f0f33cac03ae9c5b7851ba79334", "msg" : "realtek: dsa: unregister the switch during remove and probe unwind", "paths" : [ { "editType" : "edit", "file" : "target/linux/realtek/files-6.18/drivers/net/dsa/rtl83xx/rtl-otto.h" }, { "editType" : "edit", "file" : "target/linux/realtek/files-6.18/drivers/net/dsa/rtl83xx/common.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "include/kernel-build.mk", "include/image.mk", "include/subdir.mk", "target/linux/Makefile" ], "commitId" : "7cd9ef1f2223541943c1ebb9808a7a637b173864", "timestamp" : 1793119521000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "build: add optional DTB schema check target\u000a\u000aAdd target/linux/dtbs_check for manually checking selected OpenWrt\u000aDTBs against the patched kernel binding schemas. It runs only when\u000aexplicitly requested; normal builds do not invoke it.\u000a\u000aDT_SCHEMA_FILES can limit validation to specific bindings.\u000a\u000aLink: https://github.com/openwrt/openwrt/pull/26240\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-27 17:45:21 +0100", "id" : "7cd9ef1f2223541943c1ebb9808a7a637b173864", "msg" : "build: add optional DTB schema check target", "paths" : [ { "editType" : "edit", "file" : "include/subdir.mk" }, { "editType" : "edit", "file" : "include/kernel-build.mk" }, { "editType" : "edit", "file" : "include/image.mk" }, { "editType" : "edit", "file" : "target/linux/Makefile" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/realtek/rtl960x/target.mk" ], "commitId" : "dc9ce90354a2733489663c9479084477070f747d", "timestamp" : 1793169716000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/markus.stockhausen", "fullName" : "markus.stockhausen" }, "authorEmail" : "markus.stockhausen@gmx.de", "comment" : "realtek: rtl960x: change to router device type\u000a\u000aThe SoCs in rtl960x subtarget are of router variety so override the\u000adevice type from basic to router.\u000a\u000aSigned-off-by: Rustam Adilov \u000aLink: https://github.com/openwrt/openwrt/pull/26137\u000aSigned-off-by: Markus Stockhausen \u000a", "date" : "2026-10-28 07:41:56 +0100", "id" : "dc9ce90354a2733489663c9479084477070f747d", "msg" : "realtek: rtl960x: change to router device type", "paths" : [ { "editType" : "edit", "file" : "target/linux/realtek/rtl960x/target.mk" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/realtek/dts/rtl9607c-vb.dtsi", "target/linux/realtek/dts/rtl9607c-va.dtsi", "target/linux/realtek/dts/rtl9607_bt-pon_bt-g711ax.dts" ], "commitId" : "92b3ad7722ea7c2fe16173b3543eec0f50d2df07", "timestamp" : 1793169716000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/markus.stockhausen", "fullName" : "markus.stockhausen" }, "authorEmail" : "markus.stockhausen@gmx.de", "comment" : "realtek: add rtl9607c-vb and rtl9607c-va dtsi files\u000a\u000aSome devices have the an older revision of RTL9607C/RTL8198D SoC which would\u000arequire board file to override some compatibles from rtl960x dtsi file, like\u000apcie phy and potentially more in the future.\u000a\u000aAdd the separate rtl9607c-vb and rtl9607c-va dtsi files to handle it instead.\u000a\u000aChange the bt-g711ax dts file to now include the new rtl9607c-vb dtsi file.\u000a\u000aSigned-off-by: Rustam Adilov \u000aLink: https://github.com/openwrt/openwrt/pull/26137\u000aSigned-off-by: Markus Stockhausen \u000a", "date" : "2026-10-28 07:41:56 +0100", "id" : "92b3ad7722ea7c2fe16173b3543eec0f50d2df07", "msg" : "realtek: add rtl9607c-vb and rtl9607c-va dtsi files", "paths" : [ { "editType" : "add", "file" : "target/linux/realtek/dts/rtl9607c-vb.dtsi" }, { "editType" : "add", "file" : "target/linux/realtek/dts/rtl9607c-va.dtsi" }, { "editType" : "edit", "file" : "target/linux/realtek/dts/rtl9607_bt-pon_bt-g711ax.dts" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/realtek/files-6.18/drivers/net/dsa/rtl83xx/pie.c" ], "commitId" : "3947b4d959f93478a1e792afae76d55d84018945", "timestamp" : 1793172942000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/markus.stockhausen", "fullName" : "markus.stockhausen" }, "authorEmail" : "markus.stockhausen@gmx.de", "comment" : "realtek: pie: rtl930x: validate rules and propagate programming errors\u000a\u000aThe RTL930x PIE writer indexes the template array using unchecked rule\u000aand template IDs and reports success even when the table write fails.\u000aRule addition also ignores template-register read errors and the writer's\u000areturn value. TC can therefore report an offloaded rule that was not\u000ainstalled.\u000a\u000aValidate rule indices and template selectors before accessing registers,\u000abitmaps or template descriptions. Propagate register, table-acquisition\u000aand table-write errors, including writes that clear a rule.\u000a\u000aOnly enable block lookup after successfully writing the rule and return\u000alookup programming errors to the caller. Mark failed software rules\u000ainvalid but keep their IDs reserved: a failed hardware operation may\u000astill have committed and reusing the ID could overwrite an active rule.\u000a\u000aSigned-off-by: João Silva \u000aLink: https://github.com/openwrt/openwrt/pull/26227\u000aSigned-off-by: Markus Stockhausen \u000a", "date" : "2026-10-28 08:35:42 +0100", "id" : "3947b4d959f93478a1e792afae76d55d84018945", "msg" : "realtek: pie: rtl930x: validate rules and propagate programming errors", "paths" : [ { "editType" : "edit", "file" : "target/linux/realtek/files-6.18/drivers/net/dsa/rtl83xx/pie.c" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/realtek/files-6.18/drivers/net/dsa/rtl83xx/rtl-otto.h", "target/linux/realtek/files-6.18/drivers/net/dsa/rtl83xx/pie.c" ], "commitId" : "1943e280ca7f1426061cb06b3eb5cd6557d3974c", "timestamp" : 1793172942000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/markus.stockhausen", "fullName" : "markus.stockhausen" }, "authorEmail" : "markus.stockhausen@gmx.de", "comment" : "realtek: pie: rtl930x: roll back failed TC rule additions\u000a\u000aA failed table write may already have installed an active hardware rule.\u000aTC discards the failed flow without calling rule removal, so clearing\u000athe software valid flag and reserving the ID leaves that rule orphaned.\u000a\u000aInvalidate the hardware row before returning the original programming\u000aerror. Allow invalidation without reading the template selector, keeping\u000arollback independent of selector-read failures.\u000a\u000aIf rollback fails, disable and quarantine the entire PIE block until\u000areprobe. Prevent subsequent additions from re-enabling it, including\u000aafter existing rules release their IDs. Report when block disabling\u000acannot be verified and a reset is required.\u000a\u000aValidated with partial and delayed write failures, TC failure cleanup,\u000asanitizer checks, MIPS compilation and checkpatch.\u000a\u000aSigned-off-by: João Silva \u000aLink: https://github.com/openwrt/openwrt/pull/26227\u000aSigned-off-by: Markus Stockhausen \u000a", "date" : "2026-10-28 08:35:42 +0100", "id" : "1943e280ca7f1426061cb06b3eb5cd6557d3974c", "msg" : "realtek: pie: rtl930x: roll back failed TC rule additions", "paths" : [ { "editType" : "edit", "file" : "target/linux/realtek/files-6.18/drivers/net/dsa/rtl83xx/pie.c" }, { "editType" : "edit", "file" : "target/linux/realtek/files-6.18/drivers/net/dsa/rtl83xx/rtl-otto.h" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "package/network/config/netifd/files/usr/libexec/network/packet-steering.uc", "package/network/config/netifd/Makefile" ], "commitId" : "e1dd9894042a7250d96d7fae1f6d6a98060f2d78", "timestamp" : 1793178094000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "netifd: leave receive steering off on hardware RSS devices\u000a\u000aA device with at least as many receive queues as CPUs and no NAPI\u000athreads is steered like a single queue device: the default mode points\u000athe RPS mask of every queue at one CPU. qca_edma on qualcommax spreads\u000aits flows over four rings by RSS and pins the interrupts of each ring to\u000aits own CPU, so this funnels everything the hardware spread back onto\u000aCPU0.\u000a\u000aOn a Xiaomi AX3600 with flow offload off, a routed 8-flow TCP download\u000areached 737-790 Mbit/s with CPU0 at 100% softirq and the other CPUs\u000abelow 10% (6 runs). With RPS off the same download reached\u000a933-941 Mbit/s (3 runs).\u000a\u000aLeave RPS off on such a device in the default mode when its receive\u000ainterrupts, the MSI vectors of a PCI device or the GIC interrupts of its\u000adevice tree node, are pinned to every CPU between them. Interrupts named\u000afor transmit completions only are not counted. A device whose receive\u000ainterrupts land on fewer CPUs is steered as before, as are mode 2 and\u000adevices with NAPI threads.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/26222\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-28 10:01:34 +0100", "id" : "e1dd9894042a7250d96d7fae1f6d6a98060f2d78", "msg" : "netifd: leave receive steering off on hardware RSS devices", "paths" : [ { "editType" : "edit", "file" : "package/network/config/netifd/files/usr/libexec/network/packet-steering.uc" }, { "editType" : "edit", "file" : "package/network/config/netifd/Makefile" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "package/network/config/netifd/files/usr/libexec/network/packet-steering.uc" ], "commitId" : "cd6b08f2b5b99c8ac76adc58b77f75fad3bd9d78", "timestamp" : 1793178094000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "netifd: leave receive steering off on RSS Wi-Fi devices\u000a\u000amac80211 gives a Wi-Fi netdev one receive queue, also when the driver\u000aspreads received frames over several rings by RSS and says so with\u000aIEEE80211_HW_USES_RSS. ath11k hashes received frames over four REO\u000arings, and on ipq807x the interrupts of those rings are pinned to\u000aseparate CPUs, but the default steering mode points the RPS mask of the\u000aAP netdev at one CPU, moving the stack work of every received frame back\u000aonto it.\u000a\u000aOn a Xiaomi AX3600, a 4-stream TCP upload from a Wi-Fi client to a\u000awired host ran CPU0 at 80-91% with RPS on CPU0 and the other CPUs below\u000a21%. With RPS off no CPU went above 44%, at the same total CPU time and\u000a908-938 Mbit/s against 858-937 Mbit/s (3 runs each).\u000a\u000aRead the flag from mac80211's hwflags in debugfs and treat such a\u000adevice like a multi-queue RSS device: RPS stays off when its receive\u000ainterrupts are spread over every CPU.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/26222\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-28 10:01:34 +0100", "id" : "cd6b08f2b5b99c8ac76adc58b77f75fad3bd9d78", "msg" : "netifd: leave receive steering off on RSS Wi-Fi devices", "paths" : [ { "editType" : "edit", "file" : "package/network/config/netifd/files/usr/libexec/network/packet-steering.uc" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "package/network/config/netifd/files/usr/libexec/network/packet-steering.uc" ], "commitId" : "8a97b52c3a90266769c63c8efbbf1666f249e523", "timestamp" : 1793178094000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "netifd: steer receive work away from the interrupt CPUs\u000a\u000aThe default mode picks the RPS CPU of a softirq NAPI device by load\u000aalone, without knowing where its interrupts are taken. On a system with\u000aone such device the pick is CPU0, which is where its interrupts usually\u000aland, so RPS hands the stack work to the CPU that already polls the\u000adevice. A multi-queue device with every interrupt on one CPU (igb\u000awithout spread vectors, bcmgenet, ath11k over PCI on qualcommax) and a\u000asingle queue device on CPU0 run all of their receive work there.\u000a\u000aOn a Xiaomi AX3600 with every qca_edma interrupt moved to CPU0 and flow\u000aoffload off, a routed 8-flow TCP download reached 625-649 Mbit/s with\u000aRPS on CPU0 and CPU0 at 100%. RPS on one other CPU reached\u000a922-940 Mbit/s with that CPU at 97-99%. RPS over the three CPUs without\u000ainterrupts reached 929-941 Mbit/s with no CPU above 60%, and RPS over\u000aall four ran CPU0 at 93-96% (2 runs each).\u000a\u000aWhen the receive interrupts of a softirq NAPI device are pinned to some\u000abut not all CPUs, point its RPS mask in the default mode at the other\u000aCPUs and account the load on both. Devices whose interrupts cannot be\u000aread are steered as before.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/26222\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-28 10:01:34 +0100", "id" : "8a97b52c3a90266769c63c8efbbf1666f249e523", "msg" : "netifd: steer receive work away from the interrupt CPUs", "paths" : [ { "editType" : "edit", "file" : "package/network/config/netifd/files/usr/libexec/network/packet-steering.uc" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/ipq807x/base-files/etc/init.d/smp_affinity" ], "commitId" : "e54d48c11e6e45bb662339233389c94dabc87143", "timestamp" : 1793178094000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: ipq807x: re-run packet steering after pinning the IRQs\u000a\u000aThe packet steering script decides from where the receive interrupts of\u000aa device are pinned. smp_affinity pins the ath11k rings at START=93,\u000aafter packet steering ran at START=25, so steering only sees the spread\u000aif an interface event happens to reload it afterwards.\u000a\u000aReload packet steering once the interrupts are pinned.\u000a\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/26222\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-28 10:01:34 +0100", "id" : "e54d48c11e6e45bb662339233389c94dabc87143", "msg" : "qualcommax: ipq807x: re-run packet steering after pinning the IRQs", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/ipq807x/base-files/etc/init.d/smp_affinity" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "package/network/services/ustp/Makefile" ], "commitId" : "e80838fb2a84cc9537c8540382d7af0e8826c35d", "timestamp" : 1793185014000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/nbd", "fullName" : "nbd" }, "authorEmail" : "nbd@nbd.name", "comment" : "ustp: update to Git HEAD (2026-10-28)\u000a\u000a22cb867fd37e netif_utils: size sysfs path buffers for the longest interface name\u000ad13acc7252ab netif_utils: close the port file on a read error\u000adbf456dfe9ad ubus: take the bridge priority from add_bridge\u000a1be05eb1239f config: free the config of removed bridges\u000a\u000aSigned-off-by: Felix Fietkau \u000a", "date" : "2026-10-28 11:56:54 +0100", "id" : "e80838fb2a84cc9537c8540382d7af0e8826c35d", "msg" : "ustp: update to Git HEAD (2026-10-28)", "paths" : [ { "editType" : "edit", "file" : "package/network/services/ustp/Makefile" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "package/network/config/netifd/Makefile" ], "commitId" : "0747c5b48890e32876eb44659a76473ebc4ae4f5", "timestamp" : 1793185019000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/nbd", "fullName" : "nbd" }, "authorEmail" : "nbd@nbd.name", "comment" : "netifd: update to Git HEAD (2026-10-28)\u000a\u000af898a7321f80 system-linux: refuse netns move of an unresolved device\u000ab81713f8cac8 interface: defer jailed device move until the device exists\u000ac34c90f38db9 interface: drop a pending jail move when the jail config changes\u000ae9414e7f8482 bridge: take bridge-vlan configuration as part of the bridge config\u000a05574f61fff5 device: replace config devices that were removed from the config\u000a8cfc2c03959d bridge: apply member vlans from one place\u000affe88f2ce0f1 system-linux: do not leak the callback in system_bridge_vlan_check()\u000a6d4cac2b6e2f bridge: handle a failed member allocation in bridge_hotplug_add()\u000acdf3e8d097d4 vlandev: keep the old config if copying the new one fails\u000a2ac163481042 vlandev: restart the device when its vid alias changes\u000ab1d52c59304e bridge: tie the vlan handling to vlan filtering and configured vlans\u000acfaf6dd3d489 bridge: keep hotplug vlans when the config drops their vlan\u000abba15bc16284 bridge: release hotplug vlans that no member uses any more\u000a967189471692 bridge: create bridge vlans only for tracked hotplug entries\u000a569ff4f7e374 bridge: make a hotplug add set the member's hotplug vlans\u000a4a8597e9adc1 interface: keep the member when the master of a link stays the same\u000a230521b6253d bridge: handle a failed vlan allocation in bridge_hotplug_get_vlan()\u000a9a63cdaf0a0d bridge: keep range vlans when a tracked vlan of the port goes away\u000a70355bfebe6d bridge: accept range vlans of a port in the vlan check\u000a2a4c065b043a bridge: write the bridge's own extra vlans only with vlan filtering\u000a79a5da4bfce5 vlan, vlandev: forward hotplug links only to a bridge that filters vlans\u000af39e5958edb1 vlandev: check hotplug operations on vlan updates\u000a9d3dac944a60 bridge: accept auth vlans of a port in the vlan check\u000a09af141e97f3 bridge: remove all hotplug vlans of a member on hotplug del\u000a9988b0bd897e bridge: group the vlans of a member by source\u000a423123d7135b bridge: track config and hotplug ownership of members\u000a0d37020faeee device: keep the hotplug link on device replacement, iterate all devices\u000a07518dc5ad9c interface: keep an external hotplug device a hotplug main device on reload\u000aed7603dc842a interface: replace a hotplug main device with a newly linked one\u000a976eb74f9fbf extdev: do not announce a topology change for an existing member\u000aab357c2f237e interface: apply stored hotplug links again after config and device changes\u000a13002e858a70 device, bridge: add the bridge and bridge_vlan device options\u000a5f7a462bcbcb bridge: keep the default vlan 1 without bridge-vlan sections\u000ab85814b0ee17 bridge: keep auth vlans when ranges or vlans of a port are removed\u000a58768be7bf62 system-linux: check the allocations in system_bridge_vlan_check()\u000ac9041c7fa7b4 system-linux: apply the state of system_if_check() after the exchange\u000a3443837ab8cd system-linux: read the link state again on a link event\u000a721103dbeafb system-dummy: log the bridge operations\u000af077e15544ca bridge: add the bridge-vlan option local_pvid\u000a7ed52d1e649f bridge: make a kept dynamic vlan a tagged vlan of the bridge\u000a866dffb41cac ubus: walk a copy of procd's data in netifd_ubus_data_cb()\u000a8da363f546f5 main: end a helper with the netifd that started it\u000ab359355f3349 proto-ext: remove the host route when the interface goes down\u000a80e17d8b74aa build: generate ethtool-modes.h in an out-of-tree build\u000aedfe9877ee7d main: end the helpers with SIGTERM when netifd exits\u000a48e4f695773c bridge: move the vlan setup of the bridge and of a port into helpers\u000ad7d270a5a9a1 bridge: apply a change of vlan_filtering to the running bridge\u000a86d56ad76c67 config: add a procd network-device entry in the pass of its type\u000a2fec6e82409b config: add a procd network-device entry without its uci section first\u000a06fc90458560 device: add a board.json default for the pse option\u000a1816971a93c3 main: only use prctl() on linux\u000a13c7e63045ef device: add IPv6 NDP tunable device options\u000a19d2392ff5bb bridge: add per-VLAN multicast snooping\u000aadd2e9e7a29e device: add ARP and per-family neighbour options\u000a8a268294c8eb system-linux: empty the event queue before the link dump\u000ae4b847c71b50 main: wait for the helpers before exiting\u000abf7f2afae02a vlandev, macvlan: delete a stale device before creating it\u000a2922ee823259 main: start no helper while netifd waits for its exit\u000a7d1555e10a6a vrf: delete a stale device before creating it\u000a20be6d5c7908 bridge: pass the bridge priority to the userspace STP daemon\u000aade072a34c93 bridge: use a valid bridge priority as default\u000aa1e39a458f96 bridge: ignore a bridge priority above 65535\u000a\u000aFixes: https://github.com/openwrt/openwrt/issues/22905\u000aSigned-off-by: Felix Fietkau \u000a", "date" : "2026-10-28 11:56:59 +0100", "id" : "0747c5b48890e32876eb44659a76473ebc4ae4f5", "msg" : "netifd: update to Git HEAD (2026-10-28)", "paths" : [ { "editType" : "edit", "file" : "package/network/config/netifd/Makefile" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/bcm53xx/patches-6.18/348-v7.4-PCI-Expose-devm_pci_alloc_host_bridge_release.patch", "target/linux/bcm53xx/patches-6.18/351-v7.4-PCI-iproc-Use-the-EROM-outbound-window-on-BCMA.patch", "target/linux/bcm53xx/patches-6.18/351-PCI-iproc-Use-the-EROM-outbound-window-on-BCMA.patch" ], "commitId" : "79aba5e09246e52081935700b003167e9540ba63", "timestamp" : 1793189795000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "bcm53xx: backport the mainline PCIe EROM window fix\u000a\u000aThe fix carried as 351 since a36ed219f2 was reworked upstream and merged\u000afor Linux v7.4. Replace it with the two mainline commits:\u000a\u000a 39d1c645bb99 (\"PCI: Expose devm_pci_alloc_host_bridge_release()\")\u000a 9850907ee1dd (\"PCI: iproc: Use the EROM outbound window on BCMA\")\u000a\u000aThe first one adds a helper the second one uses, so it goes in as 348.\u000a\u000aLike the old 351, the upstream version requests only the memory window\u000athe enumeration ROM reports. When the DT window differs from it, the\u000adriver now prints a warning, where the old 351 dropped it silently. The\u000adts (349) carries the addresses of PCIe core revision 0x07, so boards\u000awith revision 0x01 now log \"does not match EROM window\" for controllers\u000a1 and 2. The window in use is the same as before.\u000a\u000aTested on an ASUS RT-N18U (BCM47081, core revision 0x01): the BCM4360\u000aon bcma0:7 is found and its link comes up, and /proc/iomem and the PCI\u000adevice list match the old 351.\u000a\u000aSigned-off-by: Semih Baskan \u000aLink: https://github.com/openwrt/openwrt/pull/26237\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-28 13:16:35 +0100", "id" : "79aba5e09246e52081935700b003167e9540ba63", "msg" : "bcm53xx: backport the mainline PCIe EROM window fix", "paths" : [ { "editType" : "add", "file" : "target/linux/bcm53xx/patches-6.18/351-v7.4-PCI-iproc-Use-the-EROM-outbound-window-on-BCMA.patch" }, { "editType" : "delete", "file" : "target/linux/bcm53xx/patches-6.18/351-PCI-iproc-Use-the-EROM-outbound-window-on-BCMA.patch" }, { "editType" : "add", "file" : "target/linux/bcm53xx/patches-6.18/348-v7.4-PCI-Expose-devm_pci_alloc_host_bridge_release.patch" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/generic/pending-6.18/835-03-leds-trigger-netdev-fix-sysfs-update-group-races.patch", "target/linux/generic/pending-6.18/835-05-leds-trigger-netdev-serialize-mode-interval-stores-with-trigger-lock.patch", "target/linux/generic/pending-6.18/835-01-leds-trigger-netdev-put-netdev-on-activate-error-path.patch", "target/linux/generic/pending-6.18/835-04-leds-trigger-netdev-report-invalid-interval-to-user-space.patch", "target/linux/generic/pending-6.18/835-02-leds-trigger-netdev-access-net-dev-under-trigger-data-lock-in-the-worker.patch" ], "commitId" : "e9c8d91f6dd1db454af504412fc6050cedbc454b", "timestamp" : 1793191895000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "generic: 6.18: add pending netdev LED trigger race fixes\u000a\u000aImport pending Alexander Sverdlin's v6 series [1] to fix races between\u000anetdev trigger activation and link-speed sysfs updates. The race can cause\u000aduplicate link_10 entries during boot and leave orphaned sysfs files.\u000a\u000aAdapt the link-speed and mode-store hunks to the attributes and bit\u000aoperations present in 6.18.55.\u000a\u000a[1] https://lore.kernel.org/linux-leds/20261027113349.383147-1-alexander.sverdlin@siemens.com/\u000a\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-28 13:51:35 +0100", "id" : "e9c8d91f6dd1db454af504412fc6050cedbc454b", "msg" : "generic: 6.18: add pending netdev LED trigger race fixes", "paths" : [ { "editType" : "add", "file" : "target/linux/generic/pending-6.18/835-05-leds-trigger-netdev-serialize-mode-interval-stores-with-trigger-lock.patch" }, { "editType" : "add", "file" : "target/linux/generic/pending-6.18/835-04-leds-trigger-netdev-report-invalid-interval-to-user-space.patch" }, { "editType" : "add", "file" : "target/linux/generic/pending-6.18/835-01-leds-trigger-netdev-put-netdev-on-activate-error-path.patch" }, { "editType" : "add", "file" : "target/linux/generic/pending-6.18/835-02-leds-trigger-netdev-access-net-dev-under-trigger-data-lock-in-the-worker.patch" }, { "editType" : "add", "file" : "target/linux/generic/pending-6.18/835-03-leds-trigger-netdev-fix-sysfs-update-group-races.patch" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "package/network/services/ppp/files/50-ppp-radius-servers-mode.sh", "package/network/services/ppp/files/pppd.capabilities", "package/network/services/ppp/files/ppp.sh", "package/network/services/ppp/files/50-ppp-secrets-mode.sh", "package/network/services/ppp/Makefile", "package/network/services/ppp/files/pppd_acl.json" ], "commitId" : "b855f33ea91674c5858dd5c12ed108dac4d38601", "timestamp" : 1793192222000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "ppp: run pppd in ujail as an unprivileged user\u000a\u000apppd drops to user ppp (uid 454, created via USERID) with CAP_NET_ADMIN\u000a(/dev/ppp, PPPIOCNEWUNIT, SIOCSIF*) and CAP_NET_RAW (pppoe PF_PACKET),\u000aboth ambient so ip-up/ip-down scripts keep them, with\u000aPR_SET_NO_NEW_PRIVS. The wrapper covers every proto that routes through\u000appp_generic_setup() and is skipped when /sbin/ujail, the capabilities\u000afile or the ppp user is missing. ujail exits EXIT_FAILURE on a -U name\u000ait cannot resolve, and a sysupgrade can preserve an /etc/passwd that\u000apredates the uid.\u000a\u000apppd's euid != 0 gate (EXIT_NOT_ROOT) would stop the drop. ppp 2.5.4\u000alets pppd run without root when it holds CAP_NET_ADMIN\u000a(ppp-project/ppp#608) and grants the privileged option flag on the\u000acapability, which proto_pppoe_setup() needs for its OPT_PRIV \"plugin\u000apppoe.so\", unless the exec itself raised the privileges (AT_SECURE).\u000a\u000aThe wrapper also hands /dev/ppp to the ppp user. The node is opened at\u000alink setup (generic_establish_ppp), not only in the (stubbed) kernel\u000asupport check, and ships 0600 root:root. chown keeps it reachable for\u000auid 454 alone; the kernel's ppp_open() gates on\u000ans_capable(CAP_NET_ADMIN) either way.\u000a\u000apppd mkdir_recursive()s PPP_PATH_VARRUN and writes its .pid\u000aand pppd2.tdb there, which is /var/run/pppd with OpenWrt's\u000a--localstatedir=/var; nothing in this package creates that directory\u000atoday, and uid 454 cannot do it in a root-owned /var/run, so the\u000awrapper pre-creates and chowns it. A failure to write the pid file is\u000aa logged error and a failed tdb open a warning (multilink drops),\u000aneither fatal. usepeerdns makes pppd rewrite /etc/ppp/resolv.conf, the\u000apackage's symlink to /tmp/resolv.conf.ppp, at every connect; a copy left\u000aby a root-run instance across an in-place upgrade refuses uid 454, so\u000athe wrapper removes it first. Nothing in OpenWrt reads that file, the\u000aDNS servers reach netifd through the DNS1/DNS2 environment.\u000a\u000aThe ip-up/ip-down scripts run inside the jail and report back to netifd\u000aover ubus; ubusd's ACL default is deny for every non-root uid, so the\u000ajailed notify_proto call failed silently and the interface never\u000acompleted setup while the PPP session sat connected. Ship\u000a/usr/share/acl.d/pppd.json granting uid 454 access to network.interface\u000anotify_proto and network add_dynamic (ppp6-up's dynamic-interface\u000aregistry call). The postinst reloads the ubusd ACLs, so an upgrade on a\u000arunning router does not leave the interface down. User hooks in\u000a/etc/ppp/ip-up.d and ip-down.d run the same way, as uid 454 with\u000aCAP_NET_ADMIN and CAP_NET_RAW only; a hook that needs root no longer\u000agets it.\u000a\u000aTwo things the unprivileged uid cannot reach on its own. A serial link is\u000ahanded its TTY as a pppd argument - first by proto_ppp_setup, last by\u000acomgt's proto_3g_setup - and a TTY is root:dialout 0660, which\u000aCAP_NET_ADMIN does not override, so the wrapper grants whichever\u000aargument is a character device, the same way it already grants /dev/ppp.\u000aAnd the secrets pppd authenticates from ship 0600 root:root -\u000achap-secrets here, the RADIUS server list in ppp-mod-radius - so a\u000auci-defaults script regrades them to 0640 root:ppp. That has to run on\u000athe device: a postinst runs on the build host with IPKG_INSTROOT set,\u000awhere chgrp resolves the group name through the host's /etc/group and\u000anot the target's, so the image would ship them root:root regardless.\u000a\u000aAssisted-by: Claude:claude-opus-5\u000aSigned-off-by: Julius Bairaktaris \u000aLink: https://github.com/openwrt/openwrt/pull/25537\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-28 13:57:02 +0100", "id" : "b855f33ea91674c5858dd5c12ed108dac4d38601", "msg" : "ppp: run pppd in ujail as an unprivileged user", "paths" : [ { "editType" : "add", "file" : "package/network/services/ppp/files/50-ppp-radius-servers-mode.sh" }, { "editType" : "add", "file" : "package/network/services/ppp/files/pppd.capabilities" }, { "editType" : "edit", "file" : "package/network/services/ppp/Makefile" }, { "editType" : "add", "file" : "package/network/services/ppp/files/50-ppp-secrets-mode.sh" }, { "editType" : "edit", "file" : "package/network/services/ppp/files/ppp.sh" }, { "editType" : "add", "file" : "package/network/services/ppp/files/pppd_acl.json" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "package/system/rpcd/Makefile" ], "commitId" : "8ac5e0118cfab77d0978a1ce523ec420629c8a1f", "timestamp" : 1793192606000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "rpcd: update to Git HEAD (2026-10-28)\u000a\u000a469d7852316e session: enforce session ACLs for non-root ubus callers\u000a\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-28 14:03:26 +0100", "id" : "8ac5e0118cfab77d0978a1ce523ec420629c8a1f", "msg" : "rpcd: update to Git HEAD (2026-10-28)", "paths" : [ { "editType" : "edit", "file" : "package/system/rpcd/Makefile" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "package/system/ubus/Makefile" ], "commitId" : "f79894ab5720ed505ef40e3d99fab1c87baa0036", "timestamp" : 1793192606000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "ubus: update to Git HEAD (2026-10-28)\u000a\u000adb20eff5535f libubus: add a pre-invoke callback for resolved object methods\u000a\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-28 14:03:26 +0100", "id" : "f79894ab5720ed505ef40e3d99fab1c87baa0036", "msg" : "ubus: update to Git HEAD (2026-10-28)", "paths" : [ { "editType" : "edit", "file" : "package/system/ubus/Makefile" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/qualcommax/patches-6.18/0084-03-Align-CMN-PLL-inputs-with-binding.patch", "target/linux/qualcommax/patches-6.18/0123-arm64-dts-ipq8074-add-cooling-cells-to-CPU-nodes.patch", "target/linux/qualcommax/patches-6.18/0130-arm64-dts-qcom-ipq8074-add-CPU-OPP-table.patch", "target/linux/qualcommax/patches-6.18/0121-arm64-dts-ipq8074-Add-WLAN-node.patch", "target/linux/qualcommax/patches-6.18/0120-arm64-dts-qcom-ipq8074-add-secure-WCSS-remoteproc.patch", "target/linux/qualcommax/patches-6.18/0102-arm64-dts-ipq8074-add-reserved-memory-nodes.patch", "target/linux/qualcommax/patches-6.18/0902-arm64-dts-ipq8074-add-label-to-clocks.patch", "target/linux/qualcommax/patches-6.18/0084-02-Allow-CMN-PLL-bias-clocks.patch", "target/linux/qualcommax/files/arch/arm64/boot/dts/qcom/ipq8074-ess.dtsi", "target/linux/qualcommax/patches-6.18/0147-arm64-dts-qcom-ipq8074-add-uniphy-clocks-to-gcc.patch", "target/linux/qualcommax/patches-6.18/0145-arm64-dts-qcom-ipq8074-move-DR-mode-to-board-DTS.patch", "target/linux/qualcommax/patches-6.18/0122-arm64-dts-ipq8074-add-CPU-clock.patch", "target/linux/qualcommax/patches-6.18/0084-04-Add-IPQ8074-CMN-PLL.patch", "target/linux/qualcommax/patches-6.18/0084-01-Correct-IPQ8074-NSS-NoC-rate.patch", "target/linux/qualcommax/patches-6.18/0129-arm64-dts-qcom-ipq8074-add-QFPROM-fuses.patch", "target/linux/qualcommax/patches-6.18/0146-arm64-dts-qcom-ipq8074-Flatten-usb-controller-nodes.patch" ], "commitId" : "4835f60a6d7e28db3e87b17857e1a394f8662745", "timestamp" : 1793194038000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "qualcommax: ipq8074: add pending CMN PLL clock updates\u000a\u000aImport the pending v2 CMN PLL series [1].\u000aits been reviewed and is pending merge upstream.\u000a\u000aIt corrects the NSS NoC output rate, describes the two GCC inputs, resolves\u000athem through device tree references, and adds the CMN PLL to the SoC DTSI.\u000a\u000aRemove the duplicate ESS overlay node. Refresh the later IPQ8074 DTS\u000apatches for the new node and GCC clock list, and extend the GCC binding\u000afor the target's UNIPHY clock inputs.s\u000a\u000a[1] https://patchwork.kernel.org/series/1193706/\u000a\u000aSigned-off-by: Robert Marko \u000a", "date" : "2026-10-28 14:27:18 +0100", "id" : "4835f60a6d7e28db3e87b17857e1a394f8662745", "msg" : "qualcommax: ipq8074: add pending CMN PLL clock updates", "paths" : [ { "editType" : "edit", "file" : "target/linux/qualcommax/patches-6.18/0129-arm64-dts-qcom-ipq8074-add-QFPROM-fuses.patch" }, { "editType" : "edit", "file" : "target/linux/qualcommax/patches-6.18/0146-arm64-dts-qcom-ipq8074-Flatten-usb-controller-nodes.patch" }, { "editType" : "edit", "file" : "target/linux/qualcommax/patches-6.18/0147-arm64-dts-qcom-ipq8074-add-uniphy-clocks-to-gcc.patch" }, { "editType" : "add", "file" : "target/linux/qualcommax/patches-6.18/0084-02-Allow-CMN-PLL-bias-clocks.patch" }, { "editType" : "edit", "file" : "target/linux/qualcommax/patches-6.18/0120-arm64-dts-qcom-ipq8074-add-secure-WCSS-remoteproc.patch" }, { "editType" : "edit", "file" : "target/linux/qualcommax/patches-6.18/0121-arm64-dts-ipq8074-Add-WLAN-node.patch" }, { "editType" : "edit", "file" : "target/linux/qualcommax/patches-6.18/0122-arm64-dts-ipq8074-add-CPU-clock.patch" }, { "editType" : "edit", "file" : "target/linux/qualcommax/patches-6.18/0123-arm64-dts-ipq8074-add-cooling-cells-to-CPU-nodes.patch" }, { "editType" : "add", "file" : "target/linux/qualcommax/patches-6.18/0084-04-Add-IPQ8074-CMN-PLL.patch" }, { "editType" : "edit", "file" : "target/linux/qualcommax/patches-6.18/0102-arm64-dts-ipq8074-add-reserved-memory-nodes.patch" }, { "editType" : "edit", "file" : "target/linux/qualcommax/patches-6.18/0145-arm64-dts-qcom-ipq8074-move-DR-mode-to-board-DTS.patch" }, { "editType" : "add", "file" : "target/linux/qualcommax/patches-6.18/0084-03-Align-CMN-PLL-inputs-with-binding.patch" }, { "editType" : "edit", "file" : "target/linux/qualcommax/patches-6.18/0130-arm64-dts-qcom-ipq8074-add-CPU-OPP-table.patch" }, { "editType" : "add", "file" : "target/linux/qualcommax/patches-6.18/0084-01-Correct-IPQ8074-NSS-NoC-rate.patch" }, { "editType" : "edit", "file" : "target/linux/qualcommax/files/arch/arm64/boot/dts/qcom/ipq8074-ess.dtsi" }, { "editType" : "edit", "file" : "target/linux/qualcommax/patches-6.18/0902-arm64-dts-ipq8074-add-label-to-clocks.patch" } ] }, { "_class" : "hudson.plugins.git.GitChangeSet", "affectedPaths" : [ "target/linux/mediatek/mt7622/base-files/lib/preinit/04_set_netdev_label", "target/linux/microchipsw/lan969x/base-files/lib/preinit/04_set_netdev_label", "target/linux/airoha/base-files/lib/preinit/04_set_netdev_label", "target/linux/qualcommax/ipq50xx/base-files/lib/preinit/04_set_netdev_label", "target/linux/ramips/mt7621/base-files/lib/preinit/04_set_netdev_label" ], "commitId" : "de3a3f7132b801b845e03ef533fe546b3c7a853c", "timestamp" : 1793194457000, "author" : { "absoluteUrl" : "https://taiha.net/jenkins/user/robimarko", "fullName" : "robimarko" }, "authorEmail" : "robimarko@gmail.com", "comment" : "treewide: handle name collisions in set_netdev_labels\u000a\u000aThe 04_set_netdev_label preinit scripts (airoha, mediatek/mt7622,\u000amicrochipsw/lan969x, qualcommax/ipq50xx, ramips/mt7621) rename\u000ainterfaces with a plain \"ip link set name