)]}'
{
  "log": [
    {
      "commit": "6ad246a1494807ee91e36ffcf60e4c30bb309d45",
      "tree": "55a808af978e000950dcb01e6cf0fc87b71a641f",
      "parents": [
        "2de7dfb6813facba8e376897290bedb18ef38401"
      ],
      "author": {
        "name": "Benoît Monin",
        "email": "benoit.monin@bootlin.com",
        "time": "Wed Sep 02 17:04:07 2026 +0200"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Mon Sep 28 18:06:25 2026 +0530"
      },
      "message": "platform: generic: metanoia/mt2824: configure PMA for DSP memory regions\n\nSet up the memory attribute for the XOR RAM, MRAS memory, and RSC table\nused by the DSPs as non-cacheable and bufferable during final init.\n\nSigned-off-by: Benoît Monin \u003cbenoit.monin@bootlin.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260902-mt2824-pma-v1-1-ada8f836365d@bootlin.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "2de7dfb6813facba8e376897290bedb18ef38401",
      "tree": "e33a858eb311f7998af4f8bdc6457cda356e5f0a",
      "parents": [
        "0ffaa1bc65afe89985bcf3f0c9f564c6b09e3091"
      ],
      "author": {
        "name": "Shibo Zhu",
        "email": "3499129952@qq.com",
        "time": "Mon Aug 17 23:55:37 2026 +0800"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Mon Sep 28 17:42:15 2026 +0530"
      },
      "message": "lib: sbi_mpxy: Fix shared memory size rounding\n\nThe sbi_mpxy_register_channel() currently performs ceiling division by\nPAGE_SIZE but stores the resulting page count in mpxy_shmem_size. All\nconsumers interpret mpxy_shmem_size as a byte count, so channels requiring\nmore than one page get an undersized shared memory value.\n\nAddress the above issue by using ROUNDUP() to keep the rounded value in\nbytes.\n\nFixes: ec09918426a5 (\"lib: sbi: Update MPXY framework and SBI extension as per latest spec\")\nSigned-off-by: Shibo Zhu \u003c3499129952@qq.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/tencent_14AA3BA0083001905C262A5CD64E442FC205@qq.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "0ffaa1bc65afe89985bcf3f0c9f564c6b09e3091",
      "tree": "937c6c5630f30f3cc33d803fdf4243604d0e12e4",
      "parents": [
        "5b1e739428fecb1bb107276442143fb18e4003f4"
      ],
      "author": {
        "name": "Chen Pei",
        "email": "cp0613@linux.alibaba.com",
        "time": "Mon Aug 17 16:43:03 2026 +0800"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Mon Sep 28 11:41:25 2026 +0530"
      },
      "message": "lib/utils/fdt: Reserve exactly enough FDT space for fdt_cpu_fixup\n\nfdt_cpu_fixup() reserves a hardcoded 32 extra bytes, plus 16 more per\nHART inside the \"zicntr\" loop. Those 32 bytes also have to cover every\n\"status\" property rewritten to \"disabled\", so once more than a handful\nof HARTs are disabled the fixups fail with FDT_ERR_NOSPACE (-3).\n\nBumping the reservation to a fixed worst case is not safe either:\nfdt_open_into() cannot know how much space the caller\u0027s buffer really\nhas, it just trusts the requested size. As OpenSBI does not own the\nmemory following a device tree passed in by the previous booting stage,\nover-declaring the blob risks corrupting whatever follows it.\n\nWalk the HART nodes once without modifying them instead, and reserve\nexactly what the fixups need. The read-only decision logic is factored\ninto fdt_cpu_fixup_needed() so the sizing and fixup passes cannot drift\napart, which also makes the per-HART fdt_open_into() redundant. Check\nthe return values of fdt_setprop_string() and fdt_appendprop_string()\nas well, as running out of space used to fail silently.\n\nFixes: dd9439fbace2 (\"lib: utils: Add a fdt_cpu_fixup() helper\")\nSigned-off-by: Chen Pei \u003ccp0613@linux.alibaba.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260817084303.1232-1-cp0613@linux.alibaba.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "5b1e739428fecb1bb107276442143fb18e4003f4",
      "tree": "685fa630448686b340a8b428e163f0a302b25ce0",
      "parents": [
        "f3ed2c4e5cb7e68af3b18beeb95682c0a6fa505d"
      ],
      "author": {
        "name": "Trevor Gamblin",
        "email": "tgamblin@baylibre.com",
        "time": "Wed Sep 23 17:22:14 2026 -0400"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Mon Sep 28 11:03:50 2026 +0530"
      },
      "message": "Makefile: don\u0027t grep when setting CC_SUPPORT_SAVE_RESTORE\n\nCheck exit status from the compiler/linker check directly (similar to\nOPENSBI_LD_PIE), rather than using grep to search for a pattern. This\navoids potential unrelated matches in the command.\n\nSigned-off-by: Trevor Gamblin \u003ctgamblin@baylibre.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260923-makefile_extension_grep-v2-3-0eb97b8ac076@baylibre.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "f3ed2c4e5cb7e68af3b18beeb95682c0a6fa505d",
      "tree": "baa2c2db5e4d6d18a87f14534ef3f4be1f3f0bd4",
      "parents": [
        "7b4c152c8ff9feffb81648e36180b5d65f1a7e89"
      ],
      "author": {
        "name": "Trevor Gamblin",
        "email": "tgamblin@baylibre.com",
        "time": "Wed Sep 23 17:22:13 2026 -0400"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Mon Sep 28 11:03:50 2026 +0530"
      },
      "message": "Makefile: don\u0027t grep when setting CC_SUPPORT_STRICT_ALIGN\n\nCheck exit status from the compiler/linker check directly (similar to\nOPENSBI_LD_PIE), rather than using grep to search for a pattern. This\navoids potential unrelated matches in the command.\n\nSigned-off-by: Trevor Gamblin \u003ctgamblin@baylibre.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260923-makefile_extension_grep-v2-2-0eb97b8ac076@baylibre.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "7b4c152c8ff9feffb81648e36180b5d65f1a7e89",
      "tree": "883b1277e35644456bd8c16579658cc949d85e38",
      "parents": [
        "5a175017261af97e9122fcb9d9012481cdb5aa41"
      ],
      "author": {
        "name": "Trevor Gamblin",
        "email": "tgamblin@baylibre.com",
        "time": "Wed Sep 23 17:22:12 2026 -0400"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Mon Sep 28 11:03:50 2026 +0530"
      },
      "message": "Makefile: don\u0027t grep when setting CC_SUPPORT_ZICSR_ZIFENCEI\n\nCheck exit status from the compiler/linker check directly (similar to\nOPENSBI_LD_PIE), rather than using grep to search for a pattern. This\navoids potential unrelated matches in the command.\n\nSigned-off-by: Trevor Gamblin \u003ctgamblin@baylibre.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260923-makefile_extension_grep-v2-1-0eb97b8ac076@baylibre.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "5a175017261af97e9122fcb9d9012481cdb5aa41",
      "tree": "3ef872073f887b671c9a33f8a9b704a77b677829",
      "parents": [
        "f83d6a730b083878e5dc5fa476908d2c23460918"
      ],
      "author": {
        "name": "Zong Li",
        "email": "zong.li@sifive.com",
        "time": "Fri Sep 18 01:13:13 2026 -0700"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Fri Sep 18 16:52:39 2026 +0530"
      },
      "message": "docs: firmware: document the OpenSBI firmware header\n\nAdd docs/firmware/fw_header.md describing the layout of the 128-byte\nOpenSBI firmware header, the a0/a1/a2 override flags, what a previous\nbooting stage is expected to put in each register, and an example of how\nto patch the header. Also point at it from fw.md, next to the\ndescription of the registers the previous booting stage passes.\n\nSigned-off-by: Zong Li \u003czong.li@sifive.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260918081313.659655-4-zong.li@sifive.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "f83d6a730b083878e5dc5fa476908d2c23460918",
      "tree": "a26f99ea59aa723298a2529af04f7812844f0f8d",
      "parents": [
        "1854901adcf68d51d3afd84907b88d907627f6b9"
      ],
      "author": {
        "name": "Zong Li",
        "email": "zong.li@sifive.com",
        "time": "Fri Sep 18 01:13:12 2026 -0700"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Fri Sep 18 16:52:39 2026 +0530"
      },
      "message": "firmware: fw_base.S: override a0, a1 and a2 from the firmware header\n\nWire up the flags word and the three override values in the OpenSBI\nfirmware header. When the previous booting stage sets one of the\nFW_HEADER_FLAGS_OVERRIDE_A[012] bits, the cold boot path replaces the\nmatching register with the value stored in the header before doing\nanything else with the boot arguments.\n\nThis lets a previous booting stage which cannot pass the boot arguments\nin registers hand them over by patching a few words in the firmware\nimage instead. For example, a booting stage running on a dedicated boot\nprocessor can load the OpenSBI image, patch the header with the hart id,\nthe DTB address and, for FW_DYNAMIC, the address of a struct\nfw_dynamic_info it built somewhere in DRAM, and then release the hart\nthat runs OpenSBI, without ever being able to set up that hart\u0027s\nregisters itself.\n\nThe flags word is zero in a freshly built image, so nothing is\noverridden and every existing booting stage keeps passing a0, a1 and a2\nin registers as before.\n\nSuggested-by: Anup Patel \u003canup@brainfault.org\u003e\nSigned-off-by: Zong Li \u003czong.li@sifive.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260918081313.659655-3-zong.li@sifive.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "1854901adcf68d51d3afd84907b88d907627f6b9",
      "tree": "9ff869e348c4eb524dce741462b0e93d8259d90a",
      "parents": [
        "9476b7836114b9c873ceda1452e7b486afd6d5c8"
      ],
      "author": {
        "name": "Zong Li",
        "email": "zong.li@sifive.com",
        "time": "Fri Sep 18 01:13:11 2026 -0700"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Fri Sep 18 16:52:39 2026 +0530"
      },
      "message": "firmware: fw_base.S: add a formatted OpenSBI firmware header\n\nAdd a fixed 128-byte header at the very beginning of every OpenSBI\nfirmware image, regardless of the firmware type. The header lets the\nprevious booting stage identify an OpenSBI image and, more importantly,\ngives it a well-known place inside the image to hand over information\nto OpenSBI by patching a few words instead of setting up registers.\n\nThis first patch only introduces the layout:\n\n  - a 4-byte jump over the header to _start_real, so that _start stays\n    the entry point of the image. The jump is assembled with\n    \u0027.option norvc\u0027 so that the fields behind it are always at a fixed\n    offset, even for a build with compressed instructions enabled,\n\n  - the \u0027OSBI\u0027 magic, a header version, the XLEN the firmware was built\n    for and the size of the firmware image, so that the previous booting\n    stage can validate the image and figure out how much memory it\n    occupies,\n\n  - a flags word and three 8-byte override values, which are unused\n    (and zero) for now and are wired up by the next patch.\n\nThe layout is deliberately identical for RV32 and RV64 so that the\nprevious booting stage can parse the header without knowing the XLEN of\nthe firmware upfront. Each override value is followed by a reserved\n8-byte slot so that the same layout can be extended to RV128 later.\n\nSuggested-by: Anup Patel \u003canup@brainfault.org\u003e\nSigned-off-by: Zong Li \u003czong.li@sifive.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260918081313.659655-2-zong.li@sifive.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "9476b7836114b9c873ceda1452e7b486afd6d5c8",
      "tree": "4bf2295a4f0439ace0377b92f536179b21dfd2c2",
      "parents": [
        "92b24d6acbd663d7fc035f22c8fcdec6153c22c4"
      ],
      "author": {
        "name": "Carlos López",
        "email": "carlos.lopezr4096@gmail.com",
        "time": "Wed Sep 09 18:23:23 2026 +0200"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Sep 16 09:48:30 2026 +0530"
      },
      "message": "Makefile: enable Thread Safety Analysis\n\nNow that all locking code has been properly annotated, enable clang\u0027s\nThread Safety Analysis.\n\nTwo flags are available, -Wthread-safety, and the more recent\n-Wthread-safety-pointer. The former was introduced before clang gained\nRISC-V support, so it is guaranteed to be available. The latter was\nintroduced in clang 21, so check if the provided compiler supports it\nbefore enabling it.\n\nSigned-off-by: Carlos López \u003ccarlos.lopezr4096@gmail.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260909162322.29778-16-carlos.lopezr4096@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "92b24d6acbd663d7fc035f22c8fcdec6153c22c4",
      "tree": "a0dde9456df064830a75d9b205d4a97c6eaae46e",
      "parents": [
        "8f80b2651fe714ba09ed16fb32b562ad43e2bfab"
      ],
      "author": {
        "name": "Carlos López",
        "email": "carlos.lopezr4096@gmail.com",
        "time": "Wed Sep 09 18:23:22 2026 +0200"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Sep 16 09:48:30 2026 +0530"
      },
      "message": "lib: test: disable TSA for spinlock tests\n\nThe spinlock tests deliberately use unconventional locking patterns to\ntest edge cases. Disable thread safety analysis for them so that they\ndon\u0027t generate false positives when the feature is enabled.\n\nSigned-off-by: Carlos López \u003ccarlos.lopezr4096@gmail.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260909162322.29778-15-carlos.lopezr4096@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "8f80b2651fe714ba09ed16fb32b562ad43e2bfab",
      "tree": "f4104613a5eef0b1ae639b6d2e4cae8f6b2addef",
      "parents": [
        "43bb3df782437b81aced73dc86579228d5ab9872"
      ],
      "author": {
        "name": "Carlos López",
        "email": "carlos.lopezr4096@gmail.com",
        "time": "Wed Sep 09 18:23:21 2026 +0200"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Sep 16 09:48:30 2026 +0530"
      },
      "message": "lib: sbi_sse: add TSA annotations\n\nNow that the lock relationship between an event and its owner hart is\nexplicit, add TSA annotations for the SSE handling code. This includes\nindicating which fields are protected by a spinlock, as well as\nannotating which functions need to run under a spinlock.\n\nExclude sse_local_init(), which initializes the spinlock to be acquired,\nand sse_event_get() / sse_event_put(), which perform conditional locking\nthat TSA is not able to follow properly.\n\nSigned-off-by: Carlos López \u003ccarlos.lopezr4096@gmail.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260909162322.29778-14-carlos.lopezr4096@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "43bb3df782437b81aced73dc86579228d5ab9872",
      "tree": "ee7cecaf50490d2362b9754a21ecc0acc5f6a8f0",
      "parents": [
        "fc1ae7a641b9182fb040e11efd524f475189f8ee"
      ],
      "author": {
        "name": "Carlos López",
        "email": "carlos.lopezr4096@gmail.com",
        "time": "Wed Sep 09 18:23:20 2026 +0200"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Sep 16 09:48:30 2026 +0530"
      },
      "message": "lib: sbi_sse: pass SSE hart state explicitly\n\nIn preparation to add TSA annotations, pass the SSE hart state owning an\nevent, instead of relying on sse_get_hart_state() to retrieve it from\nthe event itself. This makes the relationship more explicit, and will\nallow the compiler to reason about the state of locks.\n\nSigned-off-by: Carlos López \u003ccarlos.lopezr4096@gmail.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260909162322.29778-13-carlos.lopezr4096@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "fc1ae7a641b9182fb040e11efd524f475189f8ee",
      "tree": "371a0e847edd1a4a10b3bb747eea5b4074a9b82e",
      "parents": [
        "8740e0a38cd647e41fb83b5898ac10ce70596803"
      ],
      "author": {
        "name": "Carlos López",
        "email": "carlos.lopezr4096@gmail.com",
        "time": "Wed Sep 09 18:23:19 2026 +0200"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Sep 16 09:48:30 2026 +0530"
      },
      "message": "lib: sbi_sse: inline enable event locking\n\nIn preparation to add TSA annotations, inline sse_enabled_event_lock()\nand sse_enabled_event_unlock(). This will help the compiler reason about\nlocking of the hart state by explicitly acquiring the spinlock in the\ncallers.\n\nSigned-off-by: Carlos López \u003ccarlos.lopezr4096@gmail.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260909162322.29778-12-carlos.lopezr4096@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "8740e0a38cd647e41fb83b5898ac10ce70596803",
      "tree": "4fcdd6779dd95c544835dd9551e6f42ce3984402",
      "parents": [
        "1ba332872b5fbb50d83abef1f8f451a28b5119fb"
      ],
      "author": {
        "name": "Carlos López",
        "email": "carlos.lopezr4096@gmail.com",
        "time": "Wed Sep 09 18:23:18 2026 +0200"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Sep 16 09:48:30 2026 +0530"
      },
      "message": "lib: sbi_domain: add TSA annotations\n\nAdd Thread Safety Analysis (TSA) annotations for the domain handling\ncode. This includes indicating which fields are protected by a spinlock,\nand annotating which functions acquire a spinlock, in order to prevent\nnesting. Exclude sbi_domain_register() from analysis, as it initializes\na domain\u0027s spinlock.\n\nSigned-off-by: Carlos López \u003ccarlos.lopezr4096@gmail.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260909162322.29778-11-carlos.lopezr4096@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "1ba332872b5fbb50d83abef1f8f451a28b5119fb",
      "tree": "d23f1bf033d08a103cb1f296f9bc8da878c625c2",
      "parents": [
        "f5fd674f1817afeaca58933a71ff106c59dafa9a"
      ],
      "author": {
        "name": "Carlos López",
        "email": "carlos.lopezr4096@gmail.com",
        "time": "Wed Sep 09 18:23:17 2026 +0200"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Sep 16 09:48:30 2026 +0530"
      },
      "message": "lib: htif: add TSA annotations\n\nAdd Thread Safety Analysis (TSA) annotations for the HTIF driver\nimplementation.\n\nThis requires annotating the tohost/fromhost and htif_console_buf\nglobals to indicate that they are protected by htif_lock.\nhtif_system_reset() can be excluded since it is only called under\nspecial circumstances from sbi_system_reset(), when all other harts\nhave been stopped.\n\nSigned-off-by: Carlos López \u003ccarlos.lopezr4096@gmail.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260909162322.29778-10-carlos.lopezr4096@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "f5fd674f1817afeaca58933a71ff106c59dafa9a",
      "tree": "6badade327fd507114ca3111cb2aabd1b9531f2c",
      "parents": [
        "497ca4ccad0a8d74cdc6aa165398f53352809983"
      ],
      "author": {
        "name": "Carlos López",
        "email": "carlos.lopezr4096@gmail.com",
        "time": "Wed Sep 09 18:23:16 2026 +0200"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Sep 16 09:48:30 2026 +0530"
      },
      "message": "lib: sbi_timer: add TSA annotations\n\nAdd Thread Safety Analysis (TSA) annotations for the timer\nimplementation. This includes indicating which fields are protected\nby a spinlock, and annotating which functions must be called under a\nspinlock. Exclude sbi_timer_init(), as the spinlock cannot be acquired\nbefore it is initialized.\n\nSigned-off-by: Carlos López \u003ccarlos.lopezr4096@gmail.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260909162322.29778-9-carlos.lopezr4096@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "497ca4ccad0a8d74cdc6aa165398f53352809983",
      "tree": "b5192d1f14c9013a1299397747e787e2149bdb71",
      "parents": [
        "2d34c0e0e1fe205a60eb4185900c78b1844b2214"
      ],
      "author": {
        "name": "Carlos López",
        "email": "carlos.lopezr4096@gmail.com",
        "time": "Wed Sep 09 18:23:15 2026 +0200"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Sep 16 09:48:30 2026 +0530"
      },
      "message": "lib: rpmi: add TSA annotations\n\nAdd Thread Safety Analysis (TSA) annotations for the RPMI mailbox\nimplementation. This includes indicating which fields are protected\nby a spinlock, and annotating which functions must be called under a\nspinlock. Exclude rpmi_shmem_transport_init(), as the spinlock cannot be\nacquired before it is initialized.\n\nSigned-off-by: Carlos López \u003ccarlos.lopezr4096@gmail.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260909162322.29778-8-carlos.lopezr4096@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "2d34c0e0e1fe205a60eb4185900c78b1844b2214",
      "tree": "04a373c5d68b040fd8ffb04ba795e368a0802ddb",
      "parents": [
        "eaac726ea11e4f1a2a36353e137fc8d0d0c0a298"
      ],
      "author": {
        "name": "Carlos López",
        "email": "carlos.lopezr4096@gmail.com",
        "time": "Wed Sep 09 18:23:14 2026 +0200"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Sep 16 09:48:30 2026 +0530"
      },
      "message": "lib: sbi_scratch: add TSA annotations\n\nAdd Thread Safety Analysis (TSA) annotations for the scratch space\nimplementation. This only consists of annotating that the extra_offset\nglobal must be accessed under the extra_lock spinlock.\n\nSigned-off-by: Carlos López \u003ccarlos.lopezr4096@gmail.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260909162322.29778-7-carlos.lopezr4096@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "eaac726ea11e4f1a2a36353e137fc8d0d0c0a298",
      "tree": "a5db82b247b4fac122d35e79c7e76556735769df",
      "parents": [
        "832ede6a8fcb621097b20cb5e096260e7a3a1272"
      ],
      "author": {
        "name": "Carlos López",
        "email": "carlos.lopezr4096@gmail.com",
        "time": "Wed Sep 09 18:23:13 2026 +0200"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Sep 16 09:48:30 2026 +0530"
      },
      "message": "lib: sbi_heap: add TSA annotations\n\nAdd Thread Safety Analysis (TSA) annotations for the heap allocator.\nThis includes indicating which fields are protected by a spinlock and\nannotating functions that must be called under a spinlock. Exclude\nsbi_heap_init_new(), as the spinlock cannot be acquired before it is\ninitialized.\n\nSigned-off-by: Carlos López \u003ccarlos.lopezr4096@gmail.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260909162322.29778-6-carlos.lopezr4096@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "832ede6a8fcb621097b20cb5e096260e7a3a1272",
      "tree": "f108262c511a72e722950aa8676f7cd0f631b091",
      "parents": [
        "b8f7e9ea80b94dd3c8eb0fa434265e376e9e3330"
      ],
      "author": {
        "name": "Carlos López",
        "email": "carlos.lopezr4096@gmail.com",
        "time": "Wed Sep 09 18:23:12 2026 +0200"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Sep 16 09:48:30 2026 +0530"
      },
      "message": "lib: sbi_fifo: add TSA annotations\n\nAdd Thread Safety Analysis (TSA) annotations for the FIFO\nimplementation. This consists of indicating which fields are protected\nby a spinlock and annotating functionst that must be called under a\nspinlock. sbi_fifo_init() must be excluded, as acquiring a spinlock\nbefore it is initialized does not make sense.\n\nSigned-off-by: Carlos López \u003ccarlos.lopezr4096@gmail.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260909162322.29778-5-carlos.lopezr4096@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "b8f7e9ea80b94dd3c8eb0fa434265e376e9e3330",
      "tree": "a24ecaebfe019bfaa782096a70d76d637b41e1c9",
      "parents": [
        "8d109c7e87a30a181156890bd5d19542f3b3a24d"
      ],
      "author": {
        "name": "Carlos López",
        "email": "carlos.lopezr4096@gmail.com",
        "time": "Wed Sep 09 18:23:11 2026 +0200"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Sep 16 09:48:30 2026 +0530"
      },
      "message": "lib: sbi_locks: annotate spinlock APIs for TSA\n\nAdd Thread Safety Analysis (TSA) annotations for spinlock APIs, allowing\nthe compiler to reason about lock acquisition and release.\n\nAnnotations have a different meaning if used in header declarations vs C\nfile implementations. In header declarations, they specify the semantics\nof the given function, meaning that we must specify that the spinlock\nfunctions acquire and release a lock (\"capability\" in clang terms).\nIn function implementations, attributes affect the function body and\ncallers in the same translation unit; since the spinlock functions\ncontain inline assembly, they must be excluded from analysis.\n\nSigned-off-by: Carlos López \u003ccarlos.lopezr4096@gmail.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260909162322.29778-4-carlos.lopezr4096@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "8d109c7e87a30a181156890bd5d19542f3b3a24d",
      "tree": "597d8e0cb64b2a11d2075360c46e27e60294ba4c",
      "parents": [
        "55b6b73435ec81de612888bdd3e24bc581b3a2d2"
      ],
      "author": {
        "name": "Carlos López",
        "email": "carlos.lopezr4096@gmail.com",
        "time": "Wed Sep 09 18:23:10 2026 +0200"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Sep 16 09:48:30 2026 +0530"
      },
      "message": "include: sbi: add thread safety analysis macros\n\nAdd macros to properly guard clang\u0027s thread safety analysis attributes\nkeeping compatibility with GCC, on top of making things more readable.\n\nClang \u003c 19 does not support guarded_by() [0], so do not define the\nmacros in that case.\n\n[0] https://github.com/llvm/llvm-project/pull/94216\n\nSigned-off-by: Carlos López \u003ccarlos.lopezr4096@gmail.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260909162322.29778-3-carlos.lopezr4096@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "55b6b73435ec81de612888bdd3e24bc581b3a2d2",
      "tree": "209b47af2cfd2ad2c328c46b3d120c7a4e967f96",
      "parents": [
        "71faf5a307f06950f6c1ce5145ff640b441334da"
      ],
      "author": {
        "name": "Anup Patel",
        "email": "anup.patel@oss.qualcomm.com",
        "time": "Sat Sep 05 18:47:48 2026 +0530"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Sep 16 09:21:21 2026 +0530"
      },
      "message": "lib: sbi_domain: Introduce domain intialization order\n\nCurrently, the domain initialization order is implied by the order\nin which domains are populated by sbi_platform_domains_init() from\nsbi_domain_finalize(). This is not documented anywhere and forces\nunecessary ordering between domain DT nodes.\n\nTo address the above, introduce per-domain 32-bit integer to represent\nintialization order (aka \"init_order\") where domain with a lower\ninitialization order will be booted first and two domains must not\nhave same initialization order. For DT based domain creation, new\n\"init-order\" DT property can be used in domain DT node to specify\nthe initialization order. The ROOT domain is assumed to have lowest\ninitialization order (aka 0xffffffff).\n\nSigned-off-by: Anup Patel \u003canup.patel@oss.qualcomm.com\u003e\nReviewed-by: Pawandeep Oza \u003cpawandeep.oza@oss.qualcomm.com\u003e\nTested-by: Pawandeep Oza \u003cpawandeep.oza@oss.qualcomm.com\u003e\nReviewed-by: Yu-Chien Peter Lin \u003cpeter.lin@sifive.com\u003e\nLink: https://lore.kernel.org/r/20260905131748.922920-4-anup.patel@oss.qualcomm.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "71faf5a307f06950f6c1ce5145ff640b441334da",
      "tree": "d1944b5059c2a68af0bb5500dcda54c3a102f275",
      "parents": [
        "6dc5a896b1bb5cf1ab671458436680f80e786ad5"
      ],
      "author": {
        "name": "Anup Patel",
        "email": "anup.patel@oss.qualcomm.com",
        "time": "Sat Sep 05 18:47:47 2026 +0530"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Sep 16 09:21:21 2026 +0530"
      },
      "message": "lib: sbi_domain: Re-work boot-time assignment of HARTs to non-ROOT domains\n\nAssign a non-ROOT domain to a HART on first come first serve basis if the\nHART is listed as a possible HART of the non-ROOT domain. If no non-ROOT\ndomain list a HART as possible HART then the HART is assigned to the ROOT\ndomain.\n\nThis allows us to drop the OpenSBI specific DT property from each CPU DT\nnode (aka \"opensbi-domain\" Dt property).\n\nSigned-off-by: Anup Patel \u003canup.patel@oss.qualcomm.com\u003e\nReviewed-by: Pawandeep Oza \u003cpawandeep.oza@oss.qualcomm.com\u003e\nTested-by: Pawandeep Oza \u003cpawandeep.oza@oss.qualcomm.com\u003e\nReviewed-by: Yu-Chien Peter Lin \u003cpeter.lin@sifive.com\u003e\nLink: https://lore.kernel.org/r/20260905131748.922920-3-anup.patel@oss.qualcomm.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "6dc5a896b1bb5cf1ab671458436680f80e786ad5",
      "tree": "1fb50019f0e5181228d20125fec203b47d49a166",
      "parents": [
        "3593a5facc4c6938b90429a6973ba9ee21fc5899"
      ],
      "author": {
        "name": "Anup Patel",
        "email": "anup.patel@oss.qualcomm.com",
        "time": "Sat Sep 05 18:47:46 2026 +0530"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Sep 16 09:21:21 2026 +0530"
      },
      "message": "lib: sbi_domain: Check possible harts in sbi_domain_context_enter/exit()\n\nWhen context switching to a domain the current hart MUST be part\nof the possible harts of that domain. Add appropriate checks in\nsbi_domain_context_enter/exit() along these lines.\n\nSigned-off-by: Anup Patel \u003canup.patel@oss.qualcomm.com\u003e\nReviewed-by: Pawandeep Oza \u003cpawandeep.oza@oss.qualcomm.com\u003e\nTested-by: Pawandeep Oza \u003cpawandeep.oza@oss.qualcomm.com\u003e\nReviewed-by: Yu-Chien Peter Lin \u003cpeter.lin@sifive.com\u003e\nLink: https://lore.kernel.org/r/20260905131748.922920-2-anup.patel@oss.qualcomm.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "3593a5facc4c6938b90429a6973ba9ee21fc5899",
      "tree": "f933b0b6148c2a1e4467435aa2f1eae6af619e8e",
      "parents": [
        "35af7c8c7ee00869afc6e9e0aca64086888bd03f"
      ],
      "author": {
        "name": "Rahul Pathak",
        "email": "rahul.pathak@oss.qualcomm.com",
        "time": "Thu Aug 20 17:43:09 2026 +0530"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Sat Sep 05 12:53:15 2026 +0530"
      },
      "message": "lib: sbi_domain: Rename per-domain data to per-domain state\n\nThe per-domain sbi_domain_data hold each domain\nassociated state like hart context in sbi_domain_context,\nmpxy state in sbi_mpxy, and others like each domain backed\nby the corresponding MPT (SMMPT).\nDATA reads as a generic name, while every use stores state.\nRename functions and macros appropriately. There are no\nfunctional changes.\n\nSigned-off-by: Rahul Pathak \u003crahul.pathak@oss.qualcomm.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260820121309.2551296-3-rahul.pathak@oss.qualcomm.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "35af7c8c7ee00869afc6e9e0aca64086888bd03f",
      "tree": "2fdcb7c940c5e32308a941dea354f2a23c7606ba",
      "parents": [
        "7774b82e4aaa6fa98d1430b02d614eadd339a8a2"
      ],
      "author": {
        "name": "Rahul Pathak",
        "email": "rahul.pathak@oss.qualcomm.com",
        "time": "Thu Aug 20 17:43:08 2026 +0530"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Sat Sep 05 12:52:39 2026 +0530"
      },
      "message": "lib: sbi: Rename map_range/unmap_range functions\n\n*_map_range and *_unmap_range functions are required\nin M-Mode to get the temporary access to S-Mode and U-Mode\nregions. Thse functions only operate using the TYPE_MEMORY\nmemory protection mechanisms. Rename them to reflect the\nactual usage of these functions and let generic map/unmap_range\nnames to implement generic functions\n\nSigned-off-by: Rahul Pathak \u003crahul.pathak@oss.qualcomm.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nReviewed-by: Pawandeep Oza \u003cpawandeep.oza@oss.qualcomm.com\u003e\nLink: https://lore.kernel.org/r/20260820121309.2551296-2-rahul.pathak@oss.qualcomm.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "7774b82e4aaa6fa98d1430b02d614eadd339a8a2",
      "tree": "2565865e9e46f23af0304b2b5b65a33a07e75eb9",
      "parents": [
        "bf459f85bf75d86cbf028e136067ce96aefde221"
      ],
      "author": {
        "name": "whensun",
        "email": "theodoruswensanfebruanto025@gmail.com",
        "time": "Fri Aug 28 14:31:04 2026 -0700"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Sat Sep 05 12:33:48 2026 +0530"
      },
      "message": "lib: sbi: Fix FP state enable in sbi_fp_save\n\nsbi_fp_save() currently sets mstatus.VS before executing floating-point\nstore instructions. VS controls vector state and does not enable\nfloating-point instructions.\n\nSet mstatus.FS instead, matching sbi_fp_restore(), so floating-point\nstate can be saved when FS was previously Off.\n\nFixes: 718e1d194f5a (\"lib: sbi: Add floating-point context save/restore support.\")\nSigned-off-by: whensun \u003ctheodoruswensanfebruanto025@gmail.com\u003e\nCo-developed-by: Huamao Wu \u003chuamao.wu@spacemit.com\u003e\nSigned-off-by: Huamao Wu \u003chuamao.wu@spacemit.com\u003e\nLink: https://lore.kernel.org/r/20260828213104.84490-1-theodoruswensanfebruanto025@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "bf459f85bf75d86cbf028e136067ce96aefde221",
      "tree": "ded26a57c654598ef952495ffc7e94cf9b1a267a",
      "parents": [
        "7384968e020f3bf9d524a00c6dc5638f3383aa92"
      ],
      "author": {
        "name": "Heinrich Schuchardt",
        "email": "heinrich.schuchardt@canonical.com",
        "time": "Thu Aug 20 15:44:30 2026 +0200"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Fri Sep 04 13:55:01 2026 +0530"
      },
      "message": "lib: sbi_irqchip: error handling in sbi_irqchip_raw_handler_default()\n\nsbi_irqchip_find_handler() may return NULL.\nCheck the return value before dereferencing.\n\nAddresses-Coverity-ID: 1677408 Dereference null return value\nSigned-off-by: Heinrich Schuchardt \u003cheinrich.schuchardt@canonical.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260820134431.167336-1-heinrich.schuchardt@canonical.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "7384968e020f3bf9d524a00c6dc5638f3383aa92",
      "tree": "f3eefa7aa47e36dba313c763bfae695ce5ce250d",
      "parents": [
        "c5077d497d0f7c7c9a74218ef2c448bc1a6d8a31"
      ],
      "author": {
        "name": "Ben Zong-You Xie",
        "email": "ben717@andestech.com",
        "time": "Wed Aug 05 17:33:54 2026 +0800"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Fri Sep 04 13:44:57 2026 +0530"
      },
      "message": "lib: sbi: respect scounteren when emulating timeh\n\nbeef2f693785 (\"lib: sbi: Respect scounteren when emulating the time\nCSR\") restored it for TIME only, so on RV32 supervisor software still\ncannot restrict U-mode or VS-mode access to time: rdtime traps as\nintended while rdtimeh keeps returning the upper half.\n\nSigned-off-by: Ben Zong-You Xie \u003cben717@andestech.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260805093354.1022980-1-ben717@andestech.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "c5077d497d0f7c7c9a74218ef2c448bc1a6d8a31",
      "tree": "d75e27ed638ed00a654dda62b3dbd78cbe908819",
      "parents": [
        "d34a39df779aa55a2dd929717a32e48c3a80df66"
      ],
      "author": {
        "name": "David E. Garcia Porras",
        "email": "david.garcia@aheadcomputing.com",
        "time": "Tue Aug 18 15:00:23 2026 -0600"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Fri Sep 04 11:43:53 2026 +0530"
      },
      "message": "lib: sbi_pmu: Fix counter and event info error codes as per SBI v3.0 spec\n\nAlign the PMU extension implementation with the error codes required\nby the SBI v3.0 specification, chapter 11:\n\n - sbi_pmu_counter_start and sbi_pmu_counter_stop (secs 11.9-11.10,\n   tables 39-42): the start_flags/stop_flags bits 2:(XLEN-1) are\n   reserved and must be zero, so return SBI_ERR_INVALID_PARAM when any\n   reserved flag bit is set. Introduce SBI_PMU_START_FLAGS_MASK and\n   SBI_PMU_STOP_FLAGS_MASK for the valid bits of each function.\n\n - sbi_pmu_counter_start and sbi_pmu_counter_stop (tables 40 and 42):\n   return SBI_ERR_ALREADY_STARTED / SBI_ERR_ALREADY_STOPPED when the\n   set of counters includes a counter which is already started or\n   stopped, instead of ignoring the error returned for each counter.\n\n - sbi_pmu_event_get_info (sec 11.14, table 47): the output word must\n   indicate whether the event is supported, but firmware events were\n   only matched against the hardware event map and were always\n   reported as unsupported. Report a validated firmware event as\n   supported.\n\nSigned-off-by: David E. Garcia Porras \u003cdavid.garcia@aheadcomputing.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260818210023.466462-4-david.garcia@aheadcomputing.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "d34a39df779aa55a2dd929717a32e48c3a80df66",
      "tree": "b28cd12ecca61fe406b5606264ef5f80dfb1fe48",
      "parents": [
        "d97cfb33ae1ac0d9eebce1a03c96497a1f8f4c67"
      ],
      "author": {
        "name": "David E. Garcia Porras",
        "email": "david.garcia@aheadcomputing.com",
        "time": "Tue Aug 18 15:00:22 2026 -0600"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Thu Sep 03 13:37:34 2026 +0530"
      },
      "message": "lib: sbi_pmu: Match raw event selector only against raw event map entries\n\nsbi_pmu_event_get_info() walks hw_event_map[] to decide whether a\nrequested event is supported. For SBI_PMU_EVENT_RAW_IDX and\nSBI_PMU_EVENT_RAW_V2_IDX it compares the requested event_data against\ntemp-\u003eselect / temp-\u003eselect_mask without first checking that the map\nentry being examined is itself a raw event entry.\n\nNon-raw hardware event entries are added via sbi_pmu_add_hw_event_counter_map(),\nleave select and select_mask at zero, hence they satisfy:\n\n\ttemp-\u003eselect \u003d\u003d (event_data \u0026 temp-\u003eselect_mask)\n\nso the first non-raw entry visited will always match.\nThe issue\u0027s observability depends purely on the ordering of hw_event_map[]:\nif the platform registers its raw events last, every raw event query, including\nunsupported ones, will be reported as supported.\n\nFix it by checking event_idx against temp-\u003estart_idx and temp-\u003eend_idx before\ncomparing select/select_mask.\n\nFixes: e4345842168b (\"lib: sbi_pmu: Implement SBI PMU event info function\")\nSigned-off-by: David E. Garcia Porras \u003cdavid.garcia@aheadcomputing.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260818210023.466462-3-david.garcia@aheadcomputing.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "d97cfb33ae1ac0d9eebce1a03c96497a1f8f4c67",
      "tree": "7671e94c1364a6e3bb578a8380580d9b6a5377c8",
      "parents": [
        "35511bc6ee1c9c17b6a89b44c52e2044bb51b979"
      ],
      "author": {
        "name": "David E. Garcia Porras",
        "email": "david.garcia@aheadcomputing.com",
        "time": "Tue Aug 18 15:00:21 2026 -0600"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Thu Sep 03 13:27:45 2026 +0530"
      },
      "message": "lib: sbi_pmu: Return invalid param error for reserved event_idx bits\n\nAs per section 11.14 of the SBI specification (Function: Get PMU Event\nInfo, FID #8), Table 47, the event_idx word of an event info entry\nonly uses BIT[0:19]; BIT[20:31] are reserved for the future purpose\nand must be zero. Table 48 further requires the SBI implementation to\nreturn SBI_ERR_INVALID_PARAM if any reserved bit in an event_idx word\nis set.\n\nsbi_pmu_event_get_info() does not check the reserved bits, so a\nmalformed event_idx is silently passed on to pmu_event_validate()\ninstead of failing the call. Add SBI_PMU_EVENT_IDX_MBZ_MASK covering\nthe must-be-zero bits and return SBI_ERR_INVALID_PARAM when any of\nthem are set.\n\nFixes: e4345842168b (\"lib: sbi_pmu: Implement SBI PMU event info function\")\nSigned-off-by: David E. Garcia Porras \u003cdavid.garcia@aheadcomputing.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260818210023.466462-2-david.garcia@aheadcomputing.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "35511bc6ee1c9c17b6a89b44c52e2044bb51b979",
      "tree": "600996ef9f8d2f2a48ddec5136ce05ace05310b6",
      "parents": [
        "22a5f98b3965adb2c5b2224da7ec4590807a618a"
      ],
      "author": {
        "name": "Zhanpeng Zhang",
        "email": "zhangzhanpeng.jasper@bytedance.com",
        "time": "Fri Jul 31 15:18:20 2026 +0800"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Tue Sep 01 12:57:33 2026 +0530"
      },
      "message": "lib: sbi: sse: clear SPV for non-virtualized events\n\nSSE injection sets hstatus.SPV to the virtualization state of the\ninterrupted context before entering the supervisor handler. This patch\ncompletes that existing requirement and does not change the SSE ABI.\n\nThe existing code sets SPV when the interrupted context is virtualized,\nbut leaves it unchanged otherwise. A stale SPV value can therefore make\nan event that interrupted host execution appear to have interrupted a\nguest. Event completion can then resume with virtualization enabled.\n\nClear SPV when the interrupted context is not virtualized so the\nhandler-visible state matches the interrupted context. This also lets\nsupervisor software, such as the Linux PMU and perf code, reliably tell\nwhether an SSE interrupted host or virtualized execution.\n\nFixes: c8cdf01d8f3a (\"lib: sbi: Add support for Supervisor Software Events extension\")\nSigned-off-by: Zhanpeng Zhang \u003czhangzhanpeng.jasper@bytedance.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260731071820.7318-1-zhangzhanpeng.jasper@bytedance.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "22a5f98b3965adb2c5b2224da7ec4590807a618a",
      "tree": "cd3f9b5ac1e2d0abb8cc1843e6603b5004b9e79b",
      "parents": [
        "06af8bd61b37bbf0db93adfa9902b176d0158ea1"
      ],
      "author": {
        "name": "Randolph",
        "email": "randolph@andestech.com",
        "time": "Wed Jul 29 17:23:17 2026 +0800"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Tue Sep 01 10:59:55 2026 +0530"
      },
      "message": "platform: generic/andes: fix 32-bit shift overflow in decode_pmaaddrx()\n\ndecode_pmaaddrx() reconstructs the NAPOT region start and size from a\npmaaddr CSR value using \"1 \u003c\u003c (k + 3)\" and \"1 \u003c\u003c k\". The integer\nliteral 1 has type int, so these shifts are performed in 32-bit\nprecision. Shifting a 32-bit value by 32 or more bits is undefined\nbehavior, and on RV64 the compiler emits sllw, which truncates the\nshift amount modulo 32.\n\nAs a result, any PMA region with size \u003e\u003d 4 GiB (k \u003e\u003d 29) is decoded\nincorrectly. For example, on the Andes QiLai SoC the PCIe region\n0x1000000000 - 0x17ffffffff (pmaaddr \u003d 0x4ffffffff, k \u003d 32) is decoded\nas an 8-byte region at 0x13fffffffc.\n\nThis is not merely cosmetic: decode_pmaaddrx() is used by\nhas_pma_region_overlap() and andes_sbi_free_pma(), so overlap checks\nare performed against bogus ranges and freeing such an entry by its\nphysical address always fails.\n\nPromote the shifts to unsigned long so they are performed in the\nnative register width.\n\nFixes: aa56084c4dfb (\"platform: generic: andes: add a new Andes SBI call to set up a PMA entry\")\nSigned-off-by: Randolph Lin \u003crandolph@andestech.com\u003e\nTested-by: Benoît Monin \u003cbenoit.monin@bootlin.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260729092317.2848665-1-randolph@andestech.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "06af8bd61b37bbf0db93adfa9902b176d0158ea1",
      "tree": "fad2b34e37d574f5da78b7ba7967ccaf0cc6028c",
      "parents": [
        "0fb479930775cf86efdfd5c50ad21b4cf8353c76"
      ],
      "author": {
        "name": "Ben Zong-You Xie",
        "email": "ben717@andestech.com",
        "time": "Tue Jul 28 16:10:41 2026 +0800"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Tue Sep 01 10:46:23 2026 +0530"
      },
      "message": "lib: utils/andes: arm the SMU sleep command last\n\nOnce the sleep command is written, the next WFI puts the core to sleep,\nso everything that can fail has to run before it.\n\nAlso, use writel() so the command store cannot still be in flight at the\nWFI.\n\nSigned-off-by: Ben Zong-You Xie \u003cben717@andestech.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260728081041.2724668-8-ben717@andestech.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "0fb479930775cf86efdfd5c50ad21b4cf8353c76",
      "tree": "ef74e3e5b5b8b6b24aa23fe7aa156d21731512f4",
      "parents": [
        "e01bd7926b5d808d883df18c6ac919dfa9ab04e7"
      ],
      "author": {
        "name": "Ben Zong-You Xie",
        "email": "ben717@andestech.com",
        "time": "Tue Jul 28 16:10:40 2026 +0800"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Tue Sep 01 10:46:23 2026 +0530"
      },
      "message": "lib: utils/suspend: wait for the secondary Andes harts to sleep\n\nA secondary hart may not have reached the target sleep state by the time\nthe primary checks, so the one-shot check could fail spuriously. Poll\neach PCS status instead and give up after HART_SLEEP_TIMEOUT_MS.\n\nRework atcsmu_pcs_is_sleep() into the atcsmu_hart_is_sleep() predicate so\nsbi_timer_waitms_until() can drive it.\n\nSigned-off-by: Ben Zong-You Xie \u003cben717@andestech.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260728081041.2724668-7-ben717@andestech.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "e01bd7926b5d808d883df18c6ac919dfa9ab04e7",
      "tree": "d2bf5a2065157bd80db3b50a287766f4cb5bb981",
      "parents": [
        "9ceaf4738a62ec990dcd0e797243bb6e78508291"
      ],
      "author": {
        "name": "Ben Zong-You Xie",
        "email": "ben717@andestech.com",
        "time": "Tue Jul 28 16:10:39 2026 +0800"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Tue Sep 01 10:46:23 2026 +0530"
      },
      "message": "lib: utils/hsm: wake sleeping Andes harts with an IPI\n\nAn IPI wakes a hart from light sleep and, once MSIP is kept as an SMU\nwakeup event, from deep sleep too. Drop the WAKEUP_CMD path and the boot\nstate, hart type and sleep type conditions that selected it.\n\nSigned-off-by: Ben Zong-You Xie \u003cben717@andestech.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260728081041.2724668-6-ben717@andestech.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "9ceaf4738a62ec990dcd0e797243bb6e78508291",
      "tree": "7b5b93a0cee55a52a6517db64688e68fc23a53bf",
      "parents": [
        "733867c0c8f4941a78144eb49d8e0c5d086d8567"
      ],
      "author": {
        "name": "Ben Zong-You Xie",
        "email": "ben717@andestech.com",
        "time": "Tue Jul 28 16:10:38 2026 +0800"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Tue Sep 01 10:46:23 2026 +0530"
      },
      "message": "lib: utils/suspend: restore Andes CSRs on system-suspend resume\n\nA deep sleep resume leaves the boot hart in SUSPENDED state, so\ninit_warmboot() dispatches to init_warm_resume(), which calls the SUSP\ndevice\u0027s system_resume callback and not sbi_platform_early_init(). The\nrestore in ae350_early_init() is dead code on that path.\n\nFixes: b27ecec76b8a (\"lib: utils/suspend: add Andes ATCSMU suspend driver\")\nSigned-off-by: Ben Zong-You Xie \u003cben717@andestech.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260728081041.2724668-5-ben717@andestech.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "733867c0c8f4941a78144eb49d8e0c5d086d8567",
      "tree": "dbbfcb55e46817c7cf850e46c4d94b9b63009c74",
      "parents": [
        "4c5d9d9f93c77805dbc023cdef6a68c039b87970"
      ],
      "author": {
        "name": "Ben Zong-You Xie",
        "email": "ben717@andestech.com",
        "time": "Tue Jul 28 16:10:37 2026 +0800"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Tue Sep 01 10:46:23 2026 +0530"
      },
      "message": "platform: generic/andes: pair non-retentive CSR save/restore with a flag\n\nThe restore was gated on sbi_init_count() and the current sleep type,\nwhich is only a proxy for \"did this hart actually save its CSRs\". Track\nit explicitly instead, so restore is self-guarding and the call site\nneeds no conditions.\n\nSigned-off-by: Ben Zong-You Xie \u003cben717@andestech.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260728081041.2724668-4-ben717@andestech.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "4c5d9d9f93c77805dbc023cdef6a68c039b87970",
      "tree": "28cb04ee5c14d82ab5a65903c0831cb09e6edbc9",
      "parents": [
        "344428d8322e03cc5059a25769694f7622fb9333"
      ],
      "author": {
        "name": "Ben Zong-You Xie",
        "email": "ben717@andestech.com",
        "time": "Tue Jul 28 16:10:36 2026 +0800"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Tue Sep 01 10:46:23 2026 +0530"
      },
      "message": "lib: utils/andes: add the SMU and LLC registers to the root domain\n\nWhen a hart implements Smepmp, S-mode and U-mode cannot reach either\ndevice unless its region is shared explicitly. Suspend and resume then\ntake an access violation as soon as the supervisor touches the SMU or\nthe LLC.\n\nShare both regions, and take their sizes from the DT nodes rather than\nignoring them.\n\nSigned-off-by: Ben Zong-You Xie \u003cben717@andestech.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260728081041.2724668-3-ben717@andestech.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "344428d8322e03cc5059a25769694f7622fb9333",
      "tree": "1112095bba73b8c1034beb2409eb05a4501725dd",
      "parents": [
        "27d954538410a7fbecf0ed68ce319706d3e3b142"
      ],
      "author": {
        "name": "Ben Zong-You Xie",
        "email": "ben717@andestech.com",
        "time": "Tue Jul 28 16:10:35 2026 +0800"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Tue Sep 01 10:46:23 2026 +0530"
      },
      "message": "lib: utils/cache: fix andes_llcache_enable() return value\n\nThe driver returned 1 on success where callers expect 0, so the suspend\nto RAM aborted right after disabling the LLC and before flushing it.\n\nFixes: 82b0961821e4 (\"lib: utils/cache: add Andes last level cache controller\")\nSigned-off-by: Ben Zong-You Xie \u003cben717@andestech.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260728081041.2724668-2-ben717@andestech.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "27d954538410a7fbecf0ed68ce319706d3e3b142",
      "tree": "32458e6ec00d4d90a0ca16410e14a4c533ce92c1",
      "parents": [
        "548518e675d0ed387ff3ee2d9a2226a6e93090a4"
      ],
      "author": {
        "name": "David E. Garcia Porras",
        "email": "david.garcia@aheadcomputing.com",
        "time": "Mon Jul 27 12:30:39 2026 -0600"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Tue Sep 01 10:36:07 2026 +0530"
      },
      "message": "lib: sbi: dbtr: add platform device for per-slot trigger capabilities\n\nThe tinfo probe only discovers trigger types; WARL fields tied off\ninside an implemented trigger CSR (e.g. an mcontrol6 trigger without\nload/store address match) are not discoverable. sbi_alloc_trigger()\nalso selects the first free trigger slot irrespective of the requested\ntrigger type or configuration. Unsupported configurations are thus\nsilently dropped by the hardware while SBI reports success, where SBI\nv3.0 sections 19.4 / 19.5 require SBI_ERR_NOT_SUPPORTED.\n\nAdd an optional platform device, struct sbi_dbtr_device, with a\ntrigger_supported(idx, tdata1, tdata2, tdata3) callback reporting\nwhether the trigger slot selected by idx (the tselect value) supports\na given configuration. Use it, together with the probed per-slot\ntype_mask, to:\n\n  - allocate only trigger slots supporting the requested configuration\n    (new dbtr_find_free_slot() helper)\n  - reject unsupportable install/update requests with\n    SBI_ERR_NOT_SUPPORTED\n  - count only supporting slots in sbi_dbtr_num_trig() (SBI v3.0\n    section 19.1)\n\nThe install path now dry-runs the allocation of the whole batch before\nprogramming any trigger, so a partially-installed batch is never left\nbehind. Without a registered device, only the type matching applies (unchanged behavior).\n\nFixes: 97f234f15c96 (\"lib: sbi: Introduce the SBI debug triggers extension support\")\nSigned-off-by: David E. Garcia Porras \u003cdavid.garcia@aheadcomputing.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260727183041.258377-2-david.garcia@aheadcomputing.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "548518e675d0ed387ff3ee2d9a2226a6e93090a4",
      "tree": "b457eff4d7f786d68610ef9edb101c7b870e8c82",
      "parents": [
        "3e21ee3496684e27f1884e8947a1faf6e3195e09"
      ],
      "author": {
        "name": "David E. Garcia Porras",
        "email": "david.garcia@aheadcomputing.com",
        "time": "Mon Jul 27 12:30:41 2026 -0600"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Mon Aug 31 21:12:18 2026 +0530"
      },
      "message": "lib: sbi: dbtr: return SBI_ERR_INVALID_PARAM for invalid trigger configuration\n\nsbi_dbtr_install_trig() returns the generic SBI_ERR_FAILED when\ndbtr_trigger_valid() rejects a trigger configuration (dmode or M-mode\nbits set), where SBI v3.0 section 19.4 defines SBI_ERR_INVALID_PARAM\nfor an invalid trigger configuration entry. Return\nSBI_ERR_INVALID_PARAM instead.\n\nFixes: 97f234f15c96 (\"lib: sbi: Introduce the SBI debug triggers extension support\")\nSigned-off-by: David E. Garcia Porras \u003cdavid.garcia@aheadcomputing.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260727183041.258377-4-david.garcia@aheadcomputing.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "3e21ee3496684e27f1884e8947a1faf6e3195e09",
      "tree": "3f5bc73ff475bc014ddafb2251d4170cbb48265b",
      "parents": [
        "f95648d3955d72f77e13315a990a6135303978a5"
      ],
      "author": {
        "name": "David E. Garcia Porras",
        "email": "david.garcia@aheadcomputing.com",
        "time": "Mon Jul 27 12:30:40 2026 -0600"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Mon Aug 31 21:12:18 2026 +0530"
      },
      "message": "lib: sbi: dbtr: return SBI_ERR_BAD_RANGE for invalid trigger index range\n\nsbi_dbtr_read_trig() returns SBI_ERR_INVALID_PARAM for an out-of-range\ntrigger index range, where SBI v3.0 section 19.3 defines\nSBI_ERR_BAD_RANGE. Return SBI_ERR_BAD_RANGE, matching the equivalent\nrange check in sbi_dbtr_update_trig().\n\nFixes: 324021423d06 (\"lib: sbi: dbtr: Fix update_triggers to match SBI\")\nSigned-off-by: David E. Garcia Porras \u003cdavid.garcia@aheadcomputing.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260727183041.258377-3-david.garcia@aheadcomputing.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "f95648d3955d72f77e13315a990a6135303978a5",
      "tree": "a492dcfae7d4ada5abdb8e6736a9dd30e29bf17a",
      "parents": [
        "4e79fd7de59f1b2899092c1a84ce68c8ebc68f93"
      ],
      "author": {
        "name": "Yudistira Putra",
        "email": "pyudistira519@gmail.com",
        "time": "Sun Jul 19 06:11:25 2026 -0400"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Mon Aug 31 18:08:50 2026 +0530"
      },
      "message": "lib: sbi: clamp sbi_ecall_get_extensions_str buffer offset\n\nsbi_ecall_get_extensions_str() advanced offset by the nominal extension\nname length without checking remaining capacity. When the caller buffer\nwas smaller than the concatenated extension list, offset could pass\nexts_str_size, so (exts_str_size - offset) became negative and was\npassed to sbi_snprintf() as a large u32, and the trailing NUL write\ncould step past the caller buffer.\n\nThe helper can write beyond a caller-provided destination when the\nregistered extension list exceeds the supplied capacity.\n\nMirror the guard already used by sbi_hart_get_extensions_str(): stop\nappending when the next name would not fit. Add an SBIUNIT regression\nthat registers several extensions into a 16-byte buffer with a redzone\nand verifies no out-of-bounds write.\n\nCloses: https://github.com/riscv-software-src/opensbi/issues/416\nSigned-off-by: Yudistira Putra \u003cpyudistira519@gmail.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260719101125.190314-1-pyudistira519@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "4e79fd7de59f1b2899092c1a84ce68c8ebc68f93",
      "tree": "8a5ffddd23059dc564a43fe56f9471d2aef1490c",
      "parents": [
        "f465413402c53fd022d54c3e56a26f96cdc32ffc"
      ],
      "author": {
        "name": "Xiang W",
        "email": "wangxiang@iscas.ac.cn",
        "time": "Tue Jul 21 23:08:44 2026 +0800"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Thu Aug 20 20:46:25 2026 +0530"
      },
      "message": "platform: generic: sophgo: Move SG2042 timer memregion workaround\n\nCommit 4813a2042096 (\"lib: sbi_init: Call hart init and timer init\nbefore platform early init\") reordered the cold/warm boot sequence so\nthat sbi_timer_init() runs before sbi_platform_early_init().\n\nThe SG2042 platform workaround that merges the 16 separate timer\nregions into a single domain memregion was previously performed in\nearly_init().  After the reordering the MTIMER driver therefore adds\nthe individual regions first, defeating the purpose of the combined\nregion (and wasting PMP entries).\n\nMove the addition of the combined memregion into extensions_init(),\nwhich is invoked from sbi_hart_init() and consequently still executes\nbefore sbi_timer_init().\n\nSigned-off-by: Han Gao \u003cgaohan@iscas.ac.cn\u003e\nSigned-off-by: Xiang W \u003cwangxiang@iscas.ac.cn\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260721150844.741606-1-wangxiang@iscas.ac.cn\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "f465413402c53fd022d54c3e56a26f96cdc32ffc",
      "tree": "6969cccf1c3be9b33d4a3a3d11e0039c09fd055c",
      "parents": [
        "65be0e37c4fd226d8d0b033b04551ef3dc114037"
      ],
      "author": {
        "name": "Xiang W",
        "email": "wangxiang@iscas.ac.cn",
        "time": "Wed Jun 17 19:29:26 2026 +0800"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Thu Aug 20 20:30:48 2026 +0530"
      },
      "message": "lib: sbi: Allow platforms to override local TLB operations\n\nSome T-Head based processors, for example the Sophgo SG2044, have a\nJTLB errata that requires special handling of certain TLB maintenance\ninstructions (sfence.vma).\n\nIntroduce a sbi_tlb_local_operations structure so platforms can provide\ncustom local TLB flush implementations. Use it to implement the JTLB\nworkaround on affected SoCs (currently Sophgo SG2044).\n\nSigned-off-by: Xiang W \u003cwangxiang@iscas.ac.cn\u003e\nSigned-off-by: Han Gao \u003cgaohan@iscas.ac.cn\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260617112925.1144190-2-wangxiang@iscas.ac.cn\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "65be0e37c4fd226d8d0b033b04551ef3dc114037",
      "tree": "a55260d181c5b4ba14c1339f98d4e3f00ec9666f",
      "parents": [
        "bc7a78f41b3990546d7d238811b3ed3c3016e66f"
      ],
      "author": {
        "name": "Michael Ellerman",
        "email": "mpe@kernel.org",
        "time": "Mon Jun 22 23:26:00 2026 +1000"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Thu Aug 20 20:16:42 2026 +0530"
      },
      "message": "Makefile: Support make -s to silence the build\n\nAdd support for silencing the build with make -s, which means nothing is\nprinted unless there is an error. Similar to the way Linux does it.\n\nSigned-off-by: Michael Ellerman \u003cmpe@kernel.org\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260622-upstream-v1-1-792e1847e96e@kernel.org\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "bc7a78f41b3990546d7d238811b3ed3c3016e66f",
      "tree": "94963f7c909587a7d83681392b1f41266a79d400",
      "parents": [
        "6499c022541fd85e2eee8fc93951ebe212b494e8"
      ],
      "author": {
        "name": "Sunil V L",
        "email": "sunilvl@oss.qualcomm.com",
        "time": "Fri Jul 03 10:35:30 2026 +0530"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Thu Aug 20 17:57:47 2026 +0530"
      },
      "message": "lib/utils/cppc: Emulate REFERENCE_CTR and DELIVERED_CTR locally\n\nCurrently, REFERENCE_CTR and DELIVERED_CTR are always read by sending\nan RPMI message to the controller. This does not scale well since\nthese registers are typically polled frequently by CPU frequency\ngovernors.\n\nThe reference counter can be simply the TIME csr, so it can be read\nlocally without going over RPMI. If the fast channel is supported,\nthe delivered counter can also be emulated locally: the fast channel\nexposes the current frequency, which is multiplied by the elapsed\nTIME csr ticks since the last read to calculate the delivered\nperformance counter.\n\nSigned-off-by: Sunil V L \u003csunilvl@oss.qualcomm.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260703050530.2460037-3-sunilvl@oss.qualcomm.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "6499c022541fd85e2eee8fc93951ebe212b494e8",
      "tree": "f6016e59419514ce84855a5954384b3f4679b2a0",
      "parents": [
        "3876e7574dd7eff9241e6fb0a3822e4e51731b9f"
      ],
      "author": {
        "name": "Sunil V L",
        "email": "sunilvl@oss.qualcomm.com",
        "time": "Fri Jul 03 10:35:29 2026 +0530"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Thu Aug 20 17:57:45 2026 +0530"
      },
      "message": "lib/sbi/sbi_timer: Add sbi_timer_frequency() function\n\nDrivers and other library code sometimes need the timer frequency\noutside of lib/sbi/sbi_timer.c, but there is currently no accessor\nfor it. Add sbi_timer_frequency() to expose the frequency of the\nregistered timer device.\n\nSigned-off-by: Sunil V L \u003csunilvl@oss.qualcomm.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260703050530.2460037-2-sunilvl@oss.qualcomm.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "3876e7574dd7eff9241e6fb0a3822e4e51731b9f",
      "tree": "ba5eb0a32e609c7aa9130d0a9596422f1a0b587e",
      "parents": [
        "c5e90d4286211e5a68c157f435f12745705dfa51"
      ],
      "author": {
        "name": "Heinrich Schuchardt",
        "email": "heinrich.schuchardt@canonical.com",
        "time": "Mon Jul 06 18:05:44 2026 +0200"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Thu Aug 20 17:57:19 2026 +0530"
      },
      "message": "lib: make SBI_HARTMASK_MAX_BITS customizable, default 256\n\nLinux has raised the default number of 64bit CPUs to 256 in commit\n0fa749771993 (\"riscv: Raise default NR_CPUS for 64BIT to 256\").\nOpenSBI should use a matching value for SBI_HARTMASK_MAX_BITS.\n\nIntroduce configuration variable CONFIG_HARTMASK_MAX_BITS with\nrange 32-512 and default 256.\n\nSigned-off-by: Heinrich Schuchardt \u003cheinrich.schuchardt@canonical.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260706160544.697409-1-heinrich.schuchardt@canonical.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "c5e90d4286211e5a68c157f435f12745705dfa51",
      "tree": "eff0b65982e1ca003db453453c551af8290f17c8",
      "parents": [
        "eb6099e7aab05495bc7b73f61a5c8e89b24f6043"
      ],
      "author": {
        "name": "Ranbir Singh",
        "email": "ranbir.singh@oss.qualcomm.com",
        "time": "Fri Jul 24 13:39:20 2026 +0530"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Thu Aug 20 13:11:05 2026 +0530"
      },
      "message": "lib: utils/mpxy: logging: Use correct enum for enable notification\n\nLogging file should be using its own enum and not others.\n\nFixes: 262571217c75 (\"lib: utils/mpxy: Add RPMI MPXY driver for logging\nservice group\")\n\nSigned-off-by: Ranbir Singh \u003cranbir.singh@oss.qualcomm.com\u003e\nReviewed-by: Rahul Pathak \u003crahul.pathak@oss.qualcomm.com\u003e\nLink: https://lore.kernel.org/r/20260724080920.1629065-3-ranbir.singh@oss.qualcomm.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "eb6099e7aab05495bc7b73f61a5c8e89b24f6043",
      "tree": "c091339b71b370ca0b9ecccf359c87ecb5982e74",
      "parents": [
        "f1ab1b45efbdbb8e6ea20557b0e5139d78d15d54"
      ],
      "author": {
        "name": "Ranbir Singh",
        "email": "ranbir.singh@oss.qualcomm.com",
        "time": "Fri Jul 24 13:39:19 2026 +0530"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Thu Aug 20 13:11:05 2026 +0530"
      },
      "message": "lib: utils/mpxy: mm: Add missing RPMI_MM_SRV_ENABLE_NOTIFICATION entry\n\nAs RPMI_MM_SRV_ENABLE_NOTIFICATION is a mandatory service for RPMI MM\nservice group, it must be added in its MPXY client driver here.\n\nFixes: 0b041e58c078 (\"lib: utils: Add MPXY client driver for RPMI MM\nservice group\")\n\nSigned-off-by: Ranbir Singh \u003cranbir.singh@oss.qualcomm.com\u003e\nReviewed-by: Rahul Pathak \u003crahul.pathak@oss.qualcomm.com\u003e\nLink: https://lore.kernel.org/r/20260724080920.1629065-2-ranbir.singh@oss.qualcomm.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "f1ab1b45efbdbb8e6ea20557b0e5139d78d15d54",
      "tree": "5961454f8d87b5bba1744774c986018a2ee63070",
      "parents": [
        "50558221cf0cfc27cfc44fc18e481b61bf45dfde"
      ],
      "author": {
        "name": "Benoît Monin",
        "email": "benoit.monin@bootlin.com",
        "time": "Fri Jul 24 17:35:08 2026 +0200"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Thu Aug 20 13:02:17 2026 +0530"
      },
      "message": "lib: utils/reset: Add Metanoia MT2824 system reset driver\n\nAdd a driver for the system reset function found in Metanoia MT2824 SoC\nand enable it in the generic platform configuration.\n\nOriginally-by: Jun Chang \u003cjun.chang@metanoia-comm.com\u003e\nSigned-off-by: Benoît Monin \u003cbenoit.monin@bootlin.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260724-mt5824-v3-2-9a7d17d69974@bootlin.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "50558221cf0cfc27cfc44fc18e481b61bf45dfde",
      "tree": "9b2f1fde66066c5ed27e9770bf64a5315134b7b1",
      "parents": [
        "337c23dd66b821ac04a0f7bea313e7e7b30ecc49"
      ],
      "author": {
        "name": "Benoît Monin",
        "email": "benoit.monin@bootlin.com",
        "time": "Fri Jul 24 17:35:07 2026 +0200"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Thu Aug 20 12:12:38 2026 +0530"
      },
      "message": "platform: generic: Add Metanoia MT2824 initial support\n\nAdd the Metanoia MT2824 as a generic platform. This SoC contains two\nAX45MP cores from Andes Technology and uses Andes PMU and vendor-specific\nextensions.\n\nOriginally-by: Jun Chang \u003cjun.chang@metanoia-comm.com\u003e\nSigned-off-by: Benoît Monin \u003cbenoit.monin@bootlin.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260724-mt5824-v3-1-9a7d17d69974@bootlin.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "337c23dd66b821ac04a0f7bea313e7e7b30ecc49",
      "tree": "e4a839a849d9a650ef0d2393867981d04d8e4c5a",
      "parents": [
        "2636a9c2a827e07e29777bffd29ac05254f0bfeb"
      ],
      "author": {
        "name": "Anup Patel",
        "email": "anup.patel@oss.qualcomm.com",
        "time": "Mon Jul 27 09:55:04 2026 +0530"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Sun Aug 02 15:17:12 2026 +0530"
      },
      "message": "lib: sbi: Print list of all hart protection mechanisms at boot time\n\nInstead of printing only the best hart memory protection name at\nboot time let\u0027s print a list of all hart projection mechanisms.\n\nSigned-off-by: Anup Patel \u003canup.patel@oss.qualcomm.com\u003e\nReviewed-by: Rahul Pathak \u003crahul.pathak@oss.qualcomm.com\u003e\nLink: https://lore.kernel.org/r/20260727042504.1743250-6-anup.patel@oss.qualcomm.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "2636a9c2a827e07e29777bffd29ac05254f0bfeb",
      "tree": "a504b632c8c49f9cb64d2fa20d203a20748285b7",
      "parents": [
        "5053774c2f21029089a58ccadf21b5b3cb16c56e"
      ],
      "author": {
        "name": "Anup Patel",
        "email": "anup.patel@oss.qualcomm.com",
        "time": "Mon Jul 27 09:55:03 2026 +0530"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Sun Aug 02 15:17:12 2026 +0530"
      },
      "message": "lib: sbi: Extend hart protection abstraction to allow ID configuration\n\nThere are various ID configuration ISA extensions such as RISC-V Worlds,\nSupervisor domain ID, QoS ID, etc which need to be re-configured upon\ndomain context switch on a hart. Extend hart protection abstraction to\nsupport both memory protection and ID configuration ISA extensions.\n\nSigned-off-by: Anup Patel \u003canup.patel@oss.qualcomm.com\u003e\nReviewed-by: Rahul Pathak \u003crahul.pathak@oss.qualcomm.com\u003e\nLink: https://lore.kernel.org/r/20260727042504.1743250-5-anup.patel@oss.qualcomm.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "5053774c2f21029089a58ccadf21b5b3cb16c56e",
      "tree": "ba349983c420a2f7d97c05cd4a6a2ef76e9de8e6",
      "parents": [
        "48725e95b4d02281bee268f6a9f8b5a77eb001fb"
      ],
      "author": {
        "name": "Anup Patel",
        "email": "anup.patel@oss.qualcomm.com",
        "time": "Mon Jul 27 09:55:02 2026 +0530"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Sun Aug 02 15:17:12 2026 +0530"
      },
      "message": "lib: sbi: Add domain parameter to hart protection (un)configure()\n\nThe switch_to_next_domain_context() already has the pointers to current\nand next domains so pass them as parameters to hart protection configure()\nand unconfigure() callbacks.\n\nSigned-off-by: Anup Patel \u003canup.patel@oss.qualcomm.com\u003e\nReviewed-by: Rahul Pathak \u003crahul.pathak@oss.qualcomm.com\u003e\nLink: https://lore.kernel.org/r/20260727042504.1743250-4-anup.patel@oss.qualcomm.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "48725e95b4d02281bee268f6a9f8b5a77eb001fb",
      "tree": "3860b375d87296b496df90c29fa9c34ca2009d9c",
      "parents": [
        "622f3d44313c1d537f2301488167b42ac0680633"
      ],
      "author": {
        "name": "Anup Patel",
        "email": "anup.patel@oss.qualcomm.com",
        "time": "Mon Jul 27 09:55:01 2026 +0530"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Sun Aug 02 15:17:12 2026 +0530"
      },
      "message": "lib: sbi: Introduce sbi_hart_protection_reconfigure() function\n\nInstead of assuming that hart protection reconfiguration is unconfigure\nfollowed by configure, introduce a separate sbi_hart_protection_reconfigure()\nfunction and use it in switch_to_next_domain_context().\n\nSigned-off-by: Anup Patel \u003canup.patel@oss.qualcomm.com\u003e\nReviewed-by: Rahul Pathak \u003crahul.pathak@oss.qualcomm.com\u003e\nLink: https://lore.kernel.org/r/20260727042504.1743250-3-anup.patel@oss.qualcomm.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "622f3d44313c1d537f2301488167b42ac0680633",
      "tree": "883baa0b0755b08d59fb0f78741183678acc0b03",
      "parents": [
        "c0f87f10d1bfb9e72a84ddfafb5604ee1bfe9d04"
      ],
      "author": {
        "name": "Anup Patel",
        "email": "anup.patel@oss.qualcomm.com",
        "time": "Mon Jul 27 09:55:00 2026 +0530"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Sun Aug 02 15:17:12 2026 +0530"
      },
      "message": "lib: sbi: Fix typos related to hart protection\n\nThe hart protection is incorrectly mentioned as hart isolation\nat few places in sbi_init.c hence fix these typos.\n\nSigned-off-by: Anup Patel \u003canup.patel@oss.qualcomm.com\u003e\nReviewed-by: Rahul Pathak \u003crahul.pathak@oss.qualcomm.com\u003e\nLink: https://lore.kernel.org/r/20260727042504.1743250-2-anup.patel@oss.qualcomm.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "c0f87f10d1bfb9e72a84ddfafb5604ee1bfe9d04",
      "tree": "7f58f2b8165d8a8c77c374dfca189117bd6d9614",
      "parents": [
        "7a90f41fb966e48a160637e841b91864e7a2c81c"
      ],
      "author": {
        "name": "Oza Pawandeep",
        "email": "pawandeep.oza@oss.qualcomm.com",
        "time": "Tue Jul 21 14:57:58 2026 -0700"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Jul 22 13:58:30 2026 +0530"
      },
      "message": "lib: utils/mpxy: bind channel access to owning domain\n\nThe MPXY framework currently stores registered channels in a global\nlist and exposes them to all callers. However, the intended model is\nthat each channel is assigned to a single supervisor domain at boot and\nremains owned by that domain for its lifetime.\n\nIntroduce fixed owner-domain tracking in struct sbi_mpxy_channel and\nmake channel lookup and enumeration domain-aware. A channel is now\nvisible only when accessed from its owning domain. Also require the\nowner to be set before channel registration.\n\nThis allows MPXY to support systems where the same hart may be reused by\nmultiple domains while keeping channel ownership and visibility fixed to\nthe domain that owns the service.\n\nSigned-off-by: Oza Pawandeep \u003cpawandeep.oza@oss.qualcomm.com\u003e\nReviewed-by: Rahul Pathak \u003crahul.pathak@oss.qualcomm.com\u003e\nLink: https://lore.kernel.org/r/20260721215758.695736-1-pawandeep.oza@oss.qualcomm.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "7a90f41fb966e48a160637e841b91864e7a2c81c",
      "tree": "b0ed33368736b6497069b7dc33a3ac22e0e4a7be",
      "parents": [
        "b2ead3b037c0b864b0770891586d121c0dc68761"
      ],
      "author": {
        "name": "Pawandeep Oza",
        "email": "pawandeep.oza@oss.qualcomm.com",
        "time": "Tue Jul 21 14:48:30 2026 -0700"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Jul 22 13:20:22 2026 +0530"
      },
      "message": "lib: sbi_irqchip: fix MSI EIID gap and tail allocation logic in register_msi\n\nFix sbi_irqchip_register_msi() that prevented contiguous EIID allocation:\n\n- Gap check condition was inverted: the original `h-\u003efirst_hwirq -\n  hwirq \u003c num_hwirq` incorrectly set `found \u003d true` when the gap was\n  too small to fit num_hwirq entries. Correct to `\u003e\u003d num_hwirq` so\n  allocation only proceeds when sufficient space exists between\n  registered handlers.\n\n- Tail-space check `!found \u0026\u0026 !hwirq` never triggered after iteration\n  Replace with `(chip-\u003enum_hwirq - hwirq) \u003e\u003d num_hwirq` to correctly\n  allocate if handler list is empty.\n\nFixes: 79e63bc8348d (\"irqchip: add sbi_irqchip_register_msi support\")\n\nSigned-off-by: Oza Pawandeep \u003cpawandeep.oza@oss.qualcomm.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260721214833.687361-7-pawandeep.oza@oss.qualcomm.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "b2ead3b037c0b864b0770891586d121c0dc68761",
      "tree": "e571489573e30dfa54271ea5273efc3f9d16f931",
      "parents": [
        "193b1d9e7ecb06c6a374665bcc177f6997a58505"
      ],
      "author": {
        "name": "Pawandeep Oza",
        "email": "pawandeep.oza@oss.qualcomm.com",
        "time": "Tue Jul 21 14:48:29 2026 -0700"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Jul 22 13:20:20 2026 +0530"
      },
      "message": "lib: utils/irqchip/imsic: track IRQ enable state and restore EIE on warm init\n\nAdd an irq_state field to struct sbi_irqchip_hwirq_data with a single\nIRQ_ENABLED flag (bit 0) to track whether a hardware interrupt has been\nenabled via the irqchip framework. Set IRQ_ENABLED in\nsbi_irqchip_unmask_hwirq() when the unmask callback is invoked.\n\nAdd sbi_irqchip_get_irq_state() as a private inline accessor and expose\nsbi_irqchip_is_irq_enabled() as a public API for drivers to query the\nenabled state of a hardware interrupt by chip pointer and hwirq number.\n\nRefactor imsic_local_eix_update() to operate on a single interrupt ID\ninstead of a base+count range, simplifying the CSR bit manipulation to\na direct BIT(id) write without the inner loop. Update all call sites\naccordingly.\n\nUse sbi_irqchip_is_irq_enabled() in imsic_warm_irqchip_init() to\nrestore per-EIID EIE CSR state on warm boot and HSM resume based on\nthe saved irq_state, replacing the previous blanket disable of all\ninterrupts. This ensures that EIIDs enabled during hotplug/warminit\ncycle are correctly re-enabled on the resuming hart without requiring\nsoftware to re-register or re-unmask each interrupt.\n\nSigned-off-by: Oza Pawandeep \u003cpawandeep.oza@oss.qualcomm.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260721214833.687361-6-pawandeep.oza@oss.qualcomm.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "193b1d9e7ecb06c6a374665bcc177f6997a58505",
      "tree": "94040c126609288663484d3c673ee9d6c1030279",
      "parents": [
        "22139b889dfeaa98d4b3f1abe4fd0b7c96f11b26"
      ],
      "author": {
        "name": "Pawandeep Oza",
        "email": "pawandeep.oza@oss.qualcomm.com",
        "time": "Tue Jul 21 14:48:28 2026 -0700"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Jul 22 13:20:14 2026 +0530"
      },
      "message": "lib: utils/irqchip/imsic:migrate to irqchip framework\n\nRewrite imsic_process_hwirqs() to use container_of() for instance\nlookup, validate ownership via imsic_current_hart_owner(), mask\nIMSIC_TOPEI_ID_MASK on the raw CSR value, and dispatch non-IPI EIIDs\nthrough sbi_irqchip_process_hwirq() instead of an open-coded switch.\n\nRewrite imsic_warm_irqchip_init() to resolve the imsic_data instance\nvia container_of() from the sbi_irqchip_device pointer, and skip\ninitialization silently for harts not owned by this instance.\n\nExtend the sbi_irqchip callback set:\n- hwirq_setup: registers sbi_irqchip_raw_handler_default for the EIID\n- hwirq_cleanup: clears EIE and EIP bits for the EIID on the owner hart\n- hwirq_eoi: no-op as claim/ack is already performed by CSR_MTOPEI\n  read in imsic_process_hwirqs()\n- hwirq_set_affinity: reprograms the upstream MSI source via\n  imsic_program_msi() to retarget the interrupt to the new hart\n- hwirq_mask: clears the EIE bit for the EIID on the owner hart\n- hwirq_unmask: sets the EIE bit for the EIID on the owner hart,\n  skipping IPI (EIID 1) and non-owner harts\n\nAdd imsic_compose_msi_msg() to compute the MMIO target address for a\ngiven hart and EIID by walking the imsic_regs array and applying file\nand guest index offsets.\n\nAdd imsic_program_msi() that composes an MSI message\nand delivers it to the upstream interrupt source via\nsbi_irqchip_write_msi(), enabling APLIC TARGET register programming\nfrom the IMSIC driver.\n\nMigrate imsic_cold_irqchip_init() to the instance-based model:\npreserve any pre-configured target_harts mask set by platform or FDT\ncode before applying the imsic_device template, falling back to\nsbi_hartmask_set_all() only when no mask was provided. Register the\nirqchip and reserved EIIDs on the embedded irqchip rather than the\nglobal singleton.\n\nSigned-off-by: Oza Pawandeep \u003cpawandeep.oza@oss.qualcomm.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260721214833.687361-5-pawandeep.oza@oss.qualcomm.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "22139b889dfeaa98d4b3f1abe4fd0b7c96f11b26",
      "tree": "2c2635f614b896c0153923971be375b7f4420721",
      "parents": [
        "759910326c3ac4cf9d42133fc1399d519037eae1"
      ],
      "author": {
        "name": "Pawandeep Oza",
        "email": "pawandeep.oza@oss.qualcomm.com",
        "time": "Tue Jul 21 14:48:27 2026 -0700"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Jul 22 13:08:15 2026 +0530"
      },
      "message": "lib: utils/irqchip/imsic: embed sbi_irqchip_device in imsic data\n\nEmbed sbi_irqchip_device in struct imsic_data to enable instance-based\nirqchip registration, replacing the previous global imsic_device\nsingleton. Include sbi_irqchip.h from imsic.h to make the embedded\nstruct visible to consumers.\n\nSigned-off-by: Oza Pawandeep \u003cpawandeep.oza@oss.qualcomm.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260721214833.687361-4-pawandeep.oza@oss.qualcomm.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "759910326c3ac4cf9d42133fc1399d519037eae1",
      "tree": "f691b7839b04c18e862895bf398e67dfeb8808e3",
      "parents": [
        "5ee055944559e0823c020222a05e5cac48e0c110"
      ],
      "author": {
        "name": "Pawandeep Oza",
        "email": "pawandeep.oza@oss.qualcomm.com",
        "time": "Tue Jul 21 14:48:26 2026 -0700"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Jul 22 13:08:13 2026 +0530"
      },
      "message": "lib: utils/irqchip/aplic: add MSI mode support with IMSIC parent linking\n\nAdd parent_unique_id fields to struct aplic_data for\nIMSIC parent linking in MSI mode. Store parent IMSIC unique_id\nduring FDT parsing of the APLIC node.\n\nAdd aplic_is_msi_mode() complementing aplic_is_direct_mode() to\nconsolidate delivery mode detection. Add APLIC_TARGET_EIID() macro\nfor packing the EIID field into the TARGET register.\n\nAdd parent_unique_id and parent_irq_map fields to struct aplic_data.\nparent_unique_id identifies the upstream IMSIC irqchip device resolved\nvia sbi_irqchip_find_device() during hwirq_setup. parent_irq_map is\na per-source array allocated at cold init time to track the EIID\nassigned by the IMSIC for each APLIC source.\n\nRestore aplic_writel_msicfg() and re-introduce MSI address register\nprogramming in aplic_init(), gated on aplic_is_msi_mode(). Set the\nDOMAINCFG_DM bit to switch the hardware to MSI delivery mode when\nno IDC structures are present.\n\nAdd aplic_program_msi_target() to pack hart_index, guest_index, and\nEIID into the APLIC_TARGET register. Add aplic_write_msi() as the\nsbi_irqchip write_msi callback that extracts EIID and hart_index from\nthe MSI message and calls aplic_program_msi_target(). Add\naplic_msi_callback() as the MSI receive callback that dispatches to\nsbi_irqchip_process_hwirq() on the APLIC chip.\n\nExtend aplic_hwirq_setup() with an MSI path that resolves the parent\nIMSIC chip by parent_unique_id, registers an MSI route via\nsbi_irqchip_register_msi(), and stores the allocated EIID in\nparent_irq_map for the source being configured.\n\nExtend aplic_hwirq_set_affinity() with an MSI path that delegates\naffinity reprogramming to the parent IMSIC chip via\nsbi_irqchip_set_affinity() using the stored parent_irq_map entry.\n\nGuard warm_init, process_hwirqs, and hwirq_eoi with early returns in\nMSI mode as interrupt delivery and acknowledgement are handled by the\nIMSIC in that configuration.\n\nSigned-off-by: Oza Pawandeep \u003cpawandeep.oza@oss.qualcomm.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260721214833.687361-3-pawandeep.oza@oss.qualcomm.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "5ee055944559e0823c020222a05e5cac48e0c110",
      "tree": "000e72f56be76ff2a691d095ee785a98cb294919",
      "parents": [
        "c504877ed00bd7c38168a6574db25562e3615a38"
      ],
      "author": {
        "name": "Pawandeep Oza",
        "email": "pawandeep.oza@oss.qualcomm.com",
        "time": "Tue Jul 21 14:48:25 2026 -0700"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Jul 22 13:07:40 2026 +0530"
      },
      "message": "lib: utils/irqchip/aplic: implement direct mode irqchip callbacks\n\nAdd inline register accessors (aplic_sourcecfg_write/read,\naplic_target_write, aplic_irq_setie/clrie/clrip, aplic_idc_base/\nread/write) to consolidate all MMIO access behind typed helpers\nand eliminate open-coded address arithmetic throughout the driver.\n\nAdd field-packing macros APLIC_TARGET_HART_IDX, APLIC_TARGET_IPRIO,\nand APLIC_TARGET_GUEST_IDX for constructing\nAPLIC_TARGET register values.\n\nAdd aplic_is_direct_mode() mode helper, aplic_hwirq_flags_to_sourcecfg()\nto map generic hwirq flags to APLIC source modes, aplic_hwirq_is_delegated()\nto detect S-mode delegated sources, aplic_find_idc_index() to map a\nhart index to its IDC slot, and aplic_first_target_hart() to select\nthe first available IDC hart for initial interrupt targeting.\n\nImplement the full sbi_irqchip callback set for direct (IDC) mode:\n- warm_init: enables IDC interrupt delivery and threshold per hart\n- process_hwirqs: reads TOPI and dispatches via sbi_irqchip_process_hwirq,\n  skipping delegated sources\n- hwirq_setup: programs sourcecfg from hwirq_flags and sets the TARGET\n  register to the first available IDC hart\n- hwirq_cleanup: clears IE, IP, sourcecfg, and TARGET on teardown\n- hwirq_eoi: reads CLAIMI to acknowledge the interrupt on the IDC\n- hwirq_set_affinity: reprograms TARGET to the specified hart index\n- hwirq_mask/unmask: clears/sets IE via CLRIENUM/SETIENUM\n\nRegister the irqchip device and populate target_harts from idc_map\nonly when targets_mmode is set in aplic_cold_irqchip_init().\n\nCo-developed-by: Raymond Mao \u003craymond.mao@riscstar.com\u003e\nSigned-off-by Raymond Mao \u003craymond.mao@riscstar.com\u003e\nSigned-off-by: Oza Pawandeep \u003cpawandeep.oza@oss.qualcomm.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260721214833.687361-2-pawandeep.oza@oss.qualcomm.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "c504877ed00bd7c38168a6574db25562e3615a38",
      "tree": "eb2838cdbf6958b214d1481dc7761f4e92efb50e",
      "parents": [
        "110f28518a15e84498bcc6337242160c500857fe"
      ],
      "author": {
        "name": "Anup Patel",
        "email": "anup.patel@oss.qualcomm.com",
        "time": "Sat Jun 13 18:51:28 2026 +0530"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Tue Jul 21 11:50:46 2026 +0530"
      },
      "message": "lib: sbi: Move trap delegation setup to sbi_trap.c\n\nCurrently, the trap delegation setup is part of sbi_hart.c and\ncalled via sbi_hart_init() whereas the per-hart trap (aka exception\nand interrupt) handling is part of sbi_trap.c.\n\nMove trap delegation init to sbi_trap.c and call it via new\nsbi_trap_init() from both cold boot and warm boot path. This\nway trap delegation setup is in same place as trap handling.\n\nSigned-off-by: Anup Patel \u003canup.patel@oss.qualcomm.com\u003e\nLink: https://lore.kernel.org/r/20260613132129.2135128-2-anup.patel@oss.qualcomm.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "110f28518a15e84498bcc6337242160c500857fe",
      "tree": "a249042fd19bccc4a8809fafa13a452ede5bceda",
      "parents": [
        "92bea7bdb2334a9274a824f9592026093febec7c"
      ],
      "author": {
        "name": "Chen Pei",
        "email": "cp0613@linux.alibaba.com",
        "time": "Fri Jun 12 19:52:07 2026 +0800"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Tue Jul 21 11:28:59 2026 +0530"
      },
      "message": "lib: sbi_ipi: Validate hartids against domain, not HSM state\n\nCommit 37eaca4ab33d (\"lib: sbi_ipi: Return error for invalid hartids\")\nadded a weight check against the HSM-interruptible mask, which excludes\nharts in STOPPED state. This causes sbi_ipi_send_many() to return\nSBI_EINVAL when the kernel sends a remote fence (e.g. TLB shootdown)\nwhose hmask includes a hart that was stopped by CPU hotplug:\n\n  __sbi_rfence_v02_call: hbase \u003d [0] hmask \u003d [0x3] failed (error [-22])\n\nValidate requested hartids against the domain\u0027s assigned hartmask\ninstead. This still rejects truly invalid or cross-domain hartids, but\nallows HSM-STOPPED harts to be silently skipped — which is safe because\na stopped hart will do a full local TLB flush on its next HSM start\nbefore re-entering S-mode.\n\nFixes: 37eaca4ab33d (\"lib: sbi_ipi: Return error for invalid hartids\")\nSigned-off-by: Chen Pei \u003ccp0613@linux.alibaba.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260612115208.118132-1-cp0613@linux.alibaba.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "92bea7bdb2334a9274a824f9592026093febec7c",
      "tree": "b6051ffbc59a222b58f046eef345ae0795cdbce5",
      "parents": [
        "262571217c75c649115633d8075cb6a40d940733"
      ],
      "author": {
        "name": "David E. Garcia Porras",
        "email": "david.garcia@aheadcomputing.com",
        "time": "Wed Jul 01 12:13:34 2026 -0600"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Thu Jul 16 11:27:57 2026 +0530"
      },
      "message": "lib: sbi_irqchip: fix device lookup by caps when first is NULL\n\nWhen called with first \u003d\u003d NULL, sbi_irqchip_find_device_by_caps() should\nscan the device list from the start. Instead it always returns NULL: \"found\"\nstarts false and only flips true when an entry equals \"first\", but no entry\never equals NULL, so every entry hits \"else continue\" and nothing is checked.\n\nThis breaks MSI detection in sbi_mpxy, which calls it with first \u003d\u003d NULL:\n\n  ms-\u003emsi_avail \u003d !!sbi_irqchip_find_device_by_caps(SBI_IRQCHIP_CAPS_MSI, NULL);\n\nmsi_avail is therefore always false, so mpxy_write_std_attr() silently drops\nthe MSI attributes (MSI_ADDR_LO/HI, MSI_DATA, MSI_CONTROL) while still\nreturning success, and MSI-based MPXY notifications are never delivered.\n\nInitialize \"found\" from \"first\" so a NULL \"first\" scans from the beginning,\nand always continue in the pre-match branch so a non-NULL \"first\" resumes\nafter the given device.\n\nFixes: 8570b938444d (\"lib: sbi_irqchip: Allow irqchip drivers advertise capabilities\")\nSigned-off-by: David E. Garcia Porras \u003cdavid.garcia@aheadcomputing.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260701181334.969877-1-david.garcia@aheadcomputing.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "262571217c75c649115633d8075cb6a40d940733",
      "tree": "dd71253fd64f277c4f08aa2027337633678cdc90",
      "parents": [
        "cbf9f6734dd85a982c63e3cb5db7ffe09da839ca"
      ],
      "author": {
        "name": "Subrahmanya Lingappa",
        "email": "subrahmanya.lingappa@oss.qualcomm.com",
        "time": "Fri Jun 19 16:04:24 2026 +0530"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Jul 01 21:30:24 2026 +0530"
      },
      "message": "lib: utils/mpxy: Add RPMI MPXY driver for logging service group\n\nAdd RPMI MPXY proxy driver for LOGGING service group so that\nS-mode can leverage LOGGING service group implemented by the\nplatform microcontroller.\n\nReviewed-by: Rahul Pathak \u003crahul.pathak@oss.qualcomm.com\u003e\nSigned-off-by: Subrahmanya Lingappa \u003csubrahmanya.lingappa@oss.qualcomm.com\u003e\nLink: https://lore.kernel.org/r/20260619103424.990954-1-subrahmanya.lingappa@oss.qualcomm.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "cbf9f6734dd85a982c63e3cb5db7ffe09da839ca",
      "tree": "d6a4747810533f9da9e17428cab821ccc8b1594c",
      "parents": [
        "bf10d3bf89c05cbc4e266de8c9aff7eb55aee748"
      ],
      "author": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Tue Jun 30 19:49:54 2026 +0530"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Tue Jun 30 19:49:54 2026 +0530"
      },
      "message": "include: Bump-up version to 1.9\n\nUpdate the OpenSBI version to 1.9 as part of release preparation.\n\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "bf10d3bf89c05cbc4e266de8c9aff7eb55aee748",
      "tree": "97758ae0bf360afb32de7254cf17e044d09eb233",
      "parents": [
        "0dfe6be08778102bf2fde61c5c99ea2cceb0b73a"
      ],
      "author": {
        "name": "Zong Li",
        "email": "zong.li@sifive.com",
        "time": "Thu Jun 25 19:21:21 2026 -0700"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Sun Jun 28 14:08:14 2026 +0530"
      },
      "message": "firmware: Fix comment after relocation completion\n\nAfter relocation is done, the code is running from the load address,\nnot the link address. Fix the comment to correctly reflect this.\n\nSigned-off-by: Zong Li \u003czong.li@sifive.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260626022121.1885209-1-zong.li@sifive.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "0dfe6be08778102bf2fde61c5c99ea2cceb0b73a",
      "tree": "7cc03b6013f97ba7d27eb8e5e750c4d0b14b4c44",
      "parents": [
        "cc9b4ef8f3b465f4625ca40bf6af5e033df2bd75"
      ],
      "author": {
        "name": "liutong",
        "email": "liutong@iscas.ac.cn",
        "time": "Sun Jun 28 13:55:09 2026 +0530"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Sun Jun 28 14:04:17 2026 +0530"
      },
      "message": "lib: sbi_pmu: fix integer overflow in pmu_ctr_idx_validate\n\npmu_ctr_idx_validate() checks whether counter indices are in range\nusing cbase + sbi_fls(cmask) \u003c total_ctrs.  Both operands are unsigned\nlong, so a crafted cbase close to ULONG_MAX causes the addition to wrap\naround to a small value that passes the comparison.\n\nOnce validation is bypassed, sbi_pmu_ctr_cfg_match() with the\nSKIP_MATCH flag uses the overflowed index directly as an array subscript\ninto phs-\u003eactive_events[], producing an out-of-bounds read in M-mode.\nThrough the firmware-event code path, the same overflowed index reaches\nfw_counters_data[] and fw_counters_started, giving an attacker OOB\nwrite-zero and OOB bit-set primitives in M-mode memory.\n\nFix pmu_ctr_idx_validate() by checking for unsigned overflow before the\ncomparison, and add a secondary bounds check on cidx_first in the\nSKIP_MATCH path so that even if validation is somehow bypassed in the\nfuture, the array access remains bounded.\n\nSigned-off-by: liutong \u003cliutong@iscas.ac.cn\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260624035049.1753003-1-liutong@iscas.ac.cn\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "cc9b4ef8f3b465f4625ca40bf6af5e033df2bd75",
      "tree": "60794d79e8f5d875fe0a236da6fa6dd692240264",
      "parents": [
        "a2077d44cc5743b2fae084dd3cac9225e81bc2fb"
      ],
      "author": {
        "name": "Junhui Liu",
        "email": "junhui.liu@pigmoral.tech",
        "time": "Tue Jun 23 10:49:25 2026 +0800"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Sun Jun 28 11:16:36 2026 +0530"
      },
      "message": "platform: generic: spacemit: k1: fix wrong address definitions\n\nPMU_AP_CORE2_IDLE_CFG and PMU_AP_CORE3_IDLE_CFG are not continuous with\nPMU_AP_CORE0_IDLE_CFG and PMU_AP_CORE1_IDLE_CFG. They are at PMU AP\nbase + 0x160 and + 0x164, matching the vendor OpenSBI definitions. After\nfixing these addresses, the intermediate cluster offset macros are\nredundant now, so define the wakeup and idle registers directly as\nPMU_AP_BASE offsets. This makes the actual register addresses easier to\ninspect and compare against the vendor code.\n\nC1_RVBADDR_HI_ADDR is also corrected according to the vendor OpenSBI\ndefinition. This was tested by writing an invalid value to the corrected\naddress, which prevents cluster1 CPUs from coming online, while doing\nthe same with the old address does not affect SMP boot.\n\nFixes: 1f84ec2a (\"platform: generic: spacemit: add K1\")\nSigned-off-by: Junhui Liu \u003cjunhui.liu@pigmoral.tech\u003e\nLink: https://lore.kernel.org/r/20260623-k1-fix-addr-v1-1-3dbde8c03bd6@pigmoral.tech\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "a2077d44cc5743b2fae084dd3cac9225e81bc2fb",
      "tree": "fbc23e1e54586e83cf34f9445f4493227fd9f439",
      "parents": [
        "a8be5e9478be08f09a720f7013ec30d6e2bd9fcc"
      ],
      "author": {
        "name": "Vladimir Kondratiev",
        "email": "vladimir.kondratiev@mobileye.com",
        "time": "Thu Jun 18 13:37:13 2026 +0300"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Sun Jun 28 11:08:13 2026 +0530"
      },
      "message": "platform: generic: mips eyeq7h: fix boot with JTAG\n\nWhen JTAG is connected, internal logic leads to the bit\nMIPS_CTL0_DBG_RST_DASRT (for the debug unit) stay high and\nthis prevents normal cluster power-up.\n\nForce proper power-on reset value prior to power-up sequence.\nHold this value for about 10 usec\n\nSigned-off-by: Vladimir Kondratiev \u003cvladimir.kondratiev@mobileye.com\u003e\nLink: https://lore.kernel.org/r/20260618103713.2588984-1-vladimir.kondratiev@mobileye.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "a8be5e9478be08f09a720f7013ec30d6e2bd9fcc",
      "tree": "ec25b77714dc1492dbdb87a92ebadd183c6a5167",
      "parents": [
        "914aeddaf16d91d39112056c5dd2b013b757a424"
      ],
      "author": {
        "name": "Bo Gan",
        "email": "ganboing@gmail.com",
        "time": "Mon Jun 08 23:00:24 2026 -0700"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Jun 17 11:43:38 2026 +0530"
      },
      "message": "lib: sbi: Rework misaligned vector load/store\n\nFix the following issues with misaligned vector load/store:\n\na. Stack overflow: the mask[VLEN_MAX / 8] variable consumes 8K stack\nspace, given VLEN_MAX\u003d65536, overflowing the default-sized stack.\nThere\u0027s no need to fetch the whole mask in one go, instead, make it\non-demand. Use a 128-byte mask as local buffer to hold the sliding\nwindow of mask. For rvv load, this is allowed -- from the spec:\n\n  \"The destination vector register group for a masked vector\n   instruction cannot overlap the source mask register (v0),\n   unless the destination vector register is being written with\n   a mask value (e.g., compares) or the scalar result of a reduction\"\n\nWe don\u0027t need to worry about the mask getting overwritten.\n\nb. Maintain the value of vstart upon abort (uptrap) to avoid duplicate\nwork. After fault resolution, the instruction can restart from the\nfaulting vstart. For Fault-Only-First loads, reset vstart to 0, as\npreviously done so, to conform to spec.\n\nc. Explicitly set VS dirty in VSSTATUS with SET_VS_DIRTY() if faulting\nfrom V\u003d1, and if any vector register, including vstart/vl/vtype, gets\nchanged in the handler. It can add 1 unnecessary op to set VS dirty\nin M/SSTATUS (not VSSTATUS), where the HW already did, but for code\nsimplicity, do it anyway. The overhead should be negligible.\n\nSigned-off-by: Bo Gan \u003cganboing@gmail.com\u003e\nTested-by: Anirudh Srinivasan \u003casrinivasan@oss.tenstorrent.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260609060024.706-5-ganboing@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "914aeddaf16d91d39112056c5dd2b013b757a424",
      "tree": "5005c5fd4bee30ca6f8938c266f3feba3aeb2699",
      "parents": [
        "1475f147f646f337c85599b03bd3185fe591da39"
      ],
      "author": {
        "name": "Bo Gan",
        "email": "ganboing@gmail.com",
        "time": "Mon Jun 08 23:00:23 2026 -0700"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Jun 17 11:41:11 2026 +0530"
      },
      "message": "lib: sbi: Add variable-length unprivilege access functions\n\nsbi_load/store_loop read/write variable-length buffer unprivileged.\nBoth function use the widest aligned 8/4/2/1 byte load/stores in each\nloop to reduce the total number of iterations.\n\nAlso switch the scalar/vector misaligned handlers to make use of such\nfunctions to simplify code.\n\nMiscellaneous: remove the unnecessary [taddr] in inline assembly\n\nSigned-off-by: Bo Gan \u003cganboing@gmail.com\u003e\nTested-by: Anirudh Srinivasan \u003casrinivasan@oss.tenstorrent.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260609060024.706-4-ganboing@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "1475f147f646f337c85599b03bd3185fe591da39",
      "tree": "8ed036858c0a3314cfdd6477894a213530859c8f",
      "parents": [
        "4120e6dce250b3fcec44394dca528ee1f39f35ba"
      ],
      "author": {
        "name": "Bo Gan",
        "email": "ganboing@gmail.com",
        "time": "Mon Jun 08 23:00:22 2026 -0700"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Jun 17 11:37:19 2026 +0530"
      },
      "message": "lib: sbi: Rework and split sbi_misaligned(_v)_tinst_fixup\n\nThe load/store address offset between the uptrap and the orig_trap\ncan be derived by orig_trap-\u003etval - uptrap-\u003etval, thus refactor\nthe function prototype for simplicity.\n\nFor vector load, sbi_misaligned_v_tinst_fixup is introduced. There\u0027s\nno transformed instruction for vector load/store, so null out tinst\nif the fault is not a guest-page fault.\n\nSigned-off-by: Bo Gan \u003cganboing@gmail.com\u003e\nTested-by: Anirudh Srinivasan \u003casrinivasan@oss.tenstorrent.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260609060024.706-3-ganboing@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "4120e6dce250b3fcec44394dca528ee1f39f35ba",
      "tree": "909d4d9708acf819b6fba48191fd4f420535fdf6",
      "parents": [
        "3afe63d4e3aff1a777e3e880b8b433ca225a311c"
      ],
      "author": {
        "name": "Bo Gan",
        "email": "ganboing@gmail.com",
        "time": "Mon Jun 08 23:00:21 2026 -0700"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Jun 17 11:23:05 2026 +0530"
      },
      "message": "lib: sbi: cosmetic changes to reduce indentation\n\nIn preparation for subsequent patches.\n\nSigned-off-by: Bo Gan \u003cganboing@gmail.com\u003e\nTested-by: Anirudh Srinivasan \u003casrinivasan@oss.tenstorrent.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260609060024.706-2-ganboing@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "3afe63d4e3aff1a777e3e880b8b433ca225a311c",
      "tree": "d270778fdb2f411dcfad6b49fb824ff83a614ac6",
      "parents": [
        "d936372eab7f2cdd93d21f67d2381914589d9dd9"
      ],
      "author": {
        "name": "David E. Garcia Porras",
        "email": "david.garcia@aheadcomputing.com",
        "time": "Tue Jun 16 11:01:18 2026 -0600"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Wed Jun 17 09:42:57 2026 +0530"
      },
      "message": "lib: sbi: dbtr: do not unconditionally access tdata2/tdata3 CSRs\n\nThe current SBI DBTR extension implementation accesses tdata2 and tdata3\nwithout first checking whether either register is implemented on the\nunderlying hart. This produces an illegal instruction exception on\notherwise spec-compliant cores that legitimately omit one or both\nregisters.\n\nPer the RISC-V Debug Specification, Chapter 5 (Sdtrig ISA Extension)\nand Section 5.7 (Trigger Module Registers):\n\n  Section 5 (Sdtrig introduction):\n    \"If Sdtrig is implemented, the Trigger Module must support at least\n     one trigger. Accessing trigger CSRs that are not used by any of the\n     implemented triggers must result in an illegal instruction\n     exception. M-Mode and Debug Mode accesses to trigger CSRs that are\n     used by any of the implemented triggers must succeed, regardless of\n     the current type of the currently selected trigger.\"\n\n  Section 5.7 (Trigger Module Registers):\n    \"Attempts to access an unimplemented Trigger Module Register raise\n     an illegal instruction exception.\"\n\nPer-register optionality is also explicit:\n\n  Section 5.7.3 (Trigger Data 2, at 0x7a2):\n    \"Trigger-specific data. It is optional if no implemented triggers\n     use it.\"\n\n  Section 5.7.4 (Trigger Data 3, at 0x7a3):\n    \"Trigger-specific data. It is optional if no implemented triggers\n     use it.\"\n\n  Section 5.7.17 (Trigger Extra (RV32), at 0x7a3), which also applies\n  via textra64 on RV64:\n    \"All functionality in this register is optional. Any number of\n     upper bits of mhvalue and svalue may be tied to 0. mhselect and\n     sselect may only support 0 (ignore).\"\n\nUnconditionally accessing tdata2/tdata3 in the install/update/read/\nuninstall paths causes SBI calls to fail with an illegal instruction\nexception on hardware that does not implement one or both CSRs, even\nif the supervisor-supplied trigger configuration does not require the\nmissing CSR(s).\n\nThis patch:\n\n  1. Introduces tdata_read_safe() / tdata_write_safe() helpers that\n     wrap csr_read_allowed / csr_write_allowed so that an illegal-\n     instruction trap raised by an unimplemented CSR is caught locally\n     rather than propagated. On the read path, a trapped read yields\n     zero; on the write path, the trap is silently absorbed (writes to\n     an unimplemented CSR are no-ops by definition). Every tdata2/tdata3\n     read and write in the install/update/read/uninstall paths is\n     converted to these helpers.\n\n  2. On the install and update paths, rejects requests that program\n     a non-zero trig_tdata2 or trig_tdata3 into an unimplemented CSR\n     with SBI_ERR_NOT_SUPPORTED, matching the SBI spec\n     wording in sections 19.4 / 19.5:\n\n       \"One of the trigger configuration can\u0027t be programmed due to\n        unimplemented optional bits in tdata1, tdata2, or tdata3\n        CSRs.\"\n\n     Implementation status is probed once per call via the\n     tdata_implemented() helper. This only catches the \"whole CSR\n     unimplemented\" case; tied-off WARL bits inside an otherwise-\n     implemented CSR are not caught here and would require programming\n     the trigger and reading the value back for comparison, which can\n     be addressed separately.\n\n  3. Enable tdata3 configuration in the debug trigger install path.\n\nReferences:\n  - RISC-V Debug Specification, Chapter 5 (Sdtrig), sections 5, 5.7,\n    5.7.3, 5.7.4, 5.7.17.\n  - RISC-V SBI Specification v3.0, Chapter 19 (Debug Triggers\n    Extension), sections 19.4, 19.5.\n\nFixes: 97f234f15c96 (\"lib: sbi: Introduce the SBI debug triggers extension support\")\nSuggested-by: Nicholas Piggin \u003cnpiggin@gmail.com\u003e\nSuggested-by: Himanshu Chauhan \u003chimanshu.chauhan@oss.qualcomm.com\u003e\nSigned-off-by: David E. Garcia Porras \u003cdavid.garcia@aheadcomputing.com\u003e\nReviewed-By: Himanshu Chauhan \u003chimanshu.chauhan@oss.qualcomm.com\u003e\nLink: https://lore.kernel.org/r/20260616170118.3515676-1-david.garcia@aheadcomputing.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "d936372eab7f2cdd93d21f67d2381914589d9dd9",
      "tree": "ee818a76eb447770eca1bf249370fed1e43011bf",
      "parents": [
        "56c39d1f08cb7aa1132e115b7850715c63212d57"
      ],
      "author": {
        "name": "Oriol Catalan",
        "email": "oriol.catalan@openchip.com",
        "time": "Tue Jun 16 19:09:33 2026 +0530"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Tue Jun 16 19:21:11 2026 +0530"
      },
      "message": "lib: sbi: Make per-HART stack size configurable via Kconfig\n\nThe per-HART stack size for exception/interrupt handling is currently\nhardcoded to 8192 bytes in SBI_PLATFORM_DEFAULT_HART_STACK_SIZE. This\nmay not be sufficient for platforms with deeper call stacks (e.g. those\nenabling additional SBI extensions) or may be wasteful for minimal\nplatforms.\n\nIntroduce a HART_STACK_SIZE Kconfig option in lib/sbi/Kconfig with a\nvalid range of 8192 to 1048576 bytes and a default of 8192 bytes to\npreserve existing behavior. The SBI_PLATFORM_DEFAULT_HART_STACK_SIZE\nmacro now resolves to CONFIG_HART_STACK_SIZE, allowing all platforms\nto benefit from a single configuration knob without any source changes.\n\nSigned-off-by: Oriol Catalan \u003coriol.catalan@openchip.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/VI0P192MB3062735A6194BB6DA72083499E002@VI0P192MB3062.EURP192.PROD.OUTLOOK.COM\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "56c39d1f08cb7aa1132e115b7850715c63212d57",
      "tree": "fc22d6a8decc93b75c604e873d76fe91b9b56a63",
      "parents": [
        "3aafbf5a854058e1a02c37426bb629beb5e0c2f2"
      ],
      "author": {
        "name": "Bo Gan",
        "email": "ganboing@gmail.com",
        "time": "Fri Jun 05 04:32:13 2026 -0700"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Tue Jun 16 13:42:47 2026 +0530"
      },
      "message": "lib: sbi: Rework load/store emulator instruction decoding\n\nRehaul instruction decoding to fix the following issues:\n\n- We assume the XLEN of previous mode is the same as MXLEN. However,\n  RVC instructions decodes differently in RV32 and RV64, so shouldn\u0027t\n  have assumed that.\n- We assume it\u0027s a misaligned fault and the load/store offset is 0,\n  i.e., base address \u003d\u003d fault address, but access faults can have\n  non-0 offset (on HW supporting misaligned accesses), so platform\n  specific load/store fault handler gets the wrong base address.\n- No checking of [63:32] of tinst in RV64, which is explicitly\n  required by Privileged ISA 19.6.3. Must reject tinst with non-0\n  high 32 bits.\n\nThus, fix all the above. For misaligned load/store fault, the address\noffset should be 0, but we\u0027ll validate that on a DEBUG build. On an\noptmized build, we kill the use of base address, and use trap address\ninstead (same as before), which lets the compiler optimize out imm\nparsing and other calculations.\n\nI also analyzed the behavior of misaligned fault handler before fix.\nWith the following conditions met, it can trigger data corruption:\n\n- HW doesn\u0027t transform instruction into tinst.\n- HW doesn\u0027t support misaligned load/store, and OS doesn\u0027t enable\n  misaligned delegation, thus OpenSBI handler is in effect\n- HW supports mixed XLEN, and M mode is running RV64, and the trapping\n  mode (U/VS/VU) is running RV32.\n- The trapping instruction is c.f{l|s}w(sp).\n\nDue to the incorrect insn decoding, the trapping instruction would\nmistakenly be decoded as c.{l|s}d(sp). With this fix, c.f{l|s}w(sp)\nin RV32 is now emulated correctly.\n\nValidation:\nThe patch is validated to have fixed the issue with test cases running\non a modified version of QEMU that exposes misaligned faults [1], and\na further modified version that removes tinst transformation [2]. The\nS-mode OS is a local build of Debian Trixie 6.12 kernel that enables\nCOMPAT (RV32), and the U-mode test application exercises all integer\nand floating-point load/store (RVIFD64/32+RVC64/32) instructions with\nall possible imm values. The patch is also tested on real HW (Sifive\nP550/ESWIN EIC7700), which only supports RV64. On P550, the same test\nwas validated both in U mode and VU mode, where the host runs a 6.12\nESWIN vendor kernel that has some ESWIN SoC device driver patches [3]\napplied, and the guest runs the exact same Debian Trixie 6.12 kernel\nmentioned above.\n\n[1] https://github.com/ganboing/qemu/tree/ganboing-misalign\n[2] https://github.com/ganboing/qemu/tree/ganboing-misalign-no-tinst\n[3] https://github.com/sifiveinc/riscv-linux/tree/rel/kernel-6.12/hifive-premier-p550\n\nFixes: 7219477f7b40 (\"lib: Use MTINST CSR in misaligned load/store emulation\")\nFixes: b5ae8e8a650d (\"lib: Add misaligned load/store trap handling\")\nFixes: 4c112650bbb0 (\"lib: sbi: abstract out insn decoding to unify mem fault handlers\")\nSigned-off-by: Bo Gan \u003cganboing@gmail.com\u003e\nTested-by: Anirudh Srinivasan \u003casrinivasan@oss.tenstorrent.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260605113214.242-8-ganboing@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "3aafbf5a854058e1a02c37426bb629beb5e0c2f2",
      "tree": "ad32a3e2e7aac85293f8cc0936f23deab8b2d765",
      "parents": [
        "eba121b459889868513e5e87639d5273012728f7"
      ],
      "author": {
        "name": "Bo Gan",
        "email": "ganboing@gmail.com",
        "time": "Fri Jun 05 04:32:12 2026 -0700"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Tue Jun 16 13:25:36 2026 +0530"
      },
      "message": "Makefile: define OPENSBI_DEBUG if DEBUG builds\n\nSigned-off-by: Bo Gan \u003cganboing@gmail.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nTested-by: Anirudh Srinivasan \u003casrinivasan@oss.tenstorrent.com\u003e\nLink: https://lore.kernel.org/r/20260605113214.242-7-ganboing@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "eba121b459889868513e5e87639d5273012728f7",
      "tree": "634785c538c28f9ea84606573cb4b71478424261",
      "parents": [
        "7dfaa4678213521e714a8a7cecb7f7a722dcd18a"
      ],
      "author": {
        "name": "Bo Gan",
        "email": "ganboing@gmail.com",
        "time": "Fri Jun 05 04:32:11 2026 -0700"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Tue Jun 16 13:25:27 2026 +0530"
      },
      "message": "lib: sbi: Do not override emulator callback for vector load/store\n\nIt\u0027s wrong to override the emulator callback in sbi_trap_emulate_load/\nstore. The function must respect the callback function passed in the\nparameter. Hence, let the misaligned emulator callback decide when to\nuse sbi_misaligned_v_ld/st_emulator. To clean up things, also make the\nfollowing changes:\n\n- Add the `insn` parameter to the callback. The trapping insn has been\n  fetched by the caller already, whether transformed or directly loaded,\n  thus saving the trouble in the callback. Note that you must not rely\n  on the length of the `insn`, as it can be a transformed one from tinst\n\n- Also the `tcntx` is added, providing the callback with register values\n  to handle vector insn or other customized insns.\n\n- Clarify that the read/write length (rlen/wlen) can be 0, in which\n  case it could be a vector load/store or some customized instruction.\n  The callback is responsible to handle it accordingly.\n\nAlso fixed issues in the sbi_misaligned_v_ld/st_emulator:\na. Redirect the trap when OPENSBI_CC_SUPPORT_VECTOR is not available.\nb. Ensure the return code is \u003e0 when no faults are redirected.\n\nFixes: c2acc5e5b0d8 (\"lib: sbi_misaligned_ldst: Add handling of vector load/store\")\nSigned-off-by: Bo Gan \u003cganboing@gmail.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nTested-by: Anirudh Srinivasan \u003casrinivasan@oss.tenstorrent.com\u003e\nLink: https://lore.kernel.org/r/20260605113214.242-6-ganboing@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "7dfaa4678213521e714a8a7cecb7f7a722dcd18a",
      "tree": "efdaa950f7b78dec9b1e122e13340719ee8b878e",
      "parents": [
        "1b3424d5c0901dcea24ca3427009b299662949a0"
      ],
      "author": {
        "name": "Bo Gan",
        "email": "ganboing@gmail.com",
        "time": "Fri Jun 05 04:32:10 2026 -0700"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Tue Jun 16 12:31:01 2026 +0530"
      },
      "message": "include: sbi: set FS dirty in vsstatus when V\u003d1\n\nAccording to Privileged ISA 19.2.11:\n\nModifying the floating-point state when V\u003d1 causes both fields\n(vsstatus.FS and the HS-level sstatus.FS) to be set to 3 (Dirty)\n\nFixes: 130e65dd9d44 (\"lib: sbi: Implement SET_FS_DIRTY() to make sure the mstatus FS dirty is set\")\nSigned-off-by: Bo Gan \u003cganboing@gmail.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nTested-by: Anirudh Srinivasan \u003casrinivasan@oss.tenstorrent.com\u003e\nLink: https://lore.kernel.org/r/20260605113214.242-5-ganboing@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "1b3424d5c0901dcea24ca3427009b299662949a0",
      "tree": "5a0f65bdd5de06eb01fdb86f985489790a7ad097",
      "parents": [
        "bd986ed03d6336ddbd5ee90aca8a3ca533b1d5a4"
      ],
      "author": {
        "name": "Bo Gan",
        "email": "ganboing@gmail.com",
        "time": "Fri Jun 05 04:32:09 2026 -0700"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Tue Jun 16 12:30:22 2026 +0530"
      },
      "message": "include: sbi: Add GET_RDS_NUM/SET(_FP32/_FP64)_RDS macros\n\nThese macros can be used to decode rd\u0027 and set rd\u0027 in RVC instructions\n\nSigned-off-by: Bo Gan \u003cganboing@gmail.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nTested-by: Anirudh Srinivasan \u003casrinivasan@oss.tenstorrent.com\u003e\nLink: https://lore.kernel.org/r/20260605113214.242-4-ganboing@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "bd986ed03d6336ddbd5ee90aca8a3ca533b1d5a4",
      "tree": "fd5272ac1445e03e37821704e49f178a2a4e2269",
      "parents": [
        "3c4fa3404eb331b0c605d0de2d1b6d840d95109e"
      ],
      "author": {
        "name": "Bo Gan",
        "email": "ganboing@gmail.com",
        "time": "Fri Jun 05 04:32:08 2026 -0700"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Tue Jun 16 12:29:59 2026 +0530"
      },
      "message": "include: sbi: Add sbi_regs_prev_xlen\n\nsbi_regs_prev_xlen reports the xlen of previous mode by decoding\nfrom multiple CSRs including mstatus/hstatus/vsstatus\n\nSigned-off-by: Bo Gan \u003cganboing@gmail.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nTested-by: Anirudh Srinivasan \u003casrinivasan@oss.tenstorrent.com\u003e\nLink: https://lore.kernel.org/r/20260605113214.242-3-ganboing@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "3c4fa3404eb331b0c605d0de2d1b6d840d95109e",
      "tree": "2c43f7a1d2b4429d615781a557f959dec8915a51",
      "parents": [
        "7bff4e529e6c32df5c9364960c174fa1a682cd5d"
      ],
      "author": {
        "name": "Bo Gan",
        "email": "ganboing@gmail.com",
        "time": "Fri Jun 05 04:32:07 2026 -0700"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Tue Jun 16 12:29:29 2026 +0530"
      },
      "message": "include: sbi: Add more mstatus and instruction encoding\n\n- Add MXL encoding for calculating XLEN.\n- Add instruction encoding for c.lbu/c.sb,\n  and imm encoding for multiple RVC insn.\n\nSigned-off-by: Bo Gan \u003cganboing@gmail.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nTested-by: Anirudh Srinivasan \u003casrinivasan@oss.tenstorrent.com\u003e\nLink: https://lore.kernel.org/r/20260605113214.242-2-ganboing@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "7bff4e529e6c32df5c9364960c174fa1a682cd5d",
      "tree": "f3af882c8c83f075793094ad2fdf9addc409dd17",
      "parents": [
        "0f42eff6eaca80bd483292c8911292cbaa2341fd"
      ],
      "author": {
        "name": "Bo Gan",
        "email": "ganboing@gmail.com",
        "time": "Fri Jun 05 00:57:08 2026 -0700"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Mon Jun 15 10:38:09 2026 +0530"
      },
      "message": "platform: generic: eswin: Add eic770x_hsm and fix warm reset issues\n\nDuring warm reset, my EIC770X/Hifive Premier P550 can sometimes\nencounter memory corruption issue crashing Linux boot. Currently the\nissue is mitigated by having a sbi_printf before writing to the reset\nregister. I analyzed the issue further since then. From the SoC\ndatasheet[1], it\u0027s recommended to implement power-down flow as:\n\n  a. Designate a primary core, and let it broadcast requests to other\n     cores to execute a CEASE insn. Primary core also notifies an\n     \"Externel Agent\" to start monitoring.\n  b. Primary core waits for other cores to CEASE before it CEASEs.\n  c. \"External Agent\" waits for primary core to CEASE before resets\n     the Core Complex.\n\nIt\u0027s possible that EIC770X can trigger undefined behavior if the core\ncomplex is reset while the harts are actively running. The sbi_printf\nin the reset handler effectively hides the problem by delaying the\nreset -- by the time sbi_printf finishes, all other harts will have\nalready landed in the loop in sbi_hsm_hart_wait(), which parks the hart.\nWithout the sbi_printf, I confirmed that other harts haven\u0027t reached\nsbi_hsm_hart_wait yet before current hart resets the SoC. (by debugging)\n\nTo safely reset, and inspired by the datasheet, the warm reset logic\nin eic770x.c now use the current hart as both primary core and the\n\"External Agent\", and other harts as secondary cores. It leverages\nthe HSM framework and a new eic770x_hsm device to CEASE other harts,\nand wait for them to CEASE before resets the SoC. with the sbi_printf\nbefore reset removed, and this logic in place, stress test shows that\nthe memory corruption issue no longer occurs.\n\nThe new eic770x_hsm device is only used for the reset-CEASE logic at\nthe moment, and may be extended to a fully functional HSM device in\nthe future.\n\n[1] https://github.com/eswincomputing/EIC7700X-SoC-Technical-Reference-Manual\n\nFixes: e5797e0688c1 (\"platform: generic: eswin: add EIC7700\")\nSigned-off-by: Bo Gan \u003cganboing@gmail.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260605075708.96-3-ganboing@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "0f42eff6eaca80bd483292c8911292cbaa2341fd",
      "tree": "ec900fcea23e95ef689bea3be08a837105050fcf",
      "parents": [
        "dcb5179b5050075b575369536b4f113f6d15e7b1"
      ],
      "author": {
        "name": "Bo Gan",
        "email": "ganboing@gmail.com",
        "time": "Fri Jun 05 00:57:07 2026 -0700"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Mon Jun 15 10:38:09 2026 +0530"
      },
      "message": "include: utils/hsm: Add __noreturn attribute for sifive_cease\n\nDecorate the sifive_cease to allow more compiler optimizations\n\nSigned-off-by: Bo Gan \u003cganboing@gmail.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260605075708.96-2-ganboing@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "dcb5179b5050075b575369536b4f113f6d15e7b1",
      "tree": "7dd5fa8564494a490118585a2515a836b2334493",
      "parents": [
        "1157eb37e29500b5e30d4c70bda6979fb516c7fc"
      ],
      "author": {
        "name": "Michael Ellerman",
        "email": "mpe@kernel.org",
        "time": "Tue Jun 02 21:24:37 2026 +1000"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Mon Jun 15 10:25:43 2026 +0530"
      },
      "message": "gitignore: Ignore python cache directories\n\nRunning ./scripts/Kconfiglib/setconfig.py leaves the tree with untracked\nfiles:\n\n    Untracked files:\n      (use \"git add \u003cfile\u003e...\" to include in what will be committed)\n            scripts/Kconfiglib/__pycache__/\n\nAdd __pycache__ to .gitignore to fix it.\n\nSigned-off-by: Michael Ellerman \u003cmpe@kernel.org\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260602-fix-gitignore-v1-1-4299e2e40ee4@kernel.org\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "1157eb37e29500b5e30d4c70bda6979fb516c7fc",
      "tree": "659aea11613fb30590eb4520b2367fd45fa16528",
      "parents": [
        "dec9141a77058dc941f2d0f07107f7306a995c35"
      ],
      "author": {
        "name": "Michael Ellerman",
        "email": "mpe@kernel.org",
        "time": "Sat May 30 00:42:18 2026 +1000"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Mon Jun 15 10:21:03 2026 +0530"
      },
      "message": "lib: sbi_trap_v_ldst: Redirect unhandled traps\n\nWhen SBI is built with a compiler that doesn\u0027t support vector, the\nmisaligned vector load/store emulation is not built in, the handlers are\njust stubs.\n\nCurrently the stubs just return 0, causing sbi_trap_emulate_load() to\nreturn without incrementing mepc, meaning the instruction will just\nfault again, an infinite loop.\n\nFix the stubs to use sbi_trap_redirect(), which forwards the trap to the\nprevious mode, allowing it to be handled there.\n\nFixes: c2acc5e5 (\"lib: sbi_misaligned_ldst: Add handling of vector load/store\")\nSigned-off-by: Michael Ellerman \u003cmpe@kernel.org\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260530-trap-redirect-v1-1-45d4d333d8c9@kernel.org\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "dec9141a77058dc941f2d0f07107f7306a995c35",
      "tree": "2b974255b12b3be8e9cd93efd0d6e1e737fb3d72",
      "parents": [
        "9c9767504d4e2e48015642239cea954d17ea6bef"
      ],
      "author": {
        "name": "Inochi Amaoto",
        "email": "inochiama@gmail.com",
        "time": "Fri May 29 16:52:32 2026 +0800"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Mon Jun 15 10:10:25 2026 +0530"
      },
      "message": "lib: utils/reset: Add litex SoC reset driver\n\nLitex SoC controller supports reboot function by toggling the first\nbit of the ctrl register. Add a reset driver so other software can\nuse it.\n\nSigned-off-by: Inochi Amaoto \u003cinochiama@gmail.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260529085234.1682842-1-inochiama@gmail.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "9c9767504d4e2e48015642239cea954d17ea6bef",
      "tree": "8616a9a7824b82c7b95b31c9f0ef8a10833855c2",
      "parents": [
        "3d29f380a66c2b5608faf7e4f4b0ff3c945d0d64"
      ],
      "author": {
        "name": "Nicholas Piggin",
        "email": "npiggin@gmail.com",
        "time": "Fri May 22 17:54:17 2026 -0500"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Sat Jun 13 15:22:32 2026 +0530"
      },
      "message": "lib: sbi: Drop fw_rw_offset alignment requirement for single fw region\n\nIn a single fw region scheme, there is no separate PMP created for RW\nmemory. The checks that opensbi does for the alignment between fw_start\nand fw_rw_start (using fw_rw_offset) and the power of 2 check for\nfw_rw_offset are no longer necessary.\n\nUpdate sbi_domain_init so that these checks are only done in the non\nsingle fw region scheme.\n\nSigned-off-by: Nicholas Piggin \u003cnpiggin@gmail.com\u003e\nCo-developed-by: Anirudh Srinivasan \u003casrinivasan@oss.tenstorrent.com\u003e\nSigned-off-by: Anirudh Srinivasan \u003casrinivasan@oss.tenstorrent.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260522-fw_rw_start_alignment-v1-1-362c17331541@oss.tenstorrent.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "3d29f380a66c2b5608faf7e4f4b0ff3c945d0d64",
      "tree": "88716ad0d3fe224b64bf83f09efc011daf95d325",
      "parents": [
        "b508e6e25eb23a6a188f8cfea94f338c85c75cab"
      ],
      "author": {
        "name": "David E. Garcia Porras",
        "email": "david.garcia@aheadcomputing.com",
        "time": "Fri May 22 08:46:07 2026 -0600"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Fri Jun 12 18:59:40 2026 +0530"
      },
      "message": "lib: sbi_pmu: Honor CLEAR_VALUE/AUTO_START for all hardware event types\n\nsbi_pmu_ctr_cfg_match() only acts on SBI_PMU_CFG_FLAG_CLEAR_VALUE and\nSBI_PMU_CFG_FLAG_AUTO_START when the event type is SBI_PMU_EVENT_TYPE_HW.\nHowever, pmu_ctr_find_hw() allocates a hardware counter from the same\nhw_event_map for SBI_PMU_EVENT_TYPE_HW_CACHE, SBI_PMU_EVENT_TYPE_HW_RAW,\nand SBI_PMU_EVENT_TYPE_HW_RAW_V2 as well, and the start/clear helpers\n(pmu_ctr_start_hw, pmu_ctr_write_hw) operate on the counter index alone\nand are agnostic to the event type. As a result, when a supervisor\nconfigures a HW_CACHE/HW_RAW/HW_RAW_V2 event with these flags, the\ncounter is programmed and recorded in active_events[] but is never\ncleared or started, requiring an extra SBI call to make it count.\n\nExtend the check to cover all hardware-counter event types so that\nthe configuration flags take effect for HW_CACHE and raw events too.\n\nDeliberately avoiding using \"not FW\" logic to be explicit about\nHW-backed events only.\n\nFixes: 13d40f21 (\"lib: sbi: Add PMU support\")\nSigned-off-by: David E. Garcia Porras \u003cdavid.garcia@aheadcomputing.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260522144608.3433470-1-david.garcia@aheadcomputing.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "b508e6e25eb23a6a188f8cfea94f338c85c75cab",
      "tree": "a05677e12200df70dc5b291314ac5f87c1413b23",
      "parents": [
        "ecc92e87a990e3f45081fbf2c84ad342d9226aaa"
      ],
      "author": {
        "name": "Evgeny Voevodin",
        "email": "evvoevod@tenstorrent.com",
        "time": "Tue May 19 22:50:14 2026 +0000"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Fri Jun 12 18:59:16 2026 +0530"
      },
      "message": "lib: sbi: Apply budget restriction when polling Zkr CSR state transition\n\nZkr architecture doesn\u0027t define a time limit on state transitions\nwhich results in hanging on unresponsive or event-driven platforms.\nTo prevent this, we need to limit polling iterations and fall back\nin case the budget is over, and stack guard keeps its initial value.\nThe budget is configurable with CONFIG_ZKR_POLL_BUDGET, defaulting\nto 1000 iterations. Successful reads do not consume a try.\n\nSigned-off-by: Evgeny Voevodin \u003cevvoevod@tenstorrent.com\u003e\nReviewed-by: Anup Patel \u003canup@brainfault.org\u003e\nLink: https://lore.kernel.org/r/20260519225014.244672-1-evvoevod@tenstorrent.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    },
    {
      "commit": "ecc92e87a990e3f45081fbf2c84ad342d9226aaa",
      "tree": "2bdc9e13c6e7f2fad948f6a08374c8ebe4f6d6ee",
      "parents": [
        "a62385ab803325a27e7ff9981ca644b49de65e2b"
      ],
      "author": {
        "name": "Anup Patel",
        "email": "anup.patel@oss.qualcomm.com",
        "time": "Thu May 21 13:56:25 2026 +0530"
      },
      "committer": {
        "name": "Anup Patel",
        "email": "anup@brainfault.org",
        "time": "Fri Jun 12 16:08:29 2026 +0530"
      },
      "message": "platform: generic: Optimize extensions_init() to parse ISA extensions once\n\nInstead of parsing ISA extensions separately for each hart in the\ngeneric_extensions_init() function, it is better to parse ISA extensions\nfor all available harts in the cold boot path. Also, this allows us\nto remove fdt_isa_bitmap from scratch space and directly initialize\n\"extensions\" in struct sbi_hart_features for each hart.\n\nSigned-off-by: Anup Patel \u003canup.patel@oss.qualcomm.com\u003e\nLink: https://lore.kernel.org/r/20260521082625.1520870-3-anup.patel@oss.qualcomm.com\nSigned-off-by: Anup Patel \u003canup@brainfault.org\u003e\n"
    }
  ],
  "next": "a62385ab803325a27e7ff9981ca644b49de65e2b"
}
