)]}' { "log": [ { "commit": "c94d59a126cb9a8d1f71e3e044363d654dcd7af8", "tree": "c5c293da27c37092574953f656769a39e9e5be91", "parents": [ "1a80a098c606b285fb0a13aa992af4f86da1ff06", "1d6c39c89f617c9fec6bbae166e25b16a014f7c8" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 14 11:24:19 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 14 11:24:19 2025 -0700" }, "message": "Merge tag \u0027trace-v6.15-rc6\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace\n\nPull tracing fixes from Steven Rostedt:\n\n - Fix sample code that uses trace_array_printk()\n\n The sample code for in kernel use of trace_array (that creates an\n instance for use within the kernel) and shows how to use\n trace_array_printk() that writes into the created instance, used\n trace_printk_init_buffers(). But that function is used to initialize\n normal trace_printk() and produces the NOTICE banner which is not\n needed for use of trace_array_printk(). The function to initialize\n that is trace_array_init_printk() that takes the created trace array\n instance as a parameter.\n\n Update the sample code to reflect the proper usage.\n\n - Fix preemption count output for stacktrace event\n\n The tracing buffer shows the preempt count level when an event\n executes. Because writing the event itself disables preemption, this\n needs to be accounted for when recording. The stacktrace event did\n not account for this so the output of the stacktrace event showed\n preemption was disabled while the event that triggered the stacktrace\n shows preemption is enabled and this leads to confusion. Account for\n preemption being disabled for the stacktrace event.\n\n The same happened for stack traces triggered by function tracer.\n\n - Fix persistent ring buffer when trace_pipe is used\n\n The ring buffer swaps the reader page with the next page to read from\n the write buffer when trace_pipe is used. If there\u0027s only a page of\n data in the ring buffer, this swap will cause the \"commit\" pointer\n (last data written) to be on the reader page. If more data is written\n to the buffer, it is added to the reader page until it falls off back\n into the write buffer.\n\n If the system reboots and the commit pointer is still on the reader\n page, even if new data was written, the persistent buffer validator\n will miss finding the commit pointer because it only checks the write\n buffer and does not check the reader page. This causes the validator\n to fail the validation and clear the buffer, where the new data is\n lost.\n\n There was a check for this, but it checked the \"head pointer\", which\n was incorrect, because the \"head pointer\" always stays on the write\n buffer and is the next page to swap out for the reader page. Fix the\n logic to catch this case and allow the user to still read the data\n after reboot.\n\n* tag \u0027trace-v6.15-rc6\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:\n ring-buffer: Fix persistent buffer when commit page is the reader page\n ftrace: Fix preemption accounting for stacktrace filter command\n ftrace: Fix preemption accounting for stacktrace trigger command\n tracing: samples: Initialize trace_array_printk() with the correct function\n" }, { "commit": "1d6c39c89f617c9fec6bbae166e25b16a014f7c8", "tree": "55e51b9a5c275cdc6da245087a301195ea59dece", "parents": [ "11aff32439df6ca5b3b891b43032faf88f4a6a29" ], "author": { "name": "Steven Rostedt", "email": "rostedt@goodmis.org", "time": "Tue May 13 11:50:32 2025 -0400" }, "committer": { "name": "Steven Rostedt (Google)", "email": "rostedt@goodmis.org", "time": "Wed May 14 13:53:23 2025 -0400" }, "message": "ring-buffer: Fix persistent buffer when commit page is the reader page\n\nThe ring buffer is made up of sub buffers (sometimes called pages as they\nare by default PAGE_SIZE). It has the following \"pages\":\n\n \"tail page\" - this is the page that the next write will write to\n \"head page\" - this is the page that the reader will swap the reader page with.\n \"reader page\" - This belongs to the reader, where it will swap the head\n page from the ring buffer so that the reader does not\n race with the writer.\n\nThe writer may end up on the \"reader page\" if the ring buffer hasn\u0027t\nwritten more than one page, where the \"tail page\" and the \"head page\" are\nthe same.\n\nThe persistent ring buffer has meta data that points to where these pages\nexist so on reboot it can re-create the pointers to the cpu_buffer\ndescriptor. But when the commit page is on the reader page, the logic is\nincorrect.\n\nThe check to see if the commit page is on the reader page checked if the\nhead page was the reader page, which would never happen, as the head page\nis always in the ring buffer. The correct check would be to test if the\ncommit page is on the reader page. If that\u0027s the case, then it can exit\nout early as the commit page is only on the reader page when there\u0027s only\none page of data in the buffer. There\u0027s no reason to iterate the ring\nbuffer pages to find the \"commit page\" as it is already found.\n\nTo trigger this bug:\n\n # echo 1 \u003e /sys/kernel/tracing/instances/boot_mapped/events/syscalls/sys_enter_fchownat/enable\n # touch /tmp/x\n # chown sshd /tmp/x\n # reboot\n\nOn boot up, the dmesg will have:\n Ring buffer meta [0] is from previous boot!\n Ring buffer meta [1] is from previous boot!\n Ring buffer meta [2] is from previous boot!\n Ring buffer meta [3] is from previous boot!\n Ring buffer meta [4] commit page not found\n Ring buffer meta [5] is from previous boot!\n Ring buffer meta [6] is from previous boot!\n Ring buffer meta [7] is from previous boot!\n\nWhere the buffer on CPU 4 had a \"commit page not found\" error and that\nbuffer is cleared and reset causing the output to be empty and the data lost.\n\nWhen it works correctly, it has:\n\n # cat /sys/kernel/tracing/instances/boot_mapped/trace_pipe\n \u003c...\u003e-1137 [004] ..... 998.205323: sys_enter_fchownat: __syscall_nr\u003d0x104 (260) dfd\u003d0xffffff9c (4294967196) filename\u003d(0xffffc90000a0002c) user\u003d0x3e8 (1000) group\u003d0xffffffff (4294967295) flag\u003d0x0 (0\n\nCc: stable@vger.kernel.org\nCc: Mathieu Desnoyers \u003cmathieu.desnoyers@efficios.com\u003e\nLink: https://lore.kernel.org/20250513115032.3e0b97f7@gandalf.local.home\nFixes: 5f3b6e839f3ce (\"ring-buffer: Validate boot range memory events\")\nReported-by: Tasos Sahanidis \u003ctasos@tasossah.com\u003e\nTested-by: Tasos Sahanidis \u003ctasos@tasossah.com\u003e\nReviewed-by: Masami Hiramatsu (Google) \u003cmhiramat@kernel.org\u003e\nSigned-off-by: Steven Rostedt (Google) \u003crostedt@goodmis.org\u003e\n" }, { "commit": "11aff32439df6ca5b3b891b43032faf88f4a6a29", "tree": "d7df2cd065126ea107c52bc0afe1de3ab6748f40", "parents": [ "e333332657f615ac2b55aa35565c4a882018bbe9" ], "author": { "name": "pengdonglin", "email": "pengdonglin@xiaomi.com", "time": "Mon May 12 17:42:46 2025 +0800" }, "committer": { "name": "Steven Rostedt (Google)", "email": "rostedt@goodmis.org", "time": "Wed May 14 13:53:23 2025 -0400" }, "message": "ftrace: Fix preemption accounting for stacktrace filter command\n\nThe preemption count of the stacktrace filter command to trace ksys_read\nis consistently incorrect:\n\n$ echo ksys_read:stacktrace \u003e set_ftrace_filter\n\n \u003c...\u003e-453 [004] ...1. 38.308956: \u003cstack trace\u003e\n\u003d\u003e ksys_read\n\u003d\u003e do_syscall_64\n\u003d\u003e entry_SYSCALL_64_after_hwframe\n\nThe root cause is that the trace framework disables preemption when\ninvoking the filter command callback in function_trace_probe_call:\n\n preempt_disable_notrace();\n probe_ops-\u003efunc(ip, parent_ip, probe_opsbe-\u003etr, probe_ops, probe-\u003edata);\n preempt_enable_notrace();\n\nUse tracing_gen_ctx_dec() to account for the preempt_disable_notrace(),\nwhich will output the correct preemption count:\n\n$ echo ksys_read:stacktrace \u003e set_ftrace_filter\n\n \u003c...\u003e-410 [006] ..... 31.420396: \u003cstack trace\u003e\n\u003d\u003e ksys_read\n\u003d\u003e do_syscall_64\n\u003d\u003e entry_SYSCALL_64_after_hwframe\n\nCc: stable@vger.kernel.org\nFixes: 36590c50b2d07 (\"tracing: Merge irqflags + preempt counter.\")\nLink: https://lore.kernel.org/20250512094246.1167956-2-dolinux.peng@gmail.com\nSigned-off-by: pengdonglin \u003cdolinux.peng@gmail.com\u003e\nSigned-off-by: Steven Rostedt (Google) \u003crostedt@goodmis.org\u003e\n" }, { "commit": "e333332657f615ac2b55aa35565c4a882018bbe9", "tree": "9559a31698d22402c05037a56d1775321a1e5917", "parents": [ "1b0c192c92ea1fe2dcb178f84adf15fe37c3e7c8" ], "author": { "name": "pengdonglin", "email": "pengdonglin@xiaomi.com", "time": "Mon May 12 17:42:45 2025 +0800" }, "committer": { "name": "Steven Rostedt (Google)", "email": "rostedt@goodmis.org", "time": "Wed May 14 13:53:09 2025 -0400" }, "message": "ftrace: Fix preemption accounting for stacktrace trigger command\n\nWhen using the stacktrace trigger command to trace syscalls, the\npreemption count was consistently reported as 1 when the system call\nevent itself had 0 (\".\").\n\nFor example:\n\nroot@ubuntu22-vm:/sys/kernel/tracing/events/syscalls/sys_enter_read\n$ echo stacktrace \u003e trigger\n$ echo 1 \u003e enable\n\n sshd-416 [002] ..... 232.864910: sys_read(fd: a, buf: 556b1f3221d0, count: 8000)\n sshd-416 [002] ...1. 232.864913: \u003cstack trace\u003e\n \u003d\u003e ftrace_syscall_enter\n \u003d\u003e syscall_trace_enter\n \u003d\u003e do_syscall_64\n \u003d\u003e entry_SYSCALL_64_after_hwframe\n\nThe root cause is that the trace framework disables preemption in __DO_TRACE before\ninvoking the trigger callback.\n\nUse the tracing_gen_ctx_dec() that will accommodate for the increase of\nthe preemption count in __DO_TRACE when calling the callback. The result\nis the accurate reporting of:\n\n sshd-410 [004] ..... 210.117660: sys_read(fd: 4, buf: 559b725ba130, count: 40000)\n sshd-410 [004] ..... 210.117662: \u003cstack trace\u003e\n \u003d\u003e ftrace_syscall_enter\n \u003d\u003e syscall_trace_enter\n \u003d\u003e do_syscall_64\n \u003d\u003e entry_SYSCALL_64_after_hwframe\n\nCc: stable@vger.kernel.org\nFixes: ce33c845b030c (\"tracing: Dump stacktrace trigger to the corresponding instance\")\nLink: https://lore.kernel.org/20250512094246.1167956-1-dolinux.peng@gmail.com\nSigned-off-by: pengdonglin \u003cdolinux.peng@gmail.com\u003e\nSigned-off-by: Steven Rostedt (Google) \u003crostedt@goodmis.org\u003e\n" }, { "commit": "1a80a098c606b285fb0a13aa992af4f86da1ff06", "tree": "c7b2e187b48c8e0a5299223cef7f6160a73cba3d", "parents": [ "00f281fd9d8ddc1a83d3b188c087edac2116a089", "11854fe263eb1b9a8efa33b0c087add7719ea9b4" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 14 09:15:16 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 14 09:15:16 2025 -0700" }, "message": "Merge tag \u0027execve-v6.15-rc7\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux\n\nPull execve fix from Kees Cook:\n \"This fixes a corner case for ASLR-disabled static-PIE brk collision\n with vdso allocations:\n\n - binfmt_elf: Move brk for static PIE even if ASLR disabled\"\n\n* tag \u0027execve-v6.15-rc7\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux:\n binfmt_elf: Move brk for static PIE even if ASLR disabled\n" }, { "commit": "00f281fd9d8ddc1a83d3b188c087edac2116a089", "tree": "a6672a6990b626c8c5ab9d2966a2c9944c9f6cf4", "parents": [ "9f35e33144ae5377d6a8de86dd3bd4d995c6ac65", "15eaaa71e8ef142b122942b35653d513cfb90050" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 14 09:11:05 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 14 09:11:05 2025 -0700" }, "message": "Merge tag \u0027soc-fixes-6.15-2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc\n\nPull SoC fixes from Arnd Bergmann:\n \"These all address issues in devicetree files:\n\n - The Rockchip rk3588j are now limited the same way as the vendor\n kernel, to allow room for the industrial-grade temperature ranges.\n\n - Seven more Rockchip fixes address minor issues with specific boards\n\n - Invalid clk controller references in multiple amlogic chips, plus\n one accidentally disabled audio on clock\n\n - Two devicetree fixes for i.MX8MP boards, both for incorrect\n regulator settings\n\n - A power domain change for apple laptop touchbar, fixing\n suspend/resume problems\n\n - An incorrect DMA controller setting for sophgo cv18xx chips\"\n\n* tag \u0027soc-fixes-6.15-2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc:\n arm64: dts: amazon: Fix simple-bus node name schema warnings\n MAINTAINERS: delete email for Shiraz Hashim\n arm64: dts: imx8mp-var-som: Fix LDO5 shutdown causing SD card timeout\n arm64: dts: imx8mp: use 800MHz NoC OPP for nominal drive mode\n arm64: dts: amlogic: dreambox: fix missing clkc_audio node\n riscv: dts: sophgo: fix DMA data-width configuration for CV18xx\n arm64: dts: rockchip: fix Sige5 RTC interrupt pin\n arm64: dts: rockchip: Assign RT5616 MCLK rate on rk3588-friendlyelec-cm3588\n arm64: dts: rockchip: Align wifi node name with bindings in CB2\n arm64: dts: amlogic: g12: fix reference to unknown/untested PWM clock\n arm64: dts: amlogic: gx: fix reference to unknown/untested PWM clock\n ARM: dts: amlogic: meson8b: fix reference to unknown/untested PWM clock\n ARM: dts: amlogic: meson8: fix reference to unknown/untested PWM clock\n arm64: dts: apple: touchbar: Mark ps_dispdfr_be as always-on\n mailmap: Update email for Asahi Lina\n arm64: dts: rockchip: Fix mmc-pwrseq clock name on rock-pi-4\n arm64: dts: rockchip: Use \"regulator-fixed\" for btreg on px30-engicam for vcc3v3-btreg\n arm64: dts: rockchip: Add pinmuxing for eMMC on QNAP TS433\n arm64: dts: rockchip: Remove overdrive-mode OPPs from RK3588J SoC dtsi\n arm64: dts: rockchip: Allow Turing RK1 cooling fan to spin down\n" }, { "commit": "9f35e33144ae5377d6a8de86dd3bd4d995c6ac65", "tree": "07129e4ff4c9fd43b1b675ec512a203d0948488b", "parents": [ "405e6c37c89ef0df2bfc7a988820a3df22dacb1b" ], "author": { "name": "Eric Biggers", "email": "ebiggers@google.com", "time": "Mon May 12 19:58:39 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue May 13 14:36:08 2025 -0700" }, "message": "x86/its: Fix build errors when CONFIG_MODULES\u003dn\n\nFix several build errors when CONFIG_MODULES\u003dn, including the following:\n\n../arch/x86/kernel/alternative.c:195:25: error: incomplete definition of type \u0027struct module\u0027\n 195 | for (int i \u003d 0; i \u003c mod-\u003eits_num_pages; i++) {\n\nFixes: 872df34d7c51 (\"x86/its: Use dynamic thunks for indirect branches\")\nCc: stable@vger.kernel.org\nSigned-off-by: Eric Biggers \u003cebiggers@google.com\u003e\nAcked-by: Dave Hansen \u003cdave.hansen@intel.com\u003e\nTested-by: Steven Rostedt (Google) \u003crostedt@goodmis.org\u003e\nReviewed-by: Alexandre Chartre \u003calexandre.chartre@oracle.com\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "405e6c37c89ef0df2bfc7a988820a3df22dacb1b", "tree": "92397252cdfc65c5facd15b0f8db795530dafb59", "parents": [ "e9565e23cd89d4d5cd4388f8742130be1d6f182d", "fd837de3c9cb1a162c69bc1fb1f438467fe7f2f5" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue May 13 12:20:07 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue May 13 12:20:07 2025 -0700" }, "message": "Merge tag \u0027probes-fixes-v6.15-rc6\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace\n\nPull probes fixes from Masami Hiramatsu:\n\n - fprobe: Fix RCU warning message in list traversal\n\n fprobe_module_callback() using hlist_for_each_entry_rcu() traverse\n the fprobe list but it locks fprobe_mutex() instead of rcu lock\n because it is enough. So add lockdep_is_held() to avoid warning.\n\n - tracing: eprobe: Add missing trace_probe_log_clear for eprobe\n\n __trace_eprobe_create() uses trace_probe_log but forgot to clear it\n at exit. Add trace_probe_log_clear() calls.\n\n - tracing: probes: Fix possible race in trace_probe_log APIs\n\n trace_probe_log APIs are used in probe event (dynamic_events,\n kprobe_events and uprobe_events) creation. Only dynamic_events uses\n the dyn_event_ops_mutex mutex to serialize it. This makes kprobe and\n uprobe events to lock the same mutex to serialize its creation to\n avoid race in trace_probe_log APIs.\n\n* tag \u0027probes-fixes-v6.15-rc6\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:\n tracing: probes: Fix a possible race in trace_probe_log APIs\n tracing: add missing trace_probe_log_clear for eprobes\n tracing: fprobe: Fix RCU warning message in list traversal\n" }, { "commit": "fd837de3c9cb1a162c69bc1fb1f438467fe7f2f5", "tree": "84dd334214942fa29d0db25c9e065b1839e28073", "parents": [ "e41b5af4519f90f9a751805ede2102ae36caf5d0" ], "author": { "name": "Masami Hiramatsu (Google)", "email": "mhiramat@kernel.org", "time": "Sat May 10 12:44:41 2025 +0900" }, "committer": { "name": "Masami Hiramatsu (Google)", "email": "mhiramat@kernel.org", "time": "Tue May 13 22:23:34 2025 +0900" }, "message": "tracing: probes: Fix a possible race in trace_probe_log APIs\n\nSince the shared trace_probe_log variable can be accessed and\nmodified via probe event create operation of kprobe_events,\nuprobe_events, and dynamic_events, it should be protected.\nIn the dynamic_events, all operations are serialized by\n`dyn_event_ops_mutex`. But kprobe_events and uprobe_events\ninterfaces are not serialized.\n\nTo solve this issue, introduces dyn_event_create(), which runs\ncreate() operation under the mutex, for kprobe_events and\nuprobe_events. This also uses lockdep to check the mutex is\nheld when using trace_probe_log* APIs.\n\nLink: https://lore.kernel.org/all/174684868120.551552.3068655787654268804.stgit@devnote2/\n\nReported-by: Paul Cacheux \u003cpaulcacheux@gmail.com\u003e\nCloses: https://lore.kernel.org/all/20250510074456.805a16872b591e2971a4d221@kernel.org/\nFixes: ab105a4fb894 (\"tracing: Use tracing error_log with probe events\")\nSigned-off-by: Masami Hiramatsu (Google) \u003cmhiramat@kernel.org\u003e\n" }, { "commit": "e9565e23cd89d4d5cd4388f8742130be1d6f182d", "tree": "3ff7a6d85b24ce668411ad115f894c0096e52b51", "parents": [ "d471045e757e941ca802bca3147aab45bf770754", "428dc9fc0873989d73918d4a9cc22745b7bbc799" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon May 12 18:02:05 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon May 12 18:02:05 2025 -0700" }, "message": "Merge tag \u0027sched_ext-for-6.15-rc6-fixes\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext\n\nPull sched_ext fixes from Tejun Heo:\n \"A little bit invasive for rc6 but they\u0027re important fixes, pass tests\n fine and won\u0027t break anything outside sched_ext:\n\n - scx_bpf_cpuperf_set() calls internal functions that require the rq\n to be locked. It assumed that the BPF caller has rq locked but\n that\u0027s not always true. Fix it by tracking whether rq is currently\n held by the CPU and grabbing it if necessary\n\n - bpf_iter_scx_dsq_new() was leaving the DSQ iterator in an\n uninitialized state after an error. However, next() and destroy()\n can be called on an iterator which failed initialization and thus\n they always need to be initialized even after an init error. Fix by\n always initializing the iterator\n\n - Remove duplicate BTF_ID_FLAGS() entries\"\n\n* tag \u0027sched_ext-for-6.15-rc6-fixes\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext:\n sched_ext: bpf_iter_scx_dsq_new() should always initialize iterator\n sched_ext: Fix rq lock state in hotplug ops\n sched_ext: Remove duplicate BTF_ID_FLAGS definitions\n sched_ext: Fix missing rq lock in scx_bpf_cpuperf_set()\n sched_ext: Track currently locked rq\n" }, { "commit": "d471045e757e941ca802bca3147aab45bf770754", "tree": "a309e15f5cda44ee7066e17c68050b400016d4e7", "parents": [ "7a8bdc7fe0a7a6011de68c5ea483de7ecfb2d833", "39b5ef791d109dd54c7c2e6e87933edfcc0ad1ac" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon May 12 17:58:57 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon May 12 17:58:57 2025 -0700" }, "message": "Merge tag \u0027cgroup-for-6.15-rc6-fixes\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup\n\nPull cgroup fix from Tejun Heo:\n \"One low-risk patch to fix a cpuset bug where it over-eagerly tries to\n modify CPU affinity of kernel threads\"\n\n* tag \u0027cgroup-for-6.15-rc6-fixes\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup:\n cgroup/cpuset: Extend kthread_is_per_cpu() check to all PF_NO_SETAFFINITY tasks\n" }, { "commit": "7a8bdc7fe0a7a6011de68c5ea483de7ecfb2d833", "tree": "4c23b632801c1f961de786e26217b42c04104249", "parents": [ "8b64199a7f4fc71d2b8d9fccaab32303a7b6d753", "bfcfe6d335a967f8ea0c1980960e6f0205b5de6e" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon May 12 10:48:02 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon May 12 10:48:02 2025 -0700" }, "message": "Merge tag \u0027platform-drivers-x86-v6.15-5\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/pdx86/platform-drivers-x86\n\nPull x86 platform drivers fixes from Ilpo Järvinen:\n\n - amd/pmc: Use spurious 8042 quirk with MECHREVO Wujie 14XA\n\n - amd/pmf:\n - Ensure Smart PC policies are valid\n - Fix memory leak when the engine fails to start\n\n - amd/hsmp: Make amd_hsmp and hsmp_acpi as mutually exclusive drivers\n\n - asus-wmi: Fix wlan_ctrl_by_user detection\n\n - thinkpad_acpi: Add support for NEC Lavie X1475JAS\n\n* tag \u0027platform-drivers-x86-v6.15-5\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/pdx86/platform-drivers-x86:\n platform/x86: asus-wmi: Fix wlan_ctrl_by_user detection\n platform/x86/amd/pmc: Declare quirk_spurious_8042 for MECHREVO Wujie 14XA (GX4HRXL)\n platform/x86: thinkpad_acpi: Support also NEC Lavie X1475JAS\n platform/x86/amd/hsmp: Make amd_hsmp and hsmp_acpi as mutually exclusive drivers\n drivers/platform/x86/amd: pmf: Check for invalid Smart PC Policies\n drivers/platform/x86/amd: pmf: Check for invalid sideloaded Smart PC Policies\n" }, { "commit": "8b64199a7f4fc71d2b8d9fccaab32303a7b6d753", "tree": "ad589945d7eba2ab0befeba6ef1c34b0a58decea", "parents": [ "e238e49b18ee1bcbe4de952d06631dd3beada097", "55dd5b4db3bf04cf077a8d1712f6295d4517c337" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon May 12 10:23:20 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon May 12 10:23:20 2025 -0700" }, "message": "Merge tag \u0027udf_for_v6.15-rc7\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs\n\nPull UDF fix from Jan Kara:\n \"Fix a bug in UDF inode eviction leading to spewing pointless\n error messages\"\n\n* tag \u0027udf_for_v6.15-rc7\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:\n udf: Make sure i_lenExtents is uptodate on inode eviction\n" }, { "commit": "1b0c192c92ea1fe2dcb178f84adf15fe37c3e7c8", "tree": "b6602bba72154f420b489f37d18eb60dda80f33b", "parents": [ "82f2b0b97b36ee3fcddf0f0780a9a0825d52fec3" ], "author": { "name": "Steven Rostedt", "email": "rostedt@goodmis.org", "time": "Fri May 09 15:26:57 2025 -0400" }, "committer": { "name": "Steven Rostedt (Google)", "email": "rostedt@goodmis.org", "time": "Mon May 12 13:07:22 2025 -0400" }, "message": "tracing: samples: Initialize trace_array_printk() with the correct function\n\nWhen using trace_array_printk() on a created instance, the correct\nfunction to use to initialize it is:\n\n trace_array_init_printk()\n\nNot\n\n trace_printk_init_buffer()\n\nThe former is a proper function to use, the latter is for initializing\ntrace_printk() and causes the NOTICE banner to be displayed.\n\nCc: stable@vger.kernel.org\nCc: Masami Hiramatsu \u003cmhiramat@kernel.org\u003e\nCc: Mathieu Desnoyers \u003cmathieu.desnoyers@efficios.com\u003e\nCc: Divya Indi \u003cdivya.indi@oracle.com\u003e\nLink: https://lore.kernel.org/20250509152657.0f6744d9@gandalf.local.home\nFixes: 89ed42495ef4a (\"tracing: Sample module to demonstrate kernel access to Ftrace instances.\")\nFixes: 38ce2a9e33db6 (\"tracing: Add trace_array_init_printk() to initialize instance trace_printk() buffers\")\nSigned-off-by: Steven Rostedt (Google) \u003crostedt@goodmis.org\u003e\n" }, { "commit": "e238e49b18ee1bcbe4de952d06631dd3beada097", "tree": "90eac6691c4c0a6f9a85f30b2f84a271785d07c4", "parents": [ "627277ba7c2398dc4f95cc9be8222bb2d9477800", "04679f3c27e132c1a2d3881de2f0c5d7128de7c1" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon May 12 10:04:14 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon May 12 10:04:14 2025 -0700" }, "message": "Merge tag \u0027vfs-6.15-rc7.fixes\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull vfs fixes from Christian Brauner:\n\n - Ensure that simple_xattr_list() always includes security.* xattrs\n\n - Fix eventpoll busy loop optimization when combined with timeouts\n\n - Disable swapon() for devices with block sizes greater than page sizes\n\n - Don\u0027t call errseq_set() twice during mark_buffer_write_io_error().\n Just use mapping_set_error() which takes care to not deference\n unconditionally\n\n* tag \u0027vfs-6.15-rc7.fixes\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:\n fs: Remove redundant errseq_set call in mark_buffer_write_io_error.\n swapfile: disable swapon for bs \u003e ps devices\n fs/eventpoll: fix endless busy loop after timeout has expired\n fs/xattr.c: fix simple_xattr_list to always include security.* xattrs\n" }, { "commit": "627277ba7c2398dc4f95cc9be8222bb2d9477800", "tree": "b8d35d751139d512520fd9e5a17b9a9db3c4b928", "parents": [ "6f5bf947bab06f37ff931c359fd5770c4d9cbf87", "efe676a1a7554219eae0b0dcfe1e0cdcc9ef9aef" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun May 11 17:45:00 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun May 11 17:45:00 2025 -0700" }, "message": "Merge tag \u0027arm64_cbpf_mitigation_2025_05_08\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux\n\nPull arm64 cBPF BHB mitigation from James Morse:\n \"This adds the BHB mitigation into the code JITted for cBPF programs as\n these can be loaded by unprivileged users via features like seccomp.\n\n The existing mechanisms to disable the BHB mitigation will also\n prevent the mitigation being JITted. In addition, cBPF programs loaded\n by processes with the SYS_ADMIN capability are not mitigated as these\n could equally load an eBPF program that does the same thing.\n\n For good measure, the list of \u0027k\u0027 values for CPU\u0027s local mitigations\n is updated from the version on arm\u0027s website\"\n\n* tag \u0027arm64_cbpf_mitigation_2025_05_08\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:\n arm64: proton-pack: Add new CPUs \u0027k\u0027 values for branch mitigation\n arm64: bpf: Only mitigate cBPF programs loaded by unprivileged users\n arm64: bpf: Add BHB mitigation to the epilogue for cBPF programs\n arm64: proton-pack: Expose whether the branchy loop k value\n arm64: proton-pack: Expose whether the platform is mitigated by firmware\n arm64: insn: Add support for encoding DSB\n" }, { "commit": "6f5bf947bab06f37ff931c359fd5770c4d9cbf87", "tree": "a0a0cba694124523f4d9292f1f74bb94c5328292", "parents": [ "caf12fa9c066bb81e6a2f05dc441a89a1160c0fe", "7a9b709e7cc5ce1ffb84ce07bf6d157e1de758df" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun May 11 17:23:03 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun May 11 17:23:03 2025 -0700" }, "message": "Merge tag \u0027its-for-linus-20250509\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull x86 ITS mitigation from Dave Hansen:\n \"Mitigate Indirect Target Selection (ITS) issue.\n\n I\u0027d describe this one as a good old CPU bug where the behavior is\n _obviously_ wrong, but since it just results in bad predictions it\n wasn\u0027t wrong enough to notice. Well, the researchers noticed and also\n realized that thus bug undermined a bunch of existing indirect branch\n mitigations.\n\n Thus the unusually wide impact on this one. Details:\n\n ITS is a bug in some Intel CPUs that affects indirect branches\n including RETs in the first half of a cacheline. Due to ITS such\n branches may get wrongly predicted to a target of (direct or indirect)\n branch that is located in the second half of a cacheline. Researchers\n at VUSec found this behavior and reported to Intel.\n\n Affected processors:\n\n - Cascade Lake, Cooper Lake, Whiskey Lake V, Coffee Lake R, Comet\n Lake, Ice Lake, Tiger Lake and Rocket Lake.\n\n Scope of impact:\n\n - Guest/host isolation:\n\n When eIBRS is used for guest/host isolation, the indirect branches\n in the VMM may still be predicted with targets corresponding to\n direct branches in the guest.\n\n - Intra-mode using cBPF:\n\n cBPF can be used to poison the branch history to exploit ITS.\n Realigning the indirect branches and RETs mitigates this attack\n vector.\n\n - User/kernel:\n\n With eIBRS enabled user/kernel isolation is *not* impacted by ITS.\n\n - Indirect Branch Prediction Barrier (IBPB):\n\n Due to this bug indirect branches may be predicted with targets\n corresponding to direct branches which were executed prior to IBPB.\n This will be fixed in the microcode.\n\n Mitigation:\n\n As indirect branches in the first half of cacheline are affected, the\n mitigation is to replace those indirect branches with a call to thunk that\n is aligned to the second half of the cacheline.\n\n RETs that take prediction from RSB are not affected, but they may be\n affected by RSB-underflow condition. So, RETs in the first half of\n cacheline are also patched to a return thunk that executes the RET aligned\n to second half of cacheline\"\n\n* tag \u0027its-for-linus-20250509\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n selftest/x86/bugs: Add selftests for ITS\n x86/its: FineIBT-paranoid vs ITS\n x86/its: Use dynamic thunks for indirect branches\n x86/ibt: Keep IBT disabled during alternative patching\n mm/execmem: Unify early execmem_cache behaviour\n x86/its: Align RETs in BHB clear sequence to avoid thunking\n x86/its: Add support for RSB stuffing mitigation\n x86/its: Add \"vmexit\" option to skip mitigation on some CPUs\n x86/its: Enable Indirect Target Selection mitigation\n x86/its: Add support for ITS-safe return thunk\n x86/its: Add support for ITS-safe indirect thunk\n x86/its: Enumerate Indirect Target Selection (ITS) bug\n Documentation: x86/bugs/its: Add ITS documentation\n" }, { "commit": "caf12fa9c066bb81e6a2f05dc441a89a1160c0fe", "tree": "9795b983ab97f90fcf9a12df1eb1ca065ddae0a5", "parents": [ "82f2b0b97b36ee3fcddf0f0780a9a0825d52fec3", "073fdbe02c69c43fb7c0d547ec265c7747d4a646" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun May 11 17:17:06 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun May 11 17:17:06 2025 -0700" }, "message": "Merge tag \u0027ibti-hisory-for-linus-2025-05-06\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull x86 IBTI mitigation from Dave Hansen:\n \"Mitigate Intra-mode Branch History Injection via classic BFP programs\n\n This adds the branch history clearing mitigation to cBPF programs for\n x86. Intra-mode BHI attacks via cBPF a.k.a IBTI-History was reported\n by researchers at VUSec.\n\n For hardware that doesn\u0027t support BHI_DIS_S, the recommended\n mitigation is to run the short software sequence followed by the IBHF\n instruction after cBPF execution. On hardware that does support\n BHI_DIS_S, enable BHI_DIS_S and execute the IBHF after cBPF execution.\n\n The Indirect Branch History Fence (IBHF) is a new instruction that\n prevents indirect branch target predictions after the barrier from\n using branch history from before the barrier while BHI_DIS_S is\n enabled. On older systems this will map to a NOP. It is recommended to\n add this fence at the end of the cBPF program to support VM migration.\n This instruction is required on newer parts with BHI_NO to fully\n mitigate against these attacks.\n\n The current code disables the mitigation for anything running with the\n SYS_ADMIN capability bit set. The intention was not to waste time\n mitigating a process that has access to anything it wants anyway\"\n\n* tag \u0027ibti-hisory-for-linus-2025-05-06\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n x86/bhi: Do not set BHI_DIS_S in 32-bit mode\n x86/bpf: Add IBHF call at end of classic BPF\n x86/bpf: Call branch history clearing sequence on exit\n" }, { "commit": "82f2b0b97b36ee3fcddf0f0780a9a0825d52fec3", "tree": "aa3f6d2aac782405d894ba4f2decdb4c8a2557ae", "parents": [ "cd802e7e5f1e77ae68cd98653fb70a97189eb937" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun May 11 14:54:11 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun May 11 14:54:11 2025 -0700" }, "message": "Linux 6.15-rc6\n" }, { "commit": "cd802e7e5f1e77ae68cd98653fb70a97189eb937", "tree": "cc6a0db1c05b569d7060c907a773bd89541a7ead", "parents": [ "ecb9194d1724d1265c8cfe50fcb1dad718476e5e", "add20321af2f882ad18716a2fb7b2ce861963f76" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun May 11 11:30:13 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun May 11 11:30:13 2025 -0700" }, "message": "Merge tag \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/virt/kvm/kvm\n\nPull KVM fixes from Paolo Bonzini:\n \"ARM:\n\n - Avoid use of uninitialized memcache pointer in user_mem_abort()\n\n - Always set HCR_EL2.xMO bits when running in VHE, allowing\n interrupts to be taken while TGE\u003d0 and fixing an ugly bug on\n AmpereOne that occurs when taking an interrupt while clearing the\n xMO bits (AC03_CPU_36)\n\n - Prevent VMMs from hiding support for AArch64 at any EL virtualized\n by KVM\n\n - Save/restore the host value for HCRX_EL2 instead of restoring an\n incorrect fixed value\n\n - Make host_stage2_set_owner_locked() check that the entire requested\n range is memory rather than just the first page\n\n RISC-V:\n\n - Add missing reset of smstateen CSRs\n\n x86:\n\n - Forcibly leave SMM on SHUTDOWN interception on AMD CPUs to avoid\n causing problems due to KVM stuffing INIT on SHUTDOWN (KVM needs to\n sanitize the VMCB as its state is undefined after SHUTDOWN,\n emulating INIT is the least awful choice).\n\n - Track the valid sync/dirty fields in kvm_run as a u64 to ensure KVM\n KVM doesn\u0027t goof a sanity check in the future.\n\n - Free obsolete roots when (re)loading the MMU to fix a bug where\n pre-faulting memory can get stuck due to always encountering a\n stale root.\n\n - When dumping GHCB state, use KVM\u0027s snapshot instead of the raw GHCB\n page to print state, so that KVM doesn\u0027t print stale/wrong\n information.\n\n - When changing memory attributes (e.g. shared \u003c\u003d\u003e private), add\n potential hugepage ranges to the mmu_invalidate_range_{start,end}\n set so that KVM doesn\u0027t create a shared/private hugepage when the\n the corresponding attributes will become mixed (the attributes are\n commited *after* KVM finishes the invalidation).\n\n - Rework the SRSO mitigation to enable BP_SPEC_REDUCE only when KVM\n has at least one active VM. Effectively BP_SPEC_REDUCE when KVM is\n loaded led to very measurable performance regressions for non-KVM\n workloads\"\n\n* tag \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/virt/kvm/kvm:\n KVM: SVM: Set/clear SRSO\u0027s BP_SPEC_REDUCE on 0 \u003c\u003d\u003e 1 VM count transitions\n KVM: arm64: Fix memory check in host_stage2_set_owner_locked()\n KVM: arm64: Kill HCRX_HOST_FLAGS\n KVM: arm64: Properly save/restore HCRX_EL2\n KVM: arm64: selftest: Don\u0027t try to disable AArch64 support\n KVM: arm64: Prevent userspace from disabling AArch64 support at any virtualisable EL\n KVM: arm64: Force HCR_EL2.xMO to 1 at all times in VHE mode\n KVM: arm64: Fix uninitialized memcache pointer in user_mem_abort()\n KVM: x86/mmu: Prevent installing hugepages when mem attributes are changing\n KVM: SVM: Update dump_ghcb() to use the GHCB snapshot fields\n KVM: RISC-V: reset smstateen CSRs\n KVM: x86/mmu: Check and free obsolete roots in kvm_mmu_reload()\n KVM: x86: Check that the high 32bits are clear in kvm_arch_vcpu_ioctl_run()\n KVM: SVM: Forcibly leave SMM mode on SHUTDOWN interception\n" }, { "commit": "ecb9194d1724d1265c8cfe50fcb1dad718476e5e", "tree": "058e633a7e65880c9938f66657eadd6da2ebd618", "parents": [ "b9e62a2b8ffd2bc39b5838e163c41d8a08dae19a", "c44572e0cc13c9afff83fd333135a0aa9b27ba26" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun May 11 11:19:52 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun May 11 11:19:52 2025 -0700" }, "message": "Merge tag \u0027mips-fixes_6.15_1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/mips/linux\n\nPull MIPS fixes from Thomas Bogendoerfer:\n\n - Fix delayed timers\n\n - Fix NULL pointer deref\n\n - Fix wrong range check\n\n* tag \u0027mips-fixes_6.15_1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/mips/linux:\n MIPS: Fix MAX_REG_OFFSET\n MIPS: CPS: Fix potential NULL pointer dereferences in cps_prepare_cpus()\n MIPS: rename rollback_handler with skipover_handler\n MIPS: Move r4k_wait() to .cpuidle.text section\n MIPS: Fix idle VS timer enqueue\n" }, { "commit": "b9e62a2b8ffd2bc39b5838e163c41d8a08dae19a", "tree": "2123eacf3ead30ad69d3d787f320b9d6b086f852", "parents": [ "ac814cbbab21ad0a1ba1554312687e4abbfcfc96", "5214a9f6c0f56644acb9d2cbb58facf1856d322b" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun May 11 11:08:55 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun May 11 11:08:55 2025 -0700" }, "message": "Merge tag \u0027x86-urgent-2025-05-11\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull x86 fix from Ingo Molnar:\n \"Fix a boot regression on very old x86 CPUs without CPUID support\"\n\n* tag \u0027x86-urgent-2025-05-11\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n x86/microcode: Consolidate the loader enablement checking\n" }, { "commit": "ac814cbbab21ad0a1ba1554312687e4abbfcfc96", "tree": "6e374b49ba87e07d5cbe2ce83199e535e0d4dc61", "parents": [ "fea9123979fe892f41ca39d9d0226c1e193b0880", "94cff94634e506a4a44684bee1875d2dbf782722" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun May 11 10:33:25 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun May 11 10:33:25 2025 -0700" }, "message": "Merge tag \u0027timers-urgent-2025-05-11\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull misc timers fixes from Ingo Molnar:\n\n - Fix time keeping bugs in CLOCK_MONOTONIC_COARSE clocks\n\n - Work around absolute relocations into vDSO code that GCC erroneously\n emits in certain arm64 build environments\n\n - Fix a false positive lockdep warning in the i8253 clocksource driver\n\n* tag \u0027timers-urgent-2025-05-11\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n clocksource/i8253: Use raw_spinlock_irqsave() in clockevent_i8253_disable()\n arm64: vdso: Work around invalid absolute relocations from GCC\n timekeeping: Prevent coarse clocks going backwards\n" }, { "commit": "fea9123979fe892f41ca39d9d0226c1e193b0880", "tree": "79d3720d4ddf4abe18040107470fbc8846b6cd5b", "parents": [ "f717acc6e9977287641c84af4e397754c8403838", "93406e9d024058b3bf487656ddd0ac552e5a366e" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun May 11 10:29:29 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun May 11 10:29:29 2025 -0700" }, "message": "Merge tag \u0027input-for-v6.15-rc5\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input\n\nPull input fixes from Dmitry Torokhov:\n\n - Synaptics touchpad on multiple laptops (Dynabook Portege X30L-G,\n Dynabook Portege X30-D, TUXEDO InfinityBook Pro 14 v5, Dell Precision\n M3800, HP Elitebook 850 G1) switched from PS/2 to SMBus mode\n\n - a number of new controllers added to xpad driver: HORI Drum\n controller, PowerA Fusion Pro 4, PowerA MOGA XP-Ultra controller,\n 8BitDo Ultimate 2 Wireless Controller, 8BitDo Ultimate 3-mode\n Controller, Hyperkin DuchesS Xbox One controller\n\n - fixes to xpad driver to properly handle Mad Catz JOYTECH NEO SE\n Advanced and PDP Mirror\u0027s Edge Official controllers\n\n - fixes to xpad driver to properly handle \"Share\" button on some\n controllers\n\n - a fix for device initialization timing and for waking up the\n controller in cyttsp5 driver\n\n - a fix for hisi_powerkey driver to properly wake up from s2idle state\n\n - other assorted cleanups and fixes\n\n* tag \u0027input-for-v6.15-rc5\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:\n Input: xpad - fix xpad_device sorting\n Input: xpad - add support for several more controllers\n Input: xpad - fix Share button on Xbox One controllers\n Input: xpad - fix two controller table values\n Input: hisi_powerkey - enable system-wakeup for s2idle\n Input: synaptics - enable InterTouch on Dell Precision M3800\n Input: synaptics - enable InterTouch on TUXEDO InfinityBook Pro 14 v5\n Input: synaptics - enable InterTouch on Dynabook Portege X30L-G\n Input: synaptics - enable InterTouch on Dynabook Portege X30-D\n Input: synaptics - enable SMBus for HP Elitebook 850 G1\n Input: mtk-pmic-keys - fix possible null pointer dereference\n Input: xpad - add support for 8BitDo Ultimate 2 Wireless Controller\n Input: cyttsp5 - fix power control issue on wakeup\n MAINTAINERS: .mailmap: update Mattijs Korpershoek\u0027s email address\n dt-bindings: mediatek,mt6779-keypad: Update Mattijs\u0027 email address\n Input: stmpe-ts - use module alias instead of device table\n Input: cyttsp5 - ensure minimum reset pulse width\n Input: sparcspkr - avoid unannotated fall-through\n input/joystick: magellan: Mark __nonstring look-up table\n" }, { "commit": "f717acc6e9977287641c84af4e397754c8403838", "tree": "cd4b9a62adc2baa73983f46ed2cf3d7376859edd", "parents": [ "3ce9925823c7d6bb0e6eb951bf2db0e9e182582d", "da8bf5daa5e55a6af2b285ecda460d6454712ff4" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun May 11 10:23:53 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun May 11 10:23:53 2025 -0700" }, "message": "Merge tag \u0027fixes-2025-05-11\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/rppt/memblock\n\nPull memblock fixes from Mike Rapoport:\n\n - Mark set_high_memory() as __init to fix section mismatch\n\n - Accept memory allocated in memblock_double_array() to mitigate crash\n of SNP guests\n\n* tag \u0027fixes-2025-05-11\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/rppt/memblock:\n memblock: Accept allocated memory before use in memblock_double_array()\n mm,mm_init: Mark set_high_memory as __init\n" }, { "commit": "93406e9d024058b3bf487656ddd0ac552e5a366e", "tree": "86a842cb413ef9d6bd0eeff94cd6a09d870cdd0a", "parents": [ "7026d23cb383120712f3a214b0b33ca349cd21a0" ], "author": { "name": "Vicki Pfau", "email": "vi@endrift.com", "time": "Sat May 10 23:06:34 2025 -0700" }, "committer": { "name": "Dmitry Torokhov", "email": "dmitry.torokhov@gmail.com", "time": "Sat May 10 23:08:19 2025 -0700" }, "message": "Input: xpad - fix xpad_device sorting\n\nA recent commit put one entry in the wrong place. This just moves it to the\nright place.\n\nSigned-off-by: Vicki Pfau \u003cvi@endrift.com\u003e\nLink: https://lore.kernel.org/r/20250328234345.989761-5-vi@endrift.com\nSigned-off-by: Dmitry Torokhov \u003cdmitry.torokhov@gmail.com\u003e\n" }, { "commit": "7026d23cb383120712f3a214b0b33ca349cd21a0", "tree": "5553840759f1e0a3f89ee55640e89ed8dfdf56e8", "parents": [ "4ef46367073b107ec22f46fe5f12176e87c238e8" ], "author": { "name": "Vicki Pfau", "email": "vi@endrift.com", "time": "Sat May 10 23:00:10 2025 -0700" }, "committer": { "name": "Dmitry Torokhov", "email": "dmitry.torokhov@gmail.com", "time": "Sat May 10 23:08:18 2025 -0700" }, "message": "Input: xpad - add support for several more controllers\n\nThis adds support for several new controllers, all of which include\nShare buttons:\n\n- HORI Drum controller\n- PowerA Fusion Pro 4\n- 8BitDo Ultimate 3-mode Controller\n- Hyperkin DuchesS Xbox One controller\n- PowerA MOGA XP-Ultra controller\n\nSigned-off-by: Vicki Pfau \u003cvi@endrift.com\u003e\nLink: https://lore.kernel.org/r/20250328234345.989761-4-vi@endrift.com\nCc: stable@vger.kernel.org\nSigned-off-by: Dmitry Torokhov \u003cdmitry.torokhov@gmail.com\u003e\n" }, { "commit": "4ef46367073b107ec22f46fe5f12176e87c238e8", "tree": "3f07007dea1d7f1c24455c9631a286a624943118", "parents": [ "d05a424bea9aa3435009d5c462055008cc1545d8" ], "author": { "name": "Vicki Pfau", "email": "vi@endrift.com", "time": "Sat May 10 22:59:25 2025 -0700" }, "committer": { "name": "Dmitry Torokhov", "email": "dmitry.torokhov@gmail.com", "time": "Sat May 10 23:08:17 2025 -0700" }, "message": "Input: xpad - fix Share button on Xbox One controllers\n\nThe Share button, if present, is always one of two offsets from the end of the\nfile, depending on the presence of a specific interface. As we lack parsing for\nthe identify packet we can\u0027t automatically determine the presence of that\ninterface, but we can hardcode which of these offsets is correct for a given\ncontroller.\n\nMore controllers are probably fixable by adding the MAP_SHARE_BUTTON in the\nfuture, but for now I only added the ones that I have the ability to test\ndirectly.\n\nSigned-off-by: Vicki Pfau \u003cvi@endrift.com\u003e\nLink: https://lore.kernel.org/r/20250328234345.989761-2-vi@endrift.com\nCc: stable@vger.kernel.org\nSigned-off-by: Dmitry Torokhov \u003cdmitry.torokhov@gmail.com\u003e\n" }, { "commit": "d05a424bea9aa3435009d5c462055008cc1545d8", "tree": "cf0f1fc7287ebdfccc20244b75d337387c9b6bfb", "parents": [ "e98960bc4df931a2f4fa88f7e55370699ca4dd82" ], "author": { "name": "Vicki Pfau", "email": "vi@endrift.com", "time": "Fri Mar 28 16:43:36 2025 -0700" }, "committer": { "name": "Dmitry Torokhov", "email": "dmitry.torokhov@gmail.com", "time": "Sat May 10 22:54:18 2025 -0700" }, "message": "Input: xpad - fix two controller table values\n\nTwo controllers -- Mad Catz JOYTECH NEO SE Advanced and PDP Mirror\u0027s\nEdge Official -- were missing the value of the mapping field, and thus\nwouldn\u0027t detect properly.\n\nSigned-off-by: Vicki Pfau \u003cvi@endrift.com\u003e\nLink: https://lore.kernel.org/r/20250328234345.989761-1-vi@endrift.com\nFixes: 540602a43ae5 (\"Input: xpad - add a few new VID/PID combinations\")\nFixes: 3492321e2e60 (\"Input: xpad - add multiple supported devices\")\nCc: stable@vger.kernel.org\nSigned-off-by: Dmitry Torokhov \u003cdmitry.torokhov@gmail.com\u003e\n" }, { "commit": "e98960bc4df931a2f4fa88f7e55370699ca4dd82", "tree": "a83d08280594a6508da23d06397913c1208b1ac4", "parents": [ "a609cb4cc07aa9ab8f50466622814356c06f2c17" ], "author": { "name": "Ulf Hansson", "email": "ulf.hansson@linaro.org", "time": "Thu Mar 06 12:50:21 2025 +0100" }, "committer": { "name": "Dmitry Torokhov", "email": "dmitry.torokhov@gmail.com", "time": "Sat May 10 22:26:33 2025 -0700" }, "message": "Input: hisi_powerkey - enable system-wakeup for s2idle\n\nTo wake up the system from s2idle when pressing the power-button, let\u0027s\nconvert from using pm_wakeup_event() to pm_wakeup_dev_event(), as it allows\nus to specify the \"hard\" in-parameter, which needs to be set for s2idle.\n\nSigned-off-by: Ulf Hansson \u003culf.hansson@linaro.org\u003e\nLink: https://lore.kernel.org/r/20250306115021.797426-1-ulf.hansson@linaro.org\nSigned-off-by: Dmitry Torokhov \u003cdmitry.torokhov@gmail.com\u003e\n" }, { "commit": "3ce9925823c7d6bb0e6eb951bf2db0e9e182582d", "tree": "af9d8c84f0a376d560d677be23cdc10fa1491819", "parents": [ "345030986df8f7712f9e4c00fe61e145c8984ef3", "7b08b74f3d99f6b801250683c751d391128799ec" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 10 15:50:56 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 10 15:50:56 2025 -0700" }, "message": "Merge tag \u0027mm-hotfixes-stable-2025-05-10-14-23\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm\n\nPull misc hotfixes from Andrew Morton:\n \"22 hotfixes. 13 are cc:stable and the remainder address post-6.14\n issues or aren\u0027t considered necessary for -stable kernels.\n\n About half are for MM. Five OCFS2 fixes and a few MAINTAINERS updates\"\n\n* tag \u0027mm-hotfixes-stable-2025-05-10-14-23\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm: (22 commits)\n mm: fix folio_pte_batch() on XEN PV\n nilfs2: fix deadlock warnings caused by lock dependency in init_nilfs()\n mm/hugetlb: copy the CMA flag when demoting\n mm, swap: fix false warning for large allocation with !THP_SWAP\n selftests/mm: fix a build failure on powerpc\n selftests/mm: fix build break when compiling pkey_util.c\n mm: vmalloc: support more granular vrealloc() sizing\n tools/testing/selftests: fix guard region test tmpfs assumption\n ocfs2: stop quota recovery before disabling quotas\n ocfs2: implement handshaking with ocfs2 recovery thread\n ocfs2: switch osb-\u003edisable_recovery to enum\n mailmap: map Uwe\u0027s BayLibre addresses to a single one\n MAINTAINERS: add mm THP section\n mm/userfaultfd: fix uninitialized output field for -EAGAIN race\n selftests/mm: compaction_test: support platform with huge mount of memory\n MAINTAINERS: add core mm section\n ocfs2: fix panic in failed foilio allocation\n mm/huge_memory: fix dereferencing invalid pmd migration entry\n MAINTAINERS: add reverse mapping section\n x86: disable image size check for test builds\n ...\n" }, { "commit": "345030986df8f7712f9e4c00fe61e145c8984ef3", "tree": "62719ce1861b3b024c46267cb53db1ef1c5b3308", "parents": [ "bec6f00f120ea68ba584def5b7416287e7dd29a7", "95deee37a12364f410d22c6a8383f59738a2fef3" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 10 09:53:11 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 10 09:53:11 2025 -0700" }, "message": "Merge tag \u0027driver-core-6.15-rc6\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/driver-core/driver-core\n\nPull driver core fix from Greg KH:\n \"Here is a single driver core fix for a regression for platform devices\n that is a regression from a change that went into 6.15-rc1 that\n affected Pixel devices. It has been in linux-next for over a week with\n no reported problems\"\n\n* tag \u0027driver-core-6.15-rc6\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/driver-core/driver-core:\n platform: Fix race condition during DMA configure at IOMMU probe time\n" }, { "commit": "bec6f00f120ea68ba584def5b7416287e7dd29a7", "tree": "05d2af6caaecdd19f66b55a638da72dbff254608", "parents": [ "067dd5875d9cdc86c8db6671c695be4f72692196", "cab63934c33b12c0d1e9f4da7450928057f2c142" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 10 09:18:05 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 10 09:18:05 2025 -0700" }, "message": "Merge tag \u0027usb-6.15-rc6\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb\n\nPull USB fixes from Greg KH:\n \"Here are some small USB driver fixes for 6.15-rc6. Included in here\n are:\n\n - typec driver fixes\n\n - usbtmc ioctl fixes\n\n - xhci driver fixes\n\n - cdnsp driver fixes\n\n - some gadget driver fixes\n\n Nothing really major, just all little stuff that people have reported\n being issues. All of these have been in linux-next this week with no\n reported issues\"\n\n* tag \u0027usb-6.15-rc6\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb:\n xhci: dbc: Avoid event polling busyloop if pending rx transfers are inactive.\n usb: xhci: Don\u0027t trust the EP Context cycle bit when moving HW dequeue\n usb: usbtmc: Fix erroneous generic_read ioctl return\n usb: usbtmc: Fix erroneous wait_srq ioctl return\n usb: usbtmc: Fix erroneous get_stb ioctl error returns\n usb: typec: tcpm: delay SNK_TRY_WAIT_DEBOUNCE to SRC_TRYWAIT transition\n USB: usbtmc: use interruptible sleep in usbtmc_read\n usb: cdnsp: fix L1 resume issue for RTL_REVISION_NEW_LPM version\n usb: typec: ucsi: displayport: Fix NULL pointer access\n usb: typec: ucsi: displayport: Fix deadlock\n usb: misc: onboard_usb_dev: fix support for Cypress HX3 hubs\n usb: uhci-platform: Make the clock really optional\n usb: dwc3: gadget: Make gadget_wakeup asynchronous\n usb: gadget: Use get_status callback to set remote wakeup capability\n usb: gadget: f_ecm: Add get_status callback\n usb: host: tegra: Prevent host controller crash when OTG port is used\n usb: cdnsp: Fix issue with resuming from L1\n usb: gadget: tegra-xudc: ACK ST_RC after clearing CTRL_RUN\n" }, { "commit": "067dd5875d9cdc86c8db6671c695be4f72692196", "tree": "db87c8bd0e48a4c364ff11386f6bb217c065c016", "parents": [ "914a1fe5f818bff5abb0e792bc0addf0c7c274f4", "c6e8d85fafa7193613db37da29c0e8d6e2515b13" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 10 09:08:19 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 10 09:08:19 2025 -0700" }, "message": "Merge tag \u0027staging-6.15-rc6\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging\n\nPull staging driver fixes from Greg KH:\n \"Here are three small staging driver fixes for 6.15-rc6. These are:\n\n - bcm2835-camera driver fix\n\n - two axis-fifo driver fixes\n\n All of these have been in linux-next for a few weeks with no reported\n issues\"\n\n* tag \u0027staging-6.15-rc6\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging:\n staging: axis-fifo: Remove hardware resets for user errors\n staging: axis-fifo: Correct handling of tx_fifo_depth for size validation\n staging: bcm2835-camera: Initialise dev in v4l2_dev\n" }, { "commit": "914a1fe5f818bff5abb0e792bc0addf0c7c274f4", "tree": "82ed26a1674044494647f818594d310e8caa898e", "parents": [ "ed36b437d9b16b073f53f3a61550df0e2e987b52", "65995e97a1caacf0024bebda3332b8d1f0f443c4" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 10 08:55:15 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 10 08:55:15 2025 -0700" }, "message": "Merge tag \u0027char-misc-6.15-rc6\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc\n\nPull char/misc/IIO driver fixes from Greg KH:\n \"Here are a bunch of small driver fixes (mostly all IIO) for 6.15-rc6.\n Included in here are:\n\n - loads of tiny IIO driver fixes for reported issues\n\n - hyperv driver fix for a much-reported and worked on sysfs ring\n buffer creation bug\n\n All of these have been in linux-next for over a week (the IIO ones for\n many weeks now), with no reported issues\"\n\n* tag \u0027char-misc-6.15-rc6\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc: (30 commits)\n Drivers: hv: Make the sysfs node size for the ring buffer dynamic\n uio_hv_generic: Fix sysfs creation path for ring buffer\n iio: adis16201: Correct inclinometer channel resolution\n iio: adc: ad7606: fix serial register access\n iio: pressure: mprls0025pa: use aligned_s64 for timestamp\n iio: imu: adis16550: align buffers for timestamp\n staging: iio: adc: ad7816: Correct conditional logic for store mode\n iio: adc: ad7266: Fix potential timestamp alignment issue.\n iio: adc: ad7768-1: Fix insufficient alignment of timestamp.\n iio: adc: dln2: Use aligned_s64 for timestamp\n iio: accel: adxl355: Make timestamp 64-bit aligned using aligned_s64\n iio: temp: maxim-thermocouple: Fix potential lack of DMA safe buffer.\n iio: chemical: pms7003: use aligned_s64 for timestamp\n iio: chemical: sps30: use aligned_s64 for timestamp\n iio: imu: inv_mpu6050: align buffer for timestamp\n iio: imu: st_lsm6dsx: Fix wakeup source leaks on device unbind\n iio: adc: qcom-spmi-iadc: Fix wakeup source leaks on device unbind\n iio: accel: fxls8962af: Fix wakeup source leaks on device unbind\n iio: adc: ad7380: fix event threshold shift\n iio: hid-sensor-prox: Fix incorrect OFFSET calculation\n ...\n" }, { "commit": "ed36b437d9b16b073f53f3a61550df0e2e987b52", "tree": "cd757f9f9c05b8ce15dedc8a822142f953a28e85", "parents": [ "86c019597cd4e0fc90dfa9ebba9282b2d122c187", "b6c08bcddb2af9ea6d1846bcd36a35371e249003" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 10 08:52:41 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 10 08:52:41 2025 -0700" }, "message": "Merge tag \u0027i2c-for-6.15-rc6\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux\n\nPull i2c fixes from Wolfram Sang:\n\n - omap: use correct function to read from device tree\n\n - MAINTAINERS: remove Seth from ISMT maintainership\n\n* tag \u0027i2c-for-6.15-rc6\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:\n MAINTAINERS: Remove entry for Seth Heasley\n i2c: omap: fix deprecated of_property_read_bool() use\n" }, { "commit": "86c019597cd4e0fc90dfa9ebba9282b2d122c187", "tree": "4351e72f5733b4eca4abecd7fc442e93ad86bc38", "parents": [ "acbf235235e2b33a0b73802c14c2a1a28a3a3f04", "1f0304dfd9d217c2f8b04a9ef4b3258a66eedd27" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 10 08:44:36 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 10 08:44:36 2025 -0700" }, "message": "Merge tag \u0027for-linus-6.15a-rc6-tag\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip\n\nPull xen fixes from Juergen Gross:\n\n - A fix for the xenbus driver allowing to use a PVH Dom0 with\n Xenstore running in another domain\n\n - A fix for the xenbus driver addressing a rare race condition\n resulting in NULL dereferences and other problems\n\n - A fix for the xen-swiotlb driver fixing a problem seen on Arm\n platforms\n\n* tag \u0027for-linus-6.15a-rc6-tag\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:\n xenbus: Use kref to track req lifetime\n xenbus: Allow PVH dom0 a non-local xenstore\n xen: swiotlb: Use swiotlb bouncing if kmalloc allocation demands it\n" }, { "commit": "acbf235235e2b33a0b73802c14c2a1a28a3a3f04", "tree": "80a434bb81a7e4f54a9370ffdddf1807fefc6fe2", "parents": [ "1a33418a69cc801d48c59d7d803af5c9cd291be2", "d1ddc6f1d9f0cf887834eb54a5a68bbfeec1bb77" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 10 08:36:07 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 10 08:36:07 2025 -0700" }, "message": "Merge tag \u0027pull-fixes\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs\n\nPull mount fixes from Al Viro:\n \"A couple of races around legalize_mnt vs umount (both fairly old and\n hard to hit) plus two bugs in move_mount(2) - both around \u0027move\n detached subtree in place\u0027 logics\"\n\n* tag \u0027pull-fixes\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:\n fix IS_MNT_PROPAGATING uses\n do_move_mount(): don\u0027t leak MNTNS_PROPAGATING on failures\n do_umount(): add missing barrier before refcount checks in sync case\n __legitimize_mnt(): check for MNT_SYNC_UMOUNT should be under mount_lock\n" }, { "commit": "add20321af2f882ad18716a2fb7b2ce861963f76", "tree": "1502e15f524d22607aac420da4acf624f09b8425", "parents": [ "36867c0e94f3cb8ff6be9e63b5d40c988ad95e27", "e3417ab75ab2e7dca6372a1bfa26b1be3ac5889e" ], "author": { "name": "Paolo Bonzini", "email": "pbonzini@redhat.com", "time": "Sat May 10 11:11:06 2025 -0400" }, "committer": { "name": "Paolo Bonzini", "email": "pbonzini@redhat.com", "time": "Sat May 10 11:11:06 2025 -0400" }, "message": "Merge tag \u0027kvm-x86-fixes-6.15-rcN\u0027 of https://github.com/kvm-x86/linux into HEAD\n\nKVM x86 fixes for 6.15-rcN\n\n - Forcibly leave SMM on SHUTDOWN interception on AMD CPUs to avoid causing\n problems due to KVM stuffing INIT on SHUTDOWN (KVM needs to sanitize the\n VMCB as its state is undefined after SHUTDOWN, emulating INIT is the\n least awful choice).\n\n - Track the valid sync/dirty fields in kvm_run as a u64 to ensure KVM\n KVM doesn\u0027t goof a sanity check in the future.\n\n - Free obsolete roots when (re)loading the MMU to fix a bug where\n pre-faulting memory can get stuck due to always encountering a stale\n root.\n\n - When dumping GHCB state, use KVM\u0027s snapshot instead of the raw GHCB page\n to print state, so that KVM doesn\u0027t print stale/wrong information.\n\n - When changing memory attributes (e.g. shared \u003c\u003d\u003e private), add potential\n hugepage ranges to the mmu_invalidate_range_{start,end} set so that KVM\n doesn\u0027t create a shared/private hugepage when the the corresponding\n attributes will become mixed (the attributes are commited *after* KVM\n finishes the invalidation).\n\n - Rework the SRSO mitigation to enable BP_SPEC_REDUCE only when KVM has at\n least one active VM. Effectively BP_SPEC_REDUCE when KVM is loaded led\n to very measurable performance regressions for non-KVM workloads.\n" }, { "commit": "36867c0e94f3cb8ff6be9e63b5d40c988ad95e27", "tree": "4f478af81d10dea16f2c1c515ea20d4c60012e42", "parents": [ "6a74470b33804998632234ac95653620fdf9753a", "3949e28786cd0afcd96a46ce6629245203f629e5" ], "author": { "name": "Paolo Bonzini", "email": "pbonzini@redhat.com", "time": "Sat May 10 11:10:02 2025 -0400" }, "committer": { "name": "Paolo Bonzini", "email": "pbonzini@redhat.com", "time": "Sat May 10 11:10:02 2025 -0400" }, "message": "Merge tag \u0027kvmarm-fixes-6.15-3\u0027 of https://git.kernel.org/pub/scm/linux/kernel/git/kvmarm/kvmarm into HEAD\n\nKVM/arm64 fixes for 6.15, round #3\n\n - Avoid use of uninitialized memcache pointer in user_mem_abort()\n\n - Always set HCR_EL2.xMO bits when running in VHE, allowing interrupts\n to be taken while TGE\u003d0 and fixing an ugly bug on AmpereOne that\n occurs when taking an interrupt while clearing the xMO bits\n (AC03_CPU_36)\n\n - Prevent VMMs from hiding support for AArch64 at any EL virtualized by\n KVM\n\n - Save/restore the host value for HCRX_EL2 instead of restoring an\n incorrect fixed value\n\n - Make host_stage2_set_owner_locked() check that the entire requested\n range is memory rather than just the first page\n" }, { "commit": "6a74470b33804998632234ac95653620fdf9753a", "tree": "41b51571cd00407597eef74760a04bd2059df385", "parents": [ "92a09c47464d040866cf2b4cd052bc60555185fb", "87ec7d5249bb8ebf40261420da069fa238c21789" ], "author": { "name": "Paolo Bonzini", "email": "pbonzini@redhat.com", "time": "Sat May 10 11:09:26 2025 -0400" }, "committer": { "name": "Paolo Bonzini", "email": "pbonzini@redhat.com", "time": "Sat May 10 11:09:26 2025 -0400" }, "message": "Merge tag \u0027kvm-riscv-fixes-6.15-1\u0027 of https://github.com/kvm-riscv/linux into HEAD\n\nKVM/riscv fixes for 6.15, take #1\n\n- Add missing reset of smstateen CSRs\n" }, { "commit": "b6c08bcddb2af9ea6d1846bcd36a35371e249003", "tree": "b533a327494fc6e2a3c73a93248c58d26c4c56a8", "parents": [ "92a09c47464d040866cf2b4cd052bc60555185fb", "10aba126bc86904e2f5afeaa26354e877a593c95" ], "author": { "name": "Wolfram Sang", "email": "wsa+renesas@sang-engineering.com", "time": "Sat May 10 11:41:13 2025 +0200" }, "committer": { "name": "Wolfram Sang", "email": "wsa+renesas@sang-engineering.com", "time": "Sat May 10 11:41:13 2025 +0200" }, "message": "Merge tag \u0027i2c-host-fixes-6.15-rc6\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/andi.shyti/linux into i2c/for-current\n\ni2c-host-fixes for v6.15-rc6\n\n- omap: use correct function to read from device tree\n- MAINTAINERS: remove Seth from ISMT maintainership\n" }, { "commit": "15eaaa71e8ef142b122942b35653d513cfb90050", "tree": "c51a4db8a01d4410dc693911b66fceddbc686c6b", "parents": [ "09acc3266cdd8e32a62dd6c5052f44724111120b", "c6888983134e2ccc2db8ffd2720b0d4826d952e4" ], "author": { "name": "Arnd Bergmann", "email": "arnd@arndb.de", "time": "Sat May 10 11:10:38 2025 +0200" }, "committer": { "name": "Arnd Bergmann", "email": "arnd@arndb.de", "time": "Sat May 10 11:10:38 2025 +0200" }, "message": "Merge tag \u0027imx-fixes-6.15-2\u0027 of https://git.kernel.org/pub/scm/linux/kernel/git/shawnguo/linux into arm/fixes\n\ni.MX fixes for 6.15, 2nd round:\n\n- One more i.MX8MP nominal drive mode DT fix from Ahmad Fatoum to use\n 800MHz NoC OPP\n- A imx8mp-var-som DT change from Himanshu Bhavani to fix SD card\n timeout caused by LDO5\n\n* tag \u0027imx-fixes-6.15-2\u0027 of https://git.kernel.org/pub/scm/linux/kernel/git/shawnguo/linux:\n arm64: dts: imx8mp-var-som: Fix LDO5 shutdown causing SD card timeout\n arm64: dts: imx8mp: use 800MHz NoC OPP for nominal drive mode\n\nLink: https://lore.kernel.org/r/aB6h/woeyG1bSo12@dragon\nSigned-off-by: Arnd Bergmann \u003carnd@arndb.de\u003e\n" }, { "commit": "1a33418a69cc801d48c59d7d803af5c9cd291be2", "tree": "bc969741739c399d62d83b8e5c288309400a0e75", "parents": [ "0e1329d4045ca3606f9c06a8c47f62e758a09105", "3ca02e63edccb78ef3659bebc68579c7224a6ca2" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri May 09 16:45:21 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri May 09 16:45:21 2025 -0700" }, "message": "Merge tag \u00276.15-rc5-smb3-client-fixes\u0027 of git://git.samba.org/sfrench/cifs-2.6\n\nPull smb client fixes from Steve French:\n\n - Fix dentry leak which can cause umount crash\n\n - Add warning for parse contexts error on compounded operation\n\n* tag \u00276.15-rc5-smb3-client-fixes\u0027 of git://git.samba.org/sfrench/cifs-2.6:\n smb: client: Avoid race in open_cached_dir with lease breaks\n smb3 client: warn when parse contexts returns error on compounded operation\n" }, { "commit": "e41b5af4519f90f9a751805ede2102ae36caf5d0", "tree": "08d11fb39bf1c51405cff192d7d9f2fe9a61958b", "parents": [ "9dda18a32b4a6693fccd3f7c0738af646147b1cf" ], "author": { "name": "Paul Cacheux", "email": "paulcacheux@gmail.com", "time": "Sun May 04 20:27:52 2025 +0200" }, "committer": { "name": "Masami Hiramatsu (Google)", "email": "mhiramat@kernel.org", "time": "Sat May 10 08:44:50 2025 +0900" }, "message": "tracing: add missing trace_probe_log_clear for eprobes\n\nMake sure trace_probe_log_clear is called in the tracing\neprobe code path, matching the trace_probe_log_init call.\n\nLink: https://lore.kernel.org/all/20250504-fix-trace-probe-log-race-v3-1-9e99fec7eddc@gmail.com/\n\nSigned-off-by: Paul Cacheux \u003cpaulcacheux@gmail.com\u003e\nAcked-by: Steven Rostedt (Google) \u003crostedt@goodmis.org\u003e\nSigned-off-by: Masami Hiramatsu (Google) \u003cmhiramat@kernel.org\u003e\n" }, { "commit": "9dda18a32b4a6693fccd3f7c0738af646147b1cf", "tree": "ea4cbfde15cec674d2f6fb306fd95275e633807f", "parents": [ "0e1329d4045ca3606f9c06a8c47f62e758a09105" ], "author": { "name": "Breno Leitao", "email": "leitao@debian.org", "time": "Thu Apr 10 05:22:21 2025 -0700" }, "committer": { "name": "Masami Hiramatsu (Google)", "email": "mhiramat@kernel.org", "time": "Sat May 10 08:28:02 2025 +0900" }, "message": "tracing: fprobe: Fix RCU warning message in list traversal\n\nWhen CONFIG_PROVE_RCU_LIST is enabled, fprobe triggers the following\nwarning:\n\n WARNING: suspicious RCU usage\n kernel/trace/fprobe.c:457 RCU-list traversed in non-reader section!!\n\n other info that might help us debug this:\n\t#1: ffffffff863c4e08 (fprobe_mutex){+.+.}-{4:4}, at: fprobe_module_callback+0x7b/0x8c0\n\n Call Trace:\n\tfprobe_module_callback\n\tnotifier_call_chain\n\tblocking_notifier_call_chain\n\nThis warning occurs because fprobe_remove_node_in_module() traverses an\nRCU list using RCU primitives without holding an RCU read lock. However,\nthe function is only called from fprobe_module_callback(), which holds\nthe fprobe_mutex lock that provides sufficient protection for safely\ntraversing the list.\n\nFix the warning by specifying the locking design to the\nCONFIG_PROVE_RCU_LIST mechanism. Add the lockdep_is_held() argument to\nhlist_for_each_entry_rcu() to inform the RCU checker that fprobe_mutex\nprovides the required protection.\n\nLink: https://lore.kernel.org/all/20250410-fprobe-v1-1-068ef5f41436@debian.org/\n\nFixes: a3dc2983ca7b90 (\"tracing: fprobe: Cleanup fprobe hash when module unloading\")\nSigned-off-by: Breno Leitao \u003cleitao@debian.org\u003e\nTested-by: Antonio Quartulli \u003cantonio@mandelbit.com\u003e\nTested-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\nSigned-off-by: Masami Hiramatsu (Google) \u003cmhiramat@kernel.org\u003e\n" }, { "commit": "d1ddc6f1d9f0cf887834eb54a5a68bbfeec1bb77", "tree": "2aec0e244f081275b8791df0aaa486ac9e4bd6b2", "parents": [ "267fc3a06a37bec30cc5b4d97fb8409102bc7a9d" ], "author": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Thu May 08 15:35:51 2025 -0400" }, "committer": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Fri May 09 18:06:27 2025 -0400" }, "message": "fix IS_MNT_PROPAGATING uses\n\npropagate_mnt() does not attach anything to mounts created during\npropagate_mnt() itself. What\u0027s more, anything on -\u003emnt_slave_list\nof such new mount must also be new, so we don\u0027t need to even look\nthere.\n\nWhen move_mount() had been introduced, we\u0027ve got an additional\nclass of mounts to skip - if we are moving from anon namespace,\nwe do not want to propagate to mounts we are moving (i.e. all\nmounts in that anon namespace).\n\nUnfortunately, the part about \"everything on their -\u003emnt_slave_list\nwill also be ignorable\" is not true - if we have propagation graph\n\tA -\u003e B -\u003e C\nand do OPEN_TREE_CLONE open_tree() of B, we get\n\tA -\u003e [B \u003c-\u003e B\u0027] -\u003e C\nas propagation graph, where B\u0027 is a clone of B in our detached tree.\nMaking B private will result in\n\tA -\u003e B\u0027 -\u003e C\nC still gets propagation from A, as it would after making B private\nif we hadn\u0027t done that open_tree(), but now the propagation goes\nthrough B\u0027. Trying to move_mount() our detached tree on subdirectory\nin A should have\n\t* moved B\u0027 on that subdirectory in A\n\t* skipped the corresponding subdirectory in B\u0027 itself\n\t* copied B\u0027 on the corresponding subdirectory in C.\nAs it is, the logics in propagation_next() and friends ends up\nskipping propagation into C, since it doesn\u0027t consider anything\ndownstream of B\u0027.\n\nIOW, walking the propagation graph should only skip the -\u003emnt_slave_list\nof new mounts; the only places where the check for \"in that one\nanon namespace\" are applicable are propagate_one() (where we should\ntreat that as the same kind of thing as \"mountpoint we are looking\nat is not visible in the mount we are looking at\") and\npropagation_would_overmount(). The latter is better dealt with\nin the caller (can_move_mount_beneath()); on the first call of\npropagation_would_overmount() the test is always false, on the\nsecond it is always true in \"move from anon namespace\" case and\nalways false in \"move within our namespace\" one, so it\u0027s easier\nto just use check_mnt() before bothering with the second call and\nbe done with that.\n\nFixes: 064fe6e233e8 (\"mount: handle mount propagation for detached mount trees\")\nReviewed-by: Christian Brauner \u003cbrauner@kernel.org\u003e\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\n" }, { "commit": "267fc3a06a37bec30cc5b4d97fb8409102bc7a9d", "tree": "2add24a704f1eb17c05b8a6bb3206eeeffa142df", "parents": [ "65781e19dcfcb4aed1167d87a3ffcc2a0c071d47" ], "author": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Mon Apr 28 21:43:23 2025 -0400" }, "committer": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Fri May 09 18:06:10 2025 -0400" }, "message": "do_move_mount(): don\u0027t leak MNTNS_PROPAGATING on failures\n\nas it is, a failed move_mount(2) from anon namespace breaks\nall further propagation into that namespace, including normal\nmounts in non-anon namespaces that would otherwise propagate\nthere.\n\nFixes: 064fe6e233e8 (\"mount: handle mount propagation for detached mount trees\")\nReviewed-by: Christian Brauner \u003cbrauner@kernel.org\u003e\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\n" }, { "commit": "65781e19dcfcb4aed1167d87a3ffcc2a0c071d47", "tree": "bcb00291386cc4f07f1f1eaf15073b6877dea46e", "parents": [ "250cf3693060a5f803c5f1ddc082bb06b16112a9" ], "author": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Mon Apr 28 23:56:14 2025 -0400" }, "committer": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Fri May 09 18:05:55 2025 -0400" }, "message": "do_umount(): add missing barrier before refcount checks in sync case\n\ndo_umount() analogue of the race fixed in 119e1ef80ecf \"fix\n__legitimize_mnt()/mntput() race\". Here we want to make sure that\nif __legitimize_mnt() doesn\u0027t notice our lock_mount_hash(), we will\nnotice their refcount increment. Harder to hit than mntput_no_expire()\none, fortunately, and consequences are milder (sync umount acting\nlike umount -l on a rare race with RCU pathwalk hitting at just the\nwrong time instead of use-after-free galore mntput_no_expire()\ncounterpart used to be hit). Still a bug...\n\nFixes: 48a066e72d97 (\"RCU\u0027d vfsmounts\")\nReviewed-by: Christian Brauner \u003cbrauner@kernel.org\u003e\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\n" }, { "commit": "250cf3693060a5f803c5f1ddc082bb06b16112a9", "tree": "c884726453d713884f9b471b56f8e6d83974f482", "parents": [ "92a09c47464d040866cf2b4cd052bc60555185fb" ], "author": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Sun Apr 27 15:41:51 2025 -0400" }, "committer": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Fri May 09 18:02:12 2025 -0400" }, "message": "__legitimize_mnt(): check for MNT_SYNC_UMOUNT should be under mount_lock\n\n... or we risk stealing final mntput from sync umount - raising mnt_count\nafter umount(2) has verified that victim is not busy, but before it\nhas set MNT_SYNC_UMOUNT; in that case __legitimize_mnt() doesn\u0027t see\nthat it\u0027s safe to quietly undo mnt_count increment and leaves dropping\nthe reference to caller, where it\u0027ll be a full-blown mntput().\n\nCheck under mount_lock is needed; leaving the current one done before\ntaking that makes no sense - it\u0027s nowhere near common enough to bother\nwith.\n\nReviewed-by: Christian Brauner \u003cbrauner@kernel.org\u003e\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\n" }, { "commit": "0e1329d4045ca3606f9c06a8c47f62e758a09105", "tree": "7c4e12c7e14f4b44aff9c80784933e6571e2a906", "parents": [ "ea34704d6ad7225421cc3543906deacae35a6ea2", "5595c31c370957aabe739ac3996aedba8267603f" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri May 09 14:06:34 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri May 09 14:06:34 2025 -0700" }, "message": "Merge tag \u0027rust-fixes-6.15-2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ojeda/linux\n\nPull rust fixes from Miguel Ojeda:\n\n - Make CFI_AUTO_DEFAULT depend on !RUST or Rust \u003e\u003d 1.88.0\n\n - Clean Rust (and Clippy) lints for the upcoming Rust 1.87.0 and 1.88.0\n releases\n\n - Clean objtool warning for the upcoming Rust 1.87.0 release by adding\n one more noreturn function\n\n* tag \u0027rust-fixes-6.15-2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ojeda/linux:\n x86/Kconfig: make CFI_AUTO_DEFAULT depend on !RUST or Rust \u003e\u003d 1.88\n rust: clean Rust 1.88.0\u0027s `clippy::uninlined_format_args` lint\n rust: clean Rust 1.88.0\u0027s warning about `clippy::disallowed_macros` configuration\n rust: clean Rust 1.88.0\u0027s `unnecessary_transmutes` lint\n rust: allow Rust 1.87.0\u0027s `clippy::ptr_eq` lint\n objtool/rust: add one more `noreturn` Rust function for Rust 1.87.0\n" }, { "commit": "7a9b709e7cc5ce1ffb84ce07bf6d157e1de758df", "tree": "08b68f79eca35f07bdceaf9331f7baaf75420415", "parents": [ "e52c1dc7455d32c8a55f9949d300e5e87d011fa6" ], "author": { "name": "Pawan Gupta", "email": "pawan.kumar.gupta@linux.intel.com", "time": "Tue Dec 24 16:09:28 2024 -0800" }, "committer": { "name": "Dave Hansen", "email": "dave.hansen@linux.intel.com", "time": "Fri May 09 13:39:45 2025 -0700" }, "message": "selftest/x86/bugs: Add selftests for ITS\n\nBelow are the tests added for Indirect Target Selection (ITS):\n\n- its_sysfs.py - Check if sysfs reflects the correct mitigation status for\n the mitigation selected via the kernel cmdline.\n\n- its_permutations.py - tests mitigation selection with cmdline\n permutations with other bugs like spectre_v2 and retbleed.\n\n- its_indirect_alignment.py - verifies that for addresses in\n .retpoline_sites section that belong to lower half of cacheline are\n patched to ITS-safe thunk. Typical output looks like below:\n\n Site 49: function symbol: __x64_sys_restart_syscall+0x1f \u003c0xffffffffbb1509af\u003e\n # vmlinux: 0xffffffff813509af: jmp 0xffffffff81f5a8e0\n # kcore: 0xffffffffbb1509af: jmpq *%rax\n # ITS thunk NOT expected for site 49\n # PASSED: Found *%rax\n #\n Site 50: function symbol: __resched_curr+0xb0 \u003c0xffffffffbb181910\u003e\n # vmlinux: 0xffffffff81381910: jmp 0xffffffff81f5a8e0\n # kcore: 0xffffffffbb181910: jmp 0xffffffffc02000fc\n # ITS thunk expected for site 50\n # PASSED: Found 0xffffffffc02000fc -\u003e jmpq *%rax \u003cscattered-thunk?\u003e\n\n- its_ret_alignment.py - verifies that for addresses in .return_sites\n section that belong to lower half of cacheline are patched to\n its_return_thunk. Typical output looks like below:\n\n Site 97: function symbol: collect_event+0x48 \u003c0xffffffffbb007f18\u003e\n # vmlinux: 0xffffffff81207f18: jmp 0xffffffff81f5b500\n # kcore: 0xffffffffbb007f18: jmp 0xffffffffbbd5b560\n # PASSED: Found jmp 0xffffffffbbd5b560 \u003cits_return_thunk\u003e\n #\n Site 98: function symbol: collect_event+0xa4 \u003c0xffffffffbb007f74\u003e\n # vmlinux: 0xffffffff81207f74: jmp 0xffffffff81f5b500\n # kcore: 0xffffffffbb007f74: retq\n # PASSED: Found retq\n\nSome of these tests have dependency on tools like virtme-ng[1] and drgn[2].\nWhen the dependencies are not met, the test will be skipped.\n\n[1] https://github.com/arighi/virtme-ng\n[2] https://github.com/osandov/drgn\n\nCo-developed-by: Tao Zhang \u003ctao1.zhang@linux.intel.com\u003e\nSigned-off-by: Tao Zhang \u003ctao1.zhang@linux.intel.com\u003e\nSigned-off-by: Pawan Gupta \u003cpawan.kumar.gupta@linux.intel.com\u003e\nSigned-off-by: Dave Hansen \u003cdave.hansen@linux.intel.com\u003e\n" }, { "commit": "e52c1dc7455d32c8a55f9949d300e5e87d011fa6", "tree": "5e64ba6e1e40d504f1f8559250a6a49bf834c1df", "parents": [ "872df34d7c51a79523820ea6a14860398c639b87" ], "author": { "name": "Peter Zijlstra", "email": "peterz@infradead.org", "time": "Wed Apr 23 09:57:31 2025 +0200" }, "committer": { "name": "Dave Hansen", "email": "dave.hansen@linux.intel.com", "time": "Fri May 09 13:39:36 2025 -0700" }, "message": "x86/its: FineIBT-paranoid vs ITS\n\nFineIBT-paranoid was using the retpoline bytes for the paranoid check,\ndisabling retpolines, because all parts that have IBT also have eIBRS\nand thus don\u0027t need no stinking retpolines.\n\nExcept... ITS needs the retpolines for indirect calls must not be in\nthe first half of a cacheline :-/\n\nSo what was the paranoid call sequence:\n\n \u003cfineibt_paranoid_start\u003e:\n 0: 41 ba 78 56 34 12 mov $0x12345678, %r10d\n 6: 45 3b 53 f7 cmp -0x9(%r11), %r10d\n a: 4d 8d 5b \u003cf0\u003e lea -0x10(%r11), %r11\n e: 75 fd jne d \u003cfineibt_paranoid_start+0xd\u003e\n 10: 41 ff d3 call *%r11\n 13: 90 nop\n\nNow becomes:\n\n \u003cfineibt_paranoid_start\u003e:\n 0: 41 ba 78 56 34 12 mov $0x12345678, %r10d\n 6: 45 3b 53 f7 cmp -0x9(%r11), %r10d\n a: 4d 8d 5b f0 lea -0x10(%r11), %r11\n e: 2e e8 XX XX XX XX\tcs call __x86_indirect_paranoid_thunk_r11\n\n Where the paranoid_thunk looks like:\n\n 1d: \u003cea\u003e (bad)\n __x86_indirect_paranoid_thunk_r11:\n 1e: 75 fd jne 1d\n __x86_indirect_its_thunk_r11:\n 20: 41 ff eb jmp *%r11\n 23: cc int3\n\n[ dhansen: remove initialization to false ]\n\nSigned-off-by: Peter Zijlstra (Intel) \u003cpeterz@infradead.org\u003e\nSigned-off-by: Pawan Gupta \u003cpawan.kumar.gupta@linux.intel.com\u003e\nSigned-off-by: Dave Hansen \u003cdave.hansen@linux.intel.com\u003e\nReviewed-by: Alexandre Chartre \u003calexandre.chartre@oracle.com\u003e\n" }, { "commit": "872df34d7c51a79523820ea6a14860398c639b87", "tree": "879a171d40e26b5c2258dea65914eb18b00aeb55", "parents": [ "ebebe30794d38c51f71fe4951ba6af4159d9837d" ], "author": { "name": "Peter Zijlstra", "email": "peterz@infradead.org", "time": "Mon Oct 14 10:05:48 2024 -0700" }, "committer": { "name": "Dave Hansen", "email": "dave.hansen@linux.intel.com", "time": "Fri May 09 13:36:58 2025 -0700" }, "message": "x86/its: Use dynamic thunks for indirect branches\n\nITS mitigation moves the unsafe indirect branches to a safe thunk. This\ncould degrade the prediction accuracy as the source address of indirect\nbranches becomes same for different execution paths.\n\nTo improve the predictions, and hence the performance, assign a separate\nthunk for each indirect callsite. This is also a defense-in-depth measure\nto avoid indirect branches aliasing with each other.\n\nAs an example, 5000 dynamic thunks would utilize around 16 bits of the\naddress space, thereby gaining entropy. For a BTB that uses\n32 bits for indexing, dynamic thunks could provide better prediction\naccuracy over fixed thunks.\n\nHave ITS thunks be variable sized and use EXECMEM_MODULE_TEXT such that\nthey are both more flexible (got to extend them later) and live in 2M TLBs,\njust like kernel code, avoiding undue TLB pressure.\n\nSigned-off-by: Peter Zijlstra (Intel) \u003cpeterz@infradead.org\u003e\nSigned-off-by: Pawan Gupta \u003cpawan.kumar.gupta@linux.intel.com\u003e\nSigned-off-by: Dave Hansen \u003cdave.hansen@linux.intel.com\u003e\nReviewed-by: Alexandre Chartre \u003calexandre.chartre@oracle.com\u003e\n" }, { "commit": "ebebe30794d38c51f71fe4951ba6af4159d9837d", "tree": "aab5d921d9943e0e15ab3682f8085b96213688fa", "parents": [ "d6d1e3e6580ca35071ad474381f053cbf1fb6414" ], "author": { "name": "Pawan Gupta", "email": "pawan.kumar.gupta@linux.intel.com", "time": "Sat May 03 09:46:31 2025 -0700" }, "committer": { "name": "Dave Hansen", "email": "dave.hansen@linux.intel.com", "time": "Fri May 09 13:33:35 2025 -0700" }, "message": "x86/ibt: Keep IBT disabled during alternative patching\n\ncfi_rewrite_callers() updates the fineIBT hash matching at the caller side,\nbut except for paranoid-mode it relies on apply_retpoline() and friends for\nany ENDBR relocation. This could temporarily cause an indirect branch to\nland on a poisoned ENDBR.\n\nFor instance, with para-virtualization enabled, a simple wrmsrl() could\nhave an indirect branch pointing to native_write_msr() who\u0027s ENDBR has been\nrelocated due to fineIBT:\n\n\u003cwrmsrl\u003e:\n push %rbp\n mov %rsp,%rbp\n mov %esi,%eax\n mov %rsi,%rdx\n shr $0x20,%rdx\n mov %edi,%edi\n mov %rax,%rsi\n call *0x21e65d0(%rip) # \u003cpv_ops+0xb8\u003e\n ^^^^^^^^^^^^^^^^^^^^^^^\n\nSuch an indirect call during the alternative patching could #CP if the\ncaller is not *yet* adjusted for the new target ENDBR. To prevent a false\n #CP, keep CET-IBT disabled until all callers are patched.\n\nPatching during the module load does not need to be guarded by IBT-disable\nbecause the module code is not executed until the patching is complete.\n\nSigned-off-by: Pawan Gupta \u003cpawan.kumar.gupta@linux.intel.com\u003e\nSigned-off-by: Dave Hansen \u003cdave.hansen@linux.intel.com\u003e\nReviewed-by: Alexandre Chartre \u003calexandre.chartre@oracle.com\u003e\n" }, { "commit": "d6d1e3e6580ca35071ad474381f053cbf1fb6414", "tree": "3f7e32bd070bdc785a93423464a03c7f5a0cc11b", "parents": [ "f0cd7091cc5a032c8870b4285305d9172569d126" ], "author": { "name": "Peter Zijlstra", "email": "peterz@infradead.org", "time": "Wed Apr 23 16:30:11 2025 +0200" }, "committer": { "name": "Dave Hansen", "email": "dave.hansen@linux.intel.com", "time": "Fri May 09 13:33:20 2025 -0700" }, "message": "mm/execmem: Unify early execmem_cache behaviour\n\nEarly kernel memory is RWX, only at the end of early boot (before SMP)\ndo we mark things ROX. Have execmem_cache mirror this behaviour for\nearly users.\n\nThis avoids having to remember what code is execmem and what is not --\nwe can poke everything with impunity ;-) Also performance for not\nhaving to do endless text_poke_mm switches.\n\nSigned-off-by: Peter Zijlstra (Intel) \u003cpeterz@infradead.org\u003e\nSigned-off-by: Pawan Gupta \u003cpawan.kumar.gupta@linux.intel.com\u003e\nSigned-off-by: Dave Hansen \u003cdave.hansen@linux.intel.com\u003e\nReviewed-by: Alexandre Chartre \u003calexandre.chartre@oracle.com\u003e\n" }, { "commit": "09acc3266cdd8e32a62dd6c5052f44724111120b", "tree": "119ff4b879e80153dbf6692785ea359d9446bec0", "parents": [ "3238532ba55c9bf9b63ae17ebfc5deb320c737c3" ], "author": { "name": "Rob Herring (Arm)", "email": "robh@kernel.org", "time": "Wed Apr 09 16:02:54 2025 -0500" }, "committer": { "name": "Arnd Bergmann", "email": "arnd@arndb.de", "time": "Fri May 09 22:28:46 2025 +0200" }, "message": "arm64: dts: amazon: Fix simple-bus node name schema warnings\n\nFix a couple of node name warnings from the schema checks:\n\narch/arm64/boot/dts/amazon/alpine-v2-evp.dt.yaml: io-fabric: $nodename:0: \u0027io-fabric\u0027 does not match \u0027^(bus|soc|axi|ahb|apb)(@[0-9a-f]+)?$\u0027\narch/arm64/boot/dts/amazon/alpine-v3-evp.dt.yaml: io-fabric: $nodename:0: \u0027io-fabric\u0027 does not match \u0027^(bus|soc|axi|ahb|apb)(@[0-9a-f]+)?$\u0027\n\nSigned-off-by: Rob Herring (Arm) \u003crobh@kernel.org\u003e\nLink: https://lore.kernel.org/r/20250409210255.1541298-1-robh@kernel.org\nSigned-off-by: Arnd Bergmann \u003carnd@arndb.de\u003e\n" }, { "commit": "3238532ba55c9bf9b63ae17ebfc5deb320c737c3", "tree": "809666c1e275363c672f9fcd78c95fb71271057b", "parents": [ "33e79299f921eba4d77f2af0fd359285fa9f3fb3" ], "author": { "name": "Wolfram Sang", "email": "wsa+renesas@sang-engineering.com", "time": "Mon Mar 31 21:06:42 2025 +0200" }, "committer": { "name": "Arnd Bergmann", "email": "arnd@arndb.de", "time": "Fri May 09 22:27:27 2025 +0200" }, "message": "MAINTAINERS: delete email for Shiraz Hashim\n\nThe email address bounced. I couldn\u0027t find a newer one in recent git\nhistory (last activity 9 years ago), so delete this email entry.\n\nSigned-off-by: Wolfram Sang \u003cwsa+renesas@sang-engineering.com\u003e\nLink: https://lore.kernel.org/r/20250331190731.5094-2-wsa+renesas@sang-engineering.com\nSigned-off-by: Arnd Bergmann \u003carnd@arndb.de\u003e\n" }, { "commit": "f0cd7091cc5a032c8870b4285305d9172569d126", "tree": "95a1874f3bb98ae112efa29993f871336bc067f3", "parents": [ "facd226f7e0c8ca936ac114aba43cb3e8b94e41e" ], "author": { "name": "Pawan Gupta", "email": "pawan.kumar.gupta@linux.intel.com", "time": "Fri May 02 06:25:19 2025 -0700" }, "committer": { "name": "Dave Hansen", "email": "dave.hansen@linux.intel.com", "time": "Fri May 09 13:22:05 2025 -0700" }, "message": "x86/its: Align RETs in BHB clear sequence to avoid thunking\n\nThe software mitigation for BHI is to execute BHB clear sequence at syscall\nentry, and possibly after a cBPF program. ITS mitigation thunks RETs in the\nlower half of the cacheline. This causes the RETs in the BHB clear sequence\nto be thunked as well, adding unnecessary branches to the BHB clear\nsequence.\n\nSince the sequence is in hot path, align the RET instructions in the\nsequence to avoid thunking.\n\nThis is how disassembly clear_bhb_loop() looks like after this change:\n\n 0x44 \u003c+4\u003e: mov $0x5,%ecx\n 0x49 \u003c+9\u003e: call 0xffffffff81001d9b \u003cclear_bhb_loop+91\u003e\n 0x4e \u003c+14\u003e: jmp 0xffffffff81001de5 \u003cclear_bhb_loop+165\u003e\n 0x53 \u003c+19\u003e: int3\n ...\n 0x9b \u003c+91\u003e: call 0xffffffff81001dce \u003cclear_bhb_loop+142\u003e\n 0xa0 \u003c+96\u003e: ret\n 0xa1 \u003c+97\u003e: int3\n ...\n 0xce \u003c+142\u003e: mov $0x5,%eax\n 0xd3 \u003c+147\u003e: jmp 0xffffffff81001dd6 \u003cclear_bhb_loop+150\u003e\n 0xd5 \u003c+149\u003e: nop\n 0xd6 \u003c+150\u003e: sub $0x1,%eax\n 0xd9 \u003c+153\u003e: jne 0xffffffff81001dd3 \u003cclear_bhb_loop+147\u003e\n 0xdb \u003c+155\u003e: sub $0x1,%ecx\n 0xde \u003c+158\u003e: jne 0xffffffff81001d9b \u003cclear_bhb_loop+91\u003e\n 0xe0 \u003c+160\u003e: ret\n 0xe1 \u003c+161\u003e: int3\n 0xe2 \u003c+162\u003e: int3\n 0xe3 \u003c+163\u003e: int3\n 0xe4 \u003c+164\u003e: int3\n 0xe5 \u003c+165\u003e: lfence\n 0xe8 \u003c+168\u003e: pop %rbp\n 0xe9 \u003c+169\u003e: ret\n\nSuggested-by: Andrew Cooper \u003candrew.cooper3@citrix.com\u003e\nSigned-off-by: Pawan Gupta \u003cpawan.kumar.gupta@linux.intel.com\u003e\nSigned-off-by: Dave Hansen \u003cdave.hansen@linux.intel.com\u003e\nReviewed-by: Alexandre Chartre \u003calexandre.chartre@oracle.com\u003e\n" }, { "commit": "facd226f7e0c8ca936ac114aba43cb3e8b94e41e", "tree": "b2dd9cb6ecfa2b5c0db188ce295ede9e2fe96b69", "parents": [ "2665281a07e19550944e8354a2024635a7b2714a" ], "author": { "name": "Pawan Gupta", "email": "pawan.kumar.gupta@linux.intel.com", "time": "Mon Dec 02 12:07:08 2024 -0800" }, "committer": { "name": "Dave Hansen", "email": "dave.hansen@linux.intel.com", "time": "Fri May 09 13:22:05 2025 -0700" }, "message": "x86/its: Add support for RSB stuffing mitigation\n\nWhen retpoline mitigation is enabled for spectre-v2, enabling\ncall-depth-tracking and RSB stuffing also mitigates ITS. Add cmdline option\nindirect_target_selection\u003dstuff to allow enabling RSB stuffing mitigation.\n\nWhen retpoline mitigation is not enabled, \u003dstuff option is ignored, and\ndefault mitigation for ITS is deployed.\n\nSigned-off-by: Pawan Gupta \u003cpawan.kumar.gupta@linux.intel.com\u003e\nSigned-off-by: Dave Hansen \u003cdave.hansen@linux.intel.com\u003e\nReviewed-by: Josh Poimboeuf \u003cjpoimboe@kernel.org\u003e\nReviewed-by: Alexandre Chartre \u003calexandre.chartre@oracle.com\u003e\n" }, { "commit": "2665281a07e19550944e8354a2024635a7b2714a", "tree": "14068159db67f67490a6c672f96d0322f4c2576c", "parents": [ "f4818881c47fd91fcb6d62373c57c7844e3de1c0" ], "author": { "name": "Pawan Gupta", "email": "pawan.kumar.gupta@linux.intel.com", "time": "Mon Nov 18 09:53:12 2024 -0800" }, "committer": { "name": "Dave Hansen", "email": "dave.hansen@linux.intel.com", "time": "Fri May 09 13:22:05 2025 -0700" }, "message": "x86/its: Add \"vmexit\" option to skip mitigation on some CPUs\n\nIce Lake generation CPUs are not affected by guest/host isolation part of\nITS. If a user is only concerned about KVM guests, they can now choose a\nnew cmdline option \"vmexit\" that will not deploy the ITS mitigation when\nCPU is not affected by guest/host isolation. This saves the performance\noverhead of ITS mitigation on Ice Lake gen CPUs.\n\nWhen \"vmexit\" option selected, if the CPU is affected by ITS guest/host\nisolation, the default ITS mitigation is deployed.\n\nSigned-off-by: Pawan Gupta \u003cpawan.kumar.gupta@linux.intel.com\u003e\nSigned-off-by: Dave Hansen \u003cdave.hansen@linux.intel.com\u003e\nReviewed-by: Josh Poimboeuf \u003cjpoimboe@kernel.org\u003e\nReviewed-by: Alexandre Chartre \u003calexandre.chartre@oracle.com\u003e\n" }, { "commit": "f4818881c47fd91fcb6d62373c57c7844e3de1c0", "tree": "8fcd27d55000dfab445bb8412af95deae2011257", "parents": [ "a75bf27fe41abe658c53276a0c486c4bf9adecfc" ], "author": { "name": "Pawan Gupta", "email": "pawan.kumar.gupta@linux.intel.com", "time": "Fri Jun 21 20:23:23 2024 -0700" }, "committer": { "name": "Dave Hansen", "email": "dave.hansen@linux.intel.com", "time": "Fri May 09 13:22:05 2025 -0700" }, "message": "x86/its: Enable Indirect Target Selection mitigation\n\nIndirect Target Selection (ITS) is a bug in some pre-ADL Intel CPUs with\neIBRS. It affects prediction of indirect branch and RETs in the\nlower half of cacheline. Due to ITS such branches may get wrongly predicted\nto a target of (direct or indirect) branch that is located in the upper\nhalf of the cacheline.\n\nScope of impact\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nGuest/host isolation\n--------------------\nWhen eIBRS is used for guest/host isolation, the indirect branches in the\nVMM may still be predicted with targets corresponding to branches in the\nguest.\n\nIntra-mode\n----------\ncBPF or other native gadgets can be used for intra-mode training and\ndisclosure using ITS.\n\nUser/kernel isolation\n---------------------\nWhen eIBRS is enabled user/kernel isolation is not impacted.\n\nIndirect Branch Prediction Barrier (IBPB)\n-----------------------------------------\nAfter an IBPB, indirect branches may be predicted with targets\ncorresponding to direct branches which were executed prior to IBPB. This is\nmitigated by a microcode update.\n\nAdd cmdline parameter indirect_target_selection\u003doff|on|force to control the\nmitigation to relocate the affected branches to an ITS-safe thunk i.e.\nlocated in the upper half of cacheline. Also add the sysfs reporting.\n\nWhen retpoline mitigation is deployed, ITS safe-thunks are not needed,\nbecause retpoline sequence is already ITS-safe. Similarly, when call depth\ntracking (CDT) mitigation is deployed (retbleed\u003dstuff), ITS safe return\nthunk is not used, as CDT prevents RSB-underflow.\n\nTo not overcomplicate things, ITS mitigation is not supported with\nspectre-v2 lfence;jmp mitigation. Moreover, it is less practical to deploy\nlfence;jmp mitigation on ITS affected parts anyways.\n\nSigned-off-by: Pawan Gupta \u003cpawan.kumar.gupta@linux.intel.com\u003e\nSigned-off-by: Dave Hansen \u003cdave.hansen@linux.intel.com\u003e\nReviewed-by: Josh Poimboeuf \u003cjpoimboe@kernel.org\u003e\nReviewed-by: Alexandre Chartre \u003calexandre.chartre@oracle.com\u003e\n" }, { "commit": "a75bf27fe41abe658c53276a0c486c4bf9adecfc", "tree": "0898216e1eb51992a199696ed652042942ac11f9", "parents": [ "8754e67ad4ac692c67ff1f99c0d07156f04ae40c" ], "author": { "name": "Pawan Gupta", "email": "pawan.kumar.gupta@linux.intel.com", "time": "Fri Jun 21 21:17:21 2024 -0700" }, "committer": { "name": "Dave Hansen", "email": "dave.hansen@linux.intel.com", "time": "Fri May 09 13:22:05 2025 -0700" }, "message": "x86/its: Add support for ITS-safe return thunk\n\nRETs in the lower half of cacheline may be affected by ITS bug,\nspecifically when the RSB-underflows. Use ITS-safe return thunk for such\nRETs.\n\nRETs that are not patched:\n\n- RET in retpoline sequence does not need to be patched, because the\n sequence itself fills an RSB before RET.\n- RET in Call Depth Tracking (CDT) thunks __x86_indirect_{call|jump}_thunk\n and call_depth_return_thunk are not patched because CDT by design\n prevents RSB-underflow.\n- RETs in .init section are not reachable after init.\n- RETs that are explicitly marked safe with ANNOTATE_UNRET_SAFE.\n\nSigned-off-by: Pawan Gupta \u003cpawan.kumar.gupta@linux.intel.com\u003e\nSigned-off-by: Dave Hansen \u003cdave.hansen@linux.intel.com\u003e\nReviewed-by: Josh Poimboeuf \u003cjpoimboe@kernel.org\u003e\nReviewed-by: Alexandre Chartre \u003calexandre.chartre@oracle.com\u003e\n" }, { "commit": "8754e67ad4ac692c67ff1f99c0d07156f04ae40c", "tree": "e5b008534b11b3cdf6ddb72cbf1310ee33f85889", "parents": [ "159013a7ca18c271ff64192deb62a689b622d860" ], "author": { "name": "Pawan Gupta", "email": "pawan.kumar.gupta@linux.intel.com", "time": "Fri Jun 21 21:17:21 2024 -0700" }, "committer": { "name": "Dave Hansen", "email": "dave.hansen@linux.intel.com", "time": "Fri May 09 13:22:04 2025 -0700" }, "message": "x86/its: Add support for ITS-safe indirect thunk\n\nDue to ITS, indirect branches in the lower half of a cacheline may be\nvulnerable to branch target injection attack.\n\nIntroduce ITS-safe thunks to patch indirect branches in the lower half of\ncacheline with the thunk. Also thunk any eBPF generated indirect branches\nin emit_indirect_jump().\n\nBelow category of indirect branches are not mitigated:\n\n- Indirect branches in the .init section are not mitigated because they are\n discarded after boot.\n- Indirect branches that are explicitly marked retpoline-safe.\n\nNote that retpoline also mitigates the indirect branches against ITS. This\nis because the retpoline sequence fills an RSB entry before RET, and it\ndoes not suffer from RSB-underflow part of the ITS.\n\nSigned-off-by: Pawan Gupta \u003cpawan.kumar.gupta@linux.intel.com\u003e\nSigned-off-by: Dave Hansen \u003cdave.hansen@linux.intel.com\u003e\nReviewed-by: Josh Poimboeuf \u003cjpoimboe@kernel.org\u003e\nReviewed-by: Alexandre Chartre \u003calexandre.chartre@oracle.com\u003e\n" }, { "commit": "159013a7ca18c271ff64192deb62a689b622d860", "tree": "4abe20827c76dac6f9959c6e71539ade06b32be3", "parents": [ "1ac116ce6468670eeda39345a5585df308243dca" ], "author": { "name": "Pawan Gupta", "email": "pawan.kumar.gupta@linux.intel.com", "time": "Fri Jun 21 17:40:41 2024 -0700" }, "committer": { "name": "Dave Hansen", "email": "dave.hansen@linux.intel.com", "time": "Fri May 09 13:22:04 2025 -0700" }, "message": "x86/its: Enumerate Indirect Target Selection (ITS) bug\n\nITS bug in some pre-Alderlake Intel CPUs may allow indirect branches in the\nfirst half of a cache line get predicted to a target of a branch located in\nthe second half of the cache line.\n\nSet X86_BUG_ITS on affected CPUs. Mitigation to follow in later commits.\n\nSigned-off-by: Pawan Gupta \u003cpawan.kumar.gupta@linux.intel.com\u003e\nSigned-off-by: Dave Hansen \u003cdave.hansen@linux.intel.com\u003e\nReviewed-by: Josh Poimboeuf \u003cjpoimboe@kernel.org\u003e\nReviewed-by: Alexandre Chartre \u003calexandre.chartre@oracle.com\u003e\n" }, { "commit": "1ac116ce6468670eeda39345a5585df308243dca", "tree": "5204248db1322fdae361519ef2a6395aaaf2a4a1", "parents": [ "92a09c47464d040866cf2b4cd052bc60555185fb" ], "author": { "name": "Pawan Gupta", "email": "pawan.kumar.gupta@linux.intel.com", "time": "Fri Apr 11 15:36:38 2025 -0700" }, "committer": { "name": "Dave Hansen", "email": "dave.hansen@linux.intel.com", "time": "Fri May 09 13:22:04 2025 -0700" }, "message": "Documentation: x86/bugs/its: Add ITS documentation\n\nAdd the admin-guide for Indirect Target Selection (ITS).\n\nSigned-off-by: Pawan Gupta \u003cpawan.kumar.gupta@linux.intel.com\u003e\nSigned-off-by: Dave Hansen \u003cdave.hansen@linux.intel.com\u003e\nReviewed-by: Josh Poimboeuf \u003cjpoimboe@kernel.org\u003e\nReviewed-by: Alexandre Chartre \u003calexandre.chartre@oracle.com\u003e\n" }, { "commit": "ea34704d6ad7225421cc3543906deacae35a6ea2", "tree": "c9b21676df7a8fda0b8b775a204d3f2c0990fcf5", "parents": [ "50358c251eae19a2b2fc54b6944e362ef2fefff0", "c2c64ed09c7b44a893d22c8b8ddb3ba7265494f3" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri May 09 12:41:34 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri May 09 12:41:34 2025 -0700" }, "message": "Merge tag \u0027drm-fixes-2025-05-10\u0027 of https://gitlab.freedesktop.org/drm/kernel\n\nPull drm fixes from Dave Airlie:\n \"Weekly drm fixes, bit bigger than last week, but overall amdgpu/xe\n with some ivpu bits and a random few fixes, and dropping the\n ttm_backup struct which wrapped struct file and was recently\n frowned at.\n\n drm:\n - Fix overflow when generating wedged event\n\n ttm:\n - Fix documentation\n - Remove struct ttm_backup\n\n panel:\n - simple: Fix timings for AUO G101EVN010\n\n amdgpu:\n - DC FP fixes\n - Freesync fix\n - DMUB AUX fixes\n - VCN fix\n - Hibernation fixes\n - HDP fixes\n\n xe:\n - Prevent PF queue overflow\n - Hold all forcewake during mocs test\n - Remove GSC flush on reset path\n - Fix forcewake put on error path\n - Fix runtime warning when building without svm\n\n i915:\n - Fix oops on resume after disconnecting DP MST sinks during suspend\n - Fix SPLC num_waiters refcounting\n\n ivpu:\n - Increase timeouts\n - Fix deadlock in cmdq ioctl\n - Unlock mutices in correct order\n\n v3d:\n - Avoid memory leak in job handling\"\n\n* tag \u0027drm-fixes-2025-05-10\u0027 of https://gitlab.freedesktop.org/drm/kernel: (32 commits)\n drm/i915/dp: Fix determining SST/MST mode during MTP TU state computation\n drm/xe: Add config control for svm flush work\n drm/xe: Release force wake first then runtime power\n drm/xe/gsc: do not flush the GSC worker from the reset path\n drm/xe/tests/mocs: Hold XE_FORCEWAKE_ALL for LNCF regs\n drm/xe: Add page queue multiplier\n drm/amdgpu/hdp7: use memcfg register to post the write for HDP flush\n drm/amdgpu/hdp6: use memcfg register to post the write for HDP flush\n drm/amdgpu/hdp5.2: use memcfg register to post the write for HDP flush\n drm/amdgpu/hdp5: use memcfg register to post the write for HDP flush\n drm/amdgpu/hdp4: use memcfg register to post the write for HDP flush\n drm/amdgpu: fix pm notifier handling\n Revert \"drm/amd: Stop evicting resources on APUs in suspend\"\n drm/amdgpu/vcn: using separate VCN1_AON_SOC offset\n drm/amd/display: Fix wrong handling for AUX_DEFER case\n drm/amd/display: Copy AUX read reply data whenever length \u003e 0\n drm/amd/display: Remove incorrect checking in dmub aux handler\n drm/amd/display: Fix the checking condition in dmub aux handling\n drm/amd/display: Shift DMUB AUX reply command if necessary\n drm/amd/display: Call FP Protect Before Mode Programming/Mode Support\n ...\n" }, { "commit": "c2c64ed09c7b44a893d22c8b8ddb3ba7265494f3", "tree": "057accc489008b85293751e81c90742b9e20c0d6", "parents": [ "f7be784caf9f5ebf0ff9829e5921d42b6b5e62cf", "732b87a409667a370b87955c518e5d004de740b5" ], "author": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Sat May 10 05:07:17 2025 +1000" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Sat May 10 05:07:18 2025 +1000" }, "message": "Merge tag \u0027drm-intel-fixes-2025-05-09\u0027 of https://gitlab.freedesktop.org/drm/i915/kernel into drm-fixes\n\ndrm/i915 fixes for v6.15-rc6:\n- Fix oops on resume after disconnecting DP MST sinks during suspend\n- Fix SPLC num_waiters refcounting\n\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\nFrom: Jani Nikula \u003cjani.nikula@intel.com\u003e\nLink: https://lore.kernel.org/r/87tt5umeaw.fsf@intel.com\n" }, { "commit": "f7be784caf9f5ebf0ff9829e5921d42b6b5e62cf", "tree": "2bbae8f94542afe5235344144ab4395532a44e7b", "parents": [ "80e12f3e2a5a911fb67597da4c17af48e7cb92de", "564467e9d06c6352fac9100e8957d40a1a50234c" ], "author": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Sat May 10 05:02:38 2025 +1000" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Sat May 10 05:02:46 2025 +1000" }, "message": "Merge tag \u0027drm-xe-fixes-2025-05-09\u0027 of https://gitlab.freedesktop.org/drm/xe/kernel into drm-fixes\n\nDriver Changes:\n- Prevent PF queue overflow\n- Hold all forcewake during mocs test\n- Remove GSC flush on reset path\n- Fix forcewake put on error path\n- Fix runtime warning when building without svm\n\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n\nFrom: Lucas De Marchi \u003clucas.demarchi@intel.com\u003e\nLink: https://lore.kernel.org/r/jffqa56f2zp4i5ztz677cdspgxhnw7qfop3dd3l2epykfpfvza@q2nw6wapsphz\n" }, { "commit": "50358c251eae19a2b2fc54b6944e362ef2fefff0", "tree": "60818f46b4d5e81898db86289744ad2c3dcb7ade", "parents": [ "3013c33dcbd9b3107eef8facce0e4c69f3b7f780", "363cd2b81cfdf706bbfc9ec78db000c9b1ecc552" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri May 09 11:30:26 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri May 09 11:30:26 2025 -0700" }, "message": "Merge tag \u0027arm64-fixes\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux\n\nPull arm64 fix from Catalin Marinas:\n \"Move the arm64_use_ng_mappings variable from the .bss to the .data\n section as it is accessed very early during boot with the MMU off and\n before the .bss has been initialised.\n\n This could lead to incorrect idmap page table\"\n\n* tag \u0027arm64-fixes\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:\n arm64: cpufeature: Move arm64_use_ng_mappings to the .data section to prevent wrong idmap generation\n" }, { "commit": "3013c33dcbd9b3107eef8facce0e4c69f3b7f780", "tree": "fdc7014ed87ee3a50d38c16e60091f2d236a7672", "parents": [ "cc9f0629caee0cc54356743bf1ff54dca55275cf", "01534f3e0dd75e27ed03e5542f0c7bf6aa7130f1" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri May 09 11:17:50 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri May 09 11:17:50 2025 -0700" }, "message": "Merge tag \u0027riscv-for-linus-6.15-rc6\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux\n\nPull RISC-V fixes from Palmer Dabbelt:\n\n - The compressed half-word misaligned access instructions (c.lhu, c.lh,\n and c.sh) from the Zcb extension are now properly emulated\n\n - A series of fixes to properly emulate permissions while handling\n userspace misaligned accesses\n\n - A pair of fixes for PR_GET_TAGGED_ADDR_CTRL to avoid accessing the\n envcfg CSR on systems that don\u0027t support that CSR, and to report\n those failures up to userspace\n\n - The .rela.dyn section is no longer stripped from vmlinux, as it is\n necessary to relocate the kernel under some conditions (including\n kexec)\n\n* tag \u0027riscv-for-linus-6.15-rc6\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux:\n riscv: Disallow PR_GET_TAGGED_ADDR_CTRL without Supm\n scripts: Do not strip .rela.dyn section\n riscv: Fix kernel crash due to PR_SET_TAGGED_ADDR_CTRL\n riscv: misaligned: use get_user() instead of __get_user()\n riscv: misaligned: enable IRQs while handling misaligned accesses\n riscv: misaligned: factorize trap handling\n riscv: misaligned: Add handling for ZCB instructions\n" }, { "commit": "39b5ef791d109dd54c7c2e6e87933edfcc0ad1ac", "tree": "dd62954fdbba8052fac1065f6085fdf33b996186", "parents": [ "1bf67c8fdbda21fadd564a12dbe2b13c1ea5eda7" ], "author": { "name": "Waiman Long", "email": "longman@redhat.com", "time": "Thu May 08 15:24:13 2025 -0400" }, "committer": { "name": "Tejun Heo", "email": "tj@kernel.org", "time": "Fri May 09 07:35:14 2025 -1000" }, "message": "cgroup/cpuset: Extend kthread_is_per_cpu() check to all PF_NO_SETAFFINITY tasks\n\nCommit ec5fbdfb99d1 (\"cgroup/cpuset: Enable update_tasks_cpumask()\non top_cpuset\") enabled us to pull CPUs dedicated to child partitions\nfrom tasks in top_cpuset by ignoring per cpu kthreads. However, there\ncan be other kthreads that are not per cpu but have PF_NO_SETAFFINITY\nflag set to indicate that we shouldn\u0027t mess with their CPU affinity.\nFor other kthreads, their affinity will be changed to skip CPUs dedicated\nto child partitions whether it is an isolating or a scheduling one.\n\nAs all the per cpu kthreads have PF_NO_SETAFFINITY set, the\nPF_NO_SETAFFINITY tasks are essentially a superset of per cpu kthreads.\nFix this issue by dropping the kthread_is_per_cpu() check and checking\nthe PF_NO_SETAFFINITY flag instead.\n\nFixes: ec5fbdfb99d1 (\"cgroup/cpuset: Enable update_tasks_cpumask() on top_cpuset\")\nSigned-off-by: Waiman Long \u003clongman@redhat.com\u003e\nAcked-by: Frederic Weisbecker \u003cfrederic@kernel.org\u003e\nSigned-off-by: Tejun Heo \u003ctj@kernel.org\u003e\n" }, { "commit": "cc9f0629caee0cc54356743bf1ff54dca55275cf", "tree": "002d216020036341a7085bf1ca6cc814ce8ee2fe", "parents": [ "7380c60b2831220ca6d60d1560ecf8d9bdf06288", "dd90905d5a8a15a6d4594d15fc8ed626587187ca" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri May 09 10:34:50 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri May 09 10:34:50 2025 -0700" }, "message": "Merge tag \u0027block-6.15-20250509\u0027 of git://git.kernel.dk/linux\n\nPull block fixes from Jens Axboe:\n\n - Fix for a regression in this series for loop and read/write iterator\n handling\n\n - zone append block update tweak\n\n - remove a broken IO priority test\n\n - NVMe pull request via Christoph:\n - unblock ctrl state transition for firmware update (Daniel\n Wagner)\n\n* tag \u0027block-6.15-20250509\u0027 of git://git.kernel.dk/linux:\n block: remove test of incorrect io priority level\n nvme: unblock ctrl state transition for firmware update\n block: only update request sector if needed\n loop: Add sanity check for read/write_iter\n" }, { "commit": "7380c60b2831220ca6d60d1560ecf8d9bdf06288", "tree": "4d4db9ab669616530fd89537e374a367e0c5137a", "parents": [ "29fe5d50dfa6b61043e2e89206389ae56b5596eb", "92835cebab120f8a5f023a26a792a2ac3f816c4f" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri May 09 09:26:46 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri May 09 09:26:46 2025 -0700" }, "message": "Merge tag \u0027io_uring-6.15-20250509\u0027 of git://git.kernel.dk/linux\n\nPull io_uring fixes from Jens Axboe:\n\n - Fix for linked timeouts arming and firing wrt prep and issue of the\n request being managed by the linked timeout\n\n - Fix for a CQE ordering issue between requests with multishot and\n using the same buffer group. This is a dumbed down version for this\n release and for stable, it\u0027ll get improved for v6.16\n\n - Tweak the SQPOLL submit batch size. A previous commit made SQPOLL\n manage its own task_work and chose a tiny batch size, bump it from 8\n to 32 to fix a performance regression due to that\n\n* tag \u0027io_uring-6.15-20250509\u0027 of git://git.kernel.dk/linux:\n io_uring/sqpoll: Increase task_work submission batch size\n io_uring: ensure deferred completions are flushed for multishot\n io_uring: always arm linked timeouts prior to issue\n" }, { "commit": "29fe5d50dfa6b61043e2e89206389ae56b5596eb", "tree": "e95dce47fd34d42d7f37e9aa1fad40358e2019bb", "parents": [ "fea4e317f9e7e1f449ce90dedc27a2d2a95bee5a", "a6aeb739974ec73e5217c75a7c008a688d3d5cf1" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri May 09 09:09:49 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri May 09 09:09:49 2025 -0700" }, "message": "Merge tag \u0027modules-6.15-rc6\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/modules/linux\n\nPull modules fix from Petr Pavlu:\n \"A single fix to prevent use of an uninitialized completion pointer\n when releasing a module_kobject in specific situations.\n\n This addresses a latent bug exposed by commit f95bbfe18512 (\"drivers:\n base: handle module_kobject creation\")\"\n\n* tag \u0027modules-6.15-rc6\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/modules/linux:\n module: ensure that kobject_put() is safe for module type kobjects\n" }, { "commit": "33e79299f921eba4d77f2af0fd359285fa9f3fb3", "tree": "5e37e65d3d3f52b1fbde8dd1b62a7e22e132eb72", "parents": [ "6053915252d78f70ffc9bcc9eefb8da9c653dc82", "dc5befecbe2683ac49fc8dc76aade35e62f4cf30" ], "author": { "name": "Arnd Bergmann", "email": "arnd@arndb.de", "time": "Fri May 09 18:02:16 2025 +0200" }, "committer": { "name": "Arnd Bergmann", "email": "arnd@arndb.de", "time": "Fri May 09 18:02:16 2025 +0200" }, "message": "Merge tag \u0027asahi-soc-fixes-6.15\u0027 of https://github.com/AsahiLinux/linux into arm/fixes\n\nApple SoC fixes for 6.15\n\nThis tag contains two small commits since rc1:\n- Add a .mailmap entry requested by Asahi Lina to better filter her\n emails\n- Mark the power domains for the touchbar support introduced with 6.15\n as always on since the driver cannot initialize the touchbar from\n scratch after the domains are powered off (e.g. during suspend).\n\n* tag \u0027asahi-soc-fixes-6.15\u0027 of https://github.com/AsahiLinux/linux:\n arm64: dts: apple: touchbar: Mark ps_dispdfr_be as always-on\n mailmap: Update email for Asahi Lina\n\nLink: https://lore.kernel.org/r/20250423145047.3098-1-sven@svenpeter.dev\nSigned-off-by: Arnd Bergmann \u003carnd@arndb.de\u003e\n" }, { "commit": "6053915252d78f70ffc9bcc9eefb8da9c653dc82", "tree": "9ca0845bf63e35861c6b38d961207865f7302568", "parents": [ "e36f6de6adecafe90b52739adb9405d342bc9ad1", "3e6244429ba38f8dee3336b8b805948276b281ab" ], "author": { "name": "Arnd Bergmann", "email": "arnd@arndb.de", "time": "Fri May 09 18:01:07 2025 +0200" }, "committer": { "name": "Arnd Bergmann", "email": "arnd@arndb.de", "time": "Fri May 09 18:01:07 2025 +0200" }, "message": "Merge tag \u0027riscv-sophgo-dt-fixes-for-v6.15-rc1\u0027 of https://github.com/sophgo/linux into arm/fixes\n\nRISC-V Sophgo Devicetree fixes for v6.15-rc1\n\nJust one minor fix to correct DMA data-width\nconfiguration for CV18xx.\n\nSigned-off-by: Chen Wang \u003cunicorn_wang@outlook.com\u003e\n\n* tag \u0027riscv-sophgo-dt-fixes-for-v6.15-rc1\u0027 of https://github.com/sophgo/linux:\n riscv: dts: sophgo: fix DMA data-width configuration for CV18xx\n\nLink: https://lore.kernel.org/r/MA0P287MB2262454C19B8899BC1694D04FE832@MA0P287MB2262.INDP287.PROD.OUTLOOK.COM\nSigned-off-by: Arnd Bergmann \u003carnd@arndb.de\u003e\n" }, { "commit": "e36f6de6adecafe90b52739adb9405d342bc9ad1", "tree": "835d71110e46e0dd5dc3b3b80029bffd0b17b8f6", "parents": [ "81b7cf868aec04c95daf9996a9625ea2ce8849e0", "0f67578587bb9e5a8eecfcdf6b8a501b5bd90526" ], "author": { "name": "Arnd Bergmann", "email": "arnd@arndb.de", "time": "Fri May 09 18:00:03 2025 +0200" }, "committer": { "name": "Arnd Bergmann", "email": "arnd@arndb.de", "time": "Fri May 09 18:00:03 2025 +0200" }, "message": "Merge tag \u0027amlogic-fixes-for-v6.15\u0027 of https://git.kernel.org/pub/scm/linux/kernel/git/amlogic/linux into arm/fixes\n\nAmlogic Fixes for v6.15:\n- fix reference to unknown/untested PWM clock on ARM/ARM64 boards\n- fix missing clkc_audio node on dreambox ARM64 DT\n\n* tag \u0027amlogic-fixes-for-v6.15\u0027 of https://git.kernel.org/pub/scm/linux/kernel/git/amlogic/linux:\n arm64: dts: amlogic: dreambox: fix missing clkc_audio node\n arm64: dts: amlogic: g12: fix reference to unknown/untested PWM clock\n arm64: dts: amlogic: gx: fix reference to unknown/untested PWM clock\n ARM: dts: amlogic: meson8b: fix reference to unknown/untested PWM clock\n ARM: dts: amlogic: meson8: fix reference to unknown/untested PWM clock\n\nLink: https://lore.kernel.org/r/e9c520a1-b986-49e1-b9b1-67511c187716@linaro.org\nSigned-off-by: Arnd Bergmann \u003carnd@arndb.de\u003e\n" }, { "commit": "81b7cf868aec04c95daf9996a9625ea2ce8849e0", "tree": "ac6a1fc829613e3c754e4fd2d42f0f3311d4d48f", "parents": [ "2ef5c66cba6171feab05e62e1b22df970b238544", "4bf593be2e462623c4c34c7e3b604eb3f8f9de45" ], "author": { "name": "Arnd Bergmann", "email": "arnd@arndb.de", "time": "Fri May 09 17:57:32 2025 +0200" }, "committer": { "name": "Arnd Bergmann", "email": "arnd@arndb.de", "time": "Fri May 09 17:57:39 2025 +0200" }, "message": "Merge tag \u0027v6.15-rockchip-dtsfixes1\u0027 of https://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip into arm/fixes\n\nRemoval of operating-points above what the rk3588j soc is rated for, and\na number of smaller fixes: Turing RK1 fan can spin down again, fixed pins,\npinmuxing and clocks and some devicetree-correctnes improvements.\n\n* tag \u0027v6.15-rockchip-dtsfixes1\u0027 of https://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip:\n arm64: dts: rockchip: fix Sige5 RTC interrupt pin\n arm64: dts: rockchip: Assign RT5616 MCLK rate on rk3588-friendlyelec-cm3588\n arm64: dts: rockchip: Align wifi node name with bindings in CB2\n arm64: dts: rockchip: Fix mmc-pwrseq clock name on rock-pi-4\n arm64: dts: rockchip: Use \"regulator-fixed\" for btreg on px30-engicam for vcc3v3-btreg\n arm64: dts: rockchip: Add pinmuxing for eMMC on QNAP TS433\n arm64: dts: rockchip: Remove overdrive-mode OPPs from RK3588J SoC dtsi\n arm64: dts: rockchip: Allow Turing RK1 cooling fan to spin down\n\nLink: https://lore.kernel.org/r/2923598.88bMQJbFj6@diego\nSigned-off-by: Arnd Bergmann \u003carnd@arndb.de\u003e\n" }, { "commit": "fea4e317f9e7e1f449ce90dedc27a2d2a95bee5a", "tree": "9b5284c25eb5122b00a6e8d822d39c16536c6309", "parents": [ "9c69f88849045499e8ad114e5e13dbb3c85f4443" ], "author": { "name": "Dave Hansen", "email": "dave.hansen@linux.intel.com", "time": "Thu May 08 15:41:32 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri May 09 08:00:31 2025 -0700" }, "message": "x86/mm: Eliminate window where TLB flushes may be inadvertently skipped\n\ntl;dr: There is a window in the mm switching code where the new CR3 is\nset and the CPU should be getting TLB flushes for the new mm. But\nshould_flush_tlb() has a bug and suppresses the flush. Fix it by\nwidening the window where should_flush_tlb() sends an IPI.\n\nLong Version:\n\n\u003d\u003d\u003d History \u003d\u003d\u003d\n\nThere were a few things leading up to this.\n\nFirst, updating mm_cpumask() was observed to be too expensive, so it was\nmade lazier. But being lazy caused too many unnecessary IPIs to CPUs\ndue to the now-lazy mm_cpumask(). So code was added to cull\nmm_cpumask() periodically[2]. But that culling was a bit too aggressive\nand skipped sending TLB flushes to CPUs that need them. So here we are\nagain.\n\n\u003d\u003d\u003d Problem \u003d\u003d\u003d\n\nThe too-aggressive code in should_flush_tlb() strikes in this window:\n\n\t// Turn on IPIs for this CPU/mm combination, but only\n\t// if should_flush_tlb() agrees:\n\tcpumask_set_cpu(cpu, mm_cpumask(next));\n\n\tnext_tlb_gen \u003d atomic64_read(\u0026next-\u003econtext.tlb_gen);\n\tchoose_new_asid(next, next_tlb_gen, \u0026new_asid, \u0026need_flush);\n\tload_new_mm_cr3(need_flush);\n\t// ^ After \u0027need_flush\u0027 is set to false, IPIs *MUST*\n\t// be sent to this CPU and not be ignored.\n\n this_cpu_write(cpu_tlbstate.loaded_mm, next);\n\t// ^ Not until this point does should_flush_tlb()\n\t// become true!\n\nshould_flush_tlb() will suppress TLB flushes between load_new_mm_cr3()\nand writing to \u0027loaded_mm\u0027, which is a window where they should not be\nsuppressed. Whoops.\n\n\u003d\u003d\u003d Solution \u003d\u003d\u003d\n\nThankfully, the fuzzy \"just about to write CR3\" window is already marked\nwith loaded_mm\u003d\u003dLOADED_MM_SWITCHING. Simply checking for that state in\nshould_flush_tlb() is sufficient to ensure that the CPU is targeted with\nan IPI.\n\nThis will cause more TLB flush IPIs. But the window is relatively small\nand I do not expect this to cause any kind of measurable performance\nimpact.\n\nUpdate the comment where LOADED_MM_SWITCHING is written since it grew\nyet another user.\n\nPeter Z also raised a concern that should_flush_tlb() might not observe\n\u0027loaded_mm\u0027 and \u0027is_lazy\u0027 in the same order that switch_mm_irqs_off()\nwrites them. Add a barrier to ensure that they are observed in the\norder they are written.\n\nSigned-off-by: Dave Hansen \u003cdave.hansen@linux.intel.com\u003e\nAcked-by: Rik van Riel \u003criel@surriel.com\u003e\nLink: https://lore.kernel.org/oe-lkp/202411282207.6bd28eae-lkp@intel.com/ [1]\nFixes: 6db2526c1d69 (\"x86/mm/tlb: Only trim the mm_cpumask once a second\") [2]\nReported-by: Stephen Dolan \u003csdolan@janestreet.com\u003e\nCc: stable@vger.kernel.org\nAcked-by: Ingo Molnar \u003cmingo@kernel.org\u003e\nAcked-by: Peter Zijlstra (Intel) \u003cpeterz@infradead.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "c6888983134e2ccc2db8ffd2720b0d4826d952e4", "tree": "02ed522b0b92502ef29af39f38b0604e872fd970", "parents": [ "9379508f0661a5042ab7e0234e604038c2d63f83" ], "author": { "name": "Himanshu Bhavani", "email": "himanshu.bhavani@siliconsignals.io", "time": "Mon May 05 11:28:27 2025 +0530" }, "committer": { "name": "Shawn Guo", "email": "shawnguo@kernel.org", "time": "Fri May 09 22:48:35 2025 +0800" }, "message": "arm64: dts: imx8mp-var-som: Fix LDO5 shutdown causing SD card timeout\n\nFix SD card timeout issue caused by LDO5 regulator getting disabled\nafter boot.\n\nThe kernel log shows LDO5 being disabled, which leads to a timeout\non USDHC2:\n[ 33.760561] LDO5: disabling\n[ 81.119861] mmc1: Timeout waiting for hardware interrupt.\n\nTo prevent this, set regulator-boot-on and regulator-always-on for\nLDO5. Also add the vqmmc regulator to properly support 1.8V/3.3V\nsignaling for USDHC2 using a GPIO-controlled regulator.\n\nFixes: 6c2a1f4f71258 (\"arm64: dts: imx8mp-var-som-symphony: Add Variscite Symphony board and VAR-SOM-MX8MP SoM\")\nSigned-off-by: Himanshu Bhavani \u003chimanshu.bhavani@siliconsignals.io\u003e\nAcked-by: Tarang Raval \u003ctarang.raval@siliconsignals.io\u003e\nSigned-off-by: Shawn Guo \u003cshawnguo@kernel.org\u003e\n" }, { "commit": "92835cebab120f8a5f023a26a792a2ac3f816c4f", "tree": "4fc42b447dd3f7644184bd366aaec24e1581c155", "parents": [ "687b2bae0efff9b25e071737d6af5004e6e35af5" ], "author": { "name": "Gabriel Krisman Bertazi", "email": "krisman@suse.de", "time": "Thu May 08 14:12:03 2025 -0400" }, "committer": { "name": "Jens Axboe", "email": "axboe@kernel.dk", "time": "Fri May 09 07:56:53 2025 -0600" }, "message": "io_uring/sqpoll: Increase task_work submission batch size\n\nOur QA team reported a 10%-23%, throughput reduction on an io_uring\nsqpoll testcase doing IO to a null_blk, that I traced back to a\nreduction of the device submission queue depth utilization. It turns out\nthat, after commit af5d68f8892f (\"io_uring/sqpoll: manage task_work\nprivately\"), we capped the number of task_work entries that can be\ncompleted from a single spin of sqpoll to only 8 entries, before the\nsqpoll goes around to (potentially) sleep. While this cap doesn\u0027t drive\nthe submission side directly, it impacts the completion behavior, which\naffects the number of IO queued by fio per sqpoll cycle on the\nsubmission side, and io_uring ends up seeing less ios per sqpoll cycle.\nAs a result, block layer plugging is less effective, and we see more\ntime spent inside the block layer in profilings charts, and increased\nsubmission latency measured by fio.\n\nThere are other places that have increased overhead once sqpoll sleeps\nmore often, such as the sqpoll utilization calculation. But, in this\nmicrobenchmark, those were not representative enough in perf charts, and\ntheir removal didn\u0027t yield measurable changes in throughput. The major\noverhead comes from the fact we plug less, and less often, when submitting\nto the block layer.\n\nMy benchmark is:\n\nfio --ioengine\u003dio_uring --direct\u003d1 --iodepth\u003d128 --runtime\u003d300 --bs\u003d4k \\\n --invalidate\u003d1 --time_based --ramp_time\u003d10 --group_reporting\u003d1 \\\n --filename\u003d/dev/nullb0 --name\u003dRandomReads-direct-nullb-sqpoll-4k-1 \\\n --rw\u003drandread --numjobs\u003d1 --sqthread_poll\n\nIn one machine, tested on top of Linux 6.15-rc1, we have the following\nbaseline:\n READ: bw\u003d4994MiB/s (5236MB/s), 4994MiB/s-4994MiB/s (5236MB/s-5236MB/s), io\u003d439GiB (471GB), run\u003d90001-90001msec\n\nWith this patch:\n READ: bw\u003d5762MiB/s (6042MB/s), 5762MiB/s-5762MiB/s (6042MB/s-6042MB/s), io\u003d506GiB (544GB), run\u003d90001-90001msec\n\nwhich is a 15% improvement in measured bandwidth. The average\nsubmission latency is noticeably lowered too. As measured by\nfio:\n\nBaseline:\n lat (usec): min\u003d20, max\u003d241, avg\u003d99.81, stdev\u003d3.38\nPatched:\n lat (usec): min\u003d26, max\u003d226, avg\u003d86.48, stdev\u003d4.82\n\nIf we look at blktrace, we can also see the plugging behavior is\nimproved. In the baseline, we end up limited to plugging 8 requests in\nthe block layer regardless of the device queue depth size, while after\npatching we can drive more io, and we manage to utilize the full device\nqueue.\n\nIn the baseline, after a stabilization phase, an ordinary submission\nlooks like:\n 254,0 1 49942 0.016028795 5977 U N [iou-sqp-5976] 7\n\nAfter patching, I see consistently more requests per unplug.\n 254,0 1 4996 0.001432872 3145 U N [iou-sqp-3144] 32\n\nIdeally, the cap size would at least be the deep enough to fill the\ndevice queue, but we can\u0027t predict that behavior, or assume all IO goes\nto a single device, and thus can\u0027t guess the ideal batch size. We also\ndon\u0027t want to let the tw run unbounded, though I\u0027m not sure it would\nreally be a problem. Instead, let\u0027s just give it a more sensible value\nthat will allow for more efficient batching. I\u0027ve tested with different\ncap values, and initially proposed to increase the cap to 1024. Jens\nargued it is too big of a bump and I observed that, with 32, I\u0027m no\nlonger able to observe this bottleneck in any of my machines.\n\nFixes: af5d68f8892f (\"io_uring/sqpoll: manage task_work privately\")\nSigned-off-by: Gabriel Krisman Bertazi \u003ckrisman@suse.de\u003e\nLink: https://lore.kernel.org/r/20250508181203.3785544-1-krisman@suse.de\nSigned-off-by: Jens Axboe \u003caxboe@kernel.dk\u003e\n" }, { "commit": "04679f3c27e132c1a2d3881de2f0c5d7128de7c1", "tree": "33f6b82e7d9fab09e0eef9d0698642da7e34d55b", "parents": [ "6ba0982c3235a047c953bf73a7b014af7840c4de" ], "author": { "name": "Jeremy Bongio", "email": "jbongio@google.com", "time": "Wed May 07 12:30:10 2025 +0000" }, "committer": { "name": "Christian Brauner", "email": "brauner@kernel.org", "time": "Fri May 09 12:31:57 2025 +0200" }, "message": "fs: Remove redundant errseq_set call in mark_buffer_write_io_error.\n\nmark_buffer_write_io_error sets sb-\u003es_wb_err to -EIO twice.\nOnce in mapping_set_error and once in errseq_set.\nOnly mapping_set_error checks if bh-\u003eb_assoc_map-\u003ehost is NULL.\n\nDiscovered during null pointer dereference during writeback\nto a failing device:\n\n[\u003cffffffff9a416dc8\u003e] ? mark_buffer_write_io_error+0x98/0xc0\n[\u003cffffffff9a416dbe\u003e] ? mark_buffer_write_io_error+0x8e/0xc0\n[\u003cffffffff9ad4bda0\u003e] end_buffer_async_write+0x90/0xd0\n[\u003cffffffff9ad4e3eb\u003e] end_bio_bh_io_sync+0x2b/0x40\n[\u003cffffffff9adbafe6\u003e] blk_update_request+0x1b6/0x480\n[\u003cffffffff9adbb3d8\u003e] blk_mq_end_request+0x18/0x30\n[\u003cffffffff9adbc6aa\u003e] blk_mq_dispatch_rq_list+0x4da/0x8e0\n[\u003cffffffff9adc0a68\u003e] __blk_mq_sched_dispatch_requests+0x218/0x6a0\n[\u003cffffffff9adc07fa\u003e] blk_mq_sched_dispatch_requests+0x3a/0x80\n[\u003cffffffff9adbbb98\u003e] blk_mq_run_hw_queue+0x108/0x330\n[\u003cffffffff9adbcf58\u003e] blk_mq_flush_plug_list+0x178/0x5f0\n[\u003cffffffff9adb6741\u003e] __blk_flush_plug+0x41/0x120\n[\u003cffffffff9adb6852\u003e] blk_finish_plug+0x22/0x40\n[\u003cffffffff9ad47cb0\u003e] wb_writeback+0x150/0x280\n[\u003cffffffff9ac5343f\u003e] ? set_worker_desc+0x9f/0xc0\n[\u003cffffffff9ad4676e\u003e] wb_workfn+0x24e/0x4a0\n\nFixes: 485e9605c0573 (\"fs/buffer.c: record blockdev write errors in super_block that it backs\")\nSigned-off-by: Jeremy Bongio \u003cjbongio@google.com\u003e\nLink: https://lore.kernel.org/20250507123010.1228243-1-jbongio@google.com\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Christian Brauner \u003cbrauner@kernel.org\u003e\n" }, { "commit": "9379508f0661a5042ab7e0234e604038c2d63f83", "tree": "958b53276a8c33e95ba5f91fc090371f5e49bac2", "parents": [ "5591ce0069ddda97cdbbea596bed53e698f399c2" ], "author": { "name": "Ahmad Fatoum", "email": "a.fatoum@pengutronix.de", "time": "Tue Apr 22 09:12:35 2025 +0200" }, "committer": { "name": "Shawn Guo", "email": "shawnguo@kernel.org", "time": "Fri May 09 17:53:20 2025 +0800" }, "message": "arm64: dts: imx8mp: use 800MHz NoC OPP for nominal drive mode\n\nWhen running in nominal drive mode, the maximum allowed frequency for\nthe NoC is 800MHz, but the OPP table for the i.MX8MP interconnect device\nlisted the 1GHz operating point for the NoC, regardless of the active\nmode.\n\nThe newly introduced imx8mp-nominal.dtsi header reconfigures the clock\ncontroller to observe nominal drive mode limits, so have it modify the\nmaximum NoC OPP as well.\n\nFixes: 255fbd9eabe7 (\"arm64: dts: imx8mp: Add optional nominal drive mode DTSI\")\nSigned-off-by: Ahmad Fatoum \u003ca.fatoum@pengutronix.de\u003e\nSigned-off-by: Shawn Guo \u003cshawnguo@kernel.org\u003e\n" }, { "commit": "732b87a409667a370b87955c518e5d004de740b5", "tree": "23b1bbce8bb31d1c6e28cd383ba8882b87a50bd1", "parents": [ "936b73feab5cd7eae8fe3d08a7ac9b1f8ac68042" ], "author": { "name": "Imre Deak", "email": "imre.deak@intel.com", "time": "Wed May 07 18:19:53 2025 +0300" }, "committer": { "name": "Jani Nikula", "email": "jani.nikula@intel.com", "time": "Fri May 09 10:15:52 2025 +0300" }, "message": "drm/i915/dp: Fix determining SST/MST mode during MTP TU state computation\n\nDetermining the SST/MST mode during state computation must be done based\non the output type stored in the CRTC state, which in turn is set once\nbased on the modeset connector\u0027s SST vs. MST type and will not change as\nlong as the connector is using the CRTC. OTOH the MST mode indicated by\nthe given connector\u0027s intel_dp::is_mst flag can change independently of\nthe above output type, based on what sink is at any moment plugged to\nthe connector.\n\nFix the state computation accordingly.\n\nCc: Jani Nikula \u003cjani.nikula@intel.com\u003e\nFixes: f6971d7427c2 (\"drm/i915/mst: adapt intel_dp_mtp_tu_compute_config() for 128b/132b SST\")\nCloses: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/4607\nReviewed-by: Jani Nikula \u003cjani.nikula@intel.com\u003e\nSigned-off-by: Imre Deak \u003cimre.deak@intel.com\u003e\nLink: https://lore.kernel.org/r/20250507151953.251846-1-imre.deak@intel.com\n(cherry picked from commit 0f45696ddb2b901fbf15cb8d2e89767be481d59f)\nSigned-off-by: Jani Nikula \u003cjani.nikula@intel.com\u003e\n" }, { "commit": "da8bf5daa5e55a6af2b285ecda460d6454712ff4", "tree": "52bc01fc8c982112a061daad80c29a126f75598c", "parents": [ "42e31f0daf80d9f7bc4ab4000f2795ec3ddf5206" ], "author": { "name": "Tom Lendacky", "email": "thomas.lendacky@amd.com", "time": "Thu May 08 12:24:10 2025 -0500" }, "committer": { "name": "Mike Rapoport (Microsoft)", "email": "rppt@kernel.org", "time": "Fri May 09 08:53:12 2025 +0300" }, "message": "memblock: Accept allocated memory before use in memblock_double_array()\n\nWhen increasing the array size in memblock_double_array() and the slab\nis not yet available, a call to memblock_find_in_range() is used to\nreserve/allocate memory. However, the range returned may not have been\naccepted, which can result in a crash when booting an SNP guest:\n\n RIP: 0010:memcpy_orig+0x68/0x130\n Code: ...\n RSP: 0000:ffffffff9cc03ce8 EFLAGS: 00010006\n RAX: ff11001ff83e5000 RBX: 0000000000000000 RCX: fffffffffffff000\n RDX: 0000000000000bc0 RSI: ffffffff9dba8860 RDI: ff11001ff83e5c00\n RBP: 0000000000002000 R08: 0000000000000000 R09: 0000000000002000\n R10: 000000207fffe000 R11: 0000040000000000 R12: ffffffff9d06ef78\n R13: ff11001ff83e5000 R14: ffffffff9dba7c60 R15: 0000000000000c00\n memblock_double_array+0xff/0x310\n memblock_add_range+0x1fb/0x2f0\n memblock_reserve+0x4f/0xa0\n memblock_alloc_range_nid+0xac/0x130\n memblock_alloc_internal+0x53/0xc0\n memblock_alloc_try_nid+0x3d/0xa0\n swiotlb_init_remap+0x149/0x2f0\n mem_init+0xb/0xb0\n mm_core_init+0x8f/0x350\n start_kernel+0x17e/0x5d0\n x86_64_start_reservations+0x14/0x30\n x86_64_start_kernel+0x92/0xa0\n secondary_startup_64_no_verify+0x194/0x19b\n\nMitigate this by calling accept_memory() on the memory range returned\nbefore the slab is available.\n\nPrior to v6.12, the accept_memory() interface used a \u0027start\u0027 and \u0027end\u0027\nparameter instead of \u0027start\u0027 and \u0027size\u0027, therefore the accept_memory()\ncall must be adjusted to specify \u0027start + size\u0027 for \u0027end\u0027 when applying\nto kernels prior to v6.12.\n\nCc: stable@vger.kernel.org # see patch description, needs adjustments for \u003c\u003d 6.11\nFixes: dcdfdd40fa82 (\"mm: Add support for unaccepted memory\")\nSigned-off-by: Tom Lendacky \u003cthomas.lendacky@amd.com\u003e\nLink: https://lore.kernel.org/r/da1ac73bf4ded761e21b4e4bb5178382a580cd73.1746725050.git.thomas.lendacky@amd.com\nSigned-off-by: Mike Rapoport (Microsoft) \u003crppt@kernel.org\u003e\n" }, { "commit": "80e12f3e2a5a911fb67597da4c17af48e7cb92de", "tree": "169cf6e710ab3b4d3b9d8cee1463352b7d1e747e", "parents": [ "20a4c8142e9e16700cf9d3c43bf7808c531d9dd6", "5a11a2767731139bf87e667331aa2209e33a1d19" ], "author": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Fri May 09 11:10:41 2025 +1000" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Fri May 09 11:11:25 2025 +1000" }, "message": "Merge tag \u0027amd-drm-fixes-6.15-2025-05-08\u0027 of https://gitlab.freedesktop.org/agd5f/linux into drm-fixes\n\namd-drm-fixes-6.15-2025-05-08:\n\namdgpu:\n- DC FP fixes\n- Freesync fix\n- DMUB AUX fixes\n- VCN fix\n- Hibernation fixes\n- HDP fixes\n\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n\nFrom: Alex Deucher \u003calexander.deucher@amd.com\u003e\nLink: https://lore.kernel.org/r/20250508194102.3242372-1-alexander.deucher@amd.com\n" }, { "commit": "20a4c8142e9e16700cf9d3c43bf7808c531d9dd6", "tree": "42b7b40b6da960da60ea7563b9db7f9b950276c5", "parents": [ "92a09c47464d040866cf2b4cd052bc60555185fb", "7c6fa1797a725732981f2d77711c867166737719" ], "author": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Fri May 09 08:51:57 2025 +1000" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Fri May 09 08:52:24 2025 +1000" }, "message": "Merge tag \u0027drm-misc-fixes-2025-05-08\u0027 of https://gitlab.freedesktop.org/drm/misc/kernel into drm-fixes\n\nShort summary of fixes pull:\n\ndrm:\n- Fix overflow when generating wedged event\n\nivpu:\n- Increate timeouts\n- Fix deadlock in cmdq ioctl\n- Unlock mutices in correct order\n\npanel:\n- simple: Fix timings for AUO G101EVN010\n\nttm:\n- Fix documentation\n- Remove struct ttm_backup\n\nv3d:\n- Avoid memory leak in job handling\n\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n\nFrom: Thomas Zimmermann \u003ctzimmermann@suse.de\u003e\nLink: https://lore.kernel.org/r/20250508104939.GA76697@2a02-2454-fd5e-fd00-c110-cbf2-6528-c5be.dyn6.pyur.net\n" }, { "commit": "9c69f88849045499e8ad114e5e13dbb3c85f4443", "tree": "b5bf5522b1cf16441986e9ed7ccabfd5a039b370", "parents": [ "acaa3e726f4a29f32bca5146828565db56bc396f", "8e4d28036c293241b312b1fceafb32b994f80fcc" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 08 14:28:49 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 08 14:28:49 2025 -0700" }, "message": "Merge tag \u0027bcachefs-2025-05-08\u0027 of git://evilpiepirate.org/bcachefs\n\nPull bcachefs fixes from Kent Overstreet:\n\n - Some fixes to help with filesystem analysis: ensure superblock\n error count gets written if we go ERO, don\u0027t discard the journal\n aggressively (so it\u0027s available for list_journal -a)\n\n - Fix lost wakeup on arm causing us to get stuck when reading btree\n nodes\n\n - Fix fsck failing to exit on ctrl-c\n\n - An additional fix for filesystems with misaligned bucket sizes: we\n now ensure that allocations are properly aligned\n\n - Setting background target but not promote target will now leave that\n data cached on the foreground target, as it used to\n\n - Revert a change to when we allocate the VFS superblock, this was done\n for implementing blk_holder_ops but ended up not being needed, and\n allocating a superblock and not setting SB_BORN while we do recovery\n caused sync() calls and other things to hang\n\n - Assorted fixes for harmless error messages that caused concern to\n users\n\n* tag \u0027bcachefs-2025-05-08\u0027 of git://evilpiepirate.org/bcachefs:\n bcachefs: Don\u0027t aggressively discard the journal\n bcachefs: Ensure superblock gets written when we go ERO\n bcachefs: Filter out harmless EROFS error messages\n bcachefs: journal_shutdown is EROFS, not EIO\n bcachefs: Call bch2_fs_start before getting vfs superblock\n bcachefs: fix hung task timeout in journal read\n bcachefs: Add missing barriers before wake_up_bit()\n bcachefs: Ensure proper write alignment\n bcachefs: Improve want_cached_ptr()\n bcachefs: thread_with_stdio: fix spinning instead of exiting\n" }, { "commit": "564467e9d06c6352fac9100e8957d40a1a50234c", "tree": "1fc660de5cfe3791613fb47f3aab40f9546d9832", "parents": [ "9d271a4f5ba52520e448ab223b1a91c6e35f17c7" ], "author": { "name": "Shuicheng Lin", "email": "shuicheng.lin@intel.com", "time": "Fri May 02 17:00:52 2025 +0000" }, "committer": { "name": "Lucas De Marchi", "email": "lucas.demarchi@intel.com", "time": "Thu May 08 13:57:52 2025 -0700" }, "message": "drm/xe: Add config control for svm flush work\n\nWithout CONFIG_DRM_XE_GPUSVM set, GPU SVM is not initialized thus below\nwarning pops. Refine the flush work code to be controlled by the config\nto avoid below warning:\n\"\n[ 453.132028] ------------[ cut here ]------------\n[ 453.132527] WARNING: CPU: 9 PID: 4491 at kernel/workqueue.c:4205 __flush_work+0x379/0x3a0\n[ 453.133355] Modules linked in: xe drm_ttm_helper ttm gpu_sched drm_buddy drm_suballoc_helper drm_gpuvm drm_exec\n[ 453.134352] CPU: 9 UID: 0 PID: 4491 Comm: xe_exec_mix_mod Tainted: G U W 6.15.0-rc3+ #7 PREEMPT(full)\n[ 453.135405] Tainted: [U]\u003dUSER, [W]\u003dWARN\n...\n[ 453.136921] RIP: 0010:__flush_work+0x379/0x3a0\n[ 453.137417] Code: 8b 45 00 48 8b 55 08 89 c7 48 c1 e8 04 83 e7 08 83 e0 0f 83 cf 02 89 c6 48 0f ba 6d 00 03 e9 d5 fe ff ff 0f 0b e9 db fd ff ff \u003c0f\u003e 0b 45 31 e4 e9 d1 fd ff ff 0f 0b e9 03 ff ff ff 0f 0b e9 d6 fe\n[ 453.139250] RSP: 0018:ffffc90000c67b18 EFLAGS: 00010246\n[ 453.139782] RAX: 0000000000000000 RBX: ffff888108a24000 RCX: 0000000000002000\n[ 453.140521] RDX: 0000000000000001 RSI: 0000000000000000 RDI: ffff8881016d61c8\n[ 453.141253] RBP: ffff8881016d61c8 R08: 0000000000000000 R09: 0000000000000000\n[ 453.141985] R10: 0000000000000000 R11: 0000000008a24000 R12: 0000000000000001\n[ 453.142709] R13: 0000000000000002 R14: 0000000000000000 R15: ffff888107db8c00\n[ 453.143450] FS: 00007f44853d4c80(0000) GS:ffff8882f469b000(0000) knlGS:0000000000000000\n[ 453.144276] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033\n[ 453.144853] CR2: 00007f4487629228 CR3: 00000001016aa000 CR4: 00000000000406f0\n[ 453.145594] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000\n[ 453.146320] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400\n[ 453.147061] Call Trace:\n[ 453.147336] \u003cTASK\u003e\n[ 453.147579] ? tick_nohz_tick_stopped+0xd/0x30\n[ 453.148067] ? xas_load+0x9/0xb0\n[ 453.148435] ? xa_load+0x6f/0xb0\n[ 453.148781] __xe_vm_bind_ioctl+0xbd5/0x1500 [xe]\n[ 453.149338] ? dev_printk_emit+0x48/0x70\n[ 453.149762] ? _dev_printk+0x57/0x80\n[ 453.150148] ? drm_ioctl+0x17c/0x440\n[ 453.150544] ? __drm_dev_vprintk+0x36/0x90\n[ 453.150983] ? __pfx_xe_vm_bind_ioctl+0x10/0x10 [xe]\n[ 453.151575] ? drm_ioctl_kernel+0x9f/0xf0\n[ 453.151998] ? __pfx_xe_vm_bind_ioctl+0x10/0x10 [xe]\n[ 453.152560] drm_ioctl_kernel+0x9f/0xf0\n[ 453.152968] drm_ioctl+0x20f/0x440\n[ 453.153332] ? __pfx_xe_vm_bind_ioctl+0x10/0x10 [xe]\n[ 453.153893] ? ioctl_has_perm.constprop.0.isra.0+0xae/0x100\n[ 453.154489] ? memory_bm_test_bit+0x5/0x60\n[ 453.154935] xe_drm_ioctl+0x47/0x70 [xe]\n[ 453.155419] __x64_sys_ioctl+0x8d/0xc0\n[ 453.155824] do_syscall_64+0x47/0x110\n[ 453.156228] entry_SYSCALL_64_after_hwframe+0x76/0x7e\n\"\n\nv2 (Matt):\n refine commit message to have more details\n add Fixes tag\n move the code to xe_svm.h which already have the config\n remove a blank line per codestyle suggestion\n\nFixes: 63f6e480d115 (\"drm/xe: Add SVM garbage collector\")\nCc: Matthew Brost \u003cmatthew.brost@intel.com\u003e\nSigned-off-by: Shuicheng Lin \u003cshuicheng.lin@intel.com\u003e\nReviewed-by: Matthew Brost \u003cmatthew.brost@intel.com\u003e\nSigned-off-by: Matthew Brost \u003cmatthew.brost@intel.com\u003e\nLink: https://lore.kernel.org/r/20250502170052.1787973-1-shuicheng.lin@intel.com\n(cherry picked from commit 9d80698bcd97a5ad1088bcbb055e73fd068895e2)\nSigned-off-by: Lucas De Marchi \u003clucas.demarchi@intel.com\u003e\n" }, { "commit": "9d271a4f5ba52520e448ab223b1a91c6e35f17c7", "tree": "ab5e6da7545757534d0b95b1c04e718d0597f6b4", "parents": [ "03552d8ac0afcc080c339faa0b726e2c0e9361cb" ], "author": { "name": "Shuicheng Lin", "email": "shuicheng.lin@intel.com", "time": "Wed May 07 02:23:02 2025 +0000" }, "committer": { "name": "Lucas De Marchi", "email": "lucas.demarchi@intel.com", "time": "Thu May 08 12:44:51 2025 -0700" }, "message": "drm/xe: Release force wake first then runtime power\n\nxe_force_wake_get() is dependent on xe_pm_runtime_get(), so for\nthe release path, xe_force_wake_put() should be called first then\nxe_pm_runtime_put().\nCombine the error path and normal path together with goto.\n\nFixes: 85d547608ef5 (\"drm/xe/xe_gt_debugfs: Update handling of xe_force_wake_get return\")\nCc: Himal Prasad Ghimiray \u003chimal.prasad.ghimiray@intel.com\u003e\nCc: Rodrigo Vivi \u003crodrigo.vivi@intel.com\u003e\nSigned-off-by: Shuicheng Lin \u003cshuicheng.lin@intel.com\u003e\nReviewed-by: Himal Prasad Ghimiray \u003chimal.prasad.ghimiray@intel.com\u003e\nLink: https://lore.kernel.org/r/20250507022302.2187527-1-shuicheng.lin@intel.com\nSigned-off-by: Rodrigo Vivi \u003crodrigo.vivi@intel.com\u003e\n(cherry picked from commit 432cd94efdca06296cc5e76d673546f58aa90ee1)\nSigned-off-by: Lucas De Marchi \u003clucas.demarchi@intel.com\u003e\n" }, { "commit": "03552d8ac0afcc080c339faa0b726e2c0e9361cb", "tree": "897af38e7b7788cc9166476c669face37cb45967", "parents": [ "51c0ee84e4dc339287b2d7335f2b54d747794c83" ], "author": { "name": "Daniele Ceraolo Spurio", "email": "daniele.ceraolospurio@intel.com", "time": "Fri May 02 08:51:04 2025 -0700" }, "committer": { "name": "Lucas De Marchi", "email": "lucas.demarchi@intel.com", "time": "Thu May 08 12:44:41 2025 -0700" }, "message": "drm/xe/gsc: do not flush the GSC worker from the reset path\n\nThe workqueue used for the reset worker is marked as WQ_MEM_RECLAIM,\nwhile the GSC one isn\u0027t (and can\u0027t be as we need to do memory\nallocations in the gsc worker). Therefore, we can\u0027t flush the latter\nfrom the former.\n\nThe reason why we had such a flush was to avoid interrupting either\nthe GSC FW load or in progress GSC proxy operations. GSC proxy\noperations fall into 2 categories:\n\n1) GSC proxy init: this only happens once immediately after GSC FW load\n and does not support being interrupted. The only way to recover from\n an interruption of the proxy init is to do an FLR and re-load the GSC.\n\n2) GSC proxy request: this can happen in response to a request that\n the driver sends to the GSC. If this is interrupted, the GSC FW will\n timeout and the driver request will be failed, but overall the GSC\n will keep working fine.\n\nFlushing the work allowed us to avoid interruption in both cases (unless\nthe hang came from the GSC engine itself, in which case we\u0027re toast\nanyway). However, a failure on a proxy request is tolerable if we\u0027re in\na scenario where we\u0027re triggering a GT reset (i.e., something is already\ngone pretty wrong), so what we really need to avoid is interrupting\nthe init flow, which we can do by polling on the register that reports\nwhen the proxy init is complete (as that ensure us that all the load and\ninit operations have been completed).\n\nNote that during suspend we still want to do a flush of the worker to\nmake sure it completes any operations involving the HW before the power\nis cut.\n\nv2: fix spelling in commit msg, rename waiter function (Julia)\n\nFixes: dd0e89e5edc2 (\"drm/xe/gsc: GSC FW load\")\nCloses: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/4830\nSigned-off-by: Daniele Ceraolo Spurio \u003cdaniele.ceraolospurio@intel.com\u003e\nCc: John Harrison \u003cJohn.C.Harrison@Intel.com\u003e\nCc: Alan Previn \u003calan.previn.teres.alexis@intel.com\u003e\nCc: \u003cstable@vger.kernel.org\u003e # v6.8+\nReviewed-by: Julia Filipchuk \u003cjulia.filipchuk@intel.com\u003e\nLink: https://lore.kernel.org/r/20250502155104.2201469-1-daniele.ceraolospurio@intel.com\n(cherry picked from commit 12370bfcc4f0bdf70279ec5b570eb298963422b5)\nSigned-off-by: Lucas De Marchi \u003clucas.demarchi@intel.com\u003e\n" }, { "commit": "51c0ee84e4dc339287b2d7335f2b54d747794c83", "tree": "b3bb7111063116e330983d9d39eed0ec9a9f5ff1", "parents": [ "391008f34e711253c5983b0bf52277cc43723127" ], "author": { "name": "Tejas Upadhyay", "email": "tejas.upadhyay@intel.com", "time": "Mon Apr 28 13:53:57 2025 +0530" }, "committer": { "name": "Lucas De Marchi", "email": "lucas.demarchi@intel.com", "time": "Thu May 08 12:44:34 2025 -0700" }, "message": "drm/xe/tests/mocs: Hold XE_FORCEWAKE_ALL for LNCF regs\n\nLNCF registers report wrong values when XE_FORCEWAKE_GT\nonly is held. Holding XE_FORCEWAKE_ALL ensures correct\noperations on LNCF regs.\n\nV2(Himal):\n - Use xe_force_wake_ref_has_domain\n\nCloses: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/1999\nFixes: a6a4ea6d7d37 (\"drm/xe: Add mocs kunit\")\nReviewed-by: Himal Prasad Ghimiray \u003chimal.prasad.ghimiray@intel.com\u003e\nLink: https://patchwork.freedesktop.org/patch/msgid/20250428082357.1730068-1-tejas.upadhyay@intel.com\nSigned-off-by: Tejas Upadhyay \u003ctejas.upadhyay@intel.com\u003e\n(cherry picked from commit 70a2585e582058e94fe4381a337be42dec800337)\nSigned-off-by: Lucas De Marchi \u003clucas.demarchi@intel.com\u003e\n" }, { "commit": "391008f34e711253c5983b0bf52277cc43723127", "tree": "17b45fb9845cbb8b89a0bfce7d1c7da1debb9f35", "parents": [ "92a09c47464d040866cf2b4cd052bc60555185fb" ], "author": { "name": "Matthew Brost", "email": "matthew.brost@intel.com", "time": "Tue Apr 08 08:59:15 2025 -0700" }, "committer": { "name": "Lucas De Marchi", "email": "lucas.demarchi@intel.com", "time": "Thu May 08 12:44:20 2025 -0700" }, "message": "drm/xe: Add page queue multiplier\n\nFor an unknown reason the math to determine the PF queue size does is\nnot correct - compute UMD applications are overflowing the PF queue\nwhich is fatal. A multippier of 8 fixes the problem.\n\nFixes: 3338e4f90c14 (\"drm/xe: Use topology to determine page fault queue size\")\nCc: stable@vger.kernel.org\nSigned-off-by: Matthew Brost \u003cmatthew.brost@intel.com\u003e\nReviewed-by: Jagmeet Randhawa \u003cjagmeet.randhawa@intel.com\u003e\nLink: https://lore.kernel.org/r/20250408155915.78770-1-matthew.brost@intel.com\n(cherry picked from commit 29582e0ea75c95668d168b12406e3c56cf5a73c4)\nSigned-off-by: Lucas De Marchi \u003clucas.demarchi@intel.com\u003e\n" }, { "commit": "acaa3e726f4a29f32bca5146828565db56bc396f", "tree": "50084e7c68e7393011f1d54fc27fb1f596d6189d", "parents": [ "2c89c1b655c0b06823f4ee8b055140d8628fc4da", "c1d9dac0db168198b6f63f460665256dedad9b6e" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 08 12:09:22 2025 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 08 12:09:22 2025 -0700" }, "message": "Merge tag \u0027vfio-v6.15-rc6\u0027 of https://github.com/awilliam/linux-vfio\n\nPull vfio fix from Alex Williamson:\n\n - Fix an issue in vfio-pci huge_fault handling by aligning faults to\n the order, resulting in deterministic use of huge pages. This\n avoids a race where simultaneous aligned and unaligned faults to\n the same PMD can result in a VM_FAULT_OOM and subsequent VM crash.\n (Alex Williamson)\n\n* tag \u0027vfio-v6.15-rc6\u0027 of https://github.com/awilliam/linux-vfio:\n vfio/pci: Align huge faults to order\n" }, { "commit": "01534f3e0dd75e27ed03e5542f0c7bf6aa7130f1", "tree": "40bce945112ae5dd72edf24f2d951b2e27c765f6", "parents": [ "92a09c47464d040866cf2b4cd052bc60555185fb", "7f1c3de1370bc6a8ad5157336b258067dac0ae9c" ], "author": { "name": "Palmer Dabbelt", "email": "palmer@rivosinc.com", "time": "Thu May 08 09:40:21 2025 -0700" }, "committer": { "name": "Palmer Dabbelt", "email": "palmer@rivosinc.com", "time": "Thu May 08 09:40:21 2025 -0700" }, "message": "Merge tag \u0027riscv-fixes-6.15-rc6\u0027 of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/alexghiti/linux into fixes\n\nriscv fixes for 6.15-rc6\n\n- A fix to handle compressed halfword load/store instructions misaligned accesses\n- A fix to allow user memory access while handling a misaligned access\n- 2 fixes to return an error if the pointer masking extension is not implemented on the platform but userspace still tries to access it, which caused oops on some early platforms\n- A fix to prevent the stripping of .rela.dyn so that a vmlinux loaded by kexec can successfully boot\n\n* tag \u0027riscv-fixes-6.15-rc6\u0027 of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/alexghiti/linux:\n riscv: Disallow PR_GET_TAGGED_ADDR_CTRL without Supm\n scripts: Do not strip .rela.dyn section\n riscv: Fix kernel crash due to PR_SET_TAGGED_ADDR_CTRL\n riscv: misaligned: use get_user() instead of __get_user()\n riscv: misaligned: enable IRQs while handling misaligned accesses\n riscv: misaligned: factorize trap handling\n riscv: misaligned: Add handling for ZCB instructions\n" }, { "commit": "5a11a2767731139bf87e667331aa2209e33a1d19", "tree": "15579db5a0719a29dc82a1ac48b0a0806227ce71", "parents": [ "ca28e80abe4219c8f1a2961ae05102d70af6dc87" ], "author": { "name": "Alex Deucher", "email": "alexander.deucher@amd.com", "time": "Wed Apr 30 12:50:02 2025 -0400" }, "committer": { "name": "Alex Deucher", "email": "alexander.deucher@amd.com", "time": "Thu May 08 11:48:12 2025 -0400" }, "message": "drm/amdgpu/hdp7: use memcfg register to post the write for HDP flush\n\nReading back the remapped HDP flush register seems to cause\nproblems on some platforms. All we need is a read, so read back\nthe memcfg register.\n\nFixes: 689275140cb8 (\"drm/amdgpu/hdp7.0: do a posting read when flushing HDP\")\nReported-by: Alexey Klimov \u003calexey.klimov@linaro.org\u003e\nLink: https://lists.freedesktop.org/archives/amd-gfx/2025-April/123150.html\nCloses: https://gitlab.freedesktop.org/drm/amd/-/issues/4119\nCloses: https://gitlab.freedesktop.org/drm/amd/-/issues/3908\nReviewed-by: Felix Kuehling \u003cfelix.kuehling@amd.com\u003e\nSigned-off-by: Alex Deucher \u003calexander.deucher@amd.com\u003e\n(cherry picked from commit dbc064adfcf9095e7d895bea87b2f75c1ab23236)\nCc: stable@vger.kernel.org\n" }, { "commit": "ca28e80abe4219c8f1a2961ae05102d70af6dc87", "tree": "0f39696efe8e93de37ee1a29fc41f68b7e9839ad", "parents": [ "dbc988c689333faeeed44d5561f372ff20395304" ], "author": { "name": "Alex Deucher", "email": "alexander.deucher@amd.com", "time": "Wed Apr 30 12:48:51 2025 -0400" }, "committer": { "name": "Alex Deucher", "email": "alexander.deucher@amd.com", "time": "Thu May 08 11:47:54 2025 -0400" }, "message": "drm/amdgpu/hdp6: use memcfg register to post the write for HDP flush\n\nReading back the remapped HDP flush register seems to cause\nproblems on some platforms. All we need is a read, so read back\nthe memcfg register.\n\nFixes: abe1cbaec6cf (\"drm/amdgpu/hdp6.0: do a posting read when flushing HDP\")\nReported-by: Alexey Klimov \u003calexey.klimov@linaro.org\u003e\nLink: https://lists.freedesktop.org/archives/amd-gfx/2025-April/123150.html\nCloses: https://gitlab.freedesktop.org/drm/amd/-/issues/4119\nCloses: https://gitlab.freedesktop.org/drm/amd/-/issues/3908\nReviewed-by: Felix Kuehling \u003cfelix.kuehling@amd.com\u003e\nSigned-off-by: Alex Deucher \u003calexander.deucher@amd.com\u003e\n(cherry picked from commit 84141ff615951359c9a99696fd79a36c465ed847)\nCc: stable@vger.kernel.org\n" }, { "commit": "dbc988c689333faeeed44d5561f372ff20395304", "tree": "39fc5286b57491b8846b323837e14d5406eb85c5", "parents": [ "0e33e0f339b91eecd9558311449a3d1e728722d4" ], "author": { "name": "Alex Deucher", "email": "alexander.deucher@amd.com", "time": "Wed Apr 30 12:47:37 2025 -0400" }, "committer": { "name": "Alex Deucher", "email": "alexander.deucher@amd.com", "time": "Thu May 08 11:47:23 2025 -0400" }, "message": "drm/amdgpu/hdp5.2: use memcfg register to post the write for HDP flush\n\nReading back the remapped HDP flush register seems to cause\nproblems on some platforms. All we need is a read, so read back\nthe memcfg register.\n\nFixes: f756dbac1ce1 (\"drm/amdgpu/hdp5.2: do a posting read when flushing HDP\")\nReported-by: Alexey Klimov \u003calexey.klimov@linaro.org\u003e\nLink: https://lists.freedesktop.org/archives/amd-gfx/2025-April/123150.html\nCloses: https://gitlab.freedesktop.org/drm/amd/-/issues/4119\nCloses: https://gitlab.freedesktop.org/drm/amd/-/issues/3908\nReviewed-by: Felix Kuehling \u003cfelix.kuehling@amd.com\u003e\nSigned-off-by: Alex Deucher \u003calexander.deucher@amd.com\u003e\n(cherry picked from commit 4a89b7698e771914b4d5b571600c76e2fdcbe2a9)\nCc: stable@vger.kernel.org\n" }, { "commit": "0e33e0f339b91eecd9558311449a3d1e728722d4", "tree": "48693c5baf1830a68196b7a6fc192c6ffc714c1e", "parents": [ "f690e3974755a650259a45d71456decc9c96a282" ], "author": { "name": "Alex Deucher", "email": "alexander.deucher@amd.com", "time": "Wed Apr 30 12:46:56 2025 -0400" }, "committer": { "name": "Alex Deucher", "email": "alexander.deucher@amd.com", "time": "Thu May 08 11:46:57 2025 -0400" }, "message": "drm/amdgpu/hdp5: use memcfg register to post the write for HDP flush\n\nReading back the remapped HDP flush register seems to cause\nproblems on some platforms. All we need is a read, so read back\nthe memcfg register.\n\nFixes: cf424020e040 (\"drm/amdgpu/hdp5.0: do a posting read when flushing HDP\")\nReported-by: Alexey Klimov \u003calexey.klimov@linaro.org\u003e\nLink: https://lists.freedesktop.org/archives/amd-gfx/2025-April/123150.html\nCloses: https://gitlab.freedesktop.org/drm/amd/-/issues/4119\nCloses: https://gitlab.freedesktop.org/drm/amd/-/issues/3908\nReviewed-by: Felix Kuehling \u003cfelix.kuehling@amd.com\u003e\nSigned-off-by: Alex Deucher \u003calexander.deucher@amd.com\u003e\n(cherry picked from commit a5cb344033c7598762e89255e8ff52827abb57a4)\nCc: stable@vger.kernel.org\n" } ], "next": "2c89c1b655c0b06823f4ee8b055140d8628fc4da" }