Re: [PATCH v3 0/5] kprobes: Fixes for ftrace ops and optprobe state machine
From: Masami Hiramatsu (Google)
Date: Fri Nov 13 2026 - 04:37:54 EST
On Wed, 11 Nov 2026 23:13:21 +0900, Masami Hiramatsu (Google) <mhiramat@xxxxxxxxxx> wrote:
> Hi,
>
> Here is v3 of the remaining bug fixes for kprobes and optprobes.
> The previous version is here:
>
> https://lore.kernel.org/all/179427358690.291326.2078291446658598883.stgit@devnote2/
>
> This version restores __arm_kprobe() fix as patch 3, and update the KUnit
> test code to check it.
> This includes following fixes.
I'll pick the first 4 of this series to probes/fixes and the last
test to probes/core.
Thanks,
>
> 1. Ftrace Ops Switching for Post-Handlers (Patch 1):
> - When adding a second probe with a post_handler to an existing ftrace
> kprobe, the aggregate probe's post_handler was set but remained armed
> on kprobe_ftrace_ops. Subsequent disarming attempted to disarm from
> kprobe_ipmodify_ops, failing with -ENOENT and leaking the armed probe.
> Switch ftrace ops safely in register_aggr_kprobe() outside of
> text_mutex and cpus_read_lock.
>
> 2. Optprobe State Machine & Tasks RCU Quiescence (Patches 2-4):
> - Forcibly unoptimized probes (unoptimize_kprobe(force = true)) were
> not queued to cooling_list, causing kprobe_disarmed() to return true
> prematurely and risking use-after-free in the optinsn detour buffer
> before Tasks RCU completed.
> - When re-arming a probe that is queued on unoptimizing_list before the
> optimizer thread runs, skip arch_arm_kprobe() in __arm_kprobe() so
> the active jump instruction is not overwritten with a breakpoint.
> - When a disabled cooling probe finishes cooling and transitions to
> safely disarmed in optprobe_finalize_generation(), re-optimize any
> blocked sibling probe whose jump range covers its address.
>
> 3. KUnit Regression Tests (Patch 5):
> - Add KUnit test cases to lib/tests/test_kprobes.c covering the recent
> kprobes, kretprobes, ftrace ops, and optprobe fixes.
>
> Changes in v3:
> [03/05]
> - Restore the __arm_kprobe() fix that was accidentally dropped in v2
> while omitting the reuse_unused_kprobe() hunk superseded by commit
> 1ffd6efb1b58 ("kprobes: Arm reused optprobe on cooling list").
> [05/05]
> - Check that the jump opcode at *opt_kp.addr is preserved across
> disable_kprobe() and enable_kprobe() in
> test_optprobe_rearm_unoptimizing().
>
> Thank you,
>
> ---
> base-commit: 1ffd6efb1b58e596ed70f660182803264a4c84d9
>
> Masami Hiramatsu (Google) (5):
> kprobes: Switch ftrace ops when adding probe with post_handler
> kprobes: Fix UAF and queue cooling probes on forced unoptimization
> kprobes: Do not overwrite jump instruction when re-arming unoptimizing probe
> kprobes: Re-optimize blocked sibling probe when cooling finishes
> kprobes: Add KUnit test cases for kprobes regressions
>
>
> kernel/kprobes.c | 116 +++++++++++++++++++++++++++-----------
> lib/tests/test_kprobes.c | 139 ++++++++++++++++++++++++++++++++++++++++++++++
> 2 files changed, 220 insertions(+), 35 deletions(-)
>
> --
> Masami Hiramatsu (Google) <mhiramat@xxxxxxxxxx>
>
--
Masami Hiramatsu (Google) <mhiramat@xxxxxxxxxx>