From a5f7a5bb3b7f28ba7e4fa246775b29a0e5537255 Mon Sep 17 00:00:00 2001 From: Jinke Han Date: Tue, 8 Sep 2026 15:37:42 +0800 Subject: x86/kprobes: Fix crash when probing CS CALL instructions When using eBPF to probe CS CALL instructions within a function, a crash can be triggered. The eBPF tool probes offset 257 of the __hrtimer_run_queues() function: <__hrtimer_run_queues+249>: nopl 0x0(%rax,%rax,1) <__hrtimer_run_queues+254>: mov %r14,%rdi <__hrtimer_run_queues+257>: cs call <__x86_indirect_thunk_r12> <__hrtimer_run_queues+263>: mov %eax,%r12d <__hrtimer_run_queues+266>: xchg %ax,%ax <__hrtimer_run_queues+268>: mov %r13,%rdi Which triggers this crash: BUG: unable to handle page fault for address: 00000000000f41c9 #PF: supervisor write access in kernel mode #PF: error_code(0x0002) - not-present page PGD 0 P4D 0 Oops: 0002 [#1] SMP NOPTI CPU: 1 PID: 0 Comm: swapper/1 Kdump: loaded Tainted: P RIP: 0010:__hrtimer_run_queues+0x106/0x230 Note that __hrtimer_run_queues+0x106 is __hrtimer_run_queues+262, which is at the 6th byte of the above CS CALL instruction. Since the CS CALL instruction occupies 6 bytes, the exception occurred in the middle of that call instruction. The root cause is that when using eBPF tools to probe in the middle of a function, a kprobe with INT3 is used as the underlying implementation. During single-step emulation of the original CALL instruction, int3_emulate_call() assumes that the probed CALL instruction is 5 bytes long. However, the actual CS-prefixed CALL instruction occupies 6 bytes, so it constructs an incorrect exception return address. When the CPU returns from the kprobe handler, the next instruction to be executed is at the address of the last byte of that CS CALL instruction. Coincidentally, starting from that address, the CPU fetches and decodes a completely different instruction, which ultimately triggers a kernel crash. Fix the issue by using the actual instruction length obtained from the instruction decoder when constructing the exception return address, rather than relying on the hardcoded CALL_INSN_SIZE macro. [ mingo: Refined the changelog ] Fixes: 6256e668b7af ("x86/kprobes: Use int3 instead of debug trap for single-step") Suggested-by: Masami Hiramatsu (Google) Signed-off-by: Jinke Han Signed-off-by: Ingo Molnar Reviewed-by: Masami Hiramatsu (Google) Acked-by: Yafang Shao Acked-by: Borislav Petkov Cc: Peter Zijlstra Link: https://patch.msgid.link/20260908073742.GA10517@didi-ThinkCentre-M920t-N000 --- lib/ratelimit.c | 105 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 105 insertions(+) create mode 100644 lib/ratelimit.c (limited to 'lib/ratelimit.c') diff --git a/lib/ratelimit.c b/lib/ratelimit.c new file mode 100644 index 000000000..e2d65d3b1 --- /dev/null +++ b/lib/ratelimit.c @@ -0,0 +1,105 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * ratelimit.c - Do something with rate limit. + * + * Isolated from kernel/printk.c by Dave Young + * + * 2008-05-01 rewrite the function and use a ratelimit_state data struct as + * parameter. Now every user can use their own standalone ratelimit_state. + */ + +#include +#include +#include + +/* + * __ratelimit - rate limiting + * @rs: ratelimit_state data + * @func: name of calling function + * + * This enforces a rate limit: not more than @rs->burst callbacks + * in every @rs->interval + * + * RETURNS: + * 0 means callbacks will be suppressed. + * 1 means go ahead and do it. + */ +int ___ratelimit(struct ratelimit_state *rs, const char *func) +{ + /* Paired with WRITE_ONCE() in .proc_handler(). + * Changing two values separately could be inconsistent + * and some message could be lost. (See: net_ratelimit_state). + */ + int interval = READ_ONCE(rs->interval); + int burst = READ_ONCE(rs->burst); + unsigned long flags; + int ret = 0; + + /* + * Zero interval says never limit, otherwise, non-positive burst + * says always limit. + */ + if (interval <= 0 || burst <= 0) { + WARN_ONCE(interval < 0 || burst < 0, "Negative interval (%d) or burst (%d): Uninitialized ratelimit_state structure?\n", interval, burst); + ret = interval == 0 || burst > 0; + if (!(READ_ONCE(rs->flags) & RATELIMIT_INITIALIZED) || (!interval && !burst) || + !raw_spin_trylock_irqsave(&rs->lock, flags)) + goto nolock_ret; + + /* Force re-initialization once re-enabled. */ + rs->flags &= ~RATELIMIT_INITIALIZED; + goto unlock_ret; + } + + /* + * If we contend on this state's lock then just check if + * the current burst is used or not. It might cause + * false positive when we are past the interval and + * the current lock owner is just about to reset it. + */ + if (!raw_spin_trylock_irqsave(&rs->lock, flags)) { + if (READ_ONCE(rs->flags) & RATELIMIT_INITIALIZED && + atomic_read(&rs->rs_n_left) > 0 && atomic_dec_return(&rs->rs_n_left) >= 0) + ret = 1; + goto nolock_ret; + } + + if (!(rs->flags & RATELIMIT_INITIALIZED)) { + rs->begin = jiffies; + rs->flags |= RATELIMIT_INITIALIZED; + atomic_set(&rs->rs_n_left, rs->burst); + } + + if (time_is_before_jiffies(rs->begin + interval)) { + int m; + + /* + * Reset rs_n_left ASAP to reduce false positives + * in parallel calls, see above. + */ + atomic_set(&rs->rs_n_left, rs->burst); + rs->begin = jiffies; + + if (!(rs->flags & RATELIMIT_MSG_ON_RELEASE)) { + m = ratelimit_state_reset_miss(rs); + if (m) { + printk_deferred(KERN_WARNING + "%s: %d callbacks suppressed\n", func, m); + } + } + } + + /* Note that the burst might be taken by a parallel call. */ + if (atomic_read(&rs->rs_n_left) > 0 && atomic_dec_return(&rs->rs_n_left) >= 0) + ret = 1; + +unlock_ret: + raw_spin_unlock_irqrestore(&rs->lock, flags); + +nolock_ret: + if (!ret) + ratelimit_state_inc_miss(rs); + + return ret; +} +EXPORT_SYMBOL(___ratelimit); -- cgit v1.3.1