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 --- drivers/interconnect/bulk.c | 159 ++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 159 insertions(+) create mode 100644 drivers/interconnect/bulk.c (limited to 'drivers/interconnect/bulk.c') diff --git a/drivers/interconnect/bulk.c b/drivers/interconnect/bulk.c new file mode 100644 index 000000000..8b1d8a412 --- /dev/null +++ b/drivers/interconnect/bulk.c @@ -0,0 +1,159 @@ +// SPDX-License-Identifier: GPL-2.0 + +#include +#include +#include + +/** + * of_icc_bulk_get() - get interconnect paths + * @dev: the device requesting the path + * @num_paths: the number of icc_bulk_data + * @paths: the table with the paths we want to get + * + * Returns 0 on success or negative errno otherwise. + */ +int __must_check of_icc_bulk_get(struct device *dev, int num_paths, + struct icc_bulk_data *paths) +{ + int ret, i; + + for (i = 0; i < num_paths; i++) { + paths[i].path = of_icc_get(dev, paths[i].name); + if (IS_ERR(paths[i].path)) { + ret = PTR_ERR(paths[i].path); + if (ret != -EPROBE_DEFER) + dev_err(dev, "of_icc_get() failed on path %s (%d)\n", + paths[i].name, ret); + paths[i].path = NULL; + goto err; + } + } + + return 0; + +err: + icc_bulk_put(i, paths); + + return ret; +} +EXPORT_SYMBOL_GPL(of_icc_bulk_get); + +/** + * icc_bulk_put() - put a list of interconnect paths + * @num_paths: the number of icc_bulk_data + * @paths: the icc_bulk_data table with the paths being put + */ +void icc_bulk_put(int num_paths, struct icc_bulk_data *paths) +{ + while (--num_paths >= 0) { + icc_put(paths[num_paths].path); + paths[num_paths].path = NULL; + } +} +EXPORT_SYMBOL_GPL(icc_bulk_put); + +/** + * icc_bulk_set_bw() - set bandwidth to a set of paths + * @num_paths: the number of icc_bulk_data + * @paths: the icc_bulk_data table containing the paths and bandwidth + * + * Returns 0 on success or negative errno otherwise. + */ +int icc_bulk_set_bw(int num_paths, const struct icc_bulk_data *paths) +{ + int ret = 0; + int i; + + for (i = 0; i < num_paths; i++) { + ret = icc_set_bw(paths[i].path, paths[i].avg_bw, paths[i].peak_bw); + if (ret) { + pr_err("icc_set_bw() failed on path %s (%d)\n", paths[i].name, ret); + return ret; + } + } + + return ret; +} +EXPORT_SYMBOL_GPL(icc_bulk_set_bw); + +/** + * icc_bulk_enable() - enable a previously disabled set of paths + * @num_paths: the number of icc_bulk_data + * @paths: the icc_bulk_data table containing the paths and bandwidth + * + * Returns 0 on success or negative errno otherwise. + */ +int icc_bulk_enable(int num_paths, const struct icc_bulk_data *paths) +{ + int ret, i; + + for (i = 0; i < num_paths; i++) { + ret = icc_enable(paths[i].path); + if (ret) { + pr_err("icc_enable() failed on path %s (%d)\n", paths[i].name, ret); + goto err; + } + } + + return 0; + +err: + icc_bulk_disable(i, paths); + + return ret; +} +EXPORT_SYMBOL_GPL(icc_bulk_enable); + +/** + * icc_bulk_disable() - disable a set of interconnect paths + * @num_paths: the number of icc_bulk_data + * @paths: the icc_bulk_data table containing the paths and bandwidth + */ +void icc_bulk_disable(int num_paths, const struct icc_bulk_data *paths) +{ + while (--num_paths >= 0) + icc_disable(paths[num_paths].path); +} +EXPORT_SYMBOL_GPL(icc_bulk_disable); + +struct icc_bulk_devres { + struct icc_bulk_data *paths; + int num_paths; +}; + +static void devm_icc_bulk_release(struct device *dev, void *res) +{ + struct icc_bulk_devres *devres = res; + + icc_bulk_put(devres->num_paths, devres->paths); +} + +/** + * devm_of_icc_bulk_get() - resource managed of_icc_bulk_get + * @dev: the device requesting the path + * @num_paths: the number of icc_bulk_data + * @paths: the table with the paths we want to get + * + * Returns 0 on success or negative errno otherwise. + */ +int devm_of_icc_bulk_get(struct device *dev, int num_paths, struct icc_bulk_data *paths) +{ + struct icc_bulk_devres *devres; + int ret; + + devres = devres_alloc(devm_icc_bulk_release, sizeof(*devres), GFP_KERNEL); + if (!devres) + return -ENOMEM; + + ret = of_icc_bulk_get(dev, num_paths, paths); + if (!ret) { + devres->paths = paths; + devres->num_paths = num_paths; + devres_add(dev, devres); + } else { + devres_free(devres); + } + + return ret; +} +EXPORT_SYMBOL_GPL(devm_of_icc_bulk_get); -- cgit v1.3.1