x86/kprobes: Fix crash when probing CS CALL instructionsgrafted

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) <mhiramat@kernel.org>
Signed-off-by: Jinke Han <jinkehan@didiglobal.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Acked-by: Yafang Shao <laoar.shao@gmail.com>
Acked-by: Borislav Petkov <bp@alien8.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://patch.msgid.link/20260908073742.GA10517@didi-ThinkCentre-M920t-N000

1 files changed, 144 insertions, 0 deletions

diff --git a/samples/kfifo/dma-example.c b/samples/kfifo/dma-example.c
new file mode 100644
index 000000000..8076ac410
--- /dev/null
+++ b/samples/kfifo/dma-example.c
@@ -0,0 +1,144 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Sample fifo dma implementation
+ *
+ * Copyright (C) 2010 Stefani Seibold <stefani@seibold.net>
+ */
+
+#include <linux/init.h>
+#include <linux/kfifo.h>
+#include <linux/module.h>
+#include <linux/scatterlist.h>
+#include <linux/dma-mapping.h>
+
+/*
+ * This module shows how to handle fifo dma operations.
+ */
+
+/* fifo size in elements (bytes) */
+#define FIFO_SIZE	32
+
+static struct kfifo fifo;
+
+static int __init example_init(void)
+{
+	int			i;
+	unsigned int		ret;
+	unsigned int		nents;
+	struct scatterlist	sg[10];
+
+	printk(KERN_INFO "DMA fifo test start\n");
+
+	if (kfifo_alloc(&fifo, FIFO_SIZE, GFP_KERNEL)) {
+		printk(KERN_WARNING "error kfifo_alloc\n");
+		return -ENOMEM;
+	}
+
+	printk(KERN_INFO "queue size: %u\n", kfifo_size(&fifo));
+
+	kfifo_in(&fifo, "test", 4);
+
+	for (i = 0; i != 9; i++)
+		kfifo_put(&fifo, i);
+
+	/* kick away first byte */
+	kfifo_skip(&fifo);
+
+	printk(KERN_INFO "queue len: %u\n", kfifo_len(&fifo));
+
+	/*
+	 * Configure the kfifo buffer to receive data from DMA input.
+	 *
+	 *  .--------------------------------------.
+	 *  | 0 | 1 | 2 | ... | 12 | 13 | ... | 31 |
+	 *  |---|------------------|---------------|
+	 *   \_/ \________________/ \_____________/
+	 *    \          \                  \
+	 *     \          \_allocated data   \
+	 *      \_*free space*                \_*free space*
+	 *
+	 * We need two different SG entries: one for the free space area at the
+	 * end of the kfifo buffer (19 bytes) and another for the first free
+	 * byte at the beginning, after the kfifo_skip().
+	 */
+	sg_init_table(sg, ARRAY_SIZE(sg));
+	nents = kfifo_dma_in_prepare(&fifo, sg, ARRAY_SIZE(sg), FIFO_SIZE);
+	printk(KERN_INFO "DMA sgl entries: %d\n", nents);
+	if (!nents) {
+		/* fifo is full and no sgl was created */
+		printk(KERN_WARNING "error kfifo_dma_in_prepare\n");
+		return -EIO;
+	}
+
+	/* receive data */
+	printk(KERN_INFO "scatterlist for receive:\n");
+	for (i = 0; i < nents; i++) {
+		printk(KERN_INFO
+		"sg[%d] -> "
+		"page %p offset 0x%.8x length 0x%.8x\n",
+			i, sg_page(&sg[i]), sg[i].offset, sg[i].length);
+
+		if (sg_is_last(&sg[i]))
+			break;
+	}
+
+	/* put here your code to setup and exectute the dma operation */
+	/* ... */
+
+	/* example: zero bytes received */
+	ret = 0;
+
+	/* finish the dma operation and update the received data */
+	kfifo_dma_in_finish(&fifo, ret);
+
+	/* Prepare to transmit data, example: 8 bytes */
+	nents = kfifo_dma_out_prepare(&fifo, sg, ARRAY_SIZE(sg), 8);
+	printk(KERN_INFO "DMA sgl entries: %d\n", nents);
+	if (!nents) {
+		/* no data was available and no sgl was created */
+		printk(KERN_WARNING "error kfifo_dma_out_prepare\n");
+		return -EIO;
+	}
+
+	printk(KERN_INFO "scatterlist for transmit:\n");
+	for (i = 0; i < nents; i++) {
+		printk(KERN_INFO
+		"sg[%d] -> "
+		"page %p offset 0x%.8x length 0x%.8x\n",
+			i, sg_page(&sg[i]), sg[i].offset, sg[i].length);
+
+		if (sg_is_last(&sg[i]))
+			break;
+	}
+
+	/* put here your code to setup and exectute the dma operation */
+	/* ... */
+
+	/* example: 5 bytes transmitted */
+	ret = 5;
+
+	/* finish the dma operation and update the transmitted data */
+	kfifo_dma_out_finish(&fifo, ret);
+
+	ret = kfifo_len(&fifo);
+	printk(KERN_INFO "queue len: %u\n", kfifo_len(&fifo));
+
+	if (ret != 7) {
+		printk(KERN_WARNING "size mismatch: test failed");
+		return -EIO;
+	}
+	printk(KERN_INFO "test passed\n");
+
+	return 0;
+}
+
+static void __exit example_exit(void)
+{
+	kfifo_free(&fifo);
+}
+
+module_init(example_init);
+module_exit(example_exit);
+MODULE_DESCRIPTION("Sample fifo dma implementation");
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Stefani Seibold <stefani@seibold.net>");