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, 127 insertions, 0 deletions

diff --git a/drivers/dma/hsu/hsu.h b/drivers/dma/hsu/hsu.h
new file mode 100644
index 000000000..3bca577b9
--- /dev/null
+++ b/drivers/dma/hsu/hsu.h
@@ -0,0 +1,127 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Driver for the High Speed UART DMA
+ *
+ * Copyright (C) 2015 Intel Corporation
+ *
+ * Partially based on the bits found in drivers/tty/serial/mfd.c.
+ */
+
+#ifndef __DMA_HSU_H__
+#define __DMA_HSU_H__
+
+#include <linux/bits.h>
+#include <linux/container_of.h>
+#include <linux/io.h>
+#include <linux/types.h>
+
+#include <linux/dma/hsu.h>
+
+#include "../virt-dma.h"
+
+#define HSU_CH_SR		0x00			/* channel status */
+#define HSU_CH_CR		0x04			/* channel control */
+#define HSU_CH_DCR		0x08			/* descriptor control */
+#define HSU_CH_BSR		0x10			/* FIFO buffer size */
+#define HSU_CH_MTSR		0x14			/* minimum transfer size */
+#define HSU_CH_DxSAR(x)		(0x20 + 8 * (x))	/* desc start addr */
+#define HSU_CH_DxTSR(x)		(0x24 + 8 * (x))	/* desc transfer size */
+#define HSU_CH_D0SAR		0x20			/* desc 0 start addr */
+#define HSU_CH_D0TSR		0x24			/* desc 0 transfer size */
+#define HSU_CH_D1SAR		0x28
+#define HSU_CH_D1TSR		0x2c
+#define HSU_CH_D2SAR		0x30
+#define HSU_CH_D2TSR		0x34
+#define HSU_CH_D3SAR		0x38
+#define HSU_CH_D3TSR		0x3c
+
+#define HSU_DMA_CHAN_NR_DESC	4
+#define HSU_DMA_CHAN_LENGTH	0x40
+
+/* Bits in HSU_CH_SR */
+#define HSU_CH_SR_DESCTO(x)	BIT(8 + (x))
+#define HSU_CH_SR_DESCTO_ANY	GENMASK(11, 8)
+#define HSU_CH_SR_CHE		BIT(15)
+#define HSU_CH_SR_DESCE(x)	BIT(16 + (x))
+#define HSU_CH_SR_DESCE_ANY	GENMASK(19, 16)
+#define HSU_CH_SR_CDESC_ANY	GENMASK(31, 30)
+
+/* Bits in HSU_CH_CR */
+#define HSU_CH_CR_CHA		BIT(0)
+#define HSU_CH_CR_CHD		BIT(1)
+
+/* Bits in HSU_CH_DCR */
+#define HSU_CH_DCR_DESCA(x)	BIT(0 + (x))
+#define HSU_CH_DCR_CHSOD(x)	BIT(8 + (x))
+#define HSU_CH_DCR_CHSOTO	BIT(14)
+#define HSU_CH_DCR_CHSOE	BIT(15)
+#define HSU_CH_DCR_CHDI(x)	BIT(16 + (x))
+#define HSU_CH_DCR_CHEI		BIT(23)
+#define HSU_CH_DCR_CHTOI(x)	BIT(24 + (x))
+
+/* Bits in HSU_CH_DxTSR */
+#define HSU_CH_DxTSR_MASK	GENMASK(15, 0)
+#define HSU_CH_DxTSR_TSR(x)	((x) & HSU_CH_DxTSR_MASK)
+
+struct hsu_dma_sg {
+	dma_addr_t addr;
+	unsigned int len;
+};
+
+struct hsu_dma_desc {
+	struct virt_dma_desc vdesc;
+	enum dma_transfer_direction direction;
+	struct hsu_dma_sg *sg;
+	unsigned int nents;
+	size_t length;
+	unsigned int active;
+	enum dma_status status;
+};
+
+static inline struct hsu_dma_desc *to_hsu_dma_desc(struct virt_dma_desc *vdesc)
+{
+	return container_of(vdesc, struct hsu_dma_desc, vdesc);
+}
+
+struct hsu_dma_chan {
+	struct virt_dma_chan vchan;
+
+	void __iomem *reg;
+
+	/* hardware configuration */
+	enum dma_transfer_direction direction;
+	struct dma_slave_config config;
+
+	struct hsu_dma_desc *desc;
+};
+
+static inline struct hsu_dma_chan *to_hsu_dma_chan(struct dma_chan *chan)
+{
+	return container_of(chan, struct hsu_dma_chan, vchan.chan);
+}
+
+static inline u32 hsu_chan_readl(struct hsu_dma_chan *hsuc, int offset)
+{
+	return readl(hsuc->reg + offset);
+}
+
+static inline void hsu_chan_writel(struct hsu_dma_chan *hsuc, int offset,
+				   u32 value)
+{
+	writel(value, hsuc->reg + offset);
+}
+
+struct hsu_dma {
+	struct dma_device		dma;
+
+	/* channels */
+	struct hsu_dma_chan		*chan;
+	unsigned short			nr_channels;
+};
+
+static inline struct hsu_dma *to_hsu_dma(struct dma_device *ddev)
+{
+	return container_of(ddev, struct hsu_dma, dma);
+}
+
+#endif /* __DMA_HSU_H__ */