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, 253 insertions, 0 deletions

diff --git a/drivers/mcb/mcb-parse.c b/drivers/mcb/mcb-parse.c
new file mode 100644
index 000000000..7e09a0ea8
--- /dev/null
+++ b/drivers/mcb/mcb-parse.c
@@ -0,0 +1,253 @@
+// SPDX-License-Identifier: GPL-2.0-only
+#include <linux/types.h>
+#include <linux/ioport.h>
+#include <linux/slab.h>
+#include <linux/export.h>
+#include <linux/io.h>
+#include <linux/mcb.h>
+
+#include "mcb-internal.h"
+
+#define for_each_chameleon_cell(dtype, p)	\
+	for ((dtype) = get_next_dtype((p));	\
+	     (dtype) != CHAMELEON_DTYPE_END;	\
+	     (dtype) = get_next_dtype((p)))
+
+static inline uint32_t get_next_dtype(void __iomem *p)
+{
+	uint32_t dtype;
+
+	dtype = readl(p);
+	return dtype >> 28;
+}
+
+static int chameleon_parse_bdd(struct mcb_bus *bus,
+			struct chameleon_bar *cb,
+			void __iomem *base)
+{
+	return 0;
+}
+
+static int chameleon_parse_gdd(struct mcb_bus *bus,
+			struct chameleon_bar *cb,
+			void __iomem *base, int bar_count)
+{
+	struct chameleon_gdd __iomem *gdd =
+		(struct chameleon_gdd __iomem *) base;
+	struct mcb_device *mdev;
+	u32 dev_mapbase;
+	u32 offset;
+	u32 size;
+	int ret;
+	__le32 reg1;
+	__le32 reg2;
+
+	mdev = mcb_alloc_dev(bus);
+	if (!mdev)
+		return -ENOMEM;
+
+	reg1 = readl(&gdd->reg1);
+	reg2 = readl(&gdd->reg2);
+	offset = readl(&gdd->offset);
+	size = readl(&gdd->size);
+
+	mdev->id = GDD_DEV(reg1);
+	mdev->rev = GDD_REV(reg1);
+	mdev->var = GDD_VAR(reg1);
+	mdev->bar = GDD_BAR(reg2);
+	mdev->group = GDD_GRP(reg2);
+	mdev->inst = GDD_INS(reg2);
+
+	/*
+	 * If the BAR is missing, dev_mapbase is zero, or if the
+	 * device is IO mapped we just print a warning and go on with the
+	 * next device, instead of completely stop the gdd parser
+	 */
+	if (mdev->bar > bar_count - 1) {
+		pr_info("No BAR for 16z%03d\n", mdev->id);
+		ret = 0;
+		goto err;
+	}
+
+	dev_mapbase = cb[mdev->bar].addr;
+	if (!dev_mapbase) {
+		pr_info("BAR not assigned for 16z%03d\n", mdev->id);
+		ret = 0;
+		goto err;
+	}
+
+	if (dev_mapbase & 0x01) {
+		pr_info("IO mapped Device (16z%03d) not yet supported\n",
+			mdev->id);
+		ret = 0;
+		goto err;
+	}
+
+	pr_debug("Found a 16z%03d\n", mdev->id);
+
+	mdev->irq.start = GDD_IRQ(reg1);
+	mdev->irq.end = GDD_IRQ(reg1);
+	mdev->irq.flags = IORESOURCE_IRQ;
+
+	mdev->mem.start = dev_mapbase + offset;
+
+	mdev->mem.end = mdev->mem.start + size - 1;
+	mdev->mem.flags = IORESOURCE_MEM;
+
+	ret = mcb_device_register(bus, mdev);
+	if (ret < 0)
+		return ret;
+
+	return 0;
+
+err:
+	mcb_free_dev(mdev);
+
+	return ret;
+}
+
+static void chameleon_parse_bar(void __iomem *base,
+				struct chameleon_bar *cb, int bar_count)
+{
+	char __iomem *p = base;
+	int i;
+
+	/* skip reg1 */
+	p += sizeof(__le32);
+
+	for (i = 0; i < bar_count; i++) {
+		cb[i].addr = readl(p);
+		cb[i].size = readl(p + 4);
+
+		p += sizeof(struct chameleon_bar);
+	}
+}
+
+static int chameleon_get_bar(void __iomem **base, phys_addr_t mapbase,
+			     struct chameleon_bar **cb)
+{
+	struct chameleon_bar *c;
+	int bar_count;
+	__le32 reg;
+	u32 dtype;
+
+	/*
+	 * For those devices which are not connected
+	 * to the PCI Bus (e.g. LPC) there is a bar
+	 * descriptor located directly after the
+	 * chameleon header. This header is comparable
+	 * to a PCI header.
+	 */
+	dtype = get_next_dtype(*base);
+	if (dtype == CHAMELEON_DTYPE_BAR) {
+		reg = readl(*base);
+
+		bar_count = BAR_CNT(reg);
+		if (bar_count <= 0 || bar_count > CHAMELEON_BAR_MAX)
+			return -ENODEV;
+
+		c = kzalloc_objs(struct chameleon_bar, bar_count);
+		if (!c)
+			return -ENOMEM;
+
+		chameleon_parse_bar(*base, c, bar_count);
+		*base += BAR_DESC_SIZE(bar_count);
+	} else {
+		c = kzalloc_obj(struct chameleon_bar);
+		if (!c)
+			return -ENOMEM;
+
+		bar_count = 1;
+		c->addr = mapbase;
+	}
+
+	*cb = c;
+
+	return bar_count;
+}
+
+int chameleon_parse_cells(struct mcb_bus *bus, phys_addr_t mapbase,
+			void __iomem *base)
+{
+	struct chameleon_fpga_header *header;
+	struct chameleon_bar *cb;
+	void __iomem *p = base;
+	int num_cells = 0;
+	uint32_t dtype;
+	int bar_count;
+	int ret;
+	u32 hsize;
+	u32 table_size;
+
+	hsize = sizeof(struct chameleon_fpga_header);
+
+	header = kzalloc(hsize, GFP_KERNEL);
+	if (!header)
+		return -ENOMEM;
+
+	/* Extract header information */
+	memcpy_fromio(header, p, hsize);
+	/* We only support chameleon v2 at the moment */
+	header->magic = le16_to_cpu(header->magic);
+	if (header->magic != CHAMELEONV2_MAGIC) {
+		pr_err("Unsupported chameleon version 0x%x\n",
+				header->magic);
+		ret = -ENODEV;
+		goto free_header;
+	}
+	p += hsize;
+
+	bus->revision = header->revision;
+	bus->model = header->model;
+	bus->minor = header->minor;
+	snprintf(bus->name, CHAMELEON_FILENAME_LEN + 1, "%s",
+		 header->filename);
+
+	bar_count = chameleon_get_bar(&p, mapbase, &cb);
+	if (bar_count < 0) {
+		ret = bar_count;
+		goto free_header;
+	}
+
+	for_each_chameleon_cell(dtype, p) {
+		switch (dtype) {
+		case CHAMELEON_DTYPE_GENERAL:
+			ret = chameleon_parse_gdd(bus, cb, p, bar_count);
+			if (ret < 0)
+				goto free_bar;
+			p += sizeof(struct chameleon_gdd);
+			break;
+		case CHAMELEON_DTYPE_BRIDGE:
+			chameleon_parse_bdd(bus, cb, p);
+			p += sizeof(struct chameleon_bdd);
+			break;
+		case CHAMELEON_DTYPE_END:
+			break;
+		default:
+			pr_err("Invalid chameleon descriptor type 0x%x\n",
+				dtype);
+			ret = -EINVAL;
+			goto free_bar;
+		}
+		num_cells++;
+	}
+
+	if (num_cells == 0) {
+		ret = -EINVAL;
+		goto free_bar;
+	}
+
+	table_size = p - base;
+	pr_debug("%d cell(s) found. Chameleon table size: 0x%04x bytes\n", num_cells, table_size);
+	kfree(cb);
+	kfree(header);
+	return table_size;
+
+free_bar:
+	kfree(cb);
+free_header:
+	kfree(header);
+
+	return ret;
+}
+EXPORT_SYMBOL_NS_GPL(chameleon_parse_cells, "MCB");