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, 399 insertions, 0 deletions

diff --git a/drivers/crypto/ccp/tee-dev.c b/drivers/crypto/ccp/tee-dev.c
new file mode 100644
index 000000000..3e3645980
--- /dev/null
+++ b/drivers/crypto/ccp/tee-dev.c
@@ -0,0 +1,399 @@
+// SPDX-License-Identifier: MIT
+/*
+ * AMD Trusted Execution Environment (TEE) interface
+ *
+ * Author: Rijo Thomas <Rijo-john.Thomas@amd.com>
+ * Author: Devaraj Rangasamy <Devaraj.Rangasamy@amd.com>
+ *
+ * Copyright (C) 2019,2021 Advanced Micro Devices, Inc.
+ */
+
+#include <linux/bitfield.h>
+#include <linux/types.h>
+#include <linux/mutex.h>
+#include <linux/delay.h>
+#include <linux/slab.h>
+#include <linux/gfp.h>
+#include <linux/psp.h>
+#include <linux/psp-tee.h>
+
+#include "psp-dev.h"
+#include "tee-dev.h"
+
+static bool psp_dead;
+
+static int tee_alloc_ring(struct psp_tee_device *tee, int ring_size)
+{
+	struct ring_buf_manager *rb_mgr = &tee->rb_mgr;
+	void *start_addr;
+
+	if (!ring_size)
+		return -EINVAL;
+
+	/* We need actual physical address instead of DMA address, since
+	 * Trusted OS running on AMD Secure Processor will map this region
+	 */
+	start_addr = (void *)__get_free_pages(GFP_KERNEL, get_order(ring_size));
+	if (!start_addr)
+		return -ENOMEM;
+
+	memset(start_addr, 0x0, ring_size);
+	rb_mgr->ring_start = start_addr;
+	rb_mgr->ring_size = ring_size;
+	rb_mgr->ring_pa = __psp_pa(start_addr);
+	mutex_init(&rb_mgr->mutex);
+
+	return 0;
+}
+
+static void tee_free_ring(struct psp_tee_device *tee)
+{
+	struct ring_buf_manager *rb_mgr = &tee->rb_mgr;
+
+	if (!rb_mgr->ring_start)
+		return;
+
+	free_pages((unsigned long)rb_mgr->ring_start,
+		   get_order(rb_mgr->ring_size));
+
+	rb_mgr->ring_start = NULL;
+	rb_mgr->ring_size = 0;
+	rb_mgr->ring_pa = 0;
+	mutex_destroy(&rb_mgr->mutex);
+}
+
+static
+struct tee_init_ring_cmd *tee_alloc_cmd_buffer(struct psp_tee_device *tee)
+{
+	struct tee_init_ring_cmd *cmd;
+
+	cmd = kzalloc_obj(*cmd);
+	if (!cmd)
+		return NULL;
+
+	cmd->hi_addr = upper_32_bits(tee->rb_mgr.ring_pa);
+	cmd->low_addr = lower_32_bits(tee->rb_mgr.ring_pa);
+	cmd->size = tee->rb_mgr.ring_size;
+
+	dev_dbg(tee->dev, "tee: ring address: high = 0x%x low = 0x%x size = %u\n",
+		cmd->hi_addr, cmd->low_addr, cmd->size);
+
+	return cmd;
+}
+
+static inline void tee_free_cmd_buffer(struct tee_init_ring_cmd *cmd)
+{
+	kfree(cmd);
+}
+
+static bool tee_send_destroy_cmd(struct psp_tee_device *tee)
+{
+	unsigned int reg;
+	int ret;
+
+	ret = psp_mailbox_command(tee->psp, PSP_CMD_TEE_RING_DESTROY, NULL,
+				  TEE_DEFAULT_CMD_TIMEOUT, &reg);
+	if (ret) {
+		dev_err(tee->dev, "tee: ring destroy command timed out, disabling TEE support\n");
+		psp_dead = true;
+		return false;
+	}
+
+	if (FIELD_GET(PSP_CMDRESP_STS, reg)) {
+		dev_err(tee->dev, "tee: ring destroy command failed (%#010lx)\n",
+			FIELD_GET(PSP_CMDRESP_STS, reg));
+		psp_dead = true;
+		return false;
+	}
+
+	return true;
+}
+
+static int tee_init_ring(struct psp_tee_device *tee)
+{
+	int ring_size = MAX_RING_BUFFER_ENTRIES * sizeof(struct tee_ring_cmd);
+	struct tee_init_ring_cmd *cmd;
+	bool retry = false;
+	unsigned int reg;
+	int ret;
+
+	BUILD_BUG_ON(sizeof(struct tee_ring_cmd) != 1024);
+
+	ret = tee_alloc_ring(tee, ring_size);
+	if (ret) {
+		dev_err(tee->dev, "tee: ring allocation failed %d\n", ret);
+		return ret;
+	}
+
+	tee->rb_mgr.wptr = 0;
+
+	cmd = tee_alloc_cmd_buffer(tee);
+	if (!cmd) {
+		tee_free_ring(tee);
+		return -ENOMEM;
+	}
+
+	/* Send command buffer details to Trusted OS by writing to
+	 * CPU-PSP message registers
+	 */
+retry_init:
+	ret = psp_mailbox_command(tee->psp, PSP_CMD_TEE_RING_INIT, cmd,
+				  TEE_DEFAULT_CMD_TIMEOUT, &reg);
+	if (ret) {
+		dev_err(tee->dev, "tee: ring init command timed out, disabling TEE support\n");
+		tee_free_ring(tee);
+		psp_dead = true;
+		goto free_buf;
+	}
+
+	if (FIELD_GET(PSP_CMDRESP_STS, reg)) {
+		/*
+		 * During the hibernate resume sequence driver may have gotten loaded
+		 * but the ring not properly destroyed. If the ring doesn't work, try
+		 * to destroy and re-init once.
+		 */
+		if (!retry && FIELD_GET(PSP_CMDRESP_STS, reg) == PSP_TEE_STS_RING_BUSY) {
+			dev_info(tee->dev, "tee: ring init command failed with busy status, retrying\n");
+			if (tee_send_destroy_cmd(tee)) {
+				retry = true;
+				goto retry_init;
+			}
+		}
+		dev_err(tee->dev, "tee: ring init command failed (%#010lx)\n",
+			FIELD_GET(PSP_CMDRESP_STS, reg));
+		tee_free_ring(tee);
+		psp_dead = true;
+		ret = -EIO;
+	}
+
+free_buf:
+	tee_free_cmd_buffer(cmd);
+
+	return ret;
+}
+
+static void tee_destroy_ring(struct psp_tee_device *tee)
+{
+	if (!tee->rb_mgr.ring_start)
+		return;
+
+	if (psp_dead)
+		goto free_ring;
+
+	tee_send_destroy_cmd(tee);
+
+free_ring:
+	tee_free_ring(tee);
+}
+
+int tee_dev_init(struct psp_device *psp)
+{
+	struct device *dev = psp->dev;
+	struct psp_tee_device *tee;
+	int ret;
+
+	ret = -ENOMEM;
+	tee = devm_kzalloc(dev, sizeof(*tee), GFP_KERNEL);
+	if (!tee)
+		goto e_err;
+
+	psp->tee_data = tee;
+
+	tee->dev = dev;
+	tee->psp = psp;
+
+	tee->io_regs = psp->io_regs;
+
+	tee->vdata = (struct tee_vdata *)psp->vdata->tee;
+	if (!tee->vdata) {
+		ret = -ENODEV;
+		dev_err(dev, "tee: missing driver data\n");
+		goto e_err;
+	}
+
+	ret = tee_init_ring(tee);
+	if (ret) {
+		dev_err(dev, "tee: failed to init ring buffer\n");
+		goto e_err;
+	}
+
+	dev_notice(dev, "tee enabled\n");
+
+	return 0;
+
+e_err:
+	psp->tee_data = NULL;
+
+	dev_notice(dev, "tee initialization failed\n");
+
+	return ret;
+}
+
+void tee_dev_destroy(struct psp_device *psp)
+{
+	struct psp_tee_device *tee = psp->tee_data;
+
+	if (!tee)
+		return;
+
+	tee_destroy_ring(tee);
+}
+
+static int tee_submit_cmd(struct psp_tee_device *tee, enum tee_cmd_id cmd_id,
+			  void *buf, size_t len, struct tee_ring_cmd **resp)
+{
+	struct tee_ring_cmd *cmd;
+	int nloop = 1000, ret = 0;
+	u32 rptr;
+
+	*resp = NULL;
+
+	mutex_lock(&tee->rb_mgr.mutex);
+
+	/* Loop until empty entry found in ring buffer */
+	do {
+		/* Get pointer to ring buffer command entry */
+		cmd = (struct tee_ring_cmd *)
+			(tee->rb_mgr.ring_start + tee->rb_mgr.wptr);
+
+		rptr = ioread32(tee->io_regs + tee->vdata->ring_rptr_reg);
+
+		/* Check if ring buffer is full or command entry is waiting
+		 * for response from TEE
+		 */
+		if (!(tee->rb_mgr.wptr + sizeof(struct tee_ring_cmd) == rptr ||
+		      cmd->flag == CMD_WAITING_FOR_RESPONSE))
+			break;
+
+		dev_dbg(tee->dev, "tee: ring buffer full. rptr = %u wptr = %u\n",
+			rptr, tee->rb_mgr.wptr);
+
+		/* Wait if ring buffer is full or TEE is processing data */
+		mutex_unlock(&tee->rb_mgr.mutex);
+		schedule_timeout_interruptible(msecs_to_jiffies(10));
+		mutex_lock(&tee->rb_mgr.mutex);
+
+	} while (--nloop);
+
+	if (!nloop &&
+	    (tee->rb_mgr.wptr + sizeof(struct tee_ring_cmd) == rptr ||
+	     cmd->flag == CMD_WAITING_FOR_RESPONSE)) {
+		dev_err(tee->dev, "tee: ring buffer full. rptr = %u wptr = %u response flag %u\n",
+			rptr, tee->rb_mgr.wptr, cmd->flag);
+		ret = -EBUSY;
+		goto unlock;
+	}
+
+	/* Do not submit command if PSP got disabled while processing any
+	 * command in another thread
+	 */
+	if (psp_dead) {
+		ret = -EBUSY;
+		goto unlock;
+	}
+
+	/* Write command data into ring buffer */
+	cmd->cmd_id = cmd_id;
+	cmd->cmd_state = TEE_CMD_STATE_INIT;
+	memset(&cmd->buf[0], 0, sizeof(cmd->buf));
+	memcpy(&cmd->buf[0], buf, len);
+
+	/* Indicate driver is waiting for response */
+	cmd->flag = CMD_WAITING_FOR_RESPONSE;
+
+	/* Update local copy of write pointer */
+	tee->rb_mgr.wptr += sizeof(struct tee_ring_cmd);
+	if (tee->rb_mgr.wptr >= tee->rb_mgr.ring_size)
+		tee->rb_mgr.wptr = 0;
+
+	/* Trigger interrupt to Trusted OS */
+	iowrite32(tee->rb_mgr.wptr, tee->io_regs + tee->vdata->ring_wptr_reg);
+
+	/* The response is provided by Trusted OS in same
+	 * location as submitted data entry within ring buffer.
+	 */
+	*resp = cmd;
+
+unlock:
+	mutex_unlock(&tee->rb_mgr.mutex);
+
+	return ret;
+}
+
+static int tee_wait_cmd_completion(struct psp_tee_device *tee,
+				   struct tee_ring_cmd *resp,
+				   unsigned int timeout)
+{
+	/* ~1ms sleep per loop => nloop = timeout * 1000 */
+	int nloop = timeout * 1000;
+
+	while (--nloop) {
+		if (resp->cmd_state == TEE_CMD_STATE_COMPLETED)
+			return 0;
+
+		usleep_range(1000, 1100);
+	}
+
+	dev_err(tee->dev, "tee: command 0x%x timed out, disabling PSP\n",
+		resp->cmd_id);
+
+	psp_dead = true;
+
+	return -ETIMEDOUT;
+}
+
+int psp_tee_process_cmd(enum tee_cmd_id cmd_id, void *buf, size_t len,
+			u32 *status)
+{
+	struct psp_device *psp = psp_get_master_device();
+	struct psp_tee_device *tee;
+	struct tee_ring_cmd *resp;
+	int ret;
+
+	if (!buf || !status || !len || len > sizeof(resp->buf))
+		return -EINVAL;
+
+	*status = 0;
+
+	if (!psp || !psp->tee_data)
+		return -ENODEV;
+
+	if (psp_dead)
+		return -EBUSY;
+
+	tee = psp->tee_data;
+
+	ret = tee_submit_cmd(tee, cmd_id, buf, len, &resp);
+	if (ret)
+		return ret;
+
+	ret = tee_wait_cmd_completion(tee, resp, TEE_DEFAULT_RING_TIMEOUT);
+	if (ret) {
+		resp->flag = CMD_RESPONSE_TIMEDOUT;
+		return ret;
+	}
+
+	memcpy(buf, &resp->buf[0], len);
+	*status = resp->status;
+
+	resp->flag = CMD_RESPONSE_COPIED;
+
+	return 0;
+}
+EXPORT_SYMBOL(psp_tee_process_cmd);
+
+int psp_check_tee_status(void)
+{
+	struct psp_device *psp = psp_get_master_device();
+
+	if (!psp || !psp->tee_data)
+		return -ENODEV;
+
+	return 0;
+}
+EXPORT_SYMBOL(psp_check_tee_status);
+
+int tee_restore(struct psp_device *psp)
+{
+	return tee_init_ring(psp->tee_data);
+}