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, 246 insertions, 0 deletions

diff --git a/arch/um/os-Linux/mem.c b/arch/um/os-Linux/mem.c
new file mode 100644
index 000000000..72f302f4d
--- /dev/null
+++ b/arch/um/os-Linux/mem.c
@@ -0,0 +1,246 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (C) 2007 Jeff Dike (jdike@{addtoit,linux.intel}.com)
+ */
+
+#include <stdio.h>
+#include <stddef.h>
+#include <stdlib.h>
+#include <unistd.h>
+#include <errno.h>
+#include <fcntl.h>
+#include <string.h>
+#include <sys/stat.h>
+#include <sys/mman.h>
+#include <sys/vfs.h>
+#include <linux/magic.h>
+#include <init.h>
+#include <kern_util.h>
+#include <os.h>
+#include "internal.h"
+
+/*
+ * kasan_map_memory - maps memory from @start with a size of @len.
+ * The allocated memory is filled with zeroes upon success.
+ * @start: the start address of the memory to be mapped
+ * @len: the length of the memory to be mapped
+ *
+ * This function is used to map shadow memory for KASAN in uml
+ */
+void kasan_map_memory(void *start, size_t len)
+{
+	if (mmap(start,
+		 len,
+		 PROT_READ|PROT_WRITE,
+		 MAP_FIXED|MAP_ANONYMOUS|MAP_PRIVATE|MAP_NORESERVE,
+		 -1,
+		 0) == MAP_FAILED) {
+		os_info("Couldn't allocate shadow memory: %s\n.",
+			strerror(errno));
+		exit(1);
+	}
+
+	if (madvise(start, len, MADV_DONTDUMP)) {
+		os_info("Couldn't set MAD_DONTDUMP on shadow memory: %s\n.",
+			strerror(errno));
+		exit(1);
+	}
+
+	if (madvise(start, len, MADV_DONTFORK)) {
+		os_info("Couldn't set MADV_DONTFORK on shadow memory: %s\n.",
+			strerror(errno));
+		exit(1);
+	}
+}
+
+/* Set by make_tempfile() during early boot. */
+char *tempdir = NULL;
+
+/* Check if dir is on tmpfs. Return 0 if yes, -1 if no or error. */
+static int __init check_tmpfs(const char *dir)
+{
+	struct statfs st;
+
+	os_info("Checking if %s is on tmpfs...", dir);
+	if (statfs(dir, &st) < 0) {
+		os_info("%s\n", strerror(errno));
+	} else if (st.f_type != TMPFS_MAGIC) {
+		os_info("no\n");
+	} else {
+		os_info("OK\n");
+		return 0;
+	}
+	return -1;
+}
+
+/*
+ * Choose the tempdir to use. We want something on tmpfs so that our memory is
+ * not subject to the host's vm.dirty_ratio. If a tempdir is specified in the
+ * environment, we use that even if it's not on tmpfs, but we warn the user.
+ * Otherwise, we try common tmpfs locations, and if no tmpfs directory is found
+ * then we fall back to /tmp.
+ */
+static char * __init choose_tempdir(void)
+{
+	static const char * const vars[] = {
+		"TMPDIR",
+		"TMP",
+		"TEMP",
+		NULL
+	};
+	static const char fallback_dir[] = "/tmp";
+	static const char * const tmpfs_dirs[] = {
+		"/dev/shm",
+		fallback_dir,
+		NULL
+	};
+	int i;
+	const char *dir;
+
+	os_info("Checking environment variables for a tempdir...");
+	for (i = 0; vars[i]; i++) {
+		dir = getenv(vars[i]);
+		if ((dir != NULL) && (*dir != '\0')) {
+			os_info("%s\n", dir);
+			if (check_tmpfs(dir) >= 0)
+				goto done;
+			else
+				goto warn;
+		}
+	}
+	os_info("none found\n");
+
+	for (i = 0; tmpfs_dirs[i]; i++) {
+		dir = tmpfs_dirs[i];
+		if (check_tmpfs(dir) >= 0)
+			goto done;
+	}
+
+	dir = fallback_dir;
+warn:
+	os_warn("Warning: tempdir %s is not on tmpfs\n", dir);
+done:
+	/* Make a copy since getenv results may not remain valid forever. */
+	return strdup(dir);
+}
+
+/*
+ * Create an unlinked tempfile in a suitable tempdir. template must be the
+ * basename part of the template with a leading '/'.
+ */
+static int __init make_tempfile(const char *template)
+{
+	char *tempname;
+	int fd;
+
+	if (tempdir == NULL) {
+		tempdir = choose_tempdir();
+		if (tempdir == NULL) {
+			os_warn("Failed to choose tempdir: %s\n",
+				strerror(errno));
+			return -1;
+		}
+	}
+
+#ifdef O_TMPFILE
+	fd = open(tempdir, O_CLOEXEC | O_RDWR | O_EXCL | O_TMPFILE, 0700);
+	/*
+	 * If the running system does not support O_TMPFILE flag then retry
+	 * without it.
+	 */
+	if (fd != -1 || (errno != EINVAL && errno != EISDIR &&
+			errno != EOPNOTSUPP))
+		return fd;
+#endif
+
+	tempname = malloc(strlen(tempdir) + strlen(template) + 1);
+	if (tempname == NULL)
+		return -1;
+
+	strcpy(tempname, tempdir);
+	strcat(tempname, template);
+	fd = mkstemp(tempname);
+	if (fd < 0) {
+		os_warn("open - cannot create %s: %s\n", tempname,
+			strerror(errno));
+		goto out;
+	}
+	if (unlink(tempname) < 0) {
+		perror("unlink");
+		goto close;
+	}
+	free(tempname);
+	return fd;
+close:
+	close(fd);
+out:
+	free(tempname);
+	return -1;
+}
+
+#define TEMPNAME_TEMPLATE "/vm_file-XXXXXX"
+
+static int __init create_tmp_file(unsigned long long len)
+{
+	int fd, err;
+	char zero;
+
+	fd = make_tempfile(TEMPNAME_TEMPLATE);
+	if (fd < 0)
+		exit(1);
+
+	/*
+	 * Seek to len - 1 because writing a character there will
+	 * increase the file size by one byte, to the desired length.
+	 */
+	if (lseek64(fd, len - 1, SEEK_SET) < 0) {
+		perror("lseek64");
+		exit(1);
+	}
+
+	zero = 0;
+
+	err = write(fd, &zero, 1);
+	if (err != 1) {
+		perror("write");
+		exit(1);
+	}
+
+	return fd;
+}
+
+int __init create_mem_file(unsigned long long len)
+{
+	int err, fd;
+
+	fd = create_tmp_file(len);
+
+	err = os_set_exec_close(fd);
+	if (err < 0) {
+		errno = -err;
+		perror("exec_close");
+	}
+	return fd;
+}
+
+void __init check_tmpexec(void)
+{
+	void *addr;
+	int err, fd = create_tmp_file(UM_KERN_PAGE_SIZE);
+
+	addr = mmap(NULL, UM_KERN_PAGE_SIZE,
+		    PROT_READ | PROT_WRITE | PROT_EXEC, MAP_PRIVATE, fd, 0);
+	os_info("Checking PROT_EXEC mmap in %s...", tempdir);
+	if (addr == MAP_FAILED) {
+		err = errno;
+		os_warn("%s\n", strerror(err));
+		close(fd);
+		if (err == EPERM)
+			os_warn("%s must be not mounted noexec\n", tempdir);
+		exit(1);
+	}
+	os_info("OK\n");
+	munmap(addr, UM_KERN_PAGE_SIZE);
+
+	close(fd);
+}