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, 186 insertions, 0 deletions

diff --git a/fs/jffs2/Kconfig b/fs/jffs2/Kconfig
new file mode 100644
index 000000000..560187d61
--- /dev/null
+++ b/fs/jffs2/Kconfig
@@ -0,0 +1,186 @@
+# SPDX-License-Identifier: GPL-2.0-only
+config JFFS2_FS
+	tristate "Journalling Flash File System v2 (JFFS2) support"
+	select CRC32
+	depends on MTD
+	help
+	  JFFS2 is the second generation of the Journalling Flash File System
+	  for use on diskless embedded devices. It provides improved wear
+	  levelling, compression and support for hard links. You cannot use
+	  this on normal block devices, only on 'MTD' devices.
+
+	  Further information on the design and implementation of JFFS2 is
+	  available at <http://sources.redhat.com/jffs2/>.
+
+config JFFS2_FS_DEBUG
+	int "JFFS2 debugging verbosity (0 = quiet, 2 = noisy)"
+	depends on JFFS2_FS
+	default "0"
+	help
+	  This controls the amount of debugging messages produced by the JFFS2
+	  code. Set it to zero for use in production systems. For evaluation,
+	  testing and debugging, it's advisable to set it to one. This will
+	  enable a few assertions and will print debugging messages at the
+	  KERN_DEBUG loglevel, where they won't normally be visible. Level 2
+	  is unlikely to be useful - it enables extra debugging in certain
+	  areas which at one point needed debugging, but when the bugs were
+	  located and fixed, the detailed messages were relegated to level 2.
+
+	  If reporting bugs, please try to have available a full dump of the
+	  messages at debug level 1 while the misbehaviour was occurring.
+
+config JFFS2_FS_WRITEBUFFER
+	bool "JFFS2 write-buffering support"
+	depends on JFFS2_FS
+	default y
+	help
+	  This enables the write-buffering support in JFFS2.
+
+	  This functionality is required to support JFFS2 on the following
+	  types of flash devices:
+	    - NAND flash
+	    - NOR flash with transparent ECC
+	    - DataFlash
+
+config JFFS2_FS_WBUF_VERIFY
+	bool "Verify JFFS2 write-buffer reads"
+	depends on JFFS2_FS_WRITEBUFFER
+	default n
+	help
+	  This causes JFFS2 to read back every page written through the
+	  write-buffer, and check for errors.
+
+config JFFS2_SUMMARY
+	bool "JFFS2 summary support"
+	depends on JFFS2_FS
+	default n
+	help
+	  This feature makes it possible to use summary information
+	  for faster filesystem mount.
+
+	  The summary information can be inserted into a filesystem image
+	  by the utility 'sumtool'.
+
+	  If unsure, say 'N'.
+
+config JFFS2_FS_XATTR
+	bool "JFFS2 XATTR support"
+	depends on JFFS2_FS
+	default n
+	help
+	  Extended attributes are name:value pairs associated with inodes by
+	  the kernel or by users (see the attr(5) manual page for details).
+
+	  If unsure, say N.
+
+config JFFS2_FS_POSIX_ACL
+	bool "JFFS2 POSIX Access Control Lists"
+	depends on JFFS2_FS_XATTR
+	default y
+	select FS_POSIX_ACL
+	help
+	  Posix Access Control Lists (ACLs) support permissions for users and
+	  groups beyond the owner/group/world scheme.
+
+	  If you don't know what Access Control Lists are, say N
+
+config JFFS2_FS_SECURITY
+	bool "JFFS2 Security Labels"
+	depends on JFFS2_FS_XATTR
+	default y
+	help
+	  Security labels support alternative access control models
+	  implemented by security modules like SELinux.  This option
+	  enables an extended attribute handler for file security
+	  labels in the jffs2 filesystem.
+
+	  If you are not using a security module that requires using
+	  extended attributes for file security labels, say N.
+
+config JFFS2_COMPRESSION_OPTIONS
+	bool "Advanced compression options for JFFS2"
+	depends on JFFS2_FS
+	default n
+	help
+	  Enabling this option allows you to explicitly choose which
+	  compression modules, if any, are enabled in JFFS2. Removing
+	  compressors can mean you cannot read existing file systems,
+	  and enabling experimental compressors can mean that you
+	  write a file system which cannot be read by a standard kernel.
+
+	  If unsure, you should _definitely_ say 'N'.
+
+config JFFS2_ZLIB
+	bool "JFFS2 ZLIB compression support" if JFFS2_COMPRESSION_OPTIONS
+	select ZLIB_INFLATE
+	select ZLIB_DEFLATE
+	depends on JFFS2_FS
+	default y
+	help
+	  Zlib is designed to be a free, general-purpose, legally unencumbered,
+	  lossless data-compression library for use on virtually any computer
+	  hardware and operating system. See <http://www.gzip.org/zlib/> for
+	  further information.
+
+	  Say 'Y' if unsure.
+
+config JFFS2_LZO
+	bool "JFFS2 LZO compression support" if JFFS2_COMPRESSION_OPTIONS
+	select LZO_COMPRESS
+	select LZO_DECOMPRESS
+	depends on JFFS2_FS
+	default n
+	help
+	  minilzo-based compression. Generally works better than Zlib.
+
+	  This feature was added in July, 2007. Say 'N' if you need
+	  compatibility with older bootloaders or kernels.
+
+config JFFS2_RTIME
+	bool "JFFS2 RTIME compression support" if JFFS2_COMPRESSION_OPTIONS
+	depends on JFFS2_FS
+	default y
+	help
+	  Rtime does manage to recompress already-compressed data. Say 'Y' if unsure.
+
+config JFFS2_RUBIN
+	bool "JFFS2 RUBIN compression support" if JFFS2_COMPRESSION_OPTIONS
+	depends on JFFS2_FS
+	default n
+	help
+	  RUBINMIPS and DYNRUBIN compressors. Say 'N' if unsure.
+
+choice
+	prompt "JFFS2 default compression mode"
+	default JFFS2_CMODE_PRIORITY
+	depends on JFFS2_COMPRESSION_OPTIONS
+	depends on JFFS2_FS
+	help
+	  You can set here the default compression mode of JFFS2 from
+	  the available compression modes. Don't touch if unsure.
+
+config JFFS2_CMODE_NONE
+	bool "no compression"
+	help
+	  Uses no compression.
+
+config JFFS2_CMODE_PRIORITY
+	bool "priority"
+	help
+	  Tries the compressors in a predefined order and chooses the first
+	  successful one.
+
+config JFFS2_CMODE_SIZE
+	bool "size"
+	help
+	  Tries all compressors and chooses the one which has the smallest
+	  result.
+
+config JFFS2_CMODE_FAVOURLZO
+	bool "Favour LZO"
+	help
+	  Tries all compressors and chooses the one which has the smallest
+	  result but gives some preference to LZO (which has faster
+	  decompression) at the expense of size.
+
+endchoice