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, 214 insertions, 0 deletions

diff --git a/fs/erofs/Kconfig b/fs/erofs/Kconfig
new file mode 100644
index 000000000..37495e95a
--- /dev/null
+++ b/fs/erofs/Kconfig
@@ -0,0 +1,214 @@
+# SPDX-License-Identifier: GPL-2.0-only
+
+config EROFS_FS
+	tristate "EROFS filesystem support"
+	depends on BLOCK
+	select CRC32
+	select CRYPTO if EROFS_FS_ZIP_ACCEL
+	select CRYPTO_DEFLATE if EROFS_FS_ZIP_ACCEL
+	select FS_IOMAP
+	select LZ4_DECOMPRESS if EROFS_FS_ZIP
+	select XXHASH if EROFS_FS_XATTR
+	select XZ_DEC if EROFS_FS_ZIP_LZMA
+	select XZ_DEC_MICROLZMA if EROFS_FS_ZIP_LZMA
+	select ZLIB_INFLATE if EROFS_FS_ZIP_DEFLATE
+	select ZSTD_DECOMPRESS if EROFS_FS_ZIP_ZSTD
+	help
+	  EROFS (Enhanced Read-Only File System) is a modern, lightweight,
+	  secure read-only filesystem for various use cases, such as immutable
+	  system images, container images, application sandboxes, and datasets.
+
+	  EROFS uses a flexible, hierarchical on-disk design so that features
+	  can be enabled on demand: the core on-disk format is block-aligned in
+	  order to perform optimally on all kinds of devices, including block
+	  and memory-backed devices; the format is easy to parse and has zero
+	  metadata redundancy, unlike generic filesystems, making it ideal for
+	  filesystem auditing and remote access; inline data, random-access
+	  friendly directory data, inline/shared extended attributes and
+	  chunk-based deduplication ensure space efficiency while maintaining
+	  high performance.
+
+	  Optionally, it supports multiple devices to reference external data,
+	  enabling data sharing for container images.
+
+	  It also has advanced encoded on-disk layouts, particularly for data
+	  compression and fine-grained deduplication. It utilizes fixed-size
+	  output compression to improve storage density while keeping relatively
+	  high compression ratios. Furthermore, it implements in-place
+	  decompression to reuse file pages to keep compressed data temporarily
+	  with proper strategies, which ensures guaranteed end-to-end runtime
+	  performance under extreme memory pressure without extra cost.
+
+	  For more details, see the web pages at <https://erofs.docs.kernel.org>
+	  and the documentation at <file:Documentation/filesystems/erofs.rst>.
+
+	  To compile EROFS filesystem support as a module, choose M here. The
+	  module will be called erofs.
+
+	  If unsure, say N.
+
+config EROFS_FS_DEBUG
+	bool "EROFS debugging feature"
+	depends on EROFS_FS
+	help
+	  Print debugging messages and enable more BUG_ONs which check
+	  filesystem consistency and find potential issues aggressively,
+	  which can be used for Android eng build, for example.
+
+	  For daily use, say N.
+
+config EROFS_FS_XATTR
+	bool "EROFS extended attributes"
+	depends on EROFS_FS
+	default y
+	help
+	  Extended attributes are name:value pairs associated with inodes by
+	  the kernel or by users (see the attr(5) manual page, or visit
+	  <http://acl.bestbits.at/> for details).
+
+	  If unsure, say N.
+
+config EROFS_FS_POSIX_ACL
+	bool "EROFS Access Control Lists"
+	depends on EROFS_FS_XATTR
+	select FS_POSIX_ACL
+	default y
+	help
+	  Posix Access Control Lists (ACLs) support permissions for users and
+	  groups beyond the owner/group/world scheme.
+
+	  To learn more about Access Control Lists, visit the POSIX ACLs for
+	  Linux website <http://acl.bestbits.at/>.
+
+	  If you don't know what Access Control Lists are, say N.
+
+config EROFS_FS_SECURITY
+	bool "EROFS Security Labels"
+	depends on EROFS_FS_XATTR
+	default y
+	help
+	  Security labels provide an access control facility to support Linux
+	  Security Models (LSMs) accepted by AppArmor, SELinux, Smack and TOMOYO
+	  Linux. This option enables an extended attribute handler for file
+	  security labels in the erofs filesystem, so that it requires enabling
+	  the extended attribute support in advance.
+
+	  If you are not using a security module, say N.
+
+config EROFS_FS_BACKED_BY_FILE
+	bool "File-backed EROFS filesystem support"
+	depends on EROFS_FS
+	default y
+	help
+	  This allows EROFS to use filesystem image files directly, without
+	  the intercession of loopback block devices or likewise. It is
+	  particularly useful for container images with numerous blobs and
+	  other sandboxes, where loop devices behave intricately.  It can also
+	  be used to simplify error-prone lifetime management of unnecessary
+	  virtual block devices.
+
+	  If you don't want to enable this feature, say N.
+
+config EROFS_FS_ZIP
+	bool "EROFS Data Compression Support"
+	depends on EROFS_FS
+	default y
+	help
+	  Enable EROFS compression layouts so that filesystems containing
+	  compressed files can be parsed by the kernel.
+
+	  If you don't want to enable compression feature, say N.
+
+config EROFS_FS_ZIP_LZMA
+	bool "EROFS LZMA compressed data support"
+	depends on EROFS_FS_ZIP
+	default y
+	help
+	  Saying Y here includes support for reading EROFS file systems
+	  containing LZMA compressed data, specifically called microLZMA. It
+	  gives better compression ratios than the default LZ4 format, at the
+	  expense of more CPU overhead.
+
+	  Say N if you want to disable LZMA compression support.
+
+config EROFS_FS_ZIP_LZMA_DEFAULT_MAX_STREAMS
+	int "EROFS LZMA default maximum decompression streams" if SMP
+	depends on EROFS_FS_ZIP_LZMA
+	range 1 NR_CPUS if SMP
+	range 1 1 if !SMP
+	default 16 if SMP && NR_CPUS >= 16
+	default NR_CPUS if SMP
+	default 1
+	help
+	  By default EROFS allocates one LZMA decompression stream per CPU.
+	  Each stream can hold a dictionary of up to 8 MiB taken from the
+	  mounted image, so on systems with many CPUs this can reserve a lot
+	  of memory.  This caps the default; the lzma_streams module parameter
+	  still overrides it.
+
+	  If unsure, keep the suggested default (16 or lower).
+
+config EROFS_FS_ZIP_DEFLATE
+	bool "EROFS DEFLATE compressed data support"
+	depends on EROFS_FS_ZIP
+	help
+	  Saying Y here includes support for reading EROFS file systems
+	  containing DEFLATE compressed data.  It gives better compression
+	  ratios than the default LZ4 format, while it costs more CPU
+	  overhead.
+
+	  If unsure, say N.
+
+config EROFS_FS_ZIP_ZSTD
+	bool "EROFS Zstandard compressed data support"
+	depends on EROFS_FS_ZIP
+	help
+	  Saying Y here includes support for reading EROFS file systems
+	  containing Zstandard compressed data.  It gives better compression
+	  ratios than the default LZ4 format, while it costs more CPU
+	  overhead and memory footprint.
+
+	  If unsure, say N.
+
+config EROFS_FS_ZIP_ACCEL
+	bool "EROFS hardware decompression support"
+	depends on EROFS_FS_ZIP
+	help
+	  Saying Y here includes hardware accelerator support for reading
+	  EROFS file systems containing compressed data.  It gives better
+	  decompression speed than the software-implemented decompression, and
+	  it costs lower CPU overhead.
+
+	  Hardware accelerator support is an experimental feature for now and
+	  file systems are still readable without selecting this option.
+
+	  If unsure, say N.
+
+config EROFS_FS_PCPU_KTHREAD
+	bool "EROFS per-cpu decompression kthread workers"
+	depends on EROFS_FS_ZIP
+	help
+	  Saying Y here enables per-CPU kthread workers pool to carry out
+	  async decompression for low latencies on some architectures.
+
+	  If unsure, say N.
+
+config EROFS_FS_PCPU_KTHREAD_HIPRI
+	bool "EROFS high priority per-CPU kthread workers"
+	depends on EROFS_FS_ZIP && EROFS_FS_PCPU_KTHREAD
+	default y
+	help
+	  This permits EROFS to configure per-CPU kthread workers to run
+	  at higher priority.
+
+	  If unsure, say N.
+
+config EROFS_FS_PAGE_CACHE_SHARE
+	bool "EROFS page cache share support (experimental)"
+	depends on EROFS_FS && EROFS_FS_XATTR
+	select FS_STACK
+	help
+	  This enables page cache sharing among inodes with identical
+	  content fingerprints on the same machine.
+
+	  If unsure, say N.