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, 71 insertions, 0 deletions

diff --git a/Documentation/fb/tgafb.rst b/Documentation/fb/tgafb.rst
new file mode 100644
index 000000000..f0944da1e
--- /dev/null
+++ b/Documentation/fb/tgafb.rst
@@ -0,0 +1,71 @@
+=======================================
+tgafb - TGA graphics framebuffer driver
+=======================================
+
+This is a driver for DECChip 21030 based graphics framebuffers, a.k.a. TGA
+cards, which are usually found in older Digital Alpha systems. The
+following models are supported:
+
+- ZLxP-E1 (8bpp, 2 MB VRAM)
+- ZLxP-E2 (32bpp, 8 MB VRAM)
+- ZLxP-E3 (32bpp, 16 MB VRAM, Zbuffer)
+
+This version is an almost complete rewrite of the code written by Geert
+Uytterhoeven, which was based on the original TGA console code written by
+Jay Estabrook.
+
+Major new features since Linux 2.0.x:
+
+ * Support for multiple resolutions
+ * Support for fixed-frequency and other oddball monitors
+   (by allowing the video mode to be set at boot time)
+
+User-visible changes since Linux 2.2.x:
+
+ * Sync-on-green is now handled properly
+ * More useful information is printed on bootup
+   (this helps if people run into problems)
+
+This driver does not (yet) support the TGA2 family of framebuffers, so the
+PowerStorm 3D30/4D20 (also known as PBXGB) cards are not supported. These
+can however be used with the standard VGA Text Console driver.
+
+
+Configuration
+=============
+
+You can pass kernel command line options to tgafb with
+`video=tgafb:option1,option2:value2,option3` (multiple options should be
+separated by comma, values are separated from options by `:`).
+
+Accepted options:
+
+==========  ============================================================
+font:X      default font to use. All fonts are supported, including the
+	    SUN12x22 font which is very nice at high resolutions.
+
+mode:X      default video mode. The following video modes are supported:
+	    640x480-60, 800x600-56, 640x480-72, 800x600-60, 800x600-72,
+	    1024x768-60, 1152x864-60, 1024x768-70, 1024x768-76,
+	    1152x864-70, 1280x1024-61, 1024x768-85, 1280x1024-70,
+	    1152x864-84, 1280x1024-76, 1280x1024-85
+==========  ============================================================
+
+
+Known Issues
+============
+
+The XFree86 FBDev server has been reported not to work, since tgafb doesn't do
+mmap(). Running the standard XF86_TGA server from XFree86 3.3.x works fine for
+me, however this server does not do acceleration, which make certain operations
+quite slow. Support for acceleration is being progressively integrated in
+XFree86 4.x.
+
+When running tgafb in resolutions higher than 640x480, on switching VCs from
+tgafb to XF86_TGA 3.3.x, the entire screen is not re-drawn and must be manually
+refreshed. This is an X server problem, not a tgafb problem, and is fixed in
+XFree86 4.0.
+
+Enjoy!
+
+Martin Lucina <mato@kotelna.sk>