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, 120 insertions, 0 deletions

diff --git a/Documentation/sphinx-static/theme_overrides.css b/Documentation/sphinx-static/theme_overrides.css
new file mode 100644
index 000000000..79000b26e
--- /dev/null
+++ b/Documentation/sphinx-static/theme_overrides.css
@@ -0,0 +1,120 @@
+/* -*- coding: utf-8; mode: css -*-
+ *
+ * Sphinx HTML theme customization: read the doc
+ * Please don't add any color definition here, as the theme should
+ * work for both normal and dark modes.
+ */
+
+/* Improve contrast and increase size for easier reading. */
+
+body {
+	font-family: serif;
+	font-size: 100%;
+}
+
+h1, h2, .rst-content .toctree-wrapper p.caption, h3, h4, h5, h6, legend {
+	font-family: sans-serif;
+}
+
+div[class^="highlight"] pre {
+	font-family: monospace;
+	font-size: 100%;
+}
+
+.wy-menu-vertical {
+	font-family: sans-serif;
+}
+
+.c {
+	font-style: normal;
+}
+
+p {
+	font-size: 100%;
+}
+
+/* Interim: Code-blocks with line nos - lines and line numbers don't line up.
+ * see: https://github.com/rtfd/sphinx_rtd_theme/issues/419
+ */
+
+div[class^="highlight"] pre {
+    line-height: normal;
+}
+.rst-content .highlight > pre {
+    line-height: normal;
+}
+
+/* Keep fields from being strangely far apart due to inheirited table CSS. */
+.rst-content table.field-list th.field-name {
+    padding-top: 1px;
+    padding-bottom: 1px;
+}
+.rst-content table.field-list td.field-body {
+    padding-top: 1px;
+    padding-bottom: 1px;
+}
+
+@media screen {
+
+    /* content column
+     *
+     * RTD theme's default is 800px as max width for the content, but we have
+     * tables with tons of columns, which need the full width of the view-port.
+     */
+
+    .wy-nav-content{max-width: none; }
+
+    /* table:
+     *
+     *   - Sequences of whitespace should collapse into a single whitespace.
+     *   - make the overflow auto (scrollbar if needed)
+     *   - align caption "left" ("center" is unsuitable on vast tables)
+     */
+
+    .wy-table-responsive table td { white-space: normal; }
+    .wy-table-responsive { overflow: auto; }
+    .rst-content table.docutils caption { text-align: left; font-size: 100%; }
+
+    /* captions:
+     *
+     *   - captions should have 100% (not 85%) font size
+     *   - hide the permalink symbol as long as link is not hovered
+     */
+
+    caption, .wy-table caption, .rst-content table.field-list caption {
+        font-size: 100%;
+    }
+    caption a.headerlink { opacity: 0; }
+    caption a.headerlink:hover { opacity: 1; }
+
+    /* Menu selection and keystrokes */
+
+    span.menuselection {
+	font-family: "Courier New", Courier, monospace
+    }
+
+    code.kbd, code.kbd span {
+	font-weight: bold;
+	font-family: "Courier New", Courier, monospace
+    }
+
+    /* fix bottom margin of lists items */
+
+    .rst-content .section ul li:last-child, .rst-content .section ul li p:last-child {
+          margin-bottom: 12px;
+    }
+
+    /* inline literal: drop the borderbox, padding and red color */
+
+    code, .rst-content tt, .rst-content code {
+        color: inherit;
+        border: none;
+        padding: unset;
+        background: inherit;
+        font-size: 85%;
+    }
+
+    .rst-content tt.literal,.rst-content tt.literal,.rst-content code.literal {
+        color: inherit;
+    }
+}