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, 112 insertions, 0 deletions

diff --git a/fs/nfsd/lockd.c b/fs/nfsd/lockd.c
new file mode 100644
index 000000000..72a5b4998
--- /dev/null
+++ b/fs/nfsd/lockd.c
@@ -0,0 +1,112 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * This file contains all the stubs needed when communicating with lockd.
+ * This level of indirection is necessary so we can run nfsd+lockd without
+ * requiring the nfs client to be compiled in/loaded, and vice versa.
+ *
+ * Copyright (C) 1996, Olaf Kirch <okir@monad.swb.de>
+ */
+
+#include <linux/file.h>
+#include <linux/lockd/bind.h>
+#include "nfsd.h"
+#include "vfs.h"
+
+#define NFSDDBG_FACILITY		NFSDDBG_LOCKD
+
+/**
+ * nlm_fopen - Open an NFSD file
+ * @rqstp: NLM RPC procedure execution context
+ * @f: NFS file handle to be opened
+ * @filp: OUT: an opened struct file
+ * @flags: the POSIX open flags to use
+ *
+ * nlm_fopen() holds the dentry reference until nlm_fclose() releases it.
+ *
+ * Returns zero on success or a negative errno value if the file
+ * cannot be opened.
+ */
+static int nlm_fopen(struct svc_rqst *rqstp, struct nfs_fh *f,
+		     struct file **filp, int flags)
+{
+	__be32		nfserr;
+	int		access;
+	struct svc_fh	fh;
+
+	/* must initialize before using! but maxsize doesn't matter */
+	fh_init(&fh,0);
+	fh.fh_handle.fh_size = f->size;
+	memcpy(&fh.fh_handle.fh_raw, f->data, f->size);
+	fh.fh_export = NULL;
+
+	/*
+	 * Allow BYPASS_GSS as some client implementations use AUTH_SYS
+	 * for NLM even when GSS is used for NFS.
+	 * Allow OWNER_OVERRIDE as permission might have been changed
+	 * after the file was opened.
+	 * Pass MAY_NLM so that authentication can be completely bypassed
+	 * if NFSEXP_NOAUTHNLM is set.  Some older clients use AUTH_NULL
+	 * for NLM requests.
+	 */
+	access = (flags == O_WRONLY) ? NFSD_MAY_WRITE : NFSD_MAY_READ;
+	access |= NFSD_MAY_NLM | NFSD_MAY_OWNER_OVERRIDE | NFSD_MAY_BYPASS_GSS;
+	nfserr = nfsd_open(rqstp, &fh, S_IFREG, access, filp);
+	fh_put(&fh);
+
+	switch (nfserr) {
+	case nfs_ok:
+		break;
+	case nfserr_jukebox:
+		/*
+		 * This error can indicate a presence of a conflicting
+		 * delegation to an NLM lock request. Options are:
+		 * (1) For now, drop this request and make the client
+		 * retry. When delegation is returned, client's lock retry
+		 * will complete.
+		 * (2) NLM4_DENIED as per "spec" signals to the client
+		 * that the lock is unavailable now but client can retry.
+		 * Linux client implementation does not. It treats
+		 * NLM4_DENIED same as NLM4_FAILED and fails the request.
+		 * (3) For the future, treat this as blocked lock and try
+		 * to callback when the delegation is returned but might
+		 * not have a proper lock request to block on.
+		 */
+		return -EWOULDBLOCK;
+	case nfserr_stale:
+		return -ESTALE;
+	default:
+		return -ENOLCK;
+	}
+
+	return 0;
+}
+
+/**
+ * nlm_fclose - Close an NFSD file
+ * @filp: a struct file that was opened by nlm_fopen()
+ */
+static void
+nlm_fclose(struct file *filp)
+{
+	fput(filp);
+}
+
+static const struct nlmsvc_binding nfsd_nlm_ops = {
+	.owner		= THIS_MODULE,
+	.fopen		= nlm_fopen,		/* open file for locking */
+	.fclose		= nlm_fclose,		/* close file */
+};
+
+void
+nfsd_lockd_init(void)
+{
+	dprintk("nfsd: initializing lockd\n");
+	rcu_assign_pointer(nlmsvc_ops, &nfsd_nlm_ops);
+}
+
+void
+nfsd_lockd_shutdown(void)
+{
+	RCU_INIT_POINTER(nlmsvc_ops, NULL);
+	synchronize_rcu();
+}