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, 471 insertions, 0 deletions

diff --git a/fs/ceph/mdsmap.c b/fs/ceph/mdsmap.c
new file mode 100644
index 000000000..53079ef34
--- /dev/null
+++ b/fs/ceph/mdsmap.c
@@ -0,0 +1,471 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <linux/ceph/ceph_debug.h>
+
+#include <linux/bug.h>
+#include <linux/err.h>
+#include <linux/overflow.h>
+#include <linux/random.h>
+#include <linux/slab.h>
+#include <linux/types.h>
+
+#include <linux/ceph/messenger.h>
+#include <linux/ceph/decode.h>
+
+#include "mdsmap.h"
+#include "mds_client.h"
+#include "super.h"
+
+#define CEPH_MDS_IS_READY(i, ignore_laggy) \
+	(m->m_info[i].state > 0 && (ignore_laggy ? true : !m->m_info[i].laggy))
+
+static int __mdsmap_get_random_mds(struct ceph_mdsmap *m, bool ignore_laggy)
+{
+	int n = 0;
+	int i, j;
+
+	/* count */
+	for (i = 0; i < m->possible_max_rank; i++)
+		if (CEPH_MDS_IS_READY(i, ignore_laggy))
+			n++;
+	if (n == 0)
+		return -1;
+
+	/* pick */
+	n = get_random_u32_below(n);
+	for (j = 0, i = 0; i < m->possible_max_rank; i++) {
+		if (CEPH_MDS_IS_READY(i, ignore_laggy))
+			j++;
+		if (j > n)
+			break;
+	}
+
+	return i;
+}
+
+/*
+ * choose a random mds that is "up" (i.e. has a state > 0), or -1.
+ */
+int ceph_mdsmap_get_random_mds(struct ceph_mdsmap *m)
+{
+	int mds;
+
+	mds = __mdsmap_get_random_mds(m, false);
+	if (mds == m->possible_max_rank || mds == -1)
+		mds = __mdsmap_get_random_mds(m, true);
+
+	return mds == m->possible_max_rank ? -1 : mds;
+}
+
+#define __decode_and_drop_type(p, end, type, bad)		\
+	do {							\
+		if (*p + sizeof(type) > end)			\
+			goto bad;				\
+		*p += sizeof(type);				\
+	} while (0)
+
+#define __decode_and_drop_set(p, end, type, bad)		\
+	do {							\
+		u32 n;						\
+		size_t need;					\
+		ceph_decode_32_safe(p, end, n, bad);		\
+		need = sizeof(type) * n;			\
+		ceph_decode_need(p, end, need, bad);		\
+		*p += need;					\
+	} while (0)
+
+#define __decode_and_drop_map(p, end, ktype, vtype, bad)	\
+	do {							\
+		u32 n;						\
+		size_t need;					\
+		ceph_decode_32_safe(p, end, n, bad);		\
+		need = (sizeof(ktype) + sizeof(vtype)) * n;	\
+		ceph_decode_need(p, end, need, bad);		\
+		*p += need;					\
+	} while (0)
+
+
+static int __decode_and_drop_compat_set(void **p, void* end)
+{
+	int i;
+	/* compat, ro_compat, incompat*/
+	for (i = 0; i < 3; i++) {
+		u32 n;
+		ceph_decode_need(p, end, sizeof(u64) + sizeof(u32), bad);
+		/* mask */
+		*p += sizeof(u64);
+		/* names (map<u64, string>) */
+		n = ceph_decode_32(p);
+		while (n-- > 0) {
+			u32 len;
+			ceph_decode_need(p, end, sizeof(u64) + sizeof(u32),
+					 bad);
+			*p += sizeof(u64);
+			len = ceph_decode_32(p);
+			ceph_decode_need(p, end, len, bad);
+			*p += len;
+		}
+	}
+	return 0;
+bad:
+	return -1;
+}
+
+/*
+ * Decode an MDS map
+ *
+ * Ignore any fields we don't care about (there are quite a few of
+ * them).
+ */
+struct ceph_mdsmap *ceph_mdsmap_decode(struct ceph_mds_client *mdsc, void **p,
+				       void *end, bool msgr2)
+{
+	struct ceph_client *cl = mdsc->fsc->client;
+	struct ceph_mdsmap *m;
+	const void *start = *p;
+	int i, j, n;
+	int err;
+	u8 mdsmap_v;
+	u16 mdsmap_ev;
+	u32 target;
+	size_t export_targets_len;
+
+	m = kzalloc_obj(*m, GFP_NOFS);
+	if (!m)
+		return ERR_PTR(-ENOMEM);
+
+	ceph_decode_need(p, end, 1 + 1, bad);
+	mdsmap_v = ceph_decode_8(p);
+	*p += sizeof(u8);			/* mdsmap_cv */
+	if (mdsmap_v >= 4) {
+	       u32 mdsmap_len;
+	       ceph_decode_32_safe(p, end, mdsmap_len, bad);
+	       if (end < *p + mdsmap_len)
+		       goto bad;
+	       end = *p + mdsmap_len;
+	}
+
+	ceph_decode_need(p, end, 8*sizeof(u32) + sizeof(u64), bad);
+	m->m_epoch = ceph_decode_32(p);
+	m->m_client_epoch = ceph_decode_32(p);
+	m->m_last_failure = ceph_decode_32(p);
+	m->m_root = ceph_decode_32(p);
+	m->m_session_timeout = ceph_decode_32(p);
+	m->m_session_autoclose = ceph_decode_32(p);
+	m->m_max_file_size = ceph_decode_64(p);
+	m->m_max_mds = ceph_decode_32(p);
+
+	/*
+	 * pick out the active nodes as the m_num_active_mds, the
+	 * m_num_active_mds maybe larger than m_max_mds when decreasing
+	 * the max_mds in cluster side, in other case it should less
+	 * than or equal to m_max_mds.
+	 */
+	m->m_num_active_mds = n = ceph_decode_32(p);
+
+	/*
+	 * the possible max rank, it maybe larger than the m_num_active_mds,
+	 * for example if the mds_max == 2 in the cluster, when the MDS(0)
+	 * was laggy and being replaced by a new MDS, we will temporarily
+	 * receive a new mds map with n_num_mds == 1 and the active MDS(1),
+	 * and the mds rank >= m_num_active_mds.
+	 */
+	m->possible_max_rank = max(m->m_num_active_mds, m->m_max_mds);
+
+	m->m_info = kzalloc_objs(*m->m_info, m->possible_max_rank, GFP_NOFS);
+	if (!m->m_info)
+		goto nomem;
+
+	/* pick out active nodes from mds_info (state > 0) */
+	for (i = 0; i < n; i++) {
+		u64 global_id;
+		u32 namelen;
+		s32 mds, inc, state;
+		u8 info_v;
+		void *info_end = NULL;
+		struct ceph_entity_addr addr;
+		u32 num_export_targets;
+		void *pexport_targets = NULL;
+		struct ceph_timespec laggy_since;
+		struct ceph_mds_info *info;
+		bool laggy;
+
+		ceph_decode_need(p, end, sizeof(u64) + 1, bad);
+		global_id = ceph_decode_64(p);
+		info_v= ceph_decode_8(p);
+		if (info_v >= 4) {
+			u32 info_len;
+			ceph_decode_need(p, end, 1 + sizeof(u32), bad);
+			*p += sizeof(u8);	/* info_cv */
+			info_len = ceph_decode_32(p);
+			info_end = *p + info_len;
+			if (info_end > end)
+				goto bad;
+		}
+
+		ceph_decode_need(p, end, sizeof(u64) + sizeof(u32), bad);
+		*p += sizeof(u64);
+		namelen = ceph_decode_32(p);  /* skip mds name */
+		*p += namelen;
+
+		ceph_decode_32_safe(p, end, mds, bad);
+		ceph_decode_32_safe(p, end, inc, bad);
+		ceph_decode_32_safe(p, end, state, bad);
+		*p += sizeof(u64);		/* state_seq */
+		if (info_v >= 8)
+			err = ceph_decode_entity_addrvec(p, end, msgr2, &addr);
+		else
+			err = ceph_decode_entity_addr(p, end, &addr);
+		if (err)
+			goto corrupt;
+
+		ceph_decode_copy_safe(p, end, &laggy_since, sizeof(laggy_since),
+				      bad);
+		laggy = laggy_since.tv_sec != 0 || laggy_since.tv_nsec != 0;
+		*p += sizeof(u32);
+		ceph_decode_32_safe(p, end, namelen, bad);
+		*p += namelen;
+		if (info_v >= 2) {
+			ceph_decode_32_safe(p, end, num_export_targets, bad);
+			export_targets_len = size_mul(num_export_targets,
+						      sizeof(u32));
+			ceph_decode_need(p, end, export_targets_len, bad);
+			pexport_targets = *p;
+			*p += export_targets_len;
+		} else {
+			num_export_targets = 0;
+		}
+
+		if (info_end && *p != info_end) {
+			if (*p > info_end)
+				goto bad;
+			*p = info_end;
+		}
+
+		doutc(cl, "%d/%d %lld mds%d.%d %s %s%s\n", i+1, n, global_id,
+		      mds, inc, ceph_pr_addr(&addr),
+		      ceph_mds_state_name(state), laggy ? "(laggy)" : "");
+
+		if (mds < 0 || mds >= m->possible_max_rank) {
+			pr_warn_client(cl, "got incorrect mds(%d)\n", mds);
+			continue;
+		}
+
+		if (state <= 0) {
+			doutc(cl, "got incorrect state(%s)\n",
+			      ceph_mds_state_name(state));
+			continue;
+		}
+
+		info = &m->m_info[mds];
+		info->global_id = global_id;
+		info->state = state;
+		info->addr = addr;
+		info->laggy = laggy;
+		info->num_export_targets = num_export_targets;
+		if (num_export_targets) {
+			info->export_targets = kcalloc(num_export_targets,
+						       sizeof(u32), GFP_NOFS);
+			if (!info->export_targets)
+				goto nomem;
+			for (j = 0; j < num_export_targets; j++) {
+				target = ceph_decode_32(&pexport_targets);
+				if (target >= CEPH_MAX_MDS) {
+					err = -EIO;
+					goto corrupt;
+				}
+				info->export_targets[j] = target;
+			}
+		} else {
+			info->export_targets = NULL;
+		}
+	}
+
+	/* pg_pools */
+	ceph_decode_32_safe(p, end, n, bad);
+	m->m_num_data_pg_pools = n;
+	m->m_data_pg_pools = kcalloc(n, sizeof(u64), GFP_NOFS);
+	if (!m->m_data_pg_pools)
+		goto nomem;
+	ceph_decode_need(p, end, sizeof(u64)*(n+1), bad);
+	for (i = 0; i < n; i++)
+		m->m_data_pg_pools[i] = ceph_decode_64(p);
+	m->m_cas_pg_pool = ceph_decode_64(p);
+	m->m_enabled = m->m_epoch > 1;
+
+	mdsmap_ev = 1;
+	if (mdsmap_v >= 2) {
+		ceph_decode_16_safe(p, end, mdsmap_ev, bad_ext);
+	}
+	if (mdsmap_ev >= 3) {
+		if (__decode_and_drop_compat_set(p, end) < 0)
+			goto bad_ext;
+	}
+	/* metadata_pool */
+	if (mdsmap_ev < 5) {
+		__decode_and_drop_type(p, end, u32, bad_ext);
+	} else {
+		__decode_and_drop_type(p, end, u64, bad_ext);
+	}
+
+	/* created + modified + tableserver */
+	__decode_and_drop_type(p, end, struct ceph_timespec, bad_ext);
+	__decode_and_drop_type(p, end, struct ceph_timespec, bad_ext);
+	__decode_and_drop_type(p, end, u32, bad_ext);
+
+	/* in */
+	{
+		int num_laggy = 0;
+		ceph_decode_32_safe(p, end, n, bad_ext);
+		ceph_decode_need(p, end, sizeof(u32) * n, bad_ext);
+
+		for (i = 0; i < n; i++) {
+			s32 mds = ceph_decode_32(p);
+			if (mds >= 0 && mds < m->possible_max_rank) {
+				if (m->m_info[mds].laggy)
+					num_laggy++;
+			}
+		}
+		m->m_num_laggy = num_laggy;
+
+		if (n > m->possible_max_rank) {
+			void *new_m_info = krealloc(m->m_info,
+						    n * sizeof(*m->m_info),
+						    GFP_NOFS | __GFP_ZERO);
+			if (!new_m_info)
+				goto nomem;
+			m->m_info = new_m_info;
+		}
+		m->possible_max_rank = n;
+	}
+
+	/* inc */
+	__decode_and_drop_map(p, end, u32, u32, bad_ext);
+	/* up */
+	__decode_and_drop_map(p, end, u32, u64, bad_ext);
+	/* failed */
+	__decode_and_drop_set(p, end, u32, bad_ext);
+	/* stopped */
+	__decode_and_drop_set(p, end, u32, bad_ext);
+
+	if (mdsmap_ev >= 4) {
+		/* last_failure_osd_epoch */
+		__decode_and_drop_type(p, end, u32, bad_ext);
+	}
+	if (mdsmap_ev >= 6) {
+		/* ever_allowed_snaps */
+		__decode_and_drop_type(p, end, u8, bad_ext);
+		/* explicitly_allowed_snaps */
+		__decode_and_drop_type(p, end, u8, bad_ext);
+	}
+	if (mdsmap_ev >= 7) {
+		/* inline_data_enabled */
+		__decode_and_drop_type(p, end, u8, bad_ext);
+	}
+	if (mdsmap_ev >= 8) {
+		size_t fsname_len;
+
+		/* enabled */
+		ceph_decode_8_safe(p, end, m->m_enabled, bad_ext);
+
+		/* fs_name */
+		m->m_fs_name = ceph_extract_encoded_string(p, end,
+							   &fsname_len,
+							   GFP_NOFS);
+		if (IS_ERR(m->m_fs_name)) {
+			m->m_fs_name = NULL;
+			goto nomem;
+		}
+
+		/* validate fsname against mds_namespace */
+		if (!namespace_equals(mdsc->fsc->mount_options, m->m_fs_name,
+				      fsname_len)) {
+			pr_warn_client(cl, "fsname %s doesn't match mds_namespace %s\n",
+				       m->m_fs_name,
+				       mdsc->fsc->mount_options->mds_namespace);
+			goto bad;
+		}
+	} else {
+		m->m_enabled = false;
+		m->m_fs_name = kstrdup(CEPH_OLD_FS_NAME, GFP_NOFS);
+		if (!m->m_fs_name)
+			goto nomem;
+	}
+	/* damaged */
+	if (mdsmap_ev >= 9) {
+		size_t need;
+		ceph_decode_32_safe(p, end, n, bad_ext);
+		need = sizeof(u32) * n;
+		ceph_decode_need(p, end, need, bad_ext);
+		*p += need;
+		m->m_damaged = n > 0;
+	} else {
+		m->m_damaged = false;
+	}
+	if (mdsmap_ev >= 17) {
+		/* balancer */
+		ceph_decode_skip_string(p, end, bad_ext);
+		/* standby_count_wanted */
+		ceph_decode_skip_32(p, end, bad_ext);
+		/* old_max_mds */
+		ceph_decode_skip_32(p, end, bad_ext);
+		/* min_compat_client */
+		ceph_decode_skip_8(p, end, bad_ext);
+		/* required_client_features */
+		ceph_decode_skip_set(p, end, 64, bad_ext);
+		/* bal_rank_mask */
+		ceph_decode_skip_string(p, end, bad_ext);
+	}
+	if (mdsmap_ev >= 18) {
+		ceph_decode_64_safe(p, end, m->m_max_xattr_size, bad_ext);
+	}
+bad_ext:
+	doutc(cl, "m_enabled: %d, m_damaged: %d, m_num_laggy: %d\n",
+	      !!m->m_enabled, !!m->m_damaged, m->m_num_laggy);
+	*p = end;
+	doutc(cl, "success epoch %u\n", m->m_epoch);
+	return m;
+nomem:
+	err = -ENOMEM;
+	goto out_err;
+corrupt:
+	pr_err_client(cl, "corrupt mdsmap\n");
+	print_hex_dump(KERN_DEBUG, "mdsmap: ",
+		       DUMP_PREFIX_OFFSET, 16, 1,
+		       start, end - start, true);
+out_err:
+	ceph_mdsmap_destroy(m);
+	return ERR_PTR(err);
+bad:
+	err = -EINVAL;
+	goto corrupt;
+}
+
+void ceph_mdsmap_destroy(struct ceph_mdsmap *m)
+{
+	int i;
+
+	if (m->m_info) {
+		for (i = 0; i < m->possible_max_rank; i++)
+			kfree(m->m_info[i].export_targets);
+		kfree(m->m_info);
+	}
+	kfree(m->m_data_pg_pools);
+	kfree(m->m_fs_name);
+	kfree(m);
+}
+
+bool ceph_mdsmap_is_cluster_available(struct ceph_mdsmap *m)
+{
+	int i, nr_active = 0;
+	if (!m->m_enabled)
+		return false;
+	if (m->m_damaged)
+		return false;
+	if (m->m_num_laggy == m->m_num_active_mds)
+		return false;
+	for (i = 0; i < m->possible_max_rank; i++) {
+		if (m->m_info[i].state == CEPH_MDS_STATE_ACTIVE)
+			nr_active++;
+	}
+	return nr_active > 0;
+}