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, 188 insertions, 0 deletions

diff --git a/fs/verity/hash_algs.c b/fs/verity/hash_algs.c
new file mode 100644
index 000000000..de53e14c8
--- /dev/null
+++ b/fs/verity/hash_algs.c
@@ -0,0 +1,188 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * fs-verity hash algorithms
+ *
+ * Copyright 2019 Google LLC
+ */
+
+#include "fsverity_private.h"
+
+/* The hash algorithms supported by fs-verity */
+const struct fsverity_hash_alg fsverity_hash_algs[] = {
+	[FS_VERITY_HASH_ALG_SHA256] = {
+		.name = "sha256",
+		.digest_size = SHA256_DIGEST_SIZE,
+		.block_size = SHA256_BLOCK_SIZE,
+		.algo_id = HASH_ALGO_SHA256,
+	},
+	[FS_VERITY_HASH_ALG_SHA512] = {
+		.name = "sha512",
+		.digest_size = SHA512_DIGEST_SIZE,
+		.block_size = SHA512_BLOCK_SIZE,
+		.algo_id = HASH_ALGO_SHA512,
+	},
+};
+
+/**
+ * fsverity_get_hash_alg() - get a hash algorithm by number
+ * @inode: optional inode for logging purposes
+ * @num: the hash algorithm number
+ *
+ * Get the struct fsverity_hash_alg for the given hash algorithm number.
+ *
+ * Return: pointer to the hash alg if it's known, otherwise NULL.
+ */
+const struct fsverity_hash_alg *fsverity_get_hash_alg(const struct inode *inode,
+						      unsigned int num)
+{
+	if (num >= ARRAY_SIZE(fsverity_hash_algs) ||
+	    !fsverity_hash_algs[num].name) {
+		fsverity_warn(inode, "Unknown hash algorithm number: %u", num);
+		return NULL;
+	}
+	return &fsverity_hash_algs[num];
+}
+
+/**
+ * fsverity_prepare_hash_state() - precompute the initial hash state
+ * @alg: hash algorithm
+ * @salt: a salt which is to be prepended to all data to be hashed
+ * @salt_size: salt size in bytes
+ *
+ * Return: the kmalloc()'ed initial hash state, or NULL if out of memory.
+ */
+union fsverity_hash_ctx *
+fsverity_prepare_hash_state(const struct fsverity_hash_alg *alg,
+			    const u8 *salt, size_t salt_size)
+{
+	u8 *padded_salt = NULL;
+	size_t padded_salt_size;
+	union fsverity_hash_ctx ctx;
+	void *res = NULL;
+
+	/*
+	 * Zero-pad the salt to the next multiple of the input size of the hash
+	 * algorithm's compression function, e.g. 64 bytes for SHA-256 or 128
+	 * bytes for SHA-512.  This ensures that the hash algorithm won't have
+	 * any bytes buffered internally after processing the salt, thus making
+	 * salted hashing just as fast as unsalted hashing.
+	 */
+	padded_salt_size = round_up(salt_size, alg->block_size);
+	padded_salt = kzalloc(padded_salt_size, GFP_KERNEL);
+	if (!padded_salt)
+		return NULL;
+	memcpy(padded_salt, salt, salt_size);
+
+	switch (alg->algo_id) {
+	case HASH_ALGO_SHA256:
+		sha256_init(&ctx.sha256);
+		sha256_update(&ctx.sha256, padded_salt, padded_salt_size);
+		res = kmemdup(&ctx.sha256, sizeof(ctx.sha256), GFP_KERNEL);
+		break;
+	case HASH_ALGO_SHA512:
+		sha512_init(&ctx.sha512);
+		sha512_update(&ctx.sha512, padded_salt, padded_salt_size);
+		res = kmemdup(&ctx.sha512, sizeof(ctx.sha512), GFP_KERNEL);
+		break;
+	default:
+		WARN_ON_ONCE(1);
+	}
+	kfree(padded_salt);
+	return res;
+}
+
+/**
+ * fsverity_hash_block() - hash a single data or hash block
+ * @params: the Merkle tree's parameters
+ * @data: virtual address of a buffer containing the block to hash
+ * @out: output digest, size 'params->digest_size' bytes
+ *
+ * Hash a single data or hash block.  The hash is salted if a salt is specified
+ * in the Merkle tree parameters.
+ */
+void fsverity_hash_block(const struct merkle_tree_params *params,
+			 const void *data, u8 *out)
+{
+	union fsverity_hash_ctx ctx;
+
+	if (!params->hashstate) {
+		fsverity_hash_buffer(params->hash_alg, data, params->block_size,
+				     out);
+		return;
+	}
+
+	switch (params->hash_alg->algo_id) {
+	case HASH_ALGO_SHA256:
+		ctx.sha256 = params->hashstate->sha256;
+		sha256_update(&ctx.sha256, data, params->block_size);
+		sha256_final(&ctx.sha256, out);
+		return;
+	case HASH_ALGO_SHA512:
+		ctx.sha512 = params->hashstate->sha512;
+		sha512_update(&ctx.sha512, data, params->block_size);
+		sha512_final(&ctx.sha512, out);
+		return;
+	default:
+		BUG();
+	}
+}
+
+/**
+ * fsverity_hash_buffer() - hash some data
+ * @alg: the hash algorithm to use
+ * @data: the data to hash
+ * @size: size of data to hash, in bytes
+ * @out: output digest, size 'alg->digest_size' bytes
+ */
+void fsverity_hash_buffer(const struct fsverity_hash_alg *alg,
+			  const void *data, size_t size, u8 *out)
+{
+	switch (alg->algo_id) {
+	case HASH_ALGO_SHA256:
+		sha256(data, size, out);
+		return;
+	case HASH_ALGO_SHA512:
+		sha512(data, size, out);
+		return;
+	default:
+		BUG();
+	}
+}
+
+void __init fsverity_check_hash_algs(void)
+{
+	size_t i;
+
+	/*
+	 * Sanity check the hash algorithms (could be a build-time check, but
+	 * they're in an array)
+	 */
+	for (i = 0; i < ARRAY_SIZE(fsverity_hash_algs); i++) {
+		const struct fsverity_hash_alg *alg = &fsverity_hash_algs[i];
+
+		if (!alg->name)
+			continue;
+
+		/*
+		 * 0 must never be allocated as an FS_VERITY_HASH_ALG_* value,
+		 * as it is reserved for users that use 0 to mean unspecified or
+		 * a default value.  fs/verity/ itself doesn't care and doesn't
+		 * have a default algorithm, but some users make use of this.
+		 */
+		BUG_ON(i == 0);
+
+		BUG_ON(alg->digest_size > FS_VERITY_MAX_DIGEST_SIZE);
+
+		/*
+		 * For efficiency, the implementation currently assumes the
+		 * digest and block sizes are powers of 2.  This limitation can
+		 * be lifted if the code is updated to handle other values.
+		 */
+		BUG_ON(!is_power_of_2(alg->digest_size));
+		BUG_ON(!is_power_of_2(alg->block_size));
+
+		/* Verify that there is a valid mapping to HASH_ALGO_*. */
+		BUG_ON(alg->algo_id == 0);
+		BUG_ON(alg->digest_size != hash_digest_size[alg->algo_id]);
+	}
+}