disasm.c

  1// SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
  2/* Copyright (c) 2011-2014 PLUMgrid, http://plumgrid.com
  3 * Copyright (c) 2016 Facebook
  4 */
  5
  6#include <linux/bpf.h>
  7
  8#include "disasm.h"
  9
 10#define __BPF_FUNC_STR_FN(x) [BPF_FUNC_ ## x] = __stringify(bpf_ ## x)
 11static const char * const func_id_str[] = {
 12	__BPF_FUNC_MAPPER(__BPF_FUNC_STR_FN)
 13};
 14#undef __BPF_FUNC_STR_FN
 15
 16static const char *__func_get_name(const struct bpf_insn_cbs *cbs,
 17				   const struct bpf_insn *insn,
 18				   char *buff, size_t len)
 19{
 20	BUILD_BUG_ON(ARRAY_SIZE(func_id_str) != __BPF_FUNC_MAX_ID);
 21
 22	if (!insn->src_reg &&
 23	    insn->imm >= 0 && insn->imm < __BPF_FUNC_MAX_ID &&
 24	    func_id_str[insn->imm])
 25		return func_id_str[insn->imm];
 26
 27	if (cbs && cbs->cb_call) {
 28		const char *res;
 29
 30		res = cbs->cb_call(cbs->private_data, insn);
 31		if (res)
 32			return res;
 33	}
 34
 35	if (insn->src_reg == BPF_PSEUDO_CALL)
 36		snprintf(buff, len, "%+d", insn->imm);
 37	else if (insn->src_reg == BPF_PSEUDO_KFUNC_CALL)
 38		snprintf(buff, len, "kernel-function");
 39
 40	return buff;
 41}
 42
 43static const char *__func_imm_name(const struct bpf_insn_cbs *cbs,
 44				   const struct bpf_insn *insn,
 45				   u64 full_imm, char *buff, size_t len)
 46{
 47	if (cbs && cbs->cb_imm)
 48		return cbs->cb_imm(cbs->private_data, insn, full_imm);
 49
 50	snprintf(buff, len, "0x%llx", (unsigned long long)full_imm);
 51	return buff;
 52}
 53
 54const char *func_id_name(int id)
 55{
 56	if (id >= 0 && id < __BPF_FUNC_MAX_ID && func_id_str[id])
 57		return func_id_str[id];
 58	else
 59		return "unknown";
 60}
 61
 62const char *const bpf_class_string[8] = {
 63	[BPF_LD]    = "ld",
 64	[BPF_LDX]   = "ldx",
 65	[BPF_ST]    = "st",
 66	[BPF_STX]   = "stx",
 67	[BPF_ALU]   = "alu",
 68	[BPF_JMP]   = "jmp",
 69	[BPF_JMP32] = "jmp32",
 70	[BPF_ALU64] = "alu64",
 71};
 72
 73const char *const bpf_alu_string[16] = {
 74	[BPF_ADD >> 4]  = "+=",
 75	[BPF_SUB >> 4]  = "-=",
 76	[BPF_MUL >> 4]  = "*=",
 77	[BPF_DIV >> 4]  = "/=",
 78	[BPF_OR  >> 4]  = "|=",
 79	[BPF_AND >> 4]  = "&=",
 80	[BPF_LSH >> 4]  = "<<=",
 81	[BPF_RSH >> 4]  = ">>=",
 82	[BPF_NEG >> 4]  = "neg",
 83	[BPF_MOD >> 4]  = "%=",
 84	[BPF_XOR >> 4]  = "^=",
 85	[BPF_MOV >> 4]  = "=",
 86	[BPF_ARSH >> 4] = "s>>=",
 87	[BPF_END >> 4]  = "endian",
 88};
 89
 90static const char *const bpf_alu_sign_string[16] = {
 91	[BPF_DIV >> 4]  = "s/=",
 92	[BPF_MOD >> 4]  = "s%=",
 93};
 94
 95static const char *const bpf_movsx_string[4] = {
 96	[0] = "(s8)",
 97	[1] = "(s16)",
 98	[3] = "(s32)",
 99};
100
101static const char *const bpf_atomic_alu_string[16] = {
102	[BPF_ADD >> 4]  = "add",
103	[BPF_AND >> 4]  = "and",
104	[BPF_OR >> 4]  = "or",
105	[BPF_XOR >> 4]  = "xor",
106};
107
108static const char *const bpf_ldst_string[] = {
109	[BPF_W >> 3]  = "u32",
110	[BPF_H >> 3]  = "u16",
111	[BPF_B >> 3]  = "u8",
112	[BPF_DW >> 3] = "u64",
113};
114
115static const char *const bpf_ldsx_string[] = {
116	[BPF_W >> 3]  = "s32",
117	[BPF_H >> 3]  = "s16",
118	[BPF_B >> 3]  = "s8",
119};
120
121static const char *const bpf_jmp_string[16] = {
122	[BPF_JA >> 4]   = "jmp",
123	[BPF_JEQ >> 4]  = "==",
124	[BPF_JGT >> 4]  = ">",
125	[BPF_JLT >> 4]  = "<",
126	[BPF_JGE >> 4]  = ">=",
127	[BPF_JLE >> 4]  = "<=",
128	[BPF_JSET >> 4] = "&",
129	[BPF_JNE >> 4]  = "!=",
130	[BPF_JSGT >> 4] = "s>",
131	[BPF_JSLT >> 4] = "s<",
132	[BPF_JSGE >> 4] = "s>=",
133	[BPF_JSLE >> 4] = "s<=",
134	[BPF_CALL >> 4] = "call",
135	[BPF_EXIT >> 4] = "exit",
136};
137
138static void print_bpf_end_insn(bpf_insn_print_t verbose,
139			       void *private_data,
140			       const struct bpf_insn *insn)
141{
142	verbose(private_data, "(%02x) r%d = %s%d r%d\n",
143		insn->code, insn->dst_reg,
144		BPF_SRC(insn->code) == BPF_TO_BE ? "be" : "le",
145		insn->imm, insn->dst_reg);
146}
147
148static void print_bpf_bswap_insn(bpf_insn_print_t verbose,
149			       void *private_data,
150			       const struct bpf_insn *insn)
151{
152	verbose(private_data, "(%02x) r%d = bswap%d r%d\n",
153		insn->code, insn->dst_reg,
154		insn->imm, insn->dst_reg);
155}
156
157static bool is_sdiv_smod(const struct bpf_insn *insn)
158{
159	return (BPF_OP(insn->code)  == BPF_DIV || BPF_OP(insn->code) == BPF_MOD) &&
160	       insn->off == 1;
161}
162
163static bool is_movsx(const struct bpf_insn *insn)
164{
165	return BPF_OP(insn->code) == BPF_MOV &&
166	       (insn->off == 8 || insn->off == 16 || insn->off == 32);
167}
168
169static bool is_addr_space_cast(const struct bpf_insn *insn)
170{
171	return insn->code == (BPF_ALU64 | BPF_MOV | BPF_X) &&
172		insn->off == BPF_ADDR_SPACE_CAST;
173}
174
175/* Special (internal-only) form of mov, used to resolve per-CPU addrs:
176 * dst_reg = src_reg + <percpu_base_off>
177 * BPF_ADDR_PERCPU is used as a special insn->off value.
178 */
179#define BPF_ADDR_PERCPU	(-1)
180
181static inline bool is_mov_percpu_addr(const struct bpf_insn *insn)
182{
183	return insn->code == (BPF_ALU64 | BPF_MOV | BPF_X) && insn->off == BPF_ADDR_PERCPU;
184}
185
186void print_bpf_insn(const struct bpf_insn_cbs *cbs,
187		    const struct bpf_insn *insn,
188		    bool allow_ptr_leaks)
189{
190	const bpf_insn_print_t verbose = cbs->cb_print;
191	u8 class = BPF_CLASS(insn->code);
192
193	if (class == BPF_ALU || class == BPF_ALU64) {
194		if (BPF_OP(insn->code) == BPF_END) {
195			if (class == BPF_ALU64)
196				print_bpf_bswap_insn(verbose, cbs->private_data, insn);
197			else
198				print_bpf_end_insn(verbose, cbs->private_data, insn);
199		} else if (BPF_OP(insn->code) == BPF_NEG) {
200			verbose(cbs->private_data, "(%02x) %c%d = -%c%d\n",
201				insn->code, class == BPF_ALU ? 'w' : 'r',
202				insn->dst_reg, class == BPF_ALU ? 'w' : 'r',
203				insn->dst_reg);
204		} else if (is_addr_space_cast(insn)) {
205			verbose(cbs->private_data, "(%02x) r%d = addr_space_cast(r%d, %u, %u)\n",
206				insn->code, insn->dst_reg,
207				insn->src_reg, ((u32)insn->imm) >> 16, (u16)insn->imm);
208		} else if (is_mov_percpu_addr(insn)) {
209			verbose(cbs->private_data, "(%02x) r%d = &(void __percpu *)(r%d)\n",
210				insn->code, insn->dst_reg, insn->src_reg);
211		} else if (BPF_SRC(insn->code) == BPF_X) {
212			verbose(cbs->private_data, "(%02x) %c%d %s %s%c%d\n",
213				insn->code, class == BPF_ALU ? 'w' : 'r',
214				insn->dst_reg,
215				is_sdiv_smod(insn) ? bpf_alu_sign_string[BPF_OP(insn->code) >> 4]
216						   : bpf_alu_string[BPF_OP(insn->code) >> 4],
217				is_movsx(insn) ? bpf_movsx_string[(insn->off >> 3) - 1] : "",
218				class == BPF_ALU ? 'w' : 'r',
219				insn->src_reg);
220		} else {
221			verbose(cbs->private_data, "(%02x) %c%d %s %d\n",
222				insn->code, class == BPF_ALU ? 'w' : 'r',
223				insn->dst_reg,
224				is_sdiv_smod(insn) ? bpf_alu_sign_string[BPF_OP(insn->code) >> 4]
225						   : bpf_alu_string[BPF_OP(insn->code) >> 4],
226				insn->imm);
227		}
228	} else if (class == BPF_STX) {
229		if (BPF_MODE(insn->code) == BPF_MEM)
230			verbose(cbs->private_data, "(%02x) *(%s *)(r%d %+d) = r%d\n",
231				insn->code,
232				bpf_ldst_string[BPF_SIZE(insn->code) >> 3],
233				insn->dst_reg,
234				insn->off, insn->src_reg);
235		else if (BPF_MODE(insn->code) == BPF_ATOMIC &&
236			 (insn->imm == BPF_ADD || insn->imm == BPF_AND ||
237			  insn->imm == BPF_OR || insn->imm == BPF_XOR)) {
238			verbose(cbs->private_data, "(%02x) lock *(%s *)(r%d %+d) %s r%d\n",
239				insn->code,
240				bpf_ldst_string[BPF_SIZE(insn->code) >> 3],
241				insn->dst_reg, insn->off,
242				bpf_alu_string[BPF_OP(insn->imm) >> 4],
243				insn->src_reg);
244		} else if (BPF_MODE(insn->code) == BPF_ATOMIC &&
245			   (insn->imm == (BPF_ADD | BPF_FETCH) ||
246			    insn->imm == (BPF_AND | BPF_FETCH) ||
247			    insn->imm == (BPF_OR | BPF_FETCH) ||
248			    insn->imm == (BPF_XOR | BPF_FETCH))) {
249			verbose(cbs->private_data, "(%02x) r%d = atomic%s_fetch_%s((%s *)(r%d %+d), r%d)\n",
250				insn->code, insn->src_reg,
251				BPF_SIZE(insn->code) == BPF_DW ? "64" : "",
252				bpf_atomic_alu_string[BPF_OP(insn->imm) >> 4],
253				bpf_ldst_string[BPF_SIZE(insn->code) >> 3],
254				insn->dst_reg, insn->off, insn->src_reg);
255		} else if (BPF_MODE(insn->code) == BPF_ATOMIC &&
256			   insn->imm == BPF_CMPXCHG) {
257			verbose(cbs->private_data, "(%02x) r0 = atomic%s_cmpxchg((%s *)(r%d %+d), r0, r%d)\n",
258				insn->code,
259				BPF_SIZE(insn->code) == BPF_DW ? "64" : "",
260				bpf_ldst_string[BPF_SIZE(insn->code) >> 3],
261				insn->dst_reg, insn->off,
262				insn->src_reg);
263		} else if (BPF_MODE(insn->code) == BPF_ATOMIC &&
264			   insn->imm == BPF_XCHG) {
265			verbose(cbs->private_data, "(%02x) r%d = atomic%s_xchg((%s *)(r%d %+d), r%d)\n",
266				insn->code, insn->src_reg,
267				BPF_SIZE(insn->code) == BPF_DW ? "64" : "",
268				bpf_ldst_string[BPF_SIZE(insn->code) >> 3],
269				insn->dst_reg, insn->off, insn->src_reg);
270		} else if (BPF_MODE(insn->code) == BPF_ATOMIC &&
271			   insn->imm == BPF_LOAD_ACQ) {
272			verbose(cbs->private_data, "(%02x) r%d = load_acquire((%s *)(r%d %+d))\n",
273				insn->code, insn->dst_reg,
274				bpf_ldst_string[BPF_SIZE(insn->code) >> 3],
275				insn->src_reg, insn->off);
276		} else if (BPF_MODE(insn->code) == BPF_ATOMIC &&
277			   insn->imm == BPF_STORE_REL) {
278			verbose(cbs->private_data, "(%02x) store_release((%s *)(r%d %+d), r%d)\n",
279				insn->code,
280				bpf_ldst_string[BPF_SIZE(insn->code) >> 3],
281				insn->dst_reg, insn->off, insn->src_reg);
282		} else {
283			verbose(cbs->private_data, "BUG_%02x\n", insn->code);
284		}
285	} else if (class == BPF_ST) {
286		if (BPF_MODE(insn->code) == BPF_MEM) {
287			verbose(cbs->private_data, "(%02x) *(%s *)(r%d %+d) = %d\n",
288				insn->code,
289				bpf_ldst_string[BPF_SIZE(insn->code) >> 3],
290				insn->dst_reg,
291				insn->off, insn->imm);
292		} else if (BPF_MODE(insn->code) == 0xc0 /* BPF_NOSPEC, no UAPI */) {
293			verbose(cbs->private_data, "(%02x) nospec\n", insn->code);
294		} else {
295			verbose(cbs->private_data, "BUG_st_%02x\n", insn->code);
296		}
297	} else if (class == BPF_LDX) {
298		if (BPF_MODE(insn->code) != BPF_MEM && BPF_MODE(insn->code) != BPF_MEMSX) {
299			verbose(cbs->private_data, "BUG_ldx_%02x\n", insn->code);
300			return;
301		}
302		verbose(cbs->private_data, "(%02x) r%d = *(%s *)(r%d %+d)\n",
303			insn->code, insn->dst_reg,
304			BPF_MODE(insn->code) == BPF_MEM ?
305				 bpf_ldst_string[BPF_SIZE(insn->code) >> 3] :
306				 bpf_ldsx_string[BPF_SIZE(insn->code) >> 3],
307			insn->src_reg, insn->off);
308	} else if (class == BPF_LD) {
309		if (BPF_MODE(insn->code) == BPF_ABS) {
310			verbose(cbs->private_data, "(%02x) r0 = *(%s *)skb[%d]\n",
311				insn->code,
312				bpf_ldst_string[BPF_SIZE(insn->code) >> 3],
313				insn->imm);
314		} else if (BPF_MODE(insn->code) == BPF_IND) {
315			verbose(cbs->private_data, "(%02x) r0 = *(%s *)skb[r%d + %d]\n",
316				insn->code,
317				bpf_ldst_string[BPF_SIZE(insn->code) >> 3],
318				insn->src_reg, insn->imm);
319		} else if (BPF_MODE(insn->code) == BPF_IMM &&
320			   BPF_SIZE(insn->code) == BPF_DW) {
321			/* At this point, we already made sure that the second
322			 * part of the ldimm64 insn is accessible.
323			 */
324			u64 imm = ((u64)(insn + 1)->imm << 32) | (u32)insn->imm;
325			bool is_ptr = insn->src_reg == BPF_PSEUDO_MAP_FD ||
326				      insn->src_reg == BPF_PSEUDO_MAP_VALUE;
327			char tmp[64];
328
329			if (is_ptr && !allow_ptr_leaks)
330				imm = 0;
331
332			verbose(cbs->private_data, "(%02x) r%d = %s\n",
333				insn->code, insn->dst_reg,
334				__func_imm_name(cbs, insn, imm,
335						tmp, sizeof(tmp)));
336		} else {
337			verbose(cbs->private_data, "BUG_ld_%02x\n", insn->code);
338			return;
339		}
340	} else if (class == BPF_JMP32 || class == BPF_JMP) {
341		u8 opcode = BPF_OP(insn->code);
342
343		if (opcode == BPF_CALL) {
344			char tmp[64];
345
346			if (insn->src_reg == BPF_PSEUDO_CALL) {
347				verbose(cbs->private_data, "(%02x) call pc%s\n",
348					insn->code,
349					__func_get_name(cbs, insn,
350							tmp, sizeof(tmp)));
351			} else {
352				strcpy(tmp, "unknown");
353				verbose(cbs->private_data, "(%02x) call %s#%d\n", insn->code,
354					__func_get_name(cbs, insn,
355							tmp, sizeof(tmp)),
356					insn->imm);
357			}
358		} else if (insn->code == (BPF_JMP | BPF_JA)) {
359			verbose(cbs->private_data, "(%02x) goto pc%+d\n",
360				insn->code, insn->off);
361		} else if (insn->code == (BPF_JMP | BPF_JA | BPF_X)) {
362			verbose(cbs->private_data, "(%02x) gotox r%d\n",
363				insn->code, insn->dst_reg);
364		} else if (insn->code == (BPF_JMP | BPF_JCOND) &&
365			   insn->src_reg == BPF_MAY_GOTO) {
366			verbose(cbs->private_data, "(%02x) may_goto pc%+d\n",
367				insn->code, insn->off);
368		} else if (insn->code == (BPF_JMP32 | BPF_JA)) {
369			verbose(cbs->private_data, "(%02x) gotol pc%+d\n",
370				insn->code, insn->imm);
371		} else if (insn->code == (BPF_JMP | BPF_EXIT)) {
372			verbose(cbs->private_data, "(%02x) exit\n", insn->code);
373		} else if (BPF_SRC(insn->code) == BPF_X) {
374			verbose(cbs->private_data,
375				"(%02x) if %c%d %s %c%d goto pc%+d\n",
376				insn->code, class == BPF_JMP32 ? 'w' : 'r',
377				insn->dst_reg,
378				bpf_jmp_string[BPF_OP(insn->code) >> 4],
379				class == BPF_JMP32 ? 'w' : 'r',
380				insn->src_reg, insn->off);
381		} else {
382			verbose(cbs->private_data,
383				"(%02x) if %c%d %s 0x%x goto pc%+d\n",
384				insn->code, class == BPF_JMP32 ? 'w' : 'r',
385				insn->dst_reg,
386				bpf_jmp_string[BPF_OP(insn->code) >> 4],
387				(u32)insn->imm, insn->off);
388		}
389	} else {
390		verbose(cbs->private_data, "(%02x) %s\n",
391			insn->code, bpf_class_string[class]);
392	}
393}