diff --git a/android/configs/android-recommended.cfg b/android/configs/android-recommended.cfg index 70aaae17ad29..28610303db60 100644 --- a/android/configs/android-recommended.cfg +++ b/android/configs/android-recommended.cfg @@ -15,6 +15,7 @@ CONFIG_BLK_DEV_RAM=y CONFIG_BLK_DEV_RAM_SIZE=8192 CONFIG_CC_STACKPROTECTOR_STRONG=y CONFIG_COMPACTION=y +CONFIG_CPU_SW_DOMAIN_PAN=y CONFIG_DEBUG_RODATA=y CONFIG_DM_UEVENT=y CONFIG_DRAGONRISE_FF=y diff --git a/arch/arm/crypto/aes-ce-glue.c b/arch/arm/crypto/aes-ce-glue.c index 679c589c4828..1f7b98e1a00d 100644 --- a/arch/arm/crypto/aes-ce-glue.c +++ b/arch/arm/crypto/aes-ce-glue.c @@ -369,7 +369,7 @@ static struct crypto_alg aes_algs[] = { { .cra_blkcipher = { .min_keysize = AES_MIN_KEY_SIZE, .max_keysize = AES_MAX_KEY_SIZE, - .ivsize = AES_BLOCK_SIZE, + .ivsize = 0, .setkey = ce_aes_setkey, .encrypt = ecb_encrypt, .decrypt = ecb_decrypt, @@ -446,7 +446,7 @@ static struct crypto_alg aes_algs[] = { { .cra_ablkcipher = { .min_keysize = AES_MIN_KEY_SIZE, .max_keysize = AES_MAX_KEY_SIZE, - .ivsize = AES_BLOCK_SIZE, + .ivsize = 0, .setkey = ablk_set_key, .encrypt = ablk_encrypt, .decrypt = ablk_decrypt, diff --git a/arch/arm64/crypto/Kconfig b/arch/arm64/crypto/Kconfig index 2cf32e9887e1..de1aab4b5da8 100644 --- a/arch/arm64/crypto/Kconfig +++ b/arch/arm64/crypto/Kconfig @@ -23,6 +23,11 @@ config CRYPTO_GHASH_ARM64_CE depends on ARM64 && KERNEL_MODE_NEON select CRYPTO_HASH +config CRYPTO_POLY_HASH_ARM64_CE + tristate "poly_hash (for HEH encryption mode) using ARMv8 Crypto Extensions" + depends on ARM64 && KERNEL_MODE_NEON + select CRYPTO_HASH + config CRYPTO_AES_ARM64_CE tristate "AES core cipher using ARMv8 Crypto Extensions" depends on ARM64 && KERNEL_MODE_NEON diff --git a/arch/arm64/crypto/Makefile b/arch/arm64/crypto/Makefile index abb79b3cfcfe..f0a8f2475ea3 100644 --- a/arch/arm64/crypto/Makefile +++ b/arch/arm64/crypto/Makefile @@ -17,6 +17,9 @@ sha2-ce-y := sha2-ce-glue.o sha2-ce-core.o obj-$(CONFIG_CRYPTO_GHASH_ARM64_CE) += ghash-ce.o ghash-ce-y := ghash-ce-glue.o ghash-ce-core.o +obj-$(CONFIG_CRYPTO_POLY_HASH_ARM64_CE) += poly-hash-ce.o +poly-hash-ce-y := poly-hash-ce-glue.o poly-hash-ce-core.o + obj-$(CONFIG_CRYPTO_AES_ARM64_CE) += aes-ce-cipher.o CFLAGS_aes-ce-cipher.o += -march=armv8-a+crypto diff --git a/arch/arm64/crypto/aes-glue.c b/arch/arm64/crypto/aes-glue.c index 6a51dfccfe71..448b874a4826 100644 --- a/arch/arm64/crypto/aes-glue.c +++ b/arch/arm64/crypto/aes-glue.c @@ -294,7 +294,7 @@ static struct crypto_alg aes_algs[] = { { .cra_blkcipher = { .min_keysize = AES_MIN_KEY_SIZE, .max_keysize = AES_MAX_KEY_SIZE, - .ivsize = AES_BLOCK_SIZE, + .ivsize = 0, .setkey = aes_setkey, .encrypt = ecb_encrypt, .decrypt = ecb_decrypt, @@ -371,7 +371,7 @@ static struct crypto_alg aes_algs[] = { { .cra_ablkcipher = { .min_keysize = AES_MIN_KEY_SIZE, .max_keysize = AES_MAX_KEY_SIZE, - .ivsize = AES_BLOCK_SIZE, + .ivsize = 0, .setkey = ablk_set_key, .encrypt = ablk_encrypt, .decrypt = ablk_decrypt, diff --git a/arch/arm64/crypto/poly-hash-ce-core.S b/arch/arm64/crypto/poly-hash-ce-core.S new file mode 100644 index 000000000000..8ccb544c5526 --- /dev/null +++ b/arch/arm64/crypto/poly-hash-ce-core.S @@ -0,0 +1,163 @@ +/* + * Accelerated poly_hash implementation with ARMv8 PMULL instructions. + * + * Based on ghash-ce-core.S. + * + * Copyright (C) 2014 Linaro Ltd. + * Copyright (C) 2017 Google, Inc. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License version 2 as published + * by the Free Software Foundation. + */ + +#include +#include + + KEY .req v0 + KEY2 .req v1 + T1 .req v2 + T2 .req v3 + GSTAR .req v4 + XL .req v5 + XM .req v6 + XH .req v7 + + .text + .arch armv8-a+crypto + + /* 16-byte aligned (2**4 = 16); not required, but might as well */ + .align 4 +.Lgstar: + .quad 0x87, 0x87 + +/* + * void pmull_poly_hash_update(le128 *digest, const le128 *key, + * const u8 *src, unsigned int blocks, + * unsigned int partial); + */ +ENTRY(pmull_poly_hash_update) + + /* Load digest into XL */ + ld1 {XL.16b}, [x0] + + /* Load key into KEY */ + ld1 {KEY.16b}, [x1] + + /* Load g*(x) = g(x) + x^128 = x^7 + x^2 + x + 1 into both halves of + * GSTAR */ + adr x1, .Lgstar + ld1 {GSTAR.2d}, [x1] + + /* Set KEY2 to (KEY[1]+KEY[0]):(KEY[1]+KEY[0]). This is needed for + * Karatsuba multiplication. */ + ext KEY2.16b, KEY.16b, KEY.16b, #8 + eor KEY2.16b, KEY2.16b, KEY.16b + + /* If 'partial' is nonzero, then we're finishing a pending block and + * should go right to the multiplication. */ + cbnz w4, 1f + +0: + /* Add the next block from 'src' to the digest */ + ld1 {T1.16b}, [x2], #16 + eor XL.16b, XL.16b, T1.16b + sub w3, w3, #1 + +1: + /* + * Multiply the current 128-bit digest (a1:a0, in XL) by the 128-bit key + * (b1:b0, in KEY) using Karatsuba multiplication. + */ + + /* T1 = (a1+a0):(a1+a0) */ + ext T1.16b, XL.16b, XL.16b, #8 + eor T1.16b, T1.16b, XL.16b + + /* XH = a1 * b1 */ + pmull2 XH.1q, XL.2d, KEY.2d + + /* XL = a0 * b0 */ + pmull XL.1q, XL.1d, KEY.1d + + /* XM = (a1+a0) * (b1+b0) */ + pmull XM.1q, T1.1d, KEY2.1d + + /* XM += (XH[0]:XL[1]) + XL + XH */ + ext T1.16b, XL.16b, XH.16b, #8 + eor T2.16b, XL.16b, XH.16b + eor XM.16b, XM.16b, T1.16b + eor XM.16b, XM.16b, T2.16b + + /* + * Now the 256-bit product is in XH[1]:XM:XL[0]. It represents a + * polynomial over GF(2) with degree as large as 255. We need to + * compute its remainder modulo g(x) = x^128+x^7+x^2+x+1. For this it + * is sufficient to compute the remainder of the high half 'c(x)x^128' + * add it to the low half. To reduce the high half we use the Barrett + * reduction method. The basic idea is that we can express the + * remainder p(x) as g(x)q(x) mod x^128, where q(x) = (c(x)x^128)/g(x). + * As detailed in [1], to avoid having to divide by g(x) at runtime the + * following equivalent expression can be derived: + * + * p(x) = [ g*(x)((c(x)q+(x))/x^128) ] mod x^128 + * + * where g*(x) = x^128+g(x) = x^7+x^2+x+1, and q+(x) = x^256/g(x) = g(x) + * in this case. This is also equivalent to: + * + * p(x) = [ g*(x)((c(x)(x^128 + g*(x)))/x^128) ] mod x^128 + * = [ g*(x)(c(x) + (c(x)g*(x))/x^128) ] mod x^128 + * + * Since deg g*(x) < 64: + * + * p(x) = [ g*(x)(c(x) + ((c(x)/x^64)g*(x))/x^64) ] mod x^128 + * = [ g*(x)((c(x)/x^64)x^64 + (c(x) mod x^64) + + * ((c(x)/x^64)g*(x))/x^64) ] mod x^128 + * + * Letting t(x) = g*(x)(c(x)/x^64): + * + * p(x) = [ t(x)x^64 + g*(x)((c(x) mod x^64) + t(x)/x^64) ] mod x^128 + * + * Therefore, to do the reduction we only need to issue two 64-bit => + * 128-bit carryless multiplications: g*(x) times c(x)/x^64, and g*(x) + * times ((c(x) mod x^64) + t(x)/x^64). (Multiplication by x^64 doesn't + * count since it is simply a shift or move.) + * + * An alternate reduction method, also based on Barrett reduction and + * described in [1], uses only shifts and XORs --- no multiplications. + * However, the method with multiplications requires fewer instructions + * and is faster on processors with fast carryless multiplication. + * + * [1] "Intel Carry-Less Multiplication Instruction and its Usage for + * Computing the GCM Mode", + * https://software.intel.com/sites/default/files/managed/72/cc/clmul-wp-rev-2.02-2014-04-20.pdf + */ + + /* 256-bit product is XH[1]:XM:XL[0], so c(x) is XH[1]:XM[1] */ + + /* T1 = t(x) = g*(x)(c(x)/x^64) */ + pmull2 T1.1q, GSTAR.2d, XH.2d + + /* T2 = g*(x)((c(x) mod x^64) + t(x)/x^64) */ + eor T2.16b, XM.16b, T1.16b + pmull2 T2.1q, GSTAR.2d, T2.2d + + /* Make XL[0] be the low half of the 128-bit result by adding the low 64 + * bits of the T2 term to what was already there. The 't(x)x^64' term + * makes no difference, so skip it. */ + eor XL.16b, XL.16b, T2.16b + + /* Make XL[1] be the high half of the 128-bit result by adding the high + * 64 bits of the 't(x)x^64' and T2 terms to what was already in XM[0], + * then moving XM[0] to XL[1]. */ + eor XM.16b, XM.16b, T1.16b + ext T2.16b, T2.16b, T2.16b, #8 + eor XM.16b, XM.16b, T2.16b + mov XL.d[1], XM.d[0] + + /* If more blocks remain, then loop back to process the next block; + * else, store the digest and return. */ + cbnz w3, 0b + st1 {XL.16b}, [x0] + ret +ENDPROC(pmull_poly_hash_update) diff --git a/arch/arm64/crypto/poly-hash-ce-glue.c b/arch/arm64/crypto/poly-hash-ce-glue.c new file mode 100644 index 000000000000..e195740c9ecf --- /dev/null +++ b/arch/arm64/crypto/poly-hash-ce-glue.c @@ -0,0 +1,166 @@ +/* + * Accelerated poly_hash implementation with ARMv8 PMULL instructions. + * + * Based on ghash-ce-glue.c. + * + * poly_hash is part of the HEH (Hash-Encrypt-Hash) encryption mode, proposed in + * Internet Draft https://tools.ietf.org/html/draft-cope-heh-01. + * + * poly_hash is very similar to GHASH: both algorithms are keyed hashes which + * interpret their input data as coefficients of a polynomial over GF(2^128), + * then calculate a hash value by evaluating that polynomial at the point given + * by the key, e.g. using Horner's rule. The difference is that poly_hash uses + * the more natural "ble" convention to represent GF(2^128) elements, whereas + * GHASH uses the less natural "lle" convention (see include/crypto/gf128mul.h). + * The ble convention makes it simpler to implement GF(2^128) multiplication. + * + * Copyright (C) 2014 Linaro Ltd. + * Copyright (C) 2017 Google Inc. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License version 2 as published + * by the Free Software Foundation. + */ + +#include +#include +#include +#include +#include +#include + +/* + * Note: in this algorithm we currently use 'le128' to represent GF(2^128) + * elements, even though poly_hash-generic uses 'be128'. Both types are + * actually "wrong" because the elements are actually in 'ble' format, and there + * should be a ble type to represent this --- as well as lle, bbe, and lbe types + * for the other conventions for representing GF(2^128) elements. But + * practically it doesn't matter which type we choose here, so we just use le128 + * since it's arguably more accurate, while poly_hash-generic still has to use + * be128 because the generic GF(2^128) multiplication functions all take be128. + */ + +struct poly_hash_desc_ctx { + le128 digest; + unsigned int count; +}; + +asmlinkage void pmull_poly_hash_update(le128 *digest, const le128 *key, + const u8 *src, unsigned int blocks, + unsigned int partial); + +static int poly_hash_setkey(struct crypto_shash *tfm, + const u8 *key, unsigned int keylen) +{ + if (keylen != sizeof(le128)) { + crypto_shash_set_flags(tfm, CRYPTO_TFM_RES_BAD_KEY_LEN); + return -EINVAL; + } + + memcpy(crypto_shash_ctx(tfm), key, sizeof(le128)); + return 0; +} + +static int poly_hash_init(struct shash_desc *desc) +{ + struct poly_hash_desc_ctx *ctx = shash_desc_ctx(desc); + + ctx->digest = (le128) { 0 }; + ctx->count = 0; + return 0; +} + +static int poly_hash_update(struct shash_desc *desc, const u8 *src, + unsigned int len) +{ + struct poly_hash_desc_ctx *ctx = shash_desc_ctx(desc); + unsigned int partial = ctx->count % sizeof(le128); + u8 *dst = (u8 *)&ctx->digest + partial; + + ctx->count += len; + + /* Finishing at least one block? */ + if (partial + len >= sizeof(le128)) { + const le128 *key = crypto_shash_ctx(desc->tfm); + + if (partial) { + /* Finish the pending block. */ + unsigned int n = sizeof(le128) - partial; + + len -= n; + do { + *dst++ ^= *src++; + } while (--n); + } + + /* + * Do the real work. If 'partial' is nonzero, this starts by + * multiplying 'digest' by 'key'. Then for each additional full + * block it adds the block to 'digest' and multiplies by 'key'. + */ + kernel_neon_begin_partial(8); + pmull_poly_hash_update(&ctx->digest, key, src, + len / sizeof(le128), partial); + kernel_neon_end(); + + src += len - (len % sizeof(le128)); + len %= sizeof(le128); + dst = (u8 *)&ctx->digest; + } + + /* Continue adding the next block to 'digest'. */ + while (len--) + *dst++ ^= *src++; + return 0; +} + +static int poly_hash_final(struct shash_desc *desc, u8 *out) +{ + struct poly_hash_desc_ctx *ctx = shash_desc_ctx(desc); + unsigned int partial = ctx->count % sizeof(le128); + + /* Finish the last block if needed. */ + if (partial) { + const le128 *key = crypto_shash_ctx(desc->tfm); + + kernel_neon_begin_partial(8); + pmull_poly_hash_update(&ctx->digest, key, NULL, 0, partial); + kernel_neon_end(); + } + + memcpy(out, &ctx->digest, sizeof(le128)); + return 0; +} + +static struct shash_alg poly_hash_alg = { + .digestsize = sizeof(le128), + .init = poly_hash_init, + .update = poly_hash_update, + .final = poly_hash_final, + .setkey = poly_hash_setkey, + .descsize = sizeof(struct poly_hash_desc_ctx), + .base = { + .cra_name = "poly_hash", + .cra_driver_name = "poly_hash-ce", + .cra_priority = 300, + .cra_ctxsize = sizeof(le128), + .cra_module = THIS_MODULE, + }, +}; + +static int __init poly_hash_ce_mod_init(void) +{ + return crypto_register_shash(&poly_hash_alg); +} + +static void __exit poly_hash_ce_mod_exit(void) +{ + crypto_unregister_shash(&poly_hash_alg); +} + +MODULE_DESCRIPTION("Polynomial evaluation hash using ARMv8 Crypto Extensions"); +MODULE_AUTHOR("Eric Biggers "); +MODULE_LICENSE("GPL v2"); + +module_cpu_feature_match(PMULL, poly_hash_ce_mod_init); +module_exit(poly_hash_ce_mod_exit); diff --git a/arch/arm64/include/asm/hw_breakpoint.h b/arch/arm64/include/asm/hw_breakpoint.h index 9732908bfc8a..c72b8e201ab4 100644 --- a/arch/arm64/include/asm/hw_breakpoint.h +++ b/arch/arm64/include/asm/hw_breakpoint.h @@ -68,7 +68,11 @@ static inline void decode_ctrl_reg(u32 reg, /* Lengths */ #define ARM_BREAKPOINT_LEN_1 0x1 #define ARM_BREAKPOINT_LEN_2 0x3 +#define ARM_BREAKPOINT_LEN_3 0x7 #define ARM_BREAKPOINT_LEN_4 0xf +#define ARM_BREAKPOINT_LEN_5 0x1f +#define ARM_BREAKPOINT_LEN_6 0x3f +#define ARM_BREAKPOINT_LEN_7 0x7f #define ARM_BREAKPOINT_LEN_8 0xff /* Kernel stepping */ @@ -110,7 +114,7 @@ struct perf_event; struct pmu; extern int arch_bp_generic_fields(struct arch_hw_breakpoint_ctrl ctrl, - int *gen_len, int *gen_type); + int *gen_len, int *gen_type, int *offset); extern int arch_check_bp_in_kernelspace(struct perf_event *bp); extern int arch_validate_hwbkpt_settings(struct perf_event *bp); extern int hw_breakpoint_exceptions_notify(struct notifier_block *unused, diff --git a/arch/arm64/kernel/hw_breakpoint.c b/arch/arm64/kernel/hw_breakpoint.c index b45c95d34b83..2d2792a714ad 100644 --- a/arch/arm64/kernel/hw_breakpoint.c +++ b/arch/arm64/kernel/hw_breakpoint.c @@ -313,9 +313,21 @@ static int get_hbp_len(u8 hbp_len) case ARM_BREAKPOINT_LEN_2: len_in_bytes = 2; break; + case ARM_BREAKPOINT_LEN_3: + len_in_bytes = 3; + break; case ARM_BREAKPOINT_LEN_4: len_in_bytes = 4; break; + case ARM_BREAKPOINT_LEN_5: + len_in_bytes = 5; + break; + case ARM_BREAKPOINT_LEN_6: + len_in_bytes = 6; + break; + case ARM_BREAKPOINT_LEN_7: + len_in_bytes = 7; + break; case ARM_BREAKPOINT_LEN_8: len_in_bytes = 8; break; @@ -345,7 +357,7 @@ int arch_check_bp_in_kernelspace(struct perf_event *bp) * to generic breakpoint descriptions. */ int arch_bp_generic_fields(struct arch_hw_breakpoint_ctrl ctrl, - int *gen_len, int *gen_type) + int *gen_len, int *gen_type, int *offset) { /* Type */ switch (ctrl.type) { @@ -365,17 +377,33 @@ int arch_bp_generic_fields(struct arch_hw_breakpoint_ctrl ctrl, return -EINVAL; } + if (!ctrl.len) + return -EINVAL; + *offset = __ffs(ctrl.len); + /* Len */ - switch (ctrl.len) { + switch (ctrl.len >> *offset) { case ARM_BREAKPOINT_LEN_1: *gen_len = HW_BREAKPOINT_LEN_1; break; case ARM_BREAKPOINT_LEN_2: *gen_len = HW_BREAKPOINT_LEN_2; break; + case ARM_BREAKPOINT_LEN_3: + *gen_len = HW_BREAKPOINT_LEN_3; + break; case ARM_BREAKPOINT_LEN_4: *gen_len = HW_BREAKPOINT_LEN_4; break; + case ARM_BREAKPOINT_LEN_5: + *gen_len = HW_BREAKPOINT_LEN_5; + break; + case ARM_BREAKPOINT_LEN_6: + *gen_len = HW_BREAKPOINT_LEN_6; + break; + case ARM_BREAKPOINT_LEN_7: + *gen_len = HW_BREAKPOINT_LEN_7; + break; case ARM_BREAKPOINT_LEN_8: *gen_len = HW_BREAKPOINT_LEN_8; break; @@ -419,9 +447,21 @@ static int arch_build_bp_info(struct perf_event *bp) case HW_BREAKPOINT_LEN_2: info->ctrl.len = ARM_BREAKPOINT_LEN_2; break; + case HW_BREAKPOINT_LEN_3: + info->ctrl.len = ARM_BREAKPOINT_LEN_3; + break; case HW_BREAKPOINT_LEN_4: info->ctrl.len = ARM_BREAKPOINT_LEN_4; break; + case HW_BREAKPOINT_LEN_5: + info->ctrl.len = ARM_BREAKPOINT_LEN_5; + break; + case HW_BREAKPOINT_LEN_6: + info->ctrl.len = ARM_BREAKPOINT_LEN_6; + break; + case HW_BREAKPOINT_LEN_7: + info->ctrl.len = ARM_BREAKPOINT_LEN_7; + break; case HW_BREAKPOINT_LEN_8: info->ctrl.len = ARM_BREAKPOINT_LEN_8; break; @@ -513,18 +553,17 @@ int arch_validate_hwbkpt_settings(struct perf_event *bp) default: return -EINVAL; } - - info->address &= ~alignment_mask; - info->ctrl.len <<= offset; } else { if (info->ctrl.type == ARM_BREAKPOINT_EXECUTE) alignment_mask = 0x3; else alignment_mask = 0x7; - if (info->address & alignment_mask) - return -EINVAL; + offset = info->address & alignment_mask; } + info->address &= ~alignment_mask; + info->ctrl.len <<= offset; + /* * Disallow per-task kernel breakpoints since these would * complicate the stepping code. @@ -655,12 +694,47 @@ unlock: return 0; } +/* + * Arm64 hardware does not always report a watchpoint hit address that matches + * one of the watchpoints set. It can also report an address "near" the + * watchpoint if a single instruction access both watched and unwatched + * addresses. There is no straight-forward way, short of disassembling the + * offending instruction, to map that address back to the watchpoint. This + * function computes the distance of the memory access from the watchpoint as a + * heuristic for the likelyhood that a given access triggered the watchpoint. + * + * See Section D2.10.5 "Determining the memory location that caused a Watchpoint + * exception" of ARMv8 Architecture Reference Manual for details. + * + * The function returns the distance of the address from the bytes watched by + * the watchpoint. In case of an exact match, it returns 0. + */ +static u64 get_distance_from_watchpoint(unsigned long addr, u64 val, + struct arch_hw_breakpoint_ctrl *ctrl) +{ + u64 wp_low, wp_high; + u32 lens, lene; + + lens = __ffs(ctrl->len); + lene = __fls(ctrl->len); + + wp_low = val + lens; + wp_high = val + lene; + if (addr < wp_low) + return wp_low - addr; + else if (addr > wp_high) + return addr - wp_high; + else + return 0; +} + static int watchpoint_handler(unsigned long addr, unsigned int esr, struct pt_regs *regs) { - int i, step = 0, *kernel_step, access; + int i, step = 0, *kernel_step, access, closest_match = 0; + u64 min_dist = -1, dist; u32 ctrl_reg; - u64 val, alignment_mask; + u64 val; struct perf_event *wp, **slots; struct debug_info *debug_info; struct arch_hw_breakpoint *info; @@ -669,35 +743,15 @@ static int watchpoint_handler(unsigned long addr, unsigned int esr, slots = this_cpu_ptr(wp_on_reg); debug_info = ¤t->thread.debug; + /* + * Find all watchpoints that match the reported address. If no exact + * match is found. Attribute the hit to the closest watchpoint. + */ + rcu_read_lock(); for (i = 0; i < core_num_wrps; ++i) { - rcu_read_lock(); - wp = slots[i]; - if (wp == NULL) - goto unlock; - - info = counter_arch_bp(wp); - /* AArch32 watchpoints are either 4 or 8 bytes aligned. */ - if (is_compat_task()) { - if (info->ctrl.len == ARM_BREAKPOINT_LEN_8) - alignment_mask = 0x7; - else - alignment_mask = 0x3; - } else { - alignment_mask = 0x7; - } - - /* Check if the watchpoint value matches. */ - val = read_wb_reg(AARCH64_DBG_REG_WVR, i); - if (val != (addr & ~alignment_mask)) - goto unlock; - - /* Possible match, check the byte address select to confirm. */ - ctrl_reg = read_wb_reg(AARCH64_DBG_REG_WCR, i); - decode_ctrl_reg(ctrl_reg, &ctrl); - if (!((1 << (addr & alignment_mask)) & ctrl.len)) - goto unlock; + continue; /* * Check that the access type matches. @@ -706,18 +760,41 @@ static int watchpoint_handler(unsigned long addr, unsigned int esr, access = (esr & AARCH64_ESR_ACCESS_MASK) ? HW_BREAKPOINT_W : HW_BREAKPOINT_R; if (!(access & hw_breakpoint_type(wp))) - goto unlock; + continue; + /* Check if the watchpoint value and byte select match. */ + val = read_wb_reg(AARCH64_DBG_REG_WVR, i); + ctrl_reg = read_wb_reg(AARCH64_DBG_REG_WCR, i); + decode_ctrl_reg(ctrl_reg, &ctrl); + dist = get_distance_from_watchpoint(addr, val, &ctrl); + if (dist < min_dist) { + min_dist = dist; + closest_match = i; + } + /* Is this an exact match? */ + if (dist != 0) + continue; + + info = counter_arch_bp(wp); info->trigger = addr; perf_bp_event(wp, regs); /* Do we need to handle the stepping? */ if (!wp->overflow_handler) step = 1; - -unlock: - rcu_read_unlock(); } + if (min_dist > 0 && min_dist != -1) { + /* No exact match found. */ + wp = slots[closest_match]; + info = counter_arch_bp(wp); + info->trigger = addr; + perf_bp_event(wp, regs); + + /* Do we need to handle the stepping? */ + if (!wp->overflow_handler) + step = 1; + } + rcu_read_unlock(); if (!step) return 0; diff --git a/arch/arm64/kernel/ptrace.c b/arch/arm64/kernel/ptrace.c index 55909b2208cc..beff0fb11b6b 100644 --- a/arch/arm64/kernel/ptrace.c +++ b/arch/arm64/kernel/ptrace.c @@ -227,13 +227,13 @@ static int ptrace_hbp_fill_attr_ctrl(unsigned int note_type, struct arch_hw_breakpoint_ctrl ctrl, struct perf_event_attr *attr) { - int err, len, type, disabled = !ctrl.enabled; + int err, len, type, offset, disabled = !ctrl.enabled; attr->disabled = disabled; if (disabled) return 0; - err = arch_bp_generic_fields(ctrl, &len, &type); + err = arch_bp_generic_fields(ctrl, &len, &type, &offset); if (err) return err; @@ -252,6 +252,7 @@ static int ptrace_hbp_fill_attr_ctrl(unsigned int note_type, attr->bp_len = len; attr->bp_type = type; + attr->bp_addr += offset; return 0; } @@ -304,7 +305,7 @@ static int ptrace_hbp_get_addr(unsigned int note_type, if (IS_ERR(bp)) return PTR_ERR(bp); - *addr = bp ? bp->attr.bp_addr : 0; + *addr = bp ? counter_arch_bp(bp)->address : 0; return 0; } diff --git a/crypto/Kconfig b/crypto/Kconfig index 7240821137fd..3240d394426c 100644 --- a/crypto/Kconfig +++ b/crypto/Kconfig @@ -289,6 +289,24 @@ config CRYPTO_CBC CBC: Cipher Block Chaining mode This block cipher algorithm is required for IPSec. +config CRYPTO_HEH + tristate "HEH support" + select CRYPTO_CMAC + select CRYPTO_ECB + select CRYPTO_GF128MUL + select CRYPTO_MANAGER + select CRYPTO_POLY_HASH_ARM64_CE if ARM64 && KERNEL_MODE_NEON + help + HEH: Hash-Encrypt-Hash mode + HEH is a proposed block cipher mode of operation which extends the + strong pseudo-random permutation (SPRP) property of block ciphers to + arbitrary-length input strings. This provides a stronger notion of + security than existing block cipher modes of operation (e.g. CBC, CTR, + XTS), though it is usually less performant. Applications include disk + encryption and encryption of file names and contents. Currently, this + implementation only provides a symmetric cipher interface, so it can't + yet be used as an AEAD. + config CRYPTO_CTR tristate "CTR support" select CRYPTO_BLKCIPHER diff --git a/crypto/Makefile b/crypto/Makefile index 82fbff180ad3..35b5d5c5c65c 100644 --- a/crypto/Makefile +++ b/crypto/Makefile @@ -66,6 +66,7 @@ obj-$(CONFIG_CRYPTO_TGR192) += tgr192.o obj-$(CONFIG_CRYPTO_GF128MUL) += gf128mul.o obj-$(CONFIG_CRYPTO_ECB) += ecb.o obj-$(CONFIG_CRYPTO_CBC) += cbc.o +obj-$(CONFIG_CRYPTO_HEH) += heh.o obj-$(CONFIG_CRYPTO_PCBC) += pcbc.o obj-$(CONFIG_CRYPTO_CTS) += cts.o obj-$(CONFIG_CRYPTO_LRW) += lrw.o diff --git a/crypto/blkcipher.c b/crypto/blkcipher.c index dca7bc87dad9..7bbfadc195a6 100644 --- a/crypto/blkcipher.c +++ b/crypto/blkcipher.c @@ -373,6 +373,27 @@ int blkcipher_aead_walk_virt_block(struct blkcipher_desc *desc, } EXPORT_SYMBOL_GPL(blkcipher_aead_walk_virt_block); +/* + * This function allows ablkcipher algorithms to use the blkcipher_walk API to + * walk over their data. The specified crypto_ablkcipher tfm is used to + * initialize the struct blkcipher_walk, and the crypto_blkcipher specified in + * desc->tfm is never used so it can be left NULL. (Yes, this design is ugly, + * but it parallels blkcipher_aead_walk_virt_block() above. In the 4.10 kernel + * this is starting to be cleaned up...) + */ +int blkcipher_ablkcipher_walk_virt(struct blkcipher_desc *desc, + struct blkcipher_walk *walk, + struct crypto_ablkcipher *tfm) +{ + walk->flags &= ~BLKCIPHER_WALK_PHYS; + walk->walk_blocksize = crypto_ablkcipher_blocksize(tfm); + walk->cipher_blocksize = walk->walk_blocksize; + walk->ivsize = crypto_ablkcipher_ivsize(tfm); + walk->alignmask = crypto_ablkcipher_alignmask(tfm); + return blkcipher_walk_first(desc, walk); +} +EXPORT_SYMBOL_GPL(blkcipher_ablkcipher_walk_virt); + static int setkey_unaligned(struct crypto_tfm *tfm, const u8 *key, unsigned int keylen) { diff --git a/crypto/gf128mul.c b/crypto/gf128mul.c index 5276607c72d0..f3d9f6da0767 100644 --- a/crypto/gf128mul.c +++ b/crypto/gf128mul.c @@ -44,7 +44,7 @@ --------------------------------------------------------------------------- Issue 31/01/2006 - This file provides fast multiplication in GF(128) as required by several + This file provides fast multiplication in GF(2^128) as required by several cryptographic authentication modes */ @@ -88,37 +88,52 @@ q(0xf8), q(0xf9), q(0xfa), q(0xfb), q(0xfc), q(0xfd), q(0xfe), q(0xff) \ } -/* Given the value i in 0..255 as the byte overflow when a field element - in GHASH is multiplied by x^8, this function will return the values that - are generated in the lo 16-bit word of the field value by applying the - modular polynomial. The values lo_byte and hi_byte are returned via the - macro xp_fun(lo_byte, hi_byte) so that the values can be assembled into - memory as required by a suitable definition of this macro operating on - the table above -*/ +/* + * Given a value i in 0..255 as the byte overflow when a field element + * in GF(2^128) is multiplied by x^8, the following macro returns the + * 16-bit value that must be XOR-ed into the low-degree end of the + * product to reduce it modulo the irreducible polynomial x^128 + x^7 + + * x^2 + x + 1. + * + * There are two versions of the macro, and hence two tables: one for + * the "be" convention where the highest-order bit is the coefficient of + * the highest-degree polynomial term, and one for the "le" convention + * where the highest-order bit is the coefficient of the lowest-degree + * polynomial term. In both cases the values are stored in CPU byte + * endianness such that the coefficients are ordered consistently across + * bytes, i.e. in the "be" table bits 15..0 of the stored value + * correspond to the coefficients of x^15..x^0, and in the "le" table + * bits 15..0 correspond to the coefficients of x^0..x^15. + * + * Therefore, provided that the appropriate byte endianness conversions + * are done by the multiplication functions (and these must be in place + * anyway to support both little endian and big endian CPUs), the "be" + * table can be used for multiplications of both "bbe" and "ble" + * elements, and the "le" table can be used for multiplications of both + * "lle" and "lbe" elements. + */ -#define xx(p, q) 0x##p##q - -#define xda_bbe(i) ( \ - (i & 0x80 ? xx(43, 80) : 0) ^ (i & 0x40 ? xx(21, c0) : 0) ^ \ - (i & 0x20 ? xx(10, e0) : 0) ^ (i & 0x10 ? xx(08, 70) : 0) ^ \ - (i & 0x08 ? xx(04, 38) : 0) ^ (i & 0x04 ? xx(02, 1c) : 0) ^ \ - (i & 0x02 ? xx(01, 0e) : 0) ^ (i & 0x01 ? xx(00, 87) : 0) \ +#define xda_be(i) ( \ + (i & 0x80 ? 0x4380 : 0) ^ (i & 0x40 ? 0x21c0 : 0) ^ \ + (i & 0x20 ? 0x10e0 : 0) ^ (i & 0x10 ? 0x0870 : 0) ^ \ + (i & 0x08 ? 0x0438 : 0) ^ (i & 0x04 ? 0x021c : 0) ^ \ + (i & 0x02 ? 0x010e : 0) ^ (i & 0x01 ? 0x0087 : 0) \ ) -#define xda_lle(i) ( \ - (i & 0x80 ? xx(e1, 00) : 0) ^ (i & 0x40 ? xx(70, 80) : 0) ^ \ - (i & 0x20 ? xx(38, 40) : 0) ^ (i & 0x10 ? xx(1c, 20) : 0) ^ \ - (i & 0x08 ? xx(0e, 10) : 0) ^ (i & 0x04 ? xx(07, 08) : 0) ^ \ - (i & 0x02 ? xx(03, 84) : 0) ^ (i & 0x01 ? xx(01, c2) : 0) \ +#define xda_le(i) ( \ + (i & 0x80 ? 0xe100 : 0) ^ (i & 0x40 ? 0x7080 : 0) ^ \ + (i & 0x20 ? 0x3840 : 0) ^ (i & 0x10 ? 0x1c20 : 0) ^ \ + (i & 0x08 ? 0x0e10 : 0) ^ (i & 0x04 ? 0x0708 : 0) ^ \ + (i & 0x02 ? 0x0384 : 0) ^ (i & 0x01 ? 0x01c2 : 0) \ ) -static const u16 gf128mul_table_lle[256] = gf128mul_dat(xda_lle); -static const u16 gf128mul_table_bbe[256] = gf128mul_dat(xda_bbe); +static const u16 gf128mul_table_le[256] = gf128mul_dat(xda_le); +static const u16 gf128mul_table_be[256] = gf128mul_dat(xda_be); -/* These functions multiply a field element by x, by x^4 and by x^8 - * in the polynomial field representation. It uses 32-bit word operations - * to gain speed but compensates for machine endianess and hence works +/* + * The following functions multiply a field element by x or by x^8 in + * the polynomial field representation. They use 64-bit word operations + * to gain speed but compensate for machine endianness and hence work * correctly on both styles of machine. */ @@ -126,7 +141,7 @@ static void gf128mul_x_lle(be128 *r, const be128 *x) { u64 a = be64_to_cpu(x->a); u64 b = be64_to_cpu(x->b); - u64 _tt = gf128mul_table_lle[(b << 7) & 0xff]; + u64 _tt = gf128mul_table_le[(b << 7) & 0xff]; r->b = cpu_to_be64((b >> 1) | (a << 63)); r->a = cpu_to_be64((a >> 1) ^ (_tt << 48)); @@ -136,7 +151,7 @@ static void gf128mul_x_bbe(be128 *r, const be128 *x) { u64 a = be64_to_cpu(x->a); u64 b = be64_to_cpu(x->b); - u64 _tt = gf128mul_table_bbe[a >> 63]; + u64 _tt = gf128mul_table_be[a >> 63]; r->a = cpu_to_be64((a << 1) | (b >> 63)); r->b = cpu_to_be64((b << 1) ^ _tt); @@ -146,7 +161,7 @@ void gf128mul_x_ble(be128 *r, const be128 *x) { u64 a = le64_to_cpu(x->a); u64 b = le64_to_cpu(x->b); - u64 _tt = gf128mul_table_bbe[b >> 63]; + u64 _tt = gf128mul_table_be[b >> 63]; r->a = cpu_to_le64((a << 1) ^ _tt); r->b = cpu_to_le64((b << 1) | (a >> 63)); @@ -157,7 +172,7 @@ static void gf128mul_x8_lle(be128 *x) { u64 a = be64_to_cpu(x->a); u64 b = be64_to_cpu(x->b); - u64 _tt = gf128mul_table_lle[b & 0xff]; + u64 _tt = gf128mul_table_le[b & 0xff]; x->b = cpu_to_be64((b >> 8) | (a << 56)); x->a = cpu_to_be64((a >> 8) ^ (_tt << 48)); @@ -167,12 +182,22 @@ static void gf128mul_x8_bbe(be128 *x) { u64 a = be64_to_cpu(x->a); u64 b = be64_to_cpu(x->b); - u64 _tt = gf128mul_table_bbe[a >> 56]; + u64 _tt = gf128mul_table_be[a >> 56]; x->a = cpu_to_be64((a << 8) | (b >> 56)); x->b = cpu_to_be64((b << 8) ^ _tt); } +static void gf128mul_x8_ble(be128 *x) +{ + u64 a = le64_to_cpu(x->b); + u64 b = le64_to_cpu(x->a); + u64 _tt = gf128mul_table_be[a >> 56]; + + x->b = cpu_to_le64((a << 8) | (b >> 56)); + x->a = cpu_to_le64((b << 8) ^ _tt); +} + void gf128mul_lle(be128 *r, const be128 *b) { be128 p[8]; @@ -249,9 +274,48 @@ void gf128mul_bbe(be128 *r, const be128 *b) } EXPORT_SYMBOL(gf128mul_bbe); +void gf128mul_ble(be128 *r, const be128 *b) +{ + be128 p[8]; + int i; + + p[0] = *r; + for (i = 0; i < 7; ++i) + gf128mul_x_ble((be128 *)&p[i + 1], (be128 *)&p[i]); + + memset(r, 0, sizeof(*r)); + for (i = 0;;) { + u8 ch = ((u8 *)b)[15 - i]; + + if (ch & 0x80) + be128_xor(r, r, &p[7]); + if (ch & 0x40) + be128_xor(r, r, &p[6]); + if (ch & 0x20) + be128_xor(r, r, &p[5]); + if (ch & 0x10) + be128_xor(r, r, &p[4]); + if (ch & 0x08) + be128_xor(r, r, &p[3]); + if (ch & 0x04) + be128_xor(r, r, &p[2]); + if (ch & 0x02) + be128_xor(r, r, &p[1]); + if (ch & 0x01) + be128_xor(r, r, &p[0]); + + if (++i >= 16) + break; + + gf128mul_x8_ble(r); + } +} +EXPORT_SYMBOL(gf128mul_ble); + + /* This version uses 64k bytes of table space. A 16 byte buffer has to be multiplied by a 16 byte key - value in GF(128). If we consider a GF(128) value in + value in GF(2^128). If we consider a GF(2^128) value in the buffer's lowest byte, we can construct a table of the 256 16 byte values that result from the 256 values of this byte. This requires 4096 bytes. But we also @@ -352,8 +416,8 @@ void gf128mul_free_64k(struct gf128mul_64k *t) int i; for (i = 0; i < 16; i++) - kfree(t->t[i]); - kfree(t); + kzfree(t->t[i]); + kzfree(t); } EXPORT_SYMBOL(gf128mul_free_64k); @@ -385,7 +449,7 @@ EXPORT_SYMBOL(gf128mul_64k_bbe); /* This version uses 4k bytes of table space. A 16 byte buffer has to be multiplied by a 16 byte key - value in GF(128). If we consider a GF(128) value in a + value in GF(2^128). If we consider a GF(2^128) value in a single byte, we can construct a table of the 256 16 byte values that result from the 256 values of this byte. This requires 4096 bytes. If we take the highest byte in @@ -443,6 +507,28 @@ out: } EXPORT_SYMBOL(gf128mul_init_4k_bbe); +struct gf128mul_4k *gf128mul_init_4k_ble(const be128 *g) +{ + struct gf128mul_4k *t; + int j, k; + + t = kzalloc(sizeof(*t), GFP_KERNEL); + if (!t) + goto out; + + t->t[1] = *g; + for (j = 1; j <= 64; j <<= 1) + gf128mul_x_ble(&t->t[j + j], &t->t[j]); + + for (j = 2; j < 256; j += j) + for (k = 1; k < j; ++k) + be128_xor(&t->t[j + k], &t->t[j], &t->t[k]); + +out: + return t; +} +EXPORT_SYMBOL(gf128mul_init_4k_ble); + void gf128mul_4k_lle(be128 *a, struct gf128mul_4k *t) { u8 *ap = (u8 *)a; @@ -473,5 +559,20 @@ void gf128mul_4k_bbe(be128 *a, struct gf128mul_4k *t) } EXPORT_SYMBOL(gf128mul_4k_bbe); +void gf128mul_4k_ble(be128 *a, struct gf128mul_4k *t) +{ + u8 *ap = (u8 *)a; + be128 r[1]; + int i = 15; + + *r = t->t[ap[15]]; + while (i--) { + gf128mul_x8_ble(r); + be128_xor(r, r, &t->t[ap[i]]); + } + *a = *r; +} +EXPORT_SYMBOL(gf128mul_4k_ble); + MODULE_LICENSE("GPL"); MODULE_DESCRIPTION("Functions for multiplying elements of GF(2^128)"); diff --git a/crypto/heh.c b/crypto/heh.c new file mode 100644 index 000000000000..10c00aaf797e --- /dev/null +++ b/crypto/heh.c @@ -0,0 +1,1033 @@ +/* + * HEH: Hash-Encrypt-Hash mode + * + * Copyright (c) 2016 Google Inc. + * + * Authors: + * Alex Cope + * Eric Biggers + */ + +/* + * Hash-Encrypt-Hash (HEH) is a proposed block cipher mode of operation which + * extends the strong pseudo-random permutation (SPRP) property of block ciphers + * (e.g. AES) to arbitrary length input strings. It uses two keyed invertible + * hash functions with a layer of ECB encryption applied in-between. The + * algorithm is specified by the following Internet Draft: + * + * https://tools.ietf.org/html/draft-cope-heh-01 + * + * Although HEH can be used as either a regular symmetric cipher or as an AEAD, + * currently this module only provides it as a symmetric cipher. Additionally, + * only 16-byte nonces are supported. + */ + +#include +#include +#include +#include +#include +#include "internal.h" + +/* + * The block size is the size of GF(2^128) elements and also the required block + * size of the underlying block cipher. + */ +#define HEH_BLOCK_SIZE 16 + +struct heh_instance_ctx { + struct crypto_shash_spawn cmac; + struct crypto_shash_spawn poly_hash; + struct crypto_skcipher_spawn ecb; +}; + +struct heh_tfm_ctx { + struct crypto_shash *cmac; + struct crypto_shash *poly_hash; /* keyed with tau_key */ + struct crypto_ablkcipher *ecb; +}; + +struct heh_cmac_data { + u8 nonce[HEH_BLOCK_SIZE]; + __le32 nonce_length; + __le32 aad_length; + __le32 message_length; + __le32 padding; +}; + +struct heh_req_ctx { /* aligned to alignmask */ + be128 beta1_key; + be128 beta2_key; + union { + struct { + struct heh_cmac_data data; + struct shash_desc desc; + /* + crypto_shash_descsize(cmac) */ + } cmac; + struct { + struct shash_desc desc; + /* + crypto_shash_descsize(poly_hash) */ + } poly_hash; + struct { + u8 keystream[HEH_BLOCK_SIZE]; + u8 tmp[HEH_BLOCK_SIZE]; + struct scatterlist tmp_sgl[2]; + struct ablkcipher_request req; + /* + crypto_ablkcipher_reqsize(ecb) */ + } ecb; + } u; +}; + +/* + * Get the offset in bytes to the last full block, or equivalently the length of + * all full blocks excluding the last + */ +static inline unsigned int get_tail_offset(unsigned int len) +{ + len -= len % HEH_BLOCK_SIZE; + return len - HEH_BLOCK_SIZE; +} + +static inline struct heh_req_ctx *heh_req_ctx(struct ablkcipher_request *req) +{ + unsigned int alignmask = crypto_ablkcipher_alignmask( + crypto_ablkcipher_reqtfm(req)); + + return (void *)PTR_ALIGN((u8 *)ablkcipher_request_ctx(req), + alignmask + 1); +} + +static inline void async_done(struct crypto_async_request *areq, int err, + int (*next_step)(struct ablkcipher_request *, + u32)) +{ + struct ablkcipher_request *req = areq->data; + + if (err) + goto out; + + err = next_step(req, req->base.flags & ~CRYPTO_TFM_REQ_MAY_SLEEP); + if (err == -EINPROGRESS || + (err == -EBUSY && (req->base.flags & CRYPTO_TFM_REQ_MAY_BACKLOG))) + return; +out: + ablkcipher_request_complete(req, err); +} + +/* + * Generate the per-message "beta" keys used by the hashing layers of HEH. The + * first beta key is the CMAC of the nonce, the additional authenticated data + * (AAD), and the lengths in bytes of the nonce, AAD, and message. The nonce + * and AAD are each zero-padded to the next 16-byte block boundary, and the + * lengths are serialized as 4-byte little endian integers and zero-padded to + * the next 16-byte block boundary. + * The second beta key is the first one interpreted as an element in GF(2^128) + * and multiplied by x. + * + * Note that because the nonce and AAD may, in general, be variable-length, the + * key generation must be done by a pseudo-random function (PRF) on + * variable-length inputs. CBC-MAC does not satisfy this, as it is only a PRF + * on fixed-length inputs. CMAC remedies this flaw. Including the lengths of + * the nonce, AAD, and message is also critical to avoid collisions. + * + * That being said, this implementation does not yet operate as an AEAD and + * therefore there is never any AAD, nor are variable-length nonces supported. + */ +static int generate_betas(struct ablkcipher_request *req, + be128 *beta1_key, be128 *beta2_key) +{ + struct crypto_ablkcipher *tfm = crypto_ablkcipher_reqtfm(req); + struct heh_tfm_ctx *ctx = crypto_ablkcipher_ctx(tfm); + struct heh_req_ctx *rctx = heh_req_ctx(req); + struct heh_cmac_data *data = &rctx->u.cmac.data; + struct shash_desc *desc = &rctx->u.cmac.desc; + int err; + + BUILD_BUG_ON(sizeof(*data) != 2 * HEH_BLOCK_SIZE); + memcpy(data->nonce, req->info, HEH_BLOCK_SIZE); + data->nonce_length = cpu_to_le32(HEH_BLOCK_SIZE); + data->aad_length = cpu_to_le32(0); + data->message_length = cpu_to_le32(req->nbytes); + data->padding = cpu_to_le32(0); + + desc->tfm = ctx->cmac; + desc->flags = req->base.flags; + + err = crypto_shash_digest(desc, (const u8 *)data, sizeof(*data), + (u8 *)beta1_key); + if (err) + return err; + + gf128mul_x_ble(beta2_key, beta1_key); + return 0; +} + +/*****************************************************************************/ + +/* + * This is the generic version of poly_hash. It does the GF(2^128) + * multiplication by 'tau_key' using a precomputed table, without using any + * special CPU instructions. On some platforms, an accelerated version (with + * higher cra_priority) may be used instead. + */ + +struct poly_hash_tfm_ctx { + struct gf128mul_4k *tau_key; +}; + +struct poly_hash_desc_ctx { + be128 digest; + unsigned int count; +}; + +static int poly_hash_setkey(struct crypto_shash *tfm, + const u8 *key, unsigned int keylen) +{ + struct poly_hash_tfm_ctx *tctx = crypto_shash_ctx(tfm); + be128 key128; + + if (keylen != HEH_BLOCK_SIZE) { + crypto_shash_set_flags(tfm, CRYPTO_TFM_RES_BAD_KEY_LEN); + return -EINVAL; + } + + if (tctx->tau_key) + gf128mul_free_4k(tctx->tau_key); + memcpy(&key128, key, HEH_BLOCK_SIZE); + tctx->tau_key = gf128mul_init_4k_ble(&key128); + if (!tctx->tau_key) + return -ENOMEM; + return 0; +} + +static int poly_hash_init(struct shash_desc *desc) +{ + struct poly_hash_desc_ctx *ctx = shash_desc_ctx(desc); + + ctx->digest = (be128) { 0 }; + ctx->count = 0; + return 0; +} + +static int poly_hash_update(struct shash_desc *desc, const u8 *src, + unsigned int len) +{ + struct poly_hash_tfm_ctx *tctx = crypto_shash_ctx(desc->tfm); + struct poly_hash_desc_ctx *ctx = shash_desc_ctx(desc); + unsigned int partial = ctx->count % HEH_BLOCK_SIZE; + u8 *dst = (u8 *)&ctx->digest + partial; + + ctx->count += len; + + /* Finishing at least one block? */ + if (partial + len >= HEH_BLOCK_SIZE) { + + if (partial) { + /* Finish the pending block. */ + unsigned int n = HEH_BLOCK_SIZE - partial; + + len -= n; + do { + *dst++ ^= *src++; + } while (--n); + + gf128mul_4k_ble(&ctx->digest, tctx->tau_key); + } + + /* Process zero or more full blocks. */ + while (len >= HEH_BLOCK_SIZE) { + be128 coeff; + + memcpy(&coeff, src, HEH_BLOCK_SIZE); + be128_xor(&ctx->digest, &ctx->digest, &coeff); + src += HEH_BLOCK_SIZE; + len -= HEH_BLOCK_SIZE; + gf128mul_4k_ble(&ctx->digest, tctx->tau_key); + } + dst = (u8 *)&ctx->digest; + } + + /* Continue adding the next block to 'digest'. */ + while (len--) + *dst++ ^= *src++; + return 0; +} + +static int poly_hash_final(struct shash_desc *desc, u8 *out) +{ + struct poly_hash_desc_ctx *ctx = shash_desc_ctx(desc); + + /* Finish the last block if needed. */ + if (ctx->count % HEH_BLOCK_SIZE) { + struct poly_hash_tfm_ctx *tctx = crypto_shash_ctx(desc->tfm); + + gf128mul_4k_ble(&ctx->digest, tctx->tau_key); + } + + memcpy(out, &ctx->digest, HEH_BLOCK_SIZE); + return 0; +} + +static void poly_hash_exit(struct crypto_tfm *tfm) +{ + struct poly_hash_tfm_ctx *tctx = crypto_tfm_ctx(tfm); + + gf128mul_free_4k(tctx->tau_key); +} + +static struct shash_alg poly_hash_alg = { + .digestsize = HEH_BLOCK_SIZE, + .init = poly_hash_init, + .update = poly_hash_update, + .final = poly_hash_final, + .setkey = poly_hash_setkey, + .descsize = sizeof(struct poly_hash_desc_ctx), + .base = { + .cra_name = "poly_hash", + .cra_driver_name = "poly_hash-generic", + .cra_priority = 100, + .cra_ctxsize = sizeof(struct poly_hash_tfm_ctx), + .cra_exit = poly_hash_exit, + .cra_module = THIS_MODULE, + }, +}; + +/*****************************************************************************/ + +/* + * Split the message into 16 byte blocks, padding out the last block, and use + * the blocks as coefficients in the evaluation of a polynomial over GF(2^128) + * at the secret point 'tau_key'. For ease of implementing the higher-level + * heh_hash_inv() function, the constant and degree-1 coefficients are swapped + * if there is a partial block. + * + * Mathematically, compute: + * if (no partial block) + * k^{N-1} * m_0 + ... + k * m_{N-2} + m_{N-1} + * else if (partial block) + * k^N * m_0 + ... + k^2 * m_{N-2} + k * m_N + m_{N-1} + * + * where: + * t is tau_key + * N is the number of full blocks in the message + * m_i is the i-th full block in the message for i = 0 to N-1 inclusive + * m_N is the partial block of the message zero-padded up to 16 bytes + * + * Note that most of this is now separated out into its own keyed hash + * algorithm, to allow optimized implementations. However, we still handle the + * swapping of the last two coefficients here in the HEH template because this + * simplifies the poly_hash algorithms: they don't have to buffer an extra + * block, don't have to duplicate as much code, and are more similar to GHASH. + */ +static int poly_hash(struct ablkcipher_request *req, struct scatterlist *sgl, + be128 *hash) +{ + struct heh_req_ctx *rctx = heh_req_ctx(req); + struct crypto_ablkcipher *tfm = crypto_ablkcipher_reqtfm(req); + struct heh_tfm_ctx *ctx = crypto_ablkcipher_ctx(tfm); + struct shash_desc *desc = &rctx->u.poly_hash.desc; + unsigned int tail_offset = get_tail_offset(req->nbytes); + unsigned int tail_len = req->nbytes - tail_offset; + be128 tail[2]; + unsigned int i, n; + struct sg_mapping_iter miter; + int err; + + desc->tfm = ctx->poly_hash; + desc->flags = req->base.flags; + + /* Handle all full blocks except the last */ + err = crypto_shash_init(desc); + sg_miter_start(&miter, sgl, sg_nents(sgl), + SG_MITER_FROM_SG | SG_MITER_ATOMIC); + for (i = 0; i < tail_offset && !err; i += n) { + sg_miter_next(&miter); + n = min_t(unsigned int, miter.length, tail_offset - i); + err = crypto_shash_update(desc, miter.addr, n); + } + sg_miter_stop(&miter); + if (err) + return err; + + /* Handle the last full block and the partial block */ + scatterwalk_map_and_copy(tail, sgl, tail_offset, tail_len, 0); + + if (tail_len != HEH_BLOCK_SIZE) { + /* handle the partial block */ + memset((u8 *)tail + tail_len, 0, sizeof(tail) - tail_len); + err = crypto_shash_update(desc, (u8 *)&tail[1], HEH_BLOCK_SIZE); + if (err) + return err; + } + err = crypto_shash_final(desc, (u8 *)hash); + if (err) + return err; + be128_xor(hash, hash, &tail[0]); + return 0; +} + +/* + * Transform all full blocks except the last. + * This is used by both the hash and inverse hash phases. + */ +static int heh_tfm_blocks(struct ablkcipher_request *req, + struct scatterlist *src_sgl, + struct scatterlist *dst_sgl, unsigned int len, + const be128 *hash, const be128 *beta_key) +{ + struct crypto_ablkcipher *tfm = crypto_ablkcipher_reqtfm(req); + struct blkcipher_desc desc = { .flags = req->base.flags }; + struct blkcipher_walk walk; + be128 e = *beta_key; + int err; + unsigned int nbytes; + + blkcipher_walk_init(&walk, dst_sgl, src_sgl, len); + + err = blkcipher_ablkcipher_walk_virt(&desc, &walk, tfm); + + while ((nbytes = walk.nbytes)) { + const be128 *src = (be128 *)walk.src.virt.addr; + be128 *dst = (be128 *)walk.dst.virt.addr; + + do { + gf128mul_x_ble(&e, &e); + be128_xor(dst, src, hash); + be128_xor(dst, dst, &e); + src++; + dst++; + } while ((nbytes -= HEH_BLOCK_SIZE) >= HEH_BLOCK_SIZE); + err = blkcipher_walk_done(&desc, &walk, nbytes); + } + return err; +} + +/* + * The hash phase of HEH. Given a message, compute: + * + * (m_0 + H, ..., m_{N-2} + H, H, m_N) + (xb, x^2b, ..., x^{N-1}b, b, 0) + * + * where: + * N is the number of full blocks in the message + * m_i is the i-th full block in the message for i = 0 to N-1 inclusive + * m_N is the unpadded partial block, possibly empty + * H is the poly_hash() of the message, keyed by tau_key + * b is beta_key + * x is the element x in our representation of GF(2^128) + * + * Note that the partial block remains unchanged, but it does affect the result + * of poly_hash() and therefore the transformation of all the full blocks. + */ +static int heh_hash(struct ablkcipher_request *req, const be128 *beta_key) +{ + be128 hash; + unsigned int tail_offset = get_tail_offset(req->nbytes); + unsigned int partial_len = req->nbytes % HEH_BLOCK_SIZE; + int err; + + /* poly_hash() the full message including the partial block */ + err = poly_hash(req, req->src, &hash); + if (err) + return err; + + /* Transform all full blocks except the last */ + err = heh_tfm_blocks(req, req->src, req->dst, tail_offset, &hash, + beta_key); + if (err) + return err; + + /* Set the last full block to hash XOR beta_key */ + be128_xor(&hash, &hash, beta_key); + scatterwalk_map_and_copy(&hash, req->dst, tail_offset, HEH_BLOCK_SIZE, + 1); + + /* Copy the partial block if needed */ + if (partial_len != 0 && req->src != req->dst) { + unsigned int offs = tail_offset + HEH_BLOCK_SIZE; + + scatterwalk_map_and_copy(&hash, req->src, offs, partial_len, 0); + scatterwalk_map_and_copy(&hash, req->dst, offs, partial_len, 1); + } + return 0; +} + +/* + * The inverse hash phase of HEH. This undoes the result of heh_hash(). + */ +static int heh_hash_inv(struct ablkcipher_request *req, const be128 *beta_key) +{ + be128 hash; + be128 tmp; + struct scatterlist tmp_sgl[2]; + struct scatterlist *tail_sgl; + unsigned int tail_offset = get_tail_offset(req->nbytes); + struct scatterlist *sgl = req->dst; + int err; + + /* + * The last full block was computed as hash XOR beta_key, so XOR it with + * beta_key to recover hash. + */ + tail_sgl = scatterwalk_ffwd(tmp_sgl, sgl, tail_offset); + scatterwalk_map_and_copy(&hash, tail_sgl, 0, HEH_BLOCK_SIZE, 0); + be128_xor(&hash, &hash, beta_key); + + /* Transform all full blocks except the last */ + err = heh_tfm_blocks(req, sgl, sgl, tail_offset, &hash, beta_key); + if (err) + return err; + + /* + * Recover the last full block. We know 'hash', i.e. the poly_hash() of + * the the original message. The last full block was the constant term + * of the polynomial. To recover the last full block, temporarily zero + * it, compute the poly_hash(), and take the difference from 'hash'. + */ + memset(&tmp, 0, sizeof(tmp)); + scatterwalk_map_and_copy(&tmp, tail_sgl, 0, HEH_BLOCK_SIZE, 1); + err = poly_hash(req, sgl, &tmp); + if (err) + return err; + be128_xor(&tmp, &tmp, &hash); + scatterwalk_map_and_copy(&tmp, tail_sgl, 0, HEH_BLOCK_SIZE, 1); + return 0; +} + +static int heh_hash_inv_step(struct ablkcipher_request *req, u32 flags) +{ + struct heh_req_ctx *rctx = heh_req_ctx(req); + + return heh_hash_inv(req, &rctx->beta2_key); +} + +static int heh_ecb_step_3(struct ablkcipher_request *req, u32 flags) +{ + struct heh_req_ctx *rctx = heh_req_ctx(req); + u8 partial_block[HEH_BLOCK_SIZE] __aligned(__alignof__(u32)); + unsigned int tail_offset = get_tail_offset(req->nbytes); + unsigned int partial_offset = tail_offset + HEH_BLOCK_SIZE; + unsigned int partial_len = req->nbytes - partial_offset; + + /* + * Extract the pad in req->dst at tail_offset, and xor the partial block + * with it to create encrypted partial block + */ + scatterwalk_map_and_copy(rctx->u.ecb.keystream, req->dst, tail_offset, + HEH_BLOCK_SIZE, 0); + scatterwalk_map_and_copy(partial_block, req->dst, partial_offset, + partial_len, 0); + crypto_xor(partial_block, rctx->u.ecb.keystream, partial_len); + + /* + * Store the encrypted final block and partial block back in dst_sg + */ + scatterwalk_map_and_copy(&rctx->u.ecb.tmp, req->dst, tail_offset, + HEH_BLOCK_SIZE, 1); + scatterwalk_map_and_copy(partial_block, req->dst, partial_offset, + partial_len, 1); + + return heh_hash_inv_step(req, flags); +} + +static void heh_ecb_step_2_done(struct crypto_async_request *areq, int err) +{ + return async_done(areq, err, heh_ecb_step_3); +} + +static int heh_ecb_step_2(struct ablkcipher_request *req, u32 flags) +{ + struct heh_req_ctx *rctx = heh_req_ctx(req); + unsigned int partial_len = req->nbytes % HEH_BLOCK_SIZE; + struct scatterlist *tmp_sgl; + int err; + unsigned int tail_offset = get_tail_offset(req->nbytes); + + if (partial_len == 0) + return heh_hash_inv_step(req, flags); + + /* + * Extract the final full block, store it in tmp, and then xor that with + * the value saved in u.ecb.keystream + */ + scatterwalk_map_and_copy(rctx->u.ecb.tmp, req->dst, tail_offset, + HEH_BLOCK_SIZE, 0); + crypto_xor(rctx->u.ecb.keystream, rctx->u.ecb.tmp, HEH_BLOCK_SIZE); + + /* + * Encrypt the value in rctx->u.ecb.keystream to create the pad for the + * partial block. + * We cannot encrypt stack buffers, so re-use the dst_sg to do this + * encryption to avoid a malloc. The value at tail_offset is stored in + * tmp, and will be restored later. + */ + scatterwalk_map_and_copy(rctx->u.ecb.keystream, req->dst, tail_offset, + HEH_BLOCK_SIZE, 1); + tmp_sgl = scatterwalk_ffwd(rctx->u.ecb.tmp_sgl, req->dst, tail_offset); + ablkcipher_request_set_callback(&rctx->u.ecb.req, flags, + heh_ecb_step_2_done, req); + ablkcipher_request_set_crypt(&rctx->u.ecb.req, tmp_sgl, tmp_sgl, + HEH_BLOCK_SIZE, NULL); + err = crypto_ablkcipher_encrypt(&rctx->u.ecb.req); + if (err) + return err; + return heh_ecb_step_3(req, flags); +} + +static void heh_ecb_full_done(struct crypto_async_request *areq, int err) +{ + return async_done(areq, err, heh_ecb_step_2); +} + +/* + * The encrypt phase of HEH. This uses ECB encryption, with special handling + * for the partial block at the end if any. The source data is already in + * req->dst, so the encryption happens in-place. + * + * After the encrypt phase we continue on to the inverse hash phase. The + * functions calls are chained to support asynchronous ECB algorithms. + */ +static int heh_ecb(struct ablkcipher_request *req, bool decrypt) +{ + struct crypto_ablkcipher *tfm = crypto_ablkcipher_reqtfm(req); + struct heh_tfm_ctx *ctx = crypto_ablkcipher_ctx(tfm); + struct heh_req_ctx *rctx = heh_req_ctx(req); + struct ablkcipher_request *ecb_req = &rctx->u.ecb.req; + unsigned int tail_offset = get_tail_offset(req->nbytes); + unsigned int full_len = tail_offset + HEH_BLOCK_SIZE; + int err; + + /* + * Save the last full block before it is encrypted/decrypted. This will + * be used later to encrypt/decrypt the partial block + */ + scatterwalk_map_and_copy(rctx->u.ecb.keystream, req->dst, tail_offset, + HEH_BLOCK_SIZE, 0); + + /* Encrypt/decrypt all full blocks */ + ablkcipher_request_set_tfm(ecb_req, ctx->ecb); + ablkcipher_request_set_callback(ecb_req, req->base.flags, + heh_ecb_full_done, req); + ablkcipher_request_set_crypt(ecb_req, req->dst, req->dst, full_len, + NULL); + if (decrypt) + err = crypto_ablkcipher_decrypt(ecb_req); + else + err = crypto_ablkcipher_encrypt(ecb_req); + if (err) + return err; + + return heh_ecb_step_2(req, req->base.flags); +} + +static int heh_crypt(struct ablkcipher_request *req, bool decrypt) +{ + struct heh_req_ctx *rctx = heh_req_ctx(req); + int err; + + /* Inputs must be at least one full block */ + if (req->nbytes < HEH_BLOCK_SIZE) + return -EINVAL; + + err = generate_betas(req, &rctx->beta1_key, &rctx->beta2_key); + if (err) + return err; + + if (decrypt) + swap(rctx->beta1_key, rctx->beta2_key); + + err = heh_hash(req, &rctx->beta1_key); + if (err) + return err; + + return heh_ecb(req, decrypt); +} + +static int heh_encrypt(struct ablkcipher_request *req) +{ + return heh_crypt(req, false); +} + +static int heh_decrypt(struct ablkcipher_request *req) +{ + return heh_crypt(req, true); +} + +static int heh_setkey(struct crypto_ablkcipher *parent, const u8 *key, + unsigned int keylen) +{ + struct heh_tfm_ctx *ctx = crypto_ablkcipher_ctx(parent); + struct crypto_shash *cmac = ctx->cmac; + struct crypto_ablkcipher *ecb = ctx->ecb; + SHASH_DESC_ON_STACK(desc, cmac); + u8 *derived_keys; + u8 digest[HEH_BLOCK_SIZE]; + unsigned int i; + int err; + + /* set prf_key = key */ + crypto_shash_clear_flags(cmac, CRYPTO_TFM_REQ_MASK); + crypto_shash_set_flags(cmac, crypto_ablkcipher_get_flags(parent) & + CRYPTO_TFM_REQ_MASK); + err = crypto_shash_setkey(cmac, key, keylen); + crypto_ablkcipher_set_flags(parent, crypto_shash_get_flags(cmac) & + CRYPTO_TFM_RES_MASK); + if (err) + return err; + + /* + * Generate tau_key and ecb_key as follows: + * tau_key = cmac(prf_key, 0x00...01) + * ecb_key = cmac(prf_key, 0x00...02) || cmac(prf_key, 0x00...03) || ... + * truncated to keylen bytes + */ + derived_keys = kzalloc(round_up(HEH_BLOCK_SIZE + keylen, + HEH_BLOCK_SIZE), GFP_KERNEL); + if (!derived_keys) + return -ENOMEM; + desc->tfm = cmac; + desc->flags = (crypto_shash_get_flags(cmac) & CRYPTO_TFM_REQ_MASK); + for (i = 0; i < keylen + HEH_BLOCK_SIZE; i += HEH_BLOCK_SIZE) { + derived_keys[i + HEH_BLOCK_SIZE - 1] = + 0x01 + i / HEH_BLOCK_SIZE; + err = crypto_shash_digest(desc, derived_keys + i, + HEH_BLOCK_SIZE, digest); + if (err) + goto out; + memcpy(derived_keys + i, digest, HEH_BLOCK_SIZE); + } + + err = crypto_shash_setkey(ctx->poly_hash, derived_keys, HEH_BLOCK_SIZE); + if (err) + goto out; + + crypto_ablkcipher_clear_flags(ecb, CRYPTO_TFM_REQ_MASK); + crypto_ablkcipher_set_flags(ecb, crypto_ablkcipher_get_flags(parent) & + CRYPTO_TFM_REQ_MASK); + err = crypto_ablkcipher_setkey(ecb, derived_keys + HEH_BLOCK_SIZE, + keylen); + crypto_ablkcipher_set_flags(parent, crypto_ablkcipher_get_flags(ecb) & + CRYPTO_TFM_RES_MASK); +out: + kzfree(derived_keys); + return err; +} + +static int heh_init_tfm(struct crypto_tfm *tfm) +{ + struct crypto_instance *inst = crypto_tfm_alg_instance(tfm); + struct heh_instance_ctx *ictx = crypto_instance_ctx(inst); + struct heh_tfm_ctx *ctx = crypto_tfm_ctx(tfm); + struct crypto_shash *cmac; + struct crypto_shash *poly_hash; + struct crypto_ablkcipher *ecb; + unsigned int reqsize; + int err; + + cmac = crypto_spawn_shash(&ictx->cmac); + if (IS_ERR(cmac)) + return PTR_ERR(cmac); + + poly_hash = crypto_spawn_shash(&ictx->poly_hash); + err = PTR_ERR(poly_hash); + if (IS_ERR(poly_hash)) + goto err_free_cmac; + + ecb = crypto_spawn_skcipher(&ictx->ecb); + err = PTR_ERR(ecb); + if (IS_ERR(ecb)) + goto err_free_poly_hash; + + ctx->cmac = cmac; + ctx->poly_hash = poly_hash; + ctx->ecb = ecb; + + reqsize = crypto_tfm_alg_alignmask(tfm) & + ~(crypto_tfm_ctx_alignment() - 1); + reqsize += max3(offsetof(struct heh_req_ctx, u.cmac.desc) + + sizeof(struct shash_desc) + + crypto_shash_descsize(cmac), + offsetof(struct heh_req_ctx, u.poly_hash.desc) + + sizeof(struct shash_desc) + + crypto_shash_descsize(poly_hash), + offsetof(struct heh_req_ctx, u.ecb.req) + + sizeof(struct ablkcipher_request) + + crypto_ablkcipher_reqsize(ecb)); + tfm->crt_ablkcipher.reqsize = reqsize; + + return 0; + +err_free_poly_hash: + crypto_free_shash(poly_hash); +err_free_cmac: + crypto_free_shash(cmac); + return err; +} + +static void heh_exit_tfm(struct crypto_tfm *tfm) +{ + struct heh_tfm_ctx *ctx = crypto_tfm_ctx(tfm); + + crypto_free_shash(ctx->cmac); + crypto_free_shash(ctx->poly_hash); + crypto_free_ablkcipher(ctx->ecb); +} + +static void heh_free_instance(struct crypto_instance *inst) +{ + struct heh_instance_ctx *ctx = crypto_instance_ctx(inst); + + crypto_drop_shash(&ctx->cmac); + crypto_drop_shash(&ctx->poly_hash); + crypto_drop_skcipher(&ctx->ecb); + kfree(inst); +} + +/* + * Create an instance of HEH as a ablkcipher. + * + * This relies on underlying CMAC and ECB algorithms, usually cmac(aes) and + * ecb(aes). For performance reasons we support asynchronous ECB algorithms. + * However, we do not yet support asynchronous CMAC algorithms because CMAC is + * only used on a small fixed amount of data per request, independent of the + * request length. This would change if AEAD or variable-length nonce support + * were to be exposed. + */ +static int heh_create_common(struct crypto_template *tmpl, struct rtattr **tb, + const char *full_name, const char *cmac_name, + const char *poly_hash_name, const char *ecb_name) +{ + struct crypto_attr_type *algt; + struct crypto_instance *inst; + struct heh_instance_ctx *ctx; + struct shash_alg *cmac; + struct shash_alg *poly_hash; + struct crypto_alg *ecb; + int err; + + algt = crypto_get_attr_type(tb); + if (IS_ERR(algt)) + return PTR_ERR(algt); + + /* User must be asking for something compatible with ablkcipher */ + if ((algt->type ^ CRYPTO_ALG_TYPE_ABLKCIPHER) & algt->mask) + return -EINVAL; + + /* Allocate the ablkcipher instance */ + inst = kzalloc(sizeof(*inst) + sizeof(*ctx), GFP_KERNEL); + if (!inst) + return -ENOMEM; + + ctx = crypto_instance_ctx(inst); + + /* Set up the cmac spawn */ + ctx->cmac.base.inst = inst; + err = crypto_grab_shash(&ctx->cmac, cmac_name, 0, 0); + if (err) + goto err_free_inst; + cmac = crypto_spawn_shash_alg(&ctx->cmac); + err = -EINVAL; + if (cmac->digestsize != HEH_BLOCK_SIZE) + goto err_drop_cmac; + + /* Set up the poly_hash spawn */ + ctx->poly_hash.base.inst = inst; + err = crypto_grab_shash(&ctx->poly_hash, poly_hash_name, 0, 0); + if (err) + goto err_drop_cmac; + poly_hash = crypto_spawn_shash_alg(&ctx->poly_hash); + err = -EINVAL; + if (poly_hash->digestsize != HEH_BLOCK_SIZE) + goto err_drop_poly_hash; + + /* Set up the ecb spawn */ + ctx->ecb.base.inst = inst; + err = crypto_grab_skcipher(&ctx->ecb, ecb_name, 0, + crypto_requires_sync(algt->type, + algt->mask)); + if (err) + goto err_drop_poly_hash; + ecb = crypto_skcipher_spawn_alg(&ctx->ecb); + + /* HEH only supports block ciphers with 16 byte block size */ + err = -EINVAL; + if (ecb->cra_blocksize != HEH_BLOCK_SIZE) + goto err_drop_ecb; + + /* The underlying "ECB" algorithm must not require an IV */ + err = -EINVAL; + if ((ecb->cra_flags & CRYPTO_ALG_TYPE_MASK) == CRYPTO_ALG_TYPE_BLKCIPHER) { + if (ecb->cra_blkcipher.ivsize != 0) + goto err_drop_ecb; + } else { + if (ecb->cra_ablkcipher.ivsize != 0) + goto err_drop_ecb; + } + + /* Set the instance names */ + err = -ENAMETOOLONG; + if (snprintf(inst->alg.cra_driver_name, CRYPTO_MAX_ALG_NAME, + "heh_base(%s,%s,%s)", cmac->base.cra_driver_name, + poly_hash->base.cra_driver_name, + ecb->cra_driver_name) >= CRYPTO_MAX_ALG_NAME) + goto err_drop_ecb; + + err = -ENAMETOOLONG; + if (snprintf(inst->alg.cra_name, CRYPTO_MAX_ALG_NAME, + "%s", full_name) >= CRYPTO_MAX_ALG_NAME) + goto err_drop_ecb; + + /* Finish initializing the instance */ + + inst->alg.cra_flags = CRYPTO_ALG_TYPE_ABLKCIPHER | + (ecb->cra_flags & CRYPTO_ALG_ASYNC); + inst->alg.cra_blocksize = HEH_BLOCK_SIZE; + inst->alg.cra_ctxsize = sizeof(struct heh_tfm_ctx); + inst->alg.cra_alignmask = ecb->cra_alignmask | (__alignof__(be128) - 1); + inst->alg.cra_priority = ecb->cra_priority; + inst->alg.cra_type = &crypto_ablkcipher_type; + inst->alg.cra_init = heh_init_tfm; + inst->alg.cra_exit = heh_exit_tfm; + + inst->alg.cra_ablkcipher.setkey = heh_setkey; + inst->alg.cra_ablkcipher.encrypt = heh_encrypt; + inst->alg.cra_ablkcipher.decrypt = heh_decrypt; + if ((ecb->cra_flags & CRYPTO_ALG_TYPE_MASK) == CRYPTO_ALG_TYPE_BLKCIPHER) { + inst->alg.cra_ablkcipher.min_keysize = ecb->cra_blkcipher.min_keysize; + inst->alg.cra_ablkcipher.max_keysize = ecb->cra_blkcipher.max_keysize; + } else { + inst->alg.cra_ablkcipher.min_keysize = ecb->cra_ablkcipher.min_keysize; + inst->alg.cra_ablkcipher.max_keysize = ecb->cra_ablkcipher.max_keysize; + } + inst->alg.cra_ablkcipher.ivsize = HEH_BLOCK_SIZE; + + /* Register the instance */ + err = crypto_register_instance(tmpl, inst); + if (err) + goto err_drop_ecb; + return 0; + +err_drop_ecb: + crypto_drop_skcipher(&ctx->ecb); +err_drop_poly_hash: + crypto_drop_shash(&ctx->poly_hash); +err_drop_cmac: + crypto_drop_shash(&ctx->cmac); +err_free_inst: + kfree(inst); + return err; +} + +static int heh_create(struct crypto_template *tmpl, struct rtattr **tb) +{ + const char *cipher_name; + char full_name[CRYPTO_MAX_ALG_NAME]; + char cmac_name[CRYPTO_MAX_ALG_NAME]; + char ecb_name[CRYPTO_MAX_ALG_NAME]; + + /* Get the name of the requested block cipher (e.g. aes) */ + cipher_name = crypto_attr_alg_name(tb[1]); + if (IS_ERR(cipher_name)) + return PTR_ERR(cipher_name); + + if (snprintf(full_name, CRYPTO_MAX_ALG_NAME, "heh(%s)", cipher_name) >= + CRYPTO_MAX_ALG_NAME) + return -ENAMETOOLONG; + + if (snprintf(cmac_name, CRYPTO_MAX_ALG_NAME, "cmac(%s)", cipher_name) >= + CRYPTO_MAX_ALG_NAME) + return -ENAMETOOLONG; + + if (snprintf(ecb_name, CRYPTO_MAX_ALG_NAME, "ecb(%s)", cipher_name) >= + CRYPTO_MAX_ALG_NAME) + return -ENAMETOOLONG; + + return heh_create_common(tmpl, tb, full_name, cmac_name, "poly_hash", + ecb_name); +} + +static struct crypto_template heh_tmpl = { + .name = "heh", + .create = heh_create, + .free = heh_free_instance, + .module = THIS_MODULE, +}; + +static int heh_base_create(struct crypto_template *tmpl, struct rtattr **tb) +{ + char full_name[CRYPTO_MAX_ALG_NAME]; + const char *cmac_name; + const char *poly_hash_name; + const char *ecb_name; + + cmac_name = crypto_attr_alg_name(tb[1]); + if (IS_ERR(cmac_name)) + return PTR_ERR(cmac_name); + + poly_hash_name = crypto_attr_alg_name(tb[2]); + if (IS_ERR(poly_hash_name)) + return PTR_ERR(poly_hash_name); + + ecb_name = crypto_attr_alg_name(tb[3]); + if (IS_ERR(ecb_name)) + return PTR_ERR(ecb_name); + + if (snprintf(full_name, CRYPTO_MAX_ALG_NAME, "heh_base(%s,%s,%s)", + cmac_name, poly_hash_name, ecb_name) >= + CRYPTO_MAX_ALG_NAME) + return -ENAMETOOLONG; + + return heh_create_common(tmpl, tb, full_name, cmac_name, poly_hash_name, + ecb_name); +} + +/* + * If HEH is instantiated as "heh_base" instead of "heh", then specific + * implementations of cmac, poly_hash, and ecb can be specified instead of just + * the cipher. + */ +static struct crypto_template heh_base_tmpl = { + .name = "heh_base", + .create = heh_base_create, + .free = heh_free_instance, + .module = THIS_MODULE, +}; + +static int __init heh_module_init(void) +{ + int err; + + err = crypto_register_template(&heh_tmpl); + if (err) + return err; + + err = crypto_register_template(&heh_base_tmpl); + if (err) + goto out_undo_heh; + + err = crypto_register_shash(&poly_hash_alg); + if (err) + goto out_undo_heh_base; + + return 0; + +out_undo_heh_base: + crypto_unregister_template(&heh_base_tmpl); +out_undo_heh: + crypto_unregister_template(&heh_tmpl); + return err; +} + +static void __exit heh_module_exit(void) +{ + crypto_unregister_template(&heh_tmpl); + crypto_unregister_template(&heh_base_tmpl); + crypto_unregister_shash(&poly_hash_alg); +} + +module_init(heh_module_init); +module_exit(heh_module_exit); + +MODULE_LICENSE("GPL"); +MODULE_DESCRIPTION("Hash-Encrypt-Hash block cipher mode"); +MODULE_ALIAS_CRYPTO("heh"); +MODULE_ALIAS_CRYPTO("heh_base"); diff --git a/crypto/shash.c b/crypto/shash.c index 359754591653..9ae1e891308d 100644 --- a/crypto/shash.c +++ b/crypto/shash.c @@ -683,6 +683,14 @@ void shash_free_instance(struct crypto_instance *inst) } EXPORT_SYMBOL_GPL(shash_free_instance); +int crypto_grab_shash(struct crypto_shash_spawn *spawn, + const char *name, u32 type, u32 mask) +{ + spawn->base.frontend = &crypto_shash_type; + return crypto_grab_spawn(&spawn->base, name, type, mask); +} +EXPORT_SYMBOL_GPL(crypto_grab_shash); + int crypto_init_shash_spawn(struct crypto_shash_spawn *spawn, struct shash_alg *alg, struct crypto_instance *inst) diff --git a/crypto/testmgr.c b/crypto/testmgr.c index d4944318ca1f..8374ca8b6579 100644 --- a/crypto/testmgr.c +++ b/crypto/testmgr.c @@ -3213,6 +3213,21 @@ static const struct alg_test_desc alg_test_descs[] = { .count = GHASH_TEST_VECTORS } } + }, { + .alg = "heh(aes)", + .test = alg_test_skcipher, + .suite = { + .cipher = { + .enc = { + .vecs = aes_heh_enc_tv_template, + .count = AES_HEH_ENC_TEST_VECTORS + }, + .dec = { + .vecs = aes_heh_dec_tv_template, + .count = AES_HEH_DEC_TEST_VECTORS + } + } + } }, { .alg = "hmac(crc32)", .test = alg_test_hash, diff --git a/crypto/testmgr.h b/crypto/testmgr.h index da0a8fd765f4..97a523993bd8 100644 --- a/crypto/testmgr.h +++ b/crypto/testmgr.h @@ -14139,6 +14139,8 @@ static struct cipher_testvec cast6_xts_dec_tv_template[] = { #define AES_DEC_TEST_VECTORS 4 #define AES_CBC_ENC_TEST_VECTORS 5 #define AES_CBC_DEC_TEST_VECTORS 5 +#define AES_HEH_ENC_TEST_VECTORS 4 +#define AES_HEH_DEC_TEST_VECTORS 4 #define HMAC_MD5_ECB_CIPHER_NULL_ENC_TEST_VECTORS 2 #define HMAC_MD5_ECB_CIPHER_NULL_DEC_TEST_VECTORS 2 #define HMAC_SHA1_ECB_CIPHER_NULL_ENC_TEST_VEC 2 @@ -14511,6 +14513,198 @@ static struct cipher_testvec aes_dec_tv_template[] = { }, }; +static struct cipher_testvec aes_heh_enc_tv_template[] = { + { + .key = "\x00\x01\x02\x03\x04\x05\x06\x07" + "\x08\x09\x0A\x0B\x0C\x0D\x0E\x0F", + .klen = 16, + .iv = "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00", + .input = "\x00\x01\x02\x03\x04\x05\x06\x07" + "\x08\x09\x0A\x0B\x0C\x0D\x0E\x0F", + .ilen = 16, + .result = "\xd8\xbd\x40\xbf\xca\xe5\xee\x81" + "\x0f\x3d\x1f\x1f\xae\x89\x07\x55", + .rlen = 16, + .also_non_np = 1, + .np = 2, + .tap = { 8, 8 }, + }, { + .key = "\xa8\xda\x24\x9b\x5e\xfa\x13\xc2" + "\xc1\x94\xbf\x32\xba\x38\xa3\x77", + .klen = 16, + .iv = "\x4d\x47\x61\x37\x2b\x47\x86\xf0" + "\xd6\x47\xb5\xc2\xe8\xcf\x85\x27", + .input = "\xb8\xee\x29\xe4\xa5\xd1\xe7\x55" + "\xd0\xfd\xe7\x22\x63\x76\x36\xe2" + "\xf8\x0c\xf8\xfe\x65\x76\xe7\xca" + "\xc1\x42\xf5\xca\x5a\xa8\xac\x2a", + .ilen = 32, + .result = "\x59\xf2\x78\x4e\x10\x94\xf9\x5c" + "\x22\x23\x78\x2a\x30\x48\x11\x97" + "\xb1\xfe\x70\xc4\xef\xdf\x04\xef" + "\x16\x39\x04\xcf\xc0\x95\x9a\x98", + .rlen = 32, + .also_non_np = 1, + .np = 3, + .tap = { 16, 13, 3 }, + }, { + .key = "\x00\x01\x02\x03\x04\x05\x06\x07" + "\x08\x09\x0A\x0B\x0C\x0D\x0E\x0F", + .klen = 16, + .iv = "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00", + .input = "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00", + .ilen = 63, + .result = "\xe0\x40\xeb\xe9\x52\xbe\x65\x60" + "\xe4\x68\x68\xa3\x73\x75\xb8\x52" + "\xef\x38\x6a\x87\x25\x25\xf6\x04" + "\xe5\x8e\xbe\x14\x8b\x02\x14\x1f" + "\xa9\x73\xb7\xad\x15\xbe\x9c\xa0" + "\xd2\x8a\x2c\xdc\xd4\xe3\x05\x55" + "\x0a\xf5\xf8\x51\xee\xe5\x62\xa5" + "\x71\xa7\x7c\x15\x5d\x7a\x9e", + .rlen = 63, + .also_non_np = 1, + .np = 8, + .tap = { 20, 20, 10, 8, 2, 1, 1, 1 }, + }, { + .key = "\x00\x01\x02\x03\x04\x05\x06\x07" + "\x08\x09\x0A\x0B\x0C\x0D\x0E\x0F" + "\x00\x01\x02\x03\x04\x05\x06\x07" + "\x08\x09\x0A\x0B\x0C\x0D\x0E\x0F" + "\x00\x01\x02\x03\x04\x05\x06\x07" + "\x08\x09\x0A\x0B\x0C\x0D\x0E\x0F", + .klen = 16, + .iv = "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00", + .input = "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x01" + "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00", + .ilen = 63, + .result = "\x4b\x1a\x15\xa0\xaf\x08\x6d\x70" + "\xf0\xa7\x97\xb5\x31\x4b\x8c\xc3" + "\x4d\xf2\x7a\x9d\xdd\xd4\x15\x99" + "\x57\xad\xc6\xb1\x35\x69\xf5\x6a" + "\x2d\x70\xe4\x97\x49\xb2\x9f\x71" + "\xde\x22\xb5\x70\x8c\x69\x24\xd3" + "\xad\x80\x58\x48\x90\xe4\xed\xba" + "\x76\x3d\x71\x7c\x57\x25\x87", + .rlen = 63, + .also_non_np = 1, + .np = 8, + .tap = { 20, 20, 10, 8, 2, 1, 1, 1 }, + } +}; + +static struct cipher_testvec aes_heh_dec_tv_template[] = { + { + .key = "\x00\x01\x02\x03\x04\x05\x06\x07" + "\x08\x09\x0A\x0B\x0C\x0D\x0E\x0F", + .klen = 16, + .iv = "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00", + .input = "\xd8\xbd\x40\xbf\xca\xe5\xee\x81" + "\x0f\x3d\x1f\x1f\xae\x89\x07\x55", + .ilen = 16, + .result = "\x00\x01\x02\x03\x04\x05\x06\x07" + "\x08\x09\x0A\x0B\x0C\x0D\x0E\x0F", + .rlen = 16, + .also_non_np = 1, + .np = 2, + .tap = { 8, 8 }, + }, { + .key = "\xa8\xda\x24\x9b\x5e\xfa\x13\xc2" + "\xc1\x94\xbf\x32\xba\x38\xa3\x77", + .klen = 16, + .iv = "\x4d\x47\x61\x37\x2b\x47\x86\xf0" + "\xd6\x47\xb5\xc2\xe8\xcf\x85\x27", + .input = "\x59\xf2\x78\x4e\x10\x94\xf9\x5c" + "\x22\x23\x78\x2a\x30\x48\x11\x97" + "\xb1\xfe\x70\xc4\xef\xdf\x04\xef" + "\x16\x39\x04\xcf\xc0\x95\x9a\x98", + .ilen = 32, + .result = "\xb8\xee\x29\xe4\xa5\xd1\xe7\x55" + "\xd0\xfd\xe7\x22\x63\x76\x36\xe2" + "\xf8\x0c\xf8\xfe\x65\x76\xe7\xca" + "\xc1\x42\xf5\xca\x5a\xa8\xac\x2a", + .rlen = 32, + .also_non_np = 1, + .np = 3, + .tap = { 16, 13, 3 }, + }, { + .key = "\x00\x01\x02\x03\x04\x05\x06\x07" + "\x08\x09\x0A\x0B\x0C\x0D\x0E\x0F", + .klen = 16, + .iv = "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00", + .input = "\xe0\x40\xeb\xe9\x52\xbe\x65\x60" + "\xe4\x68\x68\xa3\x73\x75\xb8\x52" + "\xef\x38\x6a\x87\x25\x25\xf6\x04" + "\xe5\x8e\xbe\x14\x8b\x02\x14\x1f" + "\xa9\x73\xb7\xad\x15\xbe\x9c\xa0" + "\xd2\x8a\x2c\xdc\xd4\xe3\x05\x55" + "\x0a\xf5\xf8\x51\xee\xe5\x62\xa5" + "\x71\xa7\x7c\x15\x5d\x7a\x9e", + .ilen = 63, + .result = "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00", + .rlen = 63, + .also_non_np = 1, + .np = 8, + .tap = { 20, 20, 10, 8, 2, 1, 1, 1 }, + }, { + .key = "\x00\x01\x02\x03\x04\x05\x06\x07" + "\x08\x09\x0A\x0B\x0C\x0D\x0E\x0F" + "\x00\x01\x02\x03\x04\x05\x06\x07" + "\x08\x09\x0A\x0B\x0C\x0D\x0E\x0F" + "\x00\x01\x02\x03\x04\x05\x06\x07" + "\x08\x09\x0A\x0B\x0C\x0D\x0E\x0F", + .klen = 16, + .iv = "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00", + .input = "\x4b\x1a\x15\xa0\xaf\x08\x6d\x70" + "\xf0\xa7\x97\xb5\x31\x4b\x8c\xc3" + "\x4d\xf2\x7a\x9d\xdd\xd4\x15\x99" + "\x57\xad\xc6\xb1\x35\x69\xf5\x6a" + "\x2d\x70\xe4\x97\x49\xb2\x9f\x71" + "\xde\x22\xb5\x70\x8c\x69\x24\xd3" + "\xad\x80\x58\x48\x90\xe4\xed\xba" + "\x76\x3d\x71\x7c\x57\x25\x87", + .ilen = 63, + .result = "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00\x01" + "\x00\x00\x00\x00\x00\x00\x00\x00" + "\x00\x00\x00\x00\x00\x00\x00", + .rlen = 63, + .also_non_np = 1, + .np = 8, + .tap = { 20, 20, 10, 8, 2, 1, 1, 1 }, + } +}; + static struct cipher_testvec aes_cbc_enc_tv_template[] = { { /* From RFC 3602 */ .key = "\x06\xa9\x21\x40\x36\xb8\xa1\x5b" diff --git a/drivers/android/binder.c b/drivers/android/binder.c index 73ed341ded78..e6af2e819980 100644 --- a/drivers/android/binder.c +++ b/drivers/android/binder.c @@ -2047,7 +2047,7 @@ static void binder_transaction(struct binder_proc *proc, if (!IS_ALIGNED(extra_buffers_size, sizeof(u64))) { binder_user_error("%d:%d got transaction with unaligned buffers size, %lld\n", proc->pid, thread->pid, - extra_buffers_size); + (u64)extra_buffers_size); return_error = BR_FAILED_REPLY; goto err_bad_offset; } diff --git a/fs/9p/acl.c b/fs/9p/acl.c index 929b618da43b..c30c6ceac2c4 100644 --- a/fs/9p/acl.c +++ b/fs/9p/acl.c @@ -283,6 +283,7 @@ static int v9fs_xattr_set_acl(const struct xattr_handler *handler, case ACL_TYPE_ACCESS: if (acl) { struct iattr iattr; + struct posix_acl *old_acl = acl; retval = posix_acl_update_mode(inode, &iattr.ia_mode, &acl); if (retval) @@ -293,6 +294,7 @@ static int v9fs_xattr_set_acl(const struct xattr_handler *handler, * by the mode bits. So don't * update ACL. */ + posix_acl_release(old_acl); value = NULL; size = 0; } diff --git a/fs/attr.c b/fs/attr.c index d62f674a605f..11be2265a2d5 100644 --- a/fs/attr.c +++ b/fs/attr.c @@ -187,7 +187,7 @@ EXPORT_SYMBOL(setattr_copy); * the file open for write, as there can be no conflicting delegation in * that case. */ -int notify_change(struct dentry * dentry, struct iattr * attr, struct inode **delegated_inode) +int notify_change2(struct vfsmount *mnt, struct dentry * dentry, struct iattr * attr, struct inode **delegated_inode) { struct inode *inode = dentry->d_inode; umode_t mode = inode->i_mode; @@ -277,7 +277,9 @@ int notify_change(struct dentry * dentry, struct iattr * attr, struct inode **de if (error) return error; - if (inode->i_op->setattr) + if (mnt && inode->i_op->setattr2) + error = inode->i_op->setattr2(mnt, dentry, attr); + else if (inode->i_op->setattr) error = inode->i_op->setattr(dentry, attr); else error = simple_setattr(dentry, attr); @@ -290,4 +292,10 @@ int notify_change(struct dentry * dentry, struct iattr * attr, struct inode **de return error; } +EXPORT_SYMBOL(notify_change2); + +int notify_change(struct dentry * dentry, struct iattr * attr, struct inode **delegated_inode) +{ + return notify_change2(NULL, dentry, attr, delegated_inode); +} EXPORT_SYMBOL(notify_change); diff --git a/fs/coredump.c b/fs/coredump.c index 5d15c4975ba1..fe0a28da18a6 100644 --- a/fs/coredump.c +++ b/fs/coredump.c @@ -720,7 +720,7 @@ void do_coredump(const siginfo_t *siginfo) goto close_fail; if (!(cprm.file->f_mode & FMODE_CAN_WRITE)) goto close_fail; - if (do_truncate(cprm.file->f_path.dentry, 0, 0, cprm.file)) + if (do_truncate2(cprm.file->f_path.mnt, cprm.file->f_path.dentry, 0, 0, cprm.file)) goto close_fail; } diff --git a/fs/exec.c b/fs/exec.c index 3a6de10d3891..b1f5ddbf7d56 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -1132,8 +1132,10 @@ EXPORT_SYMBOL(flush_old_exec); void would_dump(struct linux_binprm *bprm, struct file *file) { struct inode *inode = file_inode(file); - if (inode_permission(inode, MAY_READ) < 0) { + + if (inode_permission2(file->f_path.mnt, inode, MAY_READ) < 0) { struct user_namespace *old, *user_ns; + bprm->interp_flags |= BINPRM_FLAGS_ENFORCE_NONDUMP; /* Ensure mm->user_ns contains the executable */ diff --git a/fs/ext4/Kconfig b/fs/ext4/Kconfig index b46e9fc64196..3c8293215603 100644 --- a/fs/ext4/Kconfig +++ b/fs/ext4/Kconfig @@ -106,6 +106,7 @@ config EXT4_ENCRYPTION select CRYPTO_ECB select CRYPTO_XTS select CRYPTO_CTS + select CRYPTO_HEH select CRYPTO_CTR select CRYPTO_SHA256 select KEYS diff --git a/fs/ext4/crypto_fname.c b/fs/ext4/crypto_fname.c index 2fbef8a14760..e2645ca9b95e 100644 --- a/fs/ext4/crypto_fname.c +++ b/fs/ext4/crypto_fname.c @@ -44,7 +44,8 @@ static void ext4_dir_crypt_complete(struct crypto_async_request *req, int res) bool ext4_valid_filenames_enc_mode(uint32_t mode) { - return (mode == EXT4_ENCRYPTION_MODE_AES_256_CTS); + return (mode == EXT4_ENCRYPTION_MODE_AES_256_CTS || + mode == EXT4_ENCRYPTION_MODE_AES_256_HEH); } static unsigned max_name_len(struct inode *inode) diff --git a/fs/ext4/crypto_key.c b/fs/ext4/crypto_key.c index 9a16d1e75a49..98d58c8ee804 100644 --- a/fs/ext4/crypto_key.c +++ b/fs/ext4/crypto_key.c @@ -29,16 +29,16 @@ static void derive_crypt_complete(struct crypto_async_request *req, int rc) } /** - * ext4_derive_key_aes() - Derive a key using AES-128-ECB + * ext4_derive_key_v1() - Derive a key using AES-128-ECB * @deriving_key: Encryption key used for derivation. * @source_key: Source key to which to apply derivation. * @derived_key: Derived key. * - * Return: Zero on success; non-zero otherwise. + * Return: 0 on success, -errno on failure */ -static int ext4_derive_key_aes(char deriving_key[EXT4_AES_128_ECB_KEY_SIZE], - char source_key[EXT4_AES_256_XTS_KEY_SIZE], - char derived_key[EXT4_AES_256_XTS_KEY_SIZE]) +static int ext4_derive_key_v1(const char deriving_key[EXT4_AES_128_ECB_KEY_SIZE], + const char source_key[EXT4_AES_256_XTS_KEY_SIZE], + char derived_key[EXT4_AES_256_XTS_KEY_SIZE]) { int res = 0; struct ablkcipher_request *req = NULL; @@ -83,6 +83,91 @@ out: return res; } +/** + * ext4_derive_key_v2() - Derive a key non-reversibly + * @nonce: the nonce associated with the file + * @master_key: the master key referenced by the file + * @derived_key: (output) the resulting derived key + * + * This function computes the following: + * derived_key[0:127] = AES-256-ENCRYPT(master_key[0:255], nonce) + * derived_key[128:255] = AES-256-ENCRYPT(master_key[0:255], nonce ^ 0x01) + * derived_key[256:383] = AES-256-ENCRYPT(master_key[256:511], nonce) + * derived_key[384:511] = AES-256-ENCRYPT(master_key[256:511], nonce ^ 0x01) + * + * 'nonce ^ 0x01' denotes flipping the low order bit of the last byte. + * + * Unlike the v1 algorithm, the v2 algorithm is "non-reversible", meaning that + * compromising a derived key does not also compromise the master key. + * + * Return: 0 on success, -errno on failure + */ +static int ext4_derive_key_v2(const char nonce[EXT4_KEY_DERIVATION_NONCE_SIZE], + const char master_key[EXT4_MAX_KEY_SIZE], + char derived_key[EXT4_MAX_KEY_SIZE]) +{ + const int noncelen = EXT4_KEY_DERIVATION_NONCE_SIZE; + struct crypto_cipher *tfm; + int err; + int i; + + /* + * Since we only use each transform for a small number of encryptions, + * requesting just "aes" turns out to be significantly faster than + * "ecb(aes)", by about a factor of two. + */ + tfm = crypto_alloc_cipher("aes", 0, 0); + if (IS_ERR(tfm)) + return PTR_ERR(tfm); + + BUILD_BUG_ON(4 * EXT4_KEY_DERIVATION_NONCE_SIZE != EXT4_MAX_KEY_SIZE); + BUILD_BUG_ON(2 * EXT4_AES_256_ECB_KEY_SIZE != EXT4_MAX_KEY_SIZE); + for (i = 0; i < 2; i++) { + memcpy(derived_key, nonce, noncelen); + memcpy(derived_key + noncelen, nonce, noncelen); + derived_key[2 * noncelen - 1] ^= 0x01; + err = crypto_cipher_setkey(tfm, master_key, + EXT4_AES_256_ECB_KEY_SIZE); + if (err) + break; + crypto_cipher_encrypt_one(tfm, derived_key, derived_key); + crypto_cipher_encrypt_one(tfm, derived_key + noncelen, + derived_key + noncelen); + master_key += EXT4_AES_256_ECB_KEY_SIZE; + derived_key += 2 * noncelen; + } + crypto_free_cipher(tfm); + return err; +} + +/** + * ext4_derive_key() - Derive a per-file key from a nonce and master key + * @ctx: the encryption context associated with the file + * @master_key: the master key referenced by the file + * @derived_key: (output) the resulting derived key + * + * Return: 0 on success, -errno on failure + */ +static int ext4_derive_key(const struct ext4_encryption_context *ctx, + const char master_key[EXT4_MAX_KEY_SIZE], + char derived_key[EXT4_MAX_KEY_SIZE]) +{ + BUILD_BUG_ON(EXT4_AES_128_ECB_KEY_SIZE != EXT4_KEY_DERIVATION_NONCE_SIZE); + BUILD_BUG_ON(EXT4_AES_256_XTS_KEY_SIZE != EXT4_MAX_KEY_SIZE); + + /* + * Although the key derivation algorithm is logically independent of the + * choice of encryption modes, in this kernel it is bundled with HEH + * encryption of filenames, which is another crypto improvement that + * requires an on-disk format change and requires userspace to specify + * different encryption policies. + */ + if (ctx->filenames_encryption_mode == EXT4_ENCRYPTION_MODE_AES_256_HEH) + return ext4_derive_key_v2(ctx->nonce, master_key, derived_key); + else + return ext4_derive_key_v1(ctx->nonce, master_key, derived_key); +} + void ext4_free_crypt_info(struct ext4_crypt_info *ci) { if (!ci) @@ -182,6 +267,9 @@ retry: case EXT4_ENCRYPTION_MODE_AES_256_CTS: cipher_str = "cts(cbc(aes))"; break; + case EXT4_ENCRYPTION_MODE_AES_256_HEH: + cipher_str = "heh(aes)"; + break; default: printk_once(KERN_WARNING "ext4: unsupported key mode %d (ino %u)\n", @@ -231,8 +319,7 @@ retry: up_read(&keyring_key->sem); goto out; } - res = ext4_derive_key_aes(ctx.nonce, master_key->raw, - raw_key); + res = ext4_derive_key(&ctx, master_key->raw, raw_key); up_read(&keyring_key->sem); if (res) goto out; diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 9f31991a5e05..8be0e4d330ab 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -589,6 +589,7 @@ enum { #define EXT4_ENCRYPTION_MODE_AES_256_GCM 2 #define EXT4_ENCRYPTION_MODE_AES_256_CBC 3 #define EXT4_ENCRYPTION_MODE_AES_256_CTS 4 +#define EXT4_ENCRYPTION_MODE_AES_256_HEH 126 #include "ext4_crypto.h" diff --git a/fs/ext4/ext4_crypto.h b/fs/ext4/ext4_crypto.h index ac7d4e813796..5a1684bd083b 100644 --- a/fs/ext4/ext4_crypto.h +++ b/fs/ext4/ext4_crypto.h @@ -58,8 +58,10 @@ struct ext4_encryption_context { #define EXT4_XTS_TWEAK_SIZE 16 #define EXT4_AES_128_ECB_KEY_SIZE 16 #define EXT4_AES_256_GCM_KEY_SIZE 32 +#define EXT4_AES_256_ECB_KEY_SIZE 32 #define EXT4_AES_256_CBC_KEY_SIZE 32 #define EXT4_AES_256_CTS_KEY_SIZE 32 +#define EXT4_AES_256_HEH_KEY_SIZE 32 #define EXT4_AES_256_XTS_KEY_SIZE 64 #define EXT4_MAX_KEY_SIZE 64 @@ -121,6 +123,8 @@ static inline int ext4_encryption_key_size(int mode) return EXT4_AES_256_CBC_KEY_SIZE; case EXT4_ENCRYPTION_MODE_AES_256_CTS: return EXT4_AES_256_CTS_KEY_SIZE; + case EXT4_ENCRYPTION_MODE_AES_256_HEH: + return EXT4_AES_256_HEH_KEY_SIZE; default: BUG(); } diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c index 43dcaab85201..98ca2bb1b829 100644 --- a/fs/ext4/inline.c +++ b/fs/ext4/inline.c @@ -503,8 +503,16 @@ int ext4_readpage_inline(struct inode *inode, struct page *page) return -EAGAIN; } - trace_android_fs_dataread_start(inode, page_offset(page), PAGE_SIZE, - current->pid, current->comm); + if (trace_android_fs_dataread_start_enabled()) { + char *path, pathbuf[MAX_TRACE_PATHBUF_LEN]; + + path = android_fstrace_get_pathname(pathbuf, + MAX_TRACE_PATHBUF_LEN, + inode); + trace_android_fs_dataread_start(inode, page_offset(page), + PAGE_SIZE, current->pid, + path, current->comm); + } /* * Current inline data can only exist in the 1st page, diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index c33fcb4f8533..790961e33961 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -1017,8 +1017,16 @@ static int ext4_write_begin(struct file *file, struct address_space *mapping, pgoff_t index; unsigned from, to; - trace_android_fs_datawrite_start(inode, pos, len, - current->pid, current->comm); + if (trace_android_fs_datawrite_start_enabled()) { + char *path, pathbuf[MAX_TRACE_PATHBUF_LEN]; + + path = android_fstrace_get_pathname(pathbuf, + MAX_TRACE_PATHBUF_LEN, + inode); + trace_android_fs_datawrite_start(inode, pos, len, + current->pid, path, + current->comm); + } trace_ext4_write_begin(inode, pos, len, flags); /* * Reserve one block more for addition to orphan list in case @@ -2732,8 +2740,16 @@ static int ext4_da_write_begin(struct file *file, struct address_space *mapping, len, flags, pagep, fsdata); } *fsdata = (void *)0; - trace_android_fs_datawrite_start(inode, pos, len, - current->pid, current->comm); + if (trace_android_fs_datawrite_start_enabled()) { + char *path, pathbuf[MAX_TRACE_PATHBUF_LEN]; + + path = android_fstrace_get_pathname(pathbuf, + MAX_TRACE_PATHBUF_LEN, + inode); + trace_android_fs_datawrite_start(inode, pos, len, + current->pid, + path, current->comm); + } trace_ext4_da_write_begin(inode, pos, len, flags); if (ext4_test_inode_state(inode, EXT4_STATE_MAY_INLINE_DATA)) { @@ -3342,16 +3358,27 @@ static ssize_t ext4_direct_IO(struct kiocb *iocb, struct iov_iter *iter, return 0; if (trace_android_fs_dataread_start_enabled() && - (iov_iter_rw(iter) == READ)) - trace_android_fs_dataread_start(inode, offset, count, - current->pid, - current->comm); - if (trace_android_fs_datawrite_start_enabled() && - (iov_iter_rw(iter) == WRITE)) - trace_android_fs_datawrite_start(inode, offset, count, - current->pid, - current->comm); + (iov_iter_rw(iter) == READ)) { + char *path, pathbuf[MAX_TRACE_PATHBUF_LEN]; + path = android_fstrace_get_pathname(pathbuf, + MAX_TRACE_PATHBUF_LEN, + inode); + trace_android_fs_dataread_start(inode, offset, count, + current->pid, path, + current->comm); + } + if (trace_android_fs_datawrite_start_enabled() && + (iov_iter_rw(iter) == WRITE)) { + char *path, pathbuf[MAX_TRACE_PATHBUF_LEN]; + + path = android_fstrace_get_pathname(pathbuf, + MAX_TRACE_PATHBUF_LEN, + inode); + trace_android_fs_datawrite_start(inode, offset, count, + current->pid, path, + current->comm); + } trace_ext4_direct_IO_enter(inode, offset, count, iov_iter_rw(iter)); if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)) ret = ext4_ext_direct_IO(iocb, iter, offset); @@ -3587,6 +3614,11 @@ static int ext4_block_truncate_page(handle_t *handle, unsigned blocksize; struct inode *inode = mapping->host; + /* If we are processing an encrypted inode during orphan list + * handling */ + if (ext4_encrypted_inode(inode) && !ext4_has_encryption_key(inode)) + return 0; + blocksize = inode->i_sb->s_blocksize; length = blocksize - (offset & (blocksize - 1)); diff --git a/fs/ext4/readpage.c b/fs/ext4/readpage.c index 1ce24a6759a0..1c5db9fd9c8f 100644 --- a/fs/ext4/readpage.c +++ b/fs/ext4/readpage.c @@ -152,11 +152,17 @@ ext4_submit_bio_read(struct bio *bio) struct page *first_page = bio->bi_io_vec[0].bv_page; if (first_page != NULL) { + char *path, pathbuf[MAX_TRACE_PATHBUF_LEN]; + + path = android_fstrace_get_pathname(pathbuf, + MAX_TRACE_PATHBUF_LEN, + first_page->mapping->host); trace_android_fs_dataread_start( first_page->mapping->host, page_offset(first_page), bio->bi_iter.bi_size, current->pid, + path, current->comm); } } diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index e692958d6e78..8936044dee4c 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -1402,8 +1402,16 @@ static int f2fs_write_begin(struct file *file, struct address_space *mapping, struct dnode_of_data dn; int err = 0; - trace_android_fs_datawrite_start(inode, pos, len, - current->pid, current->comm); + if (trace_android_fs_datawrite_start_enabled()) { + char *path, pathbuf[MAX_TRACE_PATHBUF_LEN]; + + path = android_fstrace_get_pathname(pathbuf, + MAX_TRACE_PATHBUF_LEN, + inode); + trace_android_fs_datawrite_start(inode, pos, len, + current->pid, path, + current->comm); + } trace_f2fs_write_begin(inode, pos, len, flags); f2fs_balance_fs(sbi); @@ -1587,15 +1595,27 @@ static ssize_t f2fs_direct_IO(struct kiocb *iocb, struct iov_iter *iter, trace_f2fs_direct_IO_enter(inode, offset, count, iov_iter_rw(iter)); if (trace_android_fs_dataread_start_enabled() && - (iov_iter_rw(iter) == READ)) - trace_android_fs_dataread_start(inode, offset, - count, current->pid, - current->comm); - if (trace_android_fs_datawrite_start_enabled() && - (iov_iter_rw(iter) == WRITE)) - trace_android_fs_datawrite_start(inode, offset, count, - current->pid, current->comm); + (iov_iter_rw(iter) == READ)) { + char *path, pathbuf[MAX_TRACE_PATHBUF_LEN]; + path = android_fstrace_get_pathname(pathbuf, + MAX_TRACE_PATHBUF_LEN, + inode); + trace_android_fs_dataread_start(inode, offset, + count, current->pid, path, + current->comm); + } + if (trace_android_fs_datawrite_start_enabled() && + (iov_iter_rw(iter) == WRITE)) { + char *path, pathbuf[MAX_TRACE_PATHBUF_LEN]; + + path = android_fstrace_get_pathname(pathbuf, + MAX_TRACE_PATHBUF_LEN, + inode); + trace_android_fs_datawrite_start(inode, offset, count, + current->pid, path, + current->comm); + } if (iov_iter_rw(iter) == WRITE) { __allocate_data_blocks(inode, offset, count); if (unlikely(f2fs_cp_error(F2FS_I_SB(inode)))) { diff --git a/fs/f2fs/inline.c b/fs/f2fs/inline.c index d2c5d69ba0b1..dbb2cc4df603 100644 --- a/fs/f2fs/inline.c +++ b/fs/f2fs/inline.c @@ -85,9 +85,16 @@ int f2fs_read_inline_data(struct inode *inode, struct page *page) { struct page *ipage; - trace_android_fs_dataread_start(inode, page_offset(page), - PAGE_SIZE, current->pid, - current->comm); + if (trace_android_fs_dataread_start_enabled()) { + char *path, pathbuf[MAX_TRACE_PATHBUF_LEN]; + + path = android_fstrace_get_pathname(pathbuf, + MAX_TRACE_PATHBUF_LEN, + inode); + trace_android_fs_dataread_start(inode, page_offset(page), + PAGE_SIZE, current->pid, + path, current->comm); + } ipage = get_node_page(F2FS_I_SB(inode), inode->i_ino); if (IS_ERR(ipage)) { diff --git a/fs/fs_struct.c b/fs/fs_struct.c index 7dca743b2ce1..940c683561dd 100644 --- a/fs/fs_struct.c +++ b/fs/fs_struct.c @@ -44,6 +44,7 @@ void set_fs_pwd(struct fs_struct *fs, const struct path *path) if (old_pwd.dentry) path_put(&old_pwd); } +EXPORT_SYMBOL(set_fs_pwd); static inline int replace_path(struct path *p, const struct path *old, const struct path *new) { @@ -89,6 +90,7 @@ void free_fs_struct(struct fs_struct *fs) path_put(&fs->pwd); kmem_cache_free(fs_cachep, fs); } +EXPORT_SYMBOL(free_fs_struct); void exit_fs(struct task_struct *tsk) { @@ -127,6 +129,7 @@ struct fs_struct *copy_fs_struct(struct fs_struct *old) } return fs; } +EXPORT_SYMBOL_GPL(copy_fs_struct); int unshare_fs_struct(void) { diff --git a/fs/inode.c b/fs/inode.c index b0edef500590..e89f1271c545 100644 --- a/fs/inode.c +++ b/fs/inode.c @@ -1715,7 +1715,7 @@ int dentry_needs_remove_privs(struct dentry *dentry) } EXPORT_SYMBOL(dentry_needs_remove_privs); -static int __remove_privs(struct dentry *dentry, int kill) +static int __remove_privs(struct vfsmount *mnt, struct dentry *dentry, int kill) { struct iattr newattrs; @@ -1724,7 +1724,7 @@ static int __remove_privs(struct dentry *dentry, int kill) * Note we call this on write, so notify_change will not * encounter any conflicting delegations: */ - return notify_change(dentry, &newattrs, NULL); + return notify_change2(mnt, dentry, &newattrs, NULL); } /* @@ -1746,7 +1746,7 @@ int file_remove_privs(struct file *file) if (kill < 0) return kill; if (kill) - error = __remove_privs(dentry, kill); + error = __remove_privs(file->f_path.mnt, dentry, kill); if (!error) inode_has_no_xattr(inode); diff --git a/fs/internal.h b/fs/internal.h index 71859c4d0b41..6387b35a1c0d 100644 --- a/fs/internal.h +++ b/fs/internal.h @@ -84,9 +84,11 @@ extern struct file *get_empty_filp(void); * super.c */ extern int do_remount_sb(struct super_block *, int, void *, int); +extern int do_remount_sb2(struct vfsmount *, struct super_block *, int, + void *, int); extern bool trylock_super(struct super_block *sb); extern struct dentry *mount_fs(struct file_system_type *, - int, const char *, void *); + int, const char *, struct vfsmount *, void *); extern struct super_block *user_get_super(dev_t); /* diff --git a/fs/mpage.c b/fs/mpage.c index 5c65d8942692..0fd48fdcc1b1 100644 --- a/fs/mpage.c +++ b/fs/mpage.c @@ -79,11 +79,17 @@ static struct bio *mpage_bio_submit(int rw, struct bio *bio) struct page *first_page = bio->bi_io_vec[0].bv_page; if (first_page != NULL) { + char *path, pathbuf[MAX_TRACE_PATHBUF_LEN]; + + path = android_fstrace_get_pathname(pathbuf, + MAX_TRACE_PATHBUF_LEN, + first_page->mapping->host); trace_android_fs_dataread_start( first_page->mapping->host, page_offset(first_page), bio->bi_iter.bi_size, current->pid, + path, current->comm); } } diff --git a/fs/namei.c b/fs/namei.c index 0b0acba72a71..d185869dae93 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -373,9 +373,11 @@ EXPORT_SYMBOL(generic_permission); * flag in inode->i_opflags, that says "this has not special * permission function, use the fast case". */ -static inline int do_inode_permission(struct inode *inode, int mask) +static inline int do_inode_permission(struct vfsmount *mnt, struct inode *inode, int mask) { if (unlikely(!(inode->i_opflags & IOP_FASTPERM))) { + if (likely(mnt && inode->i_op->permission2)) + return inode->i_op->permission2(mnt, inode, mask); if (likely(inode->i_op->permission)) return inode->i_op->permission(inode, mask); @@ -399,7 +401,7 @@ static inline int do_inode_permission(struct inode *inode, int mask) * This does not check for a read-only file system. You probably want * inode_permission(). */ -int __inode_permission(struct inode *inode, int mask) +int __inode_permission2(struct vfsmount *mnt, struct inode *inode, int mask) { int retval; @@ -411,7 +413,7 @@ int __inode_permission(struct inode *inode, int mask) return -EACCES; } - retval = do_inode_permission(inode, mask); + retval = do_inode_permission(mnt, inode, mask); if (retval) return retval; @@ -419,7 +421,14 @@ int __inode_permission(struct inode *inode, int mask) if (retval) return retval; - return security_inode_permission(inode, mask); + retval = security_inode_permission(inode, mask); + return retval; +} +EXPORT_SYMBOL(__inode_permission2); + +int __inode_permission(struct inode *inode, int mask) +{ + return __inode_permission2(NULL, inode, mask); } EXPORT_SYMBOL(__inode_permission); @@ -455,14 +464,20 @@ static int sb_permission(struct super_block *sb, struct inode *inode, int mask) * * When checking for MAY_APPEND, MAY_WRITE must also be set in @mask. */ -int inode_permission(struct inode *inode, int mask) +int inode_permission2(struct vfsmount *mnt, struct inode *inode, int mask) { int retval; retval = sb_permission(inode->i_sb, inode, mask); if (retval) return retval; - return __inode_permission(inode, mask); + return __inode_permission2(mnt, inode, mask); +} +EXPORT_SYMBOL(inode_permission2); + +int inode_permission(struct inode *inode, int mask) +{ + return inode_permission2(NULL, inode, mask); } EXPORT_SYMBOL(inode_permission); @@ -1645,13 +1660,13 @@ static int lookup_slow(struct nameidata *nd, struct path *path) static inline int may_lookup(struct nameidata *nd) { if (nd->flags & LOOKUP_RCU) { - int err = inode_permission(nd->inode, MAY_EXEC|MAY_NOT_BLOCK); + int err = inode_permission2(nd->path.mnt, nd->inode, MAY_EXEC|MAY_NOT_BLOCK); if (err != -ECHILD) return err; if (unlazy_walk(nd, NULL, 0)) return -ECHILD; } - return inode_permission(nd->inode, MAY_EXEC); + return inode_permission2(nd->path.mnt, nd->inode, MAY_EXEC); } static inline int handle_dots(struct nameidata *nd, int type) @@ -2005,11 +2020,12 @@ static const char *path_init(struct nameidata *nd, unsigned flags) nd->depth = 0; if (flags & LOOKUP_ROOT) { struct dentry *root = nd->root.dentry; + struct vfsmount *mnt = nd->root.mnt; struct inode *inode = root->d_inode; if (*s) { if (!d_can_lookup(root)) return ERR_PTR(-ENOTDIR); - retval = inode_permission(inode, MAY_EXEC); + retval = inode_permission2(mnt, inode, MAY_EXEC); if (retval) return ERR_PTR(retval); } @@ -2280,13 +2296,14 @@ EXPORT_SYMBOL(vfs_path_lookup); /** * lookup_one_len - filesystem helper to lookup single pathname component * @name: pathname component to lookup + * @mnt: mount we are looking up on * @base: base directory to lookup from * @len: maximum length @len should be interpreted to * * Note that this routine is purely a helper for filesystem usage and should * not be called by generic code. */ -struct dentry *lookup_one_len(const char *name, struct dentry *base, int len) +struct dentry *lookup_one_len2(const char *name, struct vfsmount *mnt, struct dentry *base, int len) { struct qstr this; unsigned int c; @@ -2320,12 +2337,18 @@ struct dentry *lookup_one_len(const char *name, struct dentry *base, int len) return ERR_PTR(err); } - err = inode_permission(base->d_inode, MAY_EXEC); + err = inode_permission2(mnt, base->d_inode, MAY_EXEC); if (err) return ERR_PTR(err); return __lookup_hash(&this, base, 0); } +EXPORT_SYMBOL(lookup_one_len2); + +struct dentry *lookup_one_len(const char *name, struct dentry *base, int len) +{ + return lookup_one_len2(name, NULL, base, len); +} EXPORT_SYMBOL(lookup_one_len); int user_path_at_empty(int dfd, const char __user *name, unsigned flags, @@ -2552,7 +2575,7 @@ EXPORT_SYMBOL(__check_sticky); * 10. We don't allow removal of NFS sillyrenamed files; it's handled by * nfs_async_unlink(). */ -static int may_delete(struct inode *dir, struct dentry *victim, bool isdir) +static int may_delete(struct vfsmount *mnt, struct inode *dir, struct dentry *victim, bool isdir) { struct inode *inode = d_backing_inode(victim); int error; @@ -2564,7 +2587,7 @@ static int may_delete(struct inode *dir, struct dentry *victim, bool isdir) BUG_ON(victim->d_parent->d_inode != dir); audit_inode_child(dir, victim, AUDIT_TYPE_CHILD_DELETE); - error = inode_permission(dir, MAY_WRITE | MAY_EXEC); + error = inode_permission2(mnt, dir, MAY_WRITE | MAY_EXEC); if (error) return error; if (IS_APPEND(dir)) @@ -2595,14 +2618,14 @@ static int may_delete(struct inode *dir, struct dentry *victim, bool isdir) * 3. We should have write and exec permissions on dir * 4. We can't do it if dir is immutable (done in permission()) */ -static inline int may_create(struct inode *dir, struct dentry *child) +static inline int may_create(struct vfsmount *mnt, struct inode *dir, struct dentry *child) { audit_inode_child(dir, child, AUDIT_TYPE_CHILD_CREATE); if (child->d_inode) return -EEXIST; if (IS_DEADDIR(dir)) return -ENOENT; - return inode_permission(dir, MAY_WRITE | MAY_EXEC); + return inode_permission2(mnt, dir, MAY_WRITE | MAY_EXEC); } /* @@ -2649,10 +2672,10 @@ void unlock_rename(struct dentry *p1, struct dentry *p2) } EXPORT_SYMBOL(unlock_rename); -int vfs_create(struct inode *dir, struct dentry *dentry, umode_t mode, - bool want_excl) +int vfs_create2(struct vfsmount *mnt, struct inode *dir, struct dentry *dentry, + umode_t mode, bool want_excl) { - int error = may_create(dir, dentry); + int error = may_create(mnt, dir, dentry); if (error) return error; @@ -2668,11 +2691,19 @@ int vfs_create(struct inode *dir, struct dentry *dentry, umode_t mode, fsnotify_create(dir, dentry); return error; } +EXPORT_SYMBOL(vfs_create2); + +int vfs_create(struct inode *dir, struct dentry *dentry, umode_t mode, + bool want_excl) +{ + return vfs_create2(NULL, dir, dentry, mode, want_excl); +} EXPORT_SYMBOL(vfs_create); static int may_open(struct path *path, int acc_mode, int flag) { struct dentry *dentry = path->dentry; + struct vfsmount *mnt = path->mnt; struct inode *inode = dentry->d_inode; int error; @@ -2701,7 +2732,7 @@ static int may_open(struct path *path, int acc_mode, int flag) break; } - error = inode_permission(inode, acc_mode); + error = inode_permission2(mnt, inode, acc_mode); if (error) return error; @@ -2736,7 +2767,7 @@ static int handle_truncate(struct file *filp) if (!error) error = security_path_truncate(path); if (!error) { - error = do_truncate(path->dentry, 0, + error = do_truncate2(path->mnt, path->dentry, 0, ATTR_MTIME|ATTR_CTIME|ATTR_OPEN, filp); } @@ -2757,7 +2788,7 @@ static int may_o_create(struct path *dir, struct dentry *dentry, umode_t mode) if (error) return error; - error = inode_permission(dir->dentry->d_inode, MAY_WRITE | MAY_EXEC); + error = inode_permission2(dir->mnt, dir->dentry->d_inode, MAY_WRITE | MAY_EXEC); if (error) return error; @@ -2943,6 +2974,7 @@ static int lookup_open(struct nameidata *nd, struct path *path, bool got_write, int *opened) { struct dentry *dir = nd->path.dentry; + struct vfsmount *mnt = nd->path.mnt; struct inode *dir_inode = dir->d_inode; struct dentry *dentry; int error; @@ -2990,7 +3022,7 @@ static int lookup_open(struct nameidata *nd, struct path *path, error = security_path_mknod(&nd->path, dentry, mode, 0); if (error) goto out_dput; - error = vfs_create(dir->d_inode, dentry, mode, + error = vfs_create2(mnt, dir->d_inode, dentry, mode, nd->flags & LOOKUP_EXCL); if (error) goto out_dput; @@ -3252,7 +3284,7 @@ static int do_tmpfile(struct nameidata *nd, unsigned flags, goto out; dir = path.dentry->d_inode; /* we want directory to be writable */ - error = inode_permission(dir, MAY_WRITE | MAY_EXEC); + error = inode_permission2(path.mnt, dir, MAY_WRITE | MAY_EXEC); if (error) goto out2; if (!dir->i_op->tmpfile) { @@ -3486,9 +3518,9 @@ inline struct dentry *user_path_create(int dfd, const char __user *pathname, } EXPORT_SYMBOL(user_path_create); -int vfs_mknod(struct inode *dir, struct dentry *dentry, umode_t mode, dev_t dev) +int vfs_mknod2(struct vfsmount *mnt, struct inode *dir, struct dentry *dentry, umode_t mode, dev_t dev) { - int error = may_create(dir, dentry); + int error = may_create(mnt, dir, dentry); if (error) return error; @@ -3512,6 +3544,12 @@ int vfs_mknod(struct inode *dir, struct dentry *dentry, umode_t mode, dev_t dev) fsnotify_create(dir, dentry); return error; } +EXPORT_SYMBOL(vfs_mknod2); + +int vfs_mknod(struct inode *dir, struct dentry *dentry, umode_t mode, dev_t dev) +{ + return vfs_mknod2(NULL, dir, dentry, mode, dev); +} EXPORT_SYMBOL(vfs_mknod); static int may_mknod(umode_t mode) @@ -3554,10 +3592,10 @@ retry: goto out; switch (mode & S_IFMT) { case 0: case S_IFREG: - error = vfs_create(path.dentry->d_inode,dentry,mode,true); + error = vfs_create2(path.mnt, path.dentry->d_inode,dentry,mode,true); break; case S_IFCHR: case S_IFBLK: - error = vfs_mknod(path.dentry->d_inode,dentry,mode, + error = vfs_mknod2(path.mnt, path.dentry->d_inode,dentry,mode, new_decode_dev(dev)); break; case S_IFIFO: case S_IFSOCK: @@ -3578,9 +3616,9 @@ SYSCALL_DEFINE3(mknod, const char __user *, filename, umode_t, mode, unsigned, d return sys_mknodat(AT_FDCWD, filename, mode, dev); } -int vfs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode) +int vfs_mkdir2(struct vfsmount *mnt, struct inode *dir, struct dentry *dentry, umode_t mode) { - int error = may_create(dir, dentry); + int error = may_create(mnt, dir, dentry); unsigned max_links = dir->i_sb->s_max_links; if (error) @@ -3602,6 +3640,12 @@ int vfs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode) fsnotify_mkdir(dir, dentry); return error; } +EXPORT_SYMBOL(vfs_mkdir2); + +int vfs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode) +{ + return vfs_mkdir2(NULL, dir, dentry, mode); +} EXPORT_SYMBOL(vfs_mkdir); SYSCALL_DEFINE3(mkdirat, int, dfd, const char __user *, pathname, umode_t, mode) @@ -3620,7 +3664,7 @@ retry: mode &= ~current_umask(); error = security_path_mkdir(&path, dentry, mode); if (!error) - error = vfs_mkdir(path.dentry->d_inode, dentry, mode); + error = vfs_mkdir2(path.mnt, path.dentry->d_inode, dentry, mode); done_path_create(&path, dentry); if (retry_estale(error, lookup_flags)) { lookup_flags |= LOOKUP_REVAL; @@ -3659,9 +3703,9 @@ void dentry_unhash(struct dentry *dentry) } EXPORT_SYMBOL(dentry_unhash); -int vfs_rmdir(struct inode *dir, struct dentry *dentry) +int vfs_rmdir2(struct vfsmount *mnt, struct inode *dir, struct dentry *dentry) { - int error = may_delete(dir, dentry, 1); + int error = may_delete(mnt, dir, dentry, 1); if (error) return error; @@ -3696,6 +3740,12 @@ out: d_delete(dentry); return error; } +EXPORT_SYMBOL(vfs_rmdir2); + +int vfs_rmdir(struct inode *dir, struct dentry *dentry) +{ + return vfs_rmdir2(NULL, dir, dentry); +} EXPORT_SYMBOL(vfs_rmdir); static long do_rmdir(int dfd, const char __user *pathname) @@ -3741,7 +3791,7 @@ retry: error = security_path_rmdir(&path, dentry); if (error) goto exit3; - error = vfs_rmdir(path.dentry->d_inode, dentry); + error = vfs_rmdir2(path.mnt, path.dentry->d_inode, dentry); exit3: dput(dentry); exit2: @@ -3780,10 +3830,10 @@ SYSCALL_DEFINE1(rmdir, const char __user *, pathname) * be appropriate for callers that expect the underlying filesystem not * to be NFS exported. */ -int vfs_unlink(struct inode *dir, struct dentry *dentry, struct inode **delegated_inode) +int vfs_unlink2(struct vfsmount *mnt, struct inode *dir, struct dentry *dentry, struct inode **delegated_inode) { struct inode *target = dentry->d_inode; - int error = may_delete(dir, dentry, 0); + int error = may_delete(mnt, dir, dentry, 0); if (error) return error; @@ -3818,6 +3868,12 @@ out: return error; } +EXPORT_SYMBOL(vfs_unlink2); + +int vfs_unlink(struct inode *dir, struct dentry *dentry, struct inode **delegated_inode) +{ + return vfs_unlink2(NULL, dir, dentry, delegated_inode); +} EXPORT_SYMBOL(vfs_unlink); /* @@ -3865,7 +3921,7 @@ retry_deleg: error = security_path_unlink(&path, dentry); if (error) goto exit2; - error = vfs_unlink(path.dentry->d_inode, dentry, &delegated_inode); + error = vfs_unlink2(path.mnt, path.dentry->d_inode, dentry, &delegated_inode); exit2: dput(dentry); } @@ -3915,9 +3971,9 @@ SYSCALL_DEFINE1(unlink, const char __user *, pathname) return do_unlinkat(AT_FDCWD, pathname); } -int vfs_symlink(struct inode *dir, struct dentry *dentry, const char *oldname) +int vfs_symlink2(struct vfsmount *mnt, struct inode *dir, struct dentry *dentry, const char *oldname) { - int error = may_create(dir, dentry); + int error = may_create(mnt, dir, dentry); if (error) return error; @@ -3934,6 +3990,12 @@ int vfs_symlink(struct inode *dir, struct dentry *dentry, const char *oldname) fsnotify_create(dir, dentry); return error; } +EXPORT_SYMBOL(vfs_symlink2); + +int vfs_symlink(struct inode *dir, struct dentry *dentry, const char *oldname) +{ + return vfs_symlink2(NULL, dir, dentry, oldname); +} EXPORT_SYMBOL(vfs_symlink); SYSCALL_DEFINE3(symlinkat, const char __user *, oldname, @@ -3956,7 +4018,7 @@ retry: error = security_path_symlink(&path, dentry, from->name); if (!error) - error = vfs_symlink(path.dentry->d_inode, dentry, from->name); + error = vfs_symlink2(path.mnt, path.dentry->d_inode, dentry, from->name); done_path_create(&path, dentry); if (retry_estale(error, lookup_flags)) { lookup_flags |= LOOKUP_REVAL; @@ -3991,7 +4053,7 @@ SYSCALL_DEFINE2(symlink, const char __user *, oldname, const char __user *, newn * be appropriate for callers that expect the underlying filesystem not * to be NFS exported. */ -int vfs_link(struct dentry *old_dentry, struct inode *dir, struct dentry *new_dentry, struct inode **delegated_inode) +int vfs_link2(struct vfsmount *mnt, struct dentry *old_dentry, struct inode *dir, struct dentry *new_dentry, struct inode **delegated_inode) { struct inode *inode = old_dentry->d_inode; unsigned max_links = dir->i_sb->s_max_links; @@ -4000,7 +4062,7 @@ int vfs_link(struct dentry *old_dentry, struct inode *dir, struct dentry *new_de if (!inode) return -ENOENT; - error = may_create(dir, new_dentry); + error = may_create(mnt, dir, new_dentry); if (error) return error; @@ -4043,6 +4105,12 @@ int vfs_link(struct dentry *old_dentry, struct inode *dir, struct dentry *new_de fsnotify_link(dir, inode, new_dentry); return error; } +EXPORT_SYMBOL(vfs_link2); + +int vfs_link(struct dentry *old_dentry, struct inode *dir, struct dentry *new_dentry, struct inode **delegated_inode) +{ + return vfs_link2(NULL, old_dentry, dir, new_dentry, delegated_inode); +} EXPORT_SYMBOL(vfs_link); /* @@ -4098,7 +4166,7 @@ retry: error = security_path_link(old_path.dentry, &new_path, new_dentry); if (error) goto out_dput; - error = vfs_link(old_path.dentry, new_path.dentry->d_inode, new_dentry, &delegated_inode); + error = vfs_link2(old_path.mnt, old_path.dentry, new_path.dentry->d_inode, new_dentry, &delegated_inode); out_dput: done_path_create(&new_path, new_dentry); if (delegated_inode) { @@ -4173,7 +4241,8 @@ SYSCALL_DEFINE2(link, const char __user *, oldname, const char __user *, newname * ->i_mutex on parents, which works but leads to some truly excessive * locking]. */ -int vfs_rename(struct inode *old_dir, struct dentry *old_dentry, +int vfs_rename2(struct vfsmount *mnt, + struct inode *old_dir, struct dentry *old_dentry, struct inode *new_dir, struct dentry *new_dentry, struct inode **delegated_inode, unsigned int flags) { @@ -4192,19 +4261,19 @@ int vfs_rename(struct inode *old_dir, struct dentry *old_dentry, if (vfs_select_inode(old_dentry, 0) == vfs_select_inode(new_dentry, 0)) return 0; - error = may_delete(old_dir, old_dentry, is_dir); + error = may_delete(mnt, old_dir, old_dentry, is_dir); if (error) return error; if (!target) { - error = may_create(new_dir, new_dentry); + error = may_create(mnt, new_dir, new_dentry); } else { new_is_dir = d_is_dir(new_dentry); if (!(flags & RENAME_EXCHANGE)) - error = may_delete(new_dir, new_dentry, is_dir); + error = may_delete(mnt, new_dir, new_dentry, is_dir); else - error = may_delete(new_dir, new_dentry, new_is_dir); + error = may_delete(mnt, new_dir, new_dentry, new_is_dir); } if (error) return error; @@ -4221,12 +4290,12 @@ int vfs_rename(struct inode *old_dir, struct dentry *old_dentry, */ if (new_dir != old_dir) { if (is_dir) { - error = inode_permission(source, MAY_WRITE); + error = inode_permission2(mnt, source, MAY_WRITE); if (error) return error; } if ((flags & RENAME_EXCHANGE) && new_is_dir) { - error = inode_permission(target, MAY_WRITE); + error = inode_permission2(mnt, target, MAY_WRITE); if (error) return error; } @@ -4309,6 +4378,14 @@ out: return error; } +EXPORT_SYMBOL(vfs_rename2); + +int vfs_rename(struct inode *old_dir, struct dentry *old_dentry, + struct inode *new_dir, struct dentry *new_dentry, + struct inode **delegated_inode, unsigned int flags) +{ + return vfs_rename2(NULL, old_dir, old_dentry, new_dir, new_dentry, delegated_inode, flags); +} EXPORT_SYMBOL(vfs_rename); SYSCALL_DEFINE5(renameat2, int, olddfd, const char __user *, oldname, @@ -4422,7 +4499,7 @@ retry_deleg: &new_path, new_dentry, flags); if (error) goto exit5; - error = vfs_rename(old_path.dentry->d_inode, old_dentry, + error = vfs_rename2(old_path.mnt, old_path.dentry->d_inode, old_dentry, new_path.dentry->d_inode, new_dentry, &delegated_inode, flags); exit5: @@ -4467,7 +4544,7 @@ SYSCALL_DEFINE2(rename, const char __user *, oldname, const char __user *, newna int vfs_whiteout(struct inode *dir, struct dentry *dentry) { - int error = may_create(dir, dentry); + int error = may_create(NULL, dir, dentry); if (error) return error; diff --git a/fs/namespace.c b/fs/namespace.c index da98a1bbd8b5..4aad64ad9ad0 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -577,6 +577,7 @@ int sb_prepare_remount_readonly(struct super_block *sb) static void free_vfsmnt(struct mount *mnt) { + kfree(mnt->mnt.data); kfree_const(mnt->mnt_devname); #ifdef CONFIG_SMP free_percpu(mnt->mnt_pcp); @@ -966,11 +967,21 @@ vfs_kern_mount(struct file_system_type *type, int flags, const char *name, void if (!mnt) return ERR_PTR(-ENOMEM); + mnt->mnt.data = NULL; + if (type->alloc_mnt_data) { + mnt->mnt.data = type->alloc_mnt_data(); + if (!mnt->mnt.data) { + mnt_free_id(mnt); + free_vfsmnt(mnt); + return ERR_PTR(-ENOMEM); + } + } if (flags & MS_KERNMOUNT) mnt->mnt.mnt_flags = MNT_INTERNAL; - root = mount_fs(type, flags, name, data); + root = mount_fs(type, flags, name, &mnt->mnt, data); if (IS_ERR(root)) { + kfree(mnt->mnt.data); mnt_free_id(mnt); free_vfsmnt(mnt); return ERR_CAST(root); @@ -998,6 +1009,14 @@ static struct mount *clone_mnt(struct mount *old, struct dentry *root, if (!mnt) return ERR_PTR(-ENOMEM); + if (sb->s_op->clone_mnt_data) { + mnt->mnt.data = sb->s_op->clone_mnt_data(old->mnt.data); + if (!mnt->mnt.data) { + err = -ENOMEM; + goto out_free; + } + } + if (flag & (CL_SLAVE | CL_PRIVATE | CL_SHARED_TO_SLAVE)) mnt->mnt_group_id = 0; /* not a peer of original */ else @@ -1066,6 +1085,7 @@ static struct mount *clone_mnt(struct mount *old, struct dentry *root, return mnt; out_free: + kfree(mnt->mnt.data); mnt_free_id(mnt); free_vfsmnt(mnt); return ERR_PTR(err); @@ -2234,8 +2254,14 @@ static int do_remount(struct path *path, int flags, int mnt_flags, err = change_mount_flags(path->mnt, flags); else if (!capable(CAP_SYS_ADMIN)) err = -EPERM; - else - err = do_remount_sb(sb, flags, data, 0); + else { + err = do_remount_sb2(path->mnt, sb, flags, data, 0); + namespace_lock(); + lock_mount_hash(); + propagate_remount(mnt); + unlock_mount_hash(); + namespace_unlock(); + } if (!err) { lock_mount_hash(); mnt_flags |= mnt->mnt.mnt_flags & ~MNT_USER_SETTABLE_MASK; diff --git a/fs/notify/fanotify/fanotify_user.c b/fs/notify/fanotify/fanotify_user.c index a64313868d3a..2958e7a81f9c 100644 --- a/fs/notify/fanotify/fanotify_user.c +++ b/fs/notify/fanotify/fanotify_user.c @@ -488,7 +488,7 @@ static int fanotify_find_path(int dfd, const char __user *filename, } /* you can only watch an inode if you have read permissions on it */ - ret = inode_permission(path->dentry->d_inode, MAY_READ); + ret = inode_permission2(path->mnt, path->dentry->d_inode, MAY_READ); if (ret) path_put(path); out: diff --git a/fs/notify/inotify/inotify_user.c b/fs/notify/inotify/inotify_user.c index e2893f17dde2..4c5b43d15e6e 100644 --- a/fs/notify/inotify/inotify_user.c +++ b/fs/notify/inotify/inotify_user.c @@ -337,7 +337,7 @@ static int inotify_find_inode(const char __user *dirname, struct path *path, uns if (error) return error; /* you can only watch an inode if you have read permissions on it */ - error = inode_permission(path->dentry->d_inode, MAY_READ); + error = inode_permission2(path->mnt, path->dentry->d_inode, MAY_READ); if (error) path_put(path); return error; diff --git a/fs/open.c b/fs/open.c index 157b9940dd73..e70cca15c976 100644 --- a/fs/open.c +++ b/fs/open.c @@ -34,8 +34,8 @@ #include "internal.h" -int do_truncate(struct dentry *dentry, loff_t length, unsigned int time_attrs, - struct file *filp) +int do_truncate2(struct vfsmount *mnt, struct dentry *dentry, loff_t length, + unsigned int time_attrs, struct file *filp) { int ret; struct iattr newattrs; @@ -60,17 +60,24 @@ int do_truncate(struct dentry *dentry, loff_t length, unsigned int time_attrs, mutex_lock(&dentry->d_inode->i_mutex); /* Note any delegations or leases have already been broken: */ - ret = notify_change(dentry, &newattrs, NULL); + ret = notify_change2(mnt, dentry, &newattrs, NULL); mutex_unlock(&dentry->d_inode->i_mutex); return ret; } +int do_truncate(struct dentry *dentry, loff_t length, unsigned int time_attrs, + struct file *filp) +{ + return do_truncate2(NULL, dentry, length, time_attrs, filp); +} long vfs_truncate(struct path *path, loff_t length) { struct inode *inode; + struct vfsmount *mnt; long error; inode = path->dentry->d_inode; + mnt = path->mnt; /* For directories it's -EISDIR, for other non-regulars - -EINVAL */ if (S_ISDIR(inode->i_mode)) @@ -82,7 +89,7 @@ long vfs_truncate(struct path *path, loff_t length) if (error) goto out; - error = inode_permission(inode, MAY_WRITE); + error = inode_permission2(mnt, inode, MAY_WRITE); if (error) goto mnt_drop_write_and_out; @@ -106,7 +113,7 @@ long vfs_truncate(struct path *path, loff_t length) if (!error) error = security_path_truncate(path); if (!error) - error = do_truncate(path->dentry, length, 0, NULL); + error = do_truncate2(mnt, path->dentry, length, 0, NULL); put_write_and_out: put_write_access(inode); @@ -155,6 +162,7 @@ static long do_sys_ftruncate(unsigned int fd, loff_t length, int small) { struct inode *inode; struct dentry *dentry; + struct vfsmount *mnt; struct fd f; int error; @@ -171,6 +179,7 @@ static long do_sys_ftruncate(unsigned int fd, loff_t length, int small) small = 0; dentry = f.file->f_path.dentry; + mnt = f.file->f_path.mnt; inode = dentry->d_inode; error = -EINVAL; if (!S_ISREG(inode->i_mode) || !(f.file->f_mode & FMODE_WRITE)) @@ -190,7 +199,7 @@ static long do_sys_ftruncate(unsigned int fd, loff_t length, int small) if (!error) error = security_path_truncate(&f.file->f_path); if (!error) - error = do_truncate(dentry, length, ATTR_MTIME|ATTR_CTIME, f.file); + error = do_truncate2(mnt, dentry, length, ATTR_MTIME|ATTR_CTIME, f.file); sb_end_write(inode->i_sb); out_putf: fdput(f); @@ -340,6 +349,7 @@ SYSCALL_DEFINE3(faccessat, int, dfd, const char __user *, filename, int, mode) struct cred *override_cred; struct path path; struct inode *inode; + struct vfsmount *mnt; int res; unsigned int lookup_flags = LOOKUP_FOLLOW; @@ -370,6 +380,7 @@ retry: goto out; inode = d_backing_inode(path.dentry); + mnt = path.mnt; if ((mode & MAY_EXEC) && S_ISREG(inode->i_mode)) { /* @@ -381,7 +392,7 @@ retry: goto out_path_release; } - res = inode_permission(inode, mode | MAY_ACCESS); + res = inode_permission2(mnt, inode, mode | MAY_ACCESS); /* SuS v2 requires we report a read only fs too */ if (res || !(mode & S_IWOTH) || special_file(inode->i_mode)) goto out_path_release; @@ -425,7 +436,7 @@ retry: if (error) goto out; - error = inode_permission(path.dentry->d_inode, MAY_EXEC | MAY_CHDIR); + error = inode_permission2(path.mnt, path.dentry->d_inode, MAY_EXEC | MAY_CHDIR); if (error) goto dput_and_out; @@ -445,6 +456,7 @@ SYSCALL_DEFINE1(fchdir, unsigned int, fd) { struct fd f = fdget_raw(fd); struct inode *inode; + struct vfsmount *mnt; int error = -EBADF; error = -EBADF; @@ -452,12 +464,13 @@ SYSCALL_DEFINE1(fchdir, unsigned int, fd) goto out; inode = file_inode(f.file); + mnt = f.file->f_path.mnt; error = -ENOTDIR; if (!S_ISDIR(inode->i_mode)) goto out_putf; - error = inode_permission(inode, MAY_EXEC | MAY_CHDIR); + error = inode_permission2(mnt, inode, MAY_EXEC | MAY_CHDIR); if (!error) set_fs_pwd(current->fs, &f.file->f_path); out_putf: @@ -476,7 +489,7 @@ retry: if (error) goto out; - error = inode_permission(path.dentry->d_inode, MAY_EXEC | MAY_CHDIR); + error = inode_permission2(path.mnt, path.dentry->d_inode, MAY_EXEC | MAY_CHDIR); if (error) goto dput_and_out; @@ -516,7 +529,7 @@ retry_deleg: goto out_unlock; newattrs.ia_mode = (mode & S_IALLUGO) | (inode->i_mode & ~S_IALLUGO); newattrs.ia_valid = ATTR_MODE | ATTR_CTIME; - error = notify_change(path->dentry, &newattrs, &delegated_inode); + error = notify_change2(path->mnt, path->dentry, &newattrs, &delegated_inode); out_unlock: mutex_unlock(&inode->i_mutex); if (delegated_inode) { @@ -596,7 +609,7 @@ retry_deleg: mutex_lock(&inode->i_mutex); error = security_path_chown(path, uid, gid); if (!error) - error = notify_change(path->dentry, &newattrs, &delegated_inode); + error = notify_change2(path->mnt, path->dentry, &newattrs, &delegated_inode); mutex_unlock(&inode->i_mutex); if (delegated_inode) { error = break_deleg_wait(&delegated_inode); diff --git a/fs/pnode.c b/fs/pnode.c index 99899705b105..cbaa998ad625 100644 --- a/fs/pnode.c +++ b/fs/pnode.c @@ -458,3 +458,32 @@ int propagate_umount(struct list_head *list) __propagate_umount(mnt); return 0; } + +/* + * Iterates over all slaves, and slaves of slaves. + */ +static struct mount *next_descendent(struct mount *root, struct mount *cur) +{ + if (!IS_MNT_NEW(cur) && !list_empty(&cur->mnt_slave_list)) + return first_slave(cur); + do { + if (cur->mnt_slave.next != &cur->mnt_master->mnt_slave_list) + return next_slave(cur); + cur = cur->mnt_master; + } while (cur != root); + return NULL; +} + +void propagate_remount(struct mount *mnt) +{ + struct mount *m = mnt; + struct super_block *sb = mnt->mnt.mnt_sb; + + if (sb->s_op->copy_mnt_data) { + m = next_descendent(mnt, m); + while (m) { + sb->s_op->copy_mnt_data(m->mnt.data, mnt->mnt.data); + m = next_descendent(mnt, m); + } + } +} diff --git a/fs/pnode.h b/fs/pnode.h index 0fcdbe7ca648..3cb58c0cdcbc 100644 --- a/fs/pnode.h +++ b/fs/pnode.h @@ -44,6 +44,7 @@ int propagate_mnt(struct mount *, struct mountpoint *, struct mount *, int propagate_umount(struct list_head *); int propagate_mount_busy(struct mount *, int); void propagate_mount_unlock(struct mount *); +void propagate_remount(struct mount *); void mnt_release_group_id(struct mount *); int get_dominating_id(struct mount *mnt, const struct path *root); unsigned int mnt_get_count(struct mount *mnt); diff --git a/fs/proc_namespace.c b/fs/proc_namespace.c index 87645955990d..961e597acfc6 100644 --- a/fs/proc_namespace.c +++ b/fs/proc_namespace.c @@ -118,7 +118,9 @@ static int show_vfsmnt(struct seq_file *m, struct vfsmount *mnt) if (err) goto out; show_mnt_opts(m, mnt); - if (sb->s_op->show_options) + if (sb->s_op->show_options2) + err = sb->s_op->show_options2(mnt, m, mnt_path.dentry); + else if (sb->s_op->show_options) err = sb->s_op->show_options(m, mnt_path.dentry); seq_puts(m, " 0 0\n"); out: @@ -178,7 +180,9 @@ static int show_mountinfo(struct seq_file *m, struct vfsmount *mnt) err = show_sb_opts(m, sb); if (err) goto out; - if (sb->s_op->show_options) + if (sb->s_op->show_options2) { + err = sb->s_op->show_options2(mnt, m, mnt->mnt_root); + } else if (sb->s_op->show_options) err = sb->s_op->show_options(m, mnt->mnt_root); seq_putc(m, '\n'); out: diff --git a/fs/sdcardfs/derived_perm.c b/fs/sdcardfs/derived_perm.c index 41e0e11b3c35..0bb442338a85 100644 --- a/fs/sdcardfs/derived_perm.c +++ b/fs/sdcardfs/derived_perm.c @@ -30,11 +30,14 @@ static void inherit_derived_state(struct inode *parent, struct inode *child) ci->userid = pi->userid; ci->d_uid = pi->d_uid; ci->under_android = pi->under_android; + ci->under_cache = pi->under_cache; + ci->under_obb = pi->under_obb; + set_top(ci, pi->top); } /* helper function for derived state */ -void setup_derived_state(struct inode *inode, perm_t perm, - userid_t userid, uid_t uid, bool under_android) +void setup_derived_state(struct inode *inode, perm_t perm, userid_t userid, + uid_t uid, bool under_android, struct inode *top) { struct sdcardfs_inode_info *info = SDCARDFS_I(inode); @@ -42,84 +45,293 @@ void setup_derived_state(struct inode *inode, perm_t perm, info->userid = userid; info->d_uid = uid; info->under_android = under_android; + info->under_cache = false; + info->under_obb = false; + set_top(info, top); } /* While renaming, there is a point where we want the path from dentry, but the name from newdentry */ -void get_derived_permission_new(struct dentry *parent, struct dentry *dentry, struct dentry *newdentry) +void get_derived_permission_new(struct dentry *parent, struct dentry *dentry, const struct qstr *name) { - struct sdcardfs_sb_info *sbi = SDCARDFS_SB(dentry->d_sb); - struct sdcardfs_inode_info *info = SDCARDFS_I(dentry->d_inode); - struct sdcardfs_inode_info *parent_info= SDCARDFS_I(parent->d_inode); + struct sdcardfs_inode_info *info = SDCARDFS_I(d_inode(dentry)); + struct sdcardfs_inode_info *parent_info= SDCARDFS_I(d_inode(parent)); appid_t appid; + struct qstr q_Android = QSTR_LITERAL("Android"); + struct qstr q_data = QSTR_LITERAL("data"); + struct qstr q_obb = QSTR_LITERAL("obb"); + struct qstr q_media = QSTR_LITERAL("media"); + struct qstr q_cache = QSTR_LITERAL("cache"); /* By default, each inode inherits from its parent. * the properties are maintained on its private fields * because the inode attributes will be modified with that of * its lower inode. - * The derived state will be updated on the last - * stage of each system call by fix_derived_permission(inode). + * These values are used by our custom permission call instead + * of using the inode permissions. */ - inherit_derived_state(parent->d_inode, dentry->d_inode); + inherit_derived_state(d_inode(parent), d_inode(dentry)); + /* Files don't get special labels */ + if (!S_ISDIR(d_inode(dentry)->i_mode)) + return; /* Derive custom permissions based on parent and current node */ switch (parent_info->perm) { case PERM_INHERIT: + case PERM_ANDROID_PACKAGE_CACHE: /* Already inherited above */ break; case PERM_PRE_ROOT: /* Legacy internal layout places users at top level */ info->perm = PERM_ROOT; - info->userid = simple_strtoul(newdentry->d_name.name, NULL, 10); + info->userid = simple_strtoul(name->name, NULL, 10); + set_top(info, &info->vfs_inode); break; case PERM_ROOT: /* Assume masked off by default. */ - if (!strcasecmp(newdentry->d_name.name, "Android")) { + if (qstr_case_eq(name, &q_Android)) { /* App-specific directories inside; let anyone traverse */ info->perm = PERM_ANDROID; info->under_android = true; + set_top(info, &info->vfs_inode); } break; case PERM_ANDROID: - if (!strcasecmp(newdentry->d_name.name, "data")) { + if (qstr_case_eq(name, &q_data)) { /* App-specific directories inside; let anyone traverse */ info->perm = PERM_ANDROID_DATA; - } else if (!strcasecmp(newdentry->d_name.name, "obb")) { + set_top(info, &info->vfs_inode); + } else if (qstr_case_eq(name, &q_obb)) { /* App-specific directories inside; let anyone traverse */ info->perm = PERM_ANDROID_OBB; + info->under_obb = true; + set_top(info, &info->vfs_inode); /* Single OBB directory is always shared */ - } else if (!strcasecmp(newdentry->d_name.name, "media")) { + } else if (qstr_case_eq(name, &q_media)) { /* App-specific directories inside; let anyone traverse */ info->perm = PERM_ANDROID_MEDIA; + set_top(info, &info->vfs_inode); } break; - case PERM_ANDROID_DATA: case PERM_ANDROID_OBB: + case PERM_ANDROID_DATA: case PERM_ANDROID_MEDIA: - appid = get_appid(sbi->pkgl_id, newdentry->d_name.name); - if (appid != 0) { + info->perm = PERM_ANDROID_PACKAGE; + appid = get_appid(name->name); + if (appid != 0 && !is_excluded(name->name, parent_info->userid)) { info->d_uid = multiuser_get_uid(parent_info->userid, appid); } + set_top(info, &info->vfs_inode); + break; + case PERM_ANDROID_PACKAGE: + if (qstr_case_eq(name, &q_cache)) { + info->perm = PERM_ANDROID_PACKAGE_CACHE; + info->under_cache = true; + } break; } } void get_derived_permission(struct dentry *parent, struct dentry *dentry) { - get_derived_permission_new(parent, dentry, dentry); + get_derived_permission_new(parent, dentry, &dentry->d_name); } -void get_derive_permissions_recursive(struct dentry *parent) { +static appid_t get_type(const char *name) { + const char *ext = strrchr(name, '.'); + appid_t id; + + if (ext && ext[0]) { + ext = &ext[1]; + id = get_ext_gid(ext); + return id?:AID_MEDIA_RW; + } + return AID_MEDIA_RW; +} + +void fixup_lower_ownership(struct dentry* dentry, const char *name) { + struct path path; + struct inode *inode; + struct inode *delegated_inode = NULL; + int error; + struct sdcardfs_inode_info *info; + struct sdcardfs_inode_info *info_top; + perm_t perm; + struct sdcardfs_sb_info *sbi = SDCARDFS_SB(dentry->d_sb); + uid_t uid = sbi->options.fs_low_uid; + gid_t gid = sbi->options.fs_low_gid; + struct iattr newattrs; + + info = SDCARDFS_I(d_inode(dentry)); + perm = info->perm; + if (info->under_obb) { + perm = PERM_ANDROID_OBB; + } else if (info->under_cache) { + perm = PERM_ANDROID_PACKAGE_CACHE; + } else if (perm == PERM_INHERIT) { + info_top = SDCARDFS_I(grab_top(info)); + perm = info_top->perm; + release_top(info); + } + + switch (perm) { + case PERM_ROOT: + case PERM_ANDROID: + case PERM_ANDROID_DATA: + case PERM_ANDROID_MEDIA: + case PERM_ANDROID_PACKAGE: + case PERM_ANDROID_PACKAGE_CACHE: + uid = multiuser_get_uid(info->userid, uid); + break; + case PERM_ANDROID_OBB: + uid = AID_MEDIA_OBB; + break; + case PERM_PRE_ROOT: + default: + break; + } + switch (perm) { + case PERM_ROOT: + case PERM_ANDROID: + case PERM_ANDROID_DATA: + case PERM_ANDROID_MEDIA: + if (S_ISDIR(d_inode(dentry)->i_mode)) + gid = multiuser_get_uid(info->userid, AID_MEDIA_RW); + else + gid = multiuser_get_uid(info->userid, get_type(name)); + break; + case PERM_ANDROID_OBB: + gid = AID_MEDIA_OBB; + break; + case PERM_ANDROID_PACKAGE: + if (info->d_uid != 0) + gid = multiuser_get_ext_gid(info->userid, info->d_uid); + else + gid = multiuser_get_uid(info->userid, uid); + break; + case PERM_ANDROID_PACKAGE_CACHE: + if (info->d_uid != 0) + gid = multiuser_get_cache_gid(info->userid, info->d_uid); + else + gid = multiuser_get_uid(info->userid, uid); + break; + case PERM_PRE_ROOT: + default: + break; + } + + sdcardfs_get_lower_path(dentry, &path); + inode = d_inode(path.dentry); + if (d_inode(path.dentry)->i_gid.val != gid || d_inode(path.dentry)->i_uid.val != uid) { +retry_deleg: + newattrs.ia_valid = ATTR_GID | ATTR_UID | ATTR_FORCE; + newattrs.ia_uid = make_kuid(current_user_ns(), uid); + newattrs.ia_gid = make_kgid(current_user_ns(), gid); + if (!S_ISDIR(inode->i_mode)) + newattrs.ia_valid |= + ATTR_KILL_SUID | ATTR_KILL_SGID | ATTR_KILL_PRIV; + mutex_lock(&inode->i_mutex); + error = security_path_chown(&path, newattrs.ia_uid, newattrs.ia_gid); + if (!error) + error = notify_change2(path.mnt, path.dentry, &newattrs, &delegated_inode); + mutex_unlock(&inode->i_mutex); + if (delegated_inode) { + error = break_deleg_wait(&delegated_inode); + if (!error) + goto retry_deleg; + } + if (error) + pr_err("sdcardfs: Failed to touch up lower fs gid/uid.\n"); + } +} + +static int descendant_may_need_fixup(struct sdcardfs_inode_info *info, struct limit_search *limit) { + if (info->perm == PERM_ROOT) + return (limit->flags & BY_USERID)?info->userid == limit->userid:1; + if (info->perm == PERM_PRE_ROOT || info->perm == PERM_ANDROID) + return 1; + return 0; +} + +static int needs_fixup(perm_t perm) { + if (perm == PERM_ANDROID_DATA || perm == PERM_ANDROID_OBB + || perm == PERM_ANDROID_MEDIA) + return 1; + return 0; +} + +void fixup_perms_recursive(struct dentry *dentry, struct limit_search *limit) { + struct dentry *child; + struct sdcardfs_inode_info *info; + if (!dget(dentry)) + return; + if (!d_inode(dentry)) { + dput(dentry); + return; + } + info = SDCARDFS_I(d_inode(dentry)); + + if (needs_fixup(info->perm)) { + spin_lock(&dentry->d_lock); + list_for_each_entry(child, &dentry->d_subdirs, d_child) { + dget(child); + if (!(limit->flags & BY_NAME) || !strncasecmp(child->d_name.name, limit->name, limit->length)) { + if (d_inode(child)) { + get_derived_permission(dentry, child); + fixup_tmp_permissions(d_inode(child)); + dput(child); + break; + } + } + dput(child); + } + spin_unlock(&dentry->d_lock); + } else if (descendant_may_need_fixup(info, limit)) { + spin_lock(&dentry->d_lock); + list_for_each_entry(child, &dentry->d_subdirs, d_child) { + fixup_perms_recursive(child, limit); + } + spin_unlock(&dentry->d_lock); + } + dput(dentry); +} + +void drop_recursive(struct dentry *parent) { struct dentry *dentry; + struct sdcardfs_inode_info *info; + if (!d_inode(parent)) + return; + info = SDCARDFS_I(d_inode(parent)); + spin_lock(&parent->d_lock); list_for_each_entry(dentry, &parent->d_subdirs, d_child) { - if (dentry->d_inode) { - mutex_lock(&dentry->d_inode->i_mutex); - get_derived_permission(parent, dentry); - fix_derived_permission(dentry->d_inode); - get_derive_permissions_recursive(dentry); - mutex_unlock(&dentry->d_inode->i_mutex); + if (d_inode(dentry)) { + if (SDCARDFS_I(d_inode(parent))->top != SDCARDFS_I(d_inode(dentry))->top) { + drop_recursive(dentry); + d_drop(dentry); + } } } + spin_unlock(&parent->d_lock); +} + +void fixup_top_recursive(struct dentry *parent) { + struct dentry *dentry; + struct sdcardfs_inode_info *info; + + if (!d_inode(parent)) + return; + info = SDCARDFS_I(d_inode(parent)); + spin_lock(&parent->d_lock); + list_for_each_entry(dentry, &parent->d_subdirs, d_child) { + if (d_inode(dentry)) { + if (SDCARDFS_I(d_inode(parent))->top != SDCARDFS_I(d_inode(dentry))->top) { + get_derived_permission(parent, dentry); + fixup_tmp_permissions(d_inode(dentry)); + fixup_top_recursive(dentry); + } + } + } + spin_unlock(&parent->d_lock); } /* main function for updating derived permission */ @@ -127,7 +339,7 @@ inline void update_derived_permission_lock(struct dentry *dentry) { struct dentry *parent; - if(!dentry || !dentry->d_inode) { + if(!dentry || !d_inode(dentry)) { printk(KERN_ERR "sdcardfs: %s: invalid dentry\n", __func__); return; } @@ -135,9 +347,8 @@ inline void update_derived_permission_lock(struct dentry *dentry) * 1. need to check whether the dentry is updated or not * 2. remove the root dentry update */ - mutex_lock(&dentry->d_inode->i_mutex); if(IS_ROOT(dentry)) { - //setup_default_pre_root_state(dentry->d_inode); + //setup_default_pre_root_state(d_inode(dentry)); } else { parent = dget_parent(dentry); if(parent) { @@ -145,19 +356,19 @@ inline void update_derived_permission_lock(struct dentry *dentry) dput(parent); } } - fix_derived_permission(dentry->d_inode); - mutex_unlock(&dentry->d_inode->i_mutex); + fixup_tmp_permissions(d_inode(dentry)); } int need_graft_path(struct dentry *dentry) { int ret = 0; struct dentry *parent = dget_parent(dentry); - struct sdcardfs_inode_info *parent_info= SDCARDFS_I(parent->d_inode); + struct sdcardfs_inode_info *parent_info= SDCARDFS_I(d_inode(parent)); struct sdcardfs_sb_info *sbi = SDCARDFS_SB(dentry->d_sb); + struct qstr obb = QSTR_LITERAL("obb"); if(parent_info->perm == PERM_ANDROID && - !strcasecmp(dentry->d_name.name, "obb")) { + qstr_case_eq(&dentry->d_name, &obb)) { /* /Android/obb is the base obbpath of DERIVED_UNIFIED */ if(!(sbi->options.multiuser == false @@ -194,7 +405,7 @@ int is_obbpath_invalid(struct dentry *dent) } else { obbpath_s = d_path(&di->lower_path, path_buf, PATH_MAX); if (d_unhashed(di->lower_path.dentry) || - strcasecmp(sbi->obbpath_s, obbpath_s)) { + !str_case_eq(sbi->obbpath_s, obbpath_s)) { ret = 1; } kfree(path_buf); @@ -212,17 +423,18 @@ int is_base_obbpath(struct dentry *dentry) { int ret = 0; struct dentry *parent = dget_parent(dentry); - struct sdcardfs_inode_info *parent_info= SDCARDFS_I(parent->d_inode); + struct sdcardfs_inode_info *parent_info= SDCARDFS_I(d_inode(parent)); struct sdcardfs_sb_info *sbi = SDCARDFS_SB(dentry->d_sb); + struct qstr q_obb = QSTR_LITERAL("obb"); spin_lock(&SDCARDFS_D(dentry)->lock); if (sbi->options.multiuser) { if(parent_info->perm == PERM_PRE_ROOT && - !strcasecmp(dentry->d_name.name, "obb")) { + qstr_case_eq(&dentry->d_name, &q_obb)) { ret = 1; } } else if (parent_info->perm == PERM_ANDROID && - !strcasecmp(dentry->d_name.name, "obb")) { + qstr_case_eq(&dentry->d_name, &q_obb)) { ret = 1; } spin_unlock(&SDCARDFS_D(dentry)->lock); diff --git a/fs/sdcardfs/file.c b/fs/sdcardfs/file.c index c249fa982d3c..23f8cd7f8877 100644 --- a/fs/sdcardfs/file.c +++ b/fs/sdcardfs/file.c @@ -216,7 +216,7 @@ static int sdcardfs_open(struct inode *inode, struct file *file) goto out_err; } - if(!check_caller_access_to_name(parent->d_inode, dentry->d_name.name)) { + if(!check_caller_access_to_name(d_inode(parent), &dentry->d_name)) { printk(KERN_INFO "%s: need to check the caller's gid in packages.list\n" " dentry: %s, task:%s\n", __func__, dentry->d_name.name, current->comm); @@ -225,7 +225,7 @@ static int sdcardfs_open(struct inode *inode, struct file *file) } /* save current_cred and override it */ - OVERRIDE_CRED(sbi, saved_cred); + OVERRIDE_CRED(sbi, saved_cred, SDCARDFS_I(inode)); file->private_data = kzalloc(sizeof(struct sdcardfs_file_info), GFP_KERNEL); diff --git a/fs/sdcardfs/inode.c b/fs/sdcardfs/inode.c index 2528da0d3ae1..68e615045616 100644 --- a/fs/sdcardfs/inode.c +++ b/fs/sdcardfs/inode.c @@ -19,18 +19,24 @@ */ #include "sdcardfs.h" +#include /* Do not directly use this function. Use OVERRIDE_CRED() instead. */ -const struct cred * override_fsids(struct sdcardfs_sb_info* sbi) +const struct cred * override_fsids(struct sdcardfs_sb_info* sbi, struct sdcardfs_inode_info *info) { struct cred * cred; const struct cred * old_cred; + uid_t uid; cred = prepare_creds(); if (!cred) return NULL; - cred->fsuid = make_kuid(&init_user_ns, sbi->options.fs_low_uid); + if (info->under_obb) + uid = AID_MEDIA_OBB; + else + uid = multiuser_get_uid(info->userid, sbi->options.fs_low_uid); + cred->fsuid = make_kuid(&init_user_ns, uid); cred->fsgid = make_kgid(&init_user_ns, sbi->options.fs_low_gid); old_cred = override_creds(cred); @@ -53,11 +59,14 @@ static int sdcardfs_create(struct inode *dir, struct dentry *dentry, { int err; struct dentry *lower_dentry; + struct vfsmount *lower_dentry_mnt; struct dentry *lower_parent_dentry = NULL; struct path lower_path; const struct cred *saved_cred = NULL; + struct fs_struct *saved_fs; + struct fs_struct *copied_fs; - if(!check_caller_access_to_name(dir, dentry->d_name.name)) { + if(!check_caller_access_to_name(dir, &dentry->d_name)) { printk(KERN_INFO "%s: need to check the caller's gid in packages.list\n" " dentry: %s, task:%s\n", __func__, dentry->d_name.name, current->comm); @@ -66,15 +75,26 @@ static int sdcardfs_create(struct inode *dir, struct dentry *dentry, } /* save current_cred and override it */ - OVERRIDE_CRED(SDCARDFS_SB(dir->i_sb), saved_cred); + OVERRIDE_CRED(SDCARDFS_SB(dir->i_sb), saved_cred, SDCARDFS_I(dir)); sdcardfs_get_lower_path(dentry, &lower_path); lower_dentry = lower_path.dentry; + lower_dentry_mnt = lower_path.mnt; lower_parent_dentry = lock_parent(lower_dentry); /* set last 16bytes of mode field to 0664 */ mode = (mode & S_IFMT) | 00664; - err = vfs_create(d_inode(lower_parent_dentry), lower_dentry, mode, want_excl); + + /* temporarily change umask for lower fs write */ + saved_fs = current->fs; + copied_fs = copy_fs_struct(current->fs); + if (!copied_fs) { + err = -ENOMEM; + goto out_unlock; + } + current->fs = copied_fs; + current->fs->umask = 0; + err = vfs_create2(lower_dentry_mnt, d_inode(lower_parent_dentry), lower_dentry, mode, want_excl); if (err) goto out; @@ -83,8 +103,12 @@ static int sdcardfs_create(struct inode *dir, struct dentry *dentry, goto out; fsstack_copy_attr_times(dir, sdcardfs_lower_inode(dir)); fsstack_copy_inode_size(dir, d_inode(lower_parent_dentry)); + fixup_lower_ownership(dentry, dentry->d_name.name); out: + current->fs = saved_fs; + free_fs_struct(copied_fs); +out_unlock: unlock_dir(lower_parent_dentry); sdcardfs_put_lower_path(dentry, &lower_path); REVERT_CRED(saved_cred); @@ -138,12 +162,13 @@ static int sdcardfs_unlink(struct inode *dir, struct dentry *dentry) { int err; struct dentry *lower_dentry; + struct vfsmount *lower_mnt; struct inode *lower_dir_inode = sdcardfs_lower_inode(dir); struct dentry *lower_dir_dentry; struct path lower_path; const struct cred *saved_cred = NULL; - if(!check_caller_access_to_name(dir, dentry->d_name.name)) { + if(!check_caller_access_to_name(dir, &dentry->d_name)) { printk(KERN_INFO "%s: need to check the caller's gid in packages.list\n" " dentry: %s, task:%s\n", __func__, dentry->d_name.name, current->comm); @@ -152,14 +177,15 @@ static int sdcardfs_unlink(struct inode *dir, struct dentry *dentry) } /* save current_cred and override it */ - OVERRIDE_CRED(SDCARDFS_SB(dir->i_sb), saved_cred); + OVERRIDE_CRED(SDCARDFS_SB(dir->i_sb), saved_cred, SDCARDFS_I(dir)); sdcardfs_get_lower_path(dentry, &lower_path); lower_dentry = lower_path.dentry; + lower_mnt = lower_path.mnt; dget(lower_dentry); lower_dir_dentry = lock_parent(lower_dentry); - err = vfs_unlink(lower_dir_inode, lower_dentry, NULL); + err = vfs_unlink2(lower_mnt, lower_dir_inode, lower_dentry, NULL); /* * Note: unlinking on top of NFS can cause silly-renamed files. @@ -240,18 +266,19 @@ static int sdcardfs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode int err; int make_nomedia_in_obb = 0; struct dentry *lower_dentry; + struct vfsmount *lower_mnt; struct dentry *lower_parent_dentry = NULL; struct path lower_path; struct sdcardfs_sb_info *sbi = SDCARDFS_SB(dentry->d_sb); const struct cred *saved_cred = NULL; struct sdcardfs_inode_info *pi = SDCARDFS_I(dir); - char *page_buf; - char *nomedia_dir_name; - char *nomedia_fullpath; - int fullpath_namelen; int touch_err = 0; + struct fs_struct *saved_fs; + struct fs_struct *copied_fs; + struct qstr q_obb = QSTR_LITERAL("obb"); + struct qstr q_data = QSTR_LITERAL("data"); - if(!check_caller_access_to_name(dir, dentry->d_name.name)) { + if(!check_caller_access_to_name(dir, &dentry->d_name)) { printk(KERN_INFO "%s: need to check the caller's gid in packages.list\n" " dentry: %s, task:%s\n", __func__, dentry->d_name.name, current->comm); @@ -260,7 +287,7 @@ static int sdcardfs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode } /* save current_cred and override it */ - OVERRIDE_CRED(SDCARDFS_SB(dir->i_sb), saved_cred); + OVERRIDE_CRED(SDCARDFS_SB(dir->i_sb), saved_cred, SDCARDFS_I(dir)); /* check disk space */ if (!check_min_free_space(dentry, 0, 1)) { @@ -272,14 +299,28 @@ static int sdcardfs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode /* the lower_dentry is negative here */ sdcardfs_get_lower_path(dentry, &lower_path); lower_dentry = lower_path.dentry; + lower_mnt = lower_path.mnt; lower_parent_dentry = lock_parent(lower_dentry); /* set last 16bytes of mode field to 0775 */ mode = (mode & S_IFMT) | 00775; - err = vfs_mkdir(d_inode(lower_parent_dentry), lower_dentry, mode); - if (err) + /* temporarily change umask for lower fs write */ + saved_fs = current->fs; + copied_fs = copy_fs_struct(current->fs); + if (!copied_fs) { + err = -ENOMEM; + unlock_dir(lower_parent_dentry); + goto out_unlock; + } + current->fs = copied_fs; + current->fs->umask = 0; + err = vfs_mkdir2(lower_mnt, d_inode(lower_parent_dentry), lower_dentry, mode); + + if (err) { + unlock_dir(lower_parent_dentry); goto out; + } /* if it is a local obb dentry, setup it with the base obbpath */ if(need_graft_path(dentry)) { @@ -301,58 +342,38 @@ static int sdcardfs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode } err = sdcardfs_interpose(dentry, dir->i_sb, &lower_path, pi->userid); - if (err) + if (err) { + unlock_dir(lower_parent_dentry); goto out; + } fsstack_copy_attr_times(dir, sdcardfs_lower_inode(dir)); fsstack_copy_inode_size(dir, d_inode(lower_parent_dentry)); /* update number of links on parent directory */ set_nlink(dir, sdcardfs_lower_inode(dir)->i_nlink); - - if ((!sbi->options.multiuser) && (!strcasecmp(dentry->d_name.name, "obb")) + fixup_lower_ownership(dentry, dentry->d_name.name); + unlock_dir(lower_parent_dentry); + if ((!sbi->options.multiuser) && (qstr_case_eq(&dentry->d_name, &q_obb)) && (pi->perm == PERM_ANDROID) && (pi->userid == 0)) make_nomedia_in_obb = 1; /* When creating /Android/data and /Android/obb, mark them as .nomedia */ if (make_nomedia_in_obb || - ((pi->perm == PERM_ANDROID) && (!strcasecmp(dentry->d_name.name, "data")))) { - - page_buf = (char *)__get_free_page(GFP_KERNEL); - if (!page_buf) { - printk(KERN_ERR "sdcardfs: failed to allocate page buf\n"); - goto out; - } - - nomedia_dir_name = d_absolute_path(&lower_path, page_buf, PAGE_SIZE); - if (IS_ERR(nomedia_dir_name)) { - free_page((unsigned long)page_buf); - printk(KERN_ERR "sdcardfs: failed to get .nomedia dir name\n"); - goto out; - } - - fullpath_namelen = page_buf + PAGE_SIZE - nomedia_dir_name - 1; - fullpath_namelen += strlen("/.nomedia"); - nomedia_fullpath = kzalloc(fullpath_namelen + 1, GFP_KERNEL); - if (!nomedia_fullpath) { - free_page((unsigned long)page_buf); - printk(KERN_ERR "sdcardfs: failed to allocate .nomedia fullpath buf\n"); - goto out; - } - - strcpy(nomedia_fullpath, nomedia_dir_name); - free_page((unsigned long)page_buf); - strcat(nomedia_fullpath, "/.nomedia"); - touch_err = touch(nomedia_fullpath, 0664); + ((pi->perm == PERM_ANDROID) && (qstr_case_eq(&dentry->d_name, &q_data)))) { + REVERT_CRED(saved_cred); + OVERRIDE_CRED(SDCARDFS_SB(dir->i_sb), saved_cred, SDCARDFS_I(d_inode(dentry))); + set_fs_pwd(current->fs, &lower_path); + touch_err = touch(".nomedia", 0664); if (touch_err) { - printk(KERN_ERR "sdcardfs: failed to touch(%s): %d\n", - nomedia_fullpath, touch_err); - kfree(nomedia_fullpath); + printk(KERN_ERR "sdcardfs: failed to create .nomedia in %s: %d\n", + lower_path.dentry->d_name.name, touch_err); goto out; } - kfree(nomedia_fullpath); } out: - unlock_dir(lower_parent_dentry); + current->fs = saved_fs; + free_fs_struct(copied_fs); +out_unlock: sdcardfs_put_lower_path(dentry, &lower_path); out_revert: REVERT_CRED(saved_cred); @@ -364,11 +385,12 @@ static int sdcardfs_rmdir(struct inode *dir, struct dentry *dentry) { struct dentry *lower_dentry; struct dentry *lower_dir_dentry; + struct vfsmount *lower_mnt; int err; struct path lower_path; const struct cred *saved_cred = NULL; - if(!check_caller_access_to_name(dir, dentry->d_name.name)) { + if(!check_caller_access_to_name(dir, &dentry->d_name)) { printk(KERN_INFO "%s: need to check the caller's gid in packages.list\n" " dentry: %s, task:%s\n", __func__, dentry->d_name.name, current->comm); @@ -377,16 +399,17 @@ static int sdcardfs_rmdir(struct inode *dir, struct dentry *dentry) } /* save current_cred and override it */ - OVERRIDE_CRED(SDCARDFS_SB(dir->i_sb), saved_cred); + OVERRIDE_CRED(SDCARDFS_SB(dir->i_sb), saved_cred, SDCARDFS_I(dir)); /* sdcardfs_get_real_lower(): in case of remove an user's obb dentry * the dentry on the original path should be deleted. */ sdcardfs_get_real_lower(dentry, &lower_path); lower_dentry = lower_path.dentry; + lower_mnt = lower_path.mnt; lower_dir_dentry = lock_parent(lower_dentry); - err = vfs_rmdir(d_inode(lower_dir_dentry), lower_dentry); + err = vfs_rmdir2(lower_mnt, d_inode(lower_dir_dentry), lower_dentry); if (err) goto out; @@ -450,13 +473,13 @@ static int sdcardfs_rename(struct inode *old_dir, struct dentry *old_dentry, struct dentry *lower_new_dentry = NULL; struct dentry *lower_old_dir_dentry = NULL; struct dentry *lower_new_dir_dentry = NULL; + struct vfsmount *lower_mnt = NULL; struct dentry *trap = NULL; - struct dentry *new_parent = NULL; struct path lower_old_path, lower_new_path; const struct cred *saved_cred = NULL; - if(!check_caller_access_to_name(old_dir, old_dentry->d_name.name) || - !check_caller_access_to_name(new_dir, new_dentry->d_name.name)) { + if(!check_caller_access_to_name(old_dir, &old_dentry->d_name) || + !check_caller_access_to_name(new_dir, &new_dentry->d_name)) { printk(KERN_INFO "%s: need to check the caller's gid in packages.list\n" " new_dentry: %s, task:%s\n", __func__, new_dentry->d_name.name, current->comm); @@ -465,12 +488,13 @@ static int sdcardfs_rename(struct inode *old_dir, struct dentry *old_dentry, } /* save current_cred and override it */ - OVERRIDE_CRED(SDCARDFS_SB(old_dir->i_sb), saved_cred); + OVERRIDE_CRED(SDCARDFS_SB(old_dir->i_sb), saved_cred, SDCARDFS_I(new_dir)); sdcardfs_get_real_lower(old_dentry, &lower_old_path); sdcardfs_get_lower_path(new_dentry, &lower_new_path); lower_old_dentry = lower_old_path.dentry; lower_new_dentry = lower_new_path.dentry; + lower_mnt = lower_old_path.mnt; lower_old_dir_dentry = dget_parent(lower_old_dentry); lower_new_dir_dentry = dget_parent(lower_new_dentry); @@ -486,7 +510,8 @@ static int sdcardfs_rename(struct inode *old_dir, struct dentry *old_dentry, goto out; } - err = vfs_rename(d_inode(lower_old_dir_dentry), lower_old_dentry, + err = vfs_rename2(lower_mnt, + d_inode(lower_old_dir_dentry), lower_old_dentry, d_inode(lower_new_dir_dentry), lower_new_dentry, NULL, 0); if (err) @@ -499,25 +524,11 @@ static int sdcardfs_rename(struct inode *old_dir, struct dentry *old_dentry, if (new_dir != old_dir) { sdcardfs_copy_and_fix_attrs(old_dir, d_inode(lower_old_dir_dentry)); fsstack_copy_inode_size(old_dir, d_inode(lower_old_dir_dentry)); - - /* update the derived permission of the old_dentry - * with its new parent - */ - new_parent = dget_parent(new_dentry); - if(new_parent) { - if(d_inode(old_dentry)) { - update_derived_permission_lock(old_dentry); - } - dput(new_parent); - } } - /* At this point, not all dentry information has been moved, so - * we pass along new_dentry for the name.*/ - mutex_lock(&d_inode(old_dentry)->i_mutex); - get_derived_permission_new(new_dentry->d_parent, old_dentry, new_dentry); - fix_derived_permission(d_inode(old_dentry)); - get_derive_permissions_recursive(old_dentry); - mutex_unlock(&d_inode(old_dentry)->i_mutex); + get_derived_permission_new(new_dentry->d_parent, old_dentry, &new_dentry->d_name); + fixup_tmp_permissions(d_inode(old_dentry)); + fixup_lower_ownership(old_dentry, new_dentry->d_name.name); + drop_recursive(old_dentry); /* Can't fixup ownership recursively :( */ out: unlock_rename(lower_old_dir_dentry, lower_new_dir_dentry); dput(lower_old_dir_dentry); @@ -586,16 +597,63 @@ static const char *sdcardfs_follow_link(struct dentry *dentry, void **cookie) } #endif -static int sdcardfs_permission(struct inode *inode, int mask) +static int sdcardfs_permission_wrn(struct inode *inode, int mask) +{ + WARN(1, "sdcardfs does not support permission. Use permission2.\n"); + return -EINVAL; +} + +void copy_attrs(struct inode *dest, const struct inode *src) +{ + dest->i_mode = src->i_mode; + dest->i_uid = src->i_uid; + dest->i_gid = src->i_gid; + dest->i_rdev = src->i_rdev; + dest->i_atime = src->i_atime; + dest->i_mtime = src->i_mtime; + dest->i_ctime = src->i_ctime; + dest->i_blkbits = src->i_blkbits; + dest->i_flags = src->i_flags; +#ifdef CONFIG_FS_POSIX_ACL + dest->i_acl = src->i_acl; +#endif +#ifdef CONFIG_SECURITY + dest->i_security = src->i_security; +#endif +} + +static int sdcardfs_permission(struct vfsmount *mnt, struct inode *inode, int mask) { int err; + struct inode tmp; + struct inode *top = grab_top(SDCARDFS_I(inode)); + + if (!top) { + release_top(SDCARDFS_I(inode)); + WARN(1, "Top value was null!\n"); + return -EINVAL; + } /* * Permission check on sdcardfs inode. * Calling process should have AID_SDCARD_RW permission + * Since generic_permission only needs i_mode, i_uid, + * i_gid, and i_sb, we can create a fake inode to pass + * this information down in. + * + * The underlying code may attempt to take locks in some + * cases for features we're not using, but if that changes, + * locks must be dealt with to avoid undefined behavior. */ - err = generic_permission(inode, mask); - + copy_attrs(&tmp, inode); + tmp.i_uid = make_kuid(&init_user_ns, SDCARDFS_I(top)->d_uid); + tmp.i_gid = make_kgid(&init_user_ns, get_gid(mnt, SDCARDFS_I(top))); + tmp.i_mode = (inode->i_mode & S_IFMT) | get_mode(mnt, SDCARDFS_I(top)); + release_top(SDCARDFS_I(inode)); + tmp.i_sb = inode->i_sb; + if (IS_POSIXACL(inode)) + printk(KERN_WARNING "%s: This may be undefined behavior... \n", __func__); + err = generic_permission(&tmp, mask); /* XXX * Original sdcardfs code calls inode_permission(lower_inode,.. ) * for checking inode permission. But doing such things here seems @@ -624,30 +682,70 @@ static int sdcardfs_permission(struct inode *inode, int mask) } -static int sdcardfs_setattr(struct dentry *dentry, struct iattr *ia) +static int sdcardfs_setattr_wrn(struct dentry *dentry, struct iattr *ia) +{ + WARN(1, "sdcardfs does not support setattr. User setattr2.\n"); + return -EINVAL; +} + +static int sdcardfs_setattr(struct vfsmount *mnt, struct dentry *dentry, struct iattr *ia) { int err; struct dentry *lower_dentry; + struct vfsmount *lower_mnt; struct inode *inode; struct inode *lower_inode; struct path lower_path; struct iattr lower_ia; struct dentry *parent; + struct inode tmp; + struct inode *top; + const struct cred *saved_cred = NULL; inode = d_inode(dentry); + top = grab_top(SDCARDFS_I(inode)); + + if (!top) { + release_top(SDCARDFS_I(inode)); + return -EINVAL; + } + + /* + * Permission check on sdcardfs inode. + * Calling process should have AID_SDCARD_RW permission + * Since generic_permission only needs i_mode, i_uid, + * i_gid, and i_sb, we can create a fake inode to pass + * this information down in. + * + * The underlying code may attempt to take locks in some + * cases for features we're not using, but if that changes, + * locks must be dealt with to avoid undefined behavior. + * + */ + copy_attrs(&tmp, inode); + tmp.i_uid = make_kuid(&init_user_ns, SDCARDFS_I(top)->d_uid); + tmp.i_gid = make_kgid(&init_user_ns, get_gid(mnt, SDCARDFS_I(top))); + tmp.i_mode = (inode->i_mode & S_IFMT) | get_mode(mnt, SDCARDFS_I(top)); + tmp.i_size = i_size_read(inode); + release_top(SDCARDFS_I(inode)); + tmp.i_sb = inode->i_sb; /* * Check if user has permission to change inode. We don't check if * this user can change the lower inode: that should happen when * calling notify_change on the lower inode. */ - err = inode_change_ok(inode, ia); + /* prepare our own lower struct iattr (with the lower file) */ + memcpy(&lower_ia, ia, sizeof(lower_ia)); + /* Allow touch updating timestamps. A previous permission check ensures + * we have write access. Changes to mode, owner, and group are ignored*/ + ia->ia_valid |= ATTR_FORCE; + err = inode_change_ok(&tmp, ia); - /* no vfs_XXX operations required, cred overriding will be skipped. wj*/ if (!err) { /* check the Android group ID */ parent = dget_parent(dentry); - if(!check_caller_access_to_name(d_inode(parent), dentry->d_name.name)) { + if(!check_caller_access_to_name(d_inode(parent), &dentry->d_name)) { printk(KERN_INFO "%s: need to check the caller's gid in packages.list\n" " dentry: %s, task:%s\n", __func__, dentry->d_name.name, current->comm); @@ -659,12 +757,14 @@ static int sdcardfs_setattr(struct dentry *dentry, struct iattr *ia) if (err) goto out_err; + /* save current_cred and override it */ + OVERRIDE_CRED(SDCARDFS_SB(dentry->d_sb), saved_cred, SDCARDFS_I(inode)); + sdcardfs_get_lower_path(dentry, &lower_path); lower_dentry = lower_path.dentry; + lower_mnt = lower_path.mnt; lower_inode = sdcardfs_lower_inode(inode); - /* prepare our own lower struct iattr (with the lower file) */ - memcpy(&lower_ia, ia, sizeof(lower_ia)); if (ia->ia_valid & ATTR_FILE) lower_ia.ia_file = sdcardfs_lower_file(ia->ia_file); @@ -681,7 +781,7 @@ static int sdcardfs_setattr(struct dentry *dentry, struct iattr *ia) if (current->mm) down_write(¤t->mm->mmap_sem); if (ia->ia_valid & ATTR_SIZE) { - err = inode_newsize_ok(inode, ia->ia_size); + err = inode_newsize_ok(&tmp, ia->ia_size); if (err) { if (current->mm) up_write(¤t->mm->mmap_sem); @@ -704,7 +804,7 @@ static int sdcardfs_setattr(struct dentry *dentry, struct iattr *ia) * tries to open(), unlink(), then ftruncate() a file. */ mutex_lock(&d_inode(lower_dentry)->i_mutex); - err = notify_change(lower_dentry, &lower_ia, /* note: lower_ia */ + err = notify_change2(lower_mnt, lower_dentry, &lower_ia, /* note: lower_ia */ NULL); mutex_unlock(&d_inode(lower_dentry)->i_mutex); if (current->mm) @@ -723,10 +823,35 @@ static int sdcardfs_setattr(struct dentry *dentry, struct iattr *ia) out: sdcardfs_put_lower_path(dentry, &lower_path); + REVERT_CRED(saved_cred); out_err: return err; } +static int sdcardfs_fillattr(struct vfsmount *mnt, struct inode *inode, struct kstat *stat) +{ + struct sdcardfs_inode_info *info = SDCARDFS_I(inode); + struct inode *top = grab_top(info); + if (!top) + return -EINVAL; + + stat->dev = inode->i_sb->s_dev; + stat->ino = inode->i_ino; + stat->mode = (inode->i_mode & S_IFMT) | get_mode(mnt, SDCARDFS_I(top)); + stat->nlink = inode->i_nlink; + stat->uid = make_kuid(&init_user_ns, SDCARDFS_I(top)->d_uid); + stat->gid = make_kgid(&init_user_ns, get_gid(mnt, SDCARDFS_I(top))); + stat->rdev = inode->i_rdev; + stat->size = i_size_read(inode); + stat->atime = inode->i_atime; + stat->mtime = inode->i_mtime; + stat->ctime = inode->i_ctime; + stat->blksize = (1 << inode->i_blkbits); + stat->blocks = inode->i_blocks; + release_top(info); + return 0; +} + static int sdcardfs_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat) { @@ -735,9 +860,10 @@ static int sdcardfs_getattr(struct vfsmount *mnt, struct dentry *dentry, struct inode *lower_inode; struct path lower_path; struct dentry *parent; + int err; parent = dget_parent(dentry); - if(!check_caller_access_to_name(d_inode(parent), dentry->d_name.name)) { + if(!check_caller_access_to_name(d_inode(parent), &dentry->d_name)) { printk(KERN_INFO "%s: need to check the caller's gid in packages.list\n" " dentry: %s, task:%s\n", __func__, dentry->d_name.name, current->comm); @@ -752,19 +878,17 @@ static int sdcardfs_getattr(struct vfsmount *mnt, struct dentry *dentry, lower_dentry = lower_path.dentry; lower_inode = sdcardfs_lower_inode(inode); - sdcardfs_copy_and_fix_attrs(inode, lower_inode); fsstack_copy_inode_size(inode, lower_inode); - - generic_fillattr(inode, stat); + err = sdcardfs_fillattr(mnt, inode, stat); sdcardfs_put_lower_path(dentry, &lower_path); - return 0; + return err; } const struct inode_operations sdcardfs_symlink_iops = { - .permission = sdcardfs_permission, - .setattr = sdcardfs_setattr, + .permission2 = sdcardfs_permission, + .setattr2 = sdcardfs_setattr, /* XXX Following operations are implemented, * but FUSE(sdcard) or FAT does not support them * These methods are *NOT* perfectly tested. @@ -777,14 +901,14 @@ const struct inode_operations sdcardfs_symlink_iops = { const struct inode_operations sdcardfs_dir_iops = { .create = sdcardfs_create, .lookup = sdcardfs_lookup, -#if 0 - .permission = sdcardfs_permission, -#endif + .permission = sdcardfs_permission_wrn, + .permission2 = sdcardfs_permission, .unlink = sdcardfs_unlink, .mkdir = sdcardfs_mkdir, .rmdir = sdcardfs_rmdir, .rename = sdcardfs_rename, - .setattr = sdcardfs_setattr, + .setattr = sdcardfs_setattr_wrn, + .setattr2 = sdcardfs_setattr, .getattr = sdcardfs_getattr, /* XXX Following operations are implemented, * but FUSE(sdcard) or FAT does not support them @@ -796,7 +920,9 @@ const struct inode_operations sdcardfs_dir_iops = { }; const struct inode_operations sdcardfs_main_iops = { - .permission = sdcardfs_permission, - .setattr = sdcardfs_setattr, + .permission = sdcardfs_permission_wrn, + .permission2 = sdcardfs_permission, + .setattr = sdcardfs_setattr_wrn, + .setattr2 = sdcardfs_setattr, .getattr = sdcardfs_getattr, }; diff --git a/fs/sdcardfs/lookup.c b/fs/sdcardfs/lookup.c index a01b06a514fd..6b595e892316 100644 --- a/fs/sdcardfs/lookup.c +++ b/fs/sdcardfs/lookup.c @@ -179,7 +179,7 @@ int sdcardfs_interpose(struct dentry *dentry, struct super_block *sb, struct inode *lower_inode; struct super_block *lower_sb; - lower_inode = lower_path->dentry->d_inode; + lower_inode = d_inode(lower_path->dentry); lower_sb = sdcardfs_lower_super(sb); /* check that the lower file system didn't cross a mount point */ @@ -219,9 +219,9 @@ static struct dentry *__sdcardfs_lookup(struct dentry *dentry, struct vfsmount *lower_dir_mnt; struct dentry *lower_dir_dentry = NULL; struct dentry *lower_dentry; - const char *name; + const struct qstr *name; struct path lower_path; - struct qstr this; + struct qstr dname; struct sdcardfs_sb_info *sbi; sbi = SDCARDFS_SB(dentry->d_sb); @@ -231,15 +231,39 @@ static struct dentry *__sdcardfs_lookup(struct dentry *dentry, if (IS_ROOT(dentry)) goto out; - name = dentry->d_name.name; + name = &dentry->d_name; /* now start the actual lookup procedure */ lower_dir_dentry = lower_parent_path->dentry; lower_dir_mnt = lower_parent_path->mnt; /* Use vfs_path_lookup to check if the dentry exists or not */ - err = vfs_path_lookup(lower_dir_dentry, lower_dir_mnt, name, 0, + err = vfs_path_lookup(lower_dir_dentry, lower_dir_mnt, name->name, 0, &lower_path); + /* check for other cases */ + if (err == -ENOENT) { + struct dentry *child; + struct dentry *match = NULL; + mutex_lock(&d_inode(lower_dir_dentry)->i_mutex); + spin_lock(&lower_dir_dentry->d_lock); + list_for_each_entry(child, &lower_dir_dentry->d_subdirs, d_child) { + if (child && d_inode(child)) { + if (qstr_case_eq(&child->d_name, name)) { + match = dget(child); + break; + } + } + } + spin_unlock(&lower_dir_dentry->d_lock); + mutex_unlock(&d_inode(lower_dir_dentry)->i_mutex); + if (match) { + err = vfs_path_lookup(lower_dir_dentry, + lower_dir_mnt, + match->d_name.name, 0, + &lower_path); + dput(match); + } + } /* no error: handle positive dentries */ if (!err) { @@ -283,14 +307,14 @@ static struct dentry *__sdcardfs_lookup(struct dentry *dentry, goto out; /* instatiate a new negative dentry */ - this.name = name; - this.len = strlen(name); - this.hash = full_name_hash(this.name, this.len); - lower_dentry = d_lookup(lower_dir_dentry, &this); + dname.name = name->name; + dname.len = name->len; + dname.hash = full_name_hash(dname.name, dname.len); + lower_dentry = d_lookup(lower_dir_dentry, &dname); if (lower_dentry) goto setup_lower; - lower_dentry = d_alloc(lower_dir_dentry, &this); + lower_dentry = d_alloc(lower_dir_dentry, &dname); if (!lower_dentry) { err = -ENOMEM; goto out; @@ -335,7 +359,7 @@ struct dentry *sdcardfs_lookup(struct inode *dir, struct dentry *dentry, parent = dget_parent(dentry); - if(!check_caller_access_to_name(parent->d_inode, dentry->d_name.name)) { + if(!check_caller_access_to_name(d_inode(parent), &dentry->d_name)) { ret = ERR_PTR(-EACCES); printk(KERN_INFO "%s: need to check the caller's gid in packages.list\n" " dentry: %s, task:%s\n", @@ -344,7 +368,7 @@ struct dentry *sdcardfs_lookup(struct inode *dir, struct dentry *dentry, } /* save current_cred and override it */ - OVERRIDE_CRED_PTR(SDCARDFS_SB(dir->i_sb), saved_cred); + OVERRIDE_CRED_PTR(SDCARDFS_SB(dir->i_sb), saved_cred, SDCARDFS_I(dir)); sdcardfs_get_lower_path(parent, &lower_parent_path); @@ -362,18 +386,17 @@ struct dentry *sdcardfs_lookup(struct inode *dir, struct dentry *dentry, } if (ret) dentry = ret; - if (dentry->d_inode) { - fsstack_copy_attr_times(dentry->d_inode, - sdcardfs_lower_inode(dentry->d_inode)); - /* get drived permission */ - mutex_lock(&dentry->d_inode->i_mutex); + if (d_inode(dentry)) { + fsstack_copy_attr_times(d_inode(dentry), + sdcardfs_lower_inode(d_inode(dentry))); + /* get derived permission */ get_derived_permission(parent, dentry); - fix_derived_permission(dentry->d_inode); - mutex_unlock(&dentry->d_inode->i_mutex); + fixup_tmp_permissions(d_inode(dentry)); + fixup_lower_ownership(dentry, dentry->d_name.name); } /* update parent directory's atime */ - fsstack_copy_attr_atime(parent->d_inode, - sdcardfs_lower_inode(parent->d_inode)); + fsstack_copy_attr_atime(d_inode(parent), + sdcardfs_lower_inode(d_inode(parent))); out: sdcardfs_put_lower_path(parent, &lower_parent_path); diff --git a/fs/sdcardfs/main.c b/fs/sdcardfs/main.c index a6522286d731..7a8eae29e44d 100644 --- a/fs/sdcardfs/main.c +++ b/fs/sdcardfs/main.c @@ -28,7 +28,6 @@ enum { Opt_fsgid, Opt_gid, Opt_debug, - Opt_lower_fs, Opt_mask, Opt_multiuser, // May need? Opt_userid, @@ -49,7 +48,8 @@ static const match_table_t sdcardfs_tokens = { }; static int parse_options(struct super_block *sb, char *options, int silent, - int *debug, struct sdcardfs_mount_options *opts) + int *debug, struct sdcardfs_vfsmount_options *vfsopts, + struct sdcardfs_mount_options *opts) { char *p; substring_t args[MAX_OPT_ARGS]; @@ -58,10 +58,10 @@ static int parse_options(struct super_block *sb, char *options, int silent, /* by default, we use AID_MEDIA_RW as uid, gid */ opts->fs_low_uid = AID_MEDIA_RW; opts->fs_low_gid = AID_MEDIA_RW; - opts->mask = 0; + vfsopts->mask = 0; opts->multiuser = false; opts->fs_user_id = 0; - opts->gid = 0; + vfsopts->gid = 0; /* by default, 0MB is reserved */ opts->reserved_mb = 0; @@ -94,7 +94,7 @@ static int parse_options(struct super_block *sb, char *options, int silent, case Opt_gid: if (match_int(&args[0], &option)) return 0; - opts->gid = option; + vfsopts->gid = option; break; case Opt_userid: if (match_int(&args[0], &option)) @@ -104,7 +104,7 @@ static int parse_options(struct super_block *sb, char *options, int silent, case Opt_mask: if (match_int(&args[0], &option)) return 0; - opts->mask = option; + vfsopts->mask = option; break; case Opt_multiuser: opts->multiuser = true; @@ -135,6 +135,65 @@ static int parse_options(struct super_block *sb, char *options, int silent, return 0; } +int parse_options_remount(struct super_block *sb, char *options, int silent, + struct sdcardfs_vfsmount_options *vfsopts) +{ + char *p; + substring_t args[MAX_OPT_ARGS]; + int option; + int debug; + + if (!options) + return 0; + + while ((p = strsep(&options, ",")) != NULL) { + int token; + if (!*p) + continue; + + token = match_token(p, sdcardfs_tokens, args); + + switch (token) { + case Opt_debug: + debug = 1; + break; + case Opt_gid: + if (match_int(&args[0], &option)) + return 0; + vfsopts->gid = option; + + break; + case Opt_mask: + if (match_int(&args[0], &option)) + return 0; + vfsopts->mask = option; + break; + case Opt_multiuser: + case Opt_userid: + case Opt_fsuid: + case Opt_fsgid: + case Opt_reserved_mb: + printk( KERN_WARNING "Option \"%s\" can't be changed during remount\n", p); + break; + /* unknown option */ + default: + if (!silent) { + printk( KERN_ERR "Unrecognized mount option \"%s\" " + "or missing value", p); + } + return -EINVAL; + } + } + + if (debug) { + printk( KERN_INFO "sdcardfs : options - debug:%d\n", debug); + printk( KERN_INFO "sdcardfs : options - gid:%d\n", vfsopts->gid); + printk( KERN_INFO "sdcardfs : options - mask:%d\n", vfsopts->mask); + } + + return 0; +} + #if 0 /* * our custom d_alloc_root work-alike @@ -172,14 +231,15 @@ EXPORT_SYMBOL_GPL(sdcardfs_super_list); * There is no need to lock the sdcardfs_super_info's rwsem as there is no * way anyone can have a reference to the superblock at this point in time. */ -static int sdcardfs_read_super(struct super_block *sb, const char *dev_name, - void *raw_data, int silent) +static int sdcardfs_read_super(struct vfsmount *mnt, struct super_block *sb, + const char *dev_name, void *raw_data, int silent) { int err = 0; int debug; struct super_block *lower_sb; struct path lower_path; struct sdcardfs_sb_info *sb_info; + struct sdcardfs_vfsmount_options *mnt_opt = mnt->data; struct inode *inode; printk(KERN_INFO "sdcardfs version 2.0\n"); @@ -193,6 +253,7 @@ static int sdcardfs_read_super(struct super_block *sb, const char *dev_name, printk(KERN_INFO "sdcardfs: dev_name -> %s\n", dev_name); printk(KERN_INFO "sdcardfs: options -> %s\n", (char *)raw_data); + printk(KERN_INFO "sdcardfs: mnt -> %p\n", mnt); /* parse lower path */ err = kern_path(dev_name, LOOKUP_FOLLOW | LOOKUP_DIRECTORY, @@ -212,7 +273,7 @@ static int sdcardfs_read_super(struct super_block *sb, const char *dev_name, sb_info = sb->s_fs_info; /* parse options */ - err = parse_options(sb, raw_data, silent, &debug, &sb_info->options); + err = parse_options(sb, raw_data, silent, &debug, mnt_opt, &sb_info->options); if (err) { printk(KERN_ERR "sdcardfs: invalid options\n"); goto out_freesbi; @@ -236,7 +297,7 @@ static int sdcardfs_read_super(struct super_block *sb, const char *dev_name, sb->s_op = &sdcardfs_sops; /* get a new inode and allocate our root dentry */ - inode = sdcardfs_iget(sb, lower_path.dentry->d_inode, 0); + inode = sdcardfs_iget(sb, d_inode(lower_path.dentry), 0); if (IS_ERR(inode)) { err = PTR_ERR(inode); goto out_sput; @@ -268,16 +329,16 @@ static int sdcardfs_read_super(struct super_block *sb, const char *dev_name, sb_info->obbpath_s = kzalloc(PATH_MAX, GFP_KERNEL); mutex_lock(&sdcardfs_super_list_lock); if(sb_info->options.multiuser) { - setup_derived_state(sb->s_root->d_inode, PERM_PRE_ROOT, sb_info->options.fs_user_id, AID_ROOT, false); + setup_derived_state(d_inode(sb->s_root), PERM_PRE_ROOT, sb_info->options.fs_user_id, AID_ROOT, false, d_inode(sb->s_root)); snprintf(sb_info->obbpath_s, PATH_MAX, "%s/obb", dev_name); /*err = prepare_dir(sb_info->obbpath_s, sb_info->options.fs_low_uid, sb_info->options.fs_low_gid, 00755);*/ } else { - setup_derived_state(sb->s_root->d_inode, PERM_ROOT, sb_info->options.fs_low_uid, AID_ROOT, false); + setup_derived_state(d_inode(sb->s_root), PERM_ROOT, sb_info->options.fs_user_id, AID_ROOT, false, d_inode(sb->s_root)); snprintf(sb_info->obbpath_s, PATH_MAX, "%s/Android/obb", dev_name); } - fix_derived_permission(sb->s_root->d_inode); + fixup_tmp_permissions(d_inode(sb->s_root)); sb_info->sb = sb; list_add(&sb_info->list, &sdcardfs_super_list); mutex_unlock(&sdcardfs_super_list_lock); @@ -306,9 +367,9 @@ out: } /* A feature which supports mount_nodev() with options */ -static struct dentry *mount_nodev_with_options(struct file_system_type *fs_type, - int flags, const char *dev_name, void *data, - int (*fill_super)(struct super_block *, const char *, void *, int)) +static struct dentry *mount_nodev_with_options(struct vfsmount *mnt, + struct file_system_type *fs_type, int flags, const char *dev_name, void *data, + int (*fill_super)(struct vfsmount *, struct super_block *, const char *, void *, int)) { int error; @@ -319,7 +380,7 @@ static struct dentry *mount_nodev_with_options(struct file_system_type *fs_type, s->s_flags = flags; - error = fill_super(s, dev_name, data, flags & MS_SILENT ? 1 : 0); + error = fill_super(mnt, s, dev_name, data, flags & MS_SILENT ? 1 : 0); if (error) { deactivate_locked_super(s); return ERR_PTR(error); @@ -328,15 +389,27 @@ static struct dentry *mount_nodev_with_options(struct file_system_type *fs_type, return dget(s->s_root); } -struct dentry *sdcardfs_mount(struct file_system_type *fs_type, int flags, +static struct dentry *sdcardfs_mount(struct vfsmount *mnt, + struct file_system_type *fs_type, int flags, const char *dev_name, void *raw_data) { /* * dev_name is a lower_path_name, * raw_data is a option string. */ - return mount_nodev_with_options(fs_type, flags, dev_name, - raw_data, sdcardfs_read_super); + return mount_nodev_with_options(mnt, fs_type, flags, dev_name, + raw_data, sdcardfs_read_super); +} + +static struct dentry *sdcardfs_mount_wrn(struct file_system_type *fs_type, int flags, + const char *dev_name, void *raw_data) +{ + WARN(1, "sdcardfs does not support mount. Use mount2.\n"); + return ERR_PTR(-EINVAL); +} + +void *sdcardfs_alloc_mnt_data(void) { + return kmalloc(sizeof(struct sdcardfs_vfsmount_options), GFP_KERNEL); } void sdcardfs_kill_sb(struct super_block *sb) { @@ -353,7 +426,9 @@ void sdcardfs_kill_sb(struct super_block *sb) { static struct file_system_type sdcardfs_fs_type = { .owner = THIS_MODULE, .name = SDCARDFS_NAME, - .mount = sdcardfs_mount, + .mount = sdcardfs_mount_wrn, + .mount2 = sdcardfs_mount, + .alloc_mnt_data = sdcardfs_alloc_mnt_data, .kill_sb = sdcardfs_kill_sb, .fs_flags = 0, }; diff --git a/fs/sdcardfs/multiuser.h b/fs/sdcardfs/multiuser.h index 923ba101dfa9..52bc20080904 100644 --- a/fs/sdcardfs/multiuser.h +++ b/fs/sdcardfs/multiuser.h @@ -18,20 +18,32 @@ * General Public License. */ -#define MULTIUSER_APP_PER_USER_RANGE 100000 +#define AID_USER_OFFSET 100000 /* offset for uid ranges for each user */ +#define AID_APP_START 10000 /* first app user */ +#define AID_APP_END 19999 /* last app user */ +#define AID_CACHE_GID_START 20000 /* start of gids for apps to mark cached data */ +#define AID_EXT_GID_START 30000 /* start of gids for apps to mark external data */ +#define AID_SHARED_GID_START 50000 /* start of gids for apps in each user to share */ typedef uid_t userid_t; typedef uid_t appid_t; -static inline userid_t multiuser_get_user_id(uid_t uid) { - return uid / MULTIUSER_APP_PER_USER_RANGE; +static inline uid_t multiuser_get_uid(userid_t user_id, appid_t app_id) { + return (user_id * AID_USER_OFFSET) + (app_id % AID_USER_OFFSET); } -static inline appid_t multiuser_get_app_id(uid_t uid) { - return uid % MULTIUSER_APP_PER_USER_RANGE; +static inline gid_t multiuser_get_cache_gid(userid_t user_id, appid_t app_id) { + if (app_id >= AID_APP_START && app_id <= AID_APP_END) { + return multiuser_get_uid(user_id, (app_id - AID_APP_START) + AID_CACHE_GID_START); + } else { + return -1; + } } -static inline uid_t multiuser_get_uid(userid_t userId, appid_t appId) { - return userId * MULTIUSER_APP_PER_USER_RANGE + (appId % MULTIUSER_APP_PER_USER_RANGE); +static inline gid_t multiuser_get_ext_gid(userid_t user_id, appid_t app_id) { + if (app_id >= AID_APP_START && app_id <= AID_APP_END) { + return multiuser_get_uid(user_id, (app_id - AID_APP_START) + AID_EXT_GID_START); + } else { + return -1; + } } - diff --git a/fs/sdcardfs/packagelist.c b/fs/sdcardfs/packagelist.c index 9c3340528eee..d96fcde041cc 100644 --- a/fs/sdcardfs/packagelist.c +++ b/fs/sdcardfs/packagelist.c @@ -21,7 +21,8 @@ #include "sdcardfs.h" #include #include - +#include +#include #include #include @@ -29,70 +30,120 @@ #include -#define STRING_BUF_SIZE (512) - struct hashtable_entry { struct hlist_node hlist; - void *key; - unsigned int value; + struct hlist_node dlist; /* for deletion cleanup */ + struct qstr key; + atomic_t value; }; -struct sb_list { - struct super_block *sb; - struct list_head list; -}; +static DEFINE_HASHTABLE(package_to_appid, 8); +static DEFINE_HASHTABLE(package_to_userid, 8); +static DEFINE_HASHTABLE(ext_to_groupid, 8); -struct packagelist_data { - DECLARE_HASHTABLE(package_to_appid,8); - struct mutex hashtable_lock; - -}; - -static struct packagelist_data *pkgl_data_all; static struct kmem_cache *hashtable_entry_cachep; -static unsigned int str_hash(const char *key) { - int i; - unsigned int h = strlen(key); - char *data = (char *)key; - - for (i = 0; i < strlen(key); i++) { - h = h * 31 + *data; - data++; - } - return h; +static void inline qstr_init(struct qstr *q, const char *name) { + q->name = name; + q->len = strlen(q->name); + q->hash = full_name_hash(q->name, q->len); } -appid_t get_appid(void *pkgl_id, const char *app_name) +static inline int qstr_copy(const struct qstr *src, struct qstr *dest) { + dest->name = kstrdup(src->name, GFP_KERNEL); + dest->hash_len = src->hash_len; + return !!dest->name; +} + + +static appid_t __get_appid(const struct qstr *key) { - struct packagelist_data *pkgl_dat = pkgl_data_all; struct hashtable_entry *hash_cur; - unsigned int hash = str_hash(app_name); + unsigned int hash = key->hash; appid_t ret_id; - mutex_lock(&pkgl_dat->hashtable_lock); - hash_for_each_possible(pkgl_dat->package_to_appid, hash_cur, hlist, hash) { - if (!strcasecmp(app_name, hash_cur->key)) { - ret_id = (appid_t)hash_cur->value; - mutex_unlock(&pkgl_dat->hashtable_lock); + rcu_read_lock(); + hash_for_each_possible_rcu(package_to_appid, hash_cur, hlist, hash) { + if (qstr_case_eq(key, &hash_cur->key)) { + ret_id = atomic_read(&hash_cur->value); + rcu_read_unlock(); return ret_id; } } - mutex_unlock(&pkgl_dat->hashtable_lock); + rcu_read_unlock(); return 0; } +appid_t get_appid(const char *key) +{ + struct qstr q; + qstr_init(&q, key); + return __get_appid(&q); +} + +static appid_t __get_ext_gid(const struct qstr *key) +{ + struct hashtable_entry *hash_cur; + unsigned int hash = key->hash; + appid_t ret_id; + + rcu_read_lock(); + hash_for_each_possible_rcu(ext_to_groupid, hash_cur, hlist, hash) { + if (qstr_case_eq(key, &hash_cur->key)) { + ret_id = atomic_read(&hash_cur->value); + rcu_read_unlock(); + return ret_id; + } + } + rcu_read_unlock(); + return 0; +} + +appid_t get_ext_gid(const char *key) +{ + struct qstr q; + qstr_init(&q, key); + return __get_ext_gid(&q); +} + +static appid_t __is_excluded(const struct qstr *app_name, userid_t user) +{ + struct hashtable_entry *hash_cur; + unsigned int hash = app_name->hash; + + rcu_read_lock(); + hash_for_each_possible_rcu(package_to_userid, hash_cur, hlist, hash) { + if (atomic_read(&hash_cur->value) == user && + qstr_case_eq(app_name, &hash_cur->key)) { + rcu_read_unlock(); + return 1; + } + } + rcu_read_unlock(); + return 0; +} + +appid_t is_excluded(const char *key, userid_t user) +{ + struct qstr q; + qstr_init(&q, key); + return __is_excluded(&q, user); +} + /* Kernel has already enforced everything we returned through * derive_permissions_locked(), so this is used to lock down access * even further, such as enforcing that apps hold sdcard_rw. */ -int check_caller_access_to_name(struct inode *parent_node, const char* name) { +int check_caller_access_to_name(struct inode *parent_node, const struct qstr *name) { + struct qstr q_autorun = QSTR_LITERAL("autorun.inf"); + struct qstr q__android_secure = QSTR_LITERAL(".android_secure"); + struct qstr q_android_secure = QSTR_LITERAL("android_secure"); /* Always block security-sensitive files at root */ if (parent_node && SDCARDFS_I(parent_node)->perm == PERM_ROOT) { - if (!strcasecmp(name, "autorun.inf") - || !strcasecmp(name, ".android_secure") - || !strcasecmp(name, "android_secure")) { + if (qstr_case_eq(name, &q_autorun) + || qstr_case_eq(name, &q__android_secure) + || qstr_case_eq(name, &q_android_secure)) { return 0; } } @@ -120,290 +171,675 @@ int open_flags_to_access_mode(int open_flags) { } } -static int insert_str_to_int_lock(struct packagelist_data *pkgl_dat, char *key, - unsigned int value) +static struct hashtable_entry *alloc_hashtable_entry(const struct qstr *key, + appid_t value) { - struct hashtable_entry *hash_cur; - struct hashtable_entry *new_entry; - unsigned int hash = str_hash(key); + struct hashtable_entry *ret = kmem_cache_alloc(hashtable_entry_cachep, + GFP_KERNEL); + if (!ret) + return NULL; - hash_for_each_possible(pkgl_dat->package_to_appid, hash_cur, hlist, hash) { - if (!strcasecmp(key, hash_cur->key)) { - hash_cur->value = value; - return 0; - } + if (!qstr_copy(key, &ret->key)) { + kmem_cache_free(hashtable_entry_cachep, ret); + return NULL; } - new_entry = kmem_cache_alloc(hashtable_entry_cachep, GFP_KERNEL); - if (!new_entry) - return -ENOMEM; - new_entry->key = kstrdup(key, GFP_KERNEL); - new_entry->value = value; - hash_add(pkgl_dat->package_to_appid, &new_entry->hlist, hash); - return 0; -} -static void fixup_perms(struct super_block *sb) { - if (sb && sb->s_magic == SDCARDFS_SUPER_MAGIC) { - mutex_lock(&sb->s_root->d_inode->i_mutex); - get_derive_permissions_recursive(sb->s_root); - mutex_unlock(&sb->s_root->d_inode->i_mutex); - } -} - -static int insert_str_to_int(struct packagelist_data *pkgl_dat, char *key, - unsigned int value) { - int ret; - struct sdcardfs_sb_info *sbinfo; - mutex_lock(&sdcardfs_super_list_lock); - mutex_lock(&pkgl_dat->hashtable_lock); - ret = insert_str_to_int_lock(pkgl_dat, key, value); - mutex_unlock(&pkgl_dat->hashtable_lock); - - list_for_each_entry(sbinfo, &sdcardfs_super_list, list) { - if (sbinfo) { - fixup_perms(sbinfo->sb); - } - } - mutex_unlock(&sdcardfs_super_list_lock); + atomic_set(&ret->value, value); return ret; } -static void remove_str_to_int_lock(struct hashtable_entry *h_entry) { - kfree(h_entry->key); - hash_del(&h_entry->hlist); - kmem_cache_free(hashtable_entry_cachep, h_entry); +static int insert_packagelist_appid_entry_locked(const struct qstr *key, appid_t value) +{ + struct hashtable_entry *hash_cur; + struct hashtable_entry *new_entry; + unsigned int hash = key->hash; + + hash_for_each_possible_rcu(package_to_appid, hash_cur, hlist, hash) { + if (qstr_case_eq(key, &hash_cur->key)) { + atomic_set(&hash_cur->value, value); + return 0; + } + } + new_entry = alloc_hashtable_entry(key, value); + if (!new_entry) + return -ENOMEM; + hash_add_rcu(package_to_appid, &new_entry->hlist, hash); + return 0; } -static void remove_str_to_int(struct packagelist_data *pkgl_dat, const char *key) +static int insert_ext_gid_entry_locked(const struct qstr *key, appid_t value) +{ + struct hashtable_entry *hash_cur; + struct hashtable_entry *new_entry; + unsigned int hash = key->hash; + + /* An extension can only belong to one gid */ + hash_for_each_possible_rcu(ext_to_groupid, hash_cur, hlist, hash) { + if (qstr_case_eq(key, &hash_cur->key)) + return -EINVAL; + } + new_entry = alloc_hashtable_entry(key, value); + if (!new_entry) + return -ENOMEM; + hash_add_rcu(ext_to_groupid, &new_entry->hlist, hash); + return 0; +} + +static int insert_userid_exclude_entry_locked(const struct qstr *key, userid_t value) +{ + struct hashtable_entry *hash_cur; + struct hashtable_entry *new_entry; + unsigned int hash = key->hash; + + /* Only insert if not already present */ + hash_for_each_possible_rcu(package_to_userid, hash_cur, hlist, hash) { + if (atomic_read(&hash_cur->value) == value && + qstr_case_eq(key, &hash_cur->key)) + return 0; + } + new_entry = alloc_hashtable_entry(key, value); + if (!new_entry) + return -ENOMEM; + hash_add_rcu(package_to_userid, &new_entry->hlist, hash); + return 0; +} + +static void fixup_all_perms_name(const struct qstr *key) { struct sdcardfs_sb_info *sbinfo; - struct hashtable_entry *hash_cur; - unsigned int hash = str_hash(key); + struct limit_search limit = { + .flags = BY_NAME, + .name = key->name, + .length = key->len, + }; + list_for_each_entry(sbinfo, &sdcardfs_super_list, list) { + if (sbinfo_has_sdcard_magic(sbinfo)) + fixup_perms_recursive(sbinfo->sb->s_root, &limit); + } +} + +static void fixup_all_perms_name_userid(const struct qstr *key, userid_t userid) +{ + struct sdcardfs_sb_info *sbinfo; + struct limit_search limit = { + .flags = BY_NAME | BY_USERID, + .name = key->name, + .length = key->len, + .userid = userid, + }; + list_for_each_entry(sbinfo, &sdcardfs_super_list, list) { + if (sbinfo_has_sdcard_magic(sbinfo)) + fixup_perms_recursive(sbinfo->sb->s_root, &limit); + } +} + +static void fixup_all_perms_userid(userid_t userid) +{ + struct sdcardfs_sb_info *sbinfo; + struct limit_search limit = { + .flags = BY_USERID, + .userid = userid, + }; + list_for_each_entry(sbinfo, &sdcardfs_super_list, list) { + if (sbinfo_has_sdcard_magic(sbinfo)) + fixup_perms_recursive(sbinfo->sb->s_root, &limit); + } +} + +static int insert_packagelist_entry(const struct qstr *key, appid_t value) +{ + int err; + mutex_lock(&sdcardfs_super_list_lock); - mutex_lock(&pkgl_dat->hashtable_lock); - hash_for_each_possible(pkgl_dat->package_to_appid, hash_cur, hlist, hash) { - if (!strcasecmp(key, hash_cur->key)) { - remove_str_to_int_lock(hash_cur); + err = insert_packagelist_appid_entry_locked(key, value); + if (!err) + fixup_all_perms_name(key); + mutex_unlock(&sdcardfs_super_list_lock); + + return err; +} + +static int insert_ext_gid_entry(const struct qstr *key, appid_t value) +{ + int err; + + mutex_lock(&sdcardfs_super_list_lock); + err = insert_ext_gid_entry_locked(key, value); + mutex_unlock(&sdcardfs_super_list_lock); + + return err; +} + +static int insert_userid_exclude_entry(const struct qstr *key, userid_t value) +{ + int err; + + mutex_lock(&sdcardfs_super_list_lock); + err = insert_userid_exclude_entry_locked(key, value); + if (!err) + fixup_all_perms_name_userid(key, value); + mutex_unlock(&sdcardfs_super_list_lock); + + return err; +} + +static void free_hashtable_entry(struct hashtable_entry *entry) +{ + kfree(entry->key.name); + hash_del_rcu(&entry->dlist); + kmem_cache_free(hashtable_entry_cachep, entry); +} + +static void remove_packagelist_entry_locked(const struct qstr *key) +{ + struct hashtable_entry *hash_cur; + unsigned int hash = key->hash; + struct hlist_node *h_t; + HLIST_HEAD(free_list); + + hash_for_each_possible_rcu(package_to_userid, hash_cur, hlist, hash) { + if (qstr_case_eq(key, &hash_cur->key)) { + hash_del_rcu(&hash_cur->hlist); + hlist_add_head(&hash_cur->dlist, &free_list); + } + } + hash_for_each_possible_rcu(package_to_appid, hash_cur, hlist, hash) { + if (qstr_case_eq(key, &hash_cur->key)) { + hash_del_rcu(&hash_cur->hlist); + hlist_add_head(&hash_cur->dlist, &free_list); break; } } - mutex_unlock(&pkgl_dat->hashtable_lock); - list_for_each_entry(sbinfo, &sdcardfs_super_list, list) { - if (sbinfo) { - fixup_perms(sbinfo->sb); - } - } + synchronize_rcu(); + hlist_for_each_entry_safe(hash_cur, h_t, &free_list, dlist) + free_hashtable_entry(hash_cur); +} + +static void remove_packagelist_entry(const struct qstr *key) +{ + mutex_lock(&sdcardfs_super_list_lock); + remove_packagelist_entry_locked(key); + fixup_all_perms_name(key); mutex_unlock(&sdcardfs_super_list_lock); return; } -static void remove_all_hashentrys(struct packagelist_data *pkgl_dat) +static void remove_ext_gid_entry_locked(const struct qstr *key, gid_t group) +{ + struct hashtable_entry *hash_cur; + unsigned int hash = key->hash; + + hash_for_each_possible_rcu(ext_to_groupid, hash_cur, hlist, hash) { + if (qstr_case_eq(key, &hash_cur->key) && atomic_read(&hash_cur->value) == group) { + hash_del_rcu(&hash_cur->hlist); + synchronize_rcu(); + free_hashtable_entry(hash_cur); + break; + } + } +} + +static void remove_ext_gid_entry(const struct qstr *key, gid_t group) +{ + mutex_lock(&sdcardfs_super_list_lock); + remove_ext_gid_entry_locked(key, group); + mutex_unlock(&sdcardfs_super_list_lock); + return; +} + +static void remove_userid_all_entry_locked(userid_t userid) { struct hashtable_entry *hash_cur; struct hlist_node *h_t; + HLIST_HEAD(free_list); int i; - mutex_lock(&pkgl_dat->hashtable_lock); - hash_for_each_safe(pkgl_dat->package_to_appid, i, h_t, hash_cur, hlist) - remove_str_to_int_lock(hash_cur); - mutex_unlock(&pkgl_dat->hashtable_lock); - hash_init(pkgl_dat->package_to_appid); -} -static struct packagelist_data * packagelist_create(void) -{ - struct packagelist_data *pkgl_dat; - - pkgl_dat = kmalloc(sizeof(*pkgl_dat), GFP_KERNEL | __GFP_ZERO); - if (!pkgl_dat) { - printk(KERN_ERR "sdcardfs: Failed to create hash\n"); - return ERR_PTR(-ENOMEM); + hash_for_each_rcu(package_to_userid, i, hash_cur, hlist) { + if (atomic_read(&hash_cur->value) == userid) { + hash_del_rcu(&hash_cur->hlist); + hlist_add_head(&hash_cur->dlist, &free_list); + } + } + synchronize_rcu(); + hlist_for_each_entry_safe(hash_cur, h_t, &free_list, dlist) { + free_hashtable_entry(hash_cur); } - - mutex_init(&pkgl_dat->hashtable_lock); - hash_init(pkgl_dat->package_to_appid); - - return pkgl_dat; } -static void packagelist_destroy(struct packagelist_data *pkgl_dat) +static void remove_userid_all_entry(userid_t userid) { - remove_all_hashentrys(pkgl_dat); - printk(KERN_INFO "sdcardfs: destroyed packagelist pkgld\n"); - kfree(pkgl_dat); + mutex_lock(&sdcardfs_super_list_lock); + remove_userid_all_entry_locked(userid); + fixup_all_perms_userid(userid); + mutex_unlock(&sdcardfs_super_list_lock); + return; } -struct package_appid { +static void remove_userid_exclude_entry_locked(const struct qstr *key, userid_t userid) +{ + struct hashtable_entry *hash_cur; + unsigned int hash = key->hash; + + hash_for_each_possible_rcu(package_to_userid, hash_cur, hlist, hash) { + if (qstr_case_eq(key, &hash_cur->key) && + atomic_read(&hash_cur->value) == userid) { + hash_del_rcu(&hash_cur->hlist); + synchronize_rcu(); + free_hashtable_entry(hash_cur); + break; + } + } +} + +static void remove_userid_exclude_entry(const struct qstr *key, userid_t userid) +{ + mutex_lock(&sdcardfs_super_list_lock); + remove_userid_exclude_entry_locked(key, userid); + fixup_all_perms_name_userid(key, userid); + mutex_unlock(&sdcardfs_super_list_lock); + return; +} + +static void packagelist_destroy(void) +{ + struct hashtable_entry *hash_cur; + struct hlist_node *h_t; + HLIST_HEAD(free_list); + int i; + mutex_lock(&sdcardfs_super_list_lock); + hash_for_each_rcu(package_to_appid, i, hash_cur, hlist) { + hash_del_rcu(&hash_cur->hlist); + hlist_add_head(&hash_cur->dlist, &free_list); + } + hash_for_each_rcu(package_to_userid, i, hash_cur, hlist) { + hash_del_rcu(&hash_cur->hlist); + hlist_add_head(&hash_cur->dlist, &free_list); + } + synchronize_rcu(); + hlist_for_each_entry_safe(hash_cur, h_t, &free_list, dlist) + free_hashtable_entry(hash_cur); + mutex_unlock(&sdcardfs_super_list_lock); + printk(KERN_INFO "sdcardfs: destroyed packagelist pkgld\n"); +} + +#define SDCARDFS_CONFIGFS_ATTR(_pfx, _name) \ +static struct configfs_attribute _pfx##attr_##_name = { \ + .ca_name = __stringify(_name), \ + .ca_mode = S_IRUGO | S_IWUGO, \ + .ca_owner = THIS_MODULE, \ + .show = _pfx##_name##_show, \ + .store = _pfx##_name##_store, \ +} + +#define SDCARDFS_CONFIGFS_ATTR_RO(_pfx, _name) \ +static struct configfs_attribute _pfx##attr_##_name = { \ + .ca_name = __stringify(_name), \ + .ca_mode = S_IRUGO, \ + .ca_owner = THIS_MODULE, \ + .show = _pfx##_name##_show, \ +} + +#define SDCARDFS_CONFIGFS_ATTR_WO(_pfx, _name) \ +static struct configfs_attribute _pfx##attr_##_name = { \ + .ca_name = __stringify(_name), \ + .ca_mode = S_IWUGO, \ + .ca_owner = THIS_MODULE, \ + .store = _pfx##_name##_store, \ +} + +struct package_details { struct config_item item; - int add_pid; + struct qstr name; }; -static inline struct package_appid *to_package_appid(struct config_item *item) +static inline struct package_details *to_package_details(struct config_item *item) { - return item ? container_of(item, struct package_appid, item) : NULL; + return item ? container_of(item, struct package_details, item) : NULL; } -static ssize_t package_appid_attr_show(struct config_item *item, - char *page) +static ssize_t package_details_appid_show(struct config_item *item, char *page) { - ssize_t count; - count = sprintf(page, "%d\n", get_appid(pkgl_data_all, item->ci_name)); - return count; + return scnprintf(page, PAGE_SIZE, "%u\n", __get_appid(&to_package_details(item)->name)); } -static ssize_t package_appid_attr_store(struct config_item *item, +static ssize_t package_details_appid_store(struct config_item *item, const char *page, size_t count) { - struct package_appid *package_appid = to_package_appid(item); - unsigned long tmp; - char *p = (char *) page; + unsigned int tmp; int ret; - tmp = simple_strtoul(p, &p, 10); - if (!p || (*p && (*p != '\n'))) - return -EINVAL; + ret = kstrtouint(page, 10, &tmp); + if (ret) + return ret; + + ret = insert_packagelist_entry(&to_package_details(item)->name, tmp); - if (tmp > INT_MAX) - return -ERANGE; - ret = insert_str_to_int(pkgl_data_all, item->ci_name, (unsigned int)tmp); - package_appid->add_pid = tmp; if (ret) return ret; return count; } -static struct configfs_attribute package_appid_attr_add_pid = { - .ca_owner = THIS_MODULE, - .ca_name = "appid", - .ca_mode = S_IRUGO | S_IWUGO, - .show = package_appid_attr_show, - .store = package_appid_attr_store, -}; +static ssize_t package_details_excluded_userids_show(struct config_item *item, + char *page) +{ + struct package_details *package_details = to_package_details(item); + struct hashtable_entry *hash_cur; + unsigned int hash = package_details->name.hash; + int count = 0; -static struct configfs_attribute *package_appid_attrs[] = { - &package_appid_attr_add_pid, + rcu_read_lock(); + hash_for_each_possible_rcu(package_to_userid, hash_cur, hlist, hash) { + if (qstr_case_eq(&package_details->name, &hash_cur->key)) + count += scnprintf(page + count, PAGE_SIZE - count, + "%d ", atomic_read(&hash_cur->value)); + } + rcu_read_unlock(); + if (count) + count--; + count += scnprintf(page + count, PAGE_SIZE - count, "\n"); + return count; +} + +static ssize_t package_details_excluded_userids_store(struct config_item *item, + const char *page, size_t count) +{ + unsigned int tmp; + int ret; + + ret = kstrtouint(page, 10, &tmp); + if (ret) + return ret; + + ret = insert_userid_exclude_entry(&to_package_details(item)->name, tmp); + + if (ret) + return ret; + + return count; +} + +static ssize_t package_details_clear_userid_store(struct config_item *item, + const char *page, size_t count) +{ + unsigned int tmp; + int ret; + + ret = kstrtouint(page, 10, &tmp); + if (ret) + return ret; + remove_userid_exclude_entry(&to_package_details(item)->name, tmp); + return count; +} + +static void package_details_release(struct config_item *item) +{ + struct package_details *package_details = to_package_details(item); + printk(KERN_INFO "sdcardfs: removing %s\n", package_details->name.name); + remove_packagelist_entry(&package_details->name); + kfree(package_details->name.name); + kfree(package_details); +} + +SDCARDFS_CONFIGFS_ATTR(package_details_, appid); +SDCARDFS_CONFIGFS_ATTR(package_details_, excluded_userids); +SDCARDFS_CONFIGFS_ATTR_WO(package_details_, clear_userid); + +static struct configfs_attribute *package_details_attrs[] = { + &package_details_attr_appid, + &package_details_attr_excluded_userids, + &package_details_attr_clear_userid, NULL, }; -static void package_appid_release(struct config_item *item) -{ - printk(KERN_INFO "sdcardfs: removing %s\n", item->ci_dentry->d_name.name); - /* item->ci_name is freed already, so we rely on the dentry */ - remove_str_to_int(pkgl_data_all, item->ci_dentry->d_name.name); - kfree(to_package_appid(item)); -} - -static struct configfs_item_operations package_appid_item_ops = { - .release = package_appid_release, +static struct configfs_item_operations package_details_item_ops = { + .release = package_details_release, }; static struct config_item_type package_appid_type = { - .ct_item_ops = &package_appid_item_ops, - .ct_attrs = package_appid_attrs, + .ct_item_ops = &package_details_item_ops, + .ct_attrs = package_details_attrs, .ct_owner = THIS_MODULE, }; - -struct sdcardfs_packages { +struct extensions_value { struct config_group group; + unsigned int num; }; -static inline struct sdcardfs_packages *to_sdcardfs_packages(struct config_item *item) +struct extension_details { + struct config_item item; + struct qstr name; + unsigned int num; +}; + +static inline struct extensions_value *to_extensions_value(struct config_item *item) { - return item ? container_of(to_config_group(item), struct sdcardfs_packages, group) : NULL; + return item ? container_of(to_config_group(item), struct extensions_value, group) : NULL; } -static struct config_item *sdcardfs_packages_make_item(struct config_group *group, const char *name) +static inline struct extension_details *to_extension_details(struct config_item *item) { - struct package_appid *package_appid; + return item ? container_of(item, struct extension_details, item) : NULL; +} - package_appid = kzalloc(sizeof(struct package_appid), GFP_KERNEL); - if (!package_appid) +static void extension_details_release(struct config_item *item) +{ + struct extension_details *extension_details = to_extension_details(item); + + printk(KERN_INFO "sdcardfs: No longer mapping %s files to gid %d\n", + extension_details->name.name, extension_details->num); + remove_ext_gid_entry(&extension_details->name, extension_details->num); + kfree(extension_details->name.name); + kfree(extension_details); +} + +static struct configfs_item_operations extension_details_item_ops = { + .release = extension_details_release, +}; + +static struct config_item_type extension_details_type = { + .ct_item_ops = &extension_details_item_ops, + .ct_owner = THIS_MODULE, +}; + +static struct config_item *extension_details_make_item(struct config_group *group, const char *name) +{ + struct extensions_value *extensions_value = to_extensions_value(&group->cg_item); + struct extension_details *extension_details = kzalloc(sizeof(struct extension_details), GFP_KERNEL); + const char *tmp; + int ret; + if (!extension_details) return ERR_PTR(-ENOMEM); - config_item_init_type_name(&package_appid->item, name, - &package_appid_type); + tmp = kstrdup(name, GFP_KERNEL); + if (!tmp) { + kfree(extension_details); + return ERR_PTR(-ENOMEM); + } + qstr_init(&extension_details->name, tmp); + ret = insert_ext_gid_entry(&extension_details->name, extensions_value->num); - package_appid->add_pid = 0; + if (ret) { + kfree(extension_details->name.name); + kfree(extension_details); + return ERR_PTR(ret); + } + config_item_init_type_name(&extension_details->item, name, &extension_details_type); - return &package_appid->item; + return &extension_details->item; } -static ssize_t packages_attr_show(struct config_item *item, - char *page) +static struct configfs_group_operations extensions_value_group_ops = { + .make_item = extension_details_make_item, +}; + +static struct config_item_type extensions_name_type = { + .ct_group_ops = &extensions_value_group_ops, + .ct_owner = THIS_MODULE, +}; + +static struct config_group *extensions_make_group(struct config_group *group, const char *name) { - struct hashtable_entry *hash_cur; - struct hlist_node *h_t; + struct extensions_value *extensions_value; + unsigned int tmp; + int ret; + + extensions_value = kzalloc(sizeof(struct extensions_value), GFP_KERNEL); + if (!extensions_value) + return ERR_PTR(-ENOMEM); + ret = kstrtouint(name, 10, &tmp); + if (ret) { + kfree(extensions_value); + return ERR_PTR(ret); + } + + extensions_value->num = tmp; + config_group_init_type_name(&extensions_value->group, name, + &extensions_name_type); + return &extensions_value->group; +} + +static void extensions_drop_group(struct config_group *group, struct config_item *item) +{ + struct extensions_value *value = to_extensions_value(item); + printk(KERN_INFO "sdcardfs: No longer mapping any files to gid %d\n", value->num); + kfree(value); +} + +static struct configfs_group_operations extensions_group_ops = { + .make_group = extensions_make_group, + .drop_item = extensions_drop_group, +}; + +static struct config_item_type extensions_type = { + .ct_group_ops = &extensions_group_ops, + .ct_owner = THIS_MODULE, +}; + +struct config_group extension_group = { + .cg_item = { + .ci_namebuf = "extensions", + .ci_type = &extensions_type, + }, +}; + +static struct config_item *packages_make_item(struct config_group *group, const char *name) +{ + struct package_details *package_details; + const char *tmp; + + package_details = kzalloc(sizeof(struct package_details), GFP_KERNEL); + if (!package_details) + return ERR_PTR(-ENOMEM); + tmp = kstrdup(name, GFP_KERNEL); + if (!tmp) { + kfree(package_details); + return ERR_PTR(-ENOMEM); + } + qstr_init(&package_details->name, tmp); + config_item_init_type_name(&package_details->item, name, + &package_appid_type); + + return &package_details->item; +} + +static ssize_t packages_list_show(struct config_item *item, char *page) +{ + struct hashtable_entry *hash_cur_app; + struct hashtable_entry *hash_cur_user; int i; int count = 0, written = 0; - char errormsg[] = "\n"; + const char errormsg[] = "\n"; + unsigned int hash; - mutex_lock(&pkgl_data_all->hashtable_lock); - hash_for_each_safe(pkgl_data_all->package_to_appid, i, h_t, hash_cur, hlist) { - written = scnprintf(page + count, PAGE_SIZE - sizeof(errormsg) - count, "%s %d\n", (char *)hash_cur->key, hash_cur->value); - if (count + written == PAGE_SIZE - sizeof(errormsg)) { + rcu_read_lock(); + hash_for_each_rcu(package_to_appid, i, hash_cur_app, hlist) { + written = scnprintf(page + count, PAGE_SIZE - sizeof(errormsg) - count, "%s %d\n", + hash_cur_app->key.name, atomic_read(&hash_cur_app->value)); + hash = hash_cur_app->key.hash; + hash_for_each_possible_rcu(package_to_userid, hash_cur_user, hlist, hash) { + if (qstr_case_eq(&hash_cur_app->key, &hash_cur_user->key)) { + written += scnprintf(page + count + written - 1, + PAGE_SIZE - sizeof(errormsg) - count - written + 1, + " %d\n", atomic_read(&hash_cur_user->value)) - 1; + } + } + if (count + written == PAGE_SIZE - sizeof(errormsg) - 1) { count += scnprintf(page + count, PAGE_SIZE - count, errormsg); break; } count += written; } - mutex_unlock(&pkgl_data_all->hashtable_lock); + rcu_read_unlock(); return count; } -static struct configfs_attribute sdcardfs_packages_attr_description = { - .ca_owner = THIS_MODULE, - .ca_name = "packages_gid.list", - .ca_mode = S_IRUGO, - .show = packages_attr_show, -}; - -static struct configfs_attribute *sdcardfs_packages_attrs[] = { - &sdcardfs_packages_attr_description, - NULL, -}; - -static void sdcardfs_packages_release(struct config_item *item) +static ssize_t packages_remove_userid_store(struct config_item *item, + const char *page, size_t count) { + unsigned int tmp; + int ret; - printk(KERN_INFO "sdcardfs: destroyed something?\n"); - kfree(to_sdcardfs_packages(item)); + ret = kstrtouint(page, 10, &tmp); + if (ret) + return ret; + remove_userid_all_entry(tmp); + return count; } -static struct configfs_item_operations sdcardfs_packages_item_ops = { - .release = sdcardfs_packages_release, +static struct configfs_attribute packages_attr_packages_gid_list = { + .ca_name = "packages_gid.list", + .ca_mode = S_IRUGO, + .ca_owner = THIS_MODULE, + .show = packages_list_show, +}; + +SDCARDFS_CONFIGFS_ATTR_WO(packages_, remove_userid); + +static struct configfs_attribute *packages_attrs[] = { + &packages_attr_packages_gid_list, + &packages_attr_remove_userid, + NULL, }; /* * Note that, since no extra work is required on ->drop_item(), * no ->drop_item() is provided. */ -static struct configfs_group_operations sdcardfs_packages_group_ops = { - .make_item = sdcardfs_packages_make_item, +static struct configfs_group_operations packages_group_ops = { + .make_item = packages_make_item, }; -static struct config_item_type sdcardfs_packages_type = { - .ct_item_ops = &sdcardfs_packages_item_ops, - .ct_group_ops = &sdcardfs_packages_group_ops, - .ct_attrs = sdcardfs_packages_attrs, +static struct config_item_type packages_type = { + .ct_group_ops = &packages_group_ops, + .ct_attrs = packages_attrs, .ct_owner = THIS_MODULE, }; -static struct configfs_subsystem sdcardfs_packages_subsys = { +struct config_group *sd_default_groups[] = { + &extension_group, + NULL, +}; + +static struct configfs_subsystem sdcardfs_packages = { .su_group = { .cg_item = { .ci_namebuf = "sdcardfs", - .ci_type = &sdcardfs_packages_type, + .ci_type = &packages_type, }, + .default_groups = sd_default_groups, }, }; static int configfs_sdcardfs_init(void) { - int ret; - struct configfs_subsystem *subsys = &sdcardfs_packages_subsys; - + int ret, i; + struct configfs_subsystem *subsys = &sdcardfs_packages; + for (i = 0; sd_default_groups[i]; i++) { + config_group_init(sd_default_groups[i]); + } config_group_init(&subsys->su_group); mutex_init(&subsys->su_mutex); ret = configfs_register_subsystem(subsys); @@ -417,7 +853,7 @@ static int configfs_sdcardfs_init(void) static void configfs_sdcardfs_exit(void) { - configfs_unregister_subsystem(&sdcardfs_packages_subsys); + configfs_unregister_subsystem(&sdcardfs_packages); } int packagelist_init(void) @@ -430,7 +866,6 @@ int packagelist_init(void) return -ENOMEM; } - pkgl_data_all = packagelist_create(); configfs_sdcardfs_init(); return 0; } @@ -438,7 +873,7 @@ int packagelist_init(void) void packagelist_exit(void) { configfs_sdcardfs_exit(); - packagelist_destroy(pkgl_data_all); + packagelist_destroy(); if (hashtable_entry_cachep) kmem_cache_destroy(hashtable_entry_cachep); } diff --git a/fs/sdcardfs/sdcardfs.h b/fs/sdcardfs/sdcardfs.h index f111f898b630..f3cced313108 100644 --- a/fs/sdcardfs/sdcardfs.h +++ b/fs/sdcardfs/sdcardfs.h @@ -65,17 +65,26 @@ #define AID_SDCARD_PICS 1033 /* external storage photos access */ #define AID_SDCARD_AV 1034 /* external storage audio/video access */ #define AID_SDCARD_ALL 1035 /* access all users external storage */ +#define AID_MEDIA_OBB 1059 /* obb files */ + +#define AID_SDCARD_IMAGE 1057 #define AID_PACKAGE_INFO 1027 -#define fix_derived_permission(x) \ + +/* + * Permissions are handled by our permission function. + * We don't want anyone who happens to look at our inode value to prematurely + * block access, so store more permissive values. These are probably never + * used. + */ +#define fixup_tmp_permissions(x) \ do { \ (x)->i_uid = make_kuid(&init_user_ns, SDCARDFS_I(x)->d_uid); \ - (x)->i_gid = make_kgid(&init_user_ns, get_gid(SDCARDFS_I(x))); \ - (x)->i_mode = ((x)->i_mode & S_IFMT) | get_mode(SDCARDFS_I(x));\ + (x)->i_gid = make_kgid(&init_user_ns, AID_SDCARD_RW); \ + (x)->i_mode = ((x)->i_mode & S_IFMT) | 0775;\ } while (0) - /* OVERRIDE_CRED() and REVERT_CRED() * OVERRID_CRED() * backup original task->cred @@ -85,12 +94,12 @@ * These two macro should be used in pair, and OVERRIDE_CRED() should be * placed at the beginning of a function, right after variable declaration. */ -#define OVERRIDE_CRED(sdcardfs_sbi, saved_cred) \ - saved_cred = override_fsids(sdcardfs_sbi); \ +#define OVERRIDE_CRED(sdcardfs_sbi, saved_cred, info) \ + saved_cred = override_fsids(sdcardfs_sbi, info); \ if (!saved_cred) { return -ENOMEM; } -#define OVERRIDE_CRED_PTR(sdcardfs_sbi, saved_cred) \ - saved_cred = override_fsids(sdcardfs_sbi); \ +#define OVERRIDE_CRED_PTR(sdcardfs_sbi, saved_cred, info) \ + saved_cred = override_fsids(sdcardfs_sbi, info); \ if (!saved_cred) { return ERR_PTR(-ENOMEM); } #define REVERT_CRED(saved_cred) revert_fsids(saved_cred) @@ -121,13 +130,18 @@ typedef enum { PERM_ANDROID_OBB, /* This node is "/Android/media" */ PERM_ANDROID_MEDIA, + /* This node is "/Android/[data|media|obb]/[package]" */ + PERM_ANDROID_PACKAGE, + /* This node is "/Android/[data|media|obb]/[package]/cache" */ + PERM_ANDROID_PACKAGE_CACHE, } perm_t; struct sdcardfs_sb_info; struct sdcardfs_mount_options; +struct sdcardfs_inode_info; /* Do not directly use this function. Use OVERRIDE_CRED() instead. */ -const struct cred * override_fsids(struct sdcardfs_sb_info* sbi); +const struct cred * override_fsids(struct sdcardfs_sb_info* sbi, struct sdcardfs_inode_info *info); /* Do not directly use this function, use REVERT_CRED() instead. */ void revert_fsids(const struct cred * old_cred); @@ -169,6 +183,10 @@ struct sdcardfs_inode_info { userid_t userid; uid_t d_uid; bool under_android; + bool under_cache; + bool under_obb; + /* top folder for ownership */ + struct inode *top; struct inode vfs_inode; }; @@ -185,12 +203,18 @@ struct sdcardfs_mount_options { uid_t fs_low_uid; gid_t fs_low_gid; userid_t fs_user_id; - gid_t gid; - mode_t mask; bool multiuser; unsigned int reserved_mb; }; +struct sdcardfs_vfsmount_options { + gid_t gid; + mode_t mask; +}; + +extern int parse_options_remount(struct super_block *sb, char *options, int silent, + struct sdcardfs_vfsmount_options *vfsopts); + /* sdcardfs super-block data in memory */ struct sdcardfs_sb_info { struct super_block *sb; @@ -321,9 +345,44 @@ static inline void sdcardfs_put_reset_##pname(const struct dentry *dent) \ SDCARDFS_DENT_FUNC(lower_path) SDCARDFS_DENT_FUNC(orig_path) -static inline int get_gid(struct sdcardfs_inode_info *info) { - struct sdcardfs_sb_info *sb_info = SDCARDFS_SB(info->vfs_inode.i_sb); - if (sb_info->options.gid == AID_SDCARD_RW) { +static inline bool sbinfo_has_sdcard_magic(struct sdcardfs_sb_info *sbinfo) +{ + return sbinfo && sbinfo->sb && sbinfo->sb->s_magic == SDCARDFS_SUPER_MAGIC; +} + +/* grab a refererence if we aren't linking to ourself */ +static inline void set_top(struct sdcardfs_inode_info *info, struct inode *top) +{ + struct inode *old_top = NULL; + BUG_ON(IS_ERR_OR_NULL(top)); + if (info->top && info->top != &info->vfs_inode) { + old_top = info->top; + } + if (top != &info->vfs_inode) + igrab(top); + info->top = top; + iput(old_top); +} + +static inline struct inode *grab_top(struct sdcardfs_inode_info *info) +{ + struct inode *top = info->top; + if (top) { + return igrab(top); + } else { + return NULL; + } +} + +static inline void release_top(struct sdcardfs_inode_info *info) +{ + iput(info->top); +} + +static inline int get_gid(struct vfsmount *mnt, struct sdcardfs_inode_info *info) { + struct sdcardfs_vfsmount_options *opts = mnt->data; + + if (opts->gid == AID_SDCARD_RW) { /* As an optimization, certain trusted system components only run * as owner but operate across all users. Since we're now handing * out the sdcard_rw GID only to trusted apps, we're okay relaxing @@ -331,14 +390,15 @@ static inline int get_gid(struct sdcardfs_inode_info *info) { * assigned to app directories are still multiuser aware. */ return AID_SDCARD_RW; } else { - return multiuser_get_uid(info->userid, sb_info->options.gid); + return multiuser_get_uid(info->userid, opts->gid); } } -static inline int get_mode(struct sdcardfs_inode_info *info) { +static inline int get_mode(struct vfsmount *mnt, struct sdcardfs_inode_info *info) { int owner_mode; int filtered_mode; - struct sdcardfs_sb_info *sb_info = SDCARDFS_SB(info->vfs_inode.i_sb); - int visible_mode = 0775 & ~sb_info->options.mask; + struct sdcardfs_vfsmount_options *opts = mnt->data; + int visible_mode = 0775 & ~opts->mask; + if (info->perm == PERM_PRE_ROOT) { /* Top of multi-user view should always be visible to ensure @@ -348,7 +408,7 @@ static inline int get_mode(struct sdcardfs_inode_info *info) { /* Block "other" access to Android directories, since only apps * belonging to a specific user should be in there; we still * leave +x open for the default view. */ - if (sb_info->options.gid == AID_SDCARD_RW) { + if (opts->gid == AID_SDCARD_RW) { visible_mode = visible_mode & ~0006; } else { visible_mode = visible_mode & ~0007; @@ -396,20 +456,34 @@ extern struct mutex sdcardfs_super_list_lock; extern struct list_head sdcardfs_super_list; /* for packagelist.c */ -extern appid_t get_appid(void *pkgl_id, const char *app_name); -extern int check_caller_access_to_name(struct inode *parent_node, const char* name); +extern appid_t get_appid(const char *app_name); +extern appid_t get_ext_gid(const char *app_name); +extern appid_t is_excluded(const char *app_name, userid_t userid); +extern int check_caller_access_to_name(struct inode *parent_node, const struct qstr* name); extern int open_flags_to_access_mode(int open_flags); extern int packagelist_init(void); extern void packagelist_exit(void); /* for derived_perm.c */ -extern void setup_derived_state(struct inode *inode, perm_t perm, - userid_t userid, uid_t uid, bool under_android); +#define BY_NAME (1 << 0) +#define BY_USERID (1 << 1) +struct limit_search { + unsigned int flags; + const char *name; + size_t length; + userid_t userid; +}; + +extern void setup_derived_state(struct inode *inode, perm_t perm, userid_t userid, + uid_t uid, bool under_android, struct inode *top); extern void get_derived_permission(struct dentry *parent, struct dentry *dentry); -extern void get_derived_permission_new(struct dentry *parent, struct dentry *dentry, struct dentry *newdentry); -extern void get_derive_permissions_recursive(struct dentry *parent); +extern void get_derived_permission_new(struct dentry *parent, struct dentry *dentry, const struct qstr *name); +extern void drop_recursive(struct dentry *parent); +extern void fixup_top_recursive(struct dentry *parent); +extern void fixup_perms_recursive(struct dentry *dentry, struct limit_search *limit); extern void update_derived_permission_lock(struct dentry *dentry); +void fixup_lower_ownership(struct dentry* dentry, const char *name); extern int need_graft_path(struct dentry *dentry); extern int is_base_obbpath(struct dentry *dentry); extern int is_obbpath_invalid(struct dentry *dentry); @@ -444,7 +518,7 @@ static inline int prepare_dir(const char *path_s, uid_t uid, gid_t gid, mode_t m goto out_unlock; } - err = vfs_mkdir(d_inode(parent.dentry), dent, mode); + err = vfs_mkdir2(parent.mnt, d_inode(parent.dentry), dent, mode); if (err) { if (err == -EEXIST) err = 0; @@ -455,7 +529,7 @@ static inline int prepare_dir(const char *path_s, uid_t uid, gid_t gid, mode_t m attrs.ia_gid = make_kgid(&init_user_ns, gid); attrs.ia_valid = ATTR_UID | ATTR_GID; mutex_lock(&d_inode(dent)->i_mutex); - notify_change(dent, &attrs, NULL); + notify_change2(parent.mnt, dent, &attrs, NULL); mutex_unlock(&d_inode(dent)->i_mutex); out_dput: @@ -513,12 +587,16 @@ static inline int check_min_free_space(struct dentry *dentry, size_t size, int d return 1; } -/* Copies attrs and maintains sdcardfs managed attrs */ +/* + * Copies attrs and maintains sdcardfs managed attrs + * Since our permission check handles all special permissions, set those to be open + */ static inline void sdcardfs_copy_and_fix_attrs(struct inode *dest, const struct inode *src) { - dest->i_mode = (src->i_mode & S_IFMT) | get_mode(SDCARDFS_I(dest)); + dest->i_mode = (src->i_mode & S_IFMT) | S_IRWXU | S_IRWXG | + S_IROTH | S_IXOTH; /* 0775 */ dest->i_uid = make_kuid(&init_user_ns, SDCARDFS_I(dest)->d_uid); - dest->i_gid = make_kgid(&init_user_ns, get_gid(SDCARDFS_I(dest))); + dest->i_gid = make_kgid(&init_user_ns, AID_SDCARD_RW); dest->i_rdev = src->i_rdev; dest->i_atime = src->i_atime; dest->i_mtime = src->i_mtime; @@ -527,4 +605,17 @@ static inline void sdcardfs_copy_and_fix_attrs(struct inode *dest, const struct dest->i_flags = src->i_flags; set_nlink(dest, src->i_nlink); } + +static inline bool str_case_eq(const char *s1, const char *s2) +{ + return !strcasecmp(s1, s2); +} + +static inline bool qstr_case_eq(const struct qstr *q1, const struct qstr *q2) +{ + return q1->len == q2->len && str_case_eq(q1->name, q2->name); +} + +#define QSTR_LITERAL(string) QSTR_INIT(string, sizeof(string)-1) + #endif /* not _SDCARDFS_H_ */ diff --git a/fs/sdcardfs/super.c b/fs/sdcardfs/super.c index 1d6490128c99..edda32b68dc0 100644 --- a/fs/sdcardfs/super.c +++ b/fs/sdcardfs/super.c @@ -108,6 +108,50 @@ static int sdcardfs_remount_fs(struct super_block *sb, int *flags, char *options return err; } +/* + * @mnt: mount point we are remounting + * @sb: superblock we are remounting + * @flags: numeric mount options + * @options: mount options string + */ +static int sdcardfs_remount_fs2(struct vfsmount *mnt, struct super_block *sb, + int *flags, char *options) +{ + int err = 0; + + /* + * The VFS will take care of "ro" and "rw" flags among others. We + * can safely accept a few flags (RDONLY, MANDLOCK), and honor + * SILENT, but anything else left over is an error. + */ + if ((*flags & ~(MS_RDONLY | MS_MANDLOCK | MS_SILENT | MS_REMOUNT)) != 0) { + printk(KERN_ERR + "sdcardfs: remount flags 0x%x unsupported\n", *flags); + err = -EINVAL; + } + printk(KERN_INFO "Remount options were %s for vfsmnt %p.\n", options, mnt); + err = parse_options_remount(sb, options, *flags & ~MS_SILENT, mnt->data); + + + return err; +} + +static void* sdcardfs_clone_mnt_data(void *data) { + struct sdcardfs_vfsmount_options* opt = kmalloc(sizeof(struct sdcardfs_vfsmount_options), GFP_KERNEL); + struct sdcardfs_vfsmount_options* old = data; + if(!opt) return NULL; + opt->gid = old->gid; + opt->mask = old->mask; + return opt; +} + +static void sdcardfs_copy_mnt_data(void *data, void *newdata) { + struct sdcardfs_vfsmount_options* old = data; + struct sdcardfs_vfsmount_options* new = newdata; + old->gid = new->gid; + old->mask = new->mask; +} + /* * Called by iput() when the inode reference count reached zero * and the inode is not hashed anywhere. Used to clear anything @@ -126,6 +170,7 @@ static void sdcardfs_evict_inode(struct inode *inode) */ lower_inode = sdcardfs_lower_inode(inode); sdcardfs_set_lower_inode(inode, NULL); + set_top(SDCARDFS_I(inode), inode); iput(lower_inode); } @@ -190,19 +235,24 @@ static void sdcardfs_umount_begin(struct super_block *sb) lower_sb->s_op->umount_begin(lower_sb); } -static int sdcardfs_show_options(struct seq_file *m, struct dentry *root) +static int sdcardfs_show_options(struct vfsmount *mnt, struct seq_file *m, struct dentry *root) { struct sdcardfs_sb_info *sbi = SDCARDFS_SB(root->d_sb); struct sdcardfs_mount_options *opts = &sbi->options; + struct sdcardfs_vfsmount_options *vfsopts = mnt->data; if (opts->fs_low_uid != 0) - seq_printf(m, ",uid=%u", opts->fs_low_uid); + seq_printf(m, ",fsuid=%u", opts->fs_low_uid); if (opts->fs_low_gid != 0) - seq_printf(m, ",gid=%u", opts->fs_low_gid); - + seq_printf(m, ",fsgid=%u", opts->fs_low_gid); + if (vfsopts->gid != 0) + seq_printf(m, ",gid=%u", vfsopts->gid); if (opts->multiuser) seq_printf(m, ",multiuser"); - + if (vfsopts->mask) + seq_printf(m, ",mask=%u", vfsopts->mask); + if (opts->fs_user_id) + seq_printf(m, ",userid=%u", opts->fs_user_id); if (opts->reserved_mb != 0) seq_printf(m, ",reserved=%uMB", opts->reserved_mb); @@ -213,9 +263,12 @@ const struct super_operations sdcardfs_sops = { .put_super = sdcardfs_put_super, .statfs = sdcardfs_statfs, .remount_fs = sdcardfs_remount_fs, + .remount_fs2 = sdcardfs_remount_fs2, + .clone_mnt_data = sdcardfs_clone_mnt_data, + .copy_mnt_data = sdcardfs_copy_mnt_data, .evict_inode = sdcardfs_evict_inode, .umount_begin = sdcardfs_umount_begin, - .show_options = sdcardfs_show_options, + .show_options2 = sdcardfs_show_options, .alloc_inode = sdcardfs_alloc_inode, .destroy_inode = sdcardfs_destroy_inode, .drop_inode = generic_delete_inode, diff --git a/fs/squashfs/Kconfig b/fs/squashfs/Kconfig index ffb093e72b6c..6dd158a216f4 100644 --- a/fs/squashfs/Kconfig +++ b/fs/squashfs/Kconfig @@ -25,34 +25,6 @@ config SQUASHFS If unsure, say N. -choice - prompt "File decompression options" - depends on SQUASHFS - help - Squashfs now supports two options for decompressing file - data. Traditionally Squashfs has decompressed into an - intermediate buffer and then memcopied it into the page cache. - Squashfs now supports the ability to decompress directly into - the page cache. - - If unsure, select "Decompress file data into an intermediate buffer" - -config SQUASHFS_FILE_CACHE - bool "Decompress file data into an intermediate buffer" - help - Decompress file data into an intermediate buffer and then - memcopy it into the page cache. - -config SQUASHFS_FILE_DIRECT - bool "Decompress files directly into the page cache" - help - Directly decompress file data into the page cache. - Doing so can significantly improve performance because - it eliminates a memcpy and it also removes the lock contention - on the single buffer. - -endchoice - choice prompt "Decompressor parallelisation options" depends on SQUASHFS diff --git a/fs/squashfs/Makefile b/fs/squashfs/Makefile index 246a6f329d89..fe51f1507ed1 100644 --- a/fs/squashfs/Makefile +++ b/fs/squashfs/Makefile @@ -5,8 +5,7 @@ obj-$(CONFIG_SQUASHFS) += squashfs.o squashfs-y += block.o cache.o dir.o export.o file.o fragment.o id.o inode.o squashfs-y += namei.o super.o symlink.o decompressor.o -squashfs-$(CONFIG_SQUASHFS_FILE_CACHE) += file_cache.o -squashfs-$(CONFIG_SQUASHFS_FILE_DIRECT) += file_direct.o page_actor.o +squashfs-y += file_direct.o page_actor.o squashfs-$(CONFIG_SQUASHFS_DECOMP_SINGLE) += decompressor_single.o squashfs-$(CONFIG_SQUASHFS_DECOMP_MULTI) += decompressor_multi.o squashfs-$(CONFIG_SQUASHFS_DECOMP_MULTI_PERCPU) += decompressor_multi_percpu.o diff --git a/fs/squashfs/block.c b/fs/squashfs/block.c index 0cea9b9236d0..2eb66decc5ab 100644 --- a/fs/squashfs/block.c +++ b/fs/squashfs/block.c @@ -28,9 +28,12 @@ #include #include +#include #include #include +#include #include +#include #include "squashfs_fs.h" #include "squashfs_fs_sb.h" @@ -38,45 +41,380 @@ #include "decompressor.h" #include "page_actor.h" -/* - * Read the metadata block length, this is stored in the first two - * bytes of the metadata block. - */ -static struct buffer_head *get_block_length(struct super_block *sb, - u64 *cur_index, int *offset, int *length) +static struct workqueue_struct *squashfs_read_wq; + +struct squashfs_read_request { + struct super_block *sb; + u64 index; + int length; + int compressed; + int offset; + u64 read_end; + struct squashfs_page_actor *output; + enum { + SQUASHFS_COPY, + SQUASHFS_DECOMPRESS, + SQUASHFS_METADATA, + } data_processing; + bool synchronous; + + /* + * If the read is synchronous, it is possible to retrieve information + * about the request by setting these pointers. + */ + int *res; + int *bytes_read; + int *bytes_uncompressed; + + int nr_buffers; + struct buffer_head **bh; + struct work_struct offload; +}; + +struct squashfs_bio_request { + struct buffer_head **bh; + int nr_buffers; +}; + +static int squashfs_bio_submit(struct squashfs_read_request *req); + +int squashfs_init_read_wq(void) { - struct squashfs_sb_info *msblk = sb->s_fs_info; - struct buffer_head *bh; + squashfs_read_wq = create_workqueue("SquashFS read wq"); + return !!squashfs_read_wq; +} - bh = sb_bread(sb, *cur_index); - if (bh == NULL) - return NULL; +void squashfs_destroy_read_wq(void) +{ + flush_workqueue(squashfs_read_wq); + destroy_workqueue(squashfs_read_wq); +} - if (msblk->devblksize - *offset == 1) { - *length = (unsigned char) bh->b_data[*offset]; - put_bh(bh); - bh = sb_bread(sb, ++(*cur_index)); - if (bh == NULL) - return NULL; - *length |= (unsigned char) bh->b_data[0] << 8; - *offset = 1; - } else { - *length = (unsigned char) bh->b_data[*offset] | - (unsigned char) bh->b_data[*offset + 1] << 8; - *offset += 2; +static void free_read_request(struct squashfs_read_request *req, int error) +{ + if (!req->synchronous) + squashfs_page_actor_free(req->output, error); + if (req->res) + *(req->res) = error; + kfree(req->bh); + kfree(req); +} - if (*offset == msblk->devblksize) { - put_bh(bh); - bh = sb_bread(sb, ++(*cur_index)); - if (bh == NULL) - return NULL; - *offset = 0; +static void squashfs_process_blocks(struct squashfs_read_request *req) +{ + int error = 0; + int bytes, i, length; + struct squashfs_sb_info *msblk = req->sb->s_fs_info; + struct squashfs_page_actor *actor = req->output; + struct buffer_head **bh = req->bh; + int nr_buffers = req->nr_buffers; + + for (i = 0; i < nr_buffers; ++i) { + if (!bh[i]) + continue; + wait_on_buffer(bh[i]); + if (!buffer_uptodate(bh[i])) + error = -EIO; + } + if (error) + goto cleanup; + + if (req->data_processing == SQUASHFS_METADATA) { + /* Extract the length of the metadata block */ + if (req->offset != msblk->devblksize - 1) + length = *((u16 *)(bh[0]->b_data + req->offset)); + else { + length = bh[0]->b_data[req->offset]; + length |= bh[1]->b_data[0] << 8; + } + req->compressed = SQUASHFS_COMPRESSED(length); + req->data_processing = req->compressed ? SQUASHFS_DECOMPRESS + : SQUASHFS_COPY; + length = SQUASHFS_COMPRESSED_SIZE(length); + if (req->index + length + 2 > req->read_end) { + for (i = 0; i < nr_buffers; ++i) + put_bh(bh[i]); + kfree(bh); + req->length = length; + req->index += 2; + squashfs_bio_submit(req); + return; + } + req->length = length; + req->offset = (req->offset + 2) % PAGE_SIZE; + if (req->offset < 2) { + put_bh(bh[0]); + ++bh; + --nr_buffers; + } + } + if (req->bytes_read) + *(req->bytes_read) = req->length; + + if (req->data_processing == SQUASHFS_COPY) { + squashfs_bh_to_actor(bh, nr_buffers, req->output, req->offset, + req->length, msblk->devblksize); + } else if (req->data_processing == SQUASHFS_DECOMPRESS) { + req->length = squashfs_decompress(msblk, bh, nr_buffers, + req->offset, req->length, actor); + if (req->length < 0) { + error = -EIO; + goto cleanup; } } - return bh; + /* Last page may have trailing bytes not filled */ + bytes = req->length % PAGE_SIZE; + if (bytes && actor->page[actor->pages - 1]) + zero_user_segment(actor->page[actor->pages - 1], bytes, + PAGE_SIZE); + +cleanup: + if (req->bytes_uncompressed) + *(req->bytes_uncompressed) = req->length; + if (error) { + for (i = 0; i < nr_buffers; ++i) + if (bh[i]) + put_bh(bh[i]); + } + free_read_request(req, error); } +static void read_wq_handler(struct work_struct *work) +{ + squashfs_process_blocks(container_of(work, + struct squashfs_read_request, offload)); +} + +static void squashfs_bio_end_io(struct bio *bio) +{ + int i; + int error = bio->bi_error; + struct squashfs_bio_request *bio_req = bio->bi_private; + + bio_put(bio); + + for (i = 0; i < bio_req->nr_buffers; ++i) { + if (!bio_req->bh[i]) + continue; + if (!error) + set_buffer_uptodate(bio_req->bh[i]); + else + clear_buffer_uptodate(bio_req->bh[i]); + unlock_buffer(bio_req->bh[i]); + } + kfree(bio_req); +} + +static int bh_is_optional(struct squashfs_read_request *req, int idx) +{ + int start_idx, end_idx; + struct squashfs_sb_info *msblk = req->sb->s_fs_info; + + start_idx = (idx * msblk->devblksize - req->offset) / PAGE_CACHE_SIZE; + end_idx = ((idx + 1) * msblk->devblksize - req->offset + 1) / PAGE_CACHE_SIZE; + if (start_idx >= req->output->pages) + return 1; + if (start_idx < 0) + start_idx = end_idx; + if (end_idx >= req->output->pages) + end_idx = start_idx; + return !req->output->page[start_idx] && !req->output->page[end_idx]; +} + +static int actor_getblks(struct squashfs_read_request *req, u64 block) +{ + int i; + + req->bh = kmalloc_array(req->nr_buffers, sizeof(*(req->bh)), GFP_NOIO); + if (!req->bh) + return -ENOMEM; + + for (i = 0; i < req->nr_buffers; ++i) { + /* + * When dealing with an uncompressed block, the actor may + * contains NULL pages. There's no need to read the buffers + * associated with these pages. + */ + if (!req->compressed && bh_is_optional(req, i)) { + req->bh[i] = NULL; + continue; + } + req->bh[i] = sb_getblk(req->sb, block + i); + if (!req->bh[i]) { + while (--i) { + if (req->bh[i]) + put_bh(req->bh[i]); + } + return -1; + } + } + return 0; +} + +static int squashfs_bio_submit(struct squashfs_read_request *req) +{ + struct bio *bio = NULL; + struct buffer_head *bh; + struct squashfs_bio_request *bio_req = NULL; + int b = 0, prev_block = 0; + struct squashfs_sb_info *msblk = req->sb->s_fs_info; + + u64 read_start = round_down(req->index, msblk->devblksize); + u64 read_end = round_up(req->index + req->length, msblk->devblksize); + sector_t block = read_start >> msblk->devblksize_log2; + sector_t block_end = read_end >> msblk->devblksize_log2; + int offset = read_start - round_down(req->index, PAGE_SIZE); + int nr_buffers = block_end - block; + int blksz = msblk->devblksize; + int bio_max_pages = nr_buffers > BIO_MAX_PAGES ? BIO_MAX_PAGES + : nr_buffers; + + /* Setup the request */ + req->read_end = read_end; + req->offset = req->index - read_start; + req->nr_buffers = nr_buffers; + if (actor_getblks(req, block) < 0) + goto getblk_failed; + + /* Create and submit the BIOs */ + for (b = 0; b < nr_buffers; ++b, offset += blksz) { + bh = req->bh[b]; + if (!bh || !trylock_buffer(bh)) + continue; + if (buffer_uptodate(bh)) { + unlock_buffer(bh); + continue; + } + offset %= PAGE_SIZE; + + /* Append the buffer to the current BIO if it is contiguous */ + if (bio && bio_req && prev_block + 1 == b) { + if (bio_add_page(bio, bh->b_page, blksz, offset)) { + bio_req->nr_buffers += 1; + prev_block = b; + continue; + } + } + + /* Otherwise, submit the current BIO and create a new one */ + if (bio) + submit_bio(READ, bio); + bio_req = kcalloc(1, sizeof(struct squashfs_bio_request), + GFP_NOIO); + if (!bio_req) + goto req_alloc_failed; + bio_req->bh = &req->bh[b]; + bio = bio_alloc(GFP_NOIO, bio_max_pages); + if (!bio) + goto bio_alloc_failed; + bio->bi_bdev = req->sb->s_bdev; + bio->bi_iter.bi_sector = (block + b) + << (msblk->devblksize_log2 - 9); + bio->bi_private = bio_req; + bio->bi_end_io = squashfs_bio_end_io; + + bio_add_page(bio, bh->b_page, blksz, offset); + bio_req->nr_buffers += 1; + prev_block = b; + } + if (bio) + submit_bio(READ, bio); + + if (req->synchronous) + squashfs_process_blocks(req); + else { + INIT_WORK(&req->offload, read_wq_handler); + schedule_work(&req->offload); + } + return 0; + +bio_alloc_failed: + kfree(bio_req); +req_alloc_failed: + unlock_buffer(bh); + while (--nr_buffers >= b) + if (req->bh[nr_buffers]) + put_bh(req->bh[nr_buffers]); + while (--b >= 0) + if (req->bh[b]) + wait_on_buffer(req->bh[b]); +getblk_failed: + free_read_request(req, -ENOMEM); + return -ENOMEM; +} + +static int read_metadata_block(struct squashfs_read_request *req, + u64 *next_index) +{ + int ret, error, bytes_read = 0, bytes_uncompressed = 0; + struct squashfs_sb_info *msblk = req->sb->s_fs_info; + + if (req->index + 2 > msblk->bytes_used) { + free_read_request(req, -EINVAL); + return -EINVAL; + } + req->length = 2; + + /* Do not read beyond the end of the device */ + if (req->index + req->length > msblk->bytes_used) + req->length = msblk->bytes_used - req->index; + req->data_processing = SQUASHFS_METADATA; + + /* + * Reading metadata is always synchronous because we don't know the + * length in advance and the function is expected to update + * 'next_index' and return the length. + */ + req->synchronous = true; + req->res = &error; + req->bytes_read = &bytes_read; + req->bytes_uncompressed = &bytes_uncompressed; + + TRACE("Metadata block @ 0x%llx, %scompressed size %d, src size %d\n", + req->index, req->compressed ? "" : "un", bytes_read, + req->output->length); + + ret = squashfs_bio_submit(req); + if (ret) + return ret; + if (error) + return error; + if (next_index) + *next_index += 2 + bytes_read; + return bytes_uncompressed; +} + +static int read_data_block(struct squashfs_read_request *req, int length, + u64 *next_index, bool synchronous) +{ + int ret, error = 0, bytes_uncompressed = 0, bytes_read = 0; + + req->compressed = SQUASHFS_COMPRESSED_BLOCK(length); + req->length = length = SQUASHFS_COMPRESSED_SIZE_BLOCK(length); + req->data_processing = req->compressed ? SQUASHFS_DECOMPRESS + : SQUASHFS_COPY; + + req->synchronous = synchronous; + if (synchronous) { + req->res = &error; + req->bytes_read = &bytes_read; + req->bytes_uncompressed = &bytes_uncompressed; + } + + TRACE("Data block @ 0x%llx, %scompressed size %d, src size %d\n", + req->index, req->compressed ? "" : "un", req->length, + req->output->length); + + ret = squashfs_bio_submit(req); + if (ret) + return ret; + if (synchronous) + ret = error ? error : bytes_uncompressed; + if (next_index) + *next_index += length; + return ret; +} /* * Read and decompress a metadata block or datablock. Length is non-zero @@ -87,128 +425,50 @@ static struct buffer_head *get_block_length(struct super_block *sb, * generated a larger block - this does occasionally happen with compression * algorithms). */ -int squashfs_read_data(struct super_block *sb, u64 index, int length, - u64 *next_index, struct squashfs_page_actor *output) +static int __squashfs_read_data(struct super_block *sb, u64 index, int length, + u64 *next_index, struct squashfs_page_actor *output, bool sync) { - struct squashfs_sb_info *msblk = sb->s_fs_info; - struct buffer_head **bh; - int offset = index & ((1 << msblk->devblksize_log2) - 1); - u64 cur_index = index >> msblk->devblksize_log2; - int bytes, compressed, b = 0, k = 0, avail, i; + struct squashfs_read_request *req; - bh = kcalloc(((output->length + msblk->devblksize - 1) - >> msblk->devblksize_log2) + 1, sizeof(*bh), GFP_KERNEL); - if (bh == NULL) + req = kcalloc(1, sizeof(struct squashfs_read_request), GFP_KERNEL); + if (!req) { + if (!sync) + squashfs_page_actor_free(output, -ENOMEM); return -ENOMEM; - - if (length) { - /* - * Datablock. - */ - bytes = -offset; - compressed = SQUASHFS_COMPRESSED_BLOCK(length); - length = SQUASHFS_COMPRESSED_SIZE_BLOCK(length); - if (next_index) - *next_index = index + length; - - TRACE("Block @ 0x%llx, %scompressed size %d, src size %d\n", - index, compressed ? "" : "un", length, output->length); - - if (length < 0 || length > output->length || - (index + length) > msblk->bytes_used) - goto read_failure; - - for (b = 0; bytes < length; b++, cur_index++) { - bh[b] = sb_getblk(sb, cur_index); - if (bh[b] == NULL) - goto block_release; - bytes += msblk->devblksize; - } - ll_rw_block(READ, b, bh); - } else { - /* - * Metadata block. - */ - if ((index + 2) > msblk->bytes_used) - goto read_failure; - - bh[0] = get_block_length(sb, &cur_index, &offset, &length); - if (bh[0] == NULL) - goto read_failure; - b = 1; - - bytes = msblk->devblksize - offset; - compressed = SQUASHFS_COMPRESSED(length); - length = SQUASHFS_COMPRESSED_SIZE(length); - if (next_index) - *next_index = index + length + 2; - - TRACE("Block @ 0x%llx, %scompressed size %d\n", index, - compressed ? "" : "un", length); - - if (length < 0 || length > output->length || - (index + length) > msblk->bytes_used) - goto block_release; - - for (; bytes < length; b++) { - bh[b] = sb_getblk(sb, ++cur_index); - if (bh[b] == NULL) - goto block_release; - bytes += msblk->devblksize; - } - ll_rw_block(READ, b - 1, bh + 1); } - for (i = 0; i < b; i++) { - wait_on_buffer(bh[i]); - if (!buffer_uptodate(bh[i])) - goto block_release; + req->sb = sb; + req->index = index; + req->output = output; + + if (next_index) + *next_index = index; + + if (length) + length = read_data_block(req, length, next_index, sync); + else + length = read_metadata_block(req, next_index); + + if (length < 0) { + ERROR("squashfs_read_data failed to read block 0x%llx\n", + (unsigned long long)index); + return -EIO; } - if (compressed) { - length = squashfs_decompress(msblk, bh, b, offset, length, - output); - if (length < 0) - goto read_failure; - } else { - /* - * Block is uncompressed. - */ - int in, pg_offset = 0; - void *data = squashfs_first_page(output); - - for (bytes = length; k < b; k++) { - in = min(bytes, msblk->devblksize - offset); - bytes -= in; - while (in) { - if (pg_offset == PAGE_CACHE_SIZE) { - data = squashfs_next_page(output); - pg_offset = 0; - } - avail = min_t(int, in, PAGE_CACHE_SIZE - - pg_offset); - memcpy(data + pg_offset, bh[k]->b_data + offset, - avail); - in -= avail; - pg_offset += avail; - offset += avail; - } - offset = 0; - put_bh(bh[k]); - } - squashfs_finish_page(output); - } - - kfree(bh); return length; - -block_release: - for (; k < b; k++) - put_bh(bh[k]); - -read_failure: - ERROR("squashfs_read_data failed to read block 0x%llx\n", - (unsigned long long) index); - kfree(bh); - return -EIO; +} + +int squashfs_read_data(struct super_block *sb, u64 index, int length, + u64 *next_index, struct squashfs_page_actor *output) +{ + return __squashfs_read_data(sb, index, length, next_index, output, + true); +} + +int squashfs_read_data_async(struct super_block *sb, u64 index, int length, + u64 *next_index, struct squashfs_page_actor *output) +{ + + return __squashfs_read_data(sb, index, length, next_index, output, + false); } diff --git a/fs/squashfs/cache.c b/fs/squashfs/cache.c index 1cb70a0b2168..6785d086ab38 100644 --- a/fs/squashfs/cache.c +++ b/fs/squashfs/cache.c @@ -209,17 +209,14 @@ void squashfs_cache_put(struct squashfs_cache_entry *entry) */ void squashfs_cache_delete(struct squashfs_cache *cache) { - int i, j; + int i; if (cache == NULL) return; for (i = 0; i < cache->entries; i++) { - if (cache->entry[i].data) { - for (j = 0; j < cache->pages; j++) - kfree(cache->entry[i].data[j]); - kfree(cache->entry[i].data); - } + if (cache->entry[i].page) + free_page_array(cache->entry[i].page, cache->pages); kfree(cache->entry[i].actor); } @@ -236,7 +233,7 @@ void squashfs_cache_delete(struct squashfs_cache *cache) struct squashfs_cache *squashfs_cache_init(char *name, int entries, int block_size) { - int i, j; + int i; struct squashfs_cache *cache = kzalloc(sizeof(*cache), GFP_KERNEL); if (cache == NULL) { @@ -268,22 +265,13 @@ struct squashfs_cache *squashfs_cache_init(char *name, int entries, init_waitqueue_head(&cache->entry[i].wait_queue); entry->cache = cache; entry->block = SQUASHFS_INVALID_BLK; - entry->data = kcalloc(cache->pages, sizeof(void *), GFP_KERNEL); - if (entry->data == NULL) { + entry->page = alloc_page_array(cache->pages, GFP_KERNEL); + if (!entry->page) { ERROR("Failed to allocate %s cache entry\n", name); goto cleanup; } - - for (j = 0; j < cache->pages; j++) { - entry->data[j] = kmalloc(PAGE_CACHE_SIZE, GFP_KERNEL); - if (entry->data[j] == NULL) { - ERROR("Failed to allocate %s buffer\n", name); - goto cleanup; - } - } - - entry->actor = squashfs_page_actor_init(entry->data, - cache->pages, 0); + entry->actor = squashfs_page_actor_init(entry->page, + cache->pages, 0, NULL); if (entry->actor == NULL) { ERROR("Failed to allocate %s cache entry\n", name); goto cleanup; @@ -314,18 +302,20 @@ int squashfs_copy_data(void *buffer, struct squashfs_cache_entry *entry, return min(length, entry->length - offset); while (offset < entry->length) { - void *buff = entry->data[offset / PAGE_CACHE_SIZE] - + (offset % PAGE_CACHE_SIZE); + void *buff = kmap_atomic(entry->page[offset / PAGE_CACHE_SIZE]) + + (offset % PAGE_CACHE_SIZE); int bytes = min_t(int, entry->length - offset, PAGE_CACHE_SIZE - (offset % PAGE_CACHE_SIZE)); if (bytes >= remaining) { memcpy(buffer, buff, remaining); + kunmap_atomic(buff); remaining = 0; break; } memcpy(buffer, buff, bytes); + kunmap_atomic(buff); buffer += bytes; remaining -= bytes; offset += bytes; @@ -416,43 +406,38 @@ struct squashfs_cache_entry *squashfs_get_datablock(struct super_block *sb, void *squashfs_read_table(struct super_block *sb, u64 block, int length) { int pages = (length + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT; - int i, res; - void *table, *buffer, **data; + struct page **page; + void *buff; + int res; struct squashfs_page_actor *actor; - table = buffer = kmalloc(length, GFP_KERNEL); - if (table == NULL) + page = alloc_page_array(pages, GFP_KERNEL); + if (!page) return ERR_PTR(-ENOMEM); - data = kcalloc(pages, sizeof(void *), GFP_KERNEL); - if (data == NULL) { + actor = squashfs_page_actor_init(page, pages, length, NULL); + if (actor == NULL) { res = -ENOMEM; goto failed; } - actor = squashfs_page_actor_init(data, pages, length); - if (actor == NULL) { - res = -ENOMEM; - goto failed2; - } - - for (i = 0; i < pages; i++, buffer += PAGE_CACHE_SIZE) - data[i] = buffer; - res = squashfs_read_data(sb, block, length | SQUASHFS_COMPRESSED_BIT_BLOCK, NULL, actor); - kfree(data); - kfree(actor); - if (res < 0) - goto failed; + goto failed2; - return table; + buff = kmalloc(length, GFP_KERNEL); + if (!buff) + goto failed2; + squashfs_actor_to_buf(actor, buff, length); + squashfs_page_actor_free(actor, 0); + free_page_array(page, pages); + return buff; failed2: - kfree(data); + squashfs_page_actor_free(actor, 0); failed: - kfree(table); + free_page_array(page, pages); return ERR_PTR(res); } diff --git a/fs/squashfs/decompressor.c b/fs/squashfs/decompressor.c index e9034bf6e5ae..7de35bf297aa 100644 --- a/fs/squashfs/decompressor.c +++ b/fs/squashfs/decompressor.c @@ -24,7 +24,8 @@ #include #include #include -#include +#include +#include #include "squashfs_fs.h" #include "squashfs_fs_sb.h" @@ -94,40 +95,44 @@ const struct squashfs_decompressor *squashfs_lookup_decompressor(int id) static void *get_comp_opts(struct super_block *sb, unsigned short flags) { struct squashfs_sb_info *msblk = sb->s_fs_info; - void *buffer = NULL, *comp_opts; + void *comp_opts, *buffer = NULL; + struct page *page; struct squashfs_page_actor *actor = NULL; int length = 0; + if (!SQUASHFS_COMP_OPTS(flags)) + return squashfs_comp_opts(msblk, buffer, length); + /* * Read decompressor specific options from file system if present */ - if (SQUASHFS_COMP_OPTS(flags)) { - buffer = kmalloc(PAGE_CACHE_SIZE, GFP_KERNEL); - if (buffer == NULL) { - comp_opts = ERR_PTR(-ENOMEM); - goto out; - } - actor = squashfs_page_actor_init(&buffer, 1, 0); - if (actor == NULL) { - comp_opts = ERR_PTR(-ENOMEM); - goto out; - } + page = alloc_page(GFP_KERNEL); + if (!page) + return ERR_PTR(-ENOMEM); - length = squashfs_read_data(sb, - sizeof(struct squashfs_super_block), 0, NULL, actor); - - if (length < 0) { - comp_opts = ERR_PTR(length); - goto out; - } + actor = squashfs_page_actor_init(&page, 1, 0, NULL); + if (actor == NULL) { + comp_opts = ERR_PTR(-ENOMEM); + goto actor_error; } - comp_opts = squashfs_comp_opts(msblk, buffer, length); + length = squashfs_read_data(sb, + sizeof(struct squashfs_super_block), 0, NULL, actor); -out: - kfree(actor); - kfree(buffer); + if (length < 0) { + comp_opts = ERR_PTR(length); + goto read_error; + } + + buffer = kmap_atomic(page); + comp_opts = squashfs_comp_opts(msblk, buffer, length); + kunmap_atomic(buffer); + +read_error: + squashfs_page_actor_free(actor, 0); +actor_error: + __free_page(page); return comp_opts; } diff --git a/fs/squashfs/file.c b/fs/squashfs/file.c index e5c9689062ba..6f5ef8d7e55a 100644 --- a/fs/squashfs/file.c +++ b/fs/squashfs/file.c @@ -47,12 +47,16 @@ #include #include #include +#include #include "squashfs_fs.h" #include "squashfs_fs_sb.h" #include "squashfs_fs_i.h" #include "squashfs.h" +// Backported from 4.5 +#define lru_to_page(head) (list_entry((head)->prev, struct page, lru)) + /* * Locate cache slot in range [offset, index] for specified inode. If * there's more than one return the slot closest to index. @@ -438,6 +442,21 @@ static int squashfs_readpage_fragment(struct page *page) return res; } +static int squashfs_readpages_fragment(struct page *page, + struct list_head *readahead_pages, struct address_space *mapping) +{ + if (!page) { + page = lru_to_page(readahead_pages); + list_del(&page->lru); + if (add_to_page_cache_lru(page, mapping, page->index, + mapping_gfp_constraint(mapping, GFP_KERNEL))) { + put_page(page); + return 0; + } + } + return squashfs_readpage_fragment(page); +} + static int squashfs_readpage_sparse(struct page *page, int index, int file_end) { struct inode *inode = page->mapping->host; @@ -450,54 +469,105 @@ static int squashfs_readpage_sparse(struct page *page, int index, int file_end) return 0; } -static int squashfs_readpage(struct file *file, struct page *page) +static int squashfs_readpages_sparse(struct page *page, + struct list_head *readahead_pages, int index, int file_end, + struct address_space *mapping) { - struct inode *inode = page->mapping->host; + if (!page) { + page = lru_to_page(readahead_pages); + list_del(&page->lru); + if (add_to_page_cache_lru(page, mapping, page->index, + mapping_gfp_constraint(mapping, GFP_KERNEL))) { + put_page(page); + return 0; + } + } + return squashfs_readpage_sparse(page, index, file_end); +} + +static int __squashfs_readpages(struct file *file, struct page *page, + struct list_head *readahead_pages, unsigned int nr_pages, + struct address_space *mapping) +{ + struct inode *inode = mapping->host; struct squashfs_sb_info *msblk = inode->i_sb->s_fs_info; - int index = page->index >> (msblk->block_log - PAGE_CACHE_SHIFT); int file_end = i_size_read(inode) >> msblk->block_log; int res; - void *pageaddr; + + do { + struct page *cur_page = page ? page + : lru_to_page(readahead_pages); + int page_index = cur_page->index; + int index = page_index >> (msblk->block_log - PAGE_CACHE_SHIFT); + + if (page_index >= ((i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> + PAGE_CACHE_SHIFT)) + return 1; + + if (index < file_end || squashfs_i(inode)->fragment_block == + SQUASHFS_INVALID_BLK) { + u64 block = 0; + int bsize = read_blocklist(inode, index, &block); + + if (bsize < 0) + return -1; + + if (bsize == 0) { + res = squashfs_readpages_sparse(page, + readahead_pages, index, file_end, + mapping); + } else { + res = squashfs_readpages_block(page, + readahead_pages, &nr_pages, mapping, + page_index, block, bsize); + } + } else { + res = squashfs_readpages_fragment(page, + readahead_pages, mapping); + } + if (res) + return 0; + page = NULL; + } while (readahead_pages && !list_empty(readahead_pages)); + + return 0; +} + +static int squashfs_readpage(struct file *file, struct page *page) +{ + int ret; TRACE("Entered squashfs_readpage, page index %lx, start block %llx\n", - page->index, squashfs_i(inode)->start); + page->index, squashfs_i(page->mapping->host)->start); - if (page->index >= ((i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> - PAGE_CACHE_SHIFT)) - goto out; + get_page(page); - if (index < file_end || squashfs_i(inode)->fragment_block == - SQUASHFS_INVALID_BLK) { - u64 block = 0; - int bsize = read_blocklist(inode, index, &block); - if (bsize < 0) - goto error_out; - - if (bsize == 0) - res = squashfs_readpage_sparse(page, index, file_end); + ret = __squashfs_readpages(file, page, NULL, 1, page->mapping); + if (ret) { + flush_dcache_page(page); + if (ret < 0) + SetPageError(page); else - res = squashfs_readpage_block(page, block, bsize); - } else - res = squashfs_readpage_fragment(page); + SetPageUptodate(page); + zero_user_segment(page, 0, PAGE_CACHE_SIZE); + unlock_page(page); + put_page(page); + } - if (!res) - return 0; - -error_out: - SetPageError(page); -out: - pageaddr = kmap_atomic(page); - memset(pageaddr, 0, PAGE_CACHE_SIZE); - kunmap_atomic(pageaddr); - flush_dcache_page(page); - if (!PageError(page)) - SetPageUptodate(page); - unlock_page(page); + return 0; +} +static int squashfs_readpages(struct file *file, struct address_space *mapping, + struct list_head *pages, unsigned int nr_pages) +{ + TRACE("Entered squashfs_readpages, %u pages, first page index %lx\n", + nr_pages, lru_to_page(pages)->index); + __squashfs_readpages(file, NULL, pages, nr_pages, mapping); return 0; } const struct address_space_operations squashfs_aops = { - .readpage = squashfs_readpage + .readpage = squashfs_readpage, + .readpages = squashfs_readpages, }; diff --git a/fs/squashfs/file_cache.c b/fs/squashfs/file_cache.c deleted file mode 100644 index f2310d2a2019..000000000000 --- a/fs/squashfs/file_cache.c +++ /dev/null @@ -1,38 +0,0 @@ -/* - * Copyright (c) 2013 - * Phillip Lougher - * - * This work is licensed under the terms of the GNU GPL, version 2. See - * the COPYING file in the top-level directory. - */ - -#include -#include -#include -#include -#include -#include -#include - -#include "squashfs_fs.h" -#include "squashfs_fs_sb.h" -#include "squashfs_fs_i.h" -#include "squashfs.h" - -/* Read separately compressed datablock and memcopy into page cache */ -int squashfs_readpage_block(struct page *page, u64 block, int bsize) -{ - struct inode *i = page->mapping->host; - struct squashfs_cache_entry *buffer = squashfs_get_datablock(i->i_sb, - block, bsize); - int res = buffer->error; - - if (res) - ERROR("Unable to read page, block %llx, size %x\n", block, - bsize); - else - squashfs_copy_cache(page, buffer, buffer->length, 0); - - squashfs_cache_put(buffer); - return res; -} diff --git a/fs/squashfs/file_direct.c b/fs/squashfs/file_direct.c index 43e7a7eddac0..c97af4c6ccd0 100644 --- a/fs/squashfs/file_direct.c +++ b/fs/squashfs/file_direct.c @@ -13,6 +13,7 @@ #include #include #include +#include #include "squashfs_fs.h" #include "squashfs_fs_sb.h" @@ -20,157 +21,139 @@ #include "squashfs.h" #include "page_actor.h" -static int squashfs_read_cache(struct page *target_page, u64 block, int bsize, - int pages, struct page **page); - -/* Read separately compressed datablock directly into page cache */ -int squashfs_readpage_block(struct page *target_page, u64 block, int bsize) +// Backported from 4.5 +#define lru_to_page(head) (list_entry((head)->prev, struct page, lru)) +static void release_actor_pages(struct page **page, int pages, int error) { - struct inode *inode = target_page->mapping->host; - struct squashfs_sb_info *msblk = inode->i_sb->s_fs_info; + int i; - int file_end = (i_size_read(inode) - 1) >> PAGE_CACHE_SHIFT; - int mask = (1 << (msblk->block_log - PAGE_CACHE_SHIFT)) - 1; - int start_index = target_page->index & ~mask; - int end_index = start_index | mask; - int i, n, pages, missing_pages, bytes, res = -ENOMEM; + for (i = 0; i < pages; i++) { + if (!page[i]) + continue; + flush_dcache_page(page[i]); + if (!error) + SetPageUptodate(page[i]); + else { + SetPageError(page[i]); + zero_user_segment(page[i], 0, PAGE_CACHE_SIZE); + } + unlock_page(page[i]); + put_page(page[i]); + } + kfree(page); +} + +/* + * Create a "page actor" which will kmap and kunmap the + * page cache pages appropriately within the decompressor + */ +static struct squashfs_page_actor *actor_from_page_cache( + unsigned int actor_pages, struct page *target_page, + struct list_head *rpages, unsigned int *nr_pages, int start_index, + struct address_space *mapping) +{ struct page **page; struct squashfs_page_actor *actor; - void *pageaddr; + int i, n; + gfp_t gfp = mapping_gfp_constraint(mapping, GFP_KERNEL); - if (end_index > file_end) - end_index = file_end; + page = kmalloc_array(actor_pages, sizeof(void *), GFP_KERNEL); + if (!page) + return NULL; - pages = end_index - start_index + 1; + for (i = 0, n = start_index; i < actor_pages; i++, n++) { + if (target_page == NULL && rpages && !list_empty(rpages)) { + struct page *cur_page = lru_to_page(rpages); - page = kmalloc_array(pages, sizeof(void *), GFP_KERNEL); - if (page == NULL) - return res; + if (cur_page->index < start_index + actor_pages) { + list_del(&cur_page->lru); + --(*nr_pages); + if (add_to_page_cache_lru(cur_page, mapping, + cur_page->index, gfp)) + put_page(cur_page); + else + target_page = cur_page; + } else + rpages = NULL; + } - /* - * Create a "page actor" which will kmap and kunmap the - * page cache pages appropriately within the decompressor - */ - actor = squashfs_page_actor_init_special(page, pages, 0); - if (actor == NULL) - goto out; - - /* Try to grab all the pages covered by the Squashfs block */ - for (missing_pages = 0, i = 0, n = start_index; i < pages; i++, n++) { - page[i] = (n == target_page->index) ? target_page : - grab_cache_page_nowait(target_page->mapping, n); - - if (page[i] == NULL) { - missing_pages++; - continue; + if (target_page && target_page->index == n) { + page[i] = target_page; + target_page = NULL; + } else { + page[i] = grab_cache_page_nowait(mapping, n); + if (page[i] == NULL) + continue; } if (PageUptodate(page[i])) { unlock_page(page[i]); - page_cache_release(page[i]); + put_page(page[i]); page[i] = NULL; - missing_pages++; } } - if (missing_pages) { - /* - * Couldn't get one or more pages, this page has either - * been VM reclaimed, but others are still in the page cache - * and uptodate, or we're racing with another thread in - * squashfs_readpage also trying to grab them. Fall back to - * using an intermediate buffer. - */ - res = squashfs_read_cache(target_page, block, bsize, pages, - page); - if (res < 0) - goto mark_errored; - - goto out; + actor = squashfs_page_actor_init(page, actor_pages, 0, + release_actor_pages); + if (!actor) { + release_actor_pages(page, actor_pages, -ENOMEM); + kfree(page); + return NULL; } - - /* Decompress directly into the page cache buffers */ - res = squashfs_read_data(inode->i_sb, block, bsize, NULL, actor); - if (res < 0) - goto mark_errored; - - /* Last page may have trailing bytes not filled */ - bytes = res % PAGE_CACHE_SIZE; - if (bytes) { - pageaddr = kmap_atomic(page[pages - 1]); - memset(pageaddr + bytes, 0, PAGE_CACHE_SIZE - bytes); - kunmap_atomic(pageaddr); - } - - /* Mark pages as uptodate, unlock and release */ - for (i = 0; i < pages; i++) { - flush_dcache_page(page[i]); - SetPageUptodate(page[i]); - unlock_page(page[i]); - if (page[i] != target_page) - page_cache_release(page[i]); - } - - kfree(actor); - kfree(page); - - return 0; - -mark_errored: - /* Decompression failed, mark pages as errored. Target_page is - * dealt with by the caller - */ - for (i = 0; i < pages; i++) { - if (page[i] == NULL || page[i] == target_page) - continue; - flush_dcache_page(page[i]); - SetPageError(page[i]); - unlock_page(page[i]); - page_cache_release(page[i]); - } - -out: - kfree(actor); - kfree(page); - return res; + return actor; } +int squashfs_readpages_block(struct page *target_page, + struct list_head *readahead_pages, + unsigned int *nr_pages, + struct address_space *mapping, + int page_index, u64 block, int bsize) -static int squashfs_read_cache(struct page *target_page, u64 block, int bsize, - int pages, struct page **page) { - struct inode *i = target_page->mapping->host; - struct squashfs_cache_entry *buffer = squashfs_get_datablock(i->i_sb, - block, bsize); - int bytes = buffer->length, res = buffer->error, n, offset = 0; - void *pageaddr; + struct squashfs_page_actor *actor; + struct inode *inode = mapping->host; + struct squashfs_sb_info *msblk = inode->i_sb->s_fs_info; + int start_index, end_index, file_end, actor_pages, res; + int mask = (1 << (msblk->block_log - PAGE_CACHE_SHIFT)) - 1; - if (res) { - ERROR("Unable to read page, block %llx, size %x\n", block, - bsize); - goto out; + /* + * If readpage() is called on an uncompressed datablock, we can just + * read the pages instead of fetching the whole block. + * This greatly improves the performance when a process keep doing + * random reads because we only fetch the necessary data. + * The readahead algorithm will take care of doing speculative reads + * if necessary. + * We can't read more than 1 block even if readahead provides use more + * pages because we don't know yet if the next block is compressed or + * not. + */ + if (bsize && !SQUASHFS_COMPRESSED_BLOCK(bsize)) { + u64 block_end = block + msblk->block_size; + + block += (page_index & mask) * PAGE_CACHE_SIZE; + actor_pages = (block_end - block) / PAGE_CACHE_SIZE; + if (*nr_pages < actor_pages) + actor_pages = *nr_pages; + start_index = page_index; + bsize = min_t(int, bsize, (PAGE_CACHE_SIZE * actor_pages) + | SQUASHFS_COMPRESSED_BIT_BLOCK); + } else { + file_end = (i_size_read(inode) - 1) >> PAGE_CACHE_SHIFT; + start_index = page_index & ~mask; + end_index = start_index | mask; + if (end_index > file_end) + end_index = file_end; + actor_pages = end_index - start_index + 1; } - for (n = 0; n < pages && bytes > 0; n++, - bytes -= PAGE_CACHE_SIZE, offset += PAGE_CACHE_SIZE) { - int avail = min_t(int, bytes, PAGE_CACHE_SIZE); + actor = actor_from_page_cache(actor_pages, target_page, + readahead_pages, nr_pages, start_index, + mapping); + if (!actor) + return -ENOMEM; - if (page[n] == NULL) - continue; - - pageaddr = kmap_atomic(page[n]); - squashfs_copy_data(pageaddr, buffer, offset, avail); - memset(pageaddr + avail, 0, PAGE_CACHE_SIZE - avail); - kunmap_atomic(pageaddr); - flush_dcache_page(page[n]); - SetPageUptodate(page[n]); - unlock_page(page[n]); - if (page[n] != target_page) - page_cache_release(page[n]); - } - -out: - squashfs_cache_put(buffer); - return res; + res = squashfs_read_data_async(inode->i_sb, block, bsize, NULL, + actor); + return res < 0 ? res : 0; } diff --git a/fs/squashfs/lz4_wrapper.c b/fs/squashfs/lz4_wrapper.c index c31e2bc9c081..df4fa3c7ddd0 100644 --- a/fs/squashfs/lz4_wrapper.c +++ b/fs/squashfs/lz4_wrapper.c @@ -94,39 +94,17 @@ static int lz4_uncompress(struct squashfs_sb_info *msblk, void *strm, struct buffer_head **bh, int b, int offset, int length, struct squashfs_page_actor *output) { - struct squashfs_lz4 *stream = strm; - void *buff = stream->input, *data; - int avail, i, bytes = length, res; + int res; size_t dest_len = output->length; + struct squashfs_lz4 *stream = strm; - for (i = 0; i < b; i++) { - avail = min(bytes, msblk->devblksize - offset); - memcpy(buff, bh[i]->b_data + offset, avail); - buff += avail; - bytes -= avail; - offset = 0; - put_bh(bh[i]); - } - + squashfs_bh_to_buf(bh, b, stream->input, offset, length, + msblk->devblksize); res = lz4_decompress_unknownoutputsize(stream->input, length, stream->output, &dest_len); if (res) return -EIO; - - bytes = dest_len; - data = squashfs_first_page(output); - buff = stream->output; - while (data) { - if (bytes <= PAGE_CACHE_SIZE) { - memcpy(data, buff, bytes); - break; - } - memcpy(data, buff, PAGE_CACHE_SIZE); - buff += PAGE_CACHE_SIZE; - bytes -= PAGE_CACHE_SIZE; - data = squashfs_next_page(output); - } - squashfs_finish_page(output); + squashfs_buf_to_actor(stream->output, output, dest_len); return dest_len; } diff --git a/fs/squashfs/lzo_wrapper.c b/fs/squashfs/lzo_wrapper.c index 244b9fbfff7b..2c844d53a59e 100644 --- a/fs/squashfs/lzo_wrapper.c +++ b/fs/squashfs/lzo_wrapper.c @@ -79,45 +79,19 @@ static int lzo_uncompress(struct squashfs_sb_info *msblk, void *strm, struct buffer_head **bh, int b, int offset, int length, struct squashfs_page_actor *output) { - struct squashfs_lzo *stream = strm; - void *buff = stream->input, *data; - int avail, i, bytes = length, res; + int res; size_t out_len = output->length; + struct squashfs_lzo *stream = strm; - for (i = 0; i < b; i++) { - avail = min(bytes, msblk->devblksize - offset); - memcpy(buff, bh[i]->b_data + offset, avail); - buff += avail; - bytes -= avail; - offset = 0; - put_bh(bh[i]); - } - + squashfs_bh_to_buf(bh, b, stream->input, offset, length, + msblk->devblksize); res = lzo1x_decompress_safe(stream->input, (size_t)length, stream->output, &out_len); if (res != LZO_E_OK) - goto failed; + return -EIO; + squashfs_buf_to_actor(stream->output, output, out_len); - res = bytes = (int)out_len; - data = squashfs_first_page(output); - buff = stream->output; - while (data) { - if (bytes <= PAGE_CACHE_SIZE) { - memcpy(data, buff, bytes); - break; - } else { - memcpy(data, buff, PAGE_CACHE_SIZE); - buff += PAGE_CACHE_SIZE; - bytes -= PAGE_CACHE_SIZE; - data = squashfs_next_page(output); - } - } - squashfs_finish_page(output); - - return res; - -failed: - return -EIO; + return out_len; } const struct squashfs_decompressor squashfs_lzo_comp_ops = { diff --git a/fs/squashfs/page_actor.c b/fs/squashfs/page_actor.c index 5a1c11f56441..53863508e400 100644 --- a/fs/squashfs/page_actor.c +++ b/fs/squashfs/page_actor.c @@ -9,79 +9,11 @@ #include #include #include +#include #include "page_actor.h" -/* - * This file contains implementations of page_actor for decompressing into - * an intermediate buffer, and for decompressing directly into the - * page cache. - * - * Calling code should avoid sleeping between calls to squashfs_first_page() - * and squashfs_finish_page(). - */ - -/* Implementation of page_actor for decompressing into intermediate buffer */ -static void *cache_first_page(struct squashfs_page_actor *actor) -{ - actor->next_page = 1; - return actor->buffer[0]; -} - -static void *cache_next_page(struct squashfs_page_actor *actor) -{ - if (actor->next_page == actor->pages) - return NULL; - - return actor->buffer[actor->next_page++]; -} - -static void cache_finish_page(struct squashfs_page_actor *actor) -{ - /* empty */ -} - -struct squashfs_page_actor *squashfs_page_actor_init(void **buffer, - int pages, int length) -{ - struct squashfs_page_actor *actor = kmalloc(sizeof(*actor), GFP_KERNEL); - - if (actor == NULL) - return NULL; - - actor->length = length ? : pages * PAGE_CACHE_SIZE; - actor->buffer = buffer; - actor->pages = pages; - actor->next_page = 0; - actor->squashfs_first_page = cache_first_page; - actor->squashfs_next_page = cache_next_page; - actor->squashfs_finish_page = cache_finish_page; - return actor; -} - -/* Implementation of page_actor for decompressing directly into page cache. */ -static void *direct_first_page(struct squashfs_page_actor *actor) -{ - actor->next_page = 1; - return actor->pageaddr = kmap_atomic(actor->page[0]); -} - -static void *direct_next_page(struct squashfs_page_actor *actor) -{ - if (actor->pageaddr) - kunmap_atomic(actor->pageaddr); - - return actor->pageaddr = actor->next_page == actor->pages ? NULL : - kmap_atomic(actor->page[actor->next_page++]); -} - -static void direct_finish_page(struct squashfs_page_actor *actor) -{ - if (actor->pageaddr) - kunmap_atomic(actor->pageaddr); -} - -struct squashfs_page_actor *squashfs_page_actor_init_special(struct page **page, - int pages, int length) +struct squashfs_page_actor *squashfs_page_actor_init(struct page **page, + int pages, int length, void (*release_pages)(struct page **, int, int)) { struct squashfs_page_actor *actor = kmalloc(sizeof(*actor), GFP_KERNEL); @@ -93,8 +25,129 @@ struct squashfs_page_actor *squashfs_page_actor_init_special(struct page **page, actor->pages = pages; actor->next_page = 0; actor->pageaddr = NULL; - actor->squashfs_first_page = direct_first_page; - actor->squashfs_next_page = direct_next_page; - actor->squashfs_finish_page = direct_finish_page; + actor->release_pages = release_pages; return actor; } + +void squashfs_page_actor_free(struct squashfs_page_actor *actor, int error) +{ + if (!actor) + return; + + if (actor->release_pages) + actor->release_pages(actor->page, actor->pages, error); + kfree(actor); +} + +void squashfs_actor_to_buf(struct squashfs_page_actor *actor, void *buf, + int length) +{ + void *pageaddr; + int pos = 0, avail, i; + + for (i = 0; i < actor->pages && pos < length; ++i) { + avail = min_t(int, length - pos, PAGE_CACHE_SIZE); + if (actor->page[i]) { + pageaddr = kmap_atomic(actor->page[i]); + memcpy(buf + pos, pageaddr, avail); + kunmap_atomic(pageaddr); + } + pos += avail; + } +} + +void squashfs_buf_to_actor(void *buf, struct squashfs_page_actor *actor, + int length) +{ + void *pageaddr; + int pos = 0, avail, i; + + for (i = 0; i < actor->pages && pos < length; ++i) { + avail = min_t(int, length - pos, PAGE_CACHE_SIZE); + if (actor->page[i]) { + pageaddr = kmap_atomic(actor->page[i]); + memcpy(pageaddr, buf + pos, avail); + kunmap_atomic(pageaddr); + } + pos += avail; + } +} + +void squashfs_bh_to_actor(struct buffer_head **bh, int nr_buffers, + struct squashfs_page_actor *actor, int offset, int length, int blksz) +{ + void *kaddr = NULL; + int bytes = 0, pgoff = 0, b = 0, p = 0, avail, i; + + while (bytes < length) { + if (actor->page[p]) { + kaddr = kmap_atomic(actor->page[p]); + while (pgoff < PAGE_CACHE_SIZE && bytes < length) { + avail = min_t(int, blksz - offset, + PAGE_CACHE_SIZE - pgoff); + memcpy(kaddr + pgoff, bh[b]->b_data + offset, + avail); + pgoff += avail; + bytes += avail; + offset = (offset + avail) % blksz; + if (!offset) { + put_bh(bh[b]); + ++b; + } + } + kunmap_atomic(kaddr); + pgoff = 0; + } else { + for (i = 0; i < PAGE_CACHE_SIZE / blksz; ++i) { + if (bh[b]) + put_bh(bh[b]); + ++b; + } + bytes += PAGE_CACHE_SIZE; + } + ++p; + } +} + +void squashfs_bh_to_buf(struct buffer_head **bh, int nr_buffers, void *buf, + int offset, int length, int blksz) +{ + int i, avail, bytes = 0; + + for (i = 0; i < nr_buffers && bytes < length; ++i) { + avail = min_t(int, length - bytes, blksz - offset); + if (bh[i]) { + memcpy(buf + bytes, bh[i]->b_data + offset, avail); + put_bh(bh[i]); + } + bytes += avail; + offset = 0; + } +} + +void free_page_array(struct page **page, int nr_pages) +{ + int i; + + for (i = 0; i < nr_pages; ++i) + __free_page(page[i]); + kfree(page); +} + +struct page **alloc_page_array(int nr_pages, int gfp_mask) +{ + int i; + struct page **page; + + page = kcalloc(nr_pages, sizeof(struct page *), gfp_mask); + if (!page) + return NULL; + for (i = 0; i < nr_pages; ++i) { + page[i] = alloc_page(gfp_mask); + if (!page[i]) { + free_page_array(page, i); + return NULL; + } + } + return page; +} diff --git a/fs/squashfs/page_actor.h b/fs/squashfs/page_actor.h index 26dd82008b82..aa1ed790b5a3 100644 --- a/fs/squashfs/page_actor.h +++ b/fs/squashfs/page_actor.h @@ -5,77 +5,61 @@ * Phillip Lougher * * This work is licensed under the terms of the GNU GPL, version 2. See - * the COPYING file in the top-level directory. + * the COPYING file in the top-level squashfsory. */ -#ifndef CONFIG_SQUASHFS_FILE_DIRECT struct squashfs_page_actor { - void **page; + struct page **page; + void *pageaddr; int pages; int length; int next_page; + void (*release_pages)(struct page **, int, int); }; -static inline struct squashfs_page_actor *squashfs_page_actor_init(void **page, - int pages, int length) -{ - struct squashfs_page_actor *actor = kmalloc(sizeof(*actor), GFP_KERNEL); +extern struct squashfs_page_actor *squashfs_page_actor_init(struct page **, + int, int, void (*)(struct page **, int, int)); +extern void squashfs_page_actor_free(struct squashfs_page_actor *, int); - if (actor == NULL) - return NULL; - - actor->length = length ? : pages * PAGE_CACHE_SIZE; - actor->page = page; - actor->pages = pages; - actor->next_page = 0; - return actor; -} +extern void squashfs_actor_to_buf(struct squashfs_page_actor *, void *, int); +extern void squashfs_buf_to_actor(void *, struct squashfs_page_actor *, int); +extern void squashfs_bh_to_actor(struct buffer_head **, int, + struct squashfs_page_actor *, int, int, int); +extern void squashfs_bh_to_buf(struct buffer_head **, int, void *, int, int, + int); +/* + * Calling code should avoid sleeping between calls to squashfs_first_page() + * and squashfs_finish_page(). + */ static inline void *squashfs_first_page(struct squashfs_page_actor *actor) { actor->next_page = 1; - return actor->page[0]; + return actor->pageaddr = actor->page[0] ? kmap_atomic(actor->page[0]) + : NULL; } static inline void *squashfs_next_page(struct squashfs_page_actor *actor) { - return actor->next_page == actor->pages ? NULL : - actor->page[actor->next_page++]; + if (!IS_ERR_OR_NULL(actor->pageaddr)) + kunmap_atomic(actor->pageaddr); + + if (actor->next_page == actor->pages) + return actor->pageaddr = ERR_PTR(-ENODATA); + + actor->pageaddr = actor->page[actor->next_page] ? + kmap_atomic(actor->page[actor->next_page]) : NULL; + ++actor->next_page; + return actor->pageaddr; } static inline void squashfs_finish_page(struct squashfs_page_actor *actor) { - /* empty */ + if (!IS_ERR_OR_NULL(actor->pageaddr)) + kunmap_atomic(actor->pageaddr); } -#else -struct squashfs_page_actor { - union { - void **buffer; - struct page **page; - }; - void *pageaddr; - void *(*squashfs_first_page)(struct squashfs_page_actor *); - void *(*squashfs_next_page)(struct squashfs_page_actor *); - void (*squashfs_finish_page)(struct squashfs_page_actor *); - int pages; - int length; - int next_page; -}; -extern struct squashfs_page_actor *squashfs_page_actor_init(void **, int, int); -extern struct squashfs_page_actor *squashfs_page_actor_init_special(struct page - **, int, int); -static inline void *squashfs_first_page(struct squashfs_page_actor *actor) -{ - return actor->squashfs_first_page(actor); -} -static inline void *squashfs_next_page(struct squashfs_page_actor *actor) -{ - return actor->squashfs_next_page(actor); -} -static inline void squashfs_finish_page(struct squashfs_page_actor *actor) -{ - actor->squashfs_finish_page(actor); -} -#endif +extern struct page **alloc_page_array(int, int); +extern void free_page_array(struct page **, int); + #endif diff --git a/fs/squashfs/squashfs.h b/fs/squashfs/squashfs.h index 887d6d270080..6093579c6c5d 100644 --- a/fs/squashfs/squashfs.h +++ b/fs/squashfs/squashfs.h @@ -28,8 +28,14 @@ #define WARNING(s, args...) pr_warn("SQUASHFS: "s, ## args) /* block.c */ +extern int squashfs_init_read_wq(void); +extern void squashfs_destroy_read_wq(void); extern int squashfs_read_data(struct super_block *, u64, int, u64 *, struct squashfs_page_actor *); +extern int squashfs_read_data(struct super_block *, u64, int, u64 *, + struct squashfs_page_actor *); +extern int squashfs_read_data_async(struct super_block *, u64, int, u64 *, + struct squashfs_page_actor *); /* cache.c */ extern struct squashfs_cache *squashfs_cache_init(char *, int, int); @@ -70,8 +76,9 @@ extern __le64 *squashfs_read_fragment_index_table(struct super_block *, void squashfs_copy_cache(struct page *, struct squashfs_cache_entry *, int, int); -/* file_xxx.c */ -extern int squashfs_readpage_block(struct page *, u64, int); +/* file_direct.c */ +extern int squashfs_readpages_block(struct page *, struct list_head *, + unsigned int *, struct address_space *, int, u64, int); /* id.c */ extern int squashfs_get_id(struct super_block *, unsigned int, unsigned int *); diff --git a/fs/squashfs/squashfs_fs_sb.h b/fs/squashfs/squashfs_fs_sb.h index 1da565cb50c3..8a6995de0277 100644 --- a/fs/squashfs/squashfs_fs_sb.h +++ b/fs/squashfs/squashfs_fs_sb.h @@ -49,7 +49,7 @@ struct squashfs_cache_entry { int num_waiters; wait_queue_head_t wait_queue; struct squashfs_cache *cache; - void **data; + struct page **page; struct squashfs_page_actor *actor; }; diff --git a/fs/squashfs/super.c b/fs/squashfs/super.c index 5056babe00df..61cd0b39ed0e 100644 --- a/fs/squashfs/super.c +++ b/fs/squashfs/super.c @@ -444,9 +444,15 @@ static int __init init_squashfs_fs(void) if (err) return err; + if (!squashfs_init_read_wq()) { + destroy_inodecache(); + return -ENOMEM; + } + err = register_filesystem(&squashfs_fs_type); if (err) { destroy_inodecache(); + squashfs_destroy_read_wq(); return err; } @@ -460,6 +466,7 @@ static void __exit exit_squashfs_fs(void) { unregister_filesystem(&squashfs_fs_type); destroy_inodecache(); + squashfs_destroy_read_wq(); } diff --git a/fs/squashfs/xz_wrapper.c b/fs/squashfs/xz_wrapper.c index c609624e4b8a..14cd373e1897 100644 --- a/fs/squashfs/xz_wrapper.c +++ b/fs/squashfs/xz_wrapper.c @@ -55,7 +55,7 @@ static void *squashfs_xz_comp_opts(struct squashfs_sb_info *msblk, struct comp_opts *opts; int err = 0, n; - opts = kmalloc(sizeof(*opts), GFP_KERNEL); + opts = kmalloc(sizeof(*opts), GFP_ATOMIC); if (opts == NULL) { err = -ENOMEM; goto out2; @@ -136,6 +136,7 @@ static int squashfs_xz_uncompress(struct squashfs_sb_info *msblk, void *strm, enum xz_ret xz_err; int avail, total = 0, k = 0; struct squashfs_xz *stream = strm; + void *buf = NULL; xz_dec_reset(stream->state); stream->buf.in_pos = 0; @@ -156,12 +157,20 @@ static int squashfs_xz_uncompress(struct squashfs_sb_info *msblk, void *strm, if (stream->buf.out_pos == stream->buf.out_size) { stream->buf.out = squashfs_next_page(output); - if (stream->buf.out != NULL) { + if (!IS_ERR(stream->buf.out)) { stream->buf.out_pos = 0; total += PAGE_CACHE_SIZE; } } + if (!stream->buf.out) { + if (!buf) { + buf = kmalloc(PAGE_CACHE_SIZE, GFP_ATOMIC); + if (!buf) + goto out; + } + stream->buf.out = buf; + } xz_err = xz_dec_run(stream->state, &stream->buf); if (stream->buf.in_pos == stream->buf.in_size && k < b) @@ -173,11 +182,13 @@ static int squashfs_xz_uncompress(struct squashfs_sb_info *msblk, void *strm, if (xz_err != XZ_STREAM_END || k < b) goto out; + kfree(buf); return total + stream->buf.out_pos; out: for (; k < b; k++) put_bh(bh[k]); + kfree(buf); return -EIO; } diff --git a/fs/squashfs/zlib_wrapper.c b/fs/squashfs/zlib_wrapper.c index 8727caba6882..09c892b5308e 100644 --- a/fs/squashfs/zlib_wrapper.c +++ b/fs/squashfs/zlib_wrapper.c @@ -66,6 +66,7 @@ static int zlib_uncompress(struct squashfs_sb_info *msblk, void *strm, struct buffer_head **bh, int b, int offset, int length, struct squashfs_page_actor *output) { + void *buf = NULL; int zlib_err, zlib_init = 0, k = 0; z_stream *stream = strm; @@ -84,10 +85,19 @@ static int zlib_uncompress(struct squashfs_sb_info *msblk, void *strm, if (stream->avail_out == 0) { stream->next_out = squashfs_next_page(output); - if (stream->next_out != NULL) + if (!IS_ERR(stream->next_out)) stream->avail_out = PAGE_CACHE_SIZE; } + if (!stream->next_out) { + if (!buf) { + buf = kmalloc(PAGE_CACHE_SIZE, GFP_ATOMIC); + if (!buf) + goto out; + } + stream->next_out = buf; + } + if (!zlib_init) { zlib_err = zlib_inflateInit(stream); if (zlib_err != Z_OK) { @@ -115,11 +125,13 @@ static int zlib_uncompress(struct squashfs_sb_info *msblk, void *strm, if (k < b) goto out; + kfree(buf); return stream->total_out; out: for (; k < b; k++) put_bh(bh[k]); + kfree(buf); return -EIO; } diff --git a/fs/super.c b/fs/super.c index b938b14f6041..c96434ea71e2 100644 --- a/fs/super.c +++ b/fs/super.c @@ -703,7 +703,8 @@ rescan: } /** - * do_remount_sb - asks filesystem to change mount options. + * do_remount_sb2 - asks filesystem to change mount options. + * @mnt: mount we are looking at * @sb: superblock in question * @flags: numeric part of options * @data: the rest of options @@ -711,7 +712,7 @@ rescan: * * Alters the mount options of a mounted file system. */ -int do_remount_sb(struct super_block *sb, int flags, void *data, int force) +int do_remount_sb2(struct vfsmount *mnt, struct super_block *sb, int flags, void *data, int force) { int retval; int remount_ro; @@ -753,7 +754,16 @@ int do_remount_sb(struct super_block *sb, int flags, void *data, int force) } } - if (sb->s_op->remount_fs) { + if (mnt && sb->s_op->remount_fs2) { + retval = sb->s_op->remount_fs2(mnt, sb, &flags, data); + if (retval) { + if (!force) + goto cancel_readonly; + /* If forced remount, go ahead despite any errors */ + WARN(1, "forced remount of a %s fs returned %i\n", + sb->s_type->name, retval); + } + } else if (sb->s_op->remount_fs) { retval = sb->s_op->remount_fs(sb, &flags, data); if (retval) { if (!force) @@ -785,6 +795,11 @@ cancel_readonly: return retval; } +int do_remount_sb(struct super_block *sb, int flags, void *data, int force) +{ + return do_remount_sb2(NULL, sb, flags, data, force); +} + static void do_emergency_remount(struct work_struct *work) { struct super_block *sb, *p = NULL; @@ -1104,7 +1119,7 @@ struct dentry *mount_single(struct file_system_type *fs_type, EXPORT_SYMBOL(mount_single); struct dentry * -mount_fs(struct file_system_type *type, int flags, const char *name, void *data) +mount_fs(struct file_system_type *type, int flags, const char *name, struct vfsmount *mnt, void *data) { struct dentry *root; struct super_block *sb; @@ -1121,7 +1136,10 @@ mount_fs(struct file_system_type *type, int flags, const char *name, void *data) goto out_free_secdata; } - root = type->mount(type, flags, name, data); + if (type->mount2) + root = type->mount2(mnt, type, flags, name, data); + else + root = type->mount(type, flags, name, data); if (IS_ERR(root)) { error = PTR_ERR(root); goto out_free_secdata; diff --git a/fs/utimes.c b/fs/utimes.c index cb771c30d102..a35e909cf8e3 100644 --- a/fs/utimes.c +++ b/fs/utimes.c @@ -91,7 +91,7 @@ static int utimes_common(struct path *path, struct timespec *times) } retry_deleg: mutex_lock(&inode->i_mutex); - error = notify_change(path->dentry, &newattrs, &delegated_inode); + error = notify_change2(path->mnt, path->dentry, &newattrs, &delegated_inode); mutex_unlock(&inode->i_mutex); if (delegated_inode) { error = break_deleg_wait(&delegated_inode); diff --git a/include/crypto/algapi.h b/include/crypto/algapi.h index c9fe145f7dd3..04661e1fb625 100644 --- a/include/crypto/algapi.h +++ b/include/crypto/algapi.h @@ -202,6 +202,9 @@ int blkcipher_aead_walk_virt_block(struct blkcipher_desc *desc, struct blkcipher_walk *walk, struct crypto_aead *tfm, unsigned int blocksize); +int blkcipher_ablkcipher_walk_virt(struct blkcipher_desc *desc, + struct blkcipher_walk *walk, + struct crypto_ablkcipher *tfm); int ablkcipher_walk_done(struct ablkcipher_request *req, struct ablkcipher_walk *walk, int err); diff --git a/include/crypto/gf128mul.h b/include/crypto/gf128mul.h index da2530e34b26..230760aef93b 100644 --- a/include/crypto/gf128mul.h +++ b/include/crypto/gf128mul.h @@ -43,7 +43,7 @@ --------------------------------------------------------------------------- Issue Date: 31/01/2006 - An implementation of field multiplication in Galois Field GF(128) + An implementation of field multiplication in Galois Field GF(2^128) */ #ifndef _CRYPTO_GF128MUL_H @@ -65,7 +65,7 @@ * are left and the lsb's are right. char b[16] is an array and b[0] is * the first octet. * - * 80000000 00000000 00000000 00000000 .... 00000000 00000000 00000000 + * 10000000 00000000 00000000 00000000 .... 00000000 00000000 00000000 * b[0] b[1] b[2] b[3] b[13] b[14] b[15] * * Every bit is a coefficient of some power of X. We can store the bits @@ -99,21 +99,21 @@ * * bbe on a little endian machine u32 x[4]: * - * MS x[0] LS MS x[1] LS + * MS x[0] LS MS x[1] LS * ms ls ms ls ms ls ms ls ms ls ms ls ms ls ms ls * 103..96 111.104 119.112 127.120 71...64 79...72 87...80 95...88 * - * MS x[2] LS MS x[3] LS + * MS x[2] LS MS x[3] LS * ms ls ms ls ms ls ms ls ms ls ms ls ms ls ms ls * 39...32 47...40 55...48 63...56 07...00 15...08 23...16 31...24 * * ble on a little endian machine * - * MS x[0] LS MS x[1] LS + * MS x[0] LS MS x[1] LS * ms ls ms ls ms ls ms ls ms ls ms ls ms ls ms ls * 31...24 23...16 15...08 07...00 63...56 55...48 47...40 39...32 * - * MS x[2] LS MS x[3] LS + * MS x[2] LS MS x[3] LS * ms ls ms ls ms ls ms ls ms ls ms ls ms ls ms ls * 95...88 87...80 79...72 71...64 127.120 199.112 111.104 103..96 * @@ -127,7 +127,7 @@ * machines this will automatically aligned to wordsize and on a 64-bit * machine also. */ -/* Multiply a GF128 field element by x. Field elements are held in arrays +/* Multiply a GF128 field element by x. Field elements are held in arrays of bytes in which field bits 8n..8n + 7 are held in byte[n], with lower indexed bits placed in the more numerically significant bit positions within bytes. @@ -135,62 +135,65 @@ On little endian machines the bit indexes translate into the bit positions within four 32-bit words in the following way - MS x[0] LS MS x[1] LS + MS x[0] LS MS x[1] LS ms ls ms ls ms ls ms ls ms ls ms ls ms ls ms ls 24...31 16...23 08...15 00...07 56...63 48...55 40...47 32...39 - MS x[2] LS MS x[3] LS + MS x[2] LS MS x[3] LS ms ls ms ls ms ls ms ls ms ls ms ls ms ls ms ls 88...95 80...87 72...79 64...71 120.127 112.119 104.111 96..103 On big endian machines the bit indexes translate into the bit positions within four 32-bit words in the following way - MS x[0] LS MS x[1] LS + MS x[0] LS MS x[1] LS ms ls ms ls ms ls ms ls ms ls ms ls ms ls ms ls 00...07 08...15 16...23 24...31 32...39 40...47 48...55 56...63 - MS x[2] LS MS x[3] LS + MS x[2] LS MS x[3] LS ms ls ms ls ms ls ms ls ms ls ms ls ms ls ms ls 64...71 72...79 80...87 88...95 96..103 104.111 112.119 120.127 */ -/* A slow generic version of gf_mul, implemented for lle and bbe - * It multiplies a and b and puts the result in a */ +/* A slow generic version of gf_mul, implemented for lle, bbe, and ble. + * It multiplies a and b and puts the result in a + */ void gf128mul_lle(be128 *a, const be128 *b); - void gf128mul_bbe(be128 *a, const be128 *b); +void gf128mul_ble(be128 *a, const be128 *b); -/* multiply by x in ble format, needed by XTS */ +/* multiply by x in ble format, needed by XTS and HEH */ void gf128mul_x_ble(be128 *a, const be128 *b); /* 4k table optimization */ - struct gf128mul_4k { be128 t[256]; }; struct gf128mul_4k *gf128mul_init_4k_lle(const be128 *g); struct gf128mul_4k *gf128mul_init_4k_bbe(const be128 *g); +struct gf128mul_4k *gf128mul_init_4k_ble(const be128 *g); void gf128mul_4k_lle(be128 *a, struct gf128mul_4k *t); void gf128mul_4k_bbe(be128 *a, struct gf128mul_4k *t); +void gf128mul_4k_ble(be128 *a, struct gf128mul_4k *t); static inline void gf128mul_free_4k(struct gf128mul_4k *t) { - kfree(t); + kzfree(t); } -/* 64k table optimization, implemented for lle and bbe */ +/* 64k table optimization, implemented for lle, ble, and bbe */ struct gf128mul_64k { struct gf128mul_4k *t[16]; }; -/* first initialize with the constant factor with which you - * want to multiply and then call gf128_64k_lle with the other - * factor in the first argument, the table in the second and a - * scratch register in the third. Afterwards *a = *r. */ +/* First initialize with the constant factor with which you + * want to multiply and then call gf128mul_64k_bbe with the other + * factor in the first argument, and the table in the second. + * Afterwards, the result is stored in *a. + */ struct gf128mul_64k *gf128mul_init_64k_lle(const be128 *g); struct gf128mul_64k *gf128mul_init_64k_bbe(const be128 *g); void gf128mul_free_64k(struct gf128mul_64k *t); diff --git a/include/crypto/internal/hash.h b/include/crypto/internal/hash.h index 3b4af1d7c7e9..476d99d0edb7 100644 --- a/include/crypto/internal/hash.h +++ b/include/crypto/internal/hash.h @@ -102,6 +102,8 @@ int shash_register_instance(struct crypto_template *tmpl, struct shash_instance *inst); void shash_free_instance(struct crypto_instance *inst); +int crypto_grab_shash(struct crypto_shash_spawn *spawn, + const char *name, u32 type, u32 mask); int crypto_init_shash_spawn(struct crypto_shash_spawn *spawn, struct shash_alg *alg, struct crypto_instance *inst); @@ -111,6 +113,12 @@ static inline void crypto_drop_shash(struct crypto_shash_spawn *spawn) crypto_drop_spawn(&spawn->base); } +static inline struct shash_alg *crypto_spawn_shash_alg( + struct crypto_shash_spawn *spawn) +{ + return container_of(spawn->base.alg, struct shash_alg, base); +} + struct shash_alg *shash_attr_alg(struct rtattr *rta, u32 type, u32 mask); int shash_ahash_update(struct ahash_request *req, struct shash_desc *desc); diff --git a/include/linux/fs.h b/include/linux/fs.h index e1a123760dbf..7f6de206cc5c 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1539,13 +1539,21 @@ extern bool inode_owner_or_capable(const struct inode *inode); * VFS helper functions.. */ extern int vfs_create(struct inode *, struct dentry *, umode_t, bool); +extern int vfs_create2(struct vfsmount *, struct inode *, struct dentry *, umode_t, bool); extern int vfs_mkdir(struct inode *, struct dentry *, umode_t); +extern int vfs_mkdir2(struct vfsmount *, struct inode *, struct dentry *, umode_t); extern int vfs_mknod(struct inode *, struct dentry *, umode_t, dev_t); +extern int vfs_mknod2(struct vfsmount *, struct inode *, struct dentry *, umode_t, dev_t); extern int vfs_symlink(struct inode *, struct dentry *, const char *); +extern int vfs_symlink2(struct vfsmount *, struct inode *, struct dentry *, const char *); extern int vfs_link(struct dentry *, struct inode *, struct dentry *, struct inode **); +extern int vfs_link2(struct vfsmount *, struct dentry *, struct inode *, struct dentry *, struct inode **); extern int vfs_rmdir(struct inode *, struct dentry *); +extern int vfs_rmdir2(struct vfsmount *, struct inode *, struct dentry *); extern int vfs_unlink(struct inode *, struct dentry *, struct inode **); +extern int vfs_unlink2(struct vfsmount *, struct inode *, struct dentry *, struct inode **); extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *, struct inode **, unsigned int); +extern int vfs_rename2(struct vfsmount *, struct inode *, struct dentry *, struct inode *, struct dentry *, struct inode **, unsigned int); extern int vfs_whiteout(struct inode *, struct dentry *); /* @@ -1671,6 +1679,7 @@ struct inode_operations { struct dentry * (*lookup) (struct inode *,struct dentry *, unsigned int); const char * (*follow_link) (struct dentry *, void **); int (*permission) (struct inode *, int); + int (*permission2) (struct vfsmount *, struct inode *, int); struct posix_acl * (*get_acl)(struct inode *, int); int (*readlink) (struct dentry *, char __user *,int); @@ -1688,6 +1697,7 @@ struct inode_operations { int (*rename2) (struct inode *, struct dentry *, struct inode *, struct dentry *, unsigned int); int (*setattr) (struct dentry *, struct iattr *); + int (*setattr2) (struct vfsmount *, struct dentry *, struct iattr *); int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *); int (*setxattr) (struct dentry *, const char *,const void *,size_t,int); ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t); @@ -1733,9 +1743,13 @@ struct super_operations { int (*unfreeze_fs) (struct super_block *); int (*statfs) (struct dentry *, struct kstatfs *); int (*remount_fs) (struct super_block *, int *, char *); + int (*remount_fs2) (struct vfsmount *, struct super_block *, int *, char *); + void *(*clone_mnt_data) (void *); + void (*copy_mnt_data) (void *, void *); void (*umount_begin) (struct super_block *); int (*show_options)(struct seq_file *, struct dentry *); + int (*show_options2)(struct vfsmount *,struct seq_file *, struct dentry *); int (*show_devname)(struct seq_file *, struct dentry *); int (*show_path)(struct seq_file *, struct dentry *); int (*show_stats)(struct seq_file *, struct dentry *); @@ -1967,6 +1981,9 @@ struct file_system_type { #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move() during rename() internally. */ struct dentry *(*mount) (struct file_system_type *, int, const char *, void *); + struct dentry *(*mount2) (struct vfsmount *, struct file_system_type *, int, + const char *, void *); + void *(*alloc_mnt_data) (void); void (*kill_sb) (struct super_block *); struct module *owner; struct file_system_type * next; @@ -2246,6 +2263,8 @@ struct filename { extern long vfs_truncate(struct path *, loff_t); extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs, struct file *filp); +extern int do_truncate2(struct vfsmount *, struct dentry *, loff_t start, + unsigned int time_attrs, struct file *filp); extern int vfs_fallocate(struct file *file, int mode, loff_t offset, loff_t len); extern long do_sys_open(int dfd, const char __user *filename, int flags, @@ -2470,8 +2489,11 @@ extern void emergency_remount(void); extern sector_t bmap(struct inode *, sector_t); #endif extern int notify_change(struct dentry *, struct iattr *, struct inode **); +extern int notify_change2(struct vfsmount *, struct dentry *, struct iattr *, struct inode **); extern int inode_permission(struct inode *, int); +extern int inode_permission2(struct vfsmount *, struct inode *, int); extern int __inode_permission(struct inode *, int); +extern int __inode_permission2(struct vfsmount *, struct inode *, int); extern int generic_permission(struct inode *, int); extern int __check_sticky(struct inode *dir, struct inode *inode); diff --git a/include/linux/mount.h b/include/linux/mount.h index f822c3c11377..0e9b0977237a 100644 --- a/include/linux/mount.h +++ b/include/linux/mount.h @@ -67,6 +67,7 @@ struct vfsmount { struct dentry *mnt_root; /* root of the mounted tree */ struct super_block *mnt_sb; /* pointer to superblock */ int mnt_flags; + void *data; }; struct file; /* forward dec */ diff --git a/include/linux/namei.h b/include/linux/namei.h index d53c25453aca..023359f18567 100644 --- a/include/linux/namei.h +++ b/include/linux/namei.h @@ -79,6 +79,7 @@ extern int vfs_path_lookup(struct dentry *, struct vfsmount *, const char *, unsigned int, struct path *); extern struct dentry *lookup_one_len(const char *, struct dentry *, int); +extern struct dentry *lookup_one_len2(const char *, struct vfsmount *mnt, struct dentry *, int); extern int follow_down_one(struct path *); extern int follow_down(struct path *); diff --git a/include/trace/events/android_fs.h b/include/trace/events/android_fs.h index 531da433a7bc..49509533d3fa 100644 --- a/include/trace/events/android_fs.h +++ b/include/trace/events/android_fs.h @@ -9,8 +9,8 @@ DEFINE_EVENT(android_fs_data_start_template, android_fs_dataread_start, TP_PROTO(struct inode *inode, loff_t offset, int bytes, - pid_t pid, char *command), - TP_ARGS(inode, offset, bytes, pid, command)); + pid_t pid, char *pathname, char *command), + TP_ARGS(inode, offset, bytes, pid, pathname, command)); DEFINE_EVENT(android_fs_data_end_template, android_fs_dataread_end, TP_PROTO(struct inode *inode, loff_t offset, int bytes), @@ -18,14 +18,48 @@ DEFINE_EVENT(android_fs_data_end_template, android_fs_dataread_end, DEFINE_EVENT(android_fs_data_start_template, android_fs_datawrite_start, TP_PROTO(struct inode *inode, loff_t offset, int bytes, - pid_t pid, char *command), - TP_ARGS(inode, offset, bytes, pid, command)); + pid_t pid, char *pathname, char *command), + TP_ARGS(inode, offset, bytes, pid, pathname, command)); DEFINE_EVENT(android_fs_data_end_template, android_fs_datawrite_end, TP_PROTO(struct inode *inode, loff_t offset, int bytes), - TP_ARGS(inode, offset, bytes)); + TP_ARGS(inode, offset, bytes)); #endif /* _TRACE_ANDROID_FS_H */ /* This part must be outside protection */ #include + +#ifndef ANDROID_FSTRACE_GET_PATHNAME +#define ANDROID_FSTRACE_GET_PATHNAME + +/* Sizes an on-stack array, so careful if sizing this up ! */ +#define MAX_TRACE_PATHBUF_LEN 256 + +static inline char * +android_fstrace_get_pathname(char *buf, int buflen, struct inode *inode) +{ + char *path; + struct dentry *d; + + /* + * d_obtain_alias() will either iput() if it locates an existing + * dentry or transfer the reference to the new dentry created. + * So get an extra reference here. + */ + ihold(inode); + d = d_obtain_alias(inode); + if (likely(!IS_ERR(d))) { + path = dentry_path_raw(d, buf, buflen); + if (unlikely(IS_ERR(path))) { + strcpy(buf, "ERROR"); + path = buf; + } + dput(d); + } else { + strcpy(buf, "ERROR"); + path = buf; + } + return path; +} +#endif diff --git a/include/trace/events/android_fs_template.h b/include/trace/events/android_fs_template.h index 618988b047c1..4e61ffe7a814 100644 --- a/include/trace/events/android_fs_template.h +++ b/include/trace/events/android_fs_template.h @@ -5,11 +5,10 @@ DECLARE_EVENT_CLASS(android_fs_data_start_template, TP_PROTO(struct inode *inode, loff_t offset, int bytes, - pid_t pid, char *command), - TP_ARGS(inode, offset, bytes, pid, command), + pid_t pid, char *pathname, char *command), + TP_ARGS(inode, offset, bytes, pid, pathname, command), TP_STRUCT__entry( - __array(char, path, MAX_FILTER_STR_VAL); - __field(char *, pathname); + __string(pathbuf, pathname); __field(loff_t, offset); __field(int, bytes); __field(loff_t, i_size); @@ -19,27 +18,7 @@ DECLARE_EVENT_CLASS(android_fs_data_start_template, ), TP_fast_assign( { - struct dentry *d; - - /* - * Grab a reference to the inode here because - * d_obtain_alias() will either drop the inode - * reference if it locates an existing dentry - * or transfer the reference to the new dentry - * created. In our case, the file is still open, - * so the dentry is guaranteed to exist (connected), - * so d_obtain_alias() drops the reference we - * grabbed here. - */ - ihold(inode); - d = d_obtain_alias(inode); - if (!IS_ERR(d)) { - __entry->pathname = dentry_path(d, - __entry->path, - MAX_FILTER_STR_VAL); - dput(d); - } else - __entry->pathname = ERR_PTR(-EINVAL); + __assign_str(pathbuf, pathname); __entry->offset = offset; __entry->bytes = bytes; __entry->i_size = i_size_read(inode); @@ -50,9 +29,8 @@ DECLARE_EVENT_CLASS(android_fs_data_start_template, ), TP_printk("entry_name %s, offset %llu, bytes %d, cmdline %s," " pid %d, i_size %llu, ino %lu", - (IS_ERR(__entry->pathname) ? "ERROR" : __entry->pathname), - __entry->offset, __entry->bytes, __get_str(cmdline), - __entry->pid, __entry->i_size, + __get_str(pathbuf), __entry->offset, __entry->bytes, + __get_str(cmdline), __entry->pid, __entry->i_size, (unsigned long) __entry->ino) ); diff --git a/include/uapi/linux/hw_breakpoint.h b/include/uapi/linux/hw_breakpoint.h index b04000a2296a..2b65efd19a46 100644 --- a/include/uapi/linux/hw_breakpoint.h +++ b/include/uapi/linux/hw_breakpoint.h @@ -4,7 +4,11 @@ enum { HW_BREAKPOINT_LEN_1 = 1, HW_BREAKPOINT_LEN_2 = 2, + HW_BREAKPOINT_LEN_3 = 3, HW_BREAKPOINT_LEN_4 = 4, + HW_BREAKPOINT_LEN_5 = 5, + HW_BREAKPOINT_LEN_6 = 6, + HW_BREAKPOINT_LEN_7 = 7, HW_BREAKPOINT_LEN_8 = 8, }; diff --git a/include/uapi/linux/magic.h b/include/uapi/linux/magic.h index cfb5c406f344..e97d5b3ccfa8 100644 --- a/include/uapi/linux/magic.h +++ b/include/uapi/linux/magic.h @@ -52,7 +52,7 @@ #define REISER2FS_SUPER_MAGIC_STRING "ReIsEr2Fs" #define REISER2FS_JR_SUPER_MAGIC_STRING "ReIsEr3Fs" -#define SDCARDFS_SUPER_MAGIC 0xb550ca10 +#define SDCARDFS_SUPER_MAGIC 0x5dca2df5 #define SMB_SUPER_MAGIC 0x517B #define CGROUP_SUPER_MAGIC 0x27e0eb diff --git a/ipc/mqueue.c b/ipc/mqueue.c index 161a1807e6ef..25b7a678f9ef 100644 --- a/ipc/mqueue.c +++ b/ipc/mqueue.c @@ -747,7 +747,7 @@ static struct file *do_create(struct ipc_namespace *ipc_ns, struct inode *dir, } mode &= ~current_umask(); - ret = vfs_create(dir, path->dentry, mode, true); + ret = vfs_create2(path->mnt, dir, path->dentry, mode, true); path->dentry->d_fsdata = NULL; if (ret) return ERR_PTR(ret); @@ -763,7 +763,7 @@ static struct file *do_open(struct path *path, int oflag) if ((oflag & O_ACCMODE) == (O_RDWR | O_WRONLY)) return ERR_PTR(-EINVAL); acc = oflag2acc[oflag & O_ACCMODE]; - if (inode_permission(d_inode(path->dentry), acc)) + if (inode_permission2(path->mnt, d_inode(path->dentry), acc)) return ERR_PTR(-EACCES); return dentry_open(path, oflag, current_cred()); } @@ -796,7 +796,7 @@ SYSCALL_DEFINE4(mq_open, const char __user *, u_name, int, oflag, umode_t, mode, ro = mnt_want_write(mnt); /* we'll drop it in any case */ error = 0; mutex_lock(&d_inode(root)->i_mutex); - path.dentry = lookup_one_len(name->name, root, strlen(name->name)); + path.dentry = lookup_one_len2(name->name, mnt, root, strlen(name->name)); if (IS_ERR(path.dentry)) { error = PTR_ERR(path.dentry); goto out_putfd; @@ -867,7 +867,7 @@ SYSCALL_DEFINE1(mq_unlink, const char __user *, u_name) if (err) goto out_name; mutex_lock_nested(&d_inode(mnt->mnt_root)->i_mutex, I_MUTEX_PARENT); - dentry = lookup_one_len(name->name, mnt->mnt_root, + dentry = lookup_one_len2(name->name, mnt, mnt->mnt_root, strlen(name->name)); if (IS_ERR(dentry)) { err = PTR_ERR(dentry); @@ -879,7 +879,7 @@ SYSCALL_DEFINE1(mq_unlink, const char __user *, u_name) err = -ENOENT; } else { ihold(inode); - err = vfs_unlink(d_inode(dentry->d_parent), dentry, NULL); + err = vfs_unlink2(mnt, d_inode(dentry->d_parent), dentry, NULL); } dput(dentry); diff --git a/security/inode.c b/security/inode.c index 16622aef9bde..0f1a041bf6cb 100644 --- a/security/inode.c +++ b/security/inode.c @@ -100,7 +100,7 @@ struct dentry *securityfs_create_file(const char *name, umode_t mode, dir = d_inode(parent); mutex_lock(&dir->i_mutex); - dentry = lookup_one_len(name, parent, strlen(name)); + dentry = lookup_one_len2(name, mount, parent, strlen(name)); if (IS_ERR(dentry)) goto out; diff --git a/security/security.c b/security/security.c index 46f405ce6b0f..ae05ab153c5a 100644 --- a/security/security.c +++ b/security/security.c @@ -498,6 +498,7 @@ int security_path_chown(struct path *path, kuid_t uid, kgid_t gid) return 0; return call_int_hook(path_chown, 0, path, uid, gid); } +EXPORT_SYMBOL(security_path_chown); int security_path_chroot(struct path *path) {