From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mail-pf1-f202.google.com (mail-pf1-f202.google.com [209.85.210.202]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 32ED22AF04 for ; Sat, 12 Apr 2025 01:10:25 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.210.202 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1744420228; cv=none; b=ZAnJHVpenSJz/ac7Utb+UvLDpbfNel47XdRAWJmtWdVvQHD1dRWlcZtxdJ9ar16o2yZWq9AZsNm+ugpCdFRKkQjELI/X+z9kH51Q1G25BsMC+pi8BRwaPHsU2Wy8UMbFADdP8ieMbu+1zimlH83NzVrOwa3M+ynd6nJjs0yF1/g= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1744420228; c=relaxed/simple; bh=CLWex6gdQgCeQ+j1Ylxw5hYzn0j5IyoX3BXUq3gS9XE=; h=Date:In-Reply-To:Mime-Version:References:Message-ID:Subject:From: To:Cc:Content-Type; b=U2FG8pL+cWxi9zv95sJF3IZOxVpGf32r9Oq04eX70x1eaBv1kHcTKJ+R12qylihrENX1559iMgRbIrgI1FHs0aWks/DO8LcucSpGdMbmT4EuTqYtuTKZ1wGzAIxjzO9b4Mj4p1DdJF6EJEEhdlwJ8FMmiY0Xm2Q0gHtpLCVs4p0= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=flex--dylanbhatch.bounces.google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=4LQsC2wO; arc=none smtp.client-ip=209.85.210.202 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=flex--dylanbhatch.bounces.google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="4LQsC2wO" Received: by mail-pf1-f202.google.com with SMTP id d2e1a72fcca58-7395d07a3dcso1836109b3a.3 for ; Fri, 11 Apr 2025 18:10:25 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1744420225; x=1745025025; darn=vger.kernel.org; h=cc:to:from:subject:message-id:references:mime-version:in-reply-to :date:from:to:cc:subject:date:message-id:reply-to; bh=UPE/w7OMV+R/BpERehO0lW1KzawWpcv3Vb+OACxnfog=; b=4LQsC2wO14y7Bf3/7k10jDCvWN6yfo+i7WjI4SzM2LB80SHKwPUhfsPpg5UVXdJbur bXK84iOZp1bJE4jXeyQNGFrteSYnprNqWPpe/HZ+0JUtrP92Q3sCYb6QpZ0CGCxnFPPw McAIGpQA5NIS/6o0xb8Rm1fuBlf4bajIvBVluwejQnCurQQf03BKmpXxQGS+Vy2xR52E BP5W2dZuX+nYYPkiMWFByENc2pRVTkzyscx8yWgd7uapnk0MvqkPypRfDlo8tgojRQrN QPT51ouJjPFbbLvrrTmp2SJMVxr4NifJ4To9TfVixF5ervCuf7a7IJHDOWY2zZad5mWd rs5Q== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1744420225; x=1745025025; h=cc:to:from:subject:message-id:references:mime-version:in-reply-to :date:x-gm-message-state:from:to:cc:subject:date:message-id:reply-to; bh=UPE/w7OMV+R/BpERehO0lW1KzawWpcv3Vb+OACxnfog=; b=SUKbaucg2z4XV67qzjeMzF9VJctP9vgkjA8OZ//flbrgln8b0+wesLYkap8WcoPsyy PN1e1JYNfjB3xl0JTrirKws1Eenuq7yLo7lhFuzWpHhlbwjgXwaJ7VNyoqtUQvKjAxhN RiQcdf8ySuBTnBmkVgpoCDGOaau6g2CX5heohSOtOL4/FzRRzxc0r81+Jvp572winF07 Z/mxiM+BNZRA0I9VoHIqtiM9rneFu7aj7qrlYXJZMEl106tTeZ4xrjfLgJhf03i2K9el B13n0VSEGpKCG9N7NxmFhWAy+1eU4Df2pY5MknsRJ3QidfkzCsjMedAbNCKOJAR9X9zw qrzg== X-Forwarded-Encrypted: i=1; AJvYcCX/7E+3azL8yOr2rODpp0vOKfERhxHOMGsK5GYtNESXpaCuVcu4Vj9/C348l9A+yv5TiYWyPXn55C1TyH0=@vger.kernel.org X-Gm-Message-State: AOJu0YwWDOgtvvx81CusTCEDEcwUX6VGNyIsISMs6ZucRs0vfz9C35By vMpnfVdvN7+RLtsSRR+yCdYbxh3mVMaO/+5Yi9mIN0miusrMBwn0ufyZXkr5am9ArAvV9mTGSVI MqRcC47elybYjtsTgNg+e0A== X-Google-Smtp-Source: AGHT+IGCCoJ+p4ug1h6JLtcNhK91NiKYTxwJFv7sDgsGsshWBvKN29loXFFysazsL5Y0Mr82eB8IfOe4BG34Ie+snA== X-Received: from pfgt19.prod.google.com ([2002:a05:6a00:1393:b0:739:56be:f58c]) (user=dylanbhatch job=prod-delivery.src-stubby-dispatcher) by 2002:a05:6a20:9f8f:b0:1f5:6c94:2cc1 with SMTP id adf61e73a8af0-201797b7eeemr6841826637.21.1744420225268; Fri, 11 Apr 2025 18:10:25 -0700 (PDT) Date: Sat, 12 Apr 2025 01:09:39 +0000 In-Reply-To: <20250412010940.1686376-1-dylanbhatch@google.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 References: <20250412010940.1686376-1-dylanbhatch@google.com> X-Mailer: git-send-email 2.49.0.604.gff1f9ca942-goog Message-ID: <20250412010940.1686376-3-dylanbhatch@google.com> Subject: [PATCH v2 2/2] arm64/module: Use text-poke API for late relocations. From: Dylan Hatch To: Daniel Borkmann , Andrii Nakryiko , Martin KaFai Lau , Eduard Zingerman , Yonghong Song , John Fastabend , KP Singh , Stanislav Fomichev , Hao Luo , Jiri Olsa , Puranjay Mohan , Xu Kuohai , "=?UTF-8?q?Philippe=20Mathieu-Daud=C3=A9?=" , Catalin Marinas , Will Deacon , "Mike Rapoport (Microsoft)" , Arnd Bergmann , Geert Uytterhoeven , Luis Chamberlain , Andrew Morton , Song Liu , Ard Biesheuvel , Mark Rutland Cc: linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, Dylan Hatch , Roman Gushchin Content-Type: text/plain; charset="UTF-8" To enable late module patching, livepatch modules need to be able to apply some of their relocations well after being loaded. In this scenario, use the text-poking API to allow this, even with STRICT_MODULE_RWX. This patch is largely based off commit 88fc078a7a8f6 ("x86/module: Use text_poke() for late relocations"). Signed-off-by: Dylan Hatch --- arch/arm64/kernel/module.c | 129 ++++++++++++++++++++++++------------- 1 file changed, 83 insertions(+), 46 deletions(-) diff --git a/arch/arm64/kernel/module.c b/arch/arm64/kernel/module.c index 06bb680bfe975..0703502d9dc37 100644 --- a/arch/arm64/kernel/module.c +++ b/arch/arm64/kernel/module.c @@ -18,11 +18,13 @@ #include #include #include +#include #include #include #include #include +#include enum aarch64_reloc_op { RELOC_OP_NONE, @@ -48,7 +50,8 @@ static u64 do_reloc(enum aarch64_reloc_op reloc_op, __le32 *place, u64 val) return 0; } -static int reloc_data(enum aarch64_reloc_op op, void *place, u64 val, int len) +static int reloc_data(enum aarch64_reloc_op op, void *place, u64 val, int len, + void *(*write)(void *dest, const void *src, size_t len)) { s64 sval = do_reloc(op, place, val); @@ -66,7 +69,7 @@ static int reloc_data(enum aarch64_reloc_op op, void *place, u64 val, int len) switch (len) { case 16: - *(s16 *)place = sval; + write(place, &sval, sizeof(s16)); switch (op) { case RELOC_OP_ABS: if (sval < 0 || sval > U16_MAX) @@ -82,7 +85,7 @@ static int reloc_data(enum aarch64_reloc_op op, void *place, u64 val, int len) } break; case 32: - *(s32 *)place = sval; + write(place, &sval, sizeof(s32)); switch (op) { case RELOC_OP_ABS: if (sval < 0 || sval > U32_MAX) @@ -98,7 +101,7 @@ static int reloc_data(enum aarch64_reloc_op op, void *place, u64 val, int len) } break; case 64: - *(s64 *)place = sval; + write(place, &sval, sizeof(s64)); break; default: pr_err("Invalid length (%d) for data relocation\n", len); @@ -113,11 +116,13 @@ enum aarch64_insn_movw_imm_type { }; static int reloc_insn_movw(enum aarch64_reloc_op op, __le32 *place, u64 val, - int lsb, enum aarch64_insn_movw_imm_type imm_type) + int lsb, enum aarch64_insn_movw_imm_type imm_type, + void *(*write)(void *dest, const void *src, size_t len)) { u64 imm; s64 sval; u32 insn = le32_to_cpu(*place); + __le32 le_insn; sval = do_reloc(op, place, val); imm = sval >> lsb; @@ -145,7 +150,8 @@ static int reloc_insn_movw(enum aarch64_reloc_op op, __le32 *place, u64 val, /* Update the instruction with the new encoding. */ insn = aarch64_insn_encode_immediate(AARCH64_INSN_IMM_16, insn, imm); - *place = cpu_to_le32(insn); + le_insn = cpu_to_le32(insn); + write(place, &le_insn, sizeof(le_insn)); if (imm > U16_MAX) return -ERANGE; @@ -154,11 +160,13 @@ static int reloc_insn_movw(enum aarch64_reloc_op op, __le32 *place, u64 val, } static int reloc_insn_imm(enum aarch64_reloc_op op, __le32 *place, u64 val, - int lsb, int len, enum aarch64_insn_imm_type imm_type) + int lsb, int len, enum aarch64_insn_imm_type imm_type, + void *(*write)(void *dest, const void *src, size_t len)) { u64 imm, imm_mask; s64 sval; u32 insn = le32_to_cpu(*place); + __le32 le_insn; /* Calculate the relocation value. */ sval = do_reloc(op, place, val); @@ -170,7 +178,8 @@ static int reloc_insn_imm(enum aarch64_reloc_op op, __le32 *place, u64 val, /* Update the instruction's immediate field. */ insn = aarch64_insn_encode_immediate(imm_type, insn, imm); - *place = cpu_to_le32(insn); + le_insn = cpu_to_le32(insn); + write(place, &le_insn, sizeof(le_insn)); /* * Extract the upper value bits (including the sign bit) and @@ -189,17 +198,19 @@ static int reloc_insn_imm(enum aarch64_reloc_op op, __le32 *place, u64 val, } static int reloc_insn_adrp(struct module *mod, Elf64_Shdr *sechdrs, - __le32 *place, u64 val) + __le32 *place, u64 val, + void *(*write)(void *dest, const void *src, size_t len)) { u32 insn; + __le32 le_insn; if (!is_forbidden_offset_for_adrp(place)) return reloc_insn_imm(RELOC_OP_PAGE, place, val, 12, 21, - AARCH64_INSN_IMM_ADR); + AARCH64_INSN_IMM_ADR, write); /* patch ADRP to ADR if it is in range */ if (!reloc_insn_imm(RELOC_OP_PREL, place, val & ~0xfff, 0, 21, - AARCH64_INSN_IMM_ADR)) { + AARCH64_INSN_IMM_ADR, write)) { insn = le32_to_cpu(*place); insn &= ~BIT(31); } else { @@ -211,15 +222,17 @@ static int reloc_insn_adrp(struct module *mod, Elf64_Shdr *sechdrs, AARCH64_INSN_BRANCH_NOLINK); } - *place = cpu_to_le32(insn); + le_insn = cpu_to_le32(insn); + write(place, &le_insn, sizeof(le_insn)); return 0; } -int apply_relocate_add(Elf64_Shdr *sechdrs, +static int __apply_relocate_add(Elf64_Shdr *sechdrs, const char *strtab, unsigned int symindex, unsigned int relsec, - struct module *me) + struct module *me, + void *(*write)(void *dest, const void *src, size_t len)) { unsigned int i; int ovf; @@ -255,23 +268,23 @@ int apply_relocate_add(Elf64_Shdr *sechdrs, /* Data relocations. */ case R_AARCH64_ABS64: overflow_check = false; - ovf = reloc_data(RELOC_OP_ABS, loc, val, 64); + ovf = reloc_data(RELOC_OP_ABS, loc, val, 64, write); break; case R_AARCH64_ABS32: - ovf = reloc_data(RELOC_OP_ABS, loc, val, 32); + ovf = reloc_data(RELOC_OP_ABS, loc, val, 32, write); break; case R_AARCH64_ABS16: - ovf = reloc_data(RELOC_OP_ABS, loc, val, 16); + ovf = reloc_data(RELOC_OP_ABS, loc, val, 16, write); break; case R_AARCH64_PREL64: overflow_check = false; - ovf = reloc_data(RELOC_OP_PREL, loc, val, 64); + ovf = reloc_data(RELOC_OP_PREL, loc, val, 64, write); break; case R_AARCH64_PREL32: - ovf = reloc_data(RELOC_OP_PREL, loc, val, 32); + ovf = reloc_data(RELOC_OP_PREL, loc, val, 32, write); break; case R_AARCH64_PREL16: - ovf = reloc_data(RELOC_OP_PREL, loc, val, 16); + ovf = reloc_data(RELOC_OP_PREL, loc, val, 16, write); break; /* MOVW instruction relocations. */ @@ -280,88 +293,88 @@ int apply_relocate_add(Elf64_Shdr *sechdrs, fallthrough; case R_AARCH64_MOVW_UABS_G0: ovf = reloc_insn_movw(RELOC_OP_ABS, loc, val, 0, - AARCH64_INSN_IMM_MOVKZ); + AARCH64_INSN_IMM_MOVKZ, write); break; case R_AARCH64_MOVW_UABS_G1_NC: overflow_check = false; fallthrough; case R_AARCH64_MOVW_UABS_G1: ovf = reloc_insn_movw(RELOC_OP_ABS, loc, val, 16, - AARCH64_INSN_IMM_MOVKZ); + AARCH64_INSN_IMM_MOVKZ, write); break; case R_AARCH64_MOVW_UABS_G2_NC: overflow_check = false; fallthrough; case R_AARCH64_MOVW_UABS_G2: ovf = reloc_insn_movw(RELOC_OP_ABS, loc, val, 32, - AARCH64_INSN_IMM_MOVKZ); + AARCH64_INSN_IMM_MOVKZ, write); break; case R_AARCH64_MOVW_UABS_G3: /* We're using the top bits so we can't overflow. */ overflow_check = false; ovf = reloc_insn_movw(RELOC_OP_ABS, loc, val, 48, - AARCH64_INSN_IMM_MOVKZ); + AARCH64_INSN_IMM_MOVKZ, write); break; case R_AARCH64_MOVW_SABS_G0: ovf = reloc_insn_movw(RELOC_OP_ABS, loc, val, 0, - AARCH64_INSN_IMM_MOVNZ); + AARCH64_INSN_IMM_MOVNZ, write); break; case R_AARCH64_MOVW_SABS_G1: ovf = reloc_insn_movw(RELOC_OP_ABS, loc, val, 16, - AARCH64_INSN_IMM_MOVNZ); + AARCH64_INSN_IMM_MOVNZ, write); break; case R_AARCH64_MOVW_SABS_G2: ovf = reloc_insn_movw(RELOC_OP_ABS, loc, val, 32, - AARCH64_INSN_IMM_MOVNZ); + AARCH64_INSN_IMM_MOVNZ, write); break; case R_AARCH64_MOVW_PREL_G0_NC: overflow_check = false; ovf = reloc_insn_movw(RELOC_OP_PREL, loc, val, 0, - AARCH64_INSN_IMM_MOVKZ); + AARCH64_INSN_IMM_MOVKZ, write); break; case R_AARCH64_MOVW_PREL_G0: ovf = reloc_insn_movw(RELOC_OP_PREL, loc, val, 0, - AARCH64_INSN_IMM_MOVNZ); + AARCH64_INSN_IMM_MOVNZ, write); break; case R_AARCH64_MOVW_PREL_G1_NC: overflow_check = false; ovf = reloc_insn_movw(RELOC_OP_PREL, loc, val, 16, - AARCH64_INSN_IMM_MOVKZ); + AARCH64_INSN_IMM_MOVKZ, write); break; case R_AARCH64_MOVW_PREL_G1: ovf = reloc_insn_movw(RELOC_OP_PREL, loc, val, 16, - AARCH64_INSN_IMM_MOVNZ); + AARCH64_INSN_IMM_MOVNZ, write); break; case R_AARCH64_MOVW_PREL_G2_NC: overflow_check = false; ovf = reloc_insn_movw(RELOC_OP_PREL, loc, val, 32, - AARCH64_INSN_IMM_MOVKZ); + AARCH64_INSN_IMM_MOVKZ, write); break; case R_AARCH64_MOVW_PREL_G2: ovf = reloc_insn_movw(RELOC_OP_PREL, loc, val, 32, - AARCH64_INSN_IMM_MOVNZ); + AARCH64_INSN_IMM_MOVNZ, write); break; case R_AARCH64_MOVW_PREL_G3: /* We're using the top bits so we can't overflow. */ overflow_check = false; ovf = reloc_insn_movw(RELOC_OP_PREL, loc, val, 48, - AARCH64_INSN_IMM_MOVNZ); + AARCH64_INSN_IMM_MOVNZ, write); break; /* Immediate instruction relocations. */ case R_AARCH64_LD_PREL_LO19: ovf = reloc_insn_imm(RELOC_OP_PREL, loc, val, 2, 19, - AARCH64_INSN_IMM_19); + AARCH64_INSN_IMM_19, write); break; case R_AARCH64_ADR_PREL_LO21: ovf = reloc_insn_imm(RELOC_OP_PREL, loc, val, 0, 21, - AARCH64_INSN_IMM_ADR); + AARCH64_INSN_IMM_ADR, write); break; case R_AARCH64_ADR_PREL_PG_HI21_NC: overflow_check = false; fallthrough; case R_AARCH64_ADR_PREL_PG_HI21: - ovf = reloc_insn_adrp(me, sechdrs, loc, val); + ovf = reloc_insn_adrp(me, sechdrs, loc, val, write); if (ovf && ovf != -ERANGE) return ovf; break; @@ -369,46 +382,46 @@ int apply_relocate_add(Elf64_Shdr *sechdrs, case R_AARCH64_LDST8_ABS_LO12_NC: overflow_check = false; ovf = reloc_insn_imm(RELOC_OP_ABS, loc, val, 0, 12, - AARCH64_INSN_IMM_12); + AARCH64_INSN_IMM_12, write); break; case R_AARCH64_LDST16_ABS_LO12_NC: overflow_check = false; ovf = reloc_insn_imm(RELOC_OP_ABS, loc, val, 1, 11, - AARCH64_INSN_IMM_12); + AARCH64_INSN_IMM_12, write); break; case R_AARCH64_LDST32_ABS_LO12_NC: overflow_check = false; ovf = reloc_insn_imm(RELOC_OP_ABS, loc, val, 2, 10, - AARCH64_INSN_IMM_12); + AARCH64_INSN_IMM_12, write); break; case R_AARCH64_LDST64_ABS_LO12_NC: overflow_check = false; ovf = reloc_insn_imm(RELOC_OP_ABS, loc, val, 3, 9, - AARCH64_INSN_IMM_12); + AARCH64_INSN_IMM_12, write); break; case R_AARCH64_LDST128_ABS_LO12_NC: overflow_check = false; ovf = reloc_insn_imm(RELOC_OP_ABS, loc, val, 4, 8, - AARCH64_INSN_IMM_12); + AARCH64_INSN_IMM_12, write); break; case R_AARCH64_TSTBR14: ovf = reloc_insn_imm(RELOC_OP_PREL, loc, val, 2, 14, - AARCH64_INSN_IMM_14); + AARCH64_INSN_IMM_14, write); break; case R_AARCH64_CONDBR19: ovf = reloc_insn_imm(RELOC_OP_PREL, loc, val, 2, 19, - AARCH64_INSN_IMM_19); + AARCH64_INSN_IMM_19, write); break; case R_AARCH64_JUMP26: case R_AARCH64_CALL26: ovf = reloc_insn_imm(RELOC_OP_PREL, loc, val, 2, 26, - AARCH64_INSN_IMM_26); + AARCH64_INSN_IMM_26, write); if (ovf == -ERANGE) { val = module_emit_plt_entry(me, sechdrs, loc, &rel[i], sym); if (!val) return -ENOEXEC; ovf = reloc_insn_imm(RELOC_OP_PREL, loc, val, 2, - 26, AARCH64_INSN_IMM_26); + 26, AARCH64_INSN_IMM_26, write); } break; @@ -431,6 +444,30 @@ int apply_relocate_add(Elf64_Shdr *sechdrs, return -ENOEXEC; } +int apply_relocate_add(Elf64_Shdr *sechdrs, + const char *strtab, + unsigned int symindex, + unsigned int relsec, + struct module *me) +{ + int ret; + bool early = me->state == MODULE_STATE_UNFORMED; + void *(*write)(void *, const void *, size_t) = memcpy; + + if (!early) { + write = text_poke; + mutex_lock(&text_mutex); + } + + ret = __apply_relocate_add(sechdrs, strtab, symindex, relsec, me, + write); + + if (!early) + mutex_unlock(&text_mutex); + + return ret; +} + static inline void __init_plt(struct plt_entry *plt, unsigned long addr) { *plt = get_plt_entry(addr, plt); -- 2.49.0.604.gff1f9ca942-goog