Received: by 2002:a25:8b91:0:0:0:0:0 with SMTP id j17csp2141098ybl; Thu, 30 Jan 2020 12:04:54 -0800 (PST) X-Google-Smtp-Source: APXvYqy0/FO+JhMdR2ZxCFPD9kLFz6QH5JawgTd7+itXGHipBaR2XaIfnpp4lQal2sfT5yDgCwCM X-Received: by 2002:a9d:5888:: with SMTP id x8mr4607547otg.361.1580414694259; Thu, 30 Jan 2020 12:04:54 -0800 (PST) ARC-Seal: i=1; a=rsa-sha256; t=1580414694; cv=none; d=google.com; s=arc-20160816; b=IfIQLsKbbhGmmpo1GpyDLJVTTttc+ES53MdDILvSloRg1LUqbetTTuPX9Y+KV9pafL u95LTc4vcJ5LIaeAG76v7rnrwypUVrC8fNipUvQW/5oEAYUW5yb6FqdM+jM8K3Z1vXPU 2GdfLTRh9T98KFD5QUz8SpawKgWV83IDVyDKzDh34TNJj6+souQkBNxnbdggguI75R+4 u4tiGGpptqoZuG2v1MWsTjxAN9zYDv/N0QqMRCiQxpzrescfAPZKjynegeBq27+Zgvci wiUmiosirK91gJIwiS1DA5ypdWWO8E2898dlun0q7xVCT2zl8QaMZ6Fmtgsd+2y0br1D jReQ== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20160816; h=list-id:precedence:sender:content-language :content-transfer-encoding:in-reply-to:mime-version:user-agent:date :message-id:from:references:cc:to:subject; bh=mfxYyjeFFT7VkpqS5LPGJieChpiJlRho/ZEdc7kzmvQ=; b=w1R7qN42OufyMX0QG1k3K1qPcNNtxUk3imH5uMdgFjYKKDSdgr8LPHk7hyy66cdfoY tYyQMdTk/dy6Z2YWF6k1cxzihXzXZhthq0Q3EQz4i/oloc2rjwpUpHYWp5dN6229yXQN CKZiAAgCK4GOPEyU0vP8+eU4a6ao+nkYbPSi6j4dcyFdPmFuINGa0A2Xz6BjYi/SzYl/ 5zDI+WdOUhcxWBfaeLPsSJDV9Wx/GERZRycBYzJ/9zf5qFQe4qOMo3G1IgR33rjr1NNg 2VL7UsKwYqHgWAAIhZzSk/OhCtwykFf+eoRh50Yyr2Mb8tbNkAr/GQjClXnCajaaG9+H UnzQ== ARC-Authentication-Results: i=1; mx.google.com; spf=pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) smtp.mailfrom=linux-kernel-owner@vger.kernel.org Return-Path: Received: from vger.kernel.org (vger.kernel.org. [209.132.180.67]) by mx.google.com with ESMTP id r5si3156069oic.19.2020.01.30.12.04.35; Thu, 30 Jan 2020 12:04:54 -0800 (PST) Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67; Authentication-Results: mx.google.com; spf=pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) smtp.mailfrom=linux-kernel-owner@vger.kernel.org Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1727235AbgA3UCK (ORCPT + 99 others); Thu, 30 Jan 2020 15:02:10 -0500 Received: from relay4-d.mail.gandi.net ([217.70.183.196]:50163 "EHLO relay4-d.mail.gandi.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726267AbgA3UCK (ORCPT ); Thu, 30 Jan 2020 15:02:10 -0500 X-Originating-IP: 79.86.19.127 Received: from [192.168.0.12] (127.19.86.79.rev.sfr.net [79.86.19.127]) (Authenticated sender: alex@ghiti.fr) by relay4-d.mail.gandi.net (Postfix) with ESMTPSA id A9715E0009; Thu, 30 Jan 2020 20:02:04 +0000 (UTC) Subject: Re: [PATCH] riscv: Introduce CONFIG_RELOCATABLE To: Zong Li Cc: Anup Patel , linux-riscv , Palmer Dabbelt , "linux-kernel@vger.kernel.org List" , Paul Walmsley References: <20200123201414.8933-1-alex@ghiti.fr> From: Alex Ghiti Message-ID: <39ffe627-d358-e1bd-dff2-653baca1fe37@ghiti.fr> Date: Thu, 30 Jan 2020 15:02:04 -0500 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101 Thunderbird/68.3.0 MIME-Version: 1.0 In-Reply-To: Content-Type: text/plain; charset=utf-8; format=flowed Content-Transfer-Encoding: 7bit Content-Language: en-US Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hi Zong, On 1/29/20 11:13 PM, Zong Li wrote: > On Fri, Jan 24, 2020 at 4:14 AM Alexandre Ghiti wrote: >> This config allows to compile the kernel as PIE and to relocate it at any >> virtual address at runtime: this paves the way to KASLR and to 4-level >> page table folding at runtime. Runtime relocation is possible since >> relocation metadata are embedded into the kernel. >> >> Note that relocating at runtime introduces an overhead even if the kernel >> is loaded at the same address it was linked at and that the compiler >> options are those used in arm64 which uses the same RELA relocation format. >> >> Signed-off-by: Alexandre Ghiti >> --- >> arch/riscv/Kconfig | 11 ++++ >> arch/riscv/Makefile | 5 +- >> arch/riscv/boot/loader.lds.S | 2 +- >> arch/riscv/include/asm/page.h | 5 +- >> arch/riscv/kernel/head.S | 3 +- >> arch/riscv/kernel/vmlinux.lds.S | 10 ++-- >> arch/riscv/mm/Makefile | 4 ++ >> arch/riscv/mm/init.c | 92 ++++++++++++++++++++++++++++----- >> 8 files changed, 110 insertions(+), 22 deletions(-) >> >> diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig >> index fa7dc03459e7..c652b4b850ce 100644 >> --- a/arch/riscv/Kconfig >> +++ b/arch/riscv/Kconfig >> @@ -163,6 +163,17 @@ config PGTABLE_LEVELS >> default 3 if 64BIT >> default 2 >> >> +config RELOCATABLE >> + bool >> + help >> + This builds a kernel as a Position Independent Executable (PIE), >> + which retains all relocation metadata required to relocate the >> + kernel binary at runtime to a different virtual address than the >> + address it was linked at. >> + Since RISCV uses the RELA relocation format, this requires a >> + relocation pass at runtime even if the kernel is loaded at the >> + same address it was linked at. >> + >> source "arch/riscv/Kconfig.socs" >> >> menu "Platform type" >> diff --git a/arch/riscv/Makefile b/arch/riscv/Makefile >> index b9009a2fbaf5..5a115cf6a9c1 100644 >> --- a/arch/riscv/Makefile >> +++ b/arch/riscv/Makefile >> @@ -9,7 +9,10 @@ >> # >> >> OBJCOPYFLAGS := -O binary >> -LDFLAGS_vmlinux := >> +ifeq ($(CONFIG_RELOCATABLE),y) >> +LDFLAGS_vmlinux := -shared -Bsymbolic -z notext -z norelro >> +KBUILD_CFLAGS += -fPIE >> +endif >> ifeq ($(CONFIG_DYNAMIC_FTRACE),y) >> LDFLAGS_vmlinux := --no-relax >> endif >> diff --git a/arch/riscv/boot/loader.lds.S b/arch/riscv/boot/loader.lds.S >> index 47a5003c2e28..a9ed218171aa 100644 >> --- a/arch/riscv/boot/loader.lds.S >> +++ b/arch/riscv/boot/loader.lds.S >> @@ -7,7 +7,7 @@ ENTRY(_start) >> >> SECTIONS >> { >> - . = PAGE_OFFSET; >> + . = CONFIG_PAGE_OFFSET; >> >> .payload : { >> *(.payload) >> diff --git a/arch/riscv/include/asm/page.h b/arch/riscv/include/asm/page.h >> index ac699246ae7e..27c95da68ecb 100644 >> --- a/arch/riscv/include/asm/page.h >> +++ b/arch/riscv/include/asm/page.h >> @@ -31,9 +31,9 @@ >> * When not using MMU this corresponds to the first free page in >> * physical memory (aligned on a page boundary). >> */ >> -#define PAGE_OFFSET _AC(CONFIG_PAGE_OFFSET, UL) >> +#define PAGE_OFFSET kernel_load_addr >> >> -#define KERN_VIRT_SIZE (-PAGE_OFFSET) >> +#define KERN_VIRT_SIZE (-_AC(CONFIG_PAGE_OFFSET, UL)) >> >> #ifndef __ASSEMBLY__ >> >> @@ -97,6 +97,7 @@ extern unsigned long pfn_base; >> #define ARCH_PFN_OFFSET (PAGE_OFFSET >> PAGE_SHIFT) >> #endif /* CONFIG_MMU */ >> >> +extern unsigned long kernel_load_addr; >> extern unsigned long max_low_pfn; >> extern unsigned long min_low_pfn; >> >> diff --git a/arch/riscv/kernel/head.S b/arch/riscv/kernel/head.S >> index 2227db63f895..5042b2b48a06 100644 >> --- a/arch/riscv/kernel/head.S >> +++ b/arch/riscv/kernel/head.S >> @@ -126,7 +126,8 @@ clear_bss_done: >> #ifdef CONFIG_MMU >> relocate: >> /* Relocate return address */ >> - li a1, PAGE_OFFSET >> + la a1, kernel_load_addr >> + REG_L a1, 0(a1) >> la a2, _start >> sub a1, a1, a2 >> add ra, ra, a1 >> diff --git a/arch/riscv/kernel/vmlinux.lds.S b/arch/riscv/kernel/vmlinux.lds.S >> index 12f42f96d46e..5095aee7c37e 100644 >> --- a/arch/riscv/kernel/vmlinux.lds.S >> +++ b/arch/riscv/kernel/vmlinux.lds.S >> @@ -4,7 +4,7 @@ >> * Copyright (C) 2017 SiFive >> */ >> >> -#define LOAD_OFFSET PAGE_OFFSET >> +#define LOAD_OFFSET CONFIG_PAGE_OFFSET >> #include >> #include >> #include >> @@ -70,9 +70,11 @@ SECTIONS >> >> EXCEPTION_TABLE(0x10) >> >> - .rel.dyn : { >> - *(.rel.dyn*) >> - } >> + .rela.dyn : ALIGN(8) { >> + __rela_dyn_start = .; >> + *(.rela .rela*) >> + __rela_dyn_end = .; >> + } >> >> _end = .; >> >> diff --git a/arch/riscv/mm/Makefile b/arch/riscv/mm/Makefile >> index a1bd95c8047a..dcd3d806243f 100644 >> --- a/arch/riscv/mm/Makefile >> +++ b/arch/riscv/mm/Makefile >> @@ -1,6 +1,10 @@ >> # SPDX-License-Identifier: GPL-2.0-only >> >> CFLAGS_init.o := -mcmodel=medany >> +ifdef CONFIG_RELOCATABLE >> +CFLAGS_init.o += -fno-pie >> +endif >> + >> ifdef CONFIG_FTRACE >> CFLAGS_REMOVE_init.o = -pg >> endif >> diff --git a/arch/riscv/mm/init.c b/arch/riscv/mm/init.c >> index 965a8cf4829c..ac9a9f69abc0 100644 >> --- a/arch/riscv/mm/init.c >> +++ b/arch/riscv/mm/init.c >> @@ -12,6 +12,9 @@ >> #include >> #include >> #include >> +#ifdef CONFIG_RELOCATABLE >> +#include >> +#endif >> >> #include >> #include >> @@ -28,6 +31,9 @@ EXPORT_SYMBOL(empty_zero_page); >> extern char _start[]; >> void *dtb_early_va; >> >> +unsigned long kernel_load_addr = _AC(CONFIG_PAGE_OFFSET, UL); >> +EXPORT_SYMBOL(kernel_load_addr); >> + >> static void __init zone_sizes_init(void) >> { >> unsigned long max_zone_pfns[MAX_NR_ZONES] = { 0, }; >> @@ -132,7 +138,8 @@ void __init setup_bootmem(void) >> phys_addr_t end = reg->base + reg->size; >> >> if (reg->base <= vmlinux_end && vmlinux_end <= end) { >> - mem_size = min(reg->size, (phys_addr_t)-PAGE_OFFSET); >> + mem_size = min(reg->size, >> + (phys_addr_t)-kernel_load_addr); >> >> /* >> * Remove memblock from the end of usable area to the >> @@ -269,7 +276,7 @@ static phys_addr_t __init alloc_pmd(uintptr_t va) >> if (mmu_enabled) >> return memblock_phys_alloc(PAGE_SIZE, PAGE_SIZE); >> >> - pmd_num = (va - PAGE_OFFSET) >> PGDIR_SHIFT; >> + pmd_num = (va - kernel_load_addr) >> PGDIR_SHIFT; >> BUG_ON(pmd_num >= NUM_EARLY_PMDS); >> return (uintptr_t)&early_pmd[pmd_num * PTRS_PER_PMD]; >> } >> @@ -370,6 +377,54 @@ static uintptr_t __init best_map_size(phys_addr_t base, phys_addr_t size) >> #error "setup_vm() is called from head.S before relocate so it should not use absolute addressing." >> #endif >> >> +#ifdef CONFIG_RELOCATABLE >> +extern unsigned long __rela_dyn_start, __rela_dyn_end; >> + >> +#ifdef CONFIG_64BIT >> +#define Elf_Rela Elf64_Rela >> +#define Elf_Addr Elf64_Addr >> +#else >> +#define Elf_Rela Elf32_Rela >> +#define Elf_Addr Elf32_Addr >> +#endif >> + >> +void __init relocate_kernel(uintptr_t load_pa) >> +{ >> + Elf_Rela *rela = (Elf_Rela *)&__rela_dyn_start; >> + uintptr_t link_addr = _AC(CONFIG_PAGE_OFFSET, UL); >> + /* >> + * This holds the offset between the linked virtual address and the >> + * relocated virtual address. >> + */ >> + uintptr_t reloc_offset = kernel_load_addr - link_addr; >> + /* >> + * This holds the offset between linked virtual address and physical >> + * address whereas va_pa_offset holds the offset between relocated >> + * virtual address and physical address. >> + */ >> + uintptr_t va_link_pa_offset = link_addr - load_pa; >> + >> + for ( ; rela < (Elf_Rela *)&__rela_dyn_end; rela++) { >> + Elf_Addr addr = (rela->r_offset - va_link_pa_offset); >> + Elf_Addr relocated_addr = rela->r_addend; >> + >> + if (rela->r_info != R_RISCV_RELATIVE) >> + continue; >> + >> + /* >> + * Make sure to not relocate vdso symbols like rt_sigreturn >> + * which are linked from the address 0 in vmlinux since >> + * vdso symbol addresses are actually used as an offset from >> + * mm->context.vdso in VDSO_OFFSET macro. >> + */ >> + if (relocated_addr >= link_addr) >> + relocated_addr += reloc_offset; >> + >> + *(Elf_Addr *)addr = relocated_addr; >> + } >> +} >> +#endif >> + >> asmlinkage void __init setup_vm(uintptr_t dtb_pa) >> { >> uintptr_t va, end_va; >> @@ -377,9 +432,20 @@ asmlinkage void __init setup_vm(uintptr_t dtb_pa) >> uintptr_t load_sz = (uintptr_t)(&_end) - load_pa; >> uintptr_t map_size = best_map_size(load_pa, MAX_EARLY_MAPPING_SIZE); >> >> - va_pa_offset = PAGE_OFFSET - load_pa; >> + va_pa_offset = kernel_load_addr - load_pa; >> pfn_base = PFN_DOWN(load_pa); >> >> +#ifdef CONFIG_RELOCATABLE >> + /* >> + * Early page table uses only one PGDIR, which makes it possible >> + * to map 1GB aligned on 1GB: if the relocation offset makes the kernel >> + * cross over a 1G boundary, raise a bug since a part of the kernel >> + * would not get mapped. >> + */ >> + BUG_ON(SZ_1G - (kernel_load_addr & (SZ_1G - 1)) < load_sz); >> + relocate_kernel(load_pa); >> +#endif >> + >> /* >> * Enforce boot alignment requirements of RV32 and >> * RV64 by only allowing PMD or PGD mappings. >> @@ -387,7 +453,7 @@ asmlinkage void __init setup_vm(uintptr_t dtb_pa) >> BUG_ON(map_size == PAGE_SIZE); >> >> /* Sanity check alignment and size */ >> - BUG_ON((PAGE_OFFSET % PGDIR_SIZE) != 0); >> + BUILD_BUG_ON((_AC(CONFIG_PAGE_OFFSET, UL) % PGDIR_SIZE) != 0); >> BUG_ON((load_pa % map_size) != 0); >> BUG_ON(load_sz > MAX_EARLY_MAPPING_SIZE); >> >> @@ -400,13 +466,13 @@ asmlinkage void __init setup_vm(uintptr_t dtb_pa) >> create_pmd_mapping(fixmap_pmd, FIXADDR_START, >> (uintptr_t)fixmap_pte, PMD_SIZE, PAGE_TABLE); >> /* Setup trampoline PGD and PMD */ >> - create_pgd_mapping(trampoline_pg_dir, PAGE_OFFSET, >> + create_pgd_mapping(trampoline_pg_dir, kernel_load_addr, >> (uintptr_t)trampoline_pmd, PGDIR_SIZE, PAGE_TABLE); >> - create_pmd_mapping(trampoline_pmd, PAGE_OFFSET, >> + create_pmd_mapping(trampoline_pmd, kernel_load_addr, >> load_pa, PMD_SIZE, PAGE_KERNEL_EXEC); >> #else >> /* Setup trampoline PGD */ >> - create_pgd_mapping(trampoline_pg_dir, PAGE_OFFSET, >> + create_pgd_mapping(trampoline_pg_dir, kernel_load_addr, >> load_pa, PGDIR_SIZE, PAGE_KERNEL_EXEC); >> #endif >> >> @@ -415,10 +481,10 @@ asmlinkage void __init setup_vm(uintptr_t dtb_pa) >> * us to reach paging_init(). We map all memory banks later >> * in setup_vm_final() below. >> */ >> - end_va = PAGE_OFFSET + load_sz; >> - for (va = PAGE_OFFSET; va < end_va; va += map_size) >> + end_va = kernel_load_addr + load_sz; >> + for (va = kernel_load_addr; va < end_va; va += map_size) >> create_pgd_mapping(early_pg_dir, va, >> - load_pa + (va - PAGE_OFFSET), >> + load_pa + (va - kernel_load_addr), >> map_size, PAGE_KERNEL_EXEC); >> >> /* Create fixed mapping for early FDT parsing */ >> @@ -457,9 +523,9 @@ static void __init setup_vm_final(void) >> break; >> if (memblock_is_nomap(reg)) >> continue; >> - if (start <= __pa(PAGE_OFFSET) && >> - __pa(PAGE_OFFSET) < end) >> - start = __pa(PAGE_OFFSET); >> + if (start <= __pa(kernel_load_addr) && >> + __pa(kernel_load_addr) < end) >> + start = __pa(kernel_load_addr); > Here should use __pa_symbol() instead of __pa() for kernel symbols. I'll add it to the v2 too, thanks. > I'm working on KASLR on top of this patch, it's work to me. If I can help, don't hesitate, even if it is simply for testing :) > > Reviewed-by: Zong Li > Tested-by: Zong Li Thanks for that, Alex > >> map_size = best_map_size(start, end - start); >> for (pa = start; pa < end; pa += map_size) { >> -- >> 2.20.1 >>