Merge patch series "riscv: Relocatable NOMMU kernels"

Samuel Holland <samuel.holland@sifive.com> says:

Currently, RISC-V NOMMU kernels are linked at CONFIG_PAGE_OFFSET, and
since they are not relocatable, must be loaded at this address as well.
CONFIG_PAGE_OFFSET is not a user-visible Kconfig option, so its value is
not obvious, and users must patch the kernel source if they want to load
it at a different address.

Make NOMMU kernels more portable by making them relocatable by default.
This allows a single kernel binary to work when loaded at any address.

* b4-shazam-merge:
  riscv: Remove CONFIG_PAGE_OFFSET
  riscv: Support CONFIG_RELOCATABLE on riscv32
  asm-generic: Always define Elf_Rel and Elf_Rela
  riscv: Support CONFIG_RELOCATABLE on NOMMU
  riscv: Allow NOMMU kernels to access all of RAM
  riscv: Remove duplicate CONFIG_PAGE_OFFSET definition

Link: https://lore.kernel.org/r/20241026171441.3047904-1-samuel.holland@sifive.com
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>

Author: palmer-dabbelt

arch/riscv/Kconfig

@@ -202,6 +202,7 @@ config RISCV
 	select PCI_DOMAINS_GENERIC if PCI
 	select PCI_ECAM if (ACPI && PCI)
 	select PCI_MSI if PCI
+	select RELOCATABLE if !MMU && !PHYS_RAM_BASE_FIXED
 	select RISCV_ALTERNATIVE if !XIP_KERNEL
 	select RISCV_APLIC
 	select RISCV_IMSIC
@@ -289,13 +290,6 @@ config MMU
 	  Select if you want MMU-based virtualised addressing space
 	  support by paged memory management. If unsure, say 'Y'.
 
-config PAGE_OFFSET
-	hex
-	default 0x80000000 if !MMU && RISCV_M_MODE
-	default 0x80200000 if !MMU
-	default 0xc0000000 if 32BIT
-	default 0xff60000000000000 if 64BIT
-
 config KASAN_SHADOW_OFFSET
 	hex
 	depends on KASAN_GENERIC
@@ -1100,7 +1094,7 @@ config PARAVIRT_TIME_ACCOUNTING
 
 config RELOCATABLE
 	bool "Build a relocatable kernel"
-	depends on MMU && 64BIT && !XIP_KERNEL
+	depends on !XIP_KERNEL
 	select MODULE_SECTIONS if MODULES
 	help
           This builds a kernel as a Position Independent Executable (PIE),

arch/riscv/Makefile

@@ -98,7 +98,6 @@ KBUILD_AFLAGS += -march=$(riscv-march-y)
 CC_FLAGS_FPU  := -march=$(shell echo $(riscv-march-y) | sed -E 's/(rv32ima|rv64ima)([^v_]*)v?/\1\2/')
 
 KBUILD_CFLAGS += -mno-save-restore
-KBUILD_CFLAGS += -DCONFIG_PAGE_OFFSET=$(CONFIG_PAGE_OFFSET)
 
 ifeq ($(CONFIG_CMODEL_MEDLOW),y)
 	KBUILD_CFLAGS += -mcmodel=medlow

arch/riscv/errata/Makefile

@@ -1,5 +1,9 @@
 ifdef CONFIG_RELOCATABLE
-KBUILD_CFLAGS += -fno-pie
+# We can't use PIC/PIE when handling early-boot errata parsing, as the kernel
+# doesn't have a GOT setup at that point.  So instead just use medany: it's
+# usually position-independent, so it should be good enough for the errata
+# handling.
+KBUILD_CFLAGS += -fno-pie -mcmodel=medany
 endif
 
 ifdef CONFIG_RISCV_ALTERNATIVE_EARLY

arch/riscv/include/asm/page.h

@@ -24,21 +24,22 @@
  * When not using MMU this corresponds to the first free page in
  * physical memory (aligned on a page boundary).
  */
-#ifdef CONFIG_64BIT
 #ifdef CONFIG_MMU
-#define PAGE_OFFSET		kernel_map.page_offset
-#else
-#define PAGE_OFFSET		_AC(CONFIG_PAGE_OFFSET, UL)
-#endif
-/*
- * By default, CONFIG_PAGE_OFFSET value corresponds to SV57 address space so
- * define the PAGE_OFFSET value for SV48 and SV39.
- */
+#ifdef CONFIG_64BIT
+#define PAGE_OFFSET_L5		_AC(0xff60000000000000, UL)
 #define PAGE_OFFSET_L4		_AC(0xffffaf8000000000, UL)
 #define PAGE_OFFSET_L3		_AC(0xffffffd600000000, UL)
+#ifdef CONFIG_XIP_KERNEL
+#define PAGE_OFFSET		PAGE_OFFSET_L3
 #else
-#define PAGE_OFFSET		_AC(CONFIG_PAGE_OFFSET, UL)
+#define PAGE_OFFSET		kernel_map.page_offset
+#endif /* CONFIG_XIP_KERNEL */
+#else
+#define PAGE_OFFSET		_AC(0xc0000000, UL)
 #endif /* CONFIG_64BIT */
+#else
+#define PAGE_OFFSET		((unsigned long)phys_ram_base)
+#endif /* CONFIG_MMU */
 
 #ifndef __ASSEMBLY__
 
@@ -95,14 +96,9 @@ typedef struct page *pgtable_t;
 #define MIN_MEMBLOCK_ADDR      0
 #endif
 
-#ifdef CONFIG_MMU
 #define ARCH_PFN_OFFSET		(PFN_DOWN((unsigned long)phys_ram_base))
-#else
-#define ARCH_PFN_OFFSET		(PAGE_OFFSET >> PAGE_SHIFT)
-#endif /* CONFIG_MMU */
 
 struct kernel_mapping {
-	unsigned long page_offset;
 	unsigned long virt_addr;
 	unsigned long virt_offset;
 	uintptr_t phys_addr;
@@ -116,6 +112,7 @@ struct kernel_mapping {
 	uintptr_t xiprom;
 	uintptr_t xiprom_sz;
 #else
+	unsigned long page_offset;
 	unsigned long va_kernel_pa_offset;
 #endif
 };

arch/riscv/include/asm/pgtable.h

@@ -12,7 +12,11 @@
 #include <asm/pgtable-bits.h>
 
 #ifndef CONFIG_MMU
-#define KERNEL_LINK_ADDR	PAGE_OFFSET
+#ifdef CONFIG_RELOCATABLE
+#define KERNEL_LINK_ADDR	UL(0)
+#else
+#define KERNEL_LINK_ADDR	_AC(CONFIG_PHYS_RAM_BASE, UL)
+#endif
 #define KERN_VIRT_SIZE		(UL(-1))
 #else
 

arch/riscv/mm/init.c

@@ -20,15 +20,13 @@
 #include <linux/dma-map-ops.h>
 #include <linux/crash_dump.h>
 #include <linux/hugetlb.h>
-#ifdef CONFIG_RELOCATABLE
-#include <linux/elf.h>
-#endif
 #include <linux/kfence.h>
 #include <linux/execmem.h>
 
 #include <asm/fixmap.h>
 #include <asm/io.h>
 #include <asm/kasan.h>
+#include <asm/module.h>
 #include <asm/numa.h>
 #include <asm/pgtable.h>
 #include <asm/sections.h>
@@ -323,6 +321,44 @@ static void __init setup_bootmem(void)
 		hugetlb_cma_reserve(PUD_SHIFT - PAGE_SHIFT);
 }
 
+#ifdef CONFIG_RELOCATABLE
+extern unsigned long __rela_dyn_start, __rela_dyn_end;
+
+static void __init relocate_kernel(void)
+{
+	Elf_Rela *rela = (Elf_Rela *)&__rela_dyn_start;
+	/*
+	 * This holds the offset between the linked virtual address and the
+	 * relocated virtual address.
+	 */
+	uintptr_t reloc_offset = kernel_map.virt_addr - KERNEL_LINK_ADDR;
+	/*
+	 * This holds the offset between kernel linked virtual address and
+	 * physical address.
+	 */
+	uintptr_t va_kernel_link_pa_offset = KERNEL_LINK_ADDR - kernel_map.phys_addr;
+
+	for ( ; rela < (Elf_Rela *)&__rela_dyn_end; rela++) {
+		Elf_Addr addr = (rela->r_offset - va_kernel_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 >= KERNEL_LINK_ADDR)
+			relocated_addr += reloc_offset;
+
+		*(Elf_Addr *)addr = relocated_addr;
+	}
+}
+#endif /* CONFIG_RELOCATABLE */
+
 #ifdef CONFIG_MMU
 struct pt_alloc_ops pt_ops __meminitdata;
 
@@ -823,6 +859,8 @@ static __init void set_satp_mode(uintptr_t dtb_pa)
 	uintptr_t set_satp_mode_pmd = ((unsigned long)set_satp_mode) & PMD_MASK;
 	u64 satp_mode_cmdline = __pi_set_satp_mode_from_cmdline(dtb_pa);
 
+	kernel_map.page_offset = PAGE_OFFSET_L5;
+
 	if (satp_mode_cmdline == SATP_MODE_57) {
 		disable_pgtable_l5();
 	} else if (satp_mode_cmdline == SATP_MODE_48) {
@@ -893,44 +931,6 @@ static __init void set_satp_mode(uintptr_t dtb_pa)
 #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;
-
-static void __init relocate_kernel(void)
-{
-	Elf64_Rela *rela = (Elf64_Rela *)&__rela_dyn_start;
-	/*
-	 * This holds the offset between the linked virtual address and the
-	 * relocated virtual address.
-	 */
-	uintptr_t reloc_offset = kernel_map.virt_addr - KERNEL_LINK_ADDR;
-	/*
-	 * This holds the offset between kernel linked virtual address and
-	 * physical address.
-	 */
-	uintptr_t va_kernel_link_pa_offset = KERNEL_LINK_ADDR - kernel_map.phys_addr;
-
-	for ( ; rela < (Elf64_Rela *)&__rela_dyn_end; rela++) {
-		Elf64_Addr addr = (rela->r_offset - va_kernel_link_pa_offset);
-		Elf64_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 >= KERNEL_LINK_ADDR)
-			relocated_addr += reloc_offset;
-
-		*(Elf64_Addr *)addr = relocated_addr;
-	}
-}
-#endif /* CONFIG_RELOCATABLE */
-
 #ifdef CONFIG_XIP_KERNEL
 static void __init create_kernel_page_table(pgd_t *pgdir,
 					    __always_unused bool early)
@@ -1108,11 +1108,6 @@ asmlinkage void __init setup_vm(uintptr_t dtb_pa)
 	kernel_map.virt_addr = KERNEL_LINK_ADDR + kernel_map.virt_offset;
 
 #ifdef CONFIG_XIP_KERNEL
-#ifdef CONFIG_64BIT
-	kernel_map.page_offset = PAGE_OFFSET_L3;
-#else
-	kernel_map.page_offset = _AC(CONFIG_PAGE_OFFSET, UL);
-#endif
 	kernel_map.xiprom = (uintptr_t)CONFIG_XIP_PHYS_ADDR;
 	kernel_map.xiprom_sz = (uintptr_t)(&_exiprom) - (uintptr_t)(&_xiprom);
 
@@ -1127,7 +1122,6 @@ asmlinkage void __init setup_vm(uintptr_t dtb_pa)
 	kernel_map.va_kernel_xip_data_pa_offset = kernel_map.virt_addr - kernel_map.phys_addr
 						+ (uintptr_t)&_sdata - (uintptr_t)&_start;
 #else
-	kernel_map.page_offset = _AC(CONFIG_PAGE_OFFSET, UL);
 	kernel_map.phys_addr = (uintptr_t)(&_start);
 	kernel_map.size = (uintptr_t)(&_end) - kernel_map.phys_addr;
 	kernel_map.va_kernel_pa_offset = kernel_map.virt_addr - kernel_map.phys_addr;
@@ -1174,7 +1168,8 @@ asmlinkage void __init setup_vm(uintptr_t dtb_pa)
 	 * makes the kernel cross over a PUD_SIZE boundary, raise a bug
 	 * since a part of the kernel would not get mapped.
 	 */
-	BUG_ON(PUD_SIZE - (kernel_map.virt_addr & (PUD_SIZE - 1)) < kernel_map.size);
+	if (IS_ENABLED(CONFIG_64BIT))
+		BUG_ON(PUD_SIZE - (kernel_map.virt_addr & (PUD_SIZE - 1)) < kernel_map.size);
 	relocate_kernel();
 #endif
 
@@ -1378,6 +1373,12 @@ asmlinkage void __init setup_vm(uintptr_t dtb_pa)
 {
 	dtb_early_va = (void *)dtb_pa;
 	dtb_early_pa = dtb_pa;
+
+#ifdef CONFIG_RELOCATABLE
+	kernel_map.virt_addr = (uintptr_t)_start;
+	kernel_map.phys_addr = (uintptr_t)_start;
+	relocate_kernel();
+#endif
 }
 
 static inline void setup_vm_final(void)

include/asm-generic/module.h

@@ -19,12 +19,8 @@ struct mod_arch_specific
 #define Elf_Dyn		Elf64_Dyn
 #define Elf_Ehdr	Elf64_Ehdr
 #define Elf_Addr	Elf64_Addr
-#ifdef CONFIG_MODULES_USE_ELF_REL
 #define Elf_Rel		Elf64_Rel
-#endif
-#ifdef CONFIG_MODULES_USE_ELF_RELA
 #define Elf_Rela	Elf64_Rela
-#endif
 #define ELF_R_TYPE(X)	ELF64_R_TYPE(X)
 #define ELF_R_SYM(X)	ELF64_R_SYM(X)
 
@@ -36,12 +32,8 @@ struct mod_arch_specific
 #define Elf_Dyn		Elf32_Dyn
 #define Elf_Ehdr	Elf32_Ehdr
 #define Elf_Addr	Elf32_Addr
-#ifdef CONFIG_MODULES_USE_ELF_REL
 #define Elf_Rel		Elf32_Rel
-#endif
-#ifdef CONFIG_MODULES_USE_ELF_RELA
 #define Elf_Rela	Elf32_Rela
-#endif
 #define ELF_R_TYPE(X)	ELF32_R_TYPE(X)
 #define ELF_R_SYM(X)	ELF32_R_SYM(X)
 #endif