From 4971531af319f8bdd9a81a87eecfb6b19f2f8c8e Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Tue, 21 Jun 2016 23:11:38 +0100 Subject: x86/efi: Test for EFI_MEMMAP functionality when iterating EFI memmap Both efi_find_mirror() and efi_fake_memmap() really want to know whether the EFI memory map is available, not just whether the machine was booted using EFI. efi_fake_memmap() even has a check for EFI_MEMMAP at the start of the function. Since we've already got other code that has this dependency, merge everything under one if() conditional, and remove the now superfluous check from efi_fake_memmap(). Tested-by: Dave Young [kexec/kdump] Tested-by: Ard Biesheuvel [arm] Acked-by: Ard Biesheuvel Cc: Taku Izumi Cc: Tony Luck Cc: Xishi Qiu Cc: Kamezawa Hiroyuki Signed-off-by: Matt Fleming --- arch/x86/kernel/setup.c | 17 ++++++++--------- 1 file changed, 8 insertions(+), 9 deletions(-) (limited to 'arch') diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index 0fa60f5f5a16..4fd69e532c15 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -1096,19 +1096,18 @@ void __init setup_arch(char **cmdline_p) memblock_set_current_limit(ISA_END_ADDRESS); memblock_x86_fill(); - if (efi_enabled(EFI_BOOT)) { + reserve_bios_regions(); + + if (efi_enabled(EFI_MEMMAP)) { efi_fake_memmap(); efi_find_mirror(); - } - - reserve_bios_regions(); - /* - * The EFI specification says that boot service code won't be called - * after ExitBootServices(). This is, in fact, a lie. - */ - if (efi_enabled(EFI_MEMMAP)) + /* + * The EFI specification says that boot service code won't be + * called after ExitBootServices(). This is, in fact, a lie. + */ efi_reserve_boot_services(); + } /* preallocate 4k for mptable mpc */ early_reserve_e820_mpc_new(); -- cgit 1.4.1 From ab72a27da4c6c19b0e3d6d7556fdd4afb581c8ac Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Mon, 20 Jun 2016 14:36:51 +0100 Subject: x86/efi: Consolidate region mapping logic EFI regions are currently mapped in two separate places. The bulk of the work is done in efi_map_regions() but when CONFIG_EFI_MIXED is enabled the additional regions that are required when operating in mixed mode are mapping in efi_setup_page_tables(). Pull everything into efi_map_regions() and refactor the test for which regions should be mapped into a should_map_region() function. Generously sprinkle comments to clarify the different cases. Acked-by: Borislav Petkov Tested-by: Dave Young [kexec/kdump] Tested-by: Ard Biesheuvel [arm] Acked-by: Ard Biesheuvel Signed-off-by: Matt Fleming --- arch/x86/platform/efi/efi.c | 50 ++++++++++++++++++++++++++++++++++++------ arch/x86/platform/efi/efi_64.c | 20 ----------------- 2 files changed, 43 insertions(+), 27 deletions(-) (limited to 'arch') diff --git a/arch/x86/platform/efi/efi.c b/arch/x86/platform/efi/efi.c index 1fbb408e2e72..625ec729b4e8 100644 --- a/arch/x86/platform/efi/efi.c +++ b/arch/x86/platform/efi/efi.c @@ -745,6 +745,46 @@ static void *efi_map_next_entry(void *entry) return entry; } +static bool should_map_region(efi_memory_desc_t *md) +{ + /* + * Runtime regions always require runtime mappings (obviously). + */ + if (md->attribute & EFI_MEMORY_RUNTIME) + return true; + + /* + * 32-bit EFI doesn't suffer from the bug that requires us to + * reserve boot services regions, and mixed mode support + * doesn't exist for 32-bit kernels. + */ + if (IS_ENABLED(CONFIG_X86_32)) + return false; + + /* + * Map all of RAM so that we can access arguments in the 1:1 + * mapping when making EFI runtime calls. + */ + if (IS_ENABLED(CONFIG_EFI_MIXED) && !efi_is_native()) { + if (md->type == EFI_CONVENTIONAL_MEMORY || + md->type == EFI_LOADER_DATA || + md->type == EFI_LOADER_CODE) + return true; + } + + /* + * Map boot services regions as a workaround for buggy + * firmware that accesses them even when they shouldn't. + * + * See efi_{reserve,free}_boot_services(). + */ + if (md->type == EFI_BOOT_SERVICES_CODE || + md->type == EFI_BOOT_SERVICES_DATA) + return true; + + return false; +} + /* * Map the efi memory ranges of the runtime services and update new_mmap with * virtual addresses. @@ -761,13 +801,9 @@ static void * __init efi_map_regions(int *count, int *pg_shift) p = NULL; while ((p = efi_map_next_entry(p))) { md = p; - if (!(md->attribute & EFI_MEMORY_RUNTIME)) { -#ifdef CONFIG_X86_64 - if (md->type != EFI_BOOT_SERVICES_CODE && - md->type != EFI_BOOT_SERVICES_DATA) -#endif - continue; - } + + if (!should_map_region(md)) + continue; efi_map_region(md); get_systab_virt_addr(md); diff --git a/arch/x86/platform/efi/efi_64.c b/arch/x86/platform/efi/efi_64.c index 677e29e29473..45434ea345e9 100644 --- a/arch/x86/platform/efi/efi_64.c +++ b/arch/x86/platform/efi/efi_64.c @@ -214,7 +214,6 @@ void efi_sync_low_kernel_mappings(void) int __init efi_setup_page_tables(unsigned long pa_memmap, unsigned num_pages) { unsigned long pfn, text; - efi_memory_desc_t *md; struct page *page; unsigned npages; pgd_t *pgd; @@ -248,25 +247,6 @@ int __init efi_setup_page_tables(unsigned long pa_memmap, unsigned num_pages) if (!IS_ENABLED(CONFIG_EFI_MIXED)) return 0; - /* - * Map all of RAM so that we can access arguments in the 1:1 - * mapping when making EFI runtime calls. - */ - for_each_efi_memory_desc(md) { - if (md->type != EFI_CONVENTIONAL_MEMORY && - md->type != EFI_LOADER_DATA && - md->type != EFI_LOADER_CODE) - continue; - - pfn = md->phys_addr >> PAGE_SHIFT; - npages = md->num_pages; - - if (kernel_map_pages_in_pgd(pgd, pfn, md->phys_addr, npages, _PAGE_RW)) { - pr_err("Failed to map 1:1 memory\n"); - return 1; - } - } - page = alloc_page(GFP_KERNEL|__GFP_DMA32); if (!page) panic("Unable to allocate EFI runtime stack < 4GB\n"); -- cgit 1.4.1 From 9479c7cebfb568f8b8b424be7f1cac120e9eea95 Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Fri, 26 Feb 2016 21:22:05 +0000 Subject: efi: Refactor efi_memmap_init_early() into arch-neutral code Every EFI architecture apart from ia64 needs to setup the EFI memory map at efi.memmap, and the code for doing that is essentially the same across all implementations. Therefore, it makes sense to factor this out into the common code under drivers/firmware/efi/. The only slight variation is the data structure out of which we pull the initial memory map information, such as physical address, memory descriptor size and version, etc. We can address this by passing a generic data structure (struct efi_memory_map_data) as the argument to efi_memmap_init_early() which contains the minimum info required for initialising the memory map. In the process, this patch also fixes a few undesirable implementation differences: - ARM and arm64 were failing to clear the EFI_MEMMAP bit when unmapping the early EFI memory map. EFI_MEMMAP indicates whether the EFI memory map is mapped (not the regions contained within) and can be traversed. It's more correct to set the bit as soon as we memremap() the passed in EFI memmap. - Rename efi_unmmap_memmap() to efi_memmap_unmap() to adhere to the regular naming scheme. This patch also uses a read-write mapping for the memory map instead of the read-only mapping currently used on ARM and arm64. x86 needs the ability to update the memory map in-place when assigning virtual addresses to regions (efi_map_region()) and tagging regions when reserving boot services (efi_reserve_boot_services()). There's no way for the generic fake_mem code to know which mapping to use without introducing some arch-specific constant/hook, so just use read-write since read-only is of dubious value for the EFI memory map. Tested-by: Dave Young [kexec/kdump] Tested-by: Ard Biesheuvel [arm] Acked-by: Ard Biesheuvel Cc: Leif Lindholm Cc: Peter Jones Cc: Borislav Petkov Cc: Mark Rutland Signed-off-by: Matt Fleming --- arch/x86/include/asm/efi.h | 1 - arch/x86/platform/efi/efi.c | 66 ++++++++++---------------------------- arch/x86/platform/efi/quirks.c | 4 +-- drivers/firmware/efi/arm-init.c | 17 +++++----- drivers/firmware/efi/arm-runtime.c | 2 +- drivers/firmware/efi/efi.c | 46 ++++++++++++++++++++++++++ drivers/firmware/efi/fake_mem.c | 15 +++++---- include/linux/efi.h | 16 +++++++++ 8 files changed, 99 insertions(+), 68 deletions(-) (limited to 'arch') diff --git a/arch/x86/include/asm/efi.h b/arch/x86/include/asm/efi.h index d0bb76d81402..4630e2bfa8fb 100644 --- a/arch/x86/include/asm/efi.h +++ b/arch/x86/include/asm/efi.h @@ -117,7 +117,6 @@ extern int __init efi_memblock_x86_reserve_range(void); extern pgd_t * __init efi_call_phys_prolog(void); extern void __init efi_call_phys_epilog(pgd_t *save_pgd); extern void __init efi_print_memmap(void); -extern void __init efi_unmap_memmap(void); extern void __init efi_memory_uc(u64 addr, unsigned long size); extern void __init efi_map_region(efi_memory_desc_t *md); extern void __init efi_map_region_fixed(efi_memory_desc_t *md); diff --git a/arch/x86/platform/efi/efi.c b/arch/x86/platform/efi/efi.c index 625ec729b4e8..5ccde8b6cdd1 100644 --- a/arch/x86/platform/efi/efi.c +++ b/arch/x86/platform/efi/efi.c @@ -172,7 +172,9 @@ static void __init do_add_efi_memmap(void) int __init efi_memblock_x86_reserve_range(void) { struct efi_info *e = &boot_params.efi_info; + struct efi_memory_map_data data; phys_addr_t pmap; + int rv; if (efi_enabled(EFI_PARAVIRT)) return 0; @@ -187,11 +189,17 @@ int __init efi_memblock_x86_reserve_range(void) #else pmap = (e->efi_memmap | ((__u64)e->efi_memmap_hi << 32)); #endif - efi.memmap.phys_map = pmap; - efi.memmap.nr_map = e->efi_memmap_size / - e->efi_memdesc_size; - efi.memmap.desc_size = e->efi_memdesc_size; - efi.memmap.desc_version = e->efi_memdesc_version; + data.phys_map = pmap; + data.size = e->efi_memmap_size; + data.desc_size = e->efi_memdesc_size; + data.desc_version = e->efi_memdesc_version; + + rv = efi_memmap_init_early(&data); + if (rv) + return rv; + + if (add_efi_memmap) + do_add_efi_memmap(); WARN(efi.memmap.desc_version != 1, "Unexpected EFI_MEMORY_DESCRIPTOR version %ld", @@ -218,19 +226,6 @@ void __init efi_print_memmap(void) } } -void __init efi_unmap_memmap(void) -{ - unsigned long size; - - clear_bit(EFI_MEMMAP, &efi.flags); - - size = efi.memmap.nr_map * efi.memmap.desc_size; - if (efi.memmap.map) { - early_memunmap(efi.memmap.map, size); - efi.memmap.map = NULL; - } -} - static int __init efi_systab_init(void *phys) { if (efi_enabled(EFI_64BIT)) { @@ -414,33 +409,6 @@ static int __init efi_runtime_init(void) return 0; } -static int __init efi_memmap_init(void) -{ - unsigned long addr, size; - - if (efi_enabled(EFI_PARAVIRT)) - return 0; - - /* Map the EFI memory map */ - size = efi.memmap.nr_map * efi.memmap.desc_size; - addr = (unsigned long)efi.memmap.phys_map; - - efi.memmap.map = early_memremap(addr, size); - if (efi.memmap.map == NULL) { - pr_err("Could not map the memory map!\n"); - return -ENOMEM; - } - - efi.memmap.map_end = efi.memmap.map + size; - - if (add_efi_memmap) - do_add_efi_memmap(); - - set_bit(EFI_MEMMAP, &efi.flags); - - return 0; -} - void __init efi_init(void) { efi_char16_t *c16; @@ -498,11 +466,11 @@ void __init efi_init(void) if (!efi_runtime_supported()) pr_info("No EFI runtime due to 32/64-bit mismatch with kernel\n"); else { - if (efi_runtime_disabled() || efi_runtime_init()) + if (efi_runtime_disabled() || efi_runtime_init()) { + efi_memmap_unmap(); return; + } } - if (efi_memmap_init()) - return; if (efi_enabled(EFI_DBG)) efi_print_memmap(); @@ -839,7 +807,7 @@ static void __init kexec_enter_virtual_mode(void) * non-native EFI */ if (!efi_is_native()) { - efi_unmap_memmap(); + efi_memmap_unmap(); clear_bit(EFI_RUNTIME_SERVICES, &efi.flags); return; } diff --git a/arch/x86/platform/efi/quirks.c b/arch/x86/platform/efi/quirks.c index 89d1146f5a6f..47b99108ff8e 100644 --- a/arch/x86/platform/efi/quirks.c +++ b/arch/x86/platform/efi/quirks.c @@ -287,7 +287,7 @@ void __init efi_free_boot_services(void) free_bootmem_late(start, size); } - efi_unmap_memmap(); + efi_memmap_unmap(); } /* @@ -365,7 +365,7 @@ void __init efi_apply_memmap_quirks(void) */ if (!efi_runtime_supported()) { pr_info("Setup done, disabling due to 32/64-bit mismatch\n"); - efi_unmap_memmap(); + efi_memmap_unmap(); } /* UV2+ BIOS has a fix for this issue. UV1 still needs the quirk. */ diff --git a/drivers/firmware/efi/arm-init.c b/drivers/firmware/efi/arm-init.c index c49d50e68aee..5a2df3fefccc 100644 --- a/drivers/firmware/efi/arm-init.c +++ b/drivers/firmware/efi/arm-init.c @@ -211,12 +211,11 @@ static __init void reserve_regions(void) memblock_mark_nomap(paddr, size); } - - set_bit(EFI_MEMMAP, &efi.flags); } void __init efi_init(void) { + struct efi_memory_map_data data; struct efi_fdt_params params; /* Grab UEFI information placed in FDT by stub */ @@ -225,9 +224,12 @@ void __init efi_init(void) efi_system_table = params.system_table; - efi.memmap.phys_map = params.mmap; - efi.memmap.map = early_memremap_ro(params.mmap, params.mmap_size); - if (efi.memmap.map == NULL) { + data.desc_version = params.desc_ver; + data.desc_size = params.desc_size; + data.size = params.mmap_size; + data.phys_map = params.mmap; + + if (efi_memmap_init_early(&data) < 0) { /* * If we are booting via UEFI, the UEFI memory map is the only * description of memory we have, so there is little point in @@ -235,9 +237,6 @@ void __init efi_init(void) */ panic("Unable to map EFI memory map.\n"); } - efi.memmap.map_end = efi.memmap.map + params.mmap_size; - efi.memmap.desc_size = params.desc_size; - efi.memmap.desc_version = params.desc_ver; WARN(efi.memmap.desc_version != 1, "Unexpected EFI_MEMORY_DESCRIPTOR version %ld", @@ -248,7 +247,7 @@ void __init efi_init(void) reserve_regions(); efi_memattr_init(); - early_memunmap(efi.memmap.map, params.mmap_size); + efi_memmap_unmap(); memblock_reserve(params.mmap & PAGE_MASK, PAGE_ALIGN(params.mmap_size + diff --git a/drivers/firmware/efi/arm-runtime.c b/drivers/firmware/efi/arm-runtime.c index c394b81fe452..eedb30351a68 100644 --- a/drivers/firmware/efi/arm-runtime.c +++ b/drivers/firmware/efi/arm-runtime.c @@ -114,7 +114,7 @@ static int __init arm_enable_runtime_services(void) pr_info("Remapping and enabling EFI services.\n"); - mapsize = efi.memmap.map_end - efi.memmap.map; + mapsize = efi.memmap.desc_size * efi.memmap.nr_map; efi.memmap.map = memremap(efi.memmap.phys_map, mapsize, MEMREMAP_WB); if (!efi.memmap.map) { diff --git a/drivers/firmware/efi/efi.c b/drivers/firmware/efi/efi.c index 5a2631af7410..c1879999abe7 100644 --- a/drivers/firmware/efi/efi.c +++ b/drivers/firmware/efi/efi.c @@ -544,6 +544,52 @@ int __init efi_config_init(efi_config_table_type_t *arch_tables) return ret; } +/** + * efi_memmap_init_early - Map the EFI memory map data structure + * @data: EFI memory map data + * + * Use early_memremap() to map the passed in EFI memory map and assign + * it to efi.memmap. + */ +int __init efi_memmap_init_early(struct efi_memory_map_data *data) +{ + struct efi_memory_map map; + + if (efi_enabled(EFI_PARAVIRT)) + return 0; + + map.phys_map = data->phys_map; + + map.map = early_memremap(data->phys_map, data->size); + if (!map.map) { + pr_err("Could not map the memory map!\n"); + return -ENOMEM; + } + + map.nr_map = data->size / data->desc_size; + map.map_end = map.map + data->size; + + map.desc_version = data->desc_version; + map.desc_size = data->desc_size; + + set_bit(EFI_MEMMAP, &efi.flags); + + efi.memmap = map; + + return 0; +} + +void __init efi_memmap_unmap(void) +{ + unsigned long size; + + size = efi.memmap.desc_size * efi.memmap.nr_map; + + early_memunmap(efi.memmap.map, size); + efi.memmap.map = NULL; + clear_bit(EFI_MEMMAP, &efi.flags); +} + #ifdef CONFIG_EFI_VARS_MODULE static int __init efi_load_efivars(void) { diff --git a/drivers/firmware/efi/fake_mem.c b/drivers/firmware/efi/fake_mem.c index c437388a7b85..939eec47139f 100644 --- a/drivers/firmware/efi/fake_mem.c +++ b/drivers/firmware/efi/fake_mem.c @@ -57,6 +57,7 @@ static int __init cmp_fake_mem(const void *x1, const void *x2) void __init efi_fake_memmap(void) { u64 start, end, m_start, m_end, m_attr; + struct efi_memory_map_data data; int new_nr_map = efi.memmap.nr_map; efi_memory_desc_t *md; phys_addr_t new_memmap_phy; @@ -180,12 +181,14 @@ void __init efi_fake_memmap(void) } /* swap into new EFI memmap */ - efi_unmap_memmap(); - efi.memmap.map = new_memmap; - efi.memmap.phys_map = new_memmap_phy; - efi.memmap.nr_map = new_nr_map; - efi.memmap.map_end = efi.memmap.map + efi.memmap.nr_map * efi.memmap.desc_size; - set_bit(EFI_MEMMAP, &efi.flags); + early_memunmap(new_memmap, efi.memmap.desc_size * new_nr_map); + efi_memmap_unmap(); + + data.phys_map = new_memmap_phy; + data.size = efi.memmap.desc_size * new_nr_map; + data.desc_version = efi.memmap.desc_version; + data.desc_size = efi.memmap.desc_size; + efi_memmap_init_early(&data); /* print new EFI memmap */ efi_print_memmap(); diff --git a/include/linux/efi.h b/include/linux/efi.h index 7f5a58225385..d862d4998580 100644 --- a/include/linux/efi.h +++ b/include/linux/efi.h @@ -669,6 +669,18 @@ typedef struct { unsigned long tables; } efi_system_table_t; +/* + * Architecture independent structure for describing a memory map for the + * benefit of efi_memmap_init_early(), saving us the need to pass four + * parameters. + */ +struct efi_memory_map_data { + phys_addr_t phys_map; + unsigned long size; + unsigned long desc_version; + unsigned long desc_size; +}; + struct efi_memory_map { phys_addr_t phys_map; void *map; @@ -900,6 +912,10 @@ static inline efi_status_t efi_query_variable_store(u32 attributes, } #endif extern void __iomem *efi_lookup_mapped_addr(u64 phys_addr); + +extern int __init efi_memmap_init_early(struct efi_memory_map_data *data); +extern void __init efi_memmap_unmap(void); + extern int efi_config_init(efi_config_table_type_t *arch_tables); #ifdef CONFIG_EFI_ESRT extern void __init efi_esrt_init(void); -- cgit 1.4.1 From dca0f971ea6fcf2f1bb78f7995adf80da9f4767f Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Sat, 27 Feb 2016 15:52:50 +0000 Subject: efi: Add efi_memmap_init_late() for permanent EFI memmap Drivers need a way to access the EFI memory map at runtime. ARM and arm64 currently provide this by remapping the EFI memory map into the vmalloc space before setting up the EFI virtual mappings. x86 does not provide this functionality which has resulted in the code in efi_mem_desc_lookup() where it will manually map individual EFI memmap entries if the memmap has already been torn down on x86, /* * If a driver calls this after efi_free_boot_services, * ->map will be NULL, and the target may also not be mapped. * So just always get our own virtual map on the CPU. * */ md = early_memremap(p, sizeof (*md)); There isn't a good reason for not providing a permanent EFI memory map for runtime queries, especially since the EFI regions are not mapped into the standard kernel page tables. Tested-by: Dave Young [kexec/kdump] Tested-by: Ard Biesheuvel [arm] Acked-by: Ard Biesheuvel Cc: Leif Lindholm Cc: Peter Jones Cc: Borislav Petkov Cc: Mark Rutland Signed-off-by: Matt Fleming --- arch/x86/platform/efi/efi.c | 44 ++++++++---- arch/x86/platform/efi/quirks.c | 2 - drivers/firmware/efi/arm-runtime.c | 4 +- drivers/firmware/efi/efi.c | 135 ++++++++++++++++++++++++++----------- include/linux/efi.h | 2 + 5 files changed, 130 insertions(+), 57 deletions(-) (limited to 'arch') diff --git a/arch/x86/platform/efi/efi.c b/arch/x86/platform/efi/efi.c index 5ccde8b6cdd1..33996987ac70 100644 --- a/arch/x86/platform/efi/efi.c +++ b/arch/x86/platform/efi/efi.c @@ -827,6 +827,19 @@ static void __init kexec_enter_virtual_mode(void) get_systab_virt_addr(md); } + /* + * Unregister the early EFI memmap from efi_init() and install + * the new EFI memory map. + */ + efi_memmap_unmap(); + + if (efi_memmap_init_late(efi.memmap.phys_map, + efi.memmap.desc_size * efi.memmap.nr_map)) { + pr_err("Failed to remap late EFI memory map\n"); + clear_bit(EFI_RUNTIME_SERVICES, &efi.flags); + return; + } + save_runtime_map(); BUG_ON(!efi.systab); @@ -888,6 +901,7 @@ static void __init __efi_enter_virtual_mode(void) int count = 0, pg_shift = 0; void *new_memmap = NULL; efi_status_t status; + phys_addr_t pa; efi.systab = NULL; @@ -905,11 +919,26 @@ static void __init __efi_enter_virtual_mode(void) return; } + pa = __pa(new_memmap); + + /* + * Unregister the early EFI memmap from efi_init() and install + * the new EFI memory map that we are about to pass to the + * firmware via SetVirtualAddressMap(). + */ + efi_memmap_unmap(); + + if (efi_memmap_init_late(pa, efi.memmap.desc_size * count)) { + pr_err("Failed to remap late EFI memory map\n"); + clear_bit(EFI_RUNTIME_SERVICES, &efi.flags); + return; + } + save_runtime_map(); BUG_ON(!efi.systab); - if (efi_setup_page_tables(__pa(new_memmap), 1 << pg_shift)) { + if (efi_setup_page_tables(pa, 1 << pg_shift)) { clear_bit(EFI_RUNTIME_SERVICES, &efi.flags); return; } @@ -921,14 +950,14 @@ static void __init __efi_enter_virtual_mode(void) efi.memmap.desc_size * count, efi.memmap.desc_size, efi.memmap.desc_version, - (efi_memory_desc_t *)__pa(new_memmap)); + (efi_memory_desc_t *)pa); } else { status = efi_thunk_set_virtual_address_map( efi_phys.set_virtual_address_map, efi.memmap.desc_size * count, efi.memmap.desc_size, efi.memmap.desc_version, - (efi_memory_desc_t *)__pa(new_memmap)); + (efi_memory_desc_t *)pa); } if (status != EFI_SUCCESS) { @@ -960,15 +989,6 @@ static void __init __efi_enter_virtual_mode(void) efi_runtime_update_mappings(); efi_dump_pagetable(); - /* - * We mapped the descriptor array into the EFI pagetable above - * but we're not unmapping it here because if we're running in - * EFI mixed mode we need all of memory to be accessible when - * we pass parameters to the EFI runtime services in the - * thunking code. - */ - free_pages((unsigned long)new_memmap, pg_shift); - /* clean DUMMY object */ efi_delete_dummy_variable(); } diff --git a/arch/x86/platform/efi/quirks.c b/arch/x86/platform/efi/quirks.c index 47b99108ff8e..9faf18874692 100644 --- a/arch/x86/platform/efi/quirks.c +++ b/arch/x86/platform/efi/quirks.c @@ -286,8 +286,6 @@ void __init efi_free_boot_services(void) free_bootmem_late(start, size); } - - efi_memmap_unmap(); } /* diff --git a/drivers/firmware/efi/arm-runtime.c b/drivers/firmware/efi/arm-runtime.c index eedb30351a68..ae001450545f 100644 --- a/drivers/firmware/efi/arm-runtime.c +++ b/drivers/firmware/efi/arm-runtime.c @@ -116,12 +116,10 @@ static int __init arm_enable_runtime_services(void) mapsize = efi.memmap.desc_size * efi.memmap.nr_map; - efi.memmap.map = memremap(efi.memmap.phys_map, mapsize, MEMREMAP_WB); - if (!efi.memmap.map) { + if (efi_memmap_init_late(efi.memmap.phys_map, mapsize)) { pr_err("Failed to remap EFI memory map\n"); return -ENOMEM; } - efi.memmap.map_end = efi.memmap.map + mapsize; if (!efi_virtmap_init()) { pr_err("UEFI virtual mapping missing or invalid -- runtime services will not be available\n"); diff --git a/drivers/firmware/efi/efi.c b/drivers/firmware/efi/efi.c index c1879999abe7..8a5e0db72b8f 100644 --- a/drivers/firmware/efi/efi.c +++ b/drivers/firmware/efi/efi.c @@ -347,56 +347,31 @@ subsys_initcall(efisubsys_init); /* * Find the efi memory descriptor for a given physical address. Given a - * physicall address, determine if it exists within an EFI Memory Map entry, + * physical address, determine if it exists within an EFI Memory Map entry, * and if so, populate the supplied memory descriptor with the appropriate * data. */ int __init efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md) { - struct efi_memory_map *map = &efi.memmap; - phys_addr_t p, e; + efi_memory_desc_t *md; if (!efi_enabled(EFI_MEMMAP)) { pr_err_once("EFI_MEMMAP is not enabled.\n"); return -EINVAL; } - if (!map) { - pr_err_once("efi.memmap is not set.\n"); - return -EINVAL; - } if (!out_md) { pr_err_once("out_md is null.\n"); return -EINVAL; } - if (WARN_ON_ONCE(!map->phys_map)) - return -EINVAL; - if (WARN_ON_ONCE(map->nr_map == 0) || WARN_ON_ONCE(map->desc_size == 0)) - return -EINVAL; - e = map->phys_map + map->nr_map * map->desc_size; - for (p = map->phys_map; p < e; p += map->desc_size) { - efi_memory_desc_t *md; + for_each_efi_memory_desc(md) { u64 size; u64 end; - /* - * If a driver calls this after efi_free_boot_services, - * ->map will be NULL, and the target may also not be mapped. - * So just always get our own virtual map on the CPU. - * - */ - md = early_memremap(p, sizeof (*md)); - if (!md) { - pr_err_once("early_memremap(%pa, %zu) failed.\n", - &p, sizeof (*md)); - return -ENOMEM; - } - if (!(md->attribute & EFI_MEMORY_RUNTIME) && md->type != EFI_BOOT_SERVICES_DATA && md->type != EFI_RUNTIME_SERVICES_DATA) { - early_memunmap(md, sizeof (*md)); continue; } @@ -404,11 +379,8 @@ int __init efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md) end = md->phys_addr + size; if (phys_addr >= md->phys_addr && phys_addr < end) { memcpy(out_md, md, sizeof(*out_md)); - early_memunmap(md, sizeof (*md)); return 0; } - - early_memunmap(md, sizeof (*md)); } pr_err_once("requested map not found.\n"); return -ENOENT; @@ -545,32 +517,49 @@ int __init efi_config_init(efi_config_table_type_t *arch_tables) } /** - * efi_memmap_init_early - Map the EFI memory map data structure + * __efi_memmap_init - Common code for mapping the EFI memory map * @data: EFI memory map data + * @late: Use early or late mapping function? * - * Use early_memremap() to map the passed in EFI memory map and assign - * it to efi.memmap. + * This function takes care of figuring out which function to use to + * map the EFI memory map in efi.memmap based on how far into the boot + * we are. + * + * During bootup @late should be %false since we only have access to + * the early_memremap*() functions as the vmalloc space isn't setup. + * Once the kernel is fully booted we can fallback to the more robust + * memremap*() API. + * + * Returns zero on success, a negative error code on failure. */ -int __init efi_memmap_init_early(struct efi_memory_map_data *data) +static int __init +__efi_memmap_init(struct efi_memory_map_data *data, bool late) { struct efi_memory_map map; + phys_addr_t phys_map; if (efi_enabled(EFI_PARAVIRT)) return 0; - map.phys_map = data->phys_map; + phys_map = data->phys_map; + + if (late) + map.map = memremap(phys_map, data->size, MEMREMAP_WB); + else + map.map = early_memremap(phys_map, data->size); - map.map = early_memremap(data->phys_map, data->size); if (!map.map) { pr_err("Could not map the memory map!\n"); return -ENOMEM; } + map.phys_map = data->phys_map; map.nr_map = data->size / data->desc_size; map.map_end = map.map + data->size; map.desc_version = data->desc_version; map.desc_size = data->desc_size; + map.late = late; set_bit(EFI_MEMMAP, &efi.flags); @@ -579,17 +568,83 @@ int __init efi_memmap_init_early(struct efi_memory_map_data *data) return 0; } +/** + * efi_memmap_init_early - Map the EFI memory map data structure + * @data: EFI memory map data + * + * Use early_memremap() to map the passed in EFI memory map and assign + * it to efi.memmap. + */ +int __init efi_memmap_init_early(struct efi_memory_map_data *data) +{ + /* Cannot go backwards */ + WARN_ON(efi.memmap.late); + + return __efi_memmap_init(data, false); +} + void __init efi_memmap_unmap(void) { - unsigned long size; + if (!efi.memmap.late) { + unsigned long size; - size = efi.memmap.desc_size * efi.memmap.nr_map; + size = efi.memmap.desc_size * efi.memmap.nr_map; + early_memunmap(efi.memmap.map, size); + } else { + memunmap(efi.memmap.map); + } - early_memunmap(efi.memmap.map, size); efi.memmap.map = NULL; clear_bit(EFI_MEMMAP, &efi.flags); } +/** + * efi_memmap_init_late - Map efi.memmap with memremap() + * @phys_addr: Physical address of the new EFI memory map + * @size: Size in bytes of the new EFI memory map + * + * Setup a mapping of the EFI memory map using ioremap_cache(). This + * function should only be called once the vmalloc space has been + * setup and is therefore not suitable for calling during early EFI + * initialise, e.g. in efi_init(). Additionally, it expects + * efi_memmap_init_early() to have already been called. + * + * The reason there are two EFI memmap initialisation + * (efi_memmap_init_early() and this late version) is because the + * early EFI memmap should be explicitly unmapped once EFI + * initialisation is complete as the fixmap space used to map the EFI + * memmap (via early_memremap()) is a scarce resource. + * + * This late mapping is intended to persist for the duration of + * runtime so that things like efi_mem_desc_lookup() and + * efi_mem_attributes() always work. + * + * Returns zero on success, a negative error code on failure. + */ +int __init efi_memmap_init_late(phys_addr_t addr, unsigned long size) +{ + struct efi_memory_map_data data = { + .phys_map = addr, + .size = size, + }; + + /* Did we forget to unmap the early EFI memmap? */ + WARN_ON(efi.memmap.map); + + /* Were we already called? */ + WARN_ON(efi.memmap.late); + + /* + * It makes no sense to allow callers to register different + * values for the following fields. Copy them out of the + * existing early EFI memmap. + */ + data.desc_version = efi.memmap.desc_version; + data.desc_size = efi.memmap.desc_size; + + return __efi_memmap_init(&data, true); +} + #ifdef CONFIG_EFI_VARS_MODULE static int __init efi_load_efivars(void) { diff --git a/include/linux/efi.h b/include/linux/efi.h index d862d4998580..f149676b2fcd 100644 --- a/include/linux/efi.h +++ b/include/linux/efi.h @@ -688,6 +688,7 @@ struct efi_memory_map { int nr_map; unsigned long desc_version; unsigned long desc_size; + bool late; }; struct efi_fdt_params { @@ -914,6 +915,7 @@ static inline efi_status_t efi_query_variable_store(u32 attributes, extern void __iomem *efi_lookup_mapped_addr(u64 phys_addr); extern int __init efi_memmap_init_early(struct efi_memory_map_data *data); +extern int __init efi_memmap_init_late(phys_addr_t addr, unsigned long size); extern void __init efi_memmap_unmap(void); extern int efi_config_init(efi_config_table_type_t *arch_tables); -- cgit 1.4.1 From 816e76129ed5fadd28e526c43397c79775194b5c Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Mon, 29 Feb 2016 21:22:52 +0000 Subject: efi: Allow drivers to reserve boot services forever Today, it is not possible for drivers to reserve EFI boot services for access after efi_free_boot_services() has been called on x86. For ARM/arm64 it can be done simply by calling memblock_reserve(). Having this ability for all three architectures is desirable for a couple of reasons, 1) It saves drivers copying data out of those regions 2) kexec reboot can now make use of things like ESRT Instead of using the standard memblock_reserve() which is insufficient to reserve the region on x86 (see efi_reserve_boot_services()), a new API is introduced in this patch; efi_mem_reserve(). efi.memmap now always represents which EFI memory regions are available. On x86 the EFI boot services regions that have not been reserved via efi_mem_reserve() will be removed from efi.memmap during efi_free_boot_services(). This has implications for kexec, since it is not possible for a newly kexec'd kernel to access the same boot services regions that the initial boot kernel had access to unless they are reserved by every kexec kernel in the chain. Tested-by: Dave Young [kexec/kdump] Tested-by: Ard Biesheuvel [arm] Acked-by: Ard Biesheuvel Cc: Leif Lindholm Cc: Peter Jones Cc: Borislav Petkov Cc: Mark Rutland Signed-off-by: Matt Fleming --- arch/x86/platform/efi/quirks.c | 121 +++++++++++++++++++++++++++++++++++++---- drivers/firmware/efi/efi.c | 30 ++++++++++ include/linux/efi.h | 1 + 3 files changed, 141 insertions(+), 11 deletions(-) (limited to 'arch') diff --git a/arch/x86/platform/efi/quirks.c b/arch/x86/platform/efi/quirks.c index 9faf18874692..f14b7a9da24b 100644 --- a/arch/x86/platform/efi/quirks.c +++ b/arch/x86/platform/efi/quirks.c @@ -163,6 +163,71 @@ efi_status_t efi_query_variable_store(u32 attributes, unsigned long size, } EXPORT_SYMBOL_GPL(efi_query_variable_store); +/* + * The UEFI specification makes it clear that the operating system is + * free to do whatever it wants with boot services code after + * ExitBootServices() has been called. Ignoring this recommendation a + * significant bunch of EFI implementations continue calling into boot + * services code (SetVirtualAddressMap). In order to work around such + * buggy implementations we reserve boot services region during EFI + * init and make sure it stays executable. Then, after + * SetVirtualAddressMap(), it is discarded. + * + * However, some boot services regions contain data that is required + * by drivers, so we need to track which memory ranges can never be + * freed. This is done by tagging those regions with the + * EFI_MEMORY_RUNTIME attribute. + * + * Any driver that wants to mark a region as reserved must use + * efi_mem_reserve() which will insert a new EFI memory descriptor + * into efi.memmap (splitting existing regions if necessary) and tag + * it with EFI_MEMORY_RUNTIME. + */ +void __init efi_arch_mem_reserve(phys_addr_t addr, u64 size) +{ + phys_addr_t new_phys, new_size; + struct efi_mem_range mr; + efi_memory_desc_t md; + int num_entries; + void *new; + + if (efi_mem_desc_lookup(addr, &md)) { + pr_err("Failed to lookup EFI memory descriptor for %pa\n", &addr); + return; + } + + if (addr + size > md.phys_addr + (md.num_pages << EFI_PAGE_SHIFT)) { + pr_err("Region spans EFI memory descriptors, %pa\n", &addr); + return; + } + + mr.range.start = addr; + mr.range.end = addr + size; + mr.attribute = md.attribute | EFI_MEMORY_RUNTIME; + + num_entries = efi_memmap_split_count(&md, &mr.range); + num_entries += efi.memmap.nr_map; + + new_size = efi.memmap.desc_size * num_entries; + + new_phys = memblock_alloc(new_size, 0); + if (!new_phys) { + pr_err("Could not allocate boot services memmap\n"); + return; + } + + new = early_memremap(new_phys, new_size); + if (!new) { + pr_err("Failed to map new boot services memmap\n"); + return; + } + + efi_memmap_insert(&efi.memmap, new, &mr); + early_memunmap(new, new_size); + + efi_memmap_install(new_phys, num_entries); +} + /* * Helper function for efi_reserve_boot_services() to figure out if we * can free regions in efi_free_boot_services(). @@ -184,15 +249,6 @@ static bool can_free_region(u64 start, u64 size) return true; } -/* - * The UEFI specification makes it clear that the operating system is free to do - * whatever it wants with boot services code after ExitBootServices() has been - * called. Ignoring this recommendation a significant bunch of EFI implementations - * continue calling into boot services code (SetVirtualAddressMap). In order to - * work around such buggy implementations we reserve boot services region during - * EFI init and make sure it stays executable. Then, after SetVirtualAddressMap(), it -* is discarded. -*/ void __init efi_reserve_boot_services(void) { efi_memory_desc_t *md; @@ -249,7 +305,10 @@ void __init efi_reserve_boot_services(void) void __init efi_free_boot_services(void) { + phys_addr_t new_phys, new_size; efi_memory_desc_t *md; + int num_entries = 0; + void *new, *new_md; for_each_efi_memory_desc(md) { unsigned long long start = md->phys_addr; @@ -257,12 +316,16 @@ void __init efi_free_boot_services(void) size_t rm_size; if (md->type != EFI_BOOT_SERVICES_CODE && - md->type != EFI_BOOT_SERVICES_DATA) + md->type != EFI_BOOT_SERVICES_DATA) { + num_entries++; continue; + } /* Do not free, someone else owns it: */ - if (md->attribute & EFI_MEMORY_RUNTIME) + if (md->attribute & EFI_MEMORY_RUNTIME) { + num_entries++; continue; + } /* * Nasty quirk: if all sub-1MB memory is used for boot @@ -286,6 +349,42 @@ void __init efi_free_boot_services(void) free_bootmem_late(start, size); } + + new_size = efi.memmap.desc_size * num_entries; + new_phys = memblock_alloc(new_size, 0); + if (!new_phys) { + pr_err("Failed to allocate new EFI memmap\n"); + return; + } + + new = memremap(new_phys, new_size, MEMREMAP_WB); + if (!new) { + pr_err("Failed to map new EFI memmap\n"); + return; + } + + /* + * Build a new EFI memmap that excludes any boot services + * regions that are not tagged EFI_MEMORY_RUNTIME, since those + * regions have now been freed. + */ + new_md = new; + for_each_efi_memory_desc(md) { + if (!(md->attribute & EFI_MEMORY_RUNTIME) && + (md->type == EFI_BOOT_SERVICES_CODE || + md->type == EFI_BOOT_SERVICES_DATA)) + continue; + + memcpy(new_md, md, efi.memmap.desc_size); + new_md += efi.memmap.desc_size; + } + + memunmap(new); + + if (efi_memmap_install(new_phys, num_entries)) { + pr_err("Could not install new EFI memmap\n"); + return; + } } /* diff --git a/drivers/firmware/efi/efi.c b/drivers/firmware/efi/efi.c index d4886fd50c16..dfe07316cae5 100644 --- a/drivers/firmware/efi/efi.c +++ b/drivers/firmware/efi/efi.c @@ -27,6 +27,7 @@ #include #include #include +#include #include @@ -396,6 +397,35 @@ u64 __init efi_mem_desc_end(efi_memory_desc_t *md) return end; } +void __init __weak efi_arch_mem_reserve(phys_addr_t addr, u64 size) {} + +/** + * efi_mem_reserve - Reserve an EFI memory region + * @addr: Physical address to reserve + * @size: Size of reservation + * + * Mark a region as reserved from general kernel allocation and + * prevent it being released by efi_free_boot_services(). + * + * This function should be called drivers once they've parsed EFI + * configuration tables to figure out where their data lives, e.g. + * efi_esrt_init(). + */ +void __init efi_mem_reserve(phys_addr_t addr, u64 size) +{ + if (!memblock_is_region_reserved(addr, size)) + memblock_reserve(addr, size); + + /* + * Some architectures (x86) reserve all boot services ranges + * until efi_free_boot_services() because of buggy firmware + * implementations. This means the above memblock_reserve() is + * superfluous on x86 and instead what it needs to do is + * ensure the @start, @size is not freed. + */ + efi_arch_mem_reserve(addr, size); +} + static __initdata efi_config_table_type_t common_tables[] = { {ACPI_20_TABLE_GUID, "ACPI 2.0", &efi.acpi20}, {ACPI_TABLE_GUID, "ACPI", &efi.acpi}, diff --git a/include/linux/efi.h b/include/linux/efi.h index 987c18f6fcae..3fe4f3c47834 100644 --- a/include/linux/efi.h +++ b/include/linux/efi.h @@ -944,6 +944,7 @@ extern u64 efi_mem_attribute (unsigned long phys_addr, unsigned long size); extern int __init efi_uart_console_only (void); extern u64 efi_mem_desc_end(efi_memory_desc_t *md); extern int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md); +extern void efi_mem_reserve(phys_addr_t addr, u64 size); extern void efi_initialize_iomem_resources(struct resource *code_resource, struct resource *data_resource, struct resource *bss_resource); extern void efi_reserve_boot_services(void); -- cgit 1.4.1 From 31ce8cc68180803aa481c0c1daac29d8eaceca9d Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Tue, 1 Mar 2016 23:02:56 +0000 Subject: efi/runtime-map: Use efi.memmap directly instead of a copy Now that efi.memmap is available all of the time there's no need to allocate and build a separate copy of the EFI memory map. Furthermore, efi.memmap contains boot services regions but only those regions that have been reserved via efi_mem_reserve(). Using efi.memmap allows us to pass boot services across kexec reboot so that the ESRT and BGRT drivers will now work. Tested-by: Dave Young [kexec/kdump] Tested-by: Ard Biesheuvel [arm] Acked-by: Ard Biesheuvel Cc: Leif Lindholm Cc: Peter Jones Cc: Borislav Petkov Cc: Mark Rutland Signed-off-by: Matt Fleming --- arch/x86/platform/efi/efi.c | 40 -------------------------------------- drivers/firmware/efi/runtime-map.c | 35 +++++++++++++-------------------- include/linux/efi.h | 4 ---- 3 files changed, 13 insertions(+), 66 deletions(-) (limited to 'arch') diff --git a/arch/x86/platform/efi/efi.c b/arch/x86/platform/efi/efi.c index 33996987ac70..342cebd1e17c 100644 --- a/arch/x86/platform/efi/efi.c +++ b/arch/x86/platform/efi/efi.c @@ -592,42 +592,6 @@ static void __init get_systab_virt_addr(efi_memory_desc_t *md) } } -static void __init save_runtime_map(void) -{ -#ifdef CONFIG_KEXEC_CORE - unsigned long desc_size; - efi_memory_desc_t *md; - void *tmp, *q = NULL; - int count = 0; - - if (efi_enabled(EFI_OLD_MEMMAP)) - return; - - desc_size = efi.memmap.desc_size; - - for_each_efi_memory_desc(md) { - if (!(md->attribute & EFI_MEMORY_RUNTIME) || - (md->type == EFI_BOOT_SERVICES_CODE) || - (md->type == EFI_BOOT_SERVICES_DATA)) - continue; - tmp = krealloc(q, (count + 1) * desc_size, GFP_KERNEL); - if (!tmp) - goto out; - q = tmp; - - memcpy(q + count * desc_size, md, desc_size); - count++; - } - - efi_runtime_map_setup(q, count, desc_size); - return; - -out: - kfree(q); - pr_err("Error saving runtime map, efi runtime on kexec non-functional!!\n"); -#endif -} - static void *realloc_pages(void *old_memmap, int old_shift) { void *ret; @@ -840,8 +804,6 @@ static void __init kexec_enter_virtual_mode(void) return; } - save_runtime_map(); - BUG_ON(!efi.systab); num_pages = ALIGN(efi.memmap.nr_map * efi.memmap.desc_size, PAGE_SIZE); @@ -934,8 +896,6 @@ static void __init __efi_enter_virtual_mode(void) return; } - save_runtime_map(); - BUG_ON(!efi.systab); if (efi_setup_page_tables(pa, 1 << pg_shift)) { diff --git a/drivers/firmware/efi/runtime-map.c b/drivers/firmware/efi/runtime-map.c index 5c55227a34c8..8e64b77aeac9 100644 --- a/drivers/firmware/efi/runtime-map.c +++ b/drivers/firmware/efi/runtime-map.c @@ -14,10 +14,6 @@ #include -static void *efi_runtime_map; -static int nr_efi_runtime_map; -static u32 efi_memdesc_size; - struct efi_runtime_map_entry { efi_memory_desc_t md; struct kobject kobj; /* kobject for each entry */ @@ -106,7 +102,8 @@ static struct kobj_type __refdata map_ktype = { static struct kset *map_kset; static struct efi_runtime_map_entry * -add_sysfs_runtime_map_entry(struct kobject *kobj, int nr) +add_sysfs_runtime_map_entry(struct kobject *kobj, int nr, + efi_memory_desc_t *md) { int ret; struct efi_runtime_map_entry *entry; @@ -124,8 +121,7 @@ add_sysfs_runtime_map_entry(struct kobject *kobj, int nr) return ERR_PTR(-ENOMEM); } - memcpy(&entry->md, efi_runtime_map + nr * efi_memdesc_size, - sizeof(efi_memory_desc_t)); + memcpy(&entry->md, md, sizeof(efi_memory_desc_t)); kobject_init(&entry->kobj, &map_ktype); entry->kobj.kset = map_kset; @@ -142,12 +138,12 @@ add_sysfs_runtime_map_entry(struct kobject *kobj, int nr) int efi_get_runtime_map_size(void) { - return nr_efi_runtime_map * efi_memdesc_size; + return efi.memmap.nr_map * efi.memmap.desc_size; } int efi_get_runtime_map_desc_size(void) { - return efi_memdesc_size; + return efi.memmap.desc_size; } int efi_runtime_map_copy(void *buf, size_t bufsz) @@ -157,38 +153,33 @@ int efi_runtime_map_copy(void *buf, size_t bufsz) if (sz > bufsz) sz = bufsz; - memcpy(buf, efi_runtime_map, sz); + memcpy(buf, efi.memmap.map, sz); return 0; } -void efi_runtime_map_setup(void *map, int nr_entries, u32 desc_size) -{ - efi_runtime_map = map; - nr_efi_runtime_map = nr_entries; - efi_memdesc_size = desc_size; -} - int __init efi_runtime_map_init(struct kobject *efi_kobj) { int i, j, ret = 0; struct efi_runtime_map_entry *entry; + efi_memory_desc_t *md; - if (!efi_runtime_map) + if (!efi_enabled(EFI_MEMMAP)) return 0; - map_entries = kzalloc(nr_efi_runtime_map * sizeof(entry), GFP_KERNEL); + map_entries = kzalloc(efi.memmap.nr_map * sizeof(entry), GFP_KERNEL); if (!map_entries) { ret = -ENOMEM; goto out; } - for (i = 0; i < nr_efi_runtime_map; i++) { - entry = add_sysfs_runtime_map_entry(efi_kobj, i); + i = 0; + for_each_efi_memory_desc(md) { + entry = add_sysfs_runtime_map_entry(efi_kobj, i, md); if (IS_ERR(entry)) { ret = PTR_ERR(entry); goto out_add_entry; } - *(map_entries + i) = entry; + *(map_entries + i++) = entry; } return 0; diff --git a/include/linux/efi.h b/include/linux/efi.h index 3fe4f3c47834..d8b555db81c7 100644 --- a/include/linux/efi.h +++ b/include/linux/efi.h @@ -1357,7 +1357,6 @@ extern int efi_capsule_update(efi_capsule_header_t *capsule, #ifdef CONFIG_EFI_RUNTIME_MAP int efi_runtime_map_init(struct kobject *); -void efi_runtime_map_setup(void *, int, u32); int efi_get_runtime_map_size(void); int efi_get_runtime_map_desc_size(void); int efi_runtime_map_copy(void *buf, size_t bufsz); @@ -1367,9 +1366,6 @@ static inline int efi_runtime_map_init(struct kobject *kobj) return 0; } -static inline void -efi_runtime_map_setup(void *map, int nr_entries, u32 desc_size) {} - static inline int efi_get_runtime_map_size(void) { return 0; -- cgit 1.4.1 From 4bc9f92e64c81192dcca1c495354bcc7c3b43e7d Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Thu, 23 Jun 2016 11:36:32 +0100 Subject: x86/efi-bgrt: Use efi_mem_reserve() to avoid copying image data MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit efi_mem_reserve() allows us to permanently mark EFI boot services regions as reserved, which means we no longer need to copy the image data out and into a separate buffer. Leaving the data in the original boot services region has the added benefit that BGRT images can now be passed across kexec reboot. Reviewed-by: Josh Triplett Tested-by: Dave Young [kexec/kdump] Tested-by: Ard Biesheuvel [arm] Acked-by: Ard Biesheuvel Cc: Leif Lindholm Cc: Peter Jones Cc: Borislav Petkov Cc: Mark Rutland Cc: Josh Boyer Cc: Andy Lutomirski Cc: Môshe van der Sterre Signed-off-by: Matt Fleming --- arch/x86/platform/efi/efi-bgrt.c | 13 ++----------- 1 file changed, 2 insertions(+), 11 deletions(-) (limited to 'arch') diff --git a/arch/x86/platform/efi/efi-bgrt.c b/arch/x86/platform/efi/efi-bgrt.c index 6a2f5691b1ab..6aad870e8962 100644 --- a/arch/x86/platform/efi/efi-bgrt.c +++ b/arch/x86/platform/efi/efi-bgrt.c @@ -82,21 +82,12 @@ void __init efi_bgrt_init(void) } bgrt_image_size = bmp_header.size; - bgrt_image = kmalloc(bgrt_image_size, GFP_KERNEL | __GFP_NOWARN); + bgrt_image = memremap(bgrt_tab->image_address, bmp_header.size, MEMREMAP_WB); if (!bgrt_image) { - pr_notice("Ignoring BGRT: failed to allocate memory for image (wanted %zu bytes)\n", - bgrt_image_size); - return; - } - - image = memremap(bgrt_tab->image_address, bmp_header.size, MEMREMAP_WB); - if (!image) { pr_notice("Ignoring BGRT: failed to map image memory\n"); - kfree(bgrt_image); bgrt_image = NULL; return; } - memcpy(bgrt_image, image, bgrt_image_size); - memunmap(image); + efi_mem_reserve(bgrt_tab->image_address, bgrt_image_size); } -- cgit 1.4.1 From ac0e94b63e65f9c6d2f3c49107118e2228236db4 Mon Sep 17 00:00:00 2001 From: Colin Ian King Date: Wed, 20 Jul 2016 11:11:06 +0100 Subject: x86/efi: Initialize status to ensure garbage is not returned on small size Although very unlikey, if size is too small or zero, then we end up with status not being set and returning garbage. Instead, initializing status to EFI_INVALID_PARAMETER to indicate that size is invalid in the calls to setup_uga32 and setup_uga64. Signed-off-by: Colin Ian King Cc: "H. Peter Anvin" Cc: Thomas Gleixner Cc: Ingo Molnar Signed-off-by: Matt Fleming --- arch/x86/boot/compressed/eboot.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'arch') diff --git a/arch/x86/boot/compressed/eboot.c b/arch/x86/boot/compressed/eboot.c index ff574dad95cc..ec6d2ef12baf 100644 --- a/arch/x86/boot/compressed/eboot.c +++ b/arch/x86/boot/compressed/eboot.c @@ -578,7 +578,7 @@ setup_uga32(void **uga_handle, unsigned long size, u32 *width, u32 *height) efi_guid_t uga_proto = EFI_UGA_PROTOCOL_GUID; unsigned long nr_ugas; u32 *handles = (u32 *)uga_handle;; - efi_status_t status; + efi_status_t status = EFI_INVALID_PARAMETER; int i; first_uga = NULL; @@ -623,7 +623,7 @@ setup_uga64(void **uga_handle, unsigned long size, u32 *width, u32 *height) efi_guid_t uga_proto = EFI_UGA_PROTOCOL_GUID; unsigned long nr_ugas; u64 *handles = (u64 *)uga_handle;; - efi_status_t status; + efi_status_t status = EFI_INVALID_PARAMETER; int i; first_uga = NULL; -- cgit 1.4.1 From 0513fe1d28e45deb39159dbeedf0660c3f0effd2 Mon Sep 17 00:00:00 2001 From: Alex Thorlton Date: Fri, 5 Aug 2016 18:59:35 -0500 Subject: x86/efi: Map in physical addresses in efi_map_region_fixed This is a simple change to add in the physical mappings as well as the virtual mappings in efi_map_region_fixed. The motivation here is to get access to EFI runtime code that is only available via the 1:1 mappings on a kexec'd kernel. The added call is essentially the kexec analog of the first __map_region that Boris put in efi_map_region in commit d2f7cbe7b26a ("x86/efi: Runtime services virtual mapping"). Signed-off-by: Alex Thorlton Cc: Russ Anderson Cc: Dimitri Sivanich Cc: Mike Travis Cc: Thomas Gleixner Cc: Ingo Molnar Cc: "H. Peter Anvin" Cc: Dave Young Cc: Borislav Petkov Signed-off-by: Matt Fleming --- arch/x86/platform/efi/efi_64.c | 1 + 1 file changed, 1 insertion(+) (limited to 'arch') diff --git a/arch/x86/platform/efi/efi_64.c b/arch/x86/platform/efi/efi_64.c index 45434ea345e9..e1ca71259468 100644 --- a/arch/x86/platform/efi/efi_64.c +++ b/arch/x86/platform/efi/efi_64.c @@ -339,6 +339,7 @@ void __init efi_map_region(efi_memory_desc_t *md) */ void __init efi_map_region_fixed(efi_memory_desc_t *md) { + __map_region(md, md->phys_addr); __map_region(md, md->virt_addr); } -- cgit 1.4.1 From 15cf7cae087a2eaf5e1feeef2bbba1b5a94c7639 Mon Sep 17 00:00:00 2001 From: Lukas Wunner Date: Mon, 15 Aug 2016 13:52:34 +0200 Subject: x86/efi: Remove unused find_bits() function Left behind by commit fc37206427ce ("efi/libstub: Move Graphics Output Protocol handling to generic code"). Signed-off-by: Lukas Wunner Cc: Ard Biesheuvel Signed-off-by: Matt Fleming --- arch/x86/boot/compressed/eboot.c | 23 ----------------------- 1 file changed, 23 deletions(-) (limited to 'arch') diff --git a/arch/x86/boot/compressed/eboot.c b/arch/x86/boot/compressed/eboot.c index ec6d2ef12baf..f7fc85bf8221 100644 --- a/arch/x86/boot/compressed/eboot.c +++ b/arch/x86/boot/compressed/eboot.c @@ -286,29 +286,6 @@ void efi_char16_printk(efi_system_table_t *table, efi_char16_t *str) } } -static void find_bits(unsigned long mask, u8 *pos, u8 *size) -{ - u8 first, len; - - first = 0; - len = 0; - - if (mask) { - while (!(mask & 0x1)) { - mask = mask >> 1; - first++; - } - - while (mask & 0x1) { - mask = mask >> 1; - len++; - } - } - - *pos = first; - *size = len; -} - static efi_status_t __setup_efi_pci32(efi_pci_io_protocol_32 *pci, struct pci_setup_rom **__rom) { -- cgit 1.4.1 From 3dad6f7f6975387f53f1a772f29f54335563d93d Mon Sep 17 00:00:00 2001 From: Ricardo Neri Date: Tue, 16 Aug 2016 17:32:31 -0700 Subject: x86/efi: Defer efi_esrt_init until after memblock_x86_fill Commit 7b02d53e7852 ("efi: Allow drivers to reserve boot services forever") introduced a new efi_mem_reserve to reserve the boot services memory regions forever. This reservation involves allocating a new EFI memory range descriptor. However, allocation can only succeed if there is memory available for the allocation. Otherwise, error such as the following may occur: esrt: Reserving ESRT space from 0x000000003dd6a000 to 0x000000003dd6a010. Kernel panic - not syncing: ERROR: Failed to allocate 0x9f0 bytes below \ 0x0. CPU: 0 PID: 0 Comm: swapper Not tainted 4.7.0-rc5+ #503 0000000000000000 ffffffff81e03ce0 ffffffff8131dae8 ffffffff81bb6c50 ffffffff81e03d70 ffffffff81e03d60 ffffffff8111f4df 0000000000000018 ffffffff81e03d70 ffffffff81e03d08 00000000000009f0 00000000000009f0 Call Trace: [] dump_stack+0x4d/0x65 [] panic+0xc5/0x206 [] memblock_alloc_base+0x29/0x2e [] memblock_alloc+0xb/0xd [] efi_arch_mem_reserve+0xbc/0x134 [] efi_mem_reserve+0x2c/0x31 [] ? efi_mem_reserve+0x2c/0x31 [] efi_esrt_init+0x19e/0x1b4 [] efi_init+0x398/0x44a [] setup_arch+0x415/0xc30 [] start_kernel+0x5b/0x3ef [] x86_64_start_reservations+0x2f/0x31 [] x86_64_start_kernel+0xea/0xed ---[ end Kernel panic - not syncing: ERROR: Failed to allocate 0x9f0 bytes below 0x0. An inspection of the memblock configuration reveals that there is no memory available for the allocation: MEMBLOCK configuration: memory size = 0x0 reserved size = 0x4f339c0 memory.cnt = 0x1 memory[0x0] [0x00000000000000-0xffffffffffffffff], 0x0 bytes on node 0\ flags: 0x0 reserved.cnt = 0x4 reserved[0x0] [0x0000000008c000-0x0000000008c9bf], 0x9c0 bytes flags: 0x0 reserved[0x1] [0x0000000009f000-0x000000000fffff], 0x61000 bytes\ flags: 0x0 reserved[0x2] [0x00000002800000-0x0000000394bfff], 0x114c000 bytes\ flags: 0x0 reserved[0x3] [0x000000304e4000-0x00000034269fff], 0x3d86000 bytes\ flags: 0x0 This situation can be avoided if we call efi_esrt_init after memblock has memory regions for the allocation. Also, the EFI ESRT driver makes use of early_memremap'pings. Therfore, we do not want to defer efi_esrt_init for too long. We must call such function while calls to early_memremap are still valid. A good place to meet the two aforementioned conditions is right after memblock_x86_fill, grouped with other EFI-related functions. Reported-by: Scott Lawson Signed-off-by: Ricardo Neri Cc: Ard Biesheuvel Cc: Peter Jones Signed-off-by: Matt Fleming --- arch/x86/kernel/setup.c | 1 + arch/x86/platform/efi/efi.c | 2 -- 2 files changed, 1 insertion(+), 2 deletions(-) (limited to 'arch') diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index 4fd69e532c15..528b8eb24a04 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -1101,6 +1101,7 @@ void __init setup_arch(char **cmdline_p) if (efi_enabled(EFI_MEMMAP)) { efi_fake_memmap(); efi_find_mirror(); + efi_esrt_init(); /* * The EFI specification says that boot service code won't be diff --git a/arch/x86/platform/efi/efi.c b/arch/x86/platform/efi/efi.c index 342cebd1e17c..0955c70897ae 100644 --- a/arch/x86/platform/efi/efi.c +++ b/arch/x86/platform/efi/efi.c @@ -474,8 +474,6 @@ void __init efi_init(void) if (efi_enabled(EFI_DBG)) efi_print_memmap(); - - efi_esrt_init(); } void __init efi_late_init(void) -- cgit 1.4.1 From 20ebc15e6c8f9772804fa10110bf074a7b1d25fa Mon Sep 17 00:00:00 2001 From: Markus Elfring Date: Thu, 25 Aug 2016 11:34:03 +0200 Subject: x86/efi: Use kmalloc_array() in efi_call_phys_prolog() * A multiplication for the size determination of a memory allocation indicated that an array data structure should be processed. Thus reuse the corresponding function "kmalloc_array". This issue was detected by using the Coccinelle software. * Replace the specification of a data type by a pointer dereference to make the corresponding size determination a bit safer according to the Linux coding style convention. Signed-off-by: Markus Elfring Reviewed-by: Paolo Bonzini Cc: "H. Peter Anvin" Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Julia Lawall Signed-off-by: Matt Fleming --- arch/x86/platform/efi/efi_64.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch') diff --git a/arch/x86/platform/efi/efi_64.c b/arch/x86/platform/efi/efi_64.c index e1ca71259468..d65cdadaa6b6 100644 --- a/arch/x86/platform/efi/efi_64.c +++ b/arch/x86/platform/efi/efi_64.c @@ -85,7 +85,7 @@ pgd_t * __init efi_call_phys_prolog(void) early_code_mapping_set_exec(1); n_pgds = DIV_ROUND_UP((max_pfn << PAGE_SHIFT), PGDIR_SIZE); - save_pgd = kmalloc(n_pgds * sizeof(pgd_t), GFP_KERNEL); + save_pgd = kmalloc_array(n_pgds, sizeof(*save_pgd), GFP_KERNEL); for (pgd = 0; pgd < n_pgds; pgd++) { save_pgd[pgd] = *pgd_offset_k(pgd * PGDIR_SIZE); -- cgit 1.4.1 From 27571616385af9c2d6a3e570b06baf86f5aa04b1 Mon Sep 17 00:00:00 2001 From: Lukas Wunner Date: Tue, 6 Sep 2016 08:05:32 +0200 Subject: x86/efi: Optimize away setup_gop32/64 if unused Commit 2c23b73c2d02 ("x86/efi: Prepare GOP handling code for reuse as generic code") introduced an efi_is_64bit() macro to x86 which previously only existed for arm arches. The macro is used to choose between the 64 bit or 32 bit code path in gop.c at runtime. However the code path that's going to be taken is known at compile time when compiling for x86_32 or for x86_64 with mixed mode disabled. Amend the macro to eliminate the unused code path in those cases. Size of gop.o text section: CONFIG_X86_32: 1758 before, 1299 after CONFIG_X86_64 && !CONFIG_EFI_MIXED: 2201 before, 1406 after CONFIG_X86_64 && CONFIG_EFI_MIXED: 2201 before and after Signed-off-by: Lukas Wunner Reviewed-by: Ard Biesheuvel Signed-off-by: Matt Fleming --- arch/x86/include/asm/efi.h | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) (limited to 'arch') diff --git a/arch/x86/include/asm/efi.h b/arch/x86/include/asm/efi.h index 4630e2bfa8fb..f14655e7726a 100644 --- a/arch/x86/include/asm/efi.h +++ b/arch/x86/include/asm/efi.h @@ -206,14 +206,23 @@ struct efi_config { __pure const struct efi_config *__efi_early(void); +static inline bool efi_is_64bit(void) +{ + if (!IS_ENABLED(CONFIG_X86_64)) + return false; + + if (!IS_ENABLED(CONFIG_EFI_MIXED)) + return true; + + return __efi_early()->is64; +} + #define efi_call_early(f, ...) \ __efi_early()->call(__efi_early()->f, __VA_ARGS__); #define __efi_call_early(f, ...) \ __efi_early()->call((unsigned long)f, __VA_ARGS__); -#define efi_is_64bit() __efi_early()->is64 - extern bool efi_reboot_required(void); #else -- cgit 1.4.1 From 0a637ee61247bd4bed9b2a07568ef7a1cfc76187 Mon Sep 17 00:00:00 2001 From: Lukas Wunner Date: Mon, 22 Aug 2016 12:01:21 +0200 Subject: x86/efi: Allow invocation of arbitrary boot services We currently allow invocation of 8 boot services with efi_call_early(). Not included are LocateHandleBuffer and LocateProtocol in particular. For graphics output or to retrieve PCI ROMs and Apple device properties, we're thus forced to use the LocateHandle + AllocatePool + LocateHandle combo, which is cumbersome and needs more code. The ARM folks allow invocation of the full set of boot services but are restricted to our 8 boot services in functions shared across arches. Thus, rather than adding just LocateHandleBuffer and LocateProtocol to struct efi_config, let's rework efi_call_early() to allow invocation of arbitrary boot services by selecting the 64 bit vs 32 bit code path in the macro itself. When compiling for 32 bit or for 64 bit without mixed mode, the unused code path is optimized away and the binary code is the same as before. But on 64 bit with mixed mode enabled, this commit adds one compare instruction to each invocation of a boot service and, depending on the code path selected, two jump instructions. (Most of the time gcc arranges the jumps in the 32 bit code path.) The result is a minuscule performance penalty and the binary code becomes slightly larger and more difficult to read when disassembled. This isn't a hot path, so these drawbacks are arguably outweighed by the attainable simplification of the C code. We have some overhead anyway for thunking or conversion between calling conventions. The 8 boot services can consequently be removed from struct efi_config. No functional change intended (for now). Example -- invocation of free_pool before (64 bit code path): 0x2d4 movq %ds:efi_early, %rdx ; efi_early 0x2db movq %ss:arg_0-0x20(%rsp), %rsi 0x2e0 xorl %eax, %eax 0x2e2 movq %ds:0x28(%rdx), %rdi ; efi_early->free_pool 0x2e6 callq *%ds:0x58(%rdx) ; efi_early->call() Example -- invocation of free_pool after (64 / 32 bit mixed code path): 0x0dc movq %ds:efi_early, %rax ; efi_early 0x0e3 cmpb $0, %ds:0x28(%rax) ; !efi_early->is64 ? 0x0e7 movq %ds:0x20(%rax), %rdx ; efi_early->call() 0x0eb movq %ds:0x10(%rax), %rax ; efi_early->boot_services 0x0ef je $0x150 0x0f1 movq %ds:0x48(%rax), %rdi ; free_pool (64 bit) 0x0f5 xorl %eax, %eax 0x0f7 callq *%rdx ... 0x150 movl %ds:0x30(%rax), %edi ; free_pool (32 bit) 0x153 jmp $0x0f5 Size of eboot.o text section: CONFIG_X86_32: 6464 before, 6318 after CONFIG_X86_64 && !CONFIG_EFI_MIXED: 7670 before, 7573 after CONFIG_X86_64 && CONFIG_EFI_MIXED: 7670 before, 8319 after Signed-off-by: Lukas Wunner Signed-off-by: Matt Fleming --- arch/x86/boot/compressed/eboot.c | 13 +------------ arch/x86/boot/compressed/head_32.S | 6 +++--- arch/x86/boot/compressed/head_64.S | 8 ++++---- arch/x86/include/asm/efi.h | 15 ++++++--------- 4 files changed, 14 insertions(+), 28 deletions(-) (limited to 'arch') diff --git a/arch/x86/boot/compressed/eboot.c b/arch/x86/boot/compressed/eboot.c index f7fc85bf8221..447a6a2df5ae 100644 --- a/arch/x86/boot/compressed/eboot.c +++ b/arch/x86/boot/compressed/eboot.c @@ -29,22 +29,11 @@ __pure const struct efi_config *__efi_early(void) static void setup_boot_services##bits(struct efi_config *c) \ { \ efi_system_table_##bits##_t *table; \ - efi_boot_services_##bits##_t *bt; \ \ table = (typeof(table))sys_table; \ \ + c->boot_services = table->boottime; \ c->text_output = table->con_out; \ - \ - bt = (typeof(bt))(unsigned long)(table->boottime); \ - \ - c->allocate_pool = bt->allocate_pool; \ - c->allocate_pages = bt->allocate_pages; \ - c->get_memory_map = bt->get_memory_map; \ - c->free_pool = bt->free_pool; \ - c->free_pages = bt->free_pages; \ - c->locate_handle = bt->locate_handle; \ - c->handle_protocol = bt->handle_protocol; \ - c->exit_boot_services = bt->exit_boot_services; \ } BOOT_SERVICES(32); BOOT_SERVICES(64); diff --git a/arch/x86/boot/compressed/head_32.S b/arch/x86/boot/compressed/head_32.S index 1038524270e7..fd0b6a272dd5 100644 --- a/arch/x86/boot/compressed/head_32.S +++ b/arch/x86/boot/compressed/head_32.S @@ -82,7 +82,7 @@ ENTRY(efi_pe_entry) /* Relocate efi_config->call() */ leal efi32_config(%esi), %eax - add %esi, 88(%eax) + add %esi, 32(%eax) pushl %eax call make_boot_params @@ -108,7 +108,7 @@ ENTRY(efi32_stub_entry) /* Relocate efi_config->call() */ leal efi32_config(%esi), %eax - add %esi, 88(%eax) + add %esi, 32(%eax) pushl %eax 2: call efi_main @@ -264,7 +264,7 @@ relocated: #ifdef CONFIG_EFI_STUB .data efi32_config: - .fill 11,8,0 + .fill 4,8,0 .long efi_call_phys .long 0 .byte 0 diff --git a/arch/x86/boot/compressed/head_64.S b/arch/x86/boot/compressed/head_64.S index 0d80a7ad65cd..efdfba21a5b2 100644 --- a/arch/x86/boot/compressed/head_64.S +++ b/arch/x86/boot/compressed/head_64.S @@ -265,7 +265,7 @@ ENTRY(efi_pe_entry) /* * Relocate efi_config->call(). */ - addq %rbp, efi64_config+88(%rip) + addq %rbp, efi64_config+32(%rip) movq %rax, %rdi call make_boot_params @@ -285,7 +285,7 @@ handover_entry: * Relocate efi_config->call(). */ movq efi_config(%rip), %rax - addq %rbp, 88(%rax) + addq %rbp, 32(%rax) 2: movq efi_config(%rip), %rdi call efi_main @@ -457,14 +457,14 @@ efi_config: #ifdef CONFIG_EFI_MIXED .global efi32_config efi32_config: - .fill 11,8,0 + .fill 4,8,0 .quad efi64_thunk .byte 0 #endif .global efi64_config efi64_config: - .fill 11,8,0 + .fill 4,8,0 .quad efi_call .byte 1 #endif /* CONFIG_EFI_STUB */ diff --git a/arch/x86/include/asm/efi.h b/arch/x86/include/asm/efi.h index f14655e7726a..389d700b961e 100644 --- a/arch/x86/include/asm/efi.h +++ b/arch/x86/include/asm/efi.h @@ -191,14 +191,7 @@ static inline efi_status_t efi_thunk_set_virtual_address_map( struct efi_config { u64 image_handle; u64 table; - u64 allocate_pool; - u64 allocate_pages; - u64 get_memory_map; - u64 free_pool; - u64 free_pages; - u64 locate_handle; - u64 handle_protocol; - u64 exit_boot_services; + u64 boot_services; u64 text_output; efi_status_t (*call)(unsigned long, ...); bool is64; @@ -218,7 +211,11 @@ static inline bool efi_is_64bit(void) } #define efi_call_early(f, ...) \ - __efi_early()->call(__efi_early()->f, __VA_ARGS__); + __efi_early()->call(efi_is_64bit() ? \ + ((efi_boot_services_64_t *)(unsigned long) \ + __efi_early()->boot_services)->f : \ + ((efi_boot_services_32_t *)(unsigned long) \ + __efi_early()->boot_services)->f, __VA_ARGS__) #define __efi_call_early(f, ...) \ __efi_early()->call((unsigned long)f, __VA_ARGS__); -- cgit 1.4.1 From 92dc33501bfba74655dbf3ec63ea82d040fd6d58 Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Fri, 16 Sep 2016 15:12:47 +0100 Subject: x86/efi: Round EFI memmap reservations to EFI_PAGE_SIZE Mike Galbraith reported that his machine started rebooting during boot after, commit 8e80632fb23f ("efi/esrt: Use efi_mem_reserve() and avoid a kmalloc()") The ESRT table on his machine is 56 bytes and at no point in the efi_arch_mem_reserve() call path is that size rounded up to EFI_PAGE_SIZE, nor is the start address on an EFI_PAGE_SIZE boundary. Since the EFI memory map only deals with whole pages, inserting an EFI memory region with 56 bytes results in a new entry covering zero pages, and completely screws up the calculations for the old regions that were trimmed. Round all sizes upwards, and start addresses downwards, to the nearest EFI_PAGE_SIZE boundary. Additionally, efi_memmap_insert() expects the mem::range::end value to be one less than the end address for the region. Reported-by: Mike Galbraith Reported-by: Mike Krinkin Tested-by: Mike Krinkin Cc: Peter Jones Cc: Ard Biesheuvel Cc: Mark Rutland Cc: Taku Izumi Signed-off-by: Matt Fleming --- arch/x86/platform/efi/quirks.c | 6 +++++- drivers/firmware/efi/memmap.c | 11 +++++++++++ 2 files changed, 16 insertions(+), 1 deletion(-) (limited to 'arch') diff --git a/arch/x86/platform/efi/quirks.c b/arch/x86/platform/efi/quirks.c index f14b7a9da24b..10aca63a50d7 100644 --- a/arch/x86/platform/efi/quirks.c +++ b/arch/x86/platform/efi/quirks.c @@ -201,8 +201,12 @@ void __init efi_arch_mem_reserve(phys_addr_t addr, u64 size) return; } + size += addr % EFI_PAGE_SIZE; + size = round_up(size, EFI_PAGE_SIZE); + addr = round_down(addr, EFI_PAGE_SIZE); + mr.range.start = addr; - mr.range.end = addr + size; + mr.range.end = addr + size - 1; mr.attribute = md.attribute | EFI_MEMORY_RUNTIME; num_entries = efi_memmap_split_count(&md, &mr.range); diff --git a/drivers/firmware/efi/memmap.c b/drivers/firmware/efi/memmap.c index cd96086fd851..f03ddecd232b 100644 --- a/drivers/firmware/efi/memmap.c +++ b/drivers/firmware/efi/memmap.c @@ -225,6 +225,17 @@ void __init efi_memmap_insert(struct efi_memory_map *old_memmap, void *buf, m_end = mem->range.end; m_attr = mem->attribute; + /* + * The EFI memory map deals with regions in EFI_PAGE_SIZE + * units. Ensure that the region described by 'mem' is aligned + * correctly. + */ + if (!IS_ALIGNED(m_start, EFI_PAGE_SIZE) || + !IS_ALIGNED(m_end + 1, EFI_PAGE_SIZE)) { + WARN_ON(1); + return; + } + for (old = old_memmap->map, new = buf; old < old_memmap->map_end; old += old_memmap->desc_size, new += old_memmap->desc_size) { -- cgit 1.4.1