In systems based on [U]EFI-conformant firmware, runtime services provide
a standardised way for the kernel to update firmware environment
variables. This is used for example by efibootmgr to update which image
should be loaded on next boot.
This patchset implements basic support for UEFI runtime services on ARM
platforms, as well as the basic underlying EFI support. It also defines
a mechanism by which the required information is passed from the
bootloader to the kernel via FDT entries.
This patchset depends on the previously submitted early_ioremap() and
generic UEFI config table scanning patchsets.
Changes from v1:
- Updated FDT bindings, based on feedback.
- Use common config table scanning and address lookup code.
- Add dependency on !CPU_BIG_ENDIAN (for now).
- Add proper efi_enabled() facility.
- Documentation updates.
Leif Lindholm (3):
Documentation: arm: add UEFI support documentation
arm: Add [U]EFI runtime services support
init: efi: arm: enable (U)EFI runtime services on arm
Documentation/arm/00-INDEX | 3 +
Documentation/arm/uefi.txt | 47 +++++
arch/arm/Kconfig | 15 ++
arch/arm/include/asm/efi.h | 22 ++
arch/arm/kernel/Makefile | 2 +
arch/arm/kernel/efi.c | 485 ++++++++++++++++++++++++++++++++++++++++++++
arch/arm/kernel/efi_phys.S | 59 ++++++
arch/arm/kernel/setup.c | 6 +
include/linux/efi.h | 2 +-
init/main.c | 4 +
10 files changed, 644 insertions(+), 1 deletion(-)
create mode 100644 Documentation/arm/uefi.txt
create mode 100644 arch/arm/include/asm/efi.h
create mode 100644 arch/arm/kernel/efi.c
create mode 100644 arch/arm/kernel/efi_phys.S
--
1.7.10.4
This patch provides documentation of the [U]EFI runtime services and
configuration features.
Cc: [email protected]
Signed-off-by: Leif Lindholm <[email protected]>
Acked-by: Rob Landley <[email protected]>
---
Documentation/arm/00-INDEX | 3 +++
Documentation/arm/uefi.txt | 47 ++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 50 insertions(+)
create mode 100644 Documentation/arm/uefi.txt
diff --git a/Documentation/arm/00-INDEX b/Documentation/arm/00-INDEX
index 4978456..87e01d1 100644
--- a/Documentation/arm/00-INDEX
+++ b/Documentation/arm/00-INDEX
@@ -36,3 +36,6 @@ nwfpe/
- NWFPE floating point emulator documentation
swp_emulation
- SWP/SWPB emulation handler/logging description
+
+uefi.txt
+ - [U]EFI configuration and runtime services documentation
diff --git a/Documentation/arm/uefi.txt b/Documentation/arm/uefi.txt
new file mode 100644
index 0000000..e6e4d41
--- /dev/null
+++ b/Documentation/arm/uefi.txt
@@ -0,0 +1,47 @@
+UEFI, the Unified Extensible Firmware Interface is a speifcication
+governing the behaviours of compatible firmware interfaces. It is
+maintained by the UEFI Forum - http://www.uefi.org/.
+
+Since UEFI is an evolution of its predecessor 'EFI', the terms EFI and
+UEFI are used somewhat interchangeably in this document and associated
+source code.
+
+The implementation depends on receiving the UEFI runtime memory map and a
+pointer to the System Table in a Flattened Device Tree - so is only available
+with CONFIG_OF.
+
+It parses the FDT /chosen node for the following parameters:
+- 'linux,efi-system-table':
+ Physical address of the system table. (required)
+ 64-bit value since an ARMv7 plattform may support LPAE, and to facilitate
+ code sharing with arm64. Top 32 bits will be ignored, since UEFI specification
+ mandates a 1:1 mapping of all RAM.
+- 'linux,efi-mmap':
+ The EFI memory map as an embedded property. (required)
+ An array of type EFI_MEMORY_DESCRIPTOR as described by the UEFI
+ specification, current version described in Linux by efi_memory_desc_t.
+ The memory map is represented in little-endian, not DT, byte order.
+ This map needs to contain at least the regions to be preserved for runtime
+ services, but would normally just be the map retreieved by calling UEFI
+ GetMemoryMap() immediately before ExitBootServices().
+- 'linux,efi-mmap-desc-size':
+ Size of each descriptor in the memory map. (override default)
+- 'linux,efi-mmap-desc-ver':
+ Memory descriptor format version. (override default)
+
+It also depends on early_memremap() to parse the UEFI configuration tables.
+
+For actually enabling [U]EFI support, enable:
+- CONFIG_EFI=y
+- CONFIG_EFI_VARS=y or m
+
+After the kernel has mapped the required regions into its address space,
+a SetVirtualAddressMap() call is made into UEFI in order to update
+relocations. This call must be performed with all the code in a 1:1
+mapping. This implementation achieves this by temporarily disabling the
+MMU for the duration of this call. This can only be done safely:
+- before secondary CPUs are brought online.
+- after early_initcalls have completed, since it uses setup_mm_for_reboot().
+
+For verbose debug messages, specify 'uefi_debug' on the kernel command
+line.
--
1.7.10.4
Since the efi_set_virtual_address_map call has strict init ordering
requirements, add an explicit hook in the required place.
Signed-off-by: Leif Lindholm <[email protected]>
---
init/main.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/init/main.c b/init/main.c
index af310af..ec6d76e 100644
--- a/init/main.c
+++ b/init/main.c
@@ -875,6 +875,10 @@ static noinline void __init kernel_init_freeable(void)
smp_prepare_cpus(setup_max_cpus);
do_pre_smp_initcalls();
+
+ if (IS_ENABLED(CONFIG_ARM) && efi_enabled(EFI_BOOT))
+ efi_enter_virtual_mode();
+
lockup_detector_init();
smp_init();
--
1.7.10.4
This patch implements basic support for UEFI runtime services in the
ARM architecture - a requirement for using efibootmgr to read and update
the system boot configuration.
It uses the generic configuration table scanning to populate ACPI and
SMBIOS pointers.
Signed-off-by: Leif Lindholm <[email protected]>
---
arch/arm/Kconfig | 15 ++
arch/arm/include/asm/efi.h | 22 ++
arch/arm/kernel/Makefile | 2 +
arch/arm/kernel/efi.c | 485 ++++++++++++++++++++++++++++++++++++++++++++
arch/arm/kernel/efi_phys.S | 59 ++++++
arch/arm/kernel/setup.c | 6 +
include/linux/efi.h | 2 +-
7 files changed, 590 insertions(+), 1 deletion(-)
create mode 100644 arch/arm/include/asm/efi.h
create mode 100644 arch/arm/kernel/efi.c
create mode 100644 arch/arm/kernel/efi_phys.S
diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig
index 5916a90..317c75d 100644
--- a/arch/arm/Kconfig
+++ b/arch/arm/Kconfig
@@ -1855,6 +1855,19 @@ config EARLY_IOREMAP
It generates its map entries in kmap region (0xfff00000) before kmap
is initialized.
+config EFI
+ bool "UEFI runtime service support"
+ depends on OF && !CPU_BIG_ENDIAN
+ select UCS2_STRING
+ select EARLY_IOREMAP
+ ---help---
+ This enables the kernel to use UEFI runtime services that are
+ available (such as the UEFI variable services).
+
+ This option is only useful on systems that have UEFI firmware.
+ However, even with this option, the resultant kernel will
+ continue to boot on non-UEFI platforms.
+
config SECCOMP
bool
prompt "Enable seccomp to safely compute untrusted bytecode"
@@ -2267,6 +2280,8 @@ source "net/Kconfig"
source "drivers/Kconfig"
+source "drivers/firmware/Kconfig"
+
source "fs/Kconfig"
source "arch/arm/Kconfig.debug"
diff --git a/arch/arm/include/asm/efi.h b/arch/arm/include/asm/efi.h
new file mode 100644
index 0000000..aead94c
--- /dev/null
+++ b/arch/arm/include/asm/efi.h
@@ -0,0 +1,22 @@
+#ifndef _ASM_ARM_EFI_H
+#define _ASM_ARM_EFI_H
+
+#include <asm/mach/map.h>
+
+extern int efi_memblock_arm_reserve_range(void);
+
+typedef efi_status_t efi_phys_call_t(u32 memory_map_size,
+ u32 descriptor_size,
+ u32 descriptor_version,
+ efi_memory_desc_t *dsc,
+ efi_set_virtual_address_map_t *f);
+
+extern efi_status_t efi_phys_call(u32, u32, u32, efi_memory_desc_t *,
+ efi_set_virtual_address_map_t *);
+
+#define efi_remap(cookie, size) __arm_ioremap((cookie), (size), MT_MEMORY)
+#define efi_ioremap(cookie, size) __arm_ioremap((cookie), (size), MT_DEVICE)
+#define efi_unmap(cookie) __arm_iounmap((cookie))
+#define efi_iounmap(cookie) __arm_iounmap((cookie))
+
+#endif /* _ASM_ARM_EFI_H */
diff --git a/arch/arm/kernel/Makefile b/arch/arm/kernel/Makefile
index 5140df5f..81b8865 100644
--- a/arch/arm/kernel/Makefile
+++ b/arch/arm/kernel/Makefile
@@ -96,4 +96,6 @@ obj-y += psci.o
obj-$(CONFIG_SMP) += psci_smp.o
endif
+obj-$(CONFIG_EFI) += efi.o efi_phys.o
+
extra-y := $(head-y) vmlinux.lds
diff --git a/arch/arm/kernel/efi.c b/arch/arm/kernel/efi.c
new file mode 100644
index 0000000..a7c2c8d
--- /dev/null
+++ b/arch/arm/kernel/efi.c
@@ -0,0 +1,485 @@
+/*
+ * Extensible Firmware Interface
+ *
+ * Based on Extensible Firmware Interface Specification version 2.3.1
+ *
+ * Copyright (C) 2013 Linaro Ltd.
+ *
+ */
+
+#include <linux/efi.h>
+#include <linux/export.h>
+#include <linux/memblock.h>
+#include <linux/of.h>
+#include <linux/of_fdt.h>
+#include <linux/sched.h>
+#include <linux/slab.h>
+
+#include <asm/cacheflush.h>
+#include <asm/efi.h>
+#include <asm/idmap.h>
+#include <asm/tlbflush.h>
+
+struct efi_memory_map memmap;
+
+static efi_runtime_services_t *runtime;
+
+static phys_addr_t efi_system_table;
+static phys_addr_t efi_boot_mmap;
+static u32 efi_boot_mmap_size;
+static u32 efi_mmap_desc_size;
+static u32 efi_mmap_desc_ver;
+
+static unsigned long arm_efi_facility;
+
+/*
+ * Default memory map descriptor information
+ */
+#define DESC_SIZE 48
+#define DESC_VER 1
+
+/*
+ * If you're planning to wire up a debugger and debug the UEFI side ...
+ */
+#undef KEEP_ALL_REGIONS
+
+/*
+ * If you need to (temporarily) support buggy firmware.
+ */
+#define KEEP_BOOT_SERVICES_REGIONS
+
+/*
+ * Returns 1 if 'facility' is enabled, 0 otherwise.
+ */
+int efi_enabled(int facility)
+{
+ return test_bit(facility, &arm_efi_facility) != 0;
+}
+EXPORT_SYMBOL(efi_enabled);
+
+static int uefi_debug __initdata;
+static int __init uefi_debug_setup(char *str)
+{
+ uefi_debug = 1;
+
+ return 0;
+}
+early_param("uefi_debug", uefi_debug_setup);
+
+static int __init fdt_find_efi_params(unsigned long node, const char *uname,
+ int depth, void *data)
+{
+ unsigned long len;
+ void *prop;
+
+ if (depth != 1 ||
+ (strcmp(uname, "chosen") != 0 && strcmp(uname, "chosen@0") != 0))
+ return 0;
+
+ pr_info("Getting EFI parameters from FDT.\n");
+
+ prop = of_get_flat_dt_prop(node, "linux,efi-system-table", &len);
+ if (!prop)
+ return 0;
+ efi_system_table = of_read_ulong(prop, len/4);
+
+ prop = of_get_flat_dt_prop(node, "linux,efi-mmap", &len);
+ if (!prop)
+ return 0;
+ efi_boot_mmap = (u32) prop;
+ efi_boot_mmap_size = len;
+
+ prop = of_get_flat_dt_prop(node, "linux,efi-mmap-desc-size", NULL);
+ if (prop)
+ efi_mmap_desc_size = of_read_ulong(prop, 1);
+ else
+ efi_mmap_desc_size = DESC_SIZE;
+
+ prop = of_get_flat_dt_prop(node, "linux,efi-mmap-desc-ver", NULL);
+ if (prop)
+ efi_mmap_desc_ver = of_read_ulong(prop, 1);
+ else
+ efi_mmap_desc_ver = DESC_VER;
+
+ if (uefi_debug) {
+ pr_info(" EFI system table @ 0x%08x\n",
+ (unsigned int) efi_system_table);
+ pr_info(" EFI mmap @ 0x%08x\n",
+ (unsigned int) efi_boot_mmap);
+ pr_info(" EFI mmap size = 0x%08x\n",
+ (unsigned int) efi_boot_mmap_size);
+ pr_info(" EFI mmap descriptor size = 0x%08x\n",
+ (unsigned int) efi_mmap_desc_size);
+ pr_info(" EFI mmap descriptor version = 0x%08x\n",
+ (unsigned int) efi_mmap_desc_ver);
+ }
+
+ return 1;
+}
+
+static int __init uefi_init(void)
+{
+ efi_char16_t *c16;
+ char vendor[100] = "unknown";
+ int i, retval;
+
+ efi.systab = early_ioremap(efi_system_table,
+ sizeof(efi_system_table_t));
+
+ /*
+ * Verify the EFI Table
+ */
+ if (efi.systab == NULL)
+ panic("Whoa! Can't find EFI system table.\n");
+ if (efi.systab->hdr.signature != EFI_SYSTEM_TABLE_SIGNATURE)
+ panic("Whoa! EFI system table signature incorrect\n");
+ if ((efi.systab->hdr.revision >> 16) == 0)
+ pr_warn("Warning: EFI system table version %d.%02d, expected 1.00 or greater\n",
+ efi.systab->hdr.revision >> 16,
+ efi.systab->hdr.revision & 0xffff);
+
+ /* Show what we know for posterity */
+ c16 = (efi_char16_t *)early_ioremap(efi.systab->fw_vendor,
+ sizeof(vendor));
+ if (c16) {
+ for (i = 0; i < (int) sizeof(vendor) - 1 && *c16; ++i)
+ vendor[i] = c16[i];
+ vendor[i] = '\0';
+ }
+
+ pr_info("EFI v%u.%.02u by %s\n",
+ efi.systab->hdr.revision >> 16,
+ efi.systab->hdr.revision & 0xffff, vendor);
+
+ retval = efi_config_init(NULL);
+ if (retval == 0)
+ set_bit(EFI_CONFIG_TABLES, &arm_efi_facility);
+
+ early_iounmap(c16, sizeof(vendor));
+ early_iounmap(efi.systab, sizeof(efi_system_table_t));
+
+ return retval;
+}
+
+static __init int is_discardable_region(efi_memory_desc_t *md)
+{
+#ifdef KEEP_ALL_REGIONS
+ return 0;
+#endif
+
+ if (md->attribute & EFI_MEMORY_RUNTIME)
+ return 0;
+
+ switch (md->type) {
+#ifdef KEEP_BOOT_SERVICES_REGIONS
+ case EFI_BOOT_SERVICES_CODE:
+ case EFI_BOOT_SERVICES_DATA:
+#endif
+ /* Keep tables around for any future kexec operations */
+ case EFI_ACPI_RECLAIM_MEMORY:
+ return 0;
+ }
+
+ return 1;
+}
+
+static __initdata struct {
+ u32 type;
+ const char *name;
+} memory_type_name_map[] = {
+ {EFI_RESERVED_TYPE, "EFI reserved"},
+ {EFI_LOADER_CODE, "EFI loader code"},
+ {EFI_LOADER_DATA, "EFI loader data"},
+ {EFI_BOOT_SERVICES_CODE, "EFI boot services code"},
+ {EFI_BOOT_SERVICES_DATA, "EFI boot services data"},
+ {EFI_RUNTIME_SERVICES_CODE, "EFI runtime services code"},
+ {EFI_RUNTIME_SERVICES_DATA, "EFI runtime services data"},
+ {EFI_CONVENTIONAL_MEMORY, "EFI conventional memory"},
+ {EFI_UNUSABLE_MEMORY, "EFI unusable memory"},
+ {EFI_ACPI_RECLAIM_MEMORY, "EFI ACPI reclaim memory"},
+ {EFI_ACPI_MEMORY_NVS, "EFI ACPI memory nvs"},
+ {EFI_MEMORY_MAPPED_IO, "EFI memory mapped I/O"},
+ {EFI_MEMORY_MAPPED_IO_PORT_SPACE, "EFI memory mapped I/O port space"},
+ {EFI_PAL_CODE, "EFI pal code"},
+ {EFI_MAX_MEMORY_TYPE, NULL},
+};
+
+static __init void remove_sections(phys_addr_t addr, unsigned long size)
+{
+ unsigned long section_offset;
+ unsigned long num_sections;
+
+ section_offset = addr - (addr & SECTION_MASK);
+ num_sections = size / SECTION_SIZE;
+ if (size % SECTION_SIZE)
+ num_sections++;
+
+ memblock_remove(addr - section_offset, num_sections * SECTION_SIZE);
+}
+
+static __init int remove_regions(void)
+{
+ efi_memory_desc_t *md;
+ int count = 0;
+ void *p;
+
+ memmap.phys_map = (void *) (u32) efi_boot_mmap;
+ memmap.desc_size = efi_mmap_desc_size;
+ memmap.desc_version = efi_mmap_desc_ver;
+ memmap.map_end = (void *) memmap.phys_map + efi_boot_mmap_size;
+ memmap.nr_map = 0;
+
+ if (uefi_debug)
+ pr_info("Processing EFI memory map:\n");
+
+ for (p = memmap.phys_map; p < memmap.map_end; p += memmap.desc_size) {
+ md = p;
+ if (is_discardable_region(md))
+ continue;
+
+ if (uefi_debug)
+ pr_info(" %8llu pages @ %016llx (%s)\n",
+ md->num_pages, md->phys_addr,
+ memory_type_name_map[md->type].name);
+
+ if (md->type != EFI_MEMORY_MAPPED_IO) {
+ remove_sections(md->phys_addr,
+ md->num_pages * PAGE_SIZE);
+ count++;
+ }
+ memmap.nr_map++;
+ }
+
+ if (uefi_debug)
+ pr_info("%d regions preserved.\n", memmap.nr_map);
+
+ return 0;
+}
+
+int __init efi_memblock_arm_reserve_range(void)
+{
+ if (!of_scan_flat_dt(fdt_find_efi_params, NULL))
+ return 0;
+
+ set_bit(EFI_BOOT, &arm_efi_facility);
+
+ uefi_init();
+
+ remove_regions();
+
+ return 0;
+}
+
+/*
+ * Disable instrrupts, enable idmap and disable caches.
+ */
+static void __init phys_call_prologue(void)
+{
+ local_irq_disable();
+
+ /* Take out a flat memory mapping. */
+ setup_mm_for_reboot();
+
+ /* Clean and invalidate caches */
+ flush_cache_all();
+
+ /* Turn off caching */
+ cpu_proc_fin();
+
+ /* Push out any further dirty data, and ensure cache is empty */
+ flush_cache_all();
+}
+
+/*
+ * Restore original memory map and re-enable interrupts.
+ */
+static void __init phys_call_epilogue(void)
+{
+ static struct mm_struct *mm = &init_mm;
+
+ /* Restore original memory mapping */
+ cpu_switch_mm(mm->pgd, mm);
+
+ /* Flush branch predictor and TLBs */
+ local_flush_bp_all();
+#ifdef CONFIG_CPU_HAS_ASID
+ local_flush_tlb_all();
+#endif
+
+ local_irq_enable();
+}
+
+/*
+ * Memory map was previously extracted from flattened device tree for
+ * reserving regions. Now we need to grab it from the unflattened tree
+ * in order to access it for remapping purposes.
+ */
+static void __init *get_runtime_mmap(void)
+{
+ struct device_node *node;
+ int len;
+ void *map;
+
+ node = of_find_node_by_path("/chosen");
+ if (node != NULL) {
+ map = (void *) of_get_property(node, "linux,efi-mmap", &len);
+
+ if (len != efi_boot_mmap_size) {
+ pr_info(" EFI mmap size mismatch!\n");
+ map = NULL;
+ }
+ } else {
+ map = NULL;
+ }
+
+ return map;
+}
+
+static int __init remap_region(efi_memory_desc_t *md, efi_memory_desc_t *entry)
+{
+ u64 va;
+ u64 paddr;
+ u64 size;
+
+ *entry = *md;
+ paddr = entry->phys_addr;
+ size = entry->num_pages << EFI_PAGE_SHIFT;
+
+ /*
+ * Map everything writeback-capable as coherent memory,
+ * anything else as device.
+ */
+ if (md->attribute & EFI_MEMORY_WB)
+ va = (u64)((u32)efi_remap(paddr, size) & 0xffffffffUL);
+ else
+ va = (u64)((u32)efi_ioremap(paddr, size) & 0xffffffffUL);
+ if (!va)
+ return 0;
+ entry->virt_addr = va;
+
+ if (uefi_debug)
+ pr_info(" %016llx-%016llx => 0x%08x : (%s)\n",
+ paddr, paddr + size - 1, (u32)va,
+ md->attribute & EFI_MEMORY_WB ? "WB" : "I/O");
+
+ return 1;
+}
+
+static int __init remap_regions(void)
+{
+ void *p, *next;
+ efi_memory_desc_t *md;
+
+ memmap.phys_map = get_runtime_mmap();
+ if (!memmap.phys_map)
+ return 0;
+ memmap.map_end = (void *)memmap.phys_map + efi_boot_mmap_size;
+
+ /* Allocate space for the physical region map */
+ memmap.map = kzalloc(memmap.nr_map * memmap.desc_size, GFP_KERNEL);
+ if (!memmap.map)
+ return 0;
+
+ next = memmap.map;
+ for (p = memmap.phys_map; p < memmap.map_end; p += memmap.desc_size) {
+ md = p;
+ if (is_discardable_region(md))
+ continue;
+
+ if (!remap_region(p, next))
+ return 0;
+
+ next += memmap.desc_size;
+ }
+
+ memmap.map_end = next;
+ efi.memmap = &memmap;
+
+ efi.systab = efi_lookup_mapped_addr(efi_system_table);
+ if (efi.systab)
+ set_bit(EFI_SYSTEM_TABLES, &arm_efi_facility);
+ /*
+ * efi.systab->runtime is a 32-bit pointer to something guaranteed by
+ * the UEFI specification to be 1:1 mapped in a 4GB address space.
+ */
+ runtime = efi_lookup_mapped_addr((u32)efi.systab->runtime);
+
+ return 1;
+}
+
+
+/*
+ * This function switches the EFI runtime services to virtual mode.
+ * This operation must be performed only once in the system's lifetime,
+ * including any kecec calls.
+ *
+ * This must be done with a 1:1 mapping. The current implementation
+ * resolves this by disabling the MMU.
+ */
+efi_status_t __init phys_set_virtual_address_map(u32 memory_map_size,
+ u32 descriptor_size,
+ u32 descriptor_version,
+ efi_memory_desc_t *dsc)
+{
+ efi_phys_call_t *phys_set_map;
+ efi_status_t status;
+
+ phys_call_prologue();
+
+ phys_set_map = (void *)(unsigned long)virt_to_phys(efi_phys_call);
+
+ /* Called with caches disabled, returns with caches enabled */
+ status = phys_set_map(memory_map_size, descriptor_size,
+ descriptor_version, dsc,
+ efi.set_virtual_address_map);
+
+ phys_call_epilogue();
+
+ return status;
+}
+
+/*
+ * Called explicitly from init/mm.c
+ */
+void __init efi_enter_virtual_mode(void)
+{
+ efi_status_t status;
+
+ if (!efi_enabled(EFI_BOOT)) {
+ pr_info("EFI services will not be available.\n");
+ return;
+ } else {
+ pr_info("Remapping and enabling EFI services.\n");
+ }
+
+ /* Map the regions we memblock_remove:d earlier into kernel
+ address space */
+ if (!remap_regions()) {
+ pr_info("Failed to remap EFI regions - runtime services will not be available.\n");
+ return;
+ }
+
+ /* Call SetVirtualAddressMap with the physical address of the map */
+ efi.set_virtual_address_map =
+ (efi_set_virtual_address_map_t *)
+ runtime->set_virtual_address_map;
+ memmap.phys_map =
+ (efi_memory_desc_t *)(u32) __virt_to_phys((u32)memmap.map);
+
+ status = phys_set_virtual_address_map(memmap.nr_map * memmap.desc_size,
+ memmap.desc_size,
+ memmap.desc_version,
+ memmap.phys_map);
+
+ if (status != EFI_SUCCESS) {
+ pr_info("Failed to set EFI virtual address map!\n");
+ return;
+ }
+
+ /* Set up function pointers for efivars */
+ efi.get_variable = (efi_get_variable_t *)runtime->get_variable;
+ efi.get_next_variable =
+ (efi_get_next_variable_t *)runtime->get_next_variable;
+ efi.set_variable = (efi_set_variable_t *)runtime->set_variable;
+ set_bit(EFI_RUNTIME_SERVICES, &arm_efi_facility);
+}
diff --git a/arch/arm/kernel/efi_phys.S b/arch/arm/kernel/efi_phys.S
new file mode 100644
index 0000000..e36cc17
--- /dev/null
+++ b/arch/arm/kernel/efi_phys.S
@@ -0,0 +1,59 @@
+/*
+ * arch/arm/kernel/efi_phys.S
+ *
+ * Copyright (C) 2013 Linaro Ltd.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/linkage.h>
+#define PAR_MASK 0xfff
+
+ .text
+@ efi_phys_call(a, b, c, d, *f)
+ .align 5
+ .pushsection .idmap.text, "ax"
+ENTRY(efi_phys_call)
+ @ Save physical context
+ mov r12, sp
+ push {r4-r5, r12, lr}
+
+ @ Extract function pointer (don't write r12 beyond this)
+ ldr r12, [sp, #16]
+
+ @ Convert sp to 32-bit physical
+ mov lr, sp
+ ldr r4, =PAR_MASK
+ and r5, lr, r4 @ Extract lower 12 bits of sp
+ mcr p15, 0, lr, c7, c8, 1 @ Write VA -> ATS1CPW
+ mrc p15, 0, lr, c7, c4, 0 @ Physical Address Register
+ mvn r4, r4
+ and lr, lr, r4 @ Clear lower 12 bits of PA
+ add lr, lr, r5 @ Calculate phys sp
+ mov sp, lr @ Update
+
+ @ Disable MMU
+ mrc p15, 0, lr, c1, c0, 0 @ ctrl register
+ bic lr, lr, #0x1 @ ...............m
+ mcr p15, 0, lr, c1, c0, 0 @ disable MMU
+ isb
+
+ @ Make call
+ blx r12
+
+ pop {r4-r5, r12, lr}
+
+ @ Enable MMU + Caches
+ mrc p15, 0, r1, c1, c0, 0 @ ctrl register
+ orr r1, r1, #0x1000 @ ...i............
+ orr r1, r1, #0x0005 @ .............c.m
+ mcr p15, 0, r1, c1, c0, 0 @ enable MMU
+ isb
+
+ @ Restore virtual sp and return
+ mov sp, r12
+ bx lr
+ENDPROC(efi_phys_call)
+ .popsection
diff --git a/arch/arm/kernel/setup.c b/arch/arm/kernel/setup.c
index b0b2360..87ff95d 100644
--- a/arch/arm/kernel/setup.c
+++ b/arch/arm/kernel/setup.c
@@ -30,6 +30,7 @@
#include <linux/bug.h>
#include <linux/compiler.h>
#include <linux/sort.h>
+#include <linux/efi.h>
#include <asm/unified.h>
#include <asm/cp15.h>
@@ -57,6 +58,7 @@
#include <asm/unwind.h>
#include <asm/memblock.h>
#include <asm/virt.h>
+#include <asm/efi.h>
#include "atags.h"
@@ -884,6 +886,10 @@ void __init setup_arch(char **cmdline_p)
sanity_check_meminfo();
arm_memblock_init(&meminfo, mdesc);
+#ifdef CONFIG_EFI
+ efi_memblock_arm_reserve_range();
+#endif
+
paging_init(mdesc);
request_standard_resources(mdesc);
diff --git a/include/linux/efi.h b/include/linux/efi.h
index c084b6d..ba4d175 100644
--- a/include/linux/efi.h
+++ b/include/linux/efi.h
@@ -644,7 +644,7 @@ extern int __init efi_setup_pcdp_console(char *);
#define EFI_64BIT 5 /* Is the firmware 64-bit? */
#ifdef CONFIG_EFI
-# ifdef CONFIG_X86
+# if defined(CONFIG_X86) || defined(CONFIG_ARM)
extern int efi_enabled(int facility);
# else
static inline int efi_enabled(int facility)
--
1.7.10.4
Adding devicetree list since you are defining bindings...
On 10/03/2013 06:24 AM, Leif Lindholm wrote:
> This patch provides documentation of the [U]EFI runtime services and
> configuration features.
>
> Cc: [email protected]
> Signed-off-by: Leif Lindholm <[email protected]>
> Acked-by: Rob Landley <[email protected]>
> ---
> Documentation/arm/00-INDEX | 3 +++
> Documentation/arm/uefi.txt | 47 ++++++++++++++++++++++++++++++++++++++++++++
> 2 files changed, 50 insertions(+)
> create mode 100644 Documentation/arm/uefi.txt
>
> diff --git a/Documentation/arm/00-INDEX b/Documentation/arm/00-INDEX
> index 4978456..87e01d1 100644
> --- a/Documentation/arm/00-INDEX
> +++ b/Documentation/arm/00-INDEX
> @@ -36,3 +36,6 @@ nwfpe/
> - NWFPE floating point emulator documentation
> swp_emulation
> - SWP/SWPB emulation handler/logging description
> +
> +uefi.txt
> + - [U]EFI configuration and runtime services documentation
> diff --git a/Documentation/arm/uefi.txt b/Documentation/arm/uefi.txt
> new file mode 100644
> index 0000000..e6e4d41
> --- /dev/null
> +++ b/Documentation/arm/uefi.txt
> @@ -0,0 +1,47 @@
> +UEFI, the Unified Extensible Firmware Interface is a speifcication
> +governing the behaviours of compatible firmware interfaces. It is
> +maintained by the UEFI Forum - http://www.uefi.org/.
> +
> +Since UEFI is an evolution of its predecessor 'EFI', the terms EFI and
> +UEFI are used somewhat interchangeably in this document and associated
> +source code.
> +
> +The implementation depends on receiving the UEFI runtime memory map and a
> +pointer to the System Table in a Flattened Device Tree - so is only available
> +with CONFIG_OF.
> +
> +It parses the FDT /chosen node for the following parameters:
DT bindings should be documented in Documentation/devicetree/bindings.
I also wonder if this would be more appropriately placed in a /firmware
node.
> +- 'linux,efi-system-table':
> + Physical address of the system table. (required)
> + 64-bit value since an ARMv7 plattform may support LPAE, and to facilitate
> + code sharing with arm64. Top 32 bits will be ignored, since UEFI specification
> + mandates a 1:1 mapping of all RAM.
> +- 'linux,efi-mmap':
> + The EFI memory map as an embedded property. (required)
> + An array of type EFI_MEMORY_DESCRIPTOR as described by the UEFI
> + specification, current version described in Linux by efi_memory_desc_t.
Is that too complex to describe here?
> + The memory map is represented in little-endian, not DT, byte order.
> + This map needs to contain at least the regions to be preserved for runtime
> + services, but would normally just be the map retreieved by calling UEFI
> + GetMemoryMap() immediately before ExitBootServices().
> +- 'linux,efi-mmap-desc-size':
> + Size of each descriptor in the memory map. (override default)
32-bit value?
> +- 'linux,efi-mmap-desc-ver':
> + Memory descriptor format version. (override default)
String? Number?
Are these all generated by UEFI at runtime or could they be statically
set in a platform's DTB?
How would other OS's get this information? Is this really linux specific?
Rob
> +
> +It also depends on early_memremap() to parse the UEFI configuration tables.
> +
> +For actually enabling [U]EFI support, enable:
> +- CONFIG_EFI=y
> +- CONFIG_EFI_VARS=y or m
> +
> +After the kernel has mapped the required regions into its address space,
> +a SetVirtualAddressMap() call is made into UEFI in order to update
> +relocations. This call must be performed with all the code in a 1:1
> +mapping. This implementation achieves this by temporarily disabling the
> +MMU for the duration of this call. This can only be done safely:
> +- before secondary CPUs are brought online.
> +- after early_initcalls have completed, since it uses setup_mm_for_reboot().
> +
> +For verbose debug messages, specify 'uefi_debug' on the kernel command
> +line.
>
Hi Leif,
On Thu, Oct 03, 2013 at 12:24:39PM +0100, Leif Lindholm wrote:
> This patch provides documentation of the [U]EFI runtime services and
> configuration features.
>
> Cc: [email protected]
> Signed-off-by: Leif Lindholm <[email protected]>
> Acked-by: Rob Landley <[email protected]>
> ---
> Documentation/arm/00-INDEX | 3 +++
> Documentation/arm/uefi.txt | 47 ++++++++++++++++++++++++++++++++++++++++++++
> 2 files changed, 50 insertions(+)
> create mode 100644 Documentation/arm/uefi.txt
>
> diff --git a/Documentation/arm/00-INDEX b/Documentation/arm/00-INDEX
> index 4978456..87e01d1 100644
> --- a/Documentation/arm/00-INDEX
> +++ b/Documentation/arm/00-INDEX
> @@ -36,3 +36,6 @@ nwfpe/
> - NWFPE floating point emulator documentation
> swp_emulation
> - SWP/SWPB emulation handler/logging description
> +
> +uefi.txt
> + - [U]EFI configuration and runtime services documentation
> diff --git a/Documentation/arm/uefi.txt b/Documentation/arm/uefi.txt
> new file mode 100644
> index 0000000..e6e4d41
> --- /dev/null
> +++ b/Documentation/arm/uefi.txt
> @@ -0,0 +1,47 @@
> +UEFI, the Unified Extensible Firmware Interface is a speifcication
> +governing the behaviours of compatible firmware interfaces. It is
> +maintained by the UEFI Forum - http://www.uefi.org/.
> +
> +Since UEFI is an evolution of its predecessor 'EFI', the terms EFI and
> +UEFI are used somewhat interchangeably in this document and associated
> +source code.
> +
> +The implementation depends on receiving the UEFI runtime memory map and a
> +pointer to the System Table in a Flattened Device Tree - so is only available
> +with CONFIG_OF.
> +
> +It parses the FDT /chosen node for the following parameters:
> +- 'linux,efi-system-table':
> + Physical address of the system table. (required)
> + 64-bit value since an ARMv7 plattform may support LPAE, and to facilitate
s/plattform/platform/
> + code sharing with arm64. Top 32 bits will be ignored, since UEFI specification
> + mandates a 1:1 mapping of all RAM.
You could use something like #size-cells to describe how big this is
going to be. Is this Linux-specific -- it looks like something provided
by EFI rather than the kernel itself.
> +- 'linux,efi-mmap':
> + The EFI memory map as an embedded property. (required)
> + An array of type EFI_MEMORY_DESCRIPTOR as described by the UEFI
> + specification, current version described in Linux by efi_memory_desc_t.
> + The memory map is represented in little-endian, not DT, byte order.
> + This map needs to contain at least the regions to be preserved for runtime
> + services, but would normally just be the map retreieved by calling UEFI
> + GetMemoryMap() immediately before ExitBootServices().
This is a little scary. If the format is so complicated, should it
really be embedded? How big is this likely to be?
Given that this is in a format defined externally, this isn't really
Linux-specific. Maybe we need an "efi" pseudo-vendor prefix.
> +- 'linux,efi-mmap-desc-size':
> + Size of each descriptor in the memory map. (override default)
What units is this in? How many u32 cells does this take up (one
presumably)?
> +- 'linux,efi-mmap-desc-ver':
> + Memory descriptor format version. (override default)
Type, format, valid values and their meaning?
Thanks,
Mark.
> +
> +It also depends on early_memremap() to parse the UEFI configuration tables.
> +
> +For actually enabling [U]EFI support, enable:
> +- CONFIG_EFI=y
> +- CONFIG_EFI_VARS=y or m
> +
> +After the kernel has mapped the required regions into its address space,
> +a SetVirtualAddressMap() call is made into UEFI in order to update
> +relocations. This call must be performed with all the code in a 1:1
> +mapping. This implementation achieves this by temporarily disabling the
> +MMU for the duration of this call. This can only be done safely:
> +- before secondary CPUs are brought online.
> +- after early_initcalls have completed, since it uses setup_mm_for_reboot().
> +
> +For verbose debug messages, specify 'uefi_debug' on the kernel command
> +line.
> --
> 1.7.10.4
>
>
`> _______________________________________________
> linux-arm-kernel mailing list
> [email protected]
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
>
On Thu, Oct 03, 2013 at 11:11:18AM -0500, Rob Herring wrote:
> Adding devicetree list since you are defining bindings...
>
> > +with CONFIG_OF.
> > +
> > +It parses the FDT /chosen node for the following parameters:
>
> DT bindings should be documented in Documentation/devicetree/bindings.
>
> I also wonder if this would be more appropriately placed in a /firmware
> node.
This is information passed to the kernel by the bootloader - not
system descriptiont - so I don't quite see why it needs different
treatment from initrd and bootargs.
Feedback on v1 was:
https://lkml.org/lkml/2013/6/26/378
and
https://lkml.org/lkml/2013/6/27/420
I don't really mind either way, but the current layout is now used
across 3 sets of kernel patches, so we need to reach some sort of
consensus. Interested parties so far: me, you, Grant, Arnd, Mark.
> > +- 'linux,efi-mmap':
> > + The EFI memory map as an embedded property. (required)
> > + An array of type EFI_MEMORY_DESCRIPTOR as described by the UEFI
> > + specification, current version described in Linux by efi_memory_desc_t.
>
> Is that too complex to describe here?
No, just felt a bit redundant, and also not architecture-specific.
> > + The memory map is represented in little-endian, not DT, byte order.
> > + This map needs to contain at least the regions to be preserved for runtime
> > + services, but would normally just be the map retreieved by calling UEFI
> > + GetMemoryMap() immediately before ExitBootServices().
> > +- 'linux,efi-mmap-desc-size':
> > + Size of each descriptor in the memory map. (override default)
>
> 32-bit value?
Value as returned by the above mentioned GetMemoryMap().
Defined in UEFI specification (and <linux/efi.h>) as 32-bit (native
int). But yes, I can be explicit.
> > +- 'linux,efi-mmap-desc-ver':
> > + Memory descriptor format version. (override default)
>
> String? Number?
Value as returned by the above mentioned GetMemoryMap().
Defined in the UEFI specification as 32-bit (uint32), not
architecture specific. And I can add that too.
> Are these all generated by UEFI at runtime or could they be statically
> set in a platform's DTB?
Generated at runtime.
This is not the platform memory map, this is the UEFI memory map,
which tells us which regions we need to preserve for runtime
services, ACPI and such.
> How would other OS's get this information? Is this really linux specific?
The way it is passed through DT is. Other operating systems might keep
boot services running for longer, and make calls into UEFI later, so
not needing to cache the data. Since boot services means the timer
interrupt is active, the ARM Linux boot protocol effectively prohibits
this.
Many of these questions are about generic UEFI mechanisms.
If they need to be documented outside the UEFI specification,
Documentation/arm is not the right place for it.
If you want, I could give a basic Documentation/uefi.txt a shot.
/
Leif
On Thu, Oct 03, 2013 at 06:10:54PM +0100, Mark Rutland wrote:
> > +The implementation depends on receiving the UEFI runtime memory map and a
> > +pointer to the System Table in a Flattened Device Tree - so is only available
> > +with CONFIG_OF.
> > +
> > +It parses the FDT /chosen node for the following parameters:
> > +- 'linux,efi-system-table':
> > + Physical address of the system table. (required)
> > + 64-bit value since an ARMv7 plattform may support LPAE, and to facilitate
>
> s/plattform/platform/
Indeed, thanks.
> > + code sharing with arm64. Top 32 bits will be ignored, since UEFI specification
> > + mandates a 1:1 mapping of all RAM.
>
> You could use something like #size-cells to describe how big this is
> going to be. Is this Linux-specific -- it looks like something provided
> by EFI rather than the kernel itself.
Provided by the bootloader, which may be the kernel's UEFI stub, or
something else (like GRUB). Extracted from UEFI by stub (or other
loader) which executes as a UEFI application.
> > +- 'linux,efi-mmap':
> > + The EFI memory map as an embedded property. (required)
> > + An array of type EFI_MEMORY_DESCRIPTOR as described by the UEFI
> > + specification, current version described in Linux by efi_memory_desc_t.
> > + The memory map is represented in little-endian, not DT, byte order.
> > + This map needs to contain at least the regions to be preserved for runtime
> > + services, but would normally just be the map retreieved by calling UEFI
> > + GetMemoryMap() immediately before ExitBootServices().
>
> This is a little scary. If the format is so complicated, should it
> really be embedded? How big is this likely to be?
With current format, and it has yet to be extended beyond v1, it is
40 bytes per entry. Number of entries will vary depending on how many
regions are allocated in UEFI (1 entry per region). I have heard of
some x86 systems with 128 or more regions, but my devboards have more
like 16, and my desktop ~60.
Having it embedded in the DT means the data passed by the bootloader
to the kernel is automatically kept together (for kexec or such).
Not vital, but nice.
> Given that this is in a format defined externally, this isn't really
> Linux-specific. Maybe we need an "efi" pseudo-vendor prefix.
A discussion for the other thread.
> > +- 'linux,efi-mmap-desc-size':
> > + Size of each descriptor in the memory map. (override default)
>
> What units is this in? How many u32 cells does this take up (one
> presumably)?
Value as returned by UEFI GetMemoryMap() boot service call.
Yes, one for ARM.
> > +- 'linux,efi-mmap-desc-ver':
> > + Memory descriptor format version. (override default)
>
> Type, format, valid values and their meaning?
Meaning defined by UEFI specification, currently providing the only
valid value as 1. 32-bit uint.
/
Leif
Hi Leif,
On Thu, Oct 03, 2013 at 06:18:02PM +0100, Leif Lindholm wrote:
> On Thu, Oct 03, 2013 at 11:11:18AM -0500, Rob Herring wrote:
> > Adding devicetree list since you are defining bindings...
> >
> > > +with CONFIG_OF.
> > > +
> > > +It parses the FDT /chosen node for the following parameters:
> >
> > DT bindings should be documented in Documentation/devicetree/bindings.
> >
> > I also wonder if this would be more appropriately placed in a /firmware
> > node.
>
> This is information passed to the kernel by the bootloader - not
> system descriptiont - so I don't quite see why it needs different
> treatment from initrd and bootargs.
That's somewhat debatable -- the description isn't a configuration
option (as bootargs and initrd are) so much as a description of the
firmware.
>
> Feedback on v1 was:
> https://lkml.org/lkml/2013/6/26/378
> and
> https://lkml.org/lkml/2013/6/27/420
>
> I don't really mind either way, but the current layout is now used
> across 3 sets of kernel patches, so we need to reach some sort of
> consensus. Interested parties so far: me, you, Grant, Arnd, Mark.
>
> > > +- 'linux,efi-mmap':
> > > + The EFI memory map as an embedded property. (required)
> > > + An array of type EFI_MEMORY_DESCRIPTOR as described by the UEFI
> > > + specification, current version described in Linux by efi_memory_desc_t.
> >
> > Is that too complex to describe here?
>
> No, just felt a bit redundant, and also not architecture-specific.
I'd point to the EFI spec and leave it at that. The particular encoding
is defined by EFI and is beyond the scope of this document (endianness
issues aside).
If this were a device with a single window of registers we'd just
describe the whole window and not the format of each individual register
or bit within them. I see no reason to do differently here.
As the efi-mmap is built by EFI, and is in a format defined by EFI, I'd
also drop the "linux" vendor prefix -- Linux has had no hand in defining
this property. It may make sense to have "efi" as a vendor prefix, but
perhaps others won't like that?
>
> > > + The memory map is represented in little-endian, not DT, byte order.
> > > + This map needs to contain at least the regions to be preserved for runtime
> > > + services, but would normally just be the map retreieved by calling UEFI
> > > + GetMemoryMap() immediately before ExitBootServices().
> > > +- 'linux,efi-mmap-desc-size':
> > > + Size of each descriptor in the memory map. (override default)
> >
> > 32-bit value?
>
> Value as returned by the above mentioned GetMemoryMap().
> Defined in UEFI specification (and <linux/efi.h>) as 32-bit (native
> int). But yes, I can be explicit.
Please do. Something like:
- efi-mmap-desc-size:
A single (u32) cell describing the size of each descriptor in the
memory map, if not the EFI default size of $EFI_MMAP_DESC_SIZE. See
$EFI_DOCUMENTATION for more information.
>
> > > +- 'linux,efi-mmap-desc-ver':
> > > + Memory descriptor format version. (override default)
> >
> > String? Number?
>
> Value as returned by the above mentioned GetMemoryMap().
> Defined in the UEFI specification as 32-bit (uint32), not
> architecture specific. And I can add that too.
How about something like:
- efi-mmap-desc-version:
A single (u32) cell describing the memory descriptor format version,
if not 1 (the current EFI version). See $EFI_DOCUMENTATION for more
information.
>
> > Are these all generated by UEFI at runtime or could they be statically
> > set in a platform's DTB?
>
> Generated at runtime.
> This is not the platform memory map, this is the UEFI memory map,
> which tells us which regions we need to preserve for runtime
> services, ACPI and such.
>
> > How would other OS's get this information? Is this really linux specific?
>
> The way it is passed through DT is. Other operating systems might keep
> boot services running for longer, and make calls into UEFI later, so
> not needing to cache the data. Since boot services means the timer
> interrupt is active, the ARM Linux boot protocol effectively prohibits
> this.
I suspect other OSs that may use DT are likely to have similar
requirements, or could at least cope with this mechanism.
I think before we can take any code in this area we need to define the
boot protocol (i.e. do we use DT to describe EFI & ACPI, what gets
described where) for EFI and/or ACPI. We have enough of a mess as it is,
and we need to carefully control anything we add to the fray. We should
discuss this at the upcoming LinuxCon, ELC, and Linaro Connect.
Thanks,
Mark.
>
> Many of these questions are about generic UEFI mechanisms.
> If they need to be documented outside the UEFI specification,
> Documentation/arm is not the right place for it.
>
> If you want, I could give a basic Documentation/uefi.txt a shot.
>
> /
> Leif
> --
> To unsubscribe from this list: send the line "unsubscribe devicetree" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
On Thu, 03 Oct, at 12:24:40PM, Leif Lindholm wrote:
> +/*
> + * If you need to (temporarily) support buggy firmware.
> + */
> +#define KEEP_BOOT_SERVICES_REGIONS
Have you seen firmware that requires this? I'm just curious more than
anything else.
> +/*
> + * Returns 1 if 'facility' is enabled, 0 otherwise.
> + */
> +int efi_enabled(int facility)
> +{
> + return test_bit(facility, &arm_efi_facility) != 0;
> +}
> +EXPORT_SYMBOL(efi_enabled);
This should move to drivers/firmware/efi/efi.c. Let me write a patch
that moves the x86 stuff out of arch/x86 and means you can get rid of
this hunk.
> +/*
> + * Called explicitly from init/mm.c
> + */
That's init/main.c.
> +void __init efi_enter_virtual_mode(void)
> +{
> + efi_status_t status;
> +
> + if (!efi_enabled(EFI_BOOT)) {
> + pr_info("EFI services will not be available.\n");
> + return;
This is dead code as PATCH 3 does,
diff --git a/init/main.c b/init/main.c
index af310af..ec6d76e 100644
--- a/init/main.c
+++ b/init/main.c
@@ -875,6 +875,10 @@ static noinline void __init
kernel_init_freeable(void)
smp_prepare_cpus(setup_max_cpus);
do_pre_smp_initcalls();
+
+ if (IS_ENABLED(CONFIG_ARM) && efi_enabled(EFI_BOOT))
+ efi_enter_virtual_mode();
+
--
Matt Fleming, Intel Open Source Technology Center
On Thu, Oct 17, 2013 at 03:07:39PM +0100, Matt Fleming wrote:
> > +/*
> > + * If you need to (temporarily) support buggy firmware.
> > + */
> > +#define KEEP_BOOT_SERVICES_REGIONS
>
> Have you seen firmware that requires this? I'm just curious more than
> anything else.
Not really.
I _think_ I saw it on a debug build of a development platform once.
That coincided with me seeing a post on linux-efi about some laptop
that broke unless boot services regions were preserved, so I decided
to put it in there for any future debugging.
> > +/*
> > + * Returns 1 if 'facility' is enabled, 0 otherwise.
> > + */
> > +int efi_enabled(int facility)
> > +{
> > + return test_bit(facility, &arm_efi_facility) != 0;
> > +}
> > +EXPORT_SYMBOL(efi_enabled);
>
> This should move to drivers/firmware/efi/efi.c. Let me write a patch
> that moves the x86 stuff out of arch/x86 and means you can get rid of
> this hunk.
Excellent, thanks!
> > +/*
> > + * Called explicitly from init/mm.c
> > + */
>
> That's init/main.c.
*cough*, right.
> > +void __init efi_enter_virtual_mode(void)
> > +{
> > + efi_status_t status;
> > +
> > + if (!efi_enabled(EFI_BOOT)) {
> > + pr_info("EFI services will not be available.\n");
> > + return;
>
> This is dead code as PATCH 3 does,
>
> diff --git a/init/main.c b/init/main.c
> index af310af..ec6d76e 100644
> --- a/init/main.c
> +++ b/init/main.c
> @@ -875,6 +875,10 @@ static noinline void __init
> kernel_init_freeable(void)
> smp_prepare_cpus(setup_max_cpus);
>
> do_pre_smp_initcalls();
> +
> + if (IS_ENABLED(CONFIG_ARM) && efi_enabled(EFI_BOOT))
> + efi_enter_virtual_mode();
> +
True.
However, this call site is likely to change in the future (preferably
to an early_initcall), if we redesign the memory mapping to be reusable
after kexec(). At which point the test in efi_enter_virtual_mode() will
make sense again.
Could I change the test in init/main.c to do
if (IS_ENABLED(CONFIG_ARM) && IS_ENABLED(CONFIG_EFI))
instead?
/
Leif
On Thu, 2013-10-17 at 16:31 +0200, Leif Lindholm wrote:
> On Thu, Oct 17, 2013 at 03:07:39PM +0100, Matt Fleming wrote:
> > > +/*
> > > + * If you need to (temporarily) support buggy firmware.
> > > + */
> > > +#define KEEP_BOOT_SERVICES_REGIONS
> >
> > Have you seen firmware that requires this? I'm just curious more than
> > anything else.
>
> Not really.
> I _think_ I saw it on a debug build of a development platform once.
> That coincided with me seeing a post on linux-efi about some laptop
> that broke unless boot services regions were preserved, so I decided
> to put it in there for any future debugging.
>
I see a fairly consistent segfault when the arm64 kernel calls the
SetVirtualAddressMap function. It doesn't happen when the boot services
regions are included in the mapping. This is with edk2 firmware. I
haven't really dug into where it goes wrong.