[PATCH v4 00/30] EFI and ACPI support for arm64

[PATCH v4 00/30] EFI and ACPI support for arm64

[Nikos Nikoleris]

Hello,

This series adds initial support for building arm64 tests as EFI
apps and running them under QEMU. Much like x86_64, we import external
dependencies from gnu-efi and adapt them to work with types and other
assumptions from kvm-unit-tests. In addition, this series adds support
for enumerating parts of the system using ACPI.

The first set of patches, moves the existing ACPI code to the common
lib path. Then, it extends definitions and functions to allow for more
robust discovery of ACPI tables. We add support for setting up the PSCI
conduit, discovering the UART, timers, GIC and cpus via ACPI. The code
retains existing behavior and gives priority to discovery through DT
when one has been provided.

In the second set of patches, we add support for getting the command
line from the EFI shell. This is a requirement for many of the
existing arm64 tests.

In the third set of patches, we import code from gnu-efi, make minor
changes and add an alternative setup sequence from arm64 systems that
boot through EFI. Finally, we add support in the build system and a
run script which is used to run an EFI app.

After this set of patches one can build arm64 EFI tests:

$> ./configure --enable-efi
$> make

And use the run script to run an EFI tests:

$> ./arm/efi/run ./arm/selftest.efi -smp 2 -m 256 -append "setup smp=2 mem=256"

Or all tests:

$> ./run_tests.sh

There are a few items that this series does not address but they would
be useful to have:
 - Support for booting the system from EL2. Currently, we assume that a
   test starts EL1. This will be required to run EFI tests on sytems
   that implement EL2.
 - Support for reading environment variables and populating __envp.
 - Support for discovering the PCI subsystem using ACPI.
 - Get rid of other assumptions (e.g., vmalloc area) that don't hold on
   real HW.
 - Various fixes related to cache maintaince to better support turn the
   MMU off.
 - Switch to a new stack and avoid relying on the one provided by EFI.

git branch: https://github.com/relokin/kvm-unit-tests/pull/new/target-efi-upstream-v4

v3: https://lore.kernel.org/all/20220630100324.3153655-1-nikos.nikoleris@arm.com/
v2: https://lore.kernel.org/kvm/20220506205605.359830-1-nikos.nikoleris@arm.com/

Changes in v4:
 - Removed patch that reworks cache maintenance when turning the MMU
   off. This is not strictly required for EFI tests running with tcg and
   will be addressed in a separate series by Alex.
 - Fix compilation for arm (Alex).
 - Convert ACPI tables to Linux style (Alex).

Changes in v3:
 - Addressed feedback from Drew, Alex and Ricardo. Many thanks for the reviews!
 - Added support for discovering the GIC through ACPI
 - Added a missing header file (<elf.h>)
 - Added support for correctly parsing the outcome of tests (./run_tests)

Thanks,

Nikos

Alexandru Elisei (3):
  arm/Makefile.common: Compile lib/acpi.c if CONFIG_EFI=y
  lib/acpi: Convert table names to Linux style
  lib: arm: Print test exit status

Andrew Jones (2):
  arm/arm64: Rename etext to _etext
  arm64: Add a new type of memory type flag MR_F_RESERVED

Nikos Nikoleris (25):
  lib: Move acpi header and implementation to lib
  x86: Move x86_64-specific EFI CFLAGS to x86_64 Makefile
  lib: Apply Lindent to acpi.{c,h}
  lib: Fix style for acpi.{c,h}
  x86: Avoid references to fields of ACPI tables
  lib/acpi: Ensure all struct definition for ACPI tables are packed
  lib/acpi: Add support for the XSDT table
  lib/acpi: Extend the definition of the FADT table
  devicetree: Check that fdt is not NULL in dt_available()
  arm64: Add support for setting up the PSCI conduit through ACPI
  arm64: Add support for discovering the UART through ACPI
  arm64: Add support for timer initialization through ACPI
  arm64: Add support for cpu initialization through ACPI
  arm64: Add support for gic initialization through ACPI
  lib/printf: Support for precision modifier in printf
  lib/printf: Add support for printing wide strings
  lib/efi: Add support for getting the cmdline
  lib: Avoid ms_abi for calls related to EFI on arm64
  arm64: Add a setup sequence for systems that boot through EFI
  arm64: Copy code from GNU-EFI
  arm64: Change GNU-EFI imported file to use defined types
  arm64: Use code from the gnu-efi when booting with EFI
  lib: Avoid external dependency in libelf
  arm64: Add support for efi in Makefile
  arm64: Add an efi/run script

 Makefile                    |   4 -
 arm/Makefile.arm            |   6 +
 arm/Makefile.arm64          |  22 ++-
 arm/Makefile.common         |  51 ++++--
 arm/cstart.S                |   1 +
 arm/cstart64.S              |   7 +
 arm/dummy.c                 |  12 ++
 arm/efi/crt0-efi-aarch64.S  | 141 +++++++++++++++++
 arm/efi/elf_aarch64_efi.lds |  63 ++++++++
 arm/efi/reloc_aarch64.c     |  93 +++++++++++
 arm/efi/run                 |  61 ++++++++
 arm/flat.lds                |   2 +-
 arm/micro-bench.c           |   4 +-
 arm/run                     |  14 +-
 arm/timer.c                 |  10 +-
 configure                   |  15 +-
 lib/acpi.c                  | 129 +++++++++++++++
 lib/acpi.h                  | 303 ++++++++++++++++++++++++++++++++++++
 lib/argv.c                  |   2 +-
 lib/argv.h                  |   1 +
 lib/arm/asm/setup.h         |   9 ++
 lib/arm/asm/timer.h         |   2 +
 lib/arm/gic.c               | 139 ++++++++++++++++-
 lib/arm/io.c                |  41 ++++-
 lib/arm/mmu.c               |   4 +
 lib/arm/psci.c              |  37 ++++-
 lib/arm/setup.c             | 264 +++++++++++++++++++++++++------
 lib/arm/timer.c             |  92 +++++++++++
 lib/devicetree.c            |   2 +-
 lib/efi.c                   | 102 ++++++++++++
 lib/elf.h                   |  57 +++++++
 lib/libcflat.h              |   1 +
 lib/linux/efi.h             |  25 +++
 lib/printf.c                | 194 +++++++++++++++++++++--
 lib/x86/acpi.c              |  82 ----------
 lib/x86/acpi.h              | 112 -------------
 lib/x86/asm/setup.h         |   2 +-
 lib/x86/setup.c             |   2 +-
 scripts/runtime.bash        |  13 +-
 x86/Makefile.common         |   2 +-
 x86/Makefile.x86_64         |   4 +
 x86/s3.c                    |  21 +--
 x86/vmexit.c                |   4 +-
 43 files changed, 1831 insertions(+), 321 deletions(-)
 create mode 100644 arm/dummy.c
 create mode 100644 arm/efi/crt0-efi-aarch64.S
 create mode 100644 arm/efi/elf_aarch64_efi.lds
 create mode 100644 arm/efi/reloc_aarch64.c
 create mode 100755 arm/efi/run
 create mode 100644 lib/acpi.c
 create mode 100644 lib/acpi.h
 create mode 100644 lib/arm/timer.c
 create mode 100644 lib/elf.h
 delete mode 100644 lib/x86/acpi.c
 delete mode 100644 lib/x86/acpi.h

-- 
2.25.1

[PATCH v4 01/30] lib: Move acpi header and implementation to lib

[Nikos Nikoleris]

Move acpi.h to lib to make it available for other architectures.

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
---
 lib/{x86 => }/acpi.c | 0
 lib/{x86 => }/acpi.h | 4 ++--
 lib/x86/asm/setup.h  | 2 +-
 x86/Makefile.common  | 2 +-
 x86/s3.c             | 2 +-
 x86/vmexit.c         | 2 +-
 6 files changed, 6 insertions(+), 6 deletions(-)
 rename lib/{x86 => }/acpi.c (100%)
 rename lib/{x86 => }/acpi.h (99%)

diff --git a/lib/x86/acpi.c b/lib/acpi.c
similarity index 100%
rename from lib/x86/acpi.c
rename to lib/acpi.c
diff --git a/lib/x86/acpi.h b/lib/acpi.h
similarity index 99%
rename from lib/x86/acpi.h
rename to lib/acpi.h
index 67ba3899..1e89840c 100644
--- a/lib/x86/acpi.h
+++ b/lib/acpi.h
@@ -1,5 +1,5 @@
-#ifndef _X86_ACPI_H_
-#define _X86_ACPI_H_
+#ifndef _ACPI_H_
+#define _ACPI_H_
 
 #include "libcflat.h"
 
diff --git a/lib/x86/asm/setup.h b/lib/x86/asm/setup.h
index 1f384274..458eac85 100644
--- a/lib/x86/asm/setup.h
+++ b/lib/x86/asm/setup.h
@@ -4,7 +4,7 @@
 unsigned long setup_tss(u8 *stacktop);
 
 #ifdef CONFIG_EFI
-#include "x86/acpi.h"
+#include "acpi.h"
 #include "x86/apic.h"
 #include "x86/processor.h"
 #include "x86/smp.h"
diff --git a/x86/Makefile.common b/x86/Makefile.common
index 365e199f..9f2bc93f 100644
--- a/x86/Makefile.common
+++ b/x86/Makefile.common
@@ -2,6 +2,7 @@
 
 all: directories test_cases
 
+cflatobjs += lib/acpi.o
 cflatobjs += lib/pci.o
 cflatobjs += lib/pci-edu.o
 cflatobjs += lib/alloc.o
@@ -18,7 +19,6 @@ cflatobjs += lib/x86/apic.o
 cflatobjs += lib/x86/atomic.o
 cflatobjs += lib/x86/desc.o
 cflatobjs += lib/x86/isr.o
-cflatobjs += lib/x86/acpi.o
 cflatobjs += lib/x86/stack.o
 cflatobjs += lib/x86/fault_test.o
 cflatobjs += lib/x86/delay.o
diff --git a/x86/s3.c b/x86/s3.c
index 6e41d0c4..378d37ae 100644
--- a/x86/s3.c
+++ b/x86/s3.c
@@ -1,5 +1,5 @@
 #include "libcflat.h"
-#include "x86/acpi.h"
+#include "acpi.h"
 #include "asm/io.h"
 
 static u32* find_resume_vector_addr(void)
diff --git a/x86/vmexit.c b/x86/vmexit.c
index b1eed8d1..865ccf6b 100644
--- a/x86/vmexit.c
+++ b/x86/vmexit.c
@@ -1,9 +1,9 @@
 #include "libcflat.h"
+#include "acpi.h"
 #include "smp.h"
 #include "pci.h"
 #include "x86/vm.h"
 #include "x86/desc.h"
-#include "x86/acpi.h"
 #include "x86/apic.h"
 #include "x86/isr.h"
 
-- 
2.25.1

[PATCH v4 02/30] x86: Move x86_64-specific EFI CFLAGS to x86_64 Makefile

[Nikos Nikoleris]

Compiler flag -macculate-outgoing-args is only needed by the x86_64
ABI. Move it to the relevant Makefile.

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Ricardo Koller <ricarkol@google.com>
---
 Makefile            | 4 ----
 x86/Makefile.x86_64 | 4 ++++
 2 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/Makefile b/Makefile
index 6ed5deac..307bc291 100644
--- a/Makefile
+++ b/Makefile
@@ -40,14 +40,10 @@ OBJDIRS += $(LIBFDT_objdir)
 
 # EFI App
 ifeq ($(CONFIG_EFI),y)
-EFI_ARCH = x86_64
 EFI_CFLAGS := -DCONFIG_EFI
 # The following CFLAGS and LDFLAGS come from:
 #   - GNU-EFI/Makefile.defaults
 #   - GNU-EFI/apps/Makefile
-# Function calls must include the number of arguments passed to the functions
-# More details: https://wiki.osdev.org/GNU-EFI
-EFI_CFLAGS += -maccumulate-outgoing-args
 # GCC defines wchar to be 32 bits, but EFI expects 16 bits
 EFI_CFLAGS += -fshort-wchar
 # EFI applications use PIC as they are loaded to dynamic addresses, not a fixed
diff --git a/x86/Makefile.x86_64 b/x86/Makefile.x86_64
index f483dead..2771a6fa 100644
--- a/x86/Makefile.x86_64
+++ b/x86/Makefile.x86_64
@@ -2,6 +2,10 @@ cstart.o = $(TEST_DIR)/cstart64.o
 bits = 64
 ldarch = elf64-x86-64
 ifeq ($(CONFIG_EFI),y)
+# Function calls must include the number of arguments passed to the functions
+# More details: https://wiki.osdev.org/GNU-EFI
+CFLAGS += -maccumulate-outgoing-args
+
 exe = efi
 bin = so
 FORMAT = efi-app-x86_64
-- 
2.25.1

[PATCH v4 03/30] arm/Makefile.common: Compile lib/acpi.c if CONFIG_EFI=y

[Nikos Nikoleris]

From: Alexandru Elisei <alexandru.elisei@arm.com>

For arm/arm64, ACPI is available only when UEFI is present. Compile
acpi.c if CONFIG_EFI is set to 'y'. For now, only x86 can configure the
tests with --enable-efi, but that will soon change.

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
---
 arm/Makefile.common | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/arm/Makefile.common b/arm/Makefile.common
index 1bbec64f..3a726c20 100644
--- a/arm/Makefile.common
+++ b/arm/Makefile.common
@@ -54,6 +54,10 @@ cflatobjs += lib/arm/smp.o
 cflatobjs += lib/arm/delay.o
 cflatobjs += lib/arm/gic.o lib/arm/gic-v2.o lib/arm/gic-v3.o
 
+ifeq ($(CONFIG_EFI),y)
+cflatobjs += lib/acpi.o
+endif
+
 OBJDIRS += lib/arm
 
 libeabi = lib/arm/libeabi.a
-- 
2.25.1

[PATCH v4 04/30] lib: Apply Lindent to acpi.{c,h}

[Nikos Nikoleris]

The change was done by modifying Linux's scripts/Lindent to use 100
columns instead of 80.

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
---
 lib/acpi.c |  70 ++++++++++++------------
 lib/acpi.h | 157 ++++++++++++++++++++++++++---------------------------
 2 files changed, 111 insertions(+), 116 deletions(-)

diff --git a/lib/acpi.c b/lib/acpi.c
index de275caf..836156a1 100644
--- a/lib/acpi.c
+++ b/lib/acpi.c
@@ -36,47 +36,45 @@ static struct rsdp_descriptor *get_rsdp(void)
 }
 #endif /* CONFIG_EFI */
 
-void* find_acpi_table_addr(u32 sig)
+void *find_acpi_table_addr(u32 sig)
 {
-    struct rsdp_descriptor *rsdp;
-    struct rsdt_descriptor_rev1 *rsdt;
-    void *end;
-    int i;
+	struct rsdp_descriptor *rsdp;
+	struct rsdt_descriptor_rev1 *rsdt;
+	void *end;
+	int i;
 
-    /* FACS is special... */
-    if (sig == FACS_SIGNATURE) {
-        struct fadt_descriptor_rev1 *fadt;
-        fadt = find_acpi_table_addr(FACP_SIGNATURE);
-        if (!fadt) {
-            return NULL;
-        }
-        return (void*)(ulong)fadt->firmware_ctrl;
-    }
+	/* FACS is special... */
+	if (sig == FACS_SIGNATURE) {
+		struct fadt_descriptor_rev1 *fadt;
+		fadt = find_acpi_table_addr(FACP_SIGNATURE);
+		if (!fadt)
+			return NULL;
 
-    rsdp = get_rsdp();
-    if (rsdp == NULL) {
-        printf("Can't find RSDP\n");
-        return 0;
-    }
+		return (void *)(ulong) fadt->firmware_ctrl;
+	}
 
-    if (sig == RSDP_SIGNATURE) {
-        return rsdp;
-    }
+	rsdp = get_rsdp();
+	if (rsdp == NULL) {
+		printf("Can't find RSDP\n");
+		return NULL;
+	}
 
-    rsdt = (void*)(ulong)rsdp->rsdt_physical_address;
-    if (!rsdt || rsdt->signature != RSDT_SIGNATURE)
-        return 0;
+	if (sig == RSDP_SIGNATURE)
+		return rsdp;
 
-    if (sig == RSDT_SIGNATURE) {
-        return rsdt;
-    }
+	rsdt = (void *)(ulong) rsdp->rsdt_physical_address;
+	if (!rsdt || rsdt->signature != RSDT_SIGNATURE)
+		return NULL;
+
+	if (sig == RSDT_SIGNATURE)
+		return rsdt;
 
-    end = (void*)rsdt + rsdt->length;
-    for (i=0; (void*)&rsdt->table_offset_entry[i] < end; i++) {
-        struct acpi_table *t = (void*)(ulong)rsdt->table_offset_entry[i];
-        if (t && t->signature == sig) {
-            return t;
-        }
-    }
-   return NULL;
+	end = (void *)rsdt + rsdt->length;
+	for (i = 0; (void *)&rsdt->table_offset_entry[i] < end; i++) {
+		struct acpi_table *t = (void *)(ulong) rsdt->table_offset_entry[i];
+		if (t && t->signature == sig) {
+			return t;
+		}
+	}
+	return NULL;
 }
diff --git a/lib/acpi.h b/lib/acpi.h
index 1e89840c..b67bbe19 100644
--- a/lib/acpi.h
+++ b/lib/acpi.h
@@ -3,7 +3,7 @@
 
 #include "libcflat.h"
 
-#define ACPI_SIGNATURE(c1, c2, c3, c4) \
+#define ACPI_SIGNATURE(c1, c2, c3, c4)				\
 	((c1) | ((c2) << 8) | ((c3) << 16) | ((c4) << 24))
 
 #define RSDP_SIGNATURE ACPI_SIGNATURE('R','S','D','P')
@@ -11,102 +11,99 @@
 #define FACP_SIGNATURE ACPI_SIGNATURE('F','A','C','P')
 #define FACS_SIGNATURE ACPI_SIGNATURE('F','A','C','S')
 
-
-#define ACPI_SIGNATURE_8BYTE(c1, c2, c3, c4, c5, c6, c7, c8) \
-	((uint64_t)(ACPI_SIGNATURE(c1, c2, c3, c4))) |       \
+#define ACPI_SIGNATURE_8BYTE(c1, c2, c3, c4, c5, c6, c7, c8)	\
+	((uint64_t)(ACPI_SIGNATURE(c1, c2, c3, c4))) |		\
 	((uint64_t)(ACPI_SIGNATURE(c5, c6, c7, c8)) << 32)
 
 #define RSDP_SIGNATURE_8BYTE (ACPI_SIGNATURE_8BYTE('R', 'S', 'D', ' ', 'P', 'T', 'R', ' '))
 
-struct rsdp_descriptor {        /* Root System Descriptor Pointer */
-    u64 signature;              /* ACPI signature, contains "RSD PTR " */
-    u8  checksum;               /* To make sum of struct == 0 */
-    u8  oem_id [6];             /* OEM identification */
-    u8  revision;               /* Must be 0 for 1.0, 2 for 2.0 */
-    u32 rsdt_physical_address;  /* 32-bit physical address of RSDT */
-    u32 length;                 /* XSDT Length in bytes including hdr */
-    u64 xsdt_physical_address;  /* 64-bit physical address of XSDT */
-    u8  extended_checksum;      /* Checksum of entire table */
-    u8  reserved [3];           /* Reserved field must be 0 */
+struct rsdp_descriptor {	/* Root System Descriptor Pointer */
+	u64 signature;		/* ACPI signature, contains "RSD PTR " */
+	u8 checksum;		/* To make sum of struct == 0 */
+	u8 oem_id[6];		/* OEM identification */
+	u8 revision;		/* Must be 0 for 1.0, 2 for 2.0 */
+	u32 rsdt_physical_address;	/* 32-bit physical address of RSDT */
+	u32 length;		/* XSDT Length in bytes including hdr */
+	u64 xsdt_physical_address;	/* 64-bit physical address of XSDT */
+	u8 extended_checksum;	/* Checksum of entire table */
+	u8 reserved[3];		/* Reserved field must be 0 */
 };
 
-#define ACPI_TABLE_HEADER_DEF   /* ACPI common table header */ \
-    u32 signature;          /* ACPI signature (4 ASCII characters) */ \
-    u32 length;                 /* Length of table, in bytes, including header */ \
-    u8  revision;               /* ACPI Specification minor version # */ \
-    u8  checksum;               /* To make sum of entire table == 0 */ \
-    u8  oem_id [6];             /* OEM identification */ \
-    u8  oem_table_id [8];       /* OEM table identification */ \
-    u32 oem_revision;           /* OEM revision number */ \
-    u8  asl_compiler_id [4];    /* ASL compiler vendor ID */ \
-    u32 asl_compiler_revision;  /* ASL compiler revision number */
+#define ACPI_TABLE_HEADER_DEF		/* ACPI common table header */			\
+	u32 signature;			/* ACPI signature (4 ASCII characters) */	\
+	u32 length;			/* Length of table, in bytes, including header */ \
+	u8  revision;			/* ACPI Specification minor version # */	\
+	u8  checksum;			/* To make sum of entire table == 0 */		\
+	u8  oem_id [6];			/* OEM identification */			\
+	u8  oem_table_id [8];		/* OEM table identification */			\
+	u32 oem_revision;		/* OEM revision number */			\
+	u8  asl_compiler_id [4];	/* ASL compiler vendor ID */			\
+	u32 asl_compiler_revision;	/* ASL compiler revision number */
 
 struct acpi_table {
-    ACPI_TABLE_HEADER_DEF
-    char data[0];
+	ACPI_TABLE_HEADER_DEF
+	char data[0];
 };
 
 struct rsdt_descriptor_rev1 {
-    ACPI_TABLE_HEADER_DEF
-    u32 table_offset_entry[0];
+	ACPI_TABLE_HEADER_DEF
+	u32 table_offset_entry[1];
 };
 
-struct fadt_descriptor_rev1
-{
-    ACPI_TABLE_HEADER_DEF     /* ACPI common table header */
-    u32 firmware_ctrl;          /* Physical address of FACS */
-    u32 dsdt;                   /* Physical address of DSDT */
-    u8  model;                  /* System Interrupt Model */
-    u8  reserved1;              /* Reserved */
-    u16 sci_int;                /* System vector of SCI interrupt */
-    u32 smi_cmd;                /* Port address of SMI command port */
-    u8  acpi_enable;            /* Value to write to smi_cmd to enable ACPI */
-    u8  acpi_disable;           /* Value to write to smi_cmd to disable ACPI */
-    u8  S4bios_req;             /* Value to write to SMI CMD to enter S4BIOS state */
-    u8  reserved2;              /* Reserved - must be zero */
-    u32 pm1a_evt_blk;           /* Port address of Power Mgt 1a acpi_event Reg Blk */
-    u32 pm1b_evt_blk;           /* Port address of Power Mgt 1b acpi_event Reg Blk */
-    u32 pm1a_cnt_blk;           /* Port address of Power Mgt 1a Control Reg Blk */
-    u32 pm1b_cnt_blk;           /* Port address of Power Mgt 1b Control Reg Blk */
-    u32 pm2_cnt_blk;            /* Port address of Power Mgt 2 Control Reg Blk */
-    u32 pm_tmr_blk;             /* Port address of Power Mgt Timer Ctrl Reg Blk */
-    u32 gpe0_blk;               /* Port addr of General Purpose acpi_event 0 Reg Blk */
-    u32 gpe1_blk;               /* Port addr of General Purpose acpi_event 1 Reg Blk */
-    u8  pm1_evt_len;            /* Byte length of ports at pm1_x_evt_blk */
-    u8  pm1_cnt_len;            /* Byte length of ports at pm1_x_cnt_blk */
-    u8  pm2_cnt_len;            /* Byte Length of ports at pm2_cnt_blk */
-    u8  pm_tmr_len;             /* Byte Length of ports at pm_tm_blk */
-    u8  gpe0_blk_len;           /* Byte Length of ports at gpe0_blk */
-    u8  gpe1_blk_len;           /* Byte Length of ports at gpe1_blk */
-    u8  gpe1_base;              /* Offset in gpe model where gpe1 events start */
-    u8  reserved3;              /* Reserved */
-    u16 plvl2_lat;              /* Worst case HW latency to enter/exit C2 state */
-    u16 plvl3_lat;              /* Worst case HW latency to enter/exit C3 state */
-    u16 flush_size;             /* Size of area read to flush caches */
-    u16 flush_stride;           /* Stride used in flushing caches */
-    u8  duty_offset;            /* Bit location of duty cycle field in p_cnt reg */
-    u8  duty_width;             /* Bit width of duty cycle field in p_cnt reg */
-    u8  day_alrm;               /* Index to day-of-month alarm in RTC CMOS RAM */
-    u8  mon_alrm;               /* Index to month-of-year alarm in RTC CMOS RAM */
-    u8  century;                /* Index to century in RTC CMOS RAM */
-    u8  reserved4;              /* Reserved */
-    u8  reserved4a;             /* Reserved */
-    u8  reserved4b;             /* Reserved */
+struct fadt_descriptor_rev1 {
+	ACPI_TABLE_HEADER_DEF	/* ACPI common table header */
+	u32 firmware_ctrl;	/* Physical address of FACS */
+	u32 dsdt;		/* Physical address of DSDT */
+	u8 model;		/* System Interrupt Model */
+	u8 reserved1;		/* Reserved */
+	u16 sci_int;		/* System vector of SCI interrupt */
+	u32 smi_cmd;		/* Port address of SMI command port */
+	u8 acpi_enable;		/* Value to write to smi_cmd to enable ACPI */
+	u8 acpi_disable;	/* Value to write to smi_cmd to disable ACPI */
+	u8 S4bios_req;		/* Value to write to SMI CMD to enter S4BIOS state */
+	u8 reserved2;		/* Reserved - must be zero */
+	u32 pm1a_evt_blk;	/* Port address of Power Mgt 1a acpi_event Reg Blk */
+	u32 pm1b_evt_blk;	/* Port address of Power Mgt 1b acpi_event Reg Blk */
+	u32 pm1a_cnt_blk;	/* Port address of Power Mgt 1a Control Reg Blk */
+	u32 pm1b_cnt_blk;	/* Port address of Power Mgt 1b Control Reg Blk */
+	u32 pm2_cnt_blk;	/* Port address of Power Mgt 2 Control Reg Blk */
+	u32 pm_tmr_blk;		/* Port address of Power Mgt Timer Ctrl Reg Blk */
+	u32 gpe0_blk;		/* Port addr of General Purpose acpi_event 0 Reg Blk */
+	u32 gpe1_blk;		/* Port addr of General Purpose acpi_event 1 Reg Blk */
+	u8 pm1_evt_len;		/* Byte length of ports at pm1_x_evt_blk */
+	u8 pm1_cnt_len;		/* Byte length of ports at pm1_x_cnt_blk */
+	u8 pm2_cnt_len;		/* Byte Length of ports at pm2_cnt_blk */
+	u8 pm_tmr_len;		/* Byte Length of ports at pm_tm_blk */
+	u8 gpe0_blk_len;	/* Byte Length of ports at gpe0_blk */
+	u8 gpe1_blk_len;	/* Byte Length of ports at gpe1_blk */
+	u8 gpe1_base;		/* Offset in gpe model where gpe1 events start */
+	u8 reserved3;		/* Reserved */
+	u16 plvl2_lat;		/* Worst case HW latency to enter/exit C2 state */
+	u16 plvl3_lat;		/* Worst case HW latency to enter/exit C3 state */
+	u16 flush_size;		/* Size of area read to flush caches */
+	u16 flush_stride;	/* Stride used in flushing caches */
+	u8 duty_offset;		/* Bit location of duty cycle field in p_cnt reg */
+	u8 duty_width;		/* Bit width of duty cycle field in p_cnt reg */
+	u8 day_alrm;		/* Index to day-of-month alarm in RTC CMOS RAM */
+	u8 mon_alrm;		/* Index to month-of-year alarm in RTC CMOS RAM */
+	u8 century;		/* Index to century in RTC CMOS RAM */
+	u8 reserved4;		/* Reserved */
+	u8 reserved4a;		/* Reserved */
+	u8 reserved4b;		/* Reserved */
 };
 
-struct facs_descriptor_rev1
-{
-    u32 signature;           /* ACPI Signature */
-    u32 length;                 /* Length of structure, in bytes */
-    u32 hardware_signature;     /* Hardware configuration signature */
-    u32 firmware_waking_vector; /* ACPI OS waking vector */
-    u32 global_lock;            /* Global Lock */
-    u32 S4bios_f        : 1;    /* Indicates if S4BIOS support is present */
-    u32 reserved1       : 31;   /* Must be 0 */
-    u8  reserved3 [40];         /* Reserved - must be zero */
+struct facs_descriptor_rev1 {
+	u32 signature;		/* ACPI Signature */
+	u32 length;		/* Length of structure, in bytes */
+	u32 hardware_signature;	/* Hardware configuration signature */
+	u32 firmware_waking_vector;	/* ACPI OS waking vector */
+	u32 global_lock;	/* Global Lock */
+	u32 S4bios_f:1;		/* Indicates if S4BIOS support is present */
+	u32 reserved1:31;	/* Must be 0 */
+	u8 reserved3[40];	/* Reserved - must be zero */
 };
 
 void set_efi_rsdp(struct rsdp_descriptor *rsdp);
-void* find_acpi_table_addr(u32 sig);
+void *find_acpi_table_addr(u32 sig);
 
 #endif
-- 
2.25.1

[PATCH v4 05/30] lib: Fix style for acpi.{c,h}

[Nikos Nikoleris]

Manually fix style issues to make the files consistent with the kernel
coding style.

Zero-length array members have been replaced with flexible array members
(for details about the motivation, consult
Documentation/process/deprecated.rst in the Linux tree, the section about
zero-length and one-element arrays).

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
[ Alex E: changes other than indentation ]
---
 lib/acpi.c | 16 ++++++++--------
 lib/acpi.h | 12 ++++++------
 2 files changed, 14 insertions(+), 14 deletions(-)

diff --git a/lib/acpi.c b/lib/acpi.c
index 836156a1..3f87711a 100644
--- a/lib/acpi.c
+++ b/lib/acpi.c
@@ -11,9 +11,9 @@ void set_efi_rsdp(struct rsdp_descriptor *rsdp)
 
 static struct rsdp_descriptor *get_rsdp(void)
 {
-	if (efi_rsdp == NULL) {
+	if (efi_rsdp == NULL)
 		printf("Can't find RSDP from UEFI, maybe set_efi_rsdp() was not called\n");
-	}
+
 	return efi_rsdp;
 }
 #else
@@ -28,9 +28,8 @@ static struct rsdp_descriptor *get_rsdp(void)
 			break;
 	}
 
-	if (addr == 0x100000) {
+	if (addr == 0x100000)
 		return NULL;
-	}
 
 	return rsdp;
 }
@@ -38,18 +37,18 @@ static struct rsdp_descriptor *get_rsdp(void)
 
 void *find_acpi_table_addr(u32 sig)
 {
-	struct rsdp_descriptor *rsdp;
 	struct rsdt_descriptor_rev1 *rsdt;
+	struct rsdp_descriptor *rsdp;
 	void *end;
 	int i;
 
 	/* FACS is special... */
 	if (sig == FACS_SIGNATURE) {
 		struct fadt_descriptor_rev1 *fadt;
+
 		fadt = find_acpi_table_addr(FACP_SIGNATURE);
 		if (!fadt)
 			return NULL;
-
 		return (void *)(ulong) fadt->firmware_ctrl;
 	}
 
@@ -72,9 +71,10 @@ void *find_acpi_table_addr(u32 sig)
 	end = (void *)rsdt + rsdt->length;
 	for (i = 0; (void *)&rsdt->table_offset_entry[i] < end; i++) {
 		struct acpi_table *t = (void *)(ulong) rsdt->table_offset_entry[i];
-		if (t && t->signature == sig) {
+
+		if (t && t->signature == sig)
 			return t;
-		}
 	}
+
 	return NULL;
 }
diff --git a/lib/acpi.h b/lib/acpi.h
index b67bbe19..fad28792 100644
--- a/lib/acpi.h
+++ b/lib/acpi.h
@@ -3,7 +3,7 @@
 
 #include "libcflat.h"
 
-#define ACPI_SIGNATURE(c1, c2, c3, c4)				\
+#define ACPI_SIGNATURE(c1, c2, c3, c4) \
 	((c1) | ((c2) << 8) | ((c3) << 16) | ((c4) << 24))
 
 #define RSDP_SIGNATURE ACPI_SIGNATURE('R','S','D','P')
@@ -11,9 +11,9 @@
 #define FACP_SIGNATURE ACPI_SIGNATURE('F','A','C','P')
 #define FACS_SIGNATURE ACPI_SIGNATURE('F','A','C','S')
 
-#define ACPI_SIGNATURE_8BYTE(c1, c2, c3, c4, c5, c6, c7, c8)	\
-	((uint64_t)(ACPI_SIGNATURE(c1, c2, c3, c4))) |		\
-	((uint64_t)(ACPI_SIGNATURE(c5, c6, c7, c8)) << 32)
+#define ACPI_SIGNATURE_8BYTE(c1, c2, c3, c4, c5, c6, c7, c8) \
+	(((uint64_t)(ACPI_SIGNATURE(c1, c2, c3, c4))) |	     \
+	 ((uint64_t)(ACPI_SIGNATURE(c5, c6, c7, c8)) << 32))
 
 #define RSDP_SIGNATURE_8BYTE (ACPI_SIGNATURE_8BYTE('R', 'S', 'D', ' ', 'P', 'T', 'R', ' '))
 
@@ -42,12 +42,12 @@ struct rsdp_descriptor {	/* Root System Descriptor Pointer */
 
 struct acpi_table {
 	ACPI_TABLE_HEADER_DEF
-	char data[0];
+	char data[];
 };
 
 struct rsdt_descriptor_rev1 {
 	ACPI_TABLE_HEADER_DEF
-	u32 table_offset_entry[1];
+	u32 table_offset_entry[];
 };
 
 struct fadt_descriptor_rev1 {
-- 
2.25.1

[PATCH v4 06/30] lib/acpi: Convert table names to Linux style

[Nikos Nikoleris]

From: Alexandru Elisei <alexandru.elisei@arm.com>

kvm-unit-tests is about to import several table struct definitions from
Linux, convert the names of the existing tables to follow the Linux style.

This is purely a cosmetic change and no functional change is intended.

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
---
 lib/acpi.c      | 16 ++++++++--------
 lib/acpi.h      | 10 +++++-----
 lib/x86/setup.c |  2 +-
 x86/s3.c        |  4 ++--
 x86/vmexit.c    |  2 +-
 5 files changed, 17 insertions(+), 17 deletions(-)

diff --git a/lib/acpi.c b/lib/acpi.c
index 3f87711a..166ffd14 100644
--- a/lib/acpi.c
+++ b/lib/acpi.c
@@ -2,14 +2,14 @@
 #include "acpi.h"
 
 #ifdef CONFIG_EFI
-struct rsdp_descriptor *efi_rsdp = NULL;
+struct acpi_table_rsdp *efi_rsdp = NULL;
 
-void set_efi_rsdp(struct rsdp_descriptor *rsdp)
+void set_efi_rsdp(struct acpi_table_rsdp *rsdp)
 {
 	efi_rsdp = rsdp;
 }
 
-static struct rsdp_descriptor *get_rsdp(void)
+static struct acpi_table_rsdp *get_rsdp(void)
 {
 	if (efi_rsdp == NULL)
 		printf("Can't find RSDP from UEFI, maybe set_efi_rsdp() was not called\n");
@@ -17,9 +17,9 @@ static struct rsdp_descriptor *get_rsdp(void)
 	return efi_rsdp;
 }
 #else
-static struct rsdp_descriptor *get_rsdp(void)
+static struct acpi_table_rsdp *get_rsdp(void)
 {
-	struct rsdp_descriptor *rsdp;
+	struct acpi_table_rsdp *rsdp;
 	unsigned long addr;
 
 	for (addr = 0xe0000; addr < 0x100000; addr += 16) {
@@ -37,14 +37,14 @@ static struct rsdp_descriptor *get_rsdp(void)
 
 void *find_acpi_table_addr(u32 sig)
 {
-	struct rsdt_descriptor_rev1 *rsdt;
-	struct rsdp_descriptor *rsdp;
+	struct acpi_table_rsdt_rev1 *rsdt;
+	struct acpi_table_rsdp *rsdp;
 	void *end;
 	int i;
 
 	/* FACS is special... */
 	if (sig == FACS_SIGNATURE) {
-		struct fadt_descriptor_rev1 *fadt;
+		struct acpi_table_fadt_rev1 *fadt;
 
 		fadt = find_acpi_table_addr(FACP_SIGNATURE);
 		if (!fadt)
diff --git a/lib/acpi.h b/lib/acpi.h
index fad28792..229a6391 100644
--- a/lib/acpi.h
+++ b/lib/acpi.h
@@ -17,7 +17,7 @@
 
 #define RSDP_SIGNATURE_8BYTE (ACPI_SIGNATURE_8BYTE('R', 'S', 'D', ' ', 'P', 'T', 'R', ' '))
 
-struct rsdp_descriptor {	/* Root System Descriptor Pointer */
+struct acpi_table_rsdp {	/* Root System Descriptor Pointer */
 	u64 signature;		/* ACPI signature, contains "RSD PTR " */
 	u8 checksum;		/* To make sum of struct == 0 */
 	u8 oem_id[6];		/* OEM identification */
@@ -45,12 +45,12 @@ struct acpi_table {
 	char data[];
 };
 
-struct rsdt_descriptor_rev1 {
+struct acpi_table_rsdt_rev1 {
 	ACPI_TABLE_HEADER_DEF
 	u32 table_offset_entry[];
 };
 
-struct fadt_descriptor_rev1 {
+struct acpi_table_fadt_rev1 {
 	ACPI_TABLE_HEADER_DEF	/* ACPI common table header */
 	u32 firmware_ctrl;	/* Physical address of FACS */
 	u32 dsdt;		/* Physical address of DSDT */
@@ -92,7 +92,7 @@ struct fadt_descriptor_rev1 {
 	u8 reserved4b;		/* Reserved */
 };
 
-struct facs_descriptor_rev1 {
+struct acpi_table_facs_rev1 {
 	u32 signature;		/* ACPI Signature */
 	u32 length;		/* Length of structure, in bytes */
 	u32 hardware_signature;	/* Hardware configuration signature */
@@ -103,7 +103,7 @@ struct facs_descriptor_rev1 {
 	u8 reserved3[40];	/* Reserved - must be zero */
 };
 
-void set_efi_rsdp(struct rsdp_descriptor *rsdp);
+void set_efi_rsdp(struct acpi_table_rsdp *rsdp);
 void *find_acpi_table_addr(u32 sig);
 
 #endif
diff --git a/lib/x86/setup.c b/lib/x86/setup.c
index dd150300..d509a248 100644
--- a/lib/x86/setup.c
+++ b/lib/x86/setup.c
@@ -245,7 +245,7 @@ static efi_status_t setup_memory_allocator(efi_bootinfo_t *efi_bootinfo)
 static efi_status_t setup_rsdp(efi_bootinfo_t *efi_bootinfo)
 {
 	efi_status_t status;
-	struct rsdp_descriptor *rsdp;
+	struct acpi_table_rsdp *rsdp;
 
 	/*
 	 * RSDP resides in an EFI_ACPI_RECLAIM_MEMORY region, which is not used
diff --git a/x86/s3.c b/x86/s3.c
index 378d37ae..96db728c 100644
--- a/x86/s3.c
+++ b/x86/s3.c
@@ -4,7 +4,7 @@
 
 static u32* find_resume_vector_addr(void)
 {
-    struct facs_descriptor_rev1 *facs = find_acpi_table_addr(FACS_SIGNATURE);
+    struct acpi_table_facs_rev1 *facs = find_acpi_table_addr(FACS_SIGNATURE);
     if (!facs)
         return 0;
     printf("FACS is at %p\n", facs);
@@ -39,7 +39,7 @@ extern char resume_start, resume_end;
 
 int main(int argc, char **argv)
 {
-	struct fadt_descriptor_rev1 *fadt = find_acpi_table_addr(FACP_SIGNATURE);
+	struct acpi_table_fadt_rev1 *fadt = find_acpi_table_addr(FACP_SIGNATURE);
 	volatile u32 *resume_vector_ptr = find_resume_vector_addr();
 	char *addr, *resume_vec = (void*)0x1000;
 
diff --git a/x86/vmexit.c b/x86/vmexit.c
index 865ccf6b..6987aca0 100644
--- a/x86/vmexit.c
+++ b/x86/vmexit.c
@@ -206,7 +206,7 @@ int pm_tmr_blk;
 static void inl_pmtimer(void)
 {
     if (!pm_tmr_blk) {
-	struct fadt_descriptor_rev1 *fadt;
+	struct acpi_table_fadt_rev1 *fadt;
 
 	fadt = find_acpi_table_addr(FACP_SIGNATURE);
 	pm_tmr_blk = fadt->pm_tmr_blk;
-- 
2.25.1

[PATCH v4 07/30] x86: Avoid references to fields of ACPI tables

[Nikos Nikoleris]

ACPI table definitions in <acpi.h> have to be packed.  However, once
we do that, direct references to members of the packed struct might
result in unaligned pointers and gcc complains about them. This change
modifies the code to avoid such references in preparation for making
the APCI table definitions packed.

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
---
 x86/s3.c | 17 +++++------------
 1 file changed, 5 insertions(+), 12 deletions(-)

diff --git a/x86/s3.c b/x86/s3.c
index 96db728c..910c57fb 100644
--- a/x86/s3.c
+++ b/x86/s3.c
@@ -2,15 +2,6 @@
 #include "acpi.h"
 #include "asm/io.h"
 
-static u32* find_resume_vector_addr(void)
-{
-    struct acpi_table_facs_rev1 *facs = find_acpi_table_addr(FACS_SIGNATURE);
-    if (!facs)
-        return 0;
-    printf("FACS is at %p\n", facs);
-    return &facs->firmware_waking_vector;
-}
-
 #define RTC_SECONDS_ALARM       1
 #define RTC_MINUTES_ALARM       3
 #define RTC_HOURS_ALARM         5
@@ -40,12 +31,14 @@ extern char resume_start, resume_end;
 int main(int argc, char **argv)
 {
 	struct acpi_table_fadt_rev1 *fadt = find_acpi_table_addr(FACP_SIGNATURE);
-	volatile u32 *resume_vector_ptr = find_resume_vector_addr();
+	struct acpi_table_facs_rev1 *facs = find_acpi_table_addr(FACS_SIGNATURE);
 	char *addr, *resume_vec = (void*)0x1000;
 
-	*resume_vector_ptr = (u32)(ulong)resume_vec;
+	assert(facs);
+	facs->firmware_waking_vector = (u32)(ulong)resume_vec;
 
-	printf("resume vector addr is %p\n", resume_vector_ptr);
+	printf("FACS is at %p\n", facs);
+	printf("resume vector addr is %p\n", &facs->firmware_waking_vector);
 	for (addr = &resume_start; addr < &resume_end; addr++)
 		*resume_vec++ = *addr;
 	printf("copy resume code from %p\n", &resume_start);
-- 
2.25.1

[PATCH v4 08/30] lib/acpi: Ensure all struct definition for ACPI tables are packed

[Nikos Nikoleris]

All ACPI table definitions are provided with precise definitions of
field sizes and offsets, make sure that no compiler optimization can
interfere with the memory layout of the corresponding structs.

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Andrew Jones <andrew.jones@linux.dev>
---
 lib/acpi.h | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/lib/acpi.h b/lib/acpi.h
index 229a6391..67d12942 100644
--- a/lib/acpi.h
+++ b/lib/acpi.h
@@ -3,6 +3,12 @@
 
 #include "libcflat.h"
 
+/*
+ * All tables and structures must be byte-packed to match the ACPI
+ * specification, since the tables are provided by the system firmware.
+ */
+#pragma pack(1)
+
 #define ACPI_SIGNATURE(c1, c2, c3, c4) \
 	((c1) | ((c2) << 8) | ((c3) << 16) | ((c4) << 24))
 
@@ -103,6 +109,8 @@ struct acpi_table_facs_rev1 {
 	u8 reserved3[40];	/* Reserved - must be zero */
 };
 
+#pragma pack(0)
+
 void set_efi_rsdp(struct acpi_table_rsdp *rsdp);
 void *find_acpi_table_addr(u32 sig);
 
-- 
2.25.1

[PATCH v4 09/30] lib/acpi: Add support for the XSDT table

[Nikos Nikoleris]

XSDT provides pointers to other ACPI tables much like RSDT. However,
contrary to RSDT that provides 32-bit addresses, XSDT provides 64-bit
pointers. ACPI requires that if XSDT is valid then it takes precedence
over RSDT.

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Andrew Jones <andrew.jones@linux.dev>
[ Alex E: Use flexible array member for XSDT struct ]
Signed-off-by: Alexandru Elisei <alexandru.elisei@arm.com>
---
 lib/acpi.c | 40 ++++++++++++++++++++++++++++++++--------
 lib/acpi.h |  6 ++++++
 2 files changed, 38 insertions(+), 8 deletions(-)

diff --git a/lib/acpi.c b/lib/acpi.c
index 166ffd14..d35f09a6 100644
--- a/lib/acpi.c
+++ b/lib/acpi.c
@@ -37,7 +37,8 @@ static struct acpi_table_rsdp *get_rsdp(void)
 
 void *find_acpi_table_addr(u32 sig)
 {
-	struct acpi_table_rsdt_rev1 *rsdt;
+	struct acpi_table_rsdt_rev1 *rsdt = NULL;
+	struct acpi_table_xsdt *xsdt = NULL;
 	struct acpi_table_rsdp *rsdp;
 	void *end;
 	int i;
@@ -62,18 +63,41 @@ void *find_acpi_table_addr(u32 sig)
 		return rsdp;
 
 	rsdt = (void *)(ulong) rsdp->rsdt_physical_address;
-	if (!rsdt || rsdt->signature != RSDT_SIGNATURE)
-		return NULL;
+	if (rsdt && rsdt->signature != RSDT_SIGNATURE)
+		rsdt = NULL;
 
 	if (sig == RSDT_SIGNATURE)
 		return rsdt;
 
-	end = (void *)rsdt + rsdt->length;
-	for (i = 0; (void *)&rsdt->table_offset_entry[i] < end; i++) {
-		struct acpi_table *t = (void *)(ulong) rsdt->table_offset_entry[i];
+	if (rsdp->revision >= 2) {
+		xsdt = (void *)rsdp->xsdt_physical_address;
+		if (xsdt && xsdt->signature != XSDT_SIGNATURE)
+			xsdt = NULL;
+	}
 
-		if (t && t->signature == sig)
-			return t;
+	if (sig == XSDT_SIGNATURE)
+		return xsdt;
+
+	/*
+	 * When the system implements APCI 2.0 and above and XSDT is valid we
+	 * have use XSDT to find other ACPI tables, otherwise, we use RSDT.
+	 */
+	if (xsdt) {
+		end = (void *)xsdt + xsdt->length;
+		for (i = 0; (void *)&xsdt->table_offset_entry[i] < end; i++) {
+			struct acpi_table *t = (void *)(ulong) xsdt->table_offset_entry[i];
+
+			if (t && t->signature == sig)
+				return t;
+		}
+	} else if (rsdt) {
+		end = (void *)rsdt + rsdt->length;
+		for (i = 0; (void *)&rsdt->table_offset_entry[i] < end; i++) {
+			struct acpi_table *t = (void *)(ulong) rsdt->table_offset_entry[i];
+
+			if (t && t->signature == sig)
+				return t;
+		}
 	}
 
 	return NULL;
diff --git a/lib/acpi.h b/lib/acpi.h
index 67d12942..621ad08a 100644
--- a/lib/acpi.h
+++ b/lib/acpi.h
@@ -14,6 +14,7 @@
 
 #define RSDP_SIGNATURE ACPI_SIGNATURE('R','S','D','P')
 #define RSDT_SIGNATURE ACPI_SIGNATURE('R','S','D','T')
+#define XSDT_SIGNATURE ACPI_SIGNATURE('X','S','D','T')
 #define FACP_SIGNATURE ACPI_SIGNATURE('F','A','C','P')
 #define FACS_SIGNATURE ACPI_SIGNATURE('F','A','C','S')
 
@@ -56,6 +57,11 @@ struct acpi_table_rsdt_rev1 {
 	u32 table_offset_entry[];
 };
 
+struct acpi_table_xsdt {
+	ACPI_TABLE_HEADER_DEF
+	u64 table_offset_entry[];
+};
+
 struct acpi_table_fadt_rev1 {
 	ACPI_TABLE_HEADER_DEF	/* ACPI common table header */
 	u32 firmware_ctrl;	/* Physical address of FACS */
-- 
2.25.1

[PATCH v4 10/30] lib/acpi: Extend the definition of the FADT table

[Nikos Nikoleris]

This change add more fields in the APCI table FADT to allow for the
discovery of the PSCI conduit in arm64 systems. The definition for
FADT is similar to the one in include/acpi/actbl.h in Linux.

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Ricardo Koller <ricarkol@google.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
---
 lib/acpi.c   |  2 +-
 lib/acpi.h   | 33 +++++++++++++++++++++++++++++----
 x86/s3.c     |  2 +-
 x86/vmexit.c |  2 +-
 4 files changed, 32 insertions(+), 7 deletions(-)

diff --git a/lib/acpi.c b/lib/acpi.c
index d35f09a6..a197f3dd 100644
--- a/lib/acpi.c
+++ b/lib/acpi.c
@@ -45,7 +45,7 @@ void *find_acpi_table_addr(u32 sig)
 
 	/* FACS is special... */
 	if (sig == FACS_SIGNATURE) {
-		struct acpi_table_fadt_rev1 *fadt;
+		struct acpi_table_fadt *fadt;
 
 		fadt = find_acpi_table_addr(FACP_SIGNATURE);
 		if (!fadt)
diff --git a/lib/acpi.h b/lib/acpi.h
index 621ad08a..38b6d5c9 100644
--- a/lib/acpi.h
+++ b/lib/acpi.h
@@ -62,7 +62,15 @@ struct acpi_table_xsdt {
 	u64 table_offset_entry[];
 };
 
-struct acpi_table_fadt_rev1 {
+struct acpi_generic_address {
+	u8 space_id;		/* Address space where struct or register exists */
+	u8 bit_width;		/* Size in bits of given register */
+	u8 bit_offset;		/* Bit offset within the register */
+	u8 access_width;	/* Minimum Access size (ACPI 3.0) */
+	u64 address;		/* 64-bit address of struct or register */
+};
+
+struct acpi_table_fadt {
 	ACPI_TABLE_HEADER_DEF	/* ACPI common table header */
 	u32 firmware_ctrl;	/* Physical address of FACS */
 	u32 dsdt;		/* Physical address of DSDT */
@@ -99,9 +107,26 @@ struct acpi_table_fadt_rev1 {
 	u8 day_alrm;		/* Index to day-of-month alarm in RTC CMOS RAM */
 	u8 mon_alrm;		/* Index to month-of-year alarm in RTC CMOS RAM */
 	u8 century;		/* Index to century in RTC CMOS RAM */
-	u8 reserved4;		/* Reserved */
-	u8 reserved4a;		/* Reserved */
-	u8 reserved4b;		/* Reserved */
+	u16 boot_flags;		/* IA-PC Boot Architecture Flags (see below for individual flags) */
+	u8 reserved;		/* Reserved, must be zero */
+	u32 flags;		/* Miscellaneous flag bits (see below for individual flags) */
+	struct acpi_generic_address reset_register;	/* 64-bit address of the Reset register */
+	u8 reset_value;		/* Value to write to the reset_register port to reset the system */
+	u16 arm_boot_flags;	/* ARM-Specific Boot Flags (see below for individual flags) (ACPI 5.1) */
+	u8 minor_revision;	/* FADT Minor Revision (ACPI 5.1) */
+	u64 Xfacs;		/* 64-bit physical address of FACS */
+	u64 Xdsdt;		/* 64-bit physical address of DSDT */
+	struct acpi_generic_address xpm1a_event_block;	/* 64-bit Extended Power Mgt 1a Event Reg Blk address */
+	struct acpi_generic_address xpm1b_event_block;	/* 64-bit Extended Power Mgt 1b Event Reg Blk address */
+	struct acpi_generic_address xpm1a_control_block;	/* 64-bit Extended Power Mgt 1a Control Reg Blk address */
+	struct acpi_generic_address xpm1b_control_block;	/* 64-bit Extended Power Mgt 1b Control Reg Blk address */
+	struct acpi_generic_address xpm2_control_block;	/* 64-bit Extended Power Mgt 2 Control Reg Blk address */
+	struct acpi_generic_address xpm_timer_block;	/* 64-bit Extended Power Mgt Timer Ctrl Reg Blk address */
+	struct acpi_generic_address xgpe0_block;	/* 64-bit Extended General Purpose Event 0 Reg Blk address */
+	struct acpi_generic_address xgpe1_block;	/* 64-bit Extended General Purpose Event 1 Reg Blk address */
+	struct acpi_generic_address sleep_control;	/* 64-bit Sleep Control register (ACPI 5.0) */
+	struct acpi_generic_address sleep_status;	/* 64-bit Sleep Status register (ACPI 5.0) */
+	u64 hypervisor_id;	/* Hypervisor Vendor ID (ACPI 6.0) */
 };
 
 struct acpi_table_facs_rev1 {
diff --git a/x86/s3.c b/x86/s3.c
index 910c57fb..6f2d6d10 100644
--- a/x86/s3.c
+++ b/x86/s3.c
@@ -30,8 +30,8 @@ extern char resume_start, resume_end;
 
 int main(int argc, char **argv)
 {
-	struct acpi_table_fadt_rev1 *fadt = find_acpi_table_addr(FACP_SIGNATURE);
 	struct acpi_table_facs_rev1 *facs = find_acpi_table_addr(FACS_SIGNATURE);
+	struct acpi_table_fadt *fadt = find_acpi_table_addr(FACP_SIGNATURE);
 	char *addr, *resume_vec = (void*)0x1000;
 
 	assert(facs);
diff --git a/x86/vmexit.c b/x86/vmexit.c
index 6987aca0..3920f441 100644
--- a/x86/vmexit.c
+++ b/x86/vmexit.c
@@ -206,7 +206,7 @@ int pm_tmr_blk;
 static void inl_pmtimer(void)
 {
     if (!pm_tmr_blk) {
-	struct acpi_table_fadt_rev1 *fadt;
+	struct acpi_table_fadt *fadt;
 
 	fadt = find_acpi_table_addr(FACP_SIGNATURE);
 	pm_tmr_blk = fadt->pm_tmr_blk;
-- 
2.25.1

[PATCH v4 11/30] devicetree: Check that fdt is not NULL in dt_available()

[Nikos Nikoleris]

Up until now, for platforms that support DT, kvm-unit-tests would
unconditionally use DT to configure the system and the code made the
assumption that the fdt will always be a valid pointer.

On Arm systems that support both ACPI and DT, kvm-unit-tests plans to
follow the same convension as the Linux kernel: attempt to configure the
system using the DT first, then fallback to ACPI.

As a result, the fdt pointer can be NULL. Check for that in dt_available().

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Andrew Jones <andrew.jones@linux.dev>
[ Alex E: Minor changes to the commit message ]
---
 lib/devicetree.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lib/devicetree.c b/lib/devicetree.c
index 78c1f6fb..3ff9d164 100644
--- a/lib/devicetree.c
+++ b/lib/devicetree.c
@@ -16,7 +16,7 @@ const void *dt_fdt(void)
 
 bool dt_available(void)
 {
-	return fdt_check_header(fdt) == 0;
+	return fdt && fdt_check_header(fdt) == 0;
 }
 
 int dt_get_nr_cells(int fdtnode, u32 *nr_address_cells, u32 *nr_size_cells)
-- 
2.25.1

[PATCH v4 12/30] arm64: Add support for setting up the PSCI conduit through ACPI

[Nikos Nikoleris]

In systems with ACPI support and when a DT is not provided, we can use
the FADT to discover whether PSCI calls need to use smc or hvc
calls. This change implements this but retains the default behavior;
we check if a valid DT is provided, if not, we try to setup the PSCI
conduit using ACPI.

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Ricardo Koller <ricarkol@google.com>
---
 arm/Makefile.arm64 |  4 ++++
 lib/acpi.h         |  5 +++++
 lib/arm/psci.c     | 37 ++++++++++++++++++++++++++++++++++++-
 3 files changed, 45 insertions(+), 1 deletion(-)

diff --git a/arm/Makefile.arm64 b/arm/Makefile.arm64
index 42e18e77..6dff6cad 100644
--- a/arm/Makefile.arm64
+++ b/arm/Makefile.arm64
@@ -25,6 +25,10 @@ cflatobjs += lib/arm64/processor.o
 cflatobjs += lib/arm64/spinlock.o
 cflatobjs += lib/arm64/gic-v3-its.o lib/arm64/gic-v3-its-cmd.o
 
+ifeq ($(CONFIG_EFI),y)
+cflatobjs += lib/acpi.o
+endif
+
 OBJDIRS += lib/arm64
 
 # arm64 specific tests
diff --git a/lib/acpi.h b/lib/acpi.h
index 38b6d5c9..5ba6e29d 100644
--- a/lib/acpi.h
+++ b/lib/acpi.h
@@ -129,6 +129,11 @@ struct acpi_table_fadt {
 	u64 hypervisor_id;	/* Hypervisor Vendor ID (ACPI 6.0) */
 };
 
+/* Masks for FADT ARM Boot Architecture Flags (arm_boot_flags) ACPI 5.1 */
+
+#define ACPI_FADT_PSCI_COMPLIANT    (1)	/* 00: [V5+] PSCI 0.2+ is implemented */
+#define ACPI_FADT_PSCI_USE_HVC      (1<<1)	/* 01: [V5+] HVC must be used instead of SMC as the PSCI conduit */
+
 struct acpi_table_facs_rev1 {
 	u32 signature;		/* ACPI Signature */
 	u32 length;		/* Length of structure, in bytes */
diff --git a/lib/arm/psci.c b/lib/arm/psci.c
index 9c031a12..bddb0787 100644
--- a/lib/arm/psci.c
+++ b/lib/arm/psci.c
@@ -56,7 +56,7 @@ void psci_system_off(void)
 	printf("CPU%d unable to do system off (error = %d)\n", smp_processor_id(), err);
 }
 
-void psci_set_conduit(void)
+static void psci_set_conduit_fdt(void)
 {
 	const void *fdt = dt_fdt();
 	const struct fdt_property *method;
@@ -75,3 +75,38 @@ void psci_set_conduit(void)
 	else
 		assert_msg(false, "Unknown PSCI conduit: %s", method->data);
 }
+
+#ifdef CONFIG_EFI
+
+#include <acpi.h>
+
+static void psci_set_conduit_acpi(void)
+{
+	struct acpi_table_fadt *fadt = find_acpi_table_addr(FACP_SIGNATURE);
+
+	assert_msg(fadt, "Unable to find ACPI FADT");
+	assert_msg(fadt->arm_boot_flags & ACPI_FADT_PSCI_COMPLIANT,
+		   "PSCI is not supported in this platform");
+
+	if (fadt->arm_boot_flags & ACPI_FADT_PSCI_USE_HVC)
+		psci_invoke = psci_invoke_hvc;
+	else
+		psci_invoke = psci_invoke_smc;
+}
+
+#else
+
+static void psci_set_conduit_acpi(void)
+{
+	assert_msg(false, "ACPI not available");
+}
+
+#endif
+
+void psci_set_conduit(void)
+{
+	if (dt_available())
+		psci_set_conduit_fdt();
+	else
+		psci_set_conduit_acpi();
+}
-- 
2.25.1

[PATCH v4 13/30] arm64: Add support for discovering the UART through ACPI

[Nikos Nikoleris]

In systems with ACPI support and when a DT is not provided, we can use
the SPCR to discover the serial port address range. This change
implements this but retains the default behavior; we check if a valid
DT is provided, if not, we try to discover the UART using ACPI.

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Ricardo Koller <ricarkol@google.com>
---
 lib/acpi.h   | 25 +++++++++++++++++++++++++
 lib/arm/io.c | 34 +++++++++++++++++++++++++++++-----
 2 files changed, 54 insertions(+), 5 deletions(-)

diff --git a/lib/acpi.h b/lib/acpi.h
index 5ba6e29d..02ca8958 100644
--- a/lib/acpi.h
+++ b/lib/acpi.h
@@ -17,6 +17,7 @@
 #define XSDT_SIGNATURE ACPI_SIGNATURE('X','S','D','T')
 #define FACP_SIGNATURE ACPI_SIGNATURE('F','A','C','P')
 #define FACS_SIGNATURE ACPI_SIGNATURE('F','A','C','S')
+#define SPCR_SIGNATURE ACPI_SIGNATURE('S','P','C','R')
 
 #define ACPI_SIGNATURE_8BYTE(c1, c2, c3, c4, c5, c6, c7, c8) \
 	(((uint64_t)(ACPI_SIGNATURE(c1, c2, c3, c4))) |	     \
@@ -145,6 +146,30 @@ struct acpi_table_facs_rev1 {
 	u8 reserved3[40];	/* Reserved - must be zero */
 };
 
+struct spcr_descriptor {
+	ACPI_TABLE_HEADER_DEF	/* ACPI common table header */
+	u8 interface_type;	/* 0=full 16550, 1=subset of 16550 */
+	u8 reserved[3];
+	struct acpi_generic_address serial_port;
+	u8 interrupt_type;
+	u8 pc_interrupt;
+	u32 interrupt;
+	u8 baud_rate;
+	u8 parity;
+	u8 stop_bits;
+	u8 flow_control;
+	u8 terminal_type;
+	u8 reserved1;
+	u16 pci_device_id;
+	u16 pci_vendor_id;
+	u8 pci_bus;
+	u8 pci_device;
+	u8 pci_function;
+	u32 pci_flags;
+	u8 pci_segment;
+	u32 reserved2;
+};
+
 #pragma pack(0)
 
 void set_efi_rsdp(struct acpi_table_rsdp *rsdp);
diff --git a/lib/arm/io.c b/lib/arm/io.c
index 343e1082..19f93490 100644
--- a/lib/arm/io.c
+++ b/lib/arm/io.c
@@ -29,7 +29,7 @@ static struct spinlock uart_lock;
 #define UART_EARLY_BASE (u8 *)(unsigned long)CONFIG_UART_EARLY_BASE
 static volatile u8 *uart0_base = UART_EARLY_BASE;
 
-static void uart0_init(void)
+static void uart0_init_fdt(void)
 {
 	/*
 	 * kvm-unit-tests uses the uart only for output. Both uart models have
@@ -65,17 +65,41 @@ static void uart0_init(void)
 	}
 
 	uart0_base = ioremap(base.addr, base.size);
+}
+
+#ifdef CONFIG_EFI
+
+#include <acpi.h>
+
+static void uart0_init_acpi(void)
+{
+	struct spcr_descriptor *spcr = find_acpi_table_addr(SPCR_SIGNATURE);
+
+	assert_msg(spcr, "Unable to find ACPI SPCR");
+	uart0_base = ioremap(spcr->serial_port.address, spcr->serial_port.bit_width);
+}
+#else
+
+static void uart0_init_acpi(void)
+{
+	assert_msg(false, "ACPI not available");
+}
+
+#endif
+
+void io_init(void)
+{
+	if (dt_available())
+		uart0_init_fdt();
+	else
+		uart0_init_acpi();
 
 	if (uart0_base != UART_EARLY_BASE) {
 		printf("WARNING: early print support may not work. "
 		       "Found uart at %p, but early base is %p.\n",
 			uart0_base, UART_EARLY_BASE);
 	}
-}
 
-void io_init(void)
-{
-	uart0_init();
 	chr_testdev_init();
 }
 
-- 
2.25.1

[PATCH v4 14/30] arm64: Add support for timer initialization through ACPI

[Nikos Nikoleris]

For systems with ACPI support, we can discover timers through the ACPI
GTDT table. This change implements the code to discover timers through
the GTDT and adds ACPI support in timer_save_state. This change
retains the default behavior; we check if a valid DT is provided, if
not, we try to discover timers using ACPI.

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
---
 arm/Makefile.common |  1 +
 arm/micro-bench.c   |  4 +-
 arm/timer.c         | 10 ++---
 lib/acpi.h          | 18 +++++++++
 lib/arm/asm/timer.h |  2 +
 lib/arm/setup.c     | 39 -------------------
 lib/arm/timer.c     | 92 +++++++++++++++++++++++++++++++++++++++++++++
 7 files changed, 120 insertions(+), 46 deletions(-)
 create mode 100644 lib/arm/timer.c

diff --git a/arm/Makefile.common b/arm/Makefile.common
index 3a726c20..c02dd906 100644
--- a/arm/Makefile.common
+++ b/arm/Makefile.common
@@ -53,6 +53,7 @@ cflatobjs += lib/arm/psci.o
 cflatobjs += lib/arm/smp.o
 cflatobjs += lib/arm/delay.o
 cflatobjs += lib/arm/gic.o lib/arm/gic-v2.o lib/arm/gic-v3.o
+cflatobjs += lib/arm/timer.o
 
 ifeq ($(CONFIG_EFI),y)
 cflatobjs += lib/acpi.o
diff --git a/arm/micro-bench.c b/arm/micro-bench.c
index 84366125..5be9b189 100644
--- a/arm/micro-bench.c
+++ b/arm/micro-bench.c
@@ -43,7 +43,7 @@ static void gic_irq_handler(struct pt_regs *regs)
 	irq_received = true;
 	gic_write_eoir(irqstat);
 
-	if (irqstat == PPI(TIMER_VTIMER_IRQ)) {
+	if (irqstat == TIMER_VTIMER_IRQ) {
 		write_sysreg((ARCH_TIMER_CTL_IMASK | ARCH_TIMER_CTL_ENABLE),
 			     cntv_ctl_el0);
 		isb();
@@ -229,7 +229,7 @@ static bool timer_prep(void)
 		assert_msg(0, "Unreachable");
 	}
 
-	writel(1 << PPI(TIMER_VTIMER_IRQ), gic_isenabler);
+	writel(1 << TIMER_VTIMER_IRQ, gic_isenabler);
 	write_sysreg(ARCH_TIMER_CTL_IMASK | ARCH_TIMER_CTL_ENABLE, cntv_ctl_el0);
 	isb();
 
diff --git a/arm/timer.c b/arm/timer.c
index c4e7b10f..a4a82812 100644
--- a/arm/timer.c
+++ b/arm/timer.c
@@ -139,7 +139,7 @@ static struct timer_info ptimer_info = {
 
 static void set_timer_irq_enabled(struct timer_info *info, bool enabled)
 {
-	u32 val = 1 << PPI(info->irq);
+	u32 val = 1 << info->irq;
 
 	if (enabled)
 		writel(val, gic_isenabler);
@@ -153,9 +153,9 @@ static void irq_handler(struct pt_regs *regs)
 	u32 irqstat = gic_read_iar();
 	u32 irqnr = gic_iar_irqnr(irqstat);
 
-	if (irqnr == PPI(vtimer_info.irq)) {
+	if (irqnr == vtimer_info.irq) {
 		info = &vtimer_info;
-	} else if (irqnr == PPI(ptimer_info.irq)) {
+	} else if (irqnr == ptimer_info.irq) {
 		info = &ptimer_info;
 	} else {
 		if (irqnr != GICC_INT_SPURIOUS)
@@ -185,9 +185,9 @@ static bool gic_timer_check_state(struct timer_info *info,
 	/* Wait for up to 1s for the GIC to sample the interrupt. */
 	for (i = 0; i < 10; i++) {
 		mdelay(100);
-		if (gic_irq_state(PPI(info->irq)) == expected_state) {
+		if (gic_irq_state(info->irq) == expected_state) {
 			mdelay(100);
-			if (gic_irq_state(PPI(info->irq)) == expected_state)
+			if (gic_irq_state(info->irq) == expected_state)
 				return true;
 		}
 	}
diff --git a/lib/acpi.h b/lib/acpi.h
index 02ca8958..9d974ee2 100644
--- a/lib/acpi.h
+++ b/lib/acpi.h
@@ -18,6 +18,7 @@
 #define FACP_SIGNATURE ACPI_SIGNATURE('F','A','C','P')
 #define FACS_SIGNATURE ACPI_SIGNATURE('F','A','C','S')
 #define SPCR_SIGNATURE ACPI_SIGNATURE('S','P','C','R')
+#define GTDT_SIGNATURE ACPI_SIGNATURE('G','T','D','T')
 
 #define ACPI_SIGNATURE_8BYTE(c1, c2, c3, c4, c5, c6, c7, c8) \
 	(((uint64_t)(ACPI_SIGNATURE(c1, c2, c3, c4))) |	     \
@@ -170,6 +171,23 @@ struct spcr_descriptor {
 	u32 reserved2;
 };
 
+struct acpi_table_gtdt {
+	ACPI_TABLE_HEADER_DEF	/* ACPI common table header */
+	u64 counter_block_addresss;
+	u32 reserved;
+	u32 secure_el1_interrupt;
+	u32 secure_el1_flags;
+	u32 non_secure_el1_interrupt;
+	u32 non_secure_el1_flags;
+	u32 virtual_timer_interrupt;
+	u32 virtual_timer_flags;
+	u32 non_secure_el2_interrupt;
+	u32 non_secure_el2_flags;
+	u64 counter_read_block_address;
+	u32 platform_timer_count;
+	u32 platform_timer_offset;
+};
+
 #pragma pack(0)
 
 void set_efi_rsdp(struct acpi_table_rsdp *rsdp);
diff --git a/lib/arm/asm/timer.h b/lib/arm/asm/timer.h
index f75cc67f..aaf839fc 100644
--- a/lib/arm/asm/timer.h
+++ b/lib/arm/asm/timer.h
@@ -27,5 +27,7 @@ extern struct timer_state __timer_state;
 #define TIMER_PTIMER_IRQ (__timer_state.ptimer.irq)
 #define TIMER_VTIMER_IRQ (__timer_state.vtimer.irq)
 
+void timer_save_state(void);
+
 #endif /* !__ASSEMBLY__ */
 #endif /* _ASMARM_TIMER_H_ */
diff --git a/lib/arm/setup.c b/lib/arm/setup.c
index bcdf0d78..1572c64e 100644
--- a/lib/arm/setup.c
+++ b/lib/arm/setup.c
@@ -35,8 +35,6 @@
 
 extern unsigned long etext;
 
-struct timer_state __timer_state;
-
 char *initrd;
 u32 initrd_size;
 
@@ -199,43 +197,6 @@ static void mem_init(phys_addr_t freemem_start)
 	page_alloc_ops_enable();
 }
 
-static void timer_save_state(void)
-{
-	const struct fdt_property *prop;
-	const void *fdt = dt_fdt();
-	int node, len;
-	u32 *data;
-
-	node = fdt_node_offset_by_compatible(fdt, -1, "arm,armv8-timer");
-	assert(node >= 0 || node == -FDT_ERR_NOTFOUND);
-
-	if (node == -FDT_ERR_NOTFOUND) {
-		__timer_state.ptimer.irq = -1;
-		__timer_state.vtimer.irq = -1;
-		return;
-	}
-
-	/*
-	 * From Linux devicetree timer binding documentation
-	 *
-	 * interrupts <type irq flags>:
-	 *	secure timer irq
-	 *	non-secure timer irq		(ptimer)
-	 *	virtual timer irq		(vtimer)
-	 *	hypervisor timer irq
-	 */
-	prop = fdt_get_property(fdt, node, "interrupts", &len);
-	assert(prop && len == (4 * 3 * sizeof(u32)));
-
-	data = (u32 *)prop->data;
-	assert(fdt32_to_cpu(data[3]) == 1 /* PPI */);
-	__timer_state.ptimer.irq = fdt32_to_cpu(data[4]);
-	__timer_state.ptimer.irq_flags = fdt32_to_cpu(data[5]);
-	assert(fdt32_to_cpu(data[6]) == 1 /* PPI */);
-	__timer_state.vtimer.irq = fdt32_to_cpu(data[7]);
-	__timer_state.vtimer.irq_flags = fdt32_to_cpu(data[8]);
-}
-
 void setup(const void *fdt, phys_addr_t freemem_start)
 {
 	void *freemem;
diff --git a/lib/arm/timer.c b/lib/arm/timer.c
new file mode 100644
index 00000000..ae702e41
--- /dev/null
+++ b/lib/arm/timer.c
@@ -0,0 +1,92 @@
+/*
+ * Initialize timers.
+ *
+ * Copyright (C) 2022, Arm Ltd., Nikos Nikoleris <nikos.nikoleris@arm.com>
+ * Copyright (C) 2014, Red Hat Inc, Andrew Jones <drjones@redhat.com>
+ *
+ * This work is licensed under the terms of the GNU LGPL, version 2.
+ */
+#include <libcflat.h>
+#include <acpi.h>
+#include <devicetree.h>
+#include <libfdt/libfdt.h>
+#include <asm/gic.h>
+#include <asm/timer.h>
+
+struct timer_state __timer_state;
+
+static void timer_save_state_fdt(void)
+{
+	const struct fdt_property *prop;
+	const void *fdt = dt_fdt();
+	int node, len;
+	u32 *data;
+
+	node = fdt_node_offset_by_compatible(fdt, -1, "arm,armv8-timer");
+	assert(node >= 0 || node == -FDT_ERR_NOTFOUND);
+
+	if (node == -FDT_ERR_NOTFOUND) {
+		__timer_state.ptimer.irq = -1;
+		__timer_state.vtimer.irq = -1;
+		return;
+	}
+
+	/*
+	 * From Linux devicetree timer binding documentation
+	 *
+	 * interrupts <type irq flags>:
+	 *      secure timer irq
+	 *      non-secure timer irq            (ptimer)
+	 *      virtual timer irq               (vtimer)
+	 *      hypervisor timer irq
+	 */
+	prop = fdt_get_property(fdt, node, "interrupts", &len);
+	assert(prop && len == (4 * 3 * sizeof(u32)));
+
+	data = (u32 *) prop->data;
+	assert(fdt32_to_cpu(data[3]) == 1 /* PPI */ );
+	__timer_state.ptimer.irq = PPI(fdt32_to_cpu(data[4]));
+	__timer_state.ptimer.irq_flags = fdt32_to_cpu(data[5]);
+	assert(fdt32_to_cpu(data[6]) == 1 /* PPI */ );
+	__timer_state.vtimer.irq = PPI(fdt32_to_cpu(data[7]));
+	__timer_state.vtimer.irq_flags = fdt32_to_cpu(data[8]);
+}
+
+#ifdef CONFIG_EFI
+
+#include <acpi.h>
+
+static void timer_save_state_acpi(void)
+{
+	struct acpi_table_gtdt *gtdt = find_acpi_table_addr(GTDT_SIGNATURE);
+
+	if (!gtdt) {
+		printf("Cannot find ACPI GTDT");
+		__timer_state.ptimer.irq = -1;
+		__timer_state.vtimer.irq = -1;
+		return;
+	}
+
+	__timer_state.ptimer.irq = gtdt->non_secure_el1_interrupt;
+	__timer_state.ptimer.irq_flags = gtdt->non_secure_el1_flags;
+
+	__timer_state.vtimer.irq = gtdt->virtual_timer_interrupt;
+	__timer_state.vtimer.irq_flags = gtdt->virtual_timer_flags;
+}
+
+#else
+
+static void timer_save_state_acpi(void)
+{
+	assert_msg(false, "ACPI not available");
+}
+
+#endif
+
+void timer_save_state(void)
+{
+	if (dt_available())
+		timer_save_state_fdt();
+	else
+		timer_save_state_acpi();
+}
-- 
2.25.1

[PATCH v4 15/30] arm64: Add support for cpu initialization through ACPI

[Nikos Nikoleris]

In systems with ACPI support and when a DT is not provided, we can use
the MADTs to discover the number of CPUs and their corresponding MIDR.
This change implements this but retains the default behavior; we check
if a valid DT is provided, if not, we try to discover the cores in the
system using ACPI.

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
---
 lib/acpi.c      | 20 ++++++++++++++++
 lib/acpi.h      | 63 +++++++++++++++++++++++++++++++++++++++++++++++++
 lib/arm/setup.c | 42 ++++++++++++++++++++++++++++++---
 3 files changed, 122 insertions(+), 3 deletions(-)

diff --git a/lib/acpi.c b/lib/acpi.c
index a197f3dd..bbe33d08 100644
--- a/lib/acpi.c
+++ b/lib/acpi.c
@@ -102,3 +102,23 @@ void *find_acpi_table_addr(u32 sig)
 
 	return NULL;
 }
+
+void acpi_table_parse_madt(enum acpi_madt_type mtype, acpi_table_handler handler)
+{
+	struct acpi_table_madt *madt;
+	struct acpi_subtable_header *header;
+	void *end;
+
+	madt = find_acpi_table_addr(MADT_SIGNATURE);
+	assert(madt);
+
+	header = (void *)(ulong) madt + sizeof(struct acpi_table_madt);
+	end = (void *)((ulong) madt + madt->length);
+
+	while ((void *)header < end) {
+		if (header->type == mtype)
+			handler(header);
+
+		header = (void *)(ulong) header + header->length;
+	}
+}
diff --git a/lib/acpi.h b/lib/acpi.h
index 9d974ee2..af02fd83 100644
--- a/lib/acpi.h
+++ b/lib/acpi.h
@@ -17,6 +17,7 @@
 #define XSDT_SIGNATURE ACPI_SIGNATURE('X','S','D','T')
 #define FACP_SIGNATURE ACPI_SIGNATURE('F','A','C','P')
 #define FACS_SIGNATURE ACPI_SIGNATURE('F','A','C','S')
+#define MADT_SIGNATURE ACPI_SIGNATURE('A','P','I','C')
 #define SPCR_SIGNATURE ACPI_SIGNATURE('S','P','C','R')
 #define GTDT_SIGNATURE ACPI_SIGNATURE('G','T','D','T')
 
@@ -147,6 +148,67 @@ struct acpi_table_facs_rev1 {
 	u8 reserved3[40];	/* Reserved - must be zero */
 };
 
+struct acpi_table_madt {
+	ACPI_TABLE_HEADER_DEF	/* ACPI common table header */
+	u32 address;		/* Physical address of local APIC */
+	u32 flags;
+};
+
+struct acpi_subtable_header {
+	u8 type;
+	u8 length;
+};
+
+typedef int (*acpi_table_handler)(struct acpi_subtable_header *header);
+
+/* 11: Generic interrupt - GICC (ACPI 5.0 + ACPI 6.0 + ACPI 6.3 changes) */
+
+struct acpi_madt_generic_interrupt {
+	u8 type;
+	u8 length;
+	u16 reserved;		/* reserved - must be zero */
+	u32 cpu_interface_number;
+	u32 uid;
+	u32 flags;
+	u32 parking_version;
+	u32 performance_interrupt;
+	u64 parked_address;
+	u64 base_address;
+	u64 gicv_base_address;
+	u64 gich_base_address;
+	u32 vgic_interrupt;
+	u64 gicr_base_address;
+	u64 arm_mpidr;
+	u8 efficiency_class;
+	u8 reserved2[1];
+	u16 spe_interrupt;	/* ACPI 6.3 */
+};
+
+/* Values for MADT subtable type in struct acpi_subtable_header */
+
+enum acpi_madt_type {
+	ACPI_MADT_TYPE_LOCAL_APIC = 0,
+	ACPI_MADT_TYPE_IO_APIC = 1,
+	ACPI_MADT_TYPE_INTERRUPT_OVERRIDE = 2,
+	ACPI_MADT_TYPE_NMI_SOURCE = 3,
+	ACPI_MADT_TYPE_LOCAL_APIC_NMI = 4,
+	ACPI_MADT_TYPE_LOCAL_APIC_OVERRIDE = 5,
+	ACPI_MADT_TYPE_IO_SAPIC = 6,
+	ACPI_MADT_TYPE_LOCAL_SAPIC = 7,
+	ACPI_MADT_TYPE_INTERRUPT_SOURCE = 8,
+	ACPI_MADT_TYPE_LOCAL_X2APIC = 9,
+	ACPI_MADT_TYPE_LOCAL_X2APIC_NMI = 10,
+	ACPI_MADT_TYPE_GENERIC_INTERRUPT = 11,
+	ACPI_MADT_TYPE_GENERIC_DISTRIBUTOR = 12,
+	ACPI_MADT_TYPE_GENERIC_MSI_FRAME = 13,
+	ACPI_MADT_TYPE_GENERIC_REDISTRIBUTOR = 14,
+	ACPI_MADT_TYPE_GENERIC_TRANSLATOR = 15,
+	ACPI_MADT_TYPE_RESERVED = 16	/* 16 and greater are reserved */
+};
+
+/* MADT Local APIC flags */
+#define ACPI_MADT_ENABLED		(1)	/* 00: Processor is usable if set */
+
 struct spcr_descriptor {
 	ACPI_TABLE_HEADER_DEF	/* ACPI common table header */
 	u8 interface_type;	/* 0=full 16550, 1=subset of 16550 */
@@ -192,5 +254,6 @@ struct acpi_table_gtdt {
 
 void set_efi_rsdp(struct acpi_table_rsdp *rsdp);
 void *find_acpi_table_addr(u32 sig);
+void acpi_table_parse_madt(enum acpi_madt_type mtype, acpi_table_handler handler);
 
 #endif
diff --git a/lib/arm/setup.c b/lib/arm/setup.c
index 1572c64e..59b0aedd 100644
--- a/lib/arm/setup.c
+++ b/lib/arm/setup.c
@@ -55,7 +55,7 @@ int mpidr_to_cpu(uint64_t mpidr)
 	return -1;
 }
 
-static void cpu_set(int fdtnode __unused, u64 regval, void *info __unused)
+static void cpu_set_fdt(int fdtnode __unused, u64 regval, void *info __unused)
 {
 	int cpu = nr_cpus++;
 
@@ -65,13 +65,49 @@ static void cpu_set(int fdtnode __unused, u64 regval, void *info __unused)
 	set_cpu_present(cpu, true);
 }
 
+#ifdef CONFIG_EFI
+
+#include <acpi.h>
+
+static int cpu_set_acpi(struct acpi_subtable_header *header)
+{
+	int cpu = nr_cpus++;
+	struct acpi_madt_generic_interrupt *gicc = (void *)header;
+
+	assert_msg(cpu < NR_CPUS, "Number cpus exceeds maximum supported (%d).", NR_CPUS);
+
+	cpus[cpu] = gicc->arm_mpidr;
+	set_cpu_present(cpu, true);
+
+	return 0;
+}
+
+static void cpu_init_acpi(void)
+{
+	acpi_table_parse_madt(ACPI_MADT_TYPE_GENERIC_INTERRUPT, cpu_set_acpi);
+}
+
+#else
+
+static void cpu_init_acpi(void)
+{
+	assert_msg(false, "ACPI not available");
+}
+
+#endif
+
 static void cpu_init(void)
 {
 	int ret;
 
 	nr_cpus = 0;
-	ret = dt_for_each_cpu_node(cpu_set, NULL);
-	assert(ret == 0);
+	if (dt_available()) {
+		ret = dt_for_each_cpu_node(cpu_set_fdt, NULL);
+		assert(ret == 0);
+	} else {
+		cpu_init_acpi();
+	}
+
 	set_cpu_online(0, true);
 }
 
-- 
2.25.1

[PATCH v4 16/30] arm64: Add support for gic initialization through ACPI

[Nikos Nikoleris]

In systems with ACPI support and when a DT is not provided, we can use
the MADTs to figure out if it implements a GICv2 or a GICv3 and
discover the GIC parameters. This change implements this but retains
the default behavior; we check if a valid DT is provided, if not, we
try to discover the cores in the system using ACPI.

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
---
 lib/acpi.c     |   9 +++-
 lib/acpi.h     |  46 +++++++++++++++-
 lib/arm/gic.c  | 139 ++++++++++++++++++++++++++++++++++++++++++++++++-
 lib/libcflat.h |   1 +
 4 files changed, 191 insertions(+), 4 deletions(-)

diff --git a/lib/acpi.c b/lib/acpi.c
index bbe33d08..760cd8b2 100644
--- a/lib/acpi.c
+++ b/lib/acpi.c
@@ -103,11 +103,12 @@ void *find_acpi_table_addr(u32 sig)
 	return NULL;
 }
 
-void acpi_table_parse_madt(enum acpi_madt_type mtype, acpi_table_handler handler)
+int acpi_table_parse_madt(enum acpi_madt_type mtype, acpi_table_handler handler)
 {
 	struct acpi_table_madt *madt;
 	struct acpi_subtable_header *header;
 	void *end;
+	int count = 0;
 
 	madt = find_acpi_table_addr(MADT_SIGNATURE);
 	assert(madt);
@@ -116,9 +117,13 @@ void acpi_table_parse_madt(enum acpi_madt_type mtype, acpi_table_handler handler
 	end = (void *)((ulong) madt + madt->length);
 
 	while ((void *)header < end) {
-		if (header->type == mtype)
+		if (header->type == mtype) {
 			handler(header);
+			count++;
+		}
 
 		header = (void *)(ulong) header + header->length;
 	}
+
+	return count;
 }
diff --git a/lib/acpi.h b/lib/acpi.h
index af02fd83..a2c0f982 100644
--- a/lib/acpi.h
+++ b/lib/acpi.h
@@ -184,6 +184,50 @@ struct acpi_madt_generic_interrupt {
 	u16 spe_interrupt;	/* ACPI 6.3 */
 };
 
+#define ACPI_MADT_ENABLED           (1)	/* 00: Processor is usable if set */
+
+/* 12: Generic Distributor (ACPI 5.0 + ACPI 6.0 changes) */
+
+struct acpi_madt_generic_distributor {
+	struct acpi_subtable_header header;
+	u16 reserved;		/* reserved - must be zero */
+	u32 gic_id;
+	u64 base_address;
+	u32 global_irq_base;
+	u8 version;
+	u8 reserved2[3];	/* reserved - must be zero */
+};
+
+/* Values for Version field above */
+
+enum acpi_madt_gic_version {
+	ACPI_MADT_GIC_VERSION_NONE = 0,
+	ACPI_MADT_GIC_VERSION_V1 = 1,
+	ACPI_MADT_GIC_VERSION_V2 = 2,
+	ACPI_MADT_GIC_VERSION_V3 = 3,
+	ACPI_MADT_GIC_VERSION_V4 = 4,
+	ACPI_MADT_GIC_VERSION_RESERVED = 5	/* 5 and greater are reserved */
+};
+
+/* 14: Generic Redistributor (ACPI 5.1) */
+
+struct acpi_madt_generic_redistributor {
+	struct acpi_subtable_header header;
+	u16 reserved;		/* reserved - must be zero */
+	u64 base_address;
+	u32 length;
+};
+
+/* 15: Generic Translator (ACPI 6.0) */
+
+struct acpi_madt_generic_translator {
+	struct acpi_subtable_header header;
+	u16 reserved;		/* reserved - must be zero */
+	u32 translation_id;
+	u64 base_address;
+	u32 reserved2;
+};
+
 /* Values for MADT subtable type in struct acpi_subtable_header */
 
 enum acpi_madt_type {
@@ -254,6 +298,6 @@ struct acpi_table_gtdt {
 
 void set_efi_rsdp(struct acpi_table_rsdp *rsdp);
 void *find_acpi_table_addr(u32 sig);
-void acpi_table_parse_madt(enum acpi_madt_type mtype, acpi_table_handler handler);
+int acpi_table_parse_madt(enum acpi_madt_type mtype, acpi_table_handler handler);
 
 #endif
diff --git a/lib/arm/gic.c b/lib/arm/gic.c
index 1bfcfcfb..8fc5596f 100644
--- a/lib/arm/gic.c
+++ b/lib/arm/gic.c
@@ -3,6 +3,7 @@
  *
  * This work is licensed under the terms of the GNU LGPL, version 2.
  */
+#include <acpi.h>
 #include <devicetree.h>
 #include <asm/gic.h>
 #include <asm/io.h>
@@ -120,7 +121,7 @@ int gic_version(void)
 	return 0;
 }
 
-int gic_init(void)
+static int gic_init_fdt(void)
 {
 	if (gicv2_init()) {
 		gic_common_ops = &gicv2_common_ops;
@@ -133,6 +134,142 @@ int gic_init(void)
 	return gic_version();
 }
 
+#ifdef CONFIG_EFI
+
+#define ACPI_GICV2_DIST_MEM_SIZE	(SZ_4K)
+#define ACPI_GIC_CPU_IF_MEM_SIZE	(SZ_8K)
+#define ACPI_GICV3_DIST_MEM_SIZE	(SZ_64K)
+#define ACPI_GICV3_ITS_MEM_SIZE		(SZ_128K)
+
+static int gic_acpi_version(struct acpi_subtable_header *header)
+{
+	struct acpi_madt_generic_distributor *dist = (void *)header;
+	int version = dist->version;
+
+	if (version == 2)
+		gic_common_ops = &gicv2_common_ops;
+	else if (version == 3)
+		gic_common_ops = &gicv3_common_ops;
+
+	return version;
+}
+
+static int gicv2_acpi_parse_madt_cpu(struct acpi_subtable_header *header)
+{
+	struct acpi_madt_generic_interrupt *gicc = (void *)header;
+	static phys_addr_t gicc_base_address;
+
+	if (!(gicc->flags & ACPI_MADT_ENABLED))
+		return 0;
+
+	if (!gicc_base_address) {
+		gicc_base_address = gicc->base_address;
+		gicv2_data.cpu_base = ioremap(gicc_base_address, ACPI_GIC_CPU_IF_MEM_SIZE);
+	}
+	assert(gicc_base_address == gicc->base_address);
+
+	return 0;
+}
+
+static int gicv2_acpi_parse_madt_dist(struct acpi_subtable_header *header)
+{
+	struct acpi_madt_generic_distributor *dist = (void *)header;
+
+	gicv2_data.dist_base = ioremap(dist->base_address, ACPI_GICV2_DIST_MEM_SIZE);
+
+	return 0;
+}
+
+static int gicv3_acpi_parse_madt_gicc(struct acpi_subtable_header *header)
+{
+	struct acpi_madt_generic_interrupt *gicc = (void *)header;
+	static phys_addr_t gicr_base_address;
+
+	if (!(gicc->flags & ACPI_MADT_ENABLED))
+		return 0;
+
+	if (!gicr_base_address) {
+		gicr_base_address = gicc->gicr_base_address;
+		gicv3_data.redist_bases[0] = ioremap(gicr_base_address, SZ_64K * 2);
+	}
+	assert(gicr_base_address == gicc->gicr_base_address);
+
+	return 0;
+}
+
+static int gicv3_acpi_parse_madt_dist(struct acpi_subtable_header *header)
+{
+	struct acpi_madt_generic_distributor *dist = (void *)header;
+
+	gicv3_data.dist_base = ioremap(dist->base_address, ACPI_GICV3_DIST_MEM_SIZE);
+
+	return 0;
+}
+
+static int gicv3_acpi_parse_madt_redist(struct acpi_subtable_header *header)
+{
+	static int i;
+	struct acpi_madt_generic_redistributor *redist = (void *)header;
+
+	gicv3_data.redist_bases[i++] = ioremap(redist->base_address, redist->length);
+
+	return 0;
+}
+
+static int gicv3_acpi_parse_madt_its(struct acpi_subtable_header *header)
+{
+	struct acpi_madt_generic_translator *its_entry = (void *)header;
+
+	its_data.base = ioremap(its_entry->base_address, ACPI_GICV3_ITS_MEM_SIZE - 1);
+
+	return 0;
+}
+
+static int gic_init_acpi(void)
+{
+	int count;
+
+	acpi_table_parse_madt(ACPI_MADT_TYPE_GENERIC_DISTRIBUTOR, gic_acpi_version);
+	if (gic_version() == 2) {
+		acpi_table_parse_madt(ACPI_MADT_TYPE_GENERIC_INTERRUPT,
+				     gicv2_acpi_parse_madt_cpu);
+		acpi_table_parse_madt(ACPI_MADT_TYPE_GENERIC_DISTRIBUTOR,
+				      gicv2_acpi_parse_madt_dist);
+	} else if (gic_version() == 3) {
+		acpi_table_parse_madt(ACPI_MADT_TYPE_GENERIC_DISTRIBUTOR,
+				      gicv3_acpi_parse_madt_dist);
+		count = acpi_table_parse_madt(ACPI_MADT_TYPE_GENERIC_REDISTRIBUTOR,
+					      gicv3_acpi_parse_madt_redist);
+		if (!count)
+			acpi_table_parse_madt(ACPI_MADT_TYPE_GENERIC_INTERRUPT,
+					      gicv3_acpi_parse_madt_gicc);
+		acpi_table_parse_madt(ACPI_MADT_TYPE_GENERIC_TRANSLATOR,
+				      gicv3_acpi_parse_madt_its);
+#ifdef __aarch64__
+		its_init();
+#endif
+	}
+
+	return gic_version();
+}
+
+#else
+
+static int gic_init_acpi(void)
+{
+	assert_msg(false, "ACPI not available");
+}
+
+#endif /* CONFIG_EFI */
+
+int gic_init(void)
+{
+	if (dt_available())
+		return gic_init_fdt();
+	else
+		return gic_init_acpi();
+}
+
 void gic_enable_defaults(void)
 {
 	if (!gic_common_ops) {
diff --git a/lib/libcflat.h b/lib/libcflat.h
index c1fd31ff..700f4352 100644
--- a/lib/libcflat.h
+++ b/lib/libcflat.h
@@ -161,6 +161,7 @@ extern void setup_vm(void);
 #define SZ_8K			(1 << 13)
 #define SZ_16K			(1 << 14)
 #define SZ_64K			(1 << 16)
+#define SZ_128K			(1 << 17)
 #define SZ_1M			(1 << 20)
 #define SZ_2M			(1 << 21)
 #define SZ_1G			(1 << 30)
-- 
2.25.1

[PATCH v4 17/30] lib/printf: Support for precision modifier in printf

[Nikos Nikoleris]

This follows the typical format of:

printf("%.Ns", *str);

Where N might be a decimal digit string or '*'. This feature is used
by a future change.

See also: man 3 printf

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
---
 lib/printf.c | 95 ++++++++++++++++++++++++++++++++++++++++++----------
 1 file changed, 78 insertions(+), 17 deletions(-)

diff --git a/lib/printf.c b/lib/printf.c
index 383799ec..d600199b 100644
--- a/lib/printf.c
+++ b/lib/printf.c
@@ -6,6 +6,7 @@
  */
 
 #include "libcflat.h"
+#include "ctype.h"
 
 #define BUFSZ 2000
 
@@ -19,6 +20,7 @@ typedef struct strprops {
 	char pad;
 	int npad;
 	bool alternate;
+	int precision;
 } strprops_t;
 
 static void addchar(pstream_t *p, char c)
@@ -43,7 +45,7 @@ static void print_str(pstream_t *p, const char *s, strprops_t props)
 		}
 	}
 
-	while (*s)
+	while (*s && props.precision--)
 		addchar(p, *s++);
 
 	if (npad < 0) {
@@ -73,12 +75,13 @@ static void print_int(pstream_t *ps, long long n, int base, strprops_t props)
 		n /= base;
 	}
 
+	while (p == buf || (p - buf < props.precision))
+		*p++ = '0';
+	props.precision = -1;
+
 	if (s)
 		*p++ = '-';
 
-	if (p == buf)
-		*p++ = '0';
-
 	for (i = 0; i < (p - buf) / 2; ++i) {
 		char tmp;
 
@@ -104,8 +107,13 @@ static void print_unsigned(pstream_t *ps, unsigned long long n, int base,
 	}
 
 	if (p == buf)
+		props.alternate = false;
+
+	while (p == buf || (p - buf < props.precision))
 		*p++ = '0';
-	else if (props.alternate && base == 16) {
+	props.precision = -1;
+
+	if (props.alternate && base == 16) {
 		if (props.pad == '0') {
 			addchar(ps, '0');
 			addchar(ps, 'x');
@@ -147,9 +155,56 @@ static int fmtnum(const char **fmt)
 	return num;
 }
 
+/*
+ * Adapted from drivers/firmware/efi/libstub/vsprintf.c
+ */
+static int skip_atoi(const char **s)
+{
+	int i = 0;
+
+	do {
+		i = i*10 + *((*s)++) - '0';
+	} while (isdigit(**s));
+
+	return i;
+}
+
+/*
+ * Adapted from drivers/firmware/efi/libstub/vsprintf.c
+ */
+static int get_int(const char **fmt, va_list *ap)
+{
+	if (isdigit(**fmt))
+		return skip_atoi(fmt);
+
+	if (**fmt == '*') {
+		++(*fmt);
+		/* it's the next argument */
+		return va_arg(*ap, int);
+	}
+	return 0;
+}
+
 int vsnprintf(char *buf, int size, const char *fmt, va_list va)
 {
 	pstream_t s;
+	va_list args;
+
+	/*
+	 * We want to pass our input va_list to helper functions by reference,
+	 * but there's an annoying edge case. If va_list was originally passed
+	 * to us by value, we could just pass &ap down to the helpers. This is
+	 * the case on, for example, X86_32.
+	 * However, on X86_64 (and possibly others), va_list is actually a
+	 * size-1 array containing a structure. Our function parameter ap has
+	 * decayed from T[1] to T*, and &ap has type T** rather than T(*)[1],
+	 * which is what will be expected by a function taking a va_list *
+	 * parameter.
+	 * One standard way to solve this mess is by creating a copy in a local
+	 * variable of type va_list and then passing a pointer to that local
+	 * copy instead, which is what we do here.
+	 */
+	va_copy(args, va);
 
 	s.buffer = buf;
 	s.remain = size - 1;
@@ -160,6 +215,7 @@ int vsnprintf(char *buf, int size, const char *fmt, va_list va)
 		strprops_t props;
 		memset(&props, 0, sizeof(props));
 		props.pad = ' ';
+		props.precision = -1;
 
 		if (f != '%') {
 			addchar(&s, f);
@@ -172,11 +228,15 @@ morefmt:
 			addchar(&s, '%');
 			break;
 		case 'c':
-			addchar(&s, va_arg(va, int));
+			addchar(&s, va_arg(args, int));
 			break;
 		case '\0':
 			--fmt;
 			break;
+		case '.':
+			props.pad = ' ';
+			props.precision = get_int(&fmt, &args);
+			goto morefmt;
 		case '#':
 			props.alternate = true;
 			goto morefmt;
@@ -204,54 +264,55 @@ morefmt:
 		case 'd':
 			switch (nlong) {
 			case 0:
-				print_int(&s, va_arg(va, int), 10, props);
+				print_int(&s, va_arg(args, int), 10, props);
 				break;
 			case 1:
-				print_int(&s, va_arg(va, long), 10, props);
+				print_int(&s, va_arg(args, long), 10, props);
 				break;
 			default:
-				print_int(&s, va_arg(va, long long), 10, props);
+				print_int(&s, va_arg(args, long long), 10, props);
 				break;
 			}
 			break;
 		case 'u':
 			switch (nlong) {
 			case 0:
-				print_unsigned(&s, va_arg(va, unsigned), 10, props);
+				print_unsigned(&s, va_arg(args, unsigned int), 10, props);
 				break;
 			case 1:
-				print_unsigned(&s, va_arg(va, unsigned long), 10, props);
+				print_unsigned(&s, va_arg(args, unsigned long), 10, props);
 				break;
 			default:
-				print_unsigned(&s, va_arg(va, unsigned long long), 10, props);
+				print_unsigned(&s, va_arg(args, unsigned long long), 10, props);
 				break;
 			}
 			break;
 		case 'x':
 			switch (nlong) {
 			case 0:
-				print_unsigned(&s, va_arg(va, unsigned), 16, props);
+				print_unsigned(&s, va_arg(args, unsigned int), 16, props);
 				break;
 			case 1:
-				print_unsigned(&s, va_arg(va, unsigned long), 16, props);
+				print_unsigned(&s, va_arg(args, unsigned long), 16, props);
 				break;
 			default:
-				print_unsigned(&s, va_arg(va, unsigned long long), 16, props);
+				print_unsigned(&s, va_arg(args, unsigned long long), 16, props);
 				break;
 			}
 			break;
 		case 'p':
 			props.alternate = true;
-			print_unsigned(&s, (unsigned long)va_arg(va, void *), 16, props);
+			print_unsigned(&s, (unsigned long)va_arg(args, void *), 16, props);
 			break;
 		case 's':
-			print_str(&s, va_arg(va, const char *), props);
+			print_str(&s, va_arg(args, const char *), props);
 			break;
 		default:
 			addchar(&s, f);
 			break;
 		}
 	}
+	va_end(args);
 	*s.buffer = 0;
 	return s.added;
 }
-- 
2.25.1

[PATCH v4 18/30] lib/printf: Add support for printing wide strings

[Nikos Nikoleris]

This change adds support for wide strings (u16*) to printf()
variants. This feature is used by a future change.

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Ricardo Koller <ricarkol@google.com>
---
 lib/printf.c | 101 ++++++++++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 100 insertions(+), 1 deletion(-)

diff --git a/lib/printf.c b/lib/printf.c
index d600199b..27041bb4 100644
--- a/lib/printf.c
+++ b/lib/printf.c
@@ -58,6 +58,102 @@ static void print_str(pstream_t *p, const char *s, strprops_t props)
 	}
 }
 
+/*
+ * Adapted from drivers/firmware/efi/libstub/vsprintf.c
+ */
+static u32 utf16_to_utf32(const u16 **s16)
+{
+	u16 c0, c1;
+
+	c0 = *(*s16)++;
+	/* not a surrogate */
+	if ((c0 & 0xf800) != 0xd800)
+		return c0;
+	/* invalid: low surrogate instead of high */
+	if (c0 & 0x0400)
+		return 0xfffd;
+	c1 = **s16;
+	/* invalid: missing low surrogate */
+	if ((c1 & 0xfc00) != 0xdc00)
+		return 0xfffd;
+	/* valid surrogate pair */
+	++(*s16);
+	return (0x10000 - (0xd800 << 10) - 0xdc00) + (c0 << 10) + c1;
+}
+
+/*
+ * Adapted from drivers/firmware/efi/libstub/vsprintf.c
+ */
+static size_t utf16s_utf8nlen(const u16 *s16, size_t maxlen)
+{
+	size_t len, clen;
+
+	for (len = 0; len < maxlen && *s16; len += clen) {
+		u16 c0 = *s16++;
+
+		/* First, get the length for a BMP character */
+		clen = 1 + (c0 >= 0x80) + (c0 >= 0x800);
+		if (len + clen > maxlen)
+			break;
+		/*
+		 * If this is a high surrogate, and we're already at maxlen, we
+		 * can't include the character if it's a valid surrogate pair.
+		 * Avoid accessing one extra word just to check if it's valid
+		 * or not.
+		 */
+		if ((c0 & 0xfc00) == 0xd800) {
+			if (len + clen == maxlen)
+				break;
+			if ((*s16 & 0xfc00) == 0xdc00) {
+				++s16;
+				++clen;
+			}
+		}
+	}
+
+	return len;
+}
+
+/*
+ * Adapted from drivers/firmware/efi/libstub/vsprintf.c
+ */
+static void print_wstring(pstream_t *p, const u16 *s, strprops_t props)
+{
+	const u16 *ws = (const u16 *)s;
+	size_t pos = 0, size = p->remain + 1, len = utf16s_utf8nlen(ws, props.precision);
+
+	while (len-- > 0) {
+		u32 c32 = utf16_to_utf32(&ws);
+		u8 *s8;
+		size_t clen;
+
+		if (c32 < 0x80) {
+			addchar(p, c32);
+			continue;
+		}
+
+		/* Number of trailing octets */
+		clen = 1 + (c32 >= 0x800) + (c32 >= 0x10000);
+
+		len -= clen;
+		s8 = (u8 *)(p->buffer - p->added + pos);
+
+		/* Avoid writing partial character */
+		addchar(p, '\0');
+		pos += clen;
+		if (pos >= size)
+			continue;
+
+		/* Set high bits of leading octet */
+		*s8 = (0xf00 >> 1) >> clen;
+		/* Write trailing octets in reverse order */
+		for (s8 += clen; clen; --clen, c32 >>= 6)
+			*s8-- = 0x80 | (c32 & 0x3f);
+		/* Set low bits of leading octet */
+		*s8 |= c32;
+	}
+}
+
 static char digits[16] = "0123456789abcdef";
 
 static void print_int(pstream_t *ps, long long n, int base, strprops_t props)
@@ -305,7 +401,10 @@ morefmt:
 			print_unsigned(&s, (unsigned long)va_arg(args, void *), 16, props);
 			break;
 		case 's':
-			print_str(&s, va_arg(args, const char *), props);
+			if (nlong)
+				print_wstring(&s, va_arg(args, const u16 *), props);
+			else
+				print_str(&s, va_arg(args, const char *), props);
 			break;
 		default:
 			addchar(&s, f);
-- 
2.25.1

[PATCH v4 19/30] lib/efi: Add support for getting the cmdline

[Nikos Nikoleris]

This change adds support for discovering the command line arguments,
as a string. Then, we parse this string to populate __argc and __argv
for EFI tests.

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
---
 lib/argv.c      |   2 +-
 lib/argv.h      |   1 +
 lib/efi.c       | 102 ++++++++++++++++++++++++++++++++++++++++++++++++
 lib/linux/efi.h |  20 ++++++++++
 4 files changed, 124 insertions(+), 1 deletion(-)

diff --git a/lib/argv.c b/lib/argv.c
index 9ffa673e..fa5ff9ae 100644
--- a/lib/argv.c
+++ b/lib/argv.c
@@ -41,7 +41,7 @@ void __setup_args(void)
 	__argc = argv - __argv;
 }
 
-static void setup_args(const char *args)
+void setup_args(const char *args)
 {
 	if (!args)
 		return;
diff --git a/lib/argv.h b/lib/argv.h
index 1fd746dc..0fa77725 100644
--- a/lib/argv.h
+++ b/lib/argv.h
@@ -9,6 +9,7 @@
 #define _ARGV_H_
 
 extern void __setup_args(void);
+extern void setup_args(const char *args);
 extern void setup_args_progname(const char *args);
 extern void setup_env(char *env, int size);
 extern void add_setup_arg(const char *arg);
diff --git a/lib/efi.c b/lib/efi.c
index 64cc9789..f524ec9b 100644
--- a/lib/efi.c
+++ b/lib/efi.c
@@ -8,6 +8,9 @@
  */
 
 #include "efi.h"
+#include <argv.h>
+#include <stdlib.h>
+#include <ctype.h>
 #include <libcflat.h>
 #include <asm/setup.h>
 
@@ -96,6 +99,80 @@ static void efi_exit(efi_status_t code)
 	efi_rs_call(reset_system, EFI_RESET_SHUTDOWN, code, 0, NULL);
 }
 
+/* Adapted from drivers/firmware/efi/libstub/efi-stub.c */
+static char *efi_convert_cmdline(struct efi_loaded_image_64 *image, int *cmd_line_len)
+{
+	const u16 *s2;
+	unsigned long cmdline_addr = 0;
+	int options_chars = image->load_options_size;
+	const u16 *options = image->load_options;
+	int options_bytes = 0, safe_options_bytes = 0;  /* UTF-8 bytes */
+	bool in_quote = false;
+	efi_status_t status;
+	const int COMMAND_LINE_SIZE = 2048;
+
+	if (options) {
+		s2 = options;
+		while (options_bytes < COMMAND_LINE_SIZE && options_chars--) {
+			u16 c = *s2++;
+
+			if (c < 0x80) {
+				if (c == L'\0' || c == L'\n')
+					break;
+				if (c == L'"')
+					in_quote = !in_quote;
+				else if (!in_quote && isspace((char)c))
+					safe_options_bytes = options_bytes;
+
+				options_bytes++;
+				continue;
+			}
+
+			/*
+			 * Get the number of UTF-8 bytes corresponding to a
+			 * UTF-16 character.
+			 * The first part handles everything in the BMP.
+			 */
+			options_bytes += 2 + (c >= 0x800);
+			/*
+			 * Add one more byte for valid surrogate pairs. Invalid
+			 * surrogates will be replaced with 0xfffd and take up
+			 * only 3 bytes.
+			 */
+			if ((c & 0xfc00) == 0xd800) {
+				/*
+				 * If the very last word is a high surrogate,
+				 * we must ignore it since we can't access the
+				 * low surrogate.
+				 */
+				if (!options_chars) {
+					options_bytes -= 3;
+				} else if ((*s2 & 0xfc00) == 0xdc00) {
+					options_bytes++;
+					options_chars--;
+					s2++;
+				}
+			}
+		}
+		if (options_bytes >= COMMAND_LINE_SIZE) {
+			options_bytes = safe_options_bytes;
+			printf("Command line is too long: truncated to %d bytes\n",
+			       options_bytes);
+		}
+	}
+
+	options_bytes++;        /* NUL termination */
+
+	status = efi_bs_call(allocate_pool, EFI_LOADER_DATA, options_bytes, (void **)&cmdline_addr);
+	if (status != EFI_SUCCESS)
+		return NULL;
+
+	snprintf((char *)cmdline_addr, options_bytes, "%.*ls", options_bytes - 1, options);
+
+	*cmd_line_len = options_bytes;
+	return (char *)cmdline_addr;
+}
+
 efi_status_t efi_main(efi_handle_t handle, efi_system_table_t *sys_tab)
 {
 	int ret;
@@ -109,6 +186,31 @@ efi_status_t efi_main(efi_handle_t handle, efi_system_table_t *sys_tab)
 	unsigned long map_size = 0, desc_size = 0, key = 0, buff_size = 0;
 	u32 desc_ver;
 
+	/* Helper variables needed to get the cmdline */
+	struct efi_loaded_image_64 *image;
+	efi_guid_t loaded_image_proto = LOADED_IMAGE_PROTOCOL_GUID;
+	char *cmdline_ptr = NULL;
+	int cmdline_size = 0;
+
+	/*
+	 * Get a handle to the loaded image protocol.  This is used to get
+	 * information about the running image, such as size and the command
+	 * line.
+	 */
+	status = efi_bs_call(handle_protocol, handle, &loaded_image_proto, (void *)&image);
+	if (status != EFI_SUCCESS) {
+		printf("Failed to get loaded image protocol\n");
+		goto efi_main_error;
+	}
+
+	cmdline_ptr = efi_convert_cmdline(image, &cmdline_size);
+	if (!cmdline_ptr) {
+		printf("getting command line via LOADED_IMAGE_PROTOCOL\n");
+		status = EFI_OUT_OF_RESOURCES;
+		goto efi_main_error;
+	}
+	setup_args(cmdline_ptr);
+
 	/* Set up efi_bootinfo */
 	efi_bootinfo.mem_map.map = &map;
 	efi_bootinfo.mem_map.map_size = &map_size;
diff --git a/lib/linux/efi.h b/lib/linux/efi.h
index 455625aa..9a1cf87b 100644
--- a/lib/linux/efi.h
+++ b/lib/linux/efi.h
@@ -60,6 +60,8 @@ typedef guid_t efi_guid_t;
 
 #define ACPI_TABLE_GUID EFI_GUID(0xeb9d2d30, 0x2d88, 0x11d3, 0x9a, 0x16, 0x00, 0x90, 0x27, 0x3f, 0xc1, 0x4d)
 
+#define LOADED_IMAGE_PROTOCOL_GUID EFI_GUID(0x5b1b31a1, 0x9562, 0x11d2,  0x8e, 0x3f, 0x00, 0xa0, 0xc9, 0x69, 0x72, 0x3b)
+
 typedef struct {
 	efi_guid_t guid;
 	void *table;
@@ -416,6 +418,24 @@ struct efi_boot_memmap {
 	unsigned long           *buff_size;
 };
 
+#define __aligned_u64 u64 __attribute__((aligned(8)))
+
+struct efi_loaded_image_64 {
+	u32			revision;
+	efi_handle_t		parent_handle;
+	efi_system_table_t	*system_table;
+	efi_handle_t		device_handle;
+	void			*file_path;
+	void			*reserved;
+	u32			load_options_size;
+	void			*load_options;
+	void			*image_base;
+	__aligned_u64		image_size;
+	unsigned int		image_code_type;
+	unsigned int		image_data_type;
+	efi_status_t		(__efiapi * unload)(efi_handle_t image_handle);
+};
+
 #define efi_bs_call(func, ...) efi_system_table->boottime->func(__VA_ARGS__)
 #define efi_rs_call(func, ...) efi_system_table->runtime->func(__VA_ARGS__)
 
-- 
2.25.1

[PATCH v4 20/30] arm/arm64: Rename etext to _etext

[Nikos Nikoleris]

From: Andrew Jones <drjones@redhat.com>

Rename etext to the more popular _etext allowing different linker
scripts to more easily be used.

Signed-off-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Ricardo Koller <ricarkol@google.com>
---
 arm/flat.lds    | 2 +-
 lib/arm/setup.c | 4 ++--
 2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/arm/flat.lds b/arm/flat.lds
index 47fcb649..9016ac9f 100644
--- a/arm/flat.lds
+++ b/arm/flat.lds
@@ -27,7 +27,7 @@ SECTIONS
     PROVIDE(_text = .);
     .text : { *(.init) *(.text) *(.text.*) }
     . = ALIGN(64K);
-    PROVIDE(etext = .);
+    PROVIDE(_etext = .);
 
     PROVIDE(reloc_start = .);
     .rela.dyn : { *(.rela.dyn) }
diff --git a/lib/arm/setup.c b/lib/arm/setup.c
index 59b0aedd..03a4098e 100644
--- a/lib/arm/setup.c
+++ b/lib/arm/setup.c
@@ -33,7 +33,7 @@
 #define NR_EXTRA_MEM_REGIONS	16
 #define NR_INITIAL_MEM_REGIONS	(MAX_DT_MEM_REGIONS + NR_EXTRA_MEM_REGIONS)
 
-extern unsigned long etext;
+extern unsigned long _etext;
 
 char *initrd;
 u32 initrd_size;
@@ -157,7 +157,7 @@ unsigned int mem_region_get_flags(phys_addr_t paddr)
 
 static void mem_regions_add_assumed(void)
 {
-	phys_addr_t code_end = (phys_addr_t)(unsigned long)&etext;
+	phys_addr_t code_end = (phys_addr_t)(unsigned long)&_etext;
 	struct mem_region *r;
 
 	r = mem_region_find(code_end - 1);
-- 
2.25.1

[PATCH v4 21/30] lib: Avoid ms_abi for calls related to EFI on arm64

[Nikos Nikoleris]

x86_64 requires that EFI calls use the ms_abi calling convention. For
arm64 this is unnecessary.

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
---
 lib/linux/efi.h | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/lib/linux/efi.h b/lib/linux/efi.h
index 9a1cf87b..53748dd4 100644
--- a/lib/linux/efi.h
+++ b/lib/linux/efi.h
@@ -33,7 +33,11 @@ typedef u16 efi_char16_t;		/* UNICODE character */
 typedef u64 efi_physical_addr_t;
 typedef void *efi_handle_t;
 
+#ifdef __x86_64__
 #define __efiapi __attribute__((ms_abi))
+#else
+#define __efiapi
+#endif
 
 /*
  * The UEFI spec and EDK2 reference implementation both define EFI_GUID as
-- 
2.25.1

[PATCH v4 22/30] arm64: Add a new type of memory type flag MR_F_RESERVED

[Nikos Nikoleris]

From: Andrew Jones <drjones@redhat.com>

This will be used by future change to add PTE entries for special EFI
memory regions.

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Ricardo Koller <ricarkol@google.com>
---
 lib/arm/asm/setup.h | 1 +
 lib/arm/mmu.c       | 4 ++++
 2 files changed, 5 insertions(+)

diff --git a/lib/arm/asm/setup.h b/lib/arm/asm/setup.h
index f0e70b11..64cd379b 100644
--- a/lib/arm/asm/setup.h
+++ b/lib/arm/asm/setup.h
@@ -15,6 +15,7 @@ extern int nr_cpus;
 
 #define MR_F_IO			(1U << 0)
 #define MR_F_CODE		(1U << 1)
+#define MR_F_RESERVED		(1U << 2)
 #define MR_F_UNKNOWN		(1U << 31)
 
 struct mem_region {
diff --git a/lib/arm/mmu.c b/lib/arm/mmu.c
index e1a72fe4..931be985 100644
--- a/lib/arm/mmu.c
+++ b/lib/arm/mmu.c
@@ -174,6 +174,10 @@ void *setup_mmu(phys_addr_t phys_end, void *unused)
 	for (r = mem_regions; r->end; ++r) {
 		if (r->flags & MR_F_IO) {
 			continue;
+		} else if (r->flags & MR_F_RESERVED) {
+			/* Reserved pages need to be writable for whatever reserved them */
+			mmu_set_range_ptes(mmu_idmap, r->start, r->start, r->end,
+					   __pgprot(PTE_WBWA));
 		} else if (r->flags & MR_F_CODE) {
 			/* armv8 requires code shared between EL1 and EL0 to be read-only */
 			mmu_set_range_ptes(mmu_idmap, r->start, r->start, r->end,
-- 
2.25.1

[PATCH v4 23/30] arm64: Add a setup sequence for systems that boot through EFI

[Nikos Nikoleris]

This change implements an alternative setup sequence for the system
when we are booting through EFI. The memory map is discovered through
EFI boot services and devices through ACPI.

This change is based on a change initially proposed by
Andrew Jones <drjones@redhat.com>

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
---
 arm/cstart.S        |   1 +
 arm/cstart64.S      |   1 +
 lib/arm/asm/setup.h |   8 ++
 lib/arm/setup.c     | 181 +++++++++++++++++++++++++++++++++++++++++++-
 lib/linux/efi.h     |   1 +
 5 files changed, 190 insertions(+), 2 deletions(-)

diff --git a/arm/cstart.S b/arm/cstart.S
index 7036e67f..3dd71ed9 100644
--- a/arm/cstart.S
+++ b/arm/cstart.S
@@ -242,6 +242,7 @@ asm_mmu_disable:
  *
  * Input r0 is the stack top, which is the exception stacks base
  */
+.globl exceptions_init
 exceptions_init:
 	mrc	p15, 0, r2, c1, c0, 0	@ read SCTLR
 	bic	r2, #CR_V		@ SCTLR.V := 0
diff --git a/arm/cstart64.S b/arm/cstart64.S
index e4ab7d06..223c1092 100644
--- a/arm/cstart64.S
+++ b/arm/cstart64.S
@@ -265,6 +265,7 @@ asm_mmu_disable:
  * Vectors
  */
 
+.globl exceptions_init
 exceptions_init:
 	adrp	x4, vector_table
 	add	x4, x4, :lo12:vector_table
diff --git a/lib/arm/asm/setup.h b/lib/arm/asm/setup.h
index 64cd379b..06069116 100644
--- a/lib/arm/asm/setup.h
+++ b/lib/arm/asm/setup.h
@@ -38,4 +38,12 @@ extern unsigned int mem_region_get_flags(phys_addr_t paddr);
 
 void setup(const void *fdt, phys_addr_t freemem_start);
 
+#ifdef CONFIG_EFI
+
+#include <efi.h>
+
+efi_status_t setup_efi(efi_bootinfo_t *efi_bootinfo);
+
+#endif
+
 #endif /* _ASMARM_SETUP_H_ */
diff --git a/lib/arm/setup.c b/lib/arm/setup.c
index 03a4098e..cab19b1e 100644
--- a/lib/arm/setup.c
+++ b/lib/arm/setup.c
@@ -33,7 +33,7 @@
 #define NR_EXTRA_MEM_REGIONS	16
 #define NR_INITIAL_MEM_REGIONS	(MAX_DT_MEM_REGIONS + NR_EXTRA_MEM_REGIONS)
 
-extern unsigned long _etext;
+extern unsigned long _text, _etext, _data, _edata;
 
 char *initrd;
 u32 initrd_size;
@@ -43,7 +43,10 @@ int nr_cpus;
 
 static struct mem_region __initial_mem_regions[NR_INITIAL_MEM_REGIONS + 1];
 struct mem_region *mem_regions = __initial_mem_regions;
-phys_addr_t __phys_offset, __phys_end;
+phys_addr_t __phys_offset = (phys_addr_t)-1, __phys_end = 0;
+
+extern void exceptions_init(void);
+extern void asm_mmu_disable(void);
 
 int mpidr_to_cpu(uint64_t mpidr)
 {
@@ -289,3 +292,177 @@ void setup(const void *fdt, phys_addr_t freemem_start)
 	if (!(auxinfo.flags & AUXINFO_MMU_OFF))
 		setup_vm();
 }
+
+#ifdef CONFIG_EFI
+
+#include <efi.h>
+
+static efi_status_t setup_rsdp(efi_bootinfo_t *efi_bootinfo)
+{
+	efi_status_t status;
+	struct acpi_table_rsdp *rsdp;
+
+	/*
+	 * RSDP resides in an EFI_ACPI_RECLAIM_MEMORY region, which is not used
+	 * by kvm-unit-tests arm64 memory allocator. So it is not necessary to
+	 * copy the data structure to another memory region to prevent
+	 * unintentional overwrite.
+	 */
+	status = efi_get_system_config_table(ACPI_20_TABLE_GUID, (void **)&rsdp);
+	if (status != EFI_SUCCESS)
+		return status;
+
+	set_efi_rsdp(rsdp);
+
+	return EFI_SUCCESS;
+}
+
+static efi_status_t efi_mem_init(efi_bootinfo_t *efi_bootinfo)
+{
+	int i;
+	unsigned long free_mem_pages = 0;
+	unsigned long free_mem_start = 0;
+	struct efi_boot_memmap *map = &(efi_bootinfo->mem_map);
+	efi_memory_desc_t *buffer = *map->map;
+	efi_memory_desc_t *d = NULL;
+	phys_addr_t base, top;
+	struct mem_region *r;
+	uintptr_t text = (uintptr_t)&_text, etext = __ALIGN((uintptr_t)&_etext, 4096);
+	uintptr_t data = (uintptr_t)&_data, edata = __ALIGN((uintptr_t)&_edata, 4096);
+
+	/*
+	 * Record the largest free EFI_CONVENTIONAL_MEMORY region
+	 * which will be used to set up the memory allocator, so that
+	 * the memory allocator can work in the largest free
+	 * continuous memory region.
+	 */
+	for (i = 0, r = &mem_regions[0]; i < *(map->map_size); i += *(map->desc_size), ++r) {
+		d = (efi_memory_desc_t *)(&((u8 *)buffer)[i]);
+
+		r->start = d->phys_addr;
+		r->end = d->phys_addr + d->num_pages * EFI_PAGE_SIZE;
+
+		switch (d->type) {
+		case EFI_RESERVED_TYPE:
+		case EFI_LOADER_DATA:
+		case EFI_BOOT_SERVICES_CODE:
+		case EFI_BOOT_SERVICES_DATA:
+		case EFI_RUNTIME_SERVICES_CODE:
+		case EFI_RUNTIME_SERVICES_DATA:
+		case EFI_UNUSABLE_MEMORY:
+		case EFI_ACPI_RECLAIM_MEMORY:
+		case EFI_ACPI_MEMORY_NVS:
+		case EFI_PAL_CODE:
+			r->flags = MR_F_RESERVED;
+			break;
+		case EFI_MEMORY_MAPPED_IO:
+		case EFI_MEMORY_MAPPED_IO_PORT_SPACE:
+			r->flags = MR_F_IO;
+			break;
+		case EFI_LOADER_CODE:
+			if (r->start <= text && r->end > text) {
+				/* This is the unit test region. Flag the code separately. */
+				phys_addr_t tmp = r->end;
+
+				assert(etext <= data);
+				assert(edata <= r->end);
+				r->flags = MR_F_CODE;
+				r->end = data;
+				++r;
+				r->start = data;
+				r->end = tmp;
+			} else {
+				r->flags = MR_F_RESERVED;
+			}
+			break;
+		case EFI_CONVENTIONAL_MEMORY:
+			if (free_mem_pages < d->num_pages) {
+				free_mem_pages = d->num_pages;
+				free_mem_start = d->phys_addr;
+			}
+			break;
+		}
+
+		if (!(r->flags & MR_F_IO)) {
+			if (r->start < __phys_offset)
+				__phys_offset = r->start;
+			if (r->end > __phys_end)
+				__phys_end = r->end;
+		}
+	}
+	__phys_end &= PHYS_MASK;
+	asm_mmu_disable();
+
+	if (free_mem_pages == 0)
+		return EFI_OUT_OF_RESOURCES;
+
+	assert(sizeof(long) == 8 || free_mem_start < (3ul << 30));
+
+	phys_alloc_init(free_mem_start, free_mem_pages << EFI_PAGE_SHIFT);
+	phys_alloc_set_minimum_alignment(SMP_CACHE_BYTES);
+
+	phys_alloc_get_unused(&base, &top);
+	base = PAGE_ALIGN(base);
+	top = top & PAGE_MASK;
+	assert(sizeof(long) == 8 || !(base >> 32));
+	if (sizeof(long) != 8 && (top >> 32) != 0)
+		top = ((uint64_t)1 << 32);
+	page_alloc_init_area(0, base >> PAGE_SHIFT, top >> PAGE_SHIFT);
+	page_alloc_ops_enable();
+
+	return EFI_SUCCESS;
+}
+
+efi_status_t setup_efi(efi_bootinfo_t *efi_bootinfo)
+{
+	efi_status_t status;
+
+	struct thread_info *ti = current_thread_info();
+
+	memset(ti, 0, sizeof(*ti));
+
+	exceptions_init();
+
+	status = efi_mem_init(efi_bootinfo);
+	if (status != EFI_SUCCESS) {
+		printf("Failed to initialize memory: ");
+		switch (status) {
+		case EFI_OUT_OF_RESOURCES:
+			printf("No free memory region\n");
+			break;
+		default:
+			printf("Unknown error\n");
+			break;
+		}
+		return status;
+	}
+
+	status = setup_rsdp(efi_bootinfo);
+	if (status != EFI_SUCCESS) {
+		printf("Cannot find RSDP in EFI system table\n");
+		return status;
+	}
+
+	psci_set_conduit();
+	cpu_init();
+	/* cpu_init must be called before thread_info_init */
+	thread_info_init(current_thread_info(), 0);
+	/* mem_init must be called before io_init */
+	io_init();
+
+	timer_save_state();
+	if (initrd) {
+		/* environ is currently the only file in the initrd */
+		char *env = malloc(initrd_size);
+
+		memcpy(env, initrd, initrd_size);
+		setup_env(env, initrd_size);
+	}
+
+	if (!(auxinfo.flags & AUXINFO_MMU_OFF))
+		setup_vm();
+
+	return EFI_SUCCESS;
+}
+
+#endif
diff --git a/lib/linux/efi.h b/lib/linux/efi.h
index 53748dd4..89f9a9e0 100644
--- a/lib/linux/efi.h
+++ b/lib/linux/efi.h
@@ -63,6 +63,7 @@ typedef guid_t efi_guid_t;
 	(c) & 0xff, ((c) >> 8) & 0xff, d } }
 
 #define ACPI_TABLE_GUID EFI_GUID(0xeb9d2d30, 0x2d88, 0x11d3, 0x9a, 0x16, 0x00, 0x90, 0x27, 0x3f, 0xc1, 0x4d)
+#define ACPI_20_TABLE_GUID EFI_GUID(0x8868e871, 0xe4f1, 0x11d3,  0xbc, 0x22, 0x00, 0x80, 0xc7, 0x3c, 0x88, 0x81)
 
 #define LOADED_IMAGE_PROTOCOL_GUID EFI_GUID(0x5b1b31a1, 0x9562, 0x11d2,  0x8e, 0x3f, 0x00, 0xa0, 0xc9, 0x69, 0x72, 0x3b)
 
-- 
2.25.1

[PATCH v4 24/30] arm64: Copy code from GNU-EFI

[Nikos Nikoleris]

This change adds unmodified dependencies that we need from GNU-EFI in
order to build arm64 EFI apps.

GNU-EFI sources from  https://git.code.sf.net/p/gnu-efi/code v3.0.14

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Ricardo Koller <ricarkol@google.com>
---
 arm/efi/crt0-efi-aarch64.S  | 130 ++++++++++++++++++++++++++++++++++++
 arm/efi/elf_aarch64_efi.lds |  63 +++++++++++++++++
 arm/efi/reloc_aarch64.c     |  97 +++++++++++++++++++++++++++
 3 files changed, 290 insertions(+)
 create mode 100644 arm/efi/crt0-efi-aarch64.S
 create mode 100644 arm/efi/elf_aarch64_efi.lds
 create mode 100644 arm/efi/reloc_aarch64.c

diff --git a/arm/efi/crt0-efi-aarch64.S b/arm/efi/crt0-efi-aarch64.S
new file mode 100644
index 00000000..d50e78dd
--- /dev/null
+++ b/arm/efi/crt0-efi-aarch64.S
@@ -0,0 +1,130 @@
+/*
+ * crt0-efi-aarch64.S - PE/COFF header for AArch64 EFI applications
+ *
+ * Copright (C) 2014 Linaro Ltd. <ard.biesheuvel@linaro.org>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice and this list of conditions, without modification.
+ * 2. The name of the author may not be used to endorse or promote products
+ *    derived from this software without specific prior written permission.
+ *
+ * Alternatively, this software may be distributed under the terms of the
+ * GNU General Public License as published by the Free Software Foundation;
+ * either version 2 of the License, or (at your option) any later version.
+ */
+
+	.section	.text.head
+
+	/*
+	 * Magic "MZ" signature for PE/COFF
+	 */
+	.globl	ImageBase
+ImageBase:
+	.ascii	"MZ"
+	.skip	58				// 'MZ' + pad + offset == 64
+	.long	pe_header - ImageBase		// Offset to the PE header.
+pe_header:
+	.ascii	"PE"
+	.short 	0
+coff_header:
+	.short	0xaa64				// AArch64
+	.short	2				// nr_sections
+	.long	0 				// TimeDateStamp
+	.long	0				// PointerToSymbolTable
+	.long	0				// NumberOfSymbols
+	.short	section_table - optional_header	// SizeOfOptionalHeader
+	.short	0x206				// Characteristics.
+						// IMAGE_FILE_DEBUG_STRIPPED |
+						// IMAGE_FILE_EXECUTABLE_IMAGE |
+						// IMAGE_FILE_LINE_NUMS_STRIPPED
+optional_header:
+	.short	0x20b				// PE32+ format
+	.byte	0x02				// MajorLinkerVersion
+	.byte	0x14				// MinorLinkerVersion
+	.long	_data - _start			// SizeOfCode
+	.long	_data_size			// SizeOfInitializedData
+	.long	0				// SizeOfUninitializedData
+	.long	_start - ImageBase		// AddressOfEntryPoint
+	.long	_start - ImageBase		// BaseOfCode
+
+extra_header_fields:
+	.quad	0				// ImageBase
+	.long	0x1000				// SectionAlignment
+	.long	0x200				// FileAlignment
+	.short	0				// MajorOperatingSystemVersion
+	.short	0				// MinorOperatingSystemVersion
+	.short	0				// MajorImageVersion
+	.short	0				// MinorImageVersion
+	.short	0				// MajorSubsystemVersion
+	.short	0				// MinorSubsystemVersion
+	.long	0				// Win32VersionValue
+
+	.long	_edata - ImageBase		// SizeOfImage
+
+	// Everything before the kernel image is considered part of the header
+	.long	_start - ImageBase		// SizeOfHeaders
+	.long	0				// CheckSum
+	.short	EFI_SUBSYSTEM			// Subsystem
+	.short	0				// DllCharacteristics
+	.quad	0				// SizeOfStackReserve
+	.quad	0				// SizeOfStackCommit
+	.quad	0				// SizeOfHeapReserve
+	.quad	0				// SizeOfHeapCommit
+	.long	0				// LoaderFlags
+	.long	0x6				// NumberOfRvaAndSizes
+
+	.quad	0				// ExportTable
+	.quad	0				// ImportTable
+	.quad	0				// ResourceTable
+	.quad	0				// ExceptionTable
+	.quad	0				// CertificationTable
+	.quad	0				// BaseRelocationTable
+
+	// Section table
+section_table:
+	.ascii	".text\0\0\0"
+	.long	_data - _start		// VirtualSize
+	.long	_start - ImageBase	// VirtualAddress
+	.long	_data - _start		// SizeOfRawData
+	.long	_start - ImageBase	// PointerToRawData
+
+	.long	0		// PointerToRelocations (0 for executables)
+	.long	0		// PointerToLineNumbers (0 for executables)
+	.short	0		// NumberOfRelocations  (0 for executables)
+	.short	0		// NumberOfLineNumbers  (0 for executables)
+	.long	0x60000020	// Characteristics (section flags)
+
+	.ascii	".data\0\0\0"
+	.long	_data_size		// VirtualSize
+	.long	_data - ImageBase	// VirtualAddress
+	.long	_data_size		// SizeOfRawData
+	.long	_data - ImageBase	// PointerToRawData
+
+	.long	0		// PointerToRelocations (0 for executables)
+	.long	0		// PointerToLineNumbers (0 for executables)
+	.short	0		// NumberOfRelocations  (0 for executables)
+	.short	0		// NumberOfLineNumbers  (0 for executables)
+	.long	0xc0000040	// Characteristics (section flags)
+
+	.align		12
+_start:
+	stp		x29, x30, [sp, #-32]!
+	mov		x29, sp
+
+	stp		x0, x1, [sp, #16]
+	mov		x2, x0
+	mov		x3, x1
+	adr		x0, ImageBase
+	adrp		x1, _DYNAMIC
+	add		x1, x1, #:lo12:_DYNAMIC
+	bl		_relocate
+	cbnz		x0, 0f
+
+	ldp		x0, x1, [sp, #16]
+	bl		efi_main
+
+0:	ldp		x29, x30, [sp], #32
+	ret
diff --git a/arm/efi/elf_aarch64_efi.lds b/arm/efi/elf_aarch64_efi.lds
new file mode 100644
index 00000000..836d9825
--- /dev/null
+++ b/arm/efi/elf_aarch64_efi.lds
@@ -0,0 +1,63 @@
+OUTPUT_FORMAT("elf64-littleaarch64", "elf64-littleaarch64", "elf64-littleaarch64")
+OUTPUT_ARCH(aarch64)
+ENTRY(_start)
+SECTIONS
+{
+  .text 0x0 : {
+    _text = .;
+    *(.text.head)
+    *(.text)
+    *(.text.*)
+    *(.gnu.linkonce.t.*)
+    *(.srodata)
+    *(.rodata*)
+    . = ALIGN(16);
+  }
+  _etext = .;
+  _text_size = . - _text;
+  .dynamic  : { *(.dynamic) }
+  .data : ALIGN(4096)
+  {
+   _data = .;
+   *(.sdata)
+   *(.data)
+   *(.data1)
+   *(.data.*)
+   *(.got.plt)
+   *(.got)
+
+   /* the EFI loader doesn't seem to like a .bss section, so we stick
+      it all into .data: */
+   . = ALIGN(16);
+   _bss = .;
+   *(.sbss)
+   *(.scommon)
+   *(.dynbss)
+   *(.bss)
+   *(COMMON)
+   . = ALIGN(16);
+   _bss_end = .;
+  }
+
+  .rela.dyn : { *(.rela.dyn) }
+  .rela.plt : { *(.rela.plt) }
+  .rela.got : { *(.rela.got) }
+  .rela.data : { *(.rela.data) *(.rela.data*) }
+  . = ALIGN(512);
+  _edata = .;
+  _data_size = . - _data;
+
+  . = ALIGN(4096);
+  .dynsym   : { *(.dynsym) }
+  . = ALIGN(4096);
+  .dynstr   : { *(.dynstr) }
+  . = ALIGN(4096);
+  .note.gnu.build-id : { *(.note.gnu.build-id) }
+  /DISCARD/ :
+  {
+    *(.rel.reloc)
+    *(.eh_frame)
+    *(.note.GNU-stack)
+  }
+  .comment 0 : { *(.comment) }
+}
diff --git a/arm/efi/reloc_aarch64.c b/arm/efi/reloc_aarch64.c
new file mode 100644
index 00000000..08672796
--- /dev/null
+++ b/arm/efi/reloc_aarch64.c
@@ -0,0 +1,97 @@
+/* reloc_aarch64.c - position independent x86 ELF shared object relocator
+   Copyright (C) 2014 Linaro Ltd. <ard.biesheuvel@linaro.org>
+   Copyright (C) 1999 Hewlett-Packard Co.
+	Contributed by David Mosberger <davidm@hpl.hp.com>.
+
+    All rights reserved.
+
+    Redistribution and use in source and binary forms, with or without
+    modification, are permitted provided that the following conditions
+    are met:
+
+    * Redistributions of source code must retain the above copyright
+      notice, this list of conditions and the following disclaimer.
+    * Redistributions in binary form must reproduce the above
+      copyright notice, this list of conditions and the following
+      disclaimer in the documentation and/or other materials
+      provided with the distribution.
+    * Neither the name of Hewlett-Packard Co. nor the names of its
+      contributors may be used to endorse or promote products derived
+      from this software without specific prior written permission.
+
+    THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND
+    CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES,
+    INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
+    MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+    DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS
+    BE LIABLE FOR ANYDIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY,
+    OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+    PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+    PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+    THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR
+    TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF
+    THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+    SUCH DAMAGE.
+*/
+
+#include <efi.h>
+#include <efilib.h>
+
+#include <elf.h>
+
+EFI_STATUS _relocate (long ldbase, Elf64_Dyn *dyn,
+		      EFI_HANDLE image EFI_UNUSED,
+		      EFI_SYSTEM_TABLE *systab EFI_UNUSED)
+{
+	long relsz = 0, relent = 0;
+	Elf64_Rela *rel = 0;
+	unsigned long *addr;
+	int i;
+
+	for (i = 0; dyn[i].d_tag != DT_NULL; ++i) {
+		switch (dyn[i].d_tag) {
+			case DT_RELA:
+				rel = (Elf64_Rela*)
+					((unsigned long)dyn[i].d_un.d_ptr
+					 + ldbase);
+				break;
+
+			case DT_RELASZ:
+				relsz = dyn[i].d_un.d_val;
+				break;
+
+			case DT_RELAENT:
+				relent = dyn[i].d_un.d_val;
+				break;
+
+			default:
+				break;
+		}
+	}
+
+	if (!rel && relent == 0)
+		return EFI_SUCCESS;
+
+	if (!rel || relent == 0)
+		return EFI_LOAD_ERROR;
+
+	while (relsz > 0) {
+		/* apply the relocs */
+		switch (ELF64_R_TYPE (rel->r_info)) {
+			case R_AARCH64_NONE:
+				break;
+
+			case R_AARCH64_RELATIVE:
+				addr = (unsigned long *)
+					(ldbase + rel->r_offset);
+				*addr = ldbase + rel->r_addend;
+				break;
+
+			default:
+				break;
+		}
+		rel = (Elf64_Rela*) ((char *) rel + relent);
+		relsz -= relent;
+	}
+	return EFI_SUCCESS;
+}
-- 
2.25.1

[PATCH v4 25/30] arm64: Change GNU-EFI imported code to use defined types

[Nikos Nikoleris]

Convert some types to avoid dependency on gnu-efi's <efi.h> and
<efilib.h>.

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
---
 arm/efi/reloc_aarch64.c | 9 +++------
 1 file changed, 3 insertions(+), 6 deletions(-)

diff --git a/arm/efi/reloc_aarch64.c b/arm/efi/reloc_aarch64.c
index 08672796..fa0cd6bc 100644
--- a/arm/efi/reloc_aarch64.c
+++ b/arm/efi/reloc_aarch64.c
@@ -34,14 +34,11 @@
     SUCH DAMAGE.
 */
 
-#include <efi.h>
-#include <efilib.h>
-
+#include "efi.h"
 #include <elf.h>
 
-EFI_STATUS _relocate (long ldbase, Elf64_Dyn *dyn,
-		      EFI_HANDLE image EFI_UNUSED,
-		      EFI_SYSTEM_TABLE *systab EFI_UNUSED)
+efi_status_t _relocate(long ldbase, Elf64_Dyn *dyn, efi_handle_t image,
+		       efi_system_table_t *sys_tab)
 {
 	long relsz = 0, relent = 0;
 	Elf64_Rela *rel = 0;
-- 
2.25.1

[PATCH v4 26/30] arm64: Use code from the gnu-efi when booting with EFI

[Nikos Nikoleris]

arm/efi/crt0-efi-aarch64.S defines the header and the handover
sequence from EFI to a efi_main. This change includes the whole file
in arm/cstart64.S when we compile with EFI support.

In addition, we change the handover code in arm/efi/crt0-efi-aarch64.S
to align the stack pointer. This alignment is necessary because we
make assumptions about cpu0's stack alignment and most importantly we
place its thread_info at the bottom of this stack.

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
---
 arm/cstart64.S             |  6 ++++++
 arm/efi/crt0-efi-aarch64.S | 19 +++++++++++++++----
 2 files changed, 21 insertions(+), 4 deletions(-)

diff --git a/arm/cstart64.S b/arm/cstart64.S
index 223c1092..cbd6b511 100644
--- a/arm/cstart64.S
+++ b/arm/cstart64.S
@@ -15,6 +15,10 @@
 #include <asm/thread_info.h>
 #include <asm/sysreg.h>
 
+#ifdef CONFIG_EFI
+#include "efi/crt0-efi-aarch64.S"
+#else
+
 .macro zero_range, tmp1, tmp2
 9998:	cmp	\tmp1, \tmp2
 	b.eq	9997f
@@ -107,6 +111,8 @@ start:
 	bl	exit
 	b	halt
 
+#endif
+
 .text
 
 /*
diff --git a/arm/efi/crt0-efi-aarch64.S b/arm/efi/crt0-efi-aarch64.S
index d50e78dd..5d0dc04a 100644
--- a/arm/efi/crt0-efi-aarch64.S
+++ b/arm/efi/crt0-efi-aarch64.S
@@ -111,10 +111,17 @@ section_table:
 
 	.align		12
 _start:
-	stp		x29, x30, [sp, #-32]!
+	stp		x29, x30, [sp, #-16]!
+
+	/* Align sp; this is necessary due to way we store cpu0's thread_info */
 	mov		x29, sp
+	mov		x30, sp
+	and		x30, x30, #THREAD_MASK
+	mov		sp, x30
+	str		x29, [sp, #-16]!
+
+	stp		x0, x1, [sp, #-16]!
 
-	stp		x0, x1, [sp, #16]
 	mov		x2, x0
 	mov		x3, x1
 	adr		x0, ImageBase
@@ -123,8 +130,12 @@ _start:
 	bl		_relocate
 	cbnz		x0, 0f
 
-	ldp		x0, x1, [sp, #16]
+	ldp		x0, x1, [sp], #16
 	bl		efi_main
 
-0:	ldp		x29, x30, [sp], #32
+	/* Restore sp */
+	ldr		x30, [sp], #16
+	mov             sp, x30
+
+0:	ldp		x29, x30, [sp], #16
 	ret
-- 
2.25.1

[PATCH v4 27/30] lib: Avoid external dependency in libelf

[Nikos Nikoleris]

There is just a small number of definitions we need from
uapi/linux/elf.h and asm/elf.h and the relocation code is
self-contained.

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Ricardo Koller <ricarkol@google.com>
---
 arm/efi/reloc_aarch64.c |  3 +--
 lib/elf.h               | 57 +++++++++++++++++++++++++++++++++++++++++
 2 files changed, 58 insertions(+), 2 deletions(-)
 create mode 100644 lib/elf.h

diff --git a/arm/efi/reloc_aarch64.c b/arm/efi/reloc_aarch64.c
index fa0cd6bc..3f6d9a6d 100644
--- a/arm/efi/reloc_aarch64.c
+++ b/arm/efi/reloc_aarch64.c
@@ -34,8 +34,7 @@
     SUCH DAMAGE.
 */
 
-#include "efi.h"
-#include <elf.h>
+#include <efi.h>
 
 efi_status_t _relocate(long ldbase, Elf64_Dyn *dyn, efi_handle_t image,
 		       efi_system_table_t *sys_tab)
diff --git a/lib/elf.h b/lib/elf.h
new file mode 100644
index 00000000..abd5cf4b
--- /dev/null
+++ b/lib/elf.h
@@ -0,0 +1,57 @@
+/* SPDX-License-Identifier: LGPL-2.0-or-later */
+/*
+ * Relevant definitions from uapi/linux/elf.h and asm/elf.h
+ */
+
+#ifndef _ELF_H_
+#define _ELF_H_
+
+#include <libcflat.h>
+
+/* 64-bit ELF base types. */
+typedef u64	Elf64_Addr;
+typedef u64	Elf64_Xword;
+typedef s64	Elf64_Sxword;
+
+typedef struct {
+	Elf64_Sxword d_tag;             /* entry tag value */
+	union {
+		Elf64_Xword d_val;
+		Elf64_Addr d_ptr;
+	} d_un;
+} Elf64_Dyn;
+
+typedef struct elf64_rel {
+	Elf64_Addr r_offset;    /* Location at which to apply the action */
+	Elf64_Xword r_info;     /* index and type of relocation */
+} Elf64_Rel;
+
+typedef struct elf64_rela {
+	Elf64_Addr r_offset;    /* Location at which to apply the action */
+	Elf64_Xword r_info;     /* index and type of relocation */
+	Elf64_Sxword r_addend;  /* Constant addend used to compute value */
+} Elf64_Rela;
+
+/* This is the info that is needed to parse the dynamic section of the file */
+#define DT_NULL		0
+#define DT_RELA		7
+#define DT_RELASZ	8
+#define DT_RELAENT	9
+
+/* x86 relocation types. */
+#define R_X86_64_NONE		0       /* No reloc */
+#define R_X86_64_RELATIVE	8       /* Adjust by program base */
+
+
+/*
+ * AArch64 static relocation types.
+ */
+
+/* Miscellaneous. */
+#define R_AARCH64_NONE		256
+#define R_AARCH64_RELATIVE	1027
+
+/* The following are used with relocations */
+#define ELF64_R_TYPE(i)		((i) & 0xffffffff)
+
+#endif /* _ELF_H_ */
-- 
2.25.1

[PATCH v4 28/30] arm64: Add support for efi in Makefile

[Nikos Nikoleris]

Users can now build kvm-unit-tests as efi apps by supplying an extra
argument when invoking configure:

$> ./configure --enable-efi

This patch is based on an earlier version by
Andrew Jones <drjones@redhat.com>

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Ricardo Koller <ricarkol@google.com>
---
 arm/Makefile.arm    |  6 ++++++
 arm/Makefile.arm64  | 18 ++++++++++++++----
 arm/Makefile.common | 45 ++++++++++++++++++++++++++++++++++-----------
 configure           | 15 ++++++++++++---
 4 files changed, 66 insertions(+), 18 deletions(-)

diff --git a/arm/Makefile.arm b/arm/Makefile.arm
index 01fd4c7b..2ce00f52 100644
--- a/arm/Makefile.arm
+++ b/arm/Makefile.arm
@@ -7,6 +7,10 @@ bits = 32
 ldarch = elf32-littlearm
 machine = -marm -mfpu=vfp
 
+ifeq ($(CONFIG_EFI),y)
+$(error Cannot build arm32 tests as EFI apps)
+endif
+
 # stack.o relies on frame pointers.
 KEEP_FRAME_POINTER := y
 
@@ -32,6 +36,8 @@ cflatobjs += lib/arm/stack.o
 cflatobjs += lib/ldiv32.o
 cflatobjs += lib/arm/ldivmod.o
 
+exe = flat
+
 # arm specific tests
 tests =
 
diff --git a/arm/Makefile.arm64 b/arm/Makefile.arm64
index 6dff6cad..eada7f9a 100644
--- a/arm/Makefile.arm64
+++ b/arm/Makefile.arm64
@@ -31,11 +31,21 @@ endif
 
 OBJDIRS += lib/arm64
 
+ifeq ($(CONFIG_EFI),y)
+# avoid jump tables before all relocations have been processed
+arm/efi/reloc_aarch64.o: CFLAGS += -fno-jump-tables
+cflatobjs += arm/efi/reloc_aarch64.o
+
+exe = efi
+else
+exe = flat
+endif
+
 # arm64 specific tests
-tests = $(TEST_DIR)/timer.flat
-tests += $(TEST_DIR)/micro-bench.flat
-tests += $(TEST_DIR)/cache.flat
-tests += $(TEST_DIR)/debug.flat
+tests = $(TEST_DIR)/timer.$(exe)
+tests += $(TEST_DIR)/micro-bench.$(exe)
+tests += $(TEST_DIR)/cache.$(exe)
+tests += $(TEST_DIR)/debug.$(exe)
 
 include $(SRCDIR)/$(TEST_DIR)/Makefile.common
 
diff --git a/arm/Makefile.common b/arm/Makefile.common
index c02dd906..e251f6a8 100644
--- a/arm/Makefile.common
+++ b/arm/Makefile.common
@@ -4,14 +4,14 @@
 # Authors: Andrew Jones <drjones@redhat.com>
 #
 
-tests-common  = $(TEST_DIR)/selftest.flat
-tests-common += $(TEST_DIR)/spinlock-test.flat
-tests-common += $(TEST_DIR)/pci-test.flat
-tests-common += $(TEST_DIR)/pmu.flat
-tests-common += $(TEST_DIR)/gic.flat
-tests-common += $(TEST_DIR)/psci.flat
-tests-common += $(TEST_DIR)/sieve.flat
-tests-common += $(TEST_DIR)/pl031.flat
+tests-common  = $(TEST_DIR)/selftest.$(exe)
+tests-common += $(TEST_DIR)/spinlock-test.$(exe)
+tests-common += $(TEST_DIR)/pci-test.$(exe)
+tests-common += $(TEST_DIR)/pmu.$(exe)
+tests-common += $(TEST_DIR)/gic.$(exe)
+tests-common += $(TEST_DIR)/psci.$(exe)
+tests-common += $(TEST_DIR)/sieve.$(exe)
+tests-common += $(TEST_DIR)/pl031.$(exe)
 
 tests-all = $(tests-common) $(tests)
 all: directories $(tests-all)
@@ -54,6 +54,9 @@ cflatobjs += lib/arm/smp.o
 cflatobjs += lib/arm/delay.o
 cflatobjs += lib/arm/gic.o lib/arm/gic-v2.o lib/arm/gic-v3.o
 cflatobjs += lib/arm/timer.o
+ifeq ($(CONFIG_EFI),y)
+cflatobjs += lib/efi.o
+endif
 
 ifeq ($(CONFIG_EFI),y)
 cflatobjs += lib/acpi.o
@@ -65,6 +68,25 @@ libeabi = lib/arm/libeabi.a
 eabiobjs = lib/arm/eabi_compat.o
 
 FLATLIBS = $(libcflat) $(LIBFDT_archive) $(libeabi)
+
+ifeq ($(CONFIG_EFI),y)
+%.so: EFI_LDFLAGS += -defsym=EFI_SUBSYSTEM=0xa --no-undefined
+%.so: %.o $(FLATLIBS) $(SRCDIR)/arm/efi/elf_aarch64_efi.lds $(cstart.o)
+	$(CC) $(CFLAGS) -c -o $(@:.so=.aux.o) $(SRCDIR)/lib/auxinfo.c \
+		-DPROGNAME=\"$(@:.so=.efi)\" -DAUXFLAGS=$(AUXFLAGS)
+	$(LD) $(EFI_LDFLAGS) -o $@ -T $(SRCDIR)/arm/efi/elf_aarch64_efi.lds \
+		$(filter %.o, $^) $(FLATLIBS) $(@:.so=.aux.o) \
+		$(EFI_LIBS)
+	$(RM) $(@:.so=.aux.o)
+
+%.efi: %.so
+	$(call arch_elf_check, $^)
+	$(OBJCOPY) \
+		-j .text -j .sdata -j .data -j .dynamic -j .dynsym \
+		-j .rel -j .rela -j .rel.* -j .rela.* -j .rel* -j .rela* \
+		-j .reloc \
+		-O binary $^ $@
+else
 %.elf: LDFLAGS = -nostdlib $(arch_LDFLAGS)
 %.elf: %.o $(FLATLIBS) $(SRCDIR)/arm/flat.lds $(cstart.o)
 	$(CC) $(CFLAGS) -c -o $(@:.elf=.aux.o) $(SRCDIR)/lib/auxinfo.c \
@@ -78,13 +100,14 @@ FLATLIBS = $(libcflat) $(LIBFDT_archive) $(libeabi)
 	$(call arch_elf_check, $^)
 	$(OBJCOPY) -O binary $^ $@
 	@chmod a-x $@
+endif
 
 $(libeabi): $(eabiobjs)
 	$(AR) rcs $@ $^
 
 arm_clean: asm_offsets_clean
-	$(RM) $(TEST_DIR)/*.{o,flat,elf} $(libeabi) $(eabiobjs) \
-	      $(TEST_DIR)/.*.d lib/arm/.*.d
+	$(RM) $(TEST_DIR)/*.{o,flat,elf,so,efi} $(libeabi) $(eabiobjs) \
+	      $(TEST_DIR)/.*.d $(TEST_DIR)/efi/.*.d lib/arm/.*.d
 
 generated-files = $(asm-offsets)
-$(tests-all:.flat=.o) $(cstart.o) $(cflatobjs): $(generated-files)
+$(tests-all:.$(exe)=.o) $(cstart.o) $(cflatobjs): $(generated-files)
diff --git a/configure b/configure
index c36fd290..7806dab0 100755
--- a/configure
+++ b/configure
@@ -208,14 +208,19 @@ else
     fi
 fi
 
-if [ "$efi" ] && [ "$arch" != "x86_64" ]; then
+if [ "$efi" ] && [ "$arch" != "x86_64" ] && [ "$arch" != "arm64" ]; then
     echo "--[enable|disable]-efi is not supported for $arch"
     usage
 fi
 
 if [ -z "$page_size" ]; then
-    [ "$arch" = "arm64" ] && page_size="65536"
-    [ "$arch" = "arm" ] && page_size="4096"
+    if [ "$efi" = 'y' ] && [ "$arch" = "arm64" ]; then
+        page_size="4096"
+    elif [ "$arch" = "arm64" ]; then
+        page_size="65536"
+    elif [ "$arch" = "arm" ]; then
+        page_size="4096"
+    fi
 else
     if [ "$arch" != "arm64" ]; then
         echo "--page-size is not supported for $arch"
@@ -230,6 +235,10 @@ else
         echo "arm64 doesn't support page size of $page_size"
         usage
     fi
+    if [ "$efi" = 'y' ] && [ "$page_size" != "4096" ]; then
+        echo "efi must use 4K pages"
+        exit 1
+    fi
 fi
 
 [ -z "$processor" ] && processor="$arch"
-- 
2.25.1

[PATCH v4 29/30] lib: arm: Print test exit status

[Nikos Nikoleris]

From: Alexandru Elisei <alexandru.elisei@arm.com>

The arm tests can be run under kvmtool, which doesn't emulate a chr-testdev
device. Print the test exit status to make it possible for the runner
scripts to pick it up when they have support for it.

Signed-off-by: Alexandru Elisei <alexandru.elisei@arm.com>
Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Andrew Jones <andrew.jones@linux.dev>
---
 lib/arm/io.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/lib/arm/io.c b/lib/arm/io.c
index 19f93490..c15e57c4 100644
--- a/lib/arm/io.c
+++ b/lib/arm/io.c
@@ -149,6 +149,13 @@ extern void halt(int code);
 
 void exit(int code)
 {
+	/*
+	 * Print the test return code in the following format which is
+	 * consistent with powerpc and s390x. The runner can pick it
+	 * up when chr-testdev is not present.
+	 */
+	printf("\nEXIT: STATUS=%d\n", ((code) << 1) | 1);
+
 	chr_testdev_exit(code);
 	psci_system_off();
 	halt(code);
-- 
2.25.1

[PATCH v4 30/30] arm64: Add an efi/run script

[Nikos Nikoleris]

This change adds a efi/run script inspired by the one in x86. This
script will setup a folder with the test compiled as an EFI app and a
startup.nsh script. The script launches QEMU providing an image with
EDKII and the path to the folder with the test which is executed
automatically.

For example:

$> ./arm/efi/run ./arm/selftest.efi setup smp=2 mem=256

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Ricardo Koller <ricarkol@google.com>
---
 arm/Makefile.common  |  1 +
 arm/dummy.c          | 12 +++++++++
 arm/efi/run          | 61 ++++++++++++++++++++++++++++++++++++++++++++
 arm/run              | 14 +++++++---
 scripts/runtime.bash | 13 +++++++---
 5 files changed, 94 insertions(+), 7 deletions(-)
 create mode 100644 arm/dummy.c
 create mode 100755 arm/efi/run

diff --git a/arm/Makefile.common b/arm/Makefile.common
index e251f6a8..90a6ff3a 100644
--- a/arm/Makefile.common
+++ b/arm/Makefile.common
@@ -12,6 +12,7 @@ tests-common += $(TEST_DIR)/gic.$(exe)
 tests-common += $(TEST_DIR)/psci.$(exe)
 tests-common += $(TEST_DIR)/sieve.$(exe)
 tests-common += $(TEST_DIR)/pl031.$(exe)
+tests-common += $(TEST_DIR)/dummy.$(exe)
 
 tests-all = $(tests-common) $(tests)
 all: directories $(tests-all)
diff --git a/arm/dummy.c b/arm/dummy.c
new file mode 100644
index 00000000..7033bb7c
--- /dev/null
+++ b/arm/dummy.c
@@ -0,0 +1,12 @@
+#include "libcflat.h"
+
+int main(int argc, char **argv)
+{
+	/*
+	 * scripts/runtime.bash uses this test as a canary to determine if the
+	 * basic setup is functional.  Print a magic string to let runtime.bash
+	 * know that all is well.
+	 */
+	printf("Dummy Hello World!");
+	return 0;
+}
diff --git a/arm/efi/run b/arm/efi/run
new file mode 100755
index 00000000..dfff717a
--- /dev/null
+++ b/arm/efi/run
@@ -0,0 +1,61 @@
+#!/bin/bash
+
+set -e
+
+if [ $# -eq 0 ]; then
+	echo "Usage $0 TEST_CASE [QEMU_ARGS]"
+	exit 2
+fi
+
+if [ ! -f config.mak ]; then
+	echo "run './configure --enable-efi && make' first. See ./configure -h"
+	exit 2
+fi
+source config.mak
+source scripts/arch-run.bash
+source scripts/common.bash
+
+: "${EFI_SRC:=$(realpath "$(dirname "$0")/../")}"
+: "${EFI_UEFI:=/usr/share/qemu-efi-aarch64/QEMU_EFI.fd}"
+: "${EFI_TEST:=efi-tests}"
+: "${EFI_CASE:=$(basename $1 .efi)}"
+
+if [ ! -f "$EFI_UEFI" ]; then
+	echo "UEFI firmware not found: $EFI_UEFI"
+	echo "Please install the UEFI firmware to this path"
+	echo "Or specify the correct path with the env variable EFI_UEFI"
+	exit 2
+fi
+
+# Remove the TEST_CASE from $@
+shift 1
+
+# Fish out the arguments for the test, they should be the next string
+# after the "-append" option
+qemu_args=()
+cmd_args=()
+while (( "$#" )); do
+	if [ "$1" = "-append" ]; then
+		cmd_args=$2
+		shift 2
+	else
+		qemu_args+=("$1")
+		shift 1
+	fi
+done
+
+if [ "$EFI_CASE" = "_NO_FILE_4Uhere_" ]; then
+	EFI_CASE=dummy
+fi
+
+: "${EFI_CASE_DIR:="$EFI_TEST/$EFI_CASE"}"
+mkdir -p "$EFI_CASE_DIR"
+
+cp "$EFI_SRC/$EFI_CASE.efi" "$EFI_TEST/$EFI_CASE/"
+echo "@echo -off" > "$EFI_TEST/$EFI_CASE/startup.nsh"
+echo "$EFI_CASE.efi" "${cmd_args[@]}" >> "$EFI_TEST/$EFI_CASE/startup.nsh"
+
+EFI_RUN=y $TEST_DIR/run \
+       -bios "$EFI_UEFI" \
+       -drive file.dir="$EFI_TEST/$EFI_CASE/",file.driver=vvfat,file.rw=on,format=raw,if=virtio \
+       "${qemu_args[@]}"
diff --git a/arm/run b/arm/run
index 12848912..62f845b3 100755
--- a/arm/run
+++ b/arm/run
@@ -65,8 +65,10 @@ if $qemu $M -chardev testdev,id=id -initrd . 2>&1 \
 	exit 2
 fi
 
-chr_testdev='-device virtio-serial-device'
-chr_testdev+=' -device virtconsole,chardev=ctd -chardev testdev,id=ctd'
+if [ "$EFI_RUN" != "y" ]; then
+	chr_testdev='-device virtio-serial-device'
+	chr_testdev+=' -device virtconsole,chardev=ctd -chardev testdev,id=ctd'
+fi
 
 pci_testdev=
 if $qemu $M -device '?' 2>&1 | grep pci-testdev > /dev/null; then
@@ -75,7 +77,11 @@ fi
 
 A="-accel $ACCEL"
 command="$qemu -nodefaults $M $A -cpu $processor $chr_testdev $pci_testdev"
-command+=" -display none -serial stdio -kernel"
+command+=" -display none -serial stdio"
 command="$(migration_cmd) $(timeout_cmd) $command"
 
-run_qemu $command "$@"
+if [ "$EFI_RUN" = "y" ]; then
+	ENVIRON_DEFAULT=n run_qemu_status $command "$@"
+else
+	run_qemu $command -kernel "$@"
+fi
diff --git a/scripts/runtime.bash b/scripts/runtime.bash
index f8794e9a..13eade26 100644
--- a/scripts/runtime.bash
+++ b/scripts/runtime.bash
@@ -130,11 +130,18 @@ function run()
         done
     fi
 
-    last_line=$(premature_failure > >(tail -1)) && {
+    log=$(premature_failure) && {
         skip=true
-        if [ "${CONFIG_EFI}" == "y" ] && [[ "${last_line}" =~ "Dummy Hello World!" ]]; then
-            skip=false
+        if [ "${CONFIG_EFI}" == "y" ]; then
+            if [ "$ARCH" == "x86_64" ] &&
+               [[ "$(tail -1 <<<"$log")" =~ "Dummy Hello World!" ]]; then
+                   skip=false
+            elif [ "$ARCH" == "arm64" ] &&
+               [[ "$(tail -2 <<<"$log" | head -1)" =~ "Dummy Hello World!" ]]; then
+                   skip=false
+            fi
         fi
+
         if [ ${skip} == true ]; then
             print_result "SKIP" $testname "" "$last_line"
             return 77
-- 
2.25.1

Re: [PATCH v4 04/30] lib: Apply Lindent to acpi.{c,h}

[Shaoqin Huang]

Hi Nikos,

On 2/13/23 18:17, Nikos Nikoleris wrote:
> The change was done by modifying Linux's scripts/Lindent to use 100
> columns instead of 80.
> 
> Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
> ---
>   lib/acpi.c |  70 ++++++++++++------------
>   lib/acpi.h | 157 ++++++++++++++++++++++++++---------------------------
>   2 files changed, 111 insertions(+), 116 deletions(-)
> 
> diff --git a/lib/acpi.c b/lib/acpi.c
> index de275caf..836156a1 100644
> --- a/lib/acpi.c
> +++ b/lib/acpi.c
> @@ -36,47 +36,45 @@ static struct rsdp_descriptor *get_rsdp(void)
>   }
>   #endif /* CONFIG_EFI */
>   
> -void* find_acpi_table_addr(u32 sig)
> +void *find_acpi_table_addr(u32 sig)
>   {
> -    struct rsdp_descriptor *rsdp;
> -    struct rsdt_descriptor_rev1 *rsdt;
> -    void *end;
> -    int i;
> +	struct rsdp_descriptor *rsdp;
> +	struct rsdt_descriptor_rev1 *rsdt;
> +	void *end;
> +	int i;
>   
> -    /* FACS is special... */
> -    if (sig == FACS_SIGNATURE) {
> -        struct fadt_descriptor_rev1 *fadt;
> -        fadt = find_acpi_table_addr(FACP_SIGNATURE);
> -        if (!fadt) {
> -            return NULL;
> -        }
> -        return (void*)(ulong)fadt->firmware_ctrl;
> -    }
> +	/* FACS is special... */
> +	if (sig == FACS_SIGNATURE) {
> +		struct fadt_descriptor_rev1 *fadt;
> +		fadt = find_acpi_table_addr(FACP_SIGNATURE);
> +		if (!fadt)
> +			return NULL;
>   
> -    rsdp = get_rsdp();
> -    if (rsdp == NULL) {
> -        printf("Can't find RSDP\n");
> -        return 0;
> -    }
> +		return (void *)(ulong) fadt->firmware_ctrl;
> +	}
>   
> -    if (sig == RSDP_SIGNATURE) {
> -        return rsdp;
> -    }
> +	rsdp = get_rsdp();
> +	if (rsdp == NULL) {
> +		printf("Can't find RSDP\n");
> +		return NULL;
> +	}
>   
> -    rsdt = (void*)(ulong)rsdp->rsdt_physical_address;
> -    if (!rsdt || rsdt->signature != RSDT_SIGNATURE)
> -        return 0;
> +	if (sig == RSDP_SIGNATURE)
> +		return rsdp;
>   
> -    if (sig == RSDT_SIGNATURE) {
> -        return rsdt;
> -    }
> +	rsdt = (void *)(ulong) rsdp->rsdt_physical_address;
> +	if (!rsdt || rsdt->signature != RSDT_SIGNATURE)
> +		return NULL;
> +
> +	if (sig == RSDT_SIGNATURE)
> +		return rsdt;
>   
> -    end = (void*)rsdt + rsdt->length;
> -    for (i=0; (void*)&rsdt->table_offset_entry[i] < end; i++) {
> -        struct acpi_table *t = (void*)(ulong)rsdt->table_offset_entry[i];
> -        if (t && t->signature == sig) {
> -            return t;
> -        }
> -    }
> -   return NULL;
> +	end = (void *)rsdt + rsdt->length;
> +	for (i = 0; (void *)&rsdt->table_offset_entry[i] < end; i++) {
> +		struct acpi_table *t = (void *)(ulong) rsdt->table_offset_entry[i];
> +		if (t && t->signature == sig) {
> +			return t;
> +		}
> +	}
> +	return NULL;
>   }
> diff --git a/lib/acpi.h b/lib/acpi.h
> index 1e89840c..b67bbe19 100644
> --- a/lib/acpi.h
> +++ b/lib/acpi.h
> @@ -3,7 +3,7 @@
>   
>   #include "libcflat.h"
>   
> -#define ACPI_SIGNATURE(c1, c2, c3, c4) \
> +#define ACPI_SIGNATURE(c1, c2, c3, c4)				\
>   	((c1) | ((c2) << 8) | ((c3) << 16) | ((c4) << 24))
>   
>   #define RSDP_SIGNATURE ACPI_SIGNATURE('R','S','D','P')
> @@ -11,102 +11,99 @@
>   #define FACP_SIGNATURE ACPI_SIGNATURE('F','A','C','P')
>   #define FACS_SIGNATURE ACPI_SIGNATURE('F','A','C','S')
>   
> -
> -#define ACPI_SIGNATURE_8BYTE(c1, c2, c3, c4, c5, c6, c7, c8) \
> -	((uint64_t)(ACPI_SIGNATURE(c1, c2, c3, c4))) |       \
> +#define ACPI_SIGNATURE_8BYTE(c1, c2, c3, c4, c5, c6, c7, c8)	\
> +	((uint64_t)(ACPI_SIGNATURE(c1, c2, c3, c4))) |		\
>   	((uint64_t)(ACPI_SIGNATURE(c5, c6, c7, c8)) << 32)
>   
>   #define RSDP_SIGNATURE_8BYTE (ACPI_SIGNATURE_8BYTE('R', 'S', 'D', ' ', 'P', 'T', 'R', ' '))
>   
> -struct rsdp_descriptor {        /* Root System Descriptor Pointer */
> -    u64 signature;              /* ACPI signature, contains "RSD PTR " */
> -    u8  checksum;               /* To make sum of struct == 0 */
> -    u8  oem_id [6];             /* OEM identification */
> -    u8  revision;               /* Must be 0 for 1.0, 2 for 2.0 */
> -    u32 rsdt_physical_address;  /* 32-bit physical address of RSDT */
> -    u32 length;                 /* XSDT Length in bytes including hdr */
> -    u64 xsdt_physical_address;  /* 64-bit physical address of XSDT */
> -    u8  extended_checksum;      /* Checksum of entire table */
> -    u8  reserved [3];           /* Reserved field must be 0 */
> +struct rsdp_descriptor {	/* Root System Descriptor Pointer */
> +	u64 signature;		/* ACPI signature, contains "RSD PTR " */
> +	u8 checksum;		/* To make sum of struct == 0 */
> +	u8 oem_id[6];		/* OEM identification */
> +	u8 revision;		/* Must be 0 for 1.0, 2 for 2.0 */
> +	u32 rsdt_physical_address;	/* 32-bit physical address of RSDT */
> +	u32 length;		/* XSDT Length in bytes including hdr */
> +	u64 xsdt_physical_address;	/* 64-bit physical address of XSDT */
> +	u8 extended_checksum;	/* Checksum of entire table */
> +	u8 reserved[3];		/* Reserved field must be 0 */
>   };
>   
> -#define ACPI_TABLE_HEADER_DEF   /* ACPI common table header */ \
> -    u32 signature;          /* ACPI signature (4 ASCII characters) */ \
> -    u32 length;                 /* Length of table, in bytes, including header */ \
> -    u8  revision;               /* ACPI Specification minor version # */ \
> -    u8  checksum;               /* To make sum of entire table == 0 */ \
> -    u8  oem_id [6];             /* OEM identification */ \
> -    u8  oem_table_id [8];       /* OEM table identification */ \
> -    u32 oem_revision;           /* OEM revision number */ \
> -    u8  asl_compiler_id [4];    /* ASL compiler vendor ID */ \
> -    u32 asl_compiler_revision;  /* ASL compiler revision number */
> +#define ACPI_TABLE_HEADER_DEF		/* ACPI common table header */			\
> +	u32 signature;			/* ACPI signature (4 ASCII characters) */	\
> +	u32 length;			/* Length of table, in bytes, including header */ \
> +	u8  revision;			/* ACPI Specification minor version # */	\
> +	u8  checksum;			/* To make sum of entire table == 0 */		\
> +	u8  oem_id [6];			/* OEM identification */			\
                   ^
> +	u8  oem_table_id [8];		/* OEM table identification */			\
                         ^
> +	u32 oem_revision;		/* OEM revision number */			\
> +	u8  asl_compiler_id [4];	/* ASL compiler vendor ID */			\
                            ^
nit: These space should also be deleted.

Thanks,
Shaoqin
> +	u32 asl_compiler_revision;	/* ASL compiler revision number */
>   
>   struct acpi_table {
> -    ACPI_TABLE_HEADER_DEF
> -    char data[0];
> +	ACPI_TABLE_HEADER_DEF
> +	char data[0];
>   };
>   
>   struct rsdt_descriptor_rev1 {
> -    ACPI_TABLE_HEADER_DEF
> -    u32 table_offset_entry[0];
> +	ACPI_TABLE_HEADER_DEF
> +	u32 table_offset_entry[1];
>   };
>   
> -struct fadt_descriptor_rev1
> -{
> -    ACPI_TABLE_HEADER_DEF     /* ACPI common table header */
> -    u32 firmware_ctrl;          /* Physical address of FACS */
> -    u32 dsdt;                   /* Physical address of DSDT */
> -    u8  model;                  /* System Interrupt Model */
> -    u8  reserved1;              /* Reserved */
> -    u16 sci_int;                /* System vector of SCI interrupt */
> -    u32 smi_cmd;                /* Port address of SMI command port */
> -    u8  acpi_enable;            /* Value to write to smi_cmd to enable ACPI */
> -    u8  acpi_disable;           /* Value to write to smi_cmd to disable ACPI */
> -    u8  S4bios_req;             /* Value to write to SMI CMD to enter S4BIOS state */
> -    u8  reserved2;              /* Reserved - must be zero */
> -    u32 pm1a_evt_blk;           /* Port address of Power Mgt 1a acpi_event Reg Blk */
> -    u32 pm1b_evt_blk;           /* Port address of Power Mgt 1b acpi_event Reg Blk */
> -    u32 pm1a_cnt_blk;           /* Port address of Power Mgt 1a Control Reg Blk */
> -    u32 pm1b_cnt_blk;           /* Port address of Power Mgt 1b Control Reg Blk */
> -    u32 pm2_cnt_blk;            /* Port address of Power Mgt 2 Control Reg Blk */
> -    u32 pm_tmr_blk;             /* Port address of Power Mgt Timer Ctrl Reg Blk */
> -    u32 gpe0_blk;               /* Port addr of General Purpose acpi_event 0 Reg Blk */
> -    u32 gpe1_blk;               /* Port addr of General Purpose acpi_event 1 Reg Blk */
> -    u8  pm1_evt_len;            /* Byte length of ports at pm1_x_evt_blk */
> -    u8  pm1_cnt_len;            /* Byte length of ports at pm1_x_cnt_blk */
> -    u8  pm2_cnt_len;            /* Byte Length of ports at pm2_cnt_blk */
> -    u8  pm_tmr_len;             /* Byte Length of ports at pm_tm_blk */
> -    u8  gpe0_blk_len;           /* Byte Length of ports at gpe0_blk */
> -    u8  gpe1_blk_len;           /* Byte Length of ports at gpe1_blk */
> -    u8  gpe1_base;              /* Offset in gpe model where gpe1 events start */
> -    u8  reserved3;              /* Reserved */
> -    u16 plvl2_lat;              /* Worst case HW latency to enter/exit C2 state */
> -    u16 plvl3_lat;              /* Worst case HW latency to enter/exit C3 state */
> -    u16 flush_size;             /* Size of area read to flush caches */
> -    u16 flush_stride;           /* Stride used in flushing caches */
> -    u8  duty_offset;            /* Bit location of duty cycle field in p_cnt reg */
> -    u8  duty_width;             /* Bit width of duty cycle field in p_cnt reg */
> -    u8  day_alrm;               /* Index to day-of-month alarm in RTC CMOS RAM */
> -    u8  mon_alrm;               /* Index to month-of-year alarm in RTC CMOS RAM */
> -    u8  century;                /* Index to century in RTC CMOS RAM */
> -    u8  reserved4;              /* Reserved */
> -    u8  reserved4a;             /* Reserved */
> -    u8  reserved4b;             /* Reserved */
> +struct fadt_descriptor_rev1 {
> +	ACPI_TABLE_HEADER_DEF	/* ACPI common table header */
> +	u32 firmware_ctrl;	/* Physical address of FACS */
> +	u32 dsdt;		/* Physical address of DSDT */
> +	u8 model;		/* System Interrupt Model */
> +	u8 reserved1;		/* Reserved */
> +	u16 sci_int;		/* System vector of SCI interrupt */
> +	u32 smi_cmd;		/* Port address of SMI command port */
> +	u8 acpi_enable;		/* Value to write to smi_cmd to enable ACPI */
> +	u8 acpi_disable;	/* Value to write to smi_cmd to disable ACPI */
> +	u8 S4bios_req;		/* Value to write to SMI CMD to enter S4BIOS state */
> +	u8 reserved2;		/* Reserved - must be zero */
> +	u32 pm1a_evt_blk;	/* Port address of Power Mgt 1a acpi_event Reg Blk */
> +	u32 pm1b_evt_blk;	/* Port address of Power Mgt 1b acpi_event Reg Blk */
> +	u32 pm1a_cnt_blk;	/* Port address of Power Mgt 1a Control Reg Blk */
> +	u32 pm1b_cnt_blk;	/* Port address of Power Mgt 1b Control Reg Blk */
> +	u32 pm2_cnt_blk;	/* Port address of Power Mgt 2 Control Reg Blk */
> +	u32 pm_tmr_blk;		/* Port address of Power Mgt Timer Ctrl Reg Blk */
> +	u32 gpe0_blk;		/* Port addr of General Purpose acpi_event 0 Reg Blk */
> +	u32 gpe1_blk;		/* Port addr of General Purpose acpi_event 1 Reg Blk */
> +	u8 pm1_evt_len;		/* Byte length of ports at pm1_x_evt_blk */
> +	u8 pm1_cnt_len;		/* Byte length of ports at pm1_x_cnt_blk */
> +	u8 pm2_cnt_len;		/* Byte Length of ports at pm2_cnt_blk */
> +	u8 pm_tmr_len;		/* Byte Length of ports at pm_tm_blk */
> +	u8 gpe0_blk_len;	/* Byte Length of ports at gpe0_blk */
> +	u8 gpe1_blk_len;	/* Byte Length of ports at gpe1_blk */
> +	u8 gpe1_base;		/* Offset in gpe model where gpe1 events start */
> +	u8 reserved3;		/* Reserved */
> +	u16 plvl2_lat;		/* Worst case HW latency to enter/exit C2 state */
> +	u16 plvl3_lat;		/* Worst case HW latency to enter/exit C3 state */
> +	u16 flush_size;		/* Size of area read to flush caches */
> +	u16 flush_stride;	/* Stride used in flushing caches */
> +	u8 duty_offset;		/* Bit location of duty cycle field in p_cnt reg */
> +	u8 duty_width;		/* Bit width of duty cycle field in p_cnt reg */
> +	u8 day_alrm;		/* Index to day-of-month alarm in RTC CMOS RAM */
> +	u8 mon_alrm;		/* Index to month-of-year alarm in RTC CMOS RAM */
> +	u8 century;		/* Index to century in RTC CMOS RAM */
> +	u8 reserved4;		/* Reserved */
> +	u8 reserved4a;		/* Reserved */
> +	u8 reserved4b;		/* Reserved */
>   };
>   
> -struct facs_descriptor_rev1
> -{
> -    u32 signature;           /* ACPI Signature */
> -    u32 length;                 /* Length of structure, in bytes */
> -    u32 hardware_signature;     /* Hardware configuration signature */
> -    u32 firmware_waking_vector; /* ACPI OS waking vector */
> -    u32 global_lock;            /* Global Lock */
> -    u32 S4bios_f        : 1;    /* Indicates if S4BIOS support is present */
> -    u32 reserved1       : 31;   /* Must be 0 */
> -    u8  reserved3 [40];         /* Reserved - must be zero */
> +struct facs_descriptor_rev1 {
> +	u32 signature;		/* ACPI Signature */
> +	u32 length;		/* Length of structure, in bytes */
> +	u32 hardware_signature;	/* Hardware configuration signature */
> +	u32 firmware_waking_vector;	/* ACPI OS waking vector */
> +	u32 global_lock;	/* Global Lock */
> +	u32 S4bios_f:1;		/* Indicates if S4BIOS support is present */
> +	u32 reserved1:31;	/* Must be 0 */
> +	u8 reserved3[40];	/* Reserved - must be zero */
>   };
>   
>   void set_efi_rsdp(struct rsdp_descriptor *rsdp);
> -void* find_acpi_table_addr(u32 sig);
> +void *find_acpi_table_addr(u32 sig);
>   
>   #endif

Re: [PATCH v4 12/30] arm64: Add support for setting up the PSCI conduit through ACPI

[Andrew Jones]

On Mon, Feb 13, 2023 at 10:17:41AM +0000, Nikos Nikoleris wrote:
> In systems with ACPI support and when a DT is not provided, we can use
> the FADT to discover whether PSCI calls need to use smc or hvc
> calls. This change implements this but retains the default behavior;
> we check if a valid DT is provided, if not, we try to setup the PSCI
> conduit using ACPI.
> 
> Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
> Reviewed-by: Andrew Jones <drjones@redhat.com>
> Reviewed-by: Ricardo Koller <ricarkol@google.com>
> ---
>  arm/Makefile.arm64 |  4 ++++
>  lib/acpi.h         |  5 +++++
>  lib/arm/psci.c     | 37 ++++++++++++++++++++++++++++++++++++-
>  3 files changed, 45 insertions(+), 1 deletion(-)
> 
> diff --git a/arm/Makefile.arm64 b/arm/Makefile.arm64
> index 42e18e77..6dff6cad 100644
> --- a/arm/Makefile.arm64
> +++ b/arm/Makefile.arm64
> @@ -25,6 +25,10 @@ cflatobjs += lib/arm64/processor.o
>  cflatobjs += lib/arm64/spinlock.o
>  cflatobjs += lib/arm64/gic-v3-its.o lib/arm64/gic-v3-its-cmd.o
>  
> +ifeq ($(CONFIG_EFI),y)
> +cflatobjs += lib/acpi.o
> +endif

This is already in arm/Makefile.common due to an earlier patch in
this series. Putting it in Makefile.arm64 makes more sense, though,
so let's just drop the earlier patch.

Thanks,
drew

Re: [PATCH v4 04/30] lib: Apply Lindent to acpi.{c,h}

[Andrew Jones]

On Thu, Mar 09, 2023 at 03:11:56PM +0800, Shaoqin Huang wrote:
...
> > +#define ACPI_TABLE_HEADER_DEF		/* ACPI common table header */			\
> > +	u32 signature;			/* ACPI signature (4 ASCII characters) */	\
> > +	u32 length;			/* Length of table, in bytes, including header */ \
> > +	u8  revision;			/* ACPI Specification minor version # */	\
> > +	u8  checksum;			/* To make sum of entire table == 0 */		\
> > +	u8  oem_id [6];			/* OEM identification */			\
>                   ^
> > +	u8  oem_table_id [8];		/* OEM table identification */			\
>                         ^
> > +	u32 oem_revision;		/* OEM revision number */			\
> > +	u8  asl_compiler_id [4];	/* ASL compiler vendor ID */			\
>                            ^
> nit: These space should also be deleted.
>

Yes, but the fixup belongs in the next patch that does manual style
changes.

Thanks,
drew

Re: [PATCH v4 28/30] arm64: Add support for efi in Makefile

[Andrew Jones]

On Mon, Feb 13, 2023 at 10:17:57AM +0000, Nikos Nikoleris wrote:
...
> --- a/configure
> +++ b/configure
> @@ -208,14 +208,19 @@ else
>      fi
>  fi
>  
> -if [ "$efi" ] && [ "$arch" != "x86_64" ]; then
> +if [ "$efi" ] && [ "$arch" != "x86_64" ] && [ "$arch" != "arm64" ]; then
>      echo "--[enable|disable]-efi is not supported for $arch"
>      usage
>  fi

Need to also update the usage text with

-           --[enable|disable]-efi Boot and run from UEFI (disabled by default, x86_64 only)
+           --[enable|disable]-efi Boot and run from UEFI (disabled by default, x86_64 and arm64 only)

Thanks,
drew

Re: [PATCH v4 30/30] arm64: Add an efi/run script

[Andrew Jones]

On Mon, Feb 13, 2023 at 10:17:59AM +0000, Nikos Nikoleris wrote:
> This change adds a efi/run script inspired by the one in x86. This
> script will setup a folder with the test compiled as an EFI app and a
> startup.nsh script. The script launches QEMU providing an image with
> EDKII and the path to the folder with the test which is executed
> automatically.
> 
> For example:
> 
> $> ./arm/efi/run ./arm/selftest.efi setup smp=2 mem=256

This should be

./arm/efi/run ./arm/selftest.efi -append "setup smp=2 mem=256" -smp 2 -m 256

but I can't get any tests to run through ./arm/efi/run. All of them
immediately die with a DABT_EL1. I can get the tests to run (and pass) by
manually booting into UEFI with the FAT partition pointing at the parent
directory

 $QEMU -nodefaults -machine virt -accel tcg -cpu cortex-a57 \
       -device pci-testdev -display none -serial stdio \
       -bios /usr/share/edk2/aarch64/QEMU_EFI.silent.fd \
       -drive file.dir=efi-tests/,file.driver=vvfat,file.rw=on,format=raw,if=virtio

and then, for example for the timer test, doing

 fs0:
 cd timer
 timer.efi

but the script never works.

Thanks,
drew

Re: [PATCH v4 30/30] arm64: Add an efi/run script

[Nikos Nikoleris]

Hi Drew,

On 21/03/2023 18:41, Andrew Jones wrote:
> On Mon, Feb 13, 2023 at 10:17:59AM +0000, Nikos Nikoleris wrote:
>> This change adds a efi/run script inspired by the one in x86. This
>> script will setup a folder with the test compiled as an EFI app and a
>> startup.nsh script. The script launches QEMU providing an image with
>> EDKII and the path to the folder with the test which is executed
>> automatically.
>>
>> For example:
>>
>> $> ./arm/efi/run ./arm/selftest.efi setup smp=2 mem=256
> 
> This should be
> 
> ./arm/efi/run ./arm/selftest.efi -append "setup smp=2 mem=256" -smp 2 -m 256
>

Indeed, I will update the commit message.

> but I can't get any tests to run through ./arm/efi/run. All of them
> immediately die with a DABT_EL1. I can get the tests to run (and pass) by
> manually booting into UEFI with the FAT partition pointing at the parent
> directory
> 

I suppose the DABT_EL1 is happening after the test has started and not 
while the UEFI interactive shell starts?

>   $QEMU -nodefaults -machine virt -accel tcg -cpu cortex-a57 \
>         -device pci-testdev -display none -serial stdio \
>         -bios /usr/share/edk2/aarch64/QEMU_EFI.silent.fd \
>         -drive file.dir=efi-tests/,file.driver=vvfat,file.rw=on,format=raw,if=virtio
> 

Do you hit the DABT_EL1 if you let it automatically start using the 
startup.nsh prepared by the ./arm/efi/run script? Meaning change the 
above command if you provided -drive file.dir=efi-tests/timer instead:

  $QEMU -nodefaults -machine virt -accel tcg -cpu cortex-a57 \
        -device pci-testdev -display none -serial stdio \
        -bios /usr/share/edk2/aarch64/QEMU_EFI.silent.fd \
        -drive file.dir=efi 
tests/timer,file.driver=vvfat,file.rw=on,format=raw,if=virtio

Thanks for reviewing this!

Nikos

> and then, for example for the timer test, doing
> 
>   fs0:
>   cd timer
>   timer.efi
> 
> but the script never works.
>
> Thanks,
> drew

Re: [PATCH v4 04/30] lib: Apply Lindent to acpi.{c,h}

[Nikos Nikoleris]

On 21/03/2023 17:32, Andrew Jones wrote:
> On Thu, Mar 09, 2023 at 03:11:56PM +0800, Shaoqin Huang wrote:
> ...
>>> +#define ACPI_TABLE_HEADER_DEF		/* ACPI common table header */			\
>>> +	u32 signature;			/* ACPI signature (4 ASCII characters) */	\
>>> +	u32 length;			/* Length of table, in bytes, including header */ \
>>> +	u8  revision;			/* ACPI Specification minor version # */	\
>>> +	u8  checksum;			/* To make sum of entire table == 0 */		\
>>> +	u8  oem_id [6];			/* OEM identification */			\
>>                    ^
>>> +	u8  oem_table_id [8];		/* OEM table identification */			\
>>                          ^
>>> +	u32 oem_revision;		/* OEM revision number */			\
>>> +	u8  asl_compiler_id [4];	/* ASL compiler vendor ID */			\
>>                             ^
>> nit: These space should also be deleted.
>>
> 
> Yes, but the fixup belongs in the next patch that does manual style
> changes.
> 

Thank you both, I will address this in the next patch.

Thanks,

Nikos

Re: [PATCH v4 30/30] arm64: Add an efi/run script

[Andrew Jones]

On Wed, Mar 22, 2023 at 10:02:35AM +0000, Nikos Nikoleris wrote:
> Hi Drew,
> 
> On 21/03/2023 18:41, Andrew Jones wrote:
> > On Mon, Feb 13, 2023 at 10:17:59AM +0000, Nikos Nikoleris wrote:
> > > This change adds a efi/run script inspired by the one in x86. This
> > > script will setup a folder with the test compiled as an EFI app and a
> > > startup.nsh script. The script launches QEMU providing an image with
> > > EDKII and the path to the folder with the test which is executed
> > > automatically.
> > > 
> > > For example:
> > > 
> > > $> ./arm/efi/run ./arm/selftest.efi setup smp=2 mem=256
> > 
> > This should be
> > 
> > ./arm/efi/run ./arm/selftest.efi -append "setup smp=2 mem=256" -smp 2 -m 256
> > 
> 
> Indeed, I will update the commit message.
> 
> > but I can't get any tests to run through ./arm/efi/run. All of them
> > immediately die with a DABT_EL1. I can get the tests to run (and pass) by
> > manually booting into UEFI with the FAT partition pointing at the parent
> > directory
> > 
> 
> I suppose the DABT_EL1 is happening after the test has started and not while
> the UEFI interactive shell starts?

The countdown completes and the startup script runs (I can add an echo to
check it). So it must be the test that fails.

> 
> >   $QEMU -nodefaults -machine virt -accel tcg -cpu cortex-a57 \
> >         -device pci-testdev -display none -serial stdio \
> >         -bios /usr/share/edk2/aarch64/QEMU_EFI.silent.fd \
> >         -drive file.dir=efi-tests/,file.driver=vvfat,file.rw=on,format=raw,if=virtio
> > 
> 
> Do you hit the DABT_EL1 if you let it automatically start using the
> startup.nsh prepared by the ./arm/efi/run script? Meaning change the above
> command if you provided -drive file.dir=efi-tests/timer instead:
> 
>  $QEMU -nodefaults -machine virt -accel tcg -cpu cortex-a57 \
>        -device pci-testdev -display none -serial stdio \
>        -bios /usr/share/edk2/aarch64/QEMU_EFI.silent.fd \
>        -drive file.dir=efi
> tests/timer,file.driver=vvfat,file.rw=on,format=raw,if=virtio

Yes, this is what ./arm/efi/run does, and it doesn't help to use the
command line directly.

> 
> Thanks for reviewing this!
> 
> Nikos
> 
> > and then, for example for the timer test, doing
> > 
> >   fs0:
> >   cd timer
> >   timer.efi

This actually doesn't work. I was actually doing

 fs0:
 cd timer
 ls
 timer.efi

and, believe it or not, without the 'ls' I get the dabt, with the 'ls' the
test runs and passes. Adding an 'ls' to the startup script doesn't help
the automatic execution though.

Which versions of QEMU and edk2 are you using? And what file system do you
have the efi-tests directory on?

Thanks,
drew

> > 
> > but the script never works.
> > 
> > Thanks,
> > drew

Re: [PATCH v4 30/30] arm64: Add an efi/run script

[Nikos Nikoleris]

On 22/03/2023 11:24, Andrew Jones wrote:
> On Wed, Mar 22, 2023 at 10:02:35AM +0000, Nikos Nikoleris wrote:
>> Hi Drew,
>>
>> On 21/03/2023 18:41, Andrew Jones wrote:
>>> On Mon, Feb 13, 2023 at 10:17:59AM +0000, Nikos Nikoleris wrote:
>>>> This change adds a efi/run script inspired by the one in x86. This
>>>> script will setup a folder with the test compiled as an EFI app and a
>>>> startup.nsh script. The script launches QEMU providing an image with
>>>> EDKII and the path to the folder with the test which is executed
>>>> automatically.
>>>>
>>>> For example:
>>>>
>>>> $> ./arm/efi/run ./arm/selftest.efi setup smp=2 mem=256
>>>
>>> This should be
>>>
>>> ./arm/efi/run ./arm/selftest.efi -append "setup smp=2 mem=256" -smp 2 -m 256
>>>
>>
>> Indeed, I will update the commit message.
>>
>>> but I can't get any tests to run through ./arm/efi/run. All of them
>>> immediately die with a DABT_EL1. I can get the tests to run (and pass) by
>>> manually booting into UEFI with the FAT partition pointing at the parent
>>> directory
>>>
>>
>> I suppose the DABT_EL1 is happening after the test has started and not while
>> the UEFI interactive shell starts?
> 
> The countdown completes and the startup script runs (I can add an echo to
> check it). So it must be the test that fails.
> 
>>
>>>    $QEMU -nodefaults -machine virt -accel tcg -cpu cortex-a57 \
>>>          -device pci-testdev -display none -serial stdio \
>>>          -bios /usr/share/edk2/aarch64/QEMU_EFI.silent.fd \
>>>          -drive file.dir=efi-tests/,file.driver=vvfat,file.rw=on,format=raw,if=virtio
>>>
>>
>> Do you hit the DABT_EL1 if you let it automatically start using the
>> startup.nsh prepared by the ./arm/efi/run script? Meaning change the above
>> command if you provided -drive file.dir=efi-tests/timer instead:
>>
>>   $QEMU -nodefaults -machine virt -accel tcg -cpu cortex-a57 \
>>         -device pci-testdev -display none -serial stdio \
>>         -bios /usr/share/edk2/aarch64/QEMU_EFI.silent.fd \
>>         -drive file.dir=efi
>> tests/timer,file.driver=vvfat,file.rw=on,format=raw,if=virtio
> 
> Yes, this is what ./arm/efi/run does, and it doesn't help to use the
> command line directly.
> 
>>
>> Thanks for reviewing this!
>>
>> Nikos
>>
>>> and then, for example for the timer test, doing
>>>
>>>    fs0:
>>>    cd timer
>>>    timer.efi
> 
> This actually doesn't work. I was actually doing
> 
>   fs0:
>   cd timer
>   ls
>   timer.efi
> 
> and, believe it or not, without the 'ls' I get the dabt, with the 'ls' the
> test runs and passes. Adding an 'ls' to the startup script doesn't help
> the automatic execution though.
> 
> Which versions of QEMU and edk2 are you using? And what file system do you
> have the efi-tests directory on?
> 

I am using the QEMU_EFI.fd image that comes with Ubuntu 20.04.6 
(0~20191122.bd85bf54-2ubuntu3.4) 
https://packages.ubuntu.com/focal-updates/qemu-efi-aarch64

and I've tried two different versions of QEMU

$> qemu-system-aarch64 --version 

QEMU emulator version 4.2.1 (Debian 1:4.2-3ubuntu6.24)
Copyright (c) 2003-2019 Fabrice Bellard and the QEMU Project developers

$> ../qemu/build/qemu-system-aarch64 --version
QEMU emulator version 7.0.0 (v7.0.0-dirty)
Copyright (c) 2003-2022 Fabrice Bellard and the QEMU Project developers

efi-tests is on ext4

I am happy to have a closer look if you help me reproduce your environment.

Thanks,

Nikos

Re: [PATCH v4 30/30] arm64: Add an efi/run script

[Andrew Jones]

On Wed, Mar 22, 2023 at 11:57:17AM +0000, Nikos Nikoleris wrote:
> On 22/03/2023 11:24, Andrew Jones wrote:
> > On Wed, Mar 22, 2023 at 10:02:35AM +0000, Nikos Nikoleris wrote:
> > > Hi Drew,
> > > 
> > > On 21/03/2023 18:41, Andrew Jones wrote:
> > > > On Mon, Feb 13, 2023 at 10:17:59AM +0000, Nikos Nikoleris wrote:
> > > > > This change adds a efi/run script inspired by the one in x86. This
> > > > > script will setup a folder with the test compiled as an EFI app and a
> > > > > startup.nsh script. The script launches QEMU providing an image with
> > > > > EDKII and the path to the folder with the test which is executed
> > > > > automatically.
> > > > > 
> > > > > For example:
> > > > > 
> > > > > $> ./arm/efi/run ./arm/selftest.efi setup smp=2 mem=256
> > > > 
> > > > This should be
> > > > 
> > > > ./arm/efi/run ./arm/selftest.efi -append "setup smp=2 mem=256" -smp 2 -m 256
> > > > 
> > > 
> > > Indeed, I will update the commit message.
> > > 
> > > > but I can't get any tests to run through ./arm/efi/run. All of them
> > > > immediately die with a DABT_EL1. I can get the tests to run (and pass) by
> > > > manually booting into UEFI with the FAT partition pointing at the parent
> > > > directory
> > > > 
> > > 
> > > I suppose the DABT_EL1 is happening after the test has started and not while
> > > the UEFI interactive shell starts?
> > 
> > The countdown completes and the startup script runs (I can add an echo to
> > check it). So it must be the test that fails.
> > 
> > > 
> > > >    $QEMU -nodefaults -machine virt -accel tcg -cpu cortex-a57 \
> > > >          -device pci-testdev -display none -serial stdio \
> > > >          -bios /usr/share/edk2/aarch64/QEMU_EFI.silent.fd \
> > > >          -drive file.dir=efi-tests/,file.driver=vvfat,file.rw=on,format=raw,if=virtio
> > > > 
> > > 
> > > Do you hit the DABT_EL1 if you let it automatically start using the
> > > startup.nsh prepared by the ./arm/efi/run script? Meaning change the above
> > > command if you provided -drive file.dir=efi-tests/timer instead:
> > > 
> > >   $QEMU -nodefaults -machine virt -accel tcg -cpu cortex-a57 \
> > >         -device pci-testdev -display none -serial stdio \
> > >         -bios /usr/share/edk2/aarch64/QEMU_EFI.silent.fd \
> > >         -drive file.dir=efi
> > > tests/timer,file.driver=vvfat,file.rw=on,format=raw,if=virtio
> > 
> > Yes, this is what ./arm/efi/run does, and it doesn't help to use the
> > command line directly.
> > 
> > > 
> > > Thanks for reviewing this!
> > > 
> > > Nikos
> > > 
> > > > and then, for example for the timer test, doing
> > > > 
> > > >    fs0:
> > > >    cd timer
> > > >    timer.efi
> > 
> > This actually doesn't work. I was actually doing
> > 
> >   fs0:
> >   cd timer
> >   ls
> >   timer.efi
> > 
> > and, believe it or not, without the 'ls' I get the dabt, with the 'ls' the
> > test runs and passes. Adding an 'ls' to the startup script doesn't help
> > the automatic execution though.
> > 
> > Which versions of QEMU and edk2 are you using? And what file system do you
> > have the efi-tests directory on?
> > 
> 
> I am using the QEMU_EFI.fd image that comes with Ubuntu 20.04.6
> (0~20191122.bd85bf54-2ubuntu3.4)
> https://packages.ubuntu.com/focal-updates/qemu-efi-aarch64
> 
> and I've tried two different versions of QEMU
> 
> $> qemu-system-aarch64 --version
> 
> QEMU emulator version 4.2.1 (Debian 1:4.2-3ubuntu6.24)
> Copyright (c) 2003-2019 Fabrice Bellard and the QEMU Project developers
> 
> $> ../qemu/build/qemu-system-aarch64 --version
> QEMU emulator version 7.0.0 (v7.0.0-dirty)
> Copyright (c) 2003-2022 Fabrice Bellard and the QEMU Project developers
> 
> efi-tests is on ext4
> 
> I am happy to have a closer look if you help me reproduce your environment.

I'm on Fedora 36 and the file system used for this is XFS. My QEMU version
was something pretty recent, but I didn't remember what, so I just updated
to latest master (which happens to be the same as v8.0.0-rc1 right now).
My edk2 is the one packaged with F36,
edk2-aarch64-20221117gitfff6d81270b5-14.fc36.noarch

The QEMU update to v8.0.0-rc1 didn't change anything for me (still same
failure and still same "fix" of running the test manually after doing
a manual 'ls').

Thanks,
drew

Re: [PATCH v4 30/30] arm64: Add an efi/run script

[Nikos Nikoleris]

On 22/03/2023 12:32, Andrew Jones wrote:
> On Wed, Mar 22, 2023 at 11:57:17AM +0000, Nikos Nikoleris wrote:
>> On 22/03/2023 11:24, Andrew Jones wrote:
>>> On Wed, Mar 22, 2023 at 10:02:35AM +0000, Nikos Nikoleris wrote:
>>>> Hi Drew,
>>>>
>>>> On 21/03/2023 18:41, Andrew Jones wrote:
>>>>> On Mon, Feb 13, 2023 at 10:17:59AM +0000, Nikos Nikoleris wrote:
>>>>>> This change adds a efi/run script inspired by the one in x86. This
>>>>>> script will setup a folder with the test compiled as an EFI app and a
>>>>>> startup.nsh script. The script launches QEMU providing an image with
>>>>>> EDKII and the path to the folder with the test which is executed
>>>>>> automatically.
>>>>>>
>>>>>> For example:
>>>>>>
>>>>>> $> ./arm/efi/run ./arm/selftest.efi setup smp=2 mem=256
>>>>>
>>>>> This should be
>>>>>
>>>>> ./arm/efi/run ./arm/selftest.efi -append "setup smp=2 mem=256" -smp 2 -m 256
>>>>>
>>>>
>>>> Indeed, I will update the commit message.
>>>>
>>>>> but I can't get any tests to run through ./arm/efi/run. All of them
>>>>> immediately die with a DABT_EL1. I can get the tests to run (and pass) by
>>>>> manually booting into UEFI with the FAT partition pointing at the parent
>>>>> directory
>>>>>
>>>>
>>>> I suppose the DABT_EL1 is happening after the test has started and not while
>>>> the UEFI interactive shell starts?
>>>
>>> The countdown completes and the startup script runs (I can add an echo to
>>> check it). So it must be the test that fails.
>>>
>>>>
>>>>>     $QEMU -nodefaults -machine virt -accel tcg -cpu cortex-a57 \
>>>>>           -device pci-testdev -display none -serial stdio \
>>>>>           -bios /usr/share/edk2/aarch64/QEMU_EFI.silent.fd \
>>>>>           -drive file.dir=efi-tests/,file.driver=vvfat,file.rw=on,format=raw,if=virtio
>>>>>
>>>>
>>>> Do you hit the DABT_EL1 if you let it automatically start using the
>>>> startup.nsh prepared by the ./arm/efi/run script? Meaning change the above
>>>> command if you provided -drive file.dir=efi-tests/timer instead:
>>>>
>>>>    $QEMU -nodefaults -machine virt -accel tcg -cpu cortex-a57 \
>>>>          -device pci-testdev -display none -serial stdio \
>>>>          -bios /usr/share/edk2/aarch64/QEMU_EFI.silent.fd \
>>>>          -drive file.dir=efi
>>>> tests/timer,file.driver=vvfat,file.rw=on,format=raw,if=virtio
>>>
>>> Yes, this is what ./arm/efi/run does, and it doesn't help to use the
>>> command line directly.
>>>
>>>>
>>>> Thanks for reviewing this!
>>>>
>>>> Nikos
>>>>
>>>>> and then, for example for the timer test, doing
>>>>>
>>>>>     fs0:
>>>>>     cd timer
>>>>>     timer.efi
>>>
>>> This actually doesn't work. I was actually doing
>>>
>>>    fs0:
>>>    cd timer
>>>    ls
>>>    timer.efi
>>>
>>> and, believe it or not, without the 'ls' I get the dabt, with the 'ls' the
>>> test runs and passes. Adding an 'ls' to the startup script doesn't help
>>> the automatic execution though.
>>>
>>> Which versions of QEMU and edk2 are you using? And what file system do you
>>> have the efi-tests directory on?
>>>
>>
>> I am using the QEMU_EFI.fd image that comes with Ubuntu 20.04.6
>> (0~20191122.bd85bf54-2ubuntu3.4)
>> https://packages.ubuntu.com/focal-updates/qemu-efi-aarch64
>>
>> and I've tried two different versions of QEMU
>>
>> $> qemu-system-aarch64 --version
>>
>> QEMU emulator version 4.2.1 (Debian 1:4.2-3ubuntu6.24)
>> Copyright (c) 2003-2019 Fabrice Bellard and the QEMU Project developers
>>
>> $> ../qemu/build/qemu-system-aarch64 --version
>> QEMU emulator version 7.0.0 (v7.0.0-dirty)
>> Copyright (c) 2003-2022 Fabrice Bellard and the QEMU Project developers
>>
>> efi-tests is on ext4
>>
>> I am happy to have a closer look if you help me reproduce your environment.
> 
> I'm on Fedora 36 and the file system used for this is XFS. My QEMU version
> was something pretty recent, but I didn't remember what, so I just updated
> to latest master (which happens to be the same as v8.0.0-rc1 right now).
> My edk2 is the one packaged with F36,
> edk2-aarch64-20221117gitfff6d81270b5-14.fc36.noarch
> 
> The QEMU update to v8.0.0-rc1 didn't change anything for me (still same
> failure and still same "fix" of running the test manually after doing
> a manual 'ls').
> 

Thanks Drew!

I managed to hit the DABT_EL1 when I switched to the F36 edk2. The 
problem seems to be with the initialization of the page allocation 
mechanism. mmu_idmap is allocated at 0x80000000 and phys_alloc_show() prints

phys_alloc minimum alignment: 0x40
0000000048000000-000000007c590fff [USED]
000000007c591000-000000007c590fff [FREE]

Am I wrong to expect that the address that page_alloc() returns for 
mmu_idmap should be within the [USED] range?

I'll have a closer look into this but I just wanted to check as I am not 
sure I fully understand the code/logic of lib/alloc_page.{c,h}

Thanks,

Nikos

Re: [PATCH v4 30/30] arm64: Add an efi/run script

[Andrew Jones]

On Wed, Mar 22, 2023 at 07:09:23PM +0000, Nikos Nikoleris wrote:
> On 22/03/2023 12:32, Andrew Jones wrote:
> > On Wed, Mar 22, 2023 at 11:57:17AM +0000, Nikos Nikoleris wrote:
> > > On 22/03/2023 11:24, Andrew Jones wrote:
> > > > On Wed, Mar 22, 2023 at 10:02:35AM +0000, Nikos Nikoleris wrote:
> > > > > Hi Drew,
> > > > > 
> > > > > On 21/03/2023 18:41, Andrew Jones wrote:
> > > > > > On Mon, Feb 13, 2023 at 10:17:59AM +0000, Nikos Nikoleris wrote:
> > > > > > > This change adds a efi/run script inspired by the one in x86. This
> > > > > > > script will setup a folder with the test compiled as an EFI app and a
> > > > > > > startup.nsh script. The script launches QEMU providing an image with
> > > > > > > EDKII and the path to the folder with the test which is executed
> > > > > > > automatically.
> > > > > > > 
> > > > > > > For example:
> > > > > > > 
> > > > > > > $> ./arm/efi/run ./arm/selftest.efi setup smp=2 mem=256
> > > > > > 
> > > > > > This should be
> > > > > > 
> > > > > > ./arm/efi/run ./arm/selftest.efi -append "setup smp=2 mem=256" -smp 2 -m 256
> > > > > > 
> > > > > 
> > > > > Indeed, I will update the commit message.
> > > > > 
> > > > > > but I can't get any tests to run through ./arm/efi/run. All of them
> > > > > > immediately die with a DABT_EL1. I can get the tests to run (and pass) by
> > > > > > manually booting into UEFI with the FAT partition pointing at the parent
> > > > > > directory
> > > > > > 
> > > > > 
> > > > > I suppose the DABT_EL1 is happening after the test has started and not while
> > > > > the UEFI interactive shell starts?
> > > > 
> > > > The countdown completes and the startup script runs (I can add an echo to
> > > > check it). So it must be the test that fails.
> > > > 
> > > > > 
> > > > > >     $QEMU -nodefaults -machine virt -accel tcg -cpu cortex-a57 \
> > > > > >           -device pci-testdev -display none -serial stdio \
> > > > > >           -bios /usr/share/edk2/aarch64/QEMU_EFI.silent.fd \
> > > > > >           -drive file.dir=efi-tests/,file.driver=vvfat,file.rw=on,format=raw,if=virtio
> > > > > > 
> > > > > 
> > > > > Do you hit the DABT_EL1 if you let it automatically start using the
> > > > > startup.nsh prepared by the ./arm/efi/run script? Meaning change the above
> > > > > command if you provided -drive file.dir=efi-tests/timer instead:
> > > > > 
> > > > >    $QEMU -nodefaults -machine virt -accel tcg -cpu cortex-a57 \
> > > > >          -device pci-testdev -display none -serial stdio \
> > > > >          -bios /usr/share/edk2/aarch64/QEMU_EFI.silent.fd \
> > > > >          -drive file.dir=efi
> > > > > tests/timer,file.driver=vvfat,file.rw=on,format=raw,if=virtio
> > > > 
> > > > Yes, this is what ./arm/efi/run does, and it doesn't help to use the
> > > > command line directly.
> > > > 
> > > > > 
> > > > > Thanks for reviewing this!
> > > > > 
> > > > > Nikos
> > > > > 
> > > > > > and then, for example for the timer test, doing
> > > > > > 
> > > > > >     fs0:
> > > > > >     cd timer
> > > > > >     timer.efi
> > > > 
> > > > This actually doesn't work. I was actually doing
> > > > 
> > > >    fs0:
> > > >    cd timer
> > > >    ls
> > > >    timer.efi
> > > > 
> > > > and, believe it or not, without the 'ls' I get the dabt, with the 'ls' the
> > > > test runs and passes. Adding an 'ls' to the startup script doesn't help
> > > > the automatic execution though.
> > > > 
> > > > Which versions of QEMU and edk2 are you using? And what file system do you
> > > > have the efi-tests directory on?
> > > > 
> > > 
> > > I am using the QEMU_EFI.fd image that comes with Ubuntu 20.04.6
> > > (0~20191122.bd85bf54-2ubuntu3.4)
> > > https://packages.ubuntu.com/focal-updates/qemu-efi-aarch64
> > > 
> > > and I've tried two different versions of QEMU
> > > 
> > > $> qemu-system-aarch64 --version
> > > 
> > > QEMU emulator version 4.2.1 (Debian 1:4.2-3ubuntu6.24)
> > > Copyright (c) 2003-2019 Fabrice Bellard and the QEMU Project developers
> > > 
> > > $> ../qemu/build/qemu-system-aarch64 --version
> > > QEMU emulator version 7.0.0 (v7.0.0-dirty)
> > > Copyright (c) 2003-2022 Fabrice Bellard and the QEMU Project developers
> > > 
> > > efi-tests is on ext4
> > > 
> > > I am happy to have a closer look if you help me reproduce your environment.
> > 
> > I'm on Fedora 36 and the file system used for this is XFS. My QEMU version
> > was something pretty recent, but I didn't remember what, so I just updated
> > to latest master (which happens to be the same as v8.0.0-rc1 right now).
> > My edk2 is the one packaged with F36,
> > edk2-aarch64-20221117gitfff6d81270b5-14.fc36.noarch
> > 
> > The QEMU update to v8.0.0-rc1 didn't change anything for me (still same
> > failure and still same "fix" of running the test manually after doing
> > a manual 'ls').
> > 
> 
> Thanks Drew!
> 
> I managed to hit the DABT_EL1 when I switched to the F36 edk2. The problem
> seems to be with the initialization of the page allocation mechanism.
> mmu_idmap is allocated at 0x80000000 and phys_alloc_show() prints
> 
> phys_alloc minimum alignment: 0x40
> 0000000048000000-000000007c590fff [USED]
> 000000007c591000-000000007c590fff [FREE]
> 
> Am I wrong to expect that the address that page_alloc() returns for
> mmu_idmap should be within the [USED] range?

You're not wrong. That's what I would expect as well from the sequence
of allocator initializations and allocations, and that's indeed how it
works for the non-uefi case.

> 
> I'll have a closer look into this but I just wanted to check as I am not
> sure I fully understand the code/logic of lib/alloc_page.{c,h}

I'll also try to find time to take a look. TBH, I never completely grokked
the new allocator either. If it gets in our way, then we could add a
simpler allocator implementation (just a linked list like it was before)
as an alternative and then use that instead.

Thanks,
drew

Re: [PATCH v4 30/30] arm64: Add an efi/run script

[Alexandru Elisei]

Hi,

Just a thought, but if the page allocator is giving you trouble, you could
try using the physical allocator instead, like I tried to do in my cache
maintenance series [1] (requires this patch [2] to change how the idmap is
allocated, to use pgd_alloc() instead of page_alloc()).

That patch also moves the point where the MMU is enabled to earlier in the
boot process, which might be too invasive for this series. In theory, it
shouldn't cause any problems, but that's just in theory.

[1] https://gitlab.arm.com/linux-arm/kvm-unit-tests-ae/-/commit/3385347b8157d72ea1f2b83afe0026305d89a9ea
[2] https://gitlab.arm.com/linux-arm/kvm-unit-tests-ae/-/commit/6faa52530b6a1c150dca4bb7e7caae6c70b162ee

Thanks,
Alex

On Thu, Mar 23, 2023 at 06:52:56PM +0100, Andrew Jones wrote:
> On Wed, Mar 22, 2023 at 07:09:23PM +0000, Nikos Nikoleris wrote:
> > On 22/03/2023 12:32, Andrew Jones wrote:
> > > On Wed, Mar 22, 2023 at 11:57:17AM +0000, Nikos Nikoleris wrote:
> > > > On 22/03/2023 11:24, Andrew Jones wrote:
> > > > > On Wed, Mar 22, 2023 at 10:02:35AM +0000, Nikos Nikoleris wrote:
> > > > > > Hi Drew,
> > > > > > 
> > > > > > On 21/03/2023 18:41, Andrew Jones wrote:
> > > > > > > On Mon, Feb 13, 2023 at 10:17:59AM +0000, Nikos Nikoleris wrote:
> > > > > > > > This change adds a efi/run script inspired by the one in x86. This
> > > > > > > > script will setup a folder with the test compiled as an EFI app and a
> > > > > > > > startup.nsh script. The script launches QEMU providing an image with
> > > > > > > > EDKII and the path to the folder with the test which is executed
> > > > > > > > automatically.
> > > > > > > > 
> > > > > > > > For example:
> > > > > > > > 
> > > > > > > > $> ./arm/efi/run ./arm/selftest.efi setup smp=2 mem=256
> > > > > > > 
> > > > > > > This should be
> > > > > > > 
> > > > > > > ./arm/efi/run ./arm/selftest.efi -append "setup smp=2 mem=256" -smp 2 -m 256
> > > > > > > 
> > > > > > 
> > > > > > Indeed, I will update the commit message.
> > > > > > 
> > > > > > > but I can't get any tests to run through ./arm/efi/run. All of them
> > > > > > > immediately die with a DABT_EL1. I can get the tests to run (and pass) by
> > > > > > > manually booting into UEFI with the FAT partition pointing at the parent
> > > > > > > directory
> > > > > > > 
> > > > > > 
> > > > > > I suppose the DABT_EL1 is happening after the test has started and not while
> > > > > > the UEFI interactive shell starts?
> > > > > 
> > > > > The countdown completes and the startup script runs (I can add an echo to
> > > > > check it). So it must be the test that fails.
> > > > > 
> > > > > > 
> > > > > > >     $QEMU -nodefaults -machine virt -accel tcg -cpu cortex-a57 \
> > > > > > >           -device pci-testdev -display none -serial stdio \
> > > > > > >           -bios /usr/share/edk2/aarch64/QEMU_EFI.silent.fd \
> > > > > > >           -drive file.dir=efi-tests/,file.driver=vvfat,file.rw=on,format=raw,if=virtio
> > > > > > > 
> > > > > > 
> > > > > > Do you hit the DABT_EL1 if you let it automatically start using the
> > > > > > startup.nsh prepared by the ./arm/efi/run script? Meaning change the above
> > > > > > command if you provided -drive file.dir=efi-tests/timer instead:
> > > > > > 
> > > > > >    $QEMU -nodefaults -machine virt -accel tcg -cpu cortex-a57 \
> > > > > >          -device pci-testdev -display none -serial stdio \
> > > > > >          -bios /usr/share/edk2/aarch64/QEMU_EFI.silent.fd \
> > > > > >          -drive file.dir=efi
> > > > > > tests/timer,file.driver=vvfat,file.rw=on,format=raw,if=virtio
> > > > > 
> > > > > Yes, this is what ./arm/efi/run does, and it doesn't help to use the
> > > > > command line directly.
> > > > > 
> > > > > > 
> > > > > > Thanks for reviewing this!
> > > > > > 
> > > > > > Nikos
> > > > > > 
> > > > > > > and then, for example for the timer test, doing
> > > > > > > 
> > > > > > >     fs0:
> > > > > > >     cd timer
> > > > > > >     timer.efi
> > > > > 
> > > > > This actually doesn't work. I was actually doing
> > > > > 
> > > > >    fs0:
> > > > >    cd timer
> > > > >    ls
> > > > >    timer.efi
> > > > > 
> > > > > and, believe it or not, without the 'ls' I get the dabt, with the 'ls' the
> > > > > test runs and passes. Adding an 'ls' to the startup script doesn't help
> > > > > the automatic execution though.
> > > > > 
> > > > > Which versions of QEMU and edk2 are you using? And what file system do you
> > > > > have the efi-tests directory on?
> > > > > 
> > > > 
> > > > I am using the QEMU_EFI.fd image that comes with Ubuntu 20.04.6
> > > > (0~20191122.bd85bf54-2ubuntu3.4)
> > > > https://packages.ubuntu.com/focal-updates/qemu-efi-aarch64
> > > > 
> > > > and I've tried two different versions of QEMU
> > > > 
> > > > $> qemu-system-aarch64 --version
> > > > 
> > > > QEMU emulator version 4.2.1 (Debian 1:4.2-3ubuntu6.24)
> > > > Copyright (c) 2003-2019 Fabrice Bellard and the QEMU Project developers
> > > > 
> > > > $> ../qemu/build/qemu-system-aarch64 --version
> > > > QEMU emulator version 7.0.0 (v7.0.0-dirty)
> > > > Copyright (c) 2003-2022 Fabrice Bellard and the QEMU Project developers
> > > > 
> > > > efi-tests is on ext4
> > > > 
> > > > I am happy to have a closer look if you help me reproduce your environment.
> > > 
> > > I'm on Fedora 36 and the file system used for this is XFS. My QEMU version
> > > was something pretty recent, but I didn't remember what, so I just updated
> > > to latest master (which happens to be the same as v8.0.0-rc1 right now).
> > > My edk2 is the one packaged with F36,
> > > edk2-aarch64-20221117gitfff6d81270b5-14.fc36.noarch
> > > 
> > > The QEMU update to v8.0.0-rc1 didn't change anything for me (still same
> > > failure and still same "fix" of running the test manually after doing
> > > a manual 'ls').
> > > 
> > 
> > Thanks Drew!
> > 
> > I managed to hit the DABT_EL1 when I switched to the F36 edk2. The problem
> > seems to be with the initialization of the page allocation mechanism.
> > mmu_idmap is allocated at 0x80000000 and phys_alloc_show() prints
> > 
> > phys_alloc minimum alignment: 0x40
> > 0000000048000000-000000007c590fff [USED]
> > 000000007c591000-000000007c590fff [FREE]
> > 
> > Am I wrong to expect that the address that page_alloc() returns for
> > mmu_idmap should be within the [USED] range?
> 
> You're not wrong. That's what I would expect as well from the sequence
> of allocator initializations and allocations, and that's indeed how it
> works for the non-uefi case.
> 
> > 
> > I'll have a closer look into this but I just wanted to check as I am not
> > sure I fully understand the code/logic of lib/alloc_page.{c,h}
> 
> I'll also try to find time to take a look. TBH, I never completely grokked
> the new allocator either. If it gets in our way, then we could add a
> simpler allocator implementation (just a linked list like it was before)
> as an alternative and then use that instead.
> 
> Thanks,
> drew
> 

Re: [PATCH v4 16/30] arm64: Add support for gic initialization through ACPI

[Shaoqin Huang]

Hi Nikos,

On 2/13/23 18:17, Nikos Nikoleris wrote:
> In systems with ACPI support and when a DT is not provided, we can use
> the MADTs to figure out if it implements a GICv2 or a GICv3 and
> discover the GIC parameters. This change implements this but retains
> the default behavior; we check if a valid DT is provided, if not, we
> try to discover the cores in the system using ACPI.
> 
> Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
> ---
>   lib/acpi.c     |   9 +++-
>   lib/acpi.h     |  46 +++++++++++++++-
>   lib/arm/gic.c  | 139 ++++++++++++++++++++++++++++++++++++++++++++++++-
>   lib/libcflat.h |   1 +
>   4 files changed, 191 insertions(+), 4 deletions(-)
> 
> diff --git a/lib/acpi.c b/lib/acpi.c
> index bbe33d08..760cd8b2 100644
> --- a/lib/acpi.c
> +++ b/lib/acpi.c
> @@ -103,11 +103,12 @@ void *find_acpi_table_addr(u32 sig)
>   	return NULL;
>   }
>   
> -void acpi_table_parse_madt(enum acpi_madt_type mtype, acpi_table_handler handler)
> +int acpi_table_parse_madt(enum acpi_madt_type mtype, acpi_table_handler handler)
>   {
>   	struct acpi_table_madt *madt;
>   	struct acpi_subtable_header *header;
>   	void *end;
> +	int count = 0;
>   
>   	madt = find_acpi_table_addr(MADT_SIGNATURE);
>   	assert(madt);
> @@ -116,9 +117,13 @@ void acpi_table_parse_madt(enum acpi_madt_type mtype, acpi_table_handler handler
>   	end = (void *)((ulong) madt + madt->length);
>   
>   	while ((void *)header < end) {
> -		if (header->type == mtype)
> +		if (header->type == mtype) {
>   			handler(header);
> +			count++;
> +		}
>   
>   		header = (void *)(ulong) header + header->length;
>   	}
> +
> +	return count;
>   }
> diff --git a/lib/acpi.h b/lib/acpi.h
> index af02fd83..a2c0f982 100644
> --- a/lib/acpi.h
> +++ b/lib/acpi.h
> @@ -184,6 +184,50 @@ struct acpi_madt_generic_interrupt {
>   	u16 spe_interrupt;	/* ACPI 6.3 */
>   };
>   
> +#define ACPI_MADT_ENABLED           (1)	/* 00: Processor is usable if set */
This macro has been defined in previous patch, so at here we don't need 
to add it again.

Thanks,
Shaoqin

> +
> +/* 12: Generic Distributor (ACPI 5.0 + ACPI 6.0 changes) */
> +
> +struct acpi_madt_generic_distributor {
> +	struct acpi_subtable_header header;
> +	u16 reserved;		/* reserved - must be zero */
> +	u32 gic_id;
> +	u64 base_address;
> +	u32 global_irq_base;
> +	u8 version;
> +	u8 reserved2[3];	/* reserved - must be zero */
> +};
> +
> +/* Values for Version field above */
> +
> +enum acpi_madt_gic_version {
> +	ACPI_MADT_GIC_VERSION_NONE = 0,
> +	ACPI_MADT_GIC_VERSION_V1 = 1,
> +	ACPI_MADT_GIC_VERSION_V2 = 2,
> +	ACPI_MADT_GIC_VERSION_V3 = 3,
> +	ACPI_MADT_GIC_VERSION_V4 = 4,
> +	ACPI_MADT_GIC_VERSION_RESERVED = 5	/* 5 and greater are reserved */
> +};
> +
> +/* 14: Generic Redistributor (ACPI 5.1) */
> +
> +struct acpi_madt_generic_redistributor {
> +	struct acpi_subtable_header header;
> +	u16 reserved;		/* reserved - must be zero */
> +	u64 base_address;
> +	u32 length;
> +};
> +
> +/* 15: Generic Translator (ACPI 6.0) */
> +
> +struct acpi_madt_generic_translator {
> +	struct acpi_subtable_header header;
> +	u16 reserved;		/* reserved - must be zero */
> +	u32 translation_id;
> +	u64 base_address;
> +	u32 reserved2;
> +};
> +
>   /* Values for MADT subtable type in struct acpi_subtable_header */
>   
>   enum acpi_madt_type {
> @@ -254,6 +298,6 @@ struct acpi_table_gtdt {
>   
>   void set_efi_rsdp(struct acpi_table_rsdp *rsdp);
>   void *find_acpi_table_addr(u32 sig);
> -void acpi_table_parse_madt(enum acpi_madt_type mtype, acpi_table_handler handler);
> +int acpi_table_parse_madt(enum acpi_madt_type mtype, acpi_table_handler handler);
>   
>   #endif
> diff --git a/lib/arm/gic.c b/lib/arm/gic.c
> index 1bfcfcfb..8fc5596f 100644
> --- a/lib/arm/gic.c
> +++ b/lib/arm/gic.c
> @@ -3,6 +3,7 @@
>    *
>    * This work is licensed under the terms of the GNU LGPL, version 2.
>    */
> +#include <acpi.h>
>   #include <devicetree.h>
>   #include <asm/gic.h>
>   #include <asm/io.h>
> @@ -120,7 +121,7 @@ int gic_version(void)
>   	return 0;
>   }
>   
> -int gic_init(void)
> +static int gic_init_fdt(void)
>   {
>   	if (gicv2_init()) {
>   		gic_common_ops = &gicv2_common_ops;
> @@ -133,6 +134,142 @@ int gic_init(void)
>   	return gic_version();
>   }
>   
> +#ifdef CONFIG_EFI
> +
> +#define ACPI_GICV2_DIST_MEM_SIZE	(SZ_4K)
> +#define ACPI_GIC_CPU_IF_MEM_SIZE	(SZ_8K)
> +#define ACPI_GICV3_DIST_MEM_SIZE	(SZ_64K)
> +#define ACPI_GICV3_ITS_MEM_SIZE		(SZ_128K)
> +
> +static int gic_acpi_version(struct acpi_subtable_header *header)
> +{
> +	struct acpi_madt_generic_distributor *dist = (void *)header;
> +	int version = dist->version;
> +
> +	if (version == 2)
> +		gic_common_ops = &gicv2_common_ops;
> +	else if (version == 3)
> +		gic_common_ops = &gicv3_common_ops;
> +
> +	return version;
> +}
> +
> +static int gicv2_acpi_parse_madt_cpu(struct acpi_subtable_header *header)
> +{
> +	struct acpi_madt_generic_interrupt *gicc = (void *)header;
> +	static phys_addr_t gicc_base_address;
> +
> +	if (!(gicc->flags & ACPI_MADT_ENABLED))
> +		return 0;
> +
> +	if (!gicc_base_address) {
> +		gicc_base_address = gicc->base_address;
> +		gicv2_data.cpu_base = ioremap(gicc_base_address, ACPI_GIC_CPU_IF_MEM_SIZE);
> +	}
> +	assert(gicc_base_address == gicc->base_address);
> +
> +	return 0;
> +}
> +
> +static int gicv2_acpi_parse_madt_dist(struct acpi_subtable_header *header)
> +{
> +	struct acpi_madt_generic_distributor *dist = (void *)header;
> +
> +	gicv2_data.dist_base = ioremap(dist->base_address, ACPI_GICV2_DIST_MEM_SIZE);
> +
> +	return 0;
> +}
> +
> +static int gicv3_acpi_parse_madt_gicc(struct acpi_subtable_header *header)
> +{
> +	struct acpi_madt_generic_interrupt *gicc = (void *)header;
> +	static phys_addr_t gicr_base_address;
> +
> +	if (!(gicc->flags & ACPI_MADT_ENABLED))
> +		return 0;
> +
> +	if (!gicr_base_address) {
> +		gicr_base_address = gicc->gicr_base_address;
> +		gicv3_data.redist_bases[0] = ioremap(gicr_base_address, SZ_64K * 2);
> +	}
> +	assert(gicr_base_address == gicc->gicr_base_address);
> +
> +	return 0;
> +}
> +
> +static int gicv3_acpi_parse_madt_dist(struct acpi_subtable_header *header)
> +{
> +	struct acpi_madt_generic_distributor *dist = (void *)header;
> +
> +	gicv3_data.dist_base = ioremap(dist->base_address, ACPI_GICV3_DIST_MEM_SIZE);
> +
> +	return 0;
> +}
> +
> +static int gicv3_acpi_parse_madt_redist(struct acpi_subtable_header *header)
> +{
> +	static int i;
> +	struct acpi_madt_generic_redistributor *redist = (void *)header;
> +
> +	gicv3_data.redist_bases[i++] = ioremap(redist->base_address, redist->length);
> +
> +	return 0;
> +}
> +
> +static int gicv3_acpi_parse_madt_its(struct acpi_subtable_header *header)
> +{
> +	struct acpi_madt_generic_translator *its_entry = (void *)header;
> +
> +	its_data.base = ioremap(its_entry->base_address, ACPI_GICV3_ITS_MEM_SIZE - 1);
> +
> +	return 0;
> +}
> +
> +static int gic_init_acpi(void)
> +{
> +	int count;
> +
> +	acpi_table_parse_madt(ACPI_MADT_TYPE_GENERIC_DISTRIBUTOR, gic_acpi_version);
> +	if (gic_version() == 2) {
> +		acpi_table_parse_madt(ACPI_MADT_TYPE_GENERIC_INTERRUPT,
> +				     gicv2_acpi_parse_madt_cpu);
> +		acpi_table_parse_madt(ACPI_MADT_TYPE_GENERIC_DISTRIBUTOR,
> +				      gicv2_acpi_parse_madt_dist);
> +	} else if (gic_version() == 3) {
> +		acpi_table_parse_madt(ACPI_MADT_TYPE_GENERIC_DISTRIBUTOR,
> +				      gicv3_acpi_parse_madt_dist);
> +		count = acpi_table_parse_madt(ACPI_MADT_TYPE_GENERIC_REDISTRIBUTOR,
> +					      gicv3_acpi_parse_madt_redist);
> +		if (!count)
> +			acpi_table_parse_madt(ACPI_MADT_TYPE_GENERIC_INTERRUPT,
> +					      gicv3_acpi_parse_madt_gicc);
> +		acpi_table_parse_madt(ACPI_MADT_TYPE_GENERIC_TRANSLATOR,
> +				      gicv3_acpi_parse_madt_its);
> +#ifdef __aarch64__
> +		its_init();
> +#endif
> +	}
> +
> +	return gic_version();
> +}
> +
> +#else
> +
> +static int gic_init_acpi(void)
> +{
> +	assert_msg(false, "ACPI not available");
> +}
> +
> +#endif /* CONFIG_EFI */
> +
> +int gic_init(void)
> +{
> +	if (dt_available())
> +		return gic_init_fdt();
> +	else
> +		return gic_init_acpi();
> +}
> +
>   void gic_enable_defaults(void)
>   {
>   	if (!gic_common_ops) {
> diff --git a/lib/libcflat.h b/lib/libcflat.h
> index c1fd31ff..700f4352 100644
> --- a/lib/libcflat.h
> +++ b/lib/libcflat.h
> @@ -161,6 +161,7 @@ extern void setup_vm(void);
>   #define SZ_8K			(1 << 13)
>   #define SZ_16K			(1 << 14)
>   #define SZ_64K			(1 << 16)
> +#define SZ_128K			(1 << 17)
>   #define SZ_1M			(1 << 20)
>   #define SZ_2M			(1 << 21)
>   #define SZ_1G			(1 << 30)

-- 
Shaoqin

Re: [PATCH v4 25/30] arm64: Change GNU-EFI imported code to use defined types

[Shaoqin Huang]

Hi Nikos,

On 2/13/23 18:17, Nikos Nikoleris wrote:
> Convert some types to avoid dependency on gnu-efi's <efi.h> and
> <efilib.h>.
> 
> Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
> ---
>   arm/efi/reloc_aarch64.c | 9 +++------
>   1 file changed, 3 insertions(+), 6 deletions(-)
> 
> diff --git a/arm/efi/reloc_aarch64.c b/arm/efi/reloc_aarch64.c
> index 08672796..fa0cd6bc 100644
> --- a/arm/efi/reloc_aarch64.c
> +++ b/arm/efi/reloc_aarch64.c
> @@ -34,14 +34,11 @@
>       SUCH DAMAGE.
>   */
>   
> -#include <efi.h>
> -#include <efilib.h>
> - > +#include "efi.h"

In [PATCH v4 27/30] lib: Avoid external dependency in libelf. The 
Include header file changed again.

So why not move this change early?

diff --git a/arm/efi/reloc_aarch64.c b/arm/efi/reloc_aarch64.c
index fa0cd6bc..3f6d9a6d 100644
--- a/arm/efi/reloc_aarch64.c
+++ b/arm/efi/reloc_aarch64.c
@@ -34,8 +34,7 @@
      SUCH DAMAGE.
  */

-#include "efi.h"
-#include <elf.h>
+#include <efi.h>

Thanks,
Shaoqin
>   #include <elf.h>
>   
> -EFI_STATUS _relocate (long ldbase, Elf64_Dyn *dyn,
> -		      EFI_HANDLE image EFI_UNUSED,
> -		      EFI_SYSTEM_TABLE *systab EFI_UNUSED)
> +efi_status_t _relocate(long ldbase, Elf64_Dyn *dyn, efi_handle_t image,
> +		       efi_system_table_t *sys_tab)
>   {
>   	long relsz = 0, relent = 0;
>   	Elf64_Rela *rel = 0;

-- 
Shaoqin

Re: [PATCH v4 23/30] arm64: Add a setup sequence for systems that boot through EFI

[Shaoqin Huang]

Hi Nikos,

For that DABT_EL1 error, I have some clues about how it happens. It's 
mainly because this patch includes a memory overflow. I will explain in 
the code body.

On 2/13/23 18:17, Nikos Nikoleris wrote:
> This change implements an alternative setup sequence for the system
> when we are booting through EFI. The memory map is discovered through
> EFI boot services and devices through ACPI.
> 
> This change is based on a change initially proposed by
> Andrew Jones <drjones@redhat.com>
> 
> Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
> ---
>   arm/cstart.S        |   1 +
>   arm/cstart64.S      |   1 +
>   lib/arm/asm/setup.h |   8 ++
>   lib/arm/setup.c     | 181 +++++++++++++++++++++++++++++++++++++++++++-
>   lib/linux/efi.h     |   1 +
>   5 files changed, 190 insertions(+), 2 deletions(-)
> 
> diff --git a/arm/cstart.S b/arm/cstart.S
> index 7036e67f..3dd71ed9 100644
> --- a/arm/cstart.S
> +++ b/arm/cstart.S
> @@ -242,6 +242,7 @@ asm_mmu_disable:
>    *
>    * Input r0 is the stack top, which is the exception stacks base
>    */
> +.globl exceptions_init
>   exceptions_init:
>   	mrc	p15, 0, r2, c1, c0, 0	@ read SCTLR
>   	bic	r2, #CR_V		@ SCTLR.V := 0
> diff --git a/arm/cstart64.S b/arm/cstart64.S
> index e4ab7d06..223c1092 100644
> --- a/arm/cstart64.S
> +++ b/arm/cstart64.S
> @@ -265,6 +265,7 @@ asm_mmu_disable:
>    * Vectors
>    */
>   
> +.globl exceptions_init
>   exceptions_init:
>   	adrp	x4, vector_table
>   	add	x4, x4, :lo12:vector_table
> diff --git a/lib/arm/asm/setup.h b/lib/arm/asm/setup.h
> index 64cd379b..06069116 100644
> --- a/lib/arm/asm/setup.h
> +++ b/lib/arm/asm/setup.h
> @@ -38,4 +38,12 @@ extern unsigned int mem_region_get_flags(phys_addr_t paddr);
>   
>   void setup(const void *fdt, phys_addr_t freemem_start);
>   
> +#ifdef CONFIG_EFI
> +
> +#include <efi.h>
> +
> +efi_status_t setup_efi(efi_bootinfo_t *efi_bootinfo);
> +
> +#endif
> +
>   #endif /* _ASMARM_SETUP_H_ */
> diff --git a/lib/arm/setup.c b/lib/arm/setup.c
> index 03a4098e..cab19b1e 100644
> --- a/lib/arm/setup.c
> +++ b/lib/arm/setup.c
> @@ -33,7 +33,7 @@
>   #define NR_EXTRA_MEM_REGIONS	16
>   #define NR_INITIAL_MEM_REGIONS	(MAX_DT_MEM_REGIONS + NR_EXTRA_MEM_REGIONS)
>   
> -extern unsigned long _etext;
> +extern unsigned long _text, _etext, _data, _edata;
>   
>   char *initrd;
>   u32 initrd_size;
> @@ -43,7 +43,10 @@ int nr_cpus;
>   
>   static struct mem_region __initial_mem_regions[NR_INITIAL_MEM_REGIONS + 1];
>   struct mem_region *mem_regions = __initial_mem_regions;
> -phys_addr_t __phys_offset, __phys_end;
> +phys_addr_t __phys_offset = (phys_addr_t)-1, __phys_end = 0;
> +
> +extern void exceptions_init(void);
> +extern void asm_mmu_disable(void);
>   
>   int mpidr_to_cpu(uint64_t mpidr)
>   {
> @@ -289,3 +292,177 @@ void setup(const void *fdt, phys_addr_t freemem_start)
>   	if (!(auxinfo.flags & AUXINFO_MMU_OFF))
>   		setup_vm();
>   }
> +
> +#ifdef CONFIG_EFI
> +
> +#include <efi.h>
> +
> +static efi_status_t setup_rsdp(efi_bootinfo_t *efi_bootinfo)
> +{
> +	efi_status_t status;
> +	struct acpi_table_rsdp *rsdp;
> +
> +	/*
> +	 * RSDP resides in an EFI_ACPI_RECLAIM_MEMORY region, which is not used
> +	 * by kvm-unit-tests arm64 memory allocator. So it is not necessary to
> +	 * copy the data structure to another memory region to prevent
> +	 * unintentional overwrite.
> +	 */
> +	status = efi_get_system_config_table(ACPI_20_TABLE_GUID, (void **)&rsdp);
> +	if (status != EFI_SUCCESS)
> +		return status;
> +
> +	set_efi_rsdp(rsdp);
> +
> +	return EFI_SUCCESS;
> +}
> +
> +static efi_status_t efi_mem_init(efi_bootinfo_t *efi_bootinfo)
> +{
> +	int i;
> +	unsigned long free_mem_pages = 0;
> +	unsigned long free_mem_start = 0;
> +	struct efi_boot_memmap *map = &(efi_bootinfo->mem_map);
> +	efi_memory_desc_t *buffer = *map->map;
> +	efi_memory_desc_t *d = NULL;
> +	phys_addr_t base, top;
> +	struct mem_region *r;
> +	uintptr_t text = (uintptr_t)&_text, etext = __ALIGN((uintptr_t)&_etext, 4096);
> +	uintptr_t data = (uintptr_t)&_data, edata = __ALIGN((uintptr_t)&_edata, 4096);
> +
> +	/*
> +	 * Record the largest free EFI_CONVENTIONAL_MEMORY region
> +	 * which will be used to set up the memory allocator, so that
> +	 * the memory allocator can work in the largest free
> +	 * continuous memory region.
> +	 */
> +	for (i = 0, r = &mem_regions[0]; i < *(map->map_size); i += *(map->desc_size), ++r) {

At here, we can see here use the mem_regions to record the 
efi_boot_memmap information, so we will iterate the efi_boot_memmap 
which has (*map->map_size)/(*map->desc_size) number of the structure. 
Obviously, here didn't check if the mem_regions is fulled, so when the 
efi_boot_memmap is bigger than the mem_regions, the memory overflow happens.

And when memory overflow happens, Coincidentally, the mmu_idmap is just 
follow the memory of the mem_regions, so this iteration will write to 
mmu_idmap memory, which cause the mmu_idmap not NULL, so when the first 
time the __ioremap being called, which the call trace is:

efi_main->
   setup_efi->
     io_init->
       uart0_init_acpi->
         ioremap->
	  __ioremap

	   if (mmu_enabled()) {
                    pgtable = current_thread_info()->pgtable;
            } else {
                    if (!mmu_idmap)
                            mmu_idmap = alloc_page();
                    pgtable = mmu_idmap;
            }

When it first arrive at here, the mmu_idmap should be NULL, and a new 
mmu_idmap will be allocated, but unfortunately, the mmu_idmap has been 
write to a value, so it is not NULL, so the dirty mmu_idmap will be used 
as a pgtable. Which cause the DABT_EL1 error when continue build the 
page table.

And the solution is very easy, just make the mem_regions bigger, for 
example:

static struct mem_region __initial_mem_regions[NR_INITIAL_MEM_REGIONS + 20];
struct mem_region *mem_regions = __initial_mem_regions;

After make it bigger, the DABT_EL1 error will not happen on my machine. 
Hope it works for you.

Thanks,
Shaoqin

> +		d = (efi_memory_desc_t *)(&((u8 *)buffer)[i]);
> +
> +		r->start = d->phys_addr;
> +		r->end = d->phys_addr + d->num_pages * EFI_PAGE_SIZE;
> +
> +		switch (d->type) {
> +		case EFI_RESERVED_TYPE:
> +		case EFI_LOADER_DATA:
> +		case EFI_BOOT_SERVICES_CODE:
> +		case EFI_BOOT_SERVICES_DATA:
> +		case EFI_RUNTIME_SERVICES_CODE:
> +		case EFI_RUNTIME_SERVICES_DATA:
> +		case EFI_UNUSABLE_MEMORY:
> +		case EFI_ACPI_RECLAIM_MEMORY:
> +		case EFI_ACPI_MEMORY_NVS:
> +		case EFI_PAL_CODE:
> +			r->flags = MR_F_RESERVED;
> +			break;
> +		case EFI_MEMORY_MAPPED_IO:
> +		case EFI_MEMORY_MAPPED_IO_PORT_SPACE:
> +			r->flags = MR_F_IO;
> +			break;
> +		case EFI_LOADER_CODE:
> +			if (r->start <= text && r->end > text) {
> +				/* This is the unit test region. Flag the code separately. */
> +				phys_addr_t tmp = r->end;
> +
> +				assert(etext <= data);
> +				assert(edata <= r->end);
> +				r->flags = MR_F_CODE;
> +				r->end = data;
> +				++r;
> +				r->start = data;
> +				r->end = tmp;
> +			} else {
> +				r->flags = MR_F_RESERVED;
> +			}
> +			break;
> +		case EFI_CONVENTIONAL_MEMORY:
> +			if (free_mem_pages < d->num_pages) {
> +				free_mem_pages = d->num_pages;
> +				free_mem_start = d->phys_addr;
> +			}
> +			break;
> +		}
> +
> +		if (!(r->flags & MR_F_IO)) {
> +			if (r->start < __phys_offset)
> +				__phys_offset = r->start;
> +			if (r->end > __phys_end)
> +				__phys_end = r->end;
> +		}
> +	}
> +	__phys_end &= PHYS_MASK;
> +	asm_mmu_disable();
> +
> +	if (free_mem_pages == 0)
> +		return EFI_OUT_OF_RESOURCES;
> +
> +	assert(sizeof(long) == 8 || free_mem_start < (3ul << 30));
> +
> +	phys_alloc_init(free_mem_start, free_mem_pages << EFI_PAGE_SHIFT);
> +	phys_alloc_set_minimum_alignment(SMP_CACHE_BYTES);
> +
> +	phys_alloc_get_unused(&base, &top);
> +	base = PAGE_ALIGN(base);
> +	top = top & PAGE_MASK;
> +	assert(sizeof(long) == 8 || !(base >> 32));
> +	if (sizeof(long) != 8 && (top >> 32) != 0)
> +		top = ((uint64_t)1 << 32);
> +	page_alloc_init_area(0, base >> PAGE_SHIFT, top >> PAGE_SHIFT);
> +	page_alloc_ops_enable();
> +
> +	return EFI_SUCCESS;
> +}
> +
> +efi_status_t setup_efi(efi_bootinfo_t *efi_bootinfo)
> +{
> +	efi_status_t status;
> +
> +	struct thread_info *ti = current_thread_info();
> +
> +	memset(ti, 0, sizeof(*ti));
> +
> +	exceptions_init();
> +
> +	status = efi_mem_init(efi_bootinfo);
> +	if (status != EFI_SUCCESS) {
> +		printf("Failed to initialize memory: ");
> +		switch (status) {
> +		case EFI_OUT_OF_RESOURCES:
> +			printf("No free memory region\n");
> +			break;
> +		default:
> +			printf("Unknown error\n");
> +			break;
> +		}
> +		return status;
> +	}
> +
> +	status = setup_rsdp(efi_bootinfo);
> +	if (status != EFI_SUCCESS) {
> +		printf("Cannot find RSDP in EFI system table\n");
> +		return status;
> +	}
> +
> +	psci_set_conduit();
> +	cpu_init();
> +	/* cpu_init must be called before thread_info_init */
> +	thread_info_init(current_thread_info(), 0);
> +	/* mem_init must be called before io_init */
> +	io_init();
> +
> +	timer_save_state();
> +	if (initrd) {
> +		/* environ is currently the only file in the initrd */
> +		char *env = malloc(initrd_size);
> +
> +		memcpy(env, initrd, initrd_size);
> +		setup_env(env, initrd_size);
> +	}
> +
> +	if (!(auxinfo.flags & AUXINFO_MMU_OFF))
> +		setup_vm();
> +
> +	return EFI_SUCCESS;
> +}
> +
> +#endif
> diff --git a/lib/linux/efi.h b/lib/linux/efi.h
> index 53748dd4..89f9a9e0 100644
> --- a/lib/linux/efi.h
> +++ b/lib/linux/efi.h
> @@ -63,6 +63,7 @@ typedef guid_t efi_guid_t;
>   	(c) & 0xff, ((c) >> 8) & 0xff, d } }
>   
>   #define ACPI_TABLE_GUID EFI_GUID(0xeb9d2d30, 0x2d88, 0x11d3, 0x9a, 0x16, 0x00, 0x90, 0x27, 0x3f, 0xc1, 0x4d)
> +#define ACPI_20_TABLE_GUID EFI_GUID(0x8868e871, 0xe4f1, 0x11d3,  0xbc, 0x22, 0x00, 0x80, 0xc7, 0x3c, 0x88, 0x81)
>   
>   #define LOADED_IMAGE_PROTOCOL_GUID EFI_GUID(0x5b1b31a1, 0x9562, 0x11d2,  0x8e, 0x3f, 0x00, 0xa0, 0xc9, 0x69, 0x72, 0x3b)
>   

-- 
Shaoqin

Re: [PATCH v4 23/30] arm64: Add a setup sequence for systems that boot through EFI

[Nikos Nikoleris]

Hi Shaoqin,

On 25/04/2023 08:04, Shaoqin Huang wrote:
> Hi Nikos,
> 
> For that DABT_EL1 error, I have some clues about how it happens. It's
> mainly because this patch includes a memory overflow. I will explain in
> the code body.
> 

Many thanks for this. This is the 2nd time I get caught by this :(

> On 2/13/23 18:17, Nikos Nikoleris wrote:
>> This change implements an alternative setup sequence for the system
>> when we are booting through EFI. The memory map is discovered through
>> EFI boot services and devices through ACPI.
>>
>> This change is based on a change initially proposed by
>> Andrew Jones <drjones@redhat.com>
>>
>> Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
>> ---
>>    arm/cstart.S        |   1 +
>>    arm/cstart64.S      |   1 +
>>    lib/arm/asm/setup.h |   8 ++
>>    lib/arm/setup.c     | 181 +++++++++++++++++++++++++++++++++++++++++++-
>>    lib/linux/efi.h     |   1 +
>>    5 files changed, 190 insertions(+), 2 deletions(-)
>>
>> diff --git a/arm/cstart.S b/arm/cstart.S
>> index 7036e67f..3dd71ed9 100644
>> --- a/arm/cstart.S
>> +++ b/arm/cstart.S
>> @@ -242,6 +242,7 @@ asm_mmu_disable:
>>     *
>>     * Input r0 is the stack top, which is the exception stacks base
>>     */
>> +.globl exceptions_init
>>    exceptions_init:
>>    	mrc	p15, 0, r2, c1, c0, 0	@ read SCTLR
>>    	bic	r2, #CR_V		@ SCTLR.V := 0
>> diff --git a/arm/cstart64.S b/arm/cstart64.S
>> index e4ab7d06..223c1092 100644
>> --- a/arm/cstart64.S
>> +++ b/arm/cstart64.S
>> @@ -265,6 +265,7 @@ asm_mmu_disable:
>>     * Vectors
>>     */
>>    
>> +.globl exceptions_init
>>    exceptions_init:
>>    	adrp	x4, vector_table
>>    	add	x4, x4, :lo12:vector_table
>> diff --git a/lib/arm/asm/setup.h b/lib/arm/asm/setup.h
>> index 64cd379b..06069116 100644
>> --- a/lib/arm/asm/setup.h
>> +++ b/lib/arm/asm/setup.h
>> @@ -38,4 +38,12 @@ extern unsigned int mem_region_get_flags(phys_addr_t paddr);
>>    
>>    void setup(const void *fdt, phys_addr_t freemem_start);
>>    
>> +#ifdef CONFIG_EFI
>> +
>> +#include <efi.h>
>> +
>> +efi_status_t setup_efi(efi_bootinfo_t *efi_bootinfo);
>> +
>> +#endif
>> +
>>    #endif /* _ASMARM_SETUP_H_ */
>> diff --git a/lib/arm/setup.c b/lib/arm/setup.c
>> index 03a4098e..cab19b1e 100644
>> --- a/lib/arm/setup.c
>> +++ b/lib/arm/setup.c
>> @@ -33,7 +33,7 @@
>>    #define NR_EXTRA_MEM_REGIONS	16
>>    #define NR_INITIAL_MEM_REGIONS	(MAX_DT_MEM_REGIONS + NR_EXTRA_MEM_REGIONS)
>>    
>> -extern unsigned long _etext;
>> +extern unsigned long _text, _etext, _data, _edata;
>>    
>>    char *initrd;
>>    u32 initrd_size;
>> @@ -43,7 +43,10 @@ int nr_cpus;
>>    
>>    static struct mem_region __initial_mem_regions[NR_INITIAL_MEM_REGIONS + 1];
>>    struct mem_region *mem_regions = __initial_mem_regions;
>> -phys_addr_t __phys_offset, __phys_end;
>> +phys_addr_t __phys_offset = (phys_addr_t)-1, __phys_end = 0;
>> +
>> +extern void exceptions_init(void);
>> +extern void asm_mmu_disable(void);
>>    
>>    int mpidr_to_cpu(uint64_t mpidr)
>>    {
>> @@ -289,3 +292,177 @@ void setup(const void *fdt, phys_addr_t freemem_start)
>>    	if (!(auxinfo.flags & AUXINFO_MMU_OFF))
>>    		setup_vm();
>>    }
>> +
>> +#ifdef CONFIG_EFI
>> +
>> +#include <efi.h>
>> +
>> +static efi_status_t setup_rsdp(efi_bootinfo_t *efi_bootinfo)
>> +{
>> +	efi_status_t status;
>> +	struct acpi_table_rsdp *rsdp;
>> +
>> +	/*
>> +	 * RSDP resides in an EFI_ACPI_RECLAIM_MEMORY region, which is not used
>> +	 * by kvm-unit-tests arm64 memory allocator. So it is not necessary to
>> +	 * copy the data structure to another memory region to prevent
>> +	 * unintentional overwrite.
>> +	 */
>> +	status = efi_get_system_config_table(ACPI_20_TABLE_GUID, (void **)&rsdp);
>> +	if (status != EFI_SUCCESS)
>> +		return status;
>> +
>> +	set_efi_rsdp(rsdp);
>> +
>> +	return EFI_SUCCESS;
>> +}
>> +
>> +static efi_status_t efi_mem_init(efi_bootinfo_t *efi_bootinfo)
>> +{
>> +	int i;
>> +	unsigned long free_mem_pages = 0;
>> +	unsigned long free_mem_start = 0;
>> +	struct efi_boot_memmap *map = &(efi_bootinfo->mem_map);
>> +	efi_memory_desc_t *buffer = *map->map;
>> +	efi_memory_desc_t *d = NULL;
>> +	phys_addr_t base, top;
>> +	struct mem_region *r;
>> +	uintptr_t text = (uintptr_t)&_text, etext = __ALIGN((uintptr_t)&_etext, 4096);
>> +	uintptr_t data = (uintptr_t)&_data, edata = __ALIGN((uintptr_t)&_edata, 4096);
>> +
>> +	/*
>> +	 * Record the largest free EFI_CONVENTIONAL_MEMORY region
>> +	 * which will be used to set up the memory allocator, so that
>> +	 * the memory allocator can work in the largest free
>> +	 * continuous memory region.
>> +	 */
>> +	for (i = 0, r = &mem_regions[0]; i < *(map->map_size); i += *(map->desc_size), ++r) {
> 
> At here, we can see here use the mem_regions to record the
> efi_boot_memmap information, so we will iterate the efi_boot_memmap
> which has (*map->map_size)/(*map->desc_size) number of the structure.
> Obviously, here didn't check if the mem_regions is fulled, so when the
> efi_boot_memmap is bigger than the mem_regions, the memory overflow happens.
> 
> And when memory overflow happens, Coincidentally, the mmu_idmap is just
> follow the memory of the mem_regions, so this iteration will write to
> mmu_idmap memory, which cause the mmu_idmap not NULL, so when the first
> time the __ioremap being called, which the call trace is:
> 
> efi_main->
>     setup_efi->
>       io_init->
>         uart0_init_acpi->
>           ioremap->
> 	  __ioremap
> 
> 	   if (mmu_enabled()) {
>                      pgtable = current_thread_info()->pgtable;
>              } else {
>                      if (!mmu_idmap)
>                              mmu_idmap = alloc_page();
>                      pgtable = mmu_idmap;
>              }
> 
> When it first arrive at here, the mmu_idmap should be NULL, and a new
> mmu_idmap will be allocated, but unfortunately, the mmu_idmap has been
> write to a value, so it is not NULL, so the dirty mmu_idmap will be used
> as a pgtable. Which cause the DABT_EL1 error when continue build the
> page table.
> 
> And the solution is very easy, just make the mem_regions bigger, for
> example:
> 
> static struct mem_region __initial_mem_regions[NR_INITIAL_MEM_REGIONS + 20];
> struct mem_region *mem_regions = __initial_mem_regions;
> 
> After make it bigger, the DABT_EL1 error will not happen on my machine.
> Hope it works for you.
> 

Indeed, I can confirm that this is the issue I run into. It would be 
nice if Drew can confirm as well. Just to be on the safe side in the v5 
I will apply these changes.

Thanks,

Nikos

 From a836dc91706cc9e9aee5ce6b8b659d74d98c7bd7 Mon Sep 17 00:00:00 2001
From: Nikos Nikoleris <nikos.nikoleris@arm.com>
Date: Wed, 3 Aug 2022 13:47:56 +0100
Subject: [kvm-unit-tests PATCH] fixup! arm/arm64: Add a setup sequence 
for systems that boot through EFI
X-ARM-No-Footer: FoSSMail

---
  lib/arm/setup.c | 45 ++++++++++++++++++++++++---------------------
  1 file changed, 24 insertions(+), 21 deletions(-)

diff --git a/lib/arm/setup.c b/lib/arm/setup.c
index cab19b1e..c4f495a9 100644
--- a/lib/arm/setup.c
+++ b/lib/arm/setup.c
@@ -30,7 +30,7 @@
  #include "io.h"

  #define MAX_DT_MEM_REGIONS	16
-#define NR_EXTRA_MEM_REGIONS	16
+#define NR_EXTRA_MEM_REGIONS	64
  #define NR_INITIAL_MEM_REGIONS	(MAX_DT_MEM_REGIONS + NR_EXTRA_MEM_REGIONS)

  extern unsigned long _text, _etext, _data, _edata;
@@ -326,7 +326,7 @@ static efi_status_t efi_mem_init(efi_bootinfo_t 
*efi_bootinfo)
  	efi_memory_desc_t *buffer = *map->map;
  	efi_memory_desc_t *d = NULL;
  	phys_addr_t base, top;
-	struct mem_region *r;
+	struct mem_region r;
  	uintptr_t text = (uintptr_t)&_text, etext = 
__ALIGN((uintptr_t)&_etext, 4096);
  	uintptr_t data = (uintptr_t)&_data, edata = 
__ALIGN((uintptr_t)&_edata, 4096);

@@ -336,11 +336,12 @@ static efi_status_t efi_mem_init(efi_bootinfo_t 
*efi_bootinfo)
  	 * the memory allocator can work in the largest free
  	 * continuous memory region.
  	 */
-	for (i = 0, r = &mem_regions[0]; i < *(map->map_size); i += 
*(map->desc_size), ++r) {
+	for (i = 0; i < *(map->map_size); i += *(map->desc_size)) {
  		d = (efi_memory_desc_t *)(&((u8 *)buffer)[i]);

-		r->start = d->phys_addr;
-		r->end = d->phys_addr + d->num_pages * EFI_PAGE_SIZE;
+		r.start = d->phys_addr;
+		r.end = d->phys_addr + d->num_pages * EFI_PAGE_SIZE;
+		r.flags = 0;

  		switch (d->type) {
  		case EFI_RESERVED_TYPE:
@@ -353,26 +354,27 @@ static efi_status_t efi_mem_init(efi_bootinfo_t 
*efi_bootinfo)
  		case EFI_ACPI_RECLAIM_MEMORY:
  		case EFI_ACPI_MEMORY_NVS:
  		case EFI_PAL_CODE:
-			r->flags = MR_F_RESERVED;
+			r.flags = MR_F_RESERVED;
  			break;
  		case EFI_MEMORY_MAPPED_IO:
  		case EFI_MEMORY_MAPPED_IO_PORT_SPACE:
-			r->flags = MR_F_IO;
+			r.flags = MR_F_IO;
  			break;
  		case EFI_LOADER_CODE:
-			if (r->start <= text && r->end > text) {
+			if (r.start <= text && r.end > text) {
  				/* This is the unit test region. Flag the code separately. */
-				phys_addr_t tmp = r->end;
+				phys_addr_t tmp = r.end;

  				assert(etext <= data);
-				assert(edata <= r->end);
-				r->flags = MR_F_CODE;
-				r->end = data;
-				++r;
-				r->start = data;
-				r->end = tmp;
+				assert(edata <= r.end);
+				r.flags = MR_F_CODE;
+				r.end = data;
+				mem_region_add(&r);
+				r.start = data;
+				r.end = tmp;
+				r.flags = 0;
  			} else {
-				r->flags = MR_F_RESERVED;
+				r.flags = MR_F_RESERVED;
  			}
  			break;
  		case EFI_CONVENTIONAL_MEMORY:
@@ -383,12 +385,13 @@ static efi_status_t efi_mem_init(efi_bootinfo_t 
*efi_bootinfo)
  			break;
  		}

-		if (!(r->flags & MR_F_IO)) {
-			if (r->start < __phys_offset)
-				__phys_offset = r->start;
-			if (r->end > __phys_end)
-				__phys_end = r->end;
+		if (!(r.flags & MR_F_IO)) {
+			if (r.start < __phys_offset)
+				__phys_offset = r.start;
+			if (r.end > __phys_end)
+				__phys_end = r.end;
  		}
+		mem_region_add(&r);
  	}
  	__phys_end &= PHYS_MASK;
  	asm_mmu_disable();
--
2.25.1

Re: [PATCH v4 23/30] arm64: Add a setup sequence for systems that boot through EFI

[Andrew Jones]

On Tue, Apr 25, 2023 at 10:09:51AM +0100, Nikos Nikoleris wrote:
> Hi Shaoqin,
> 
> On 25/04/2023 08:04, Shaoqin Huang wrote:
> > Hi Nikos,
> > 
> > For that DABT_EL1 error, I have some clues about how it happens. It's
> > mainly because this patch includes a memory overflow. I will explain in
> > the code body.
> > 
> 
> Many thanks for this. This is the 2nd time I get caught by this :(

Yes, thank you Shaoqin for getting to the bottom of this and reporting it!

> 
> > On 2/13/23 18:17, Nikos Nikoleris wrote:
> > > This change implements an alternative setup sequence for the system
> > > when we are booting through EFI. The memory map is discovered through
> > > EFI boot services and devices through ACPI.
> > > 
> > > This change is based on a change initially proposed by
> > > Andrew Jones <drjones@redhat.com>
> > > 
> > > Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
> > > ---
> > >    arm/cstart.S        |   1 +
> > >    arm/cstart64.S      |   1 +
> > >    lib/arm/asm/setup.h |   8 ++
> > >    lib/arm/setup.c     | 181 +++++++++++++++++++++++++++++++++++++++++++-
> > >    lib/linux/efi.h     |   1 +
> > >    5 files changed, 190 insertions(+), 2 deletions(-)
> > > 
> > > diff --git a/arm/cstart.S b/arm/cstart.S
> > > index 7036e67f..3dd71ed9 100644
> > > --- a/arm/cstart.S
> > > +++ b/arm/cstart.S
> > > @@ -242,6 +242,7 @@ asm_mmu_disable:
> > >     *
> > >     * Input r0 is the stack top, which is the exception stacks base
> > >     */
> > > +.globl exceptions_init
> > >    exceptions_init:
> > >    	mrc	p15, 0, r2, c1, c0, 0	@ read SCTLR
> > >    	bic	r2, #CR_V		@ SCTLR.V := 0
> > > diff --git a/arm/cstart64.S b/arm/cstart64.S
> > > index e4ab7d06..223c1092 100644
> > > --- a/arm/cstart64.S
> > > +++ b/arm/cstart64.S
> > > @@ -265,6 +265,7 @@ asm_mmu_disable:
> > >     * Vectors
> > >     */
> > > +.globl exceptions_init
> > >    exceptions_init:
> > >    	adrp	x4, vector_table
> > >    	add	x4, x4, :lo12:vector_table
> > > diff --git a/lib/arm/asm/setup.h b/lib/arm/asm/setup.h
> > > index 64cd379b..06069116 100644
> > > --- a/lib/arm/asm/setup.h
> > > +++ b/lib/arm/asm/setup.h
> > > @@ -38,4 +38,12 @@ extern unsigned int mem_region_get_flags(phys_addr_t paddr);
> > >    void setup(const void *fdt, phys_addr_t freemem_start);
> > > +#ifdef CONFIG_EFI
> > > +
> > > +#include <efi.h>
> > > +
> > > +efi_status_t setup_efi(efi_bootinfo_t *efi_bootinfo);
> > > +
> > > +#endif
> > > +
> > >    #endif /* _ASMARM_SETUP_H_ */
> > > diff --git a/lib/arm/setup.c b/lib/arm/setup.c
> > > index 03a4098e..cab19b1e 100644
> > > --- a/lib/arm/setup.c
> > > +++ b/lib/arm/setup.c
> > > @@ -33,7 +33,7 @@
> > >    #define NR_EXTRA_MEM_REGIONS	16
> > >    #define NR_INITIAL_MEM_REGIONS	(MAX_DT_MEM_REGIONS + NR_EXTRA_MEM_REGIONS)
> > > -extern unsigned long _etext;
> > > +extern unsigned long _text, _etext, _data, _edata;
> > >    char *initrd;
> > >    u32 initrd_size;
> > > @@ -43,7 +43,10 @@ int nr_cpus;
> > >    static struct mem_region __initial_mem_regions[NR_INITIAL_MEM_REGIONS + 1];
> > >    struct mem_region *mem_regions = __initial_mem_regions;
> > > -phys_addr_t __phys_offset, __phys_end;
> > > +phys_addr_t __phys_offset = (phys_addr_t)-1, __phys_end = 0;
> > > +
> > > +extern void exceptions_init(void);
> > > +extern void asm_mmu_disable(void);
> > >    int mpidr_to_cpu(uint64_t mpidr)
> > >    {
> > > @@ -289,3 +292,177 @@ void setup(const void *fdt, phys_addr_t freemem_start)
> > >    	if (!(auxinfo.flags & AUXINFO_MMU_OFF))
> > >    		setup_vm();
> > >    }
> > > +
> > > +#ifdef CONFIG_EFI
> > > +
> > > +#include <efi.h>
> > > +
> > > +static efi_status_t setup_rsdp(efi_bootinfo_t *efi_bootinfo)
> > > +{
> > > +	efi_status_t status;
> > > +	struct acpi_table_rsdp *rsdp;
> > > +
> > > +	/*
> > > +	 * RSDP resides in an EFI_ACPI_RECLAIM_MEMORY region, which is not used
> > > +	 * by kvm-unit-tests arm64 memory allocator. So it is not necessary to
> > > +	 * copy the data structure to another memory region to prevent
> > > +	 * unintentional overwrite.
> > > +	 */
> > > +	status = efi_get_system_config_table(ACPI_20_TABLE_GUID, (void **)&rsdp);
> > > +	if (status != EFI_SUCCESS)
> > > +		return status;
> > > +
> > > +	set_efi_rsdp(rsdp);
> > > +
> > > +	return EFI_SUCCESS;
> > > +}
> > > +
> > > +static efi_status_t efi_mem_init(efi_bootinfo_t *efi_bootinfo)
> > > +{
> > > +	int i;
> > > +	unsigned long free_mem_pages = 0;
> > > +	unsigned long free_mem_start = 0;
> > > +	struct efi_boot_memmap *map = &(efi_bootinfo->mem_map);
> > > +	efi_memory_desc_t *buffer = *map->map;
> > > +	efi_memory_desc_t *d = NULL;
> > > +	phys_addr_t base, top;
> > > +	struct mem_region *r;
> > > +	uintptr_t text = (uintptr_t)&_text, etext = __ALIGN((uintptr_t)&_etext, 4096);
> > > +	uintptr_t data = (uintptr_t)&_data, edata = __ALIGN((uintptr_t)&_edata, 4096);
> > > +
> > > +	/*
> > > +	 * Record the largest free EFI_CONVENTIONAL_MEMORY region
> > > +	 * which will be used to set up the memory allocator, so that
> > > +	 * the memory allocator can work in the largest free
> > > +	 * continuous memory region.
> > > +	 */
> > > +	for (i = 0, r = &mem_regions[0]; i < *(map->map_size); i += *(map->desc_size), ++r) {
> > 
> > At here, we can see here use the mem_regions to record the
> > efi_boot_memmap information, so we will iterate the efi_boot_memmap
> > which has (*map->map_size)/(*map->desc_size) number of the structure.
> > Obviously, here didn't check if the mem_regions is fulled, so when the
> > efi_boot_memmap is bigger than the mem_regions, the memory overflow happens.
> > 
> > And when memory overflow happens, Coincidentally, the mmu_idmap is just
> > follow the memory of the mem_regions, so this iteration will write to
> > mmu_idmap memory, which cause the mmu_idmap not NULL, so when the first
> > time the __ioremap being called, which the call trace is:
> > 
> > efi_main->
> >     setup_efi->
> >       io_init->
> >         uart0_init_acpi->
> >           ioremap->
> > 	  __ioremap
> > 
> > 	   if (mmu_enabled()) {
> >                      pgtable = current_thread_info()->pgtable;
> >              } else {
> >                      if (!mmu_idmap)
> >                              mmu_idmap = alloc_page();
> >                      pgtable = mmu_idmap;
> >              }
> > 
> > When it first arrive at here, the mmu_idmap should be NULL, and a new
> > mmu_idmap will be allocated, but unfortunately, the mmu_idmap has been
> > write to a value, so it is not NULL, so the dirty mmu_idmap will be used
> > as a pgtable. Which cause the DABT_EL1 error when continue build the
> > page table.
> > 
> > And the solution is very easy, just make the mem_regions bigger, for
> > example:
> > 
> > static struct mem_region __initial_mem_regions[NR_INITIAL_MEM_REGIONS + 20];
> > struct mem_region *mem_regions = __initial_mem_regions;
> > 
> > After make it bigger, the DABT_EL1 error will not happen on my machine.
> > Hope it works for you.
> > 
> 
> Indeed, I can confirm that this is the issue I run into. It would be nice if
> Drew can confirm as well. Just to be on the safe side in the v5 I will apply
> these changes.
> 
> Thanks,
> 
> Nikos

I gave the below patch a try and it appears to have resolved the issue.
I'll look forward to a refresh of the series which, fingers crossed,
will be finally merged!

Thanks,
drew

> 
> From a836dc91706cc9e9aee5ce6b8b659d74d98c7bd7 Mon Sep 17 00:00:00 2001
> From: Nikos Nikoleris <nikos.nikoleris@arm.com>
> Date: Wed, 3 Aug 2022 13:47:56 +0100
> Subject: [kvm-unit-tests PATCH] fixup! arm/arm64: Add a setup sequence for
> systems that boot through EFI
> X-ARM-No-Footer: FoSSMail
> 
> ---
>  lib/arm/setup.c | 45 ++++++++++++++++++++++++---------------------
>  1 file changed, 24 insertions(+), 21 deletions(-)
> 
> diff --git a/lib/arm/setup.c b/lib/arm/setup.c
> index cab19b1e..c4f495a9 100644
> --- a/lib/arm/setup.c
> +++ b/lib/arm/setup.c
> @@ -30,7 +30,7 @@
>  #include "io.h"
> 
>  #define MAX_DT_MEM_REGIONS	16
> -#define NR_EXTRA_MEM_REGIONS	16
> +#define NR_EXTRA_MEM_REGIONS	64
>  #define NR_INITIAL_MEM_REGIONS	(MAX_DT_MEM_REGIONS + NR_EXTRA_MEM_REGIONS)
> 
>  extern unsigned long _text, _etext, _data, _edata;
> @@ -326,7 +326,7 @@ static efi_status_t efi_mem_init(efi_bootinfo_t
> *efi_bootinfo)
>  	efi_memory_desc_t *buffer = *map->map;
>  	efi_memory_desc_t *d = NULL;
>  	phys_addr_t base, top;
> -	struct mem_region *r;
> +	struct mem_region r;
>  	uintptr_t text = (uintptr_t)&_text, etext = __ALIGN((uintptr_t)&_etext,
> 4096);
>  	uintptr_t data = (uintptr_t)&_data, edata = __ALIGN((uintptr_t)&_edata,
> 4096);
> 
> @@ -336,11 +336,12 @@ static efi_status_t efi_mem_init(efi_bootinfo_t
> *efi_bootinfo)
>  	 * the memory allocator can work in the largest free
>  	 * continuous memory region.
>  	 */
> -	for (i = 0, r = &mem_regions[0]; i < *(map->map_size); i +=
> *(map->desc_size), ++r) {
> +	for (i = 0; i < *(map->map_size); i += *(map->desc_size)) {
>  		d = (efi_memory_desc_t *)(&((u8 *)buffer)[i]);
> 
> -		r->start = d->phys_addr;
> -		r->end = d->phys_addr + d->num_pages * EFI_PAGE_SIZE;
> +		r.start = d->phys_addr;
> +		r.end = d->phys_addr + d->num_pages * EFI_PAGE_SIZE;
> +		r.flags = 0;
> 
>  		switch (d->type) {
>  		case EFI_RESERVED_TYPE:
> @@ -353,26 +354,27 @@ static efi_status_t efi_mem_init(efi_bootinfo_t
> *efi_bootinfo)
>  		case EFI_ACPI_RECLAIM_MEMORY:
>  		case EFI_ACPI_MEMORY_NVS:
>  		case EFI_PAL_CODE:
> -			r->flags = MR_F_RESERVED;
> +			r.flags = MR_F_RESERVED;
>  			break;
>  		case EFI_MEMORY_MAPPED_IO:
>  		case EFI_MEMORY_MAPPED_IO_PORT_SPACE:
> -			r->flags = MR_F_IO;
> +			r.flags = MR_F_IO;
>  			break;
>  		case EFI_LOADER_CODE:
> -			if (r->start <= text && r->end > text) {
> +			if (r.start <= text && r.end > text) {
>  				/* This is the unit test region. Flag the code separately. */
> -				phys_addr_t tmp = r->end;
> +				phys_addr_t tmp = r.end;
> 
>  				assert(etext <= data);
> -				assert(edata <= r->end);
> -				r->flags = MR_F_CODE;
> -				r->end = data;
> -				++r;
> -				r->start = data;
> -				r->end = tmp;
> +				assert(edata <= r.end);
> +				r.flags = MR_F_CODE;
> +				r.end = data;
> +				mem_region_add(&r);
> +				r.start = data;
> +				r.end = tmp;
> +				r.flags = 0;
>  			} else {
> -				r->flags = MR_F_RESERVED;
> +				r.flags = MR_F_RESERVED;
>  			}
>  			break;
>  		case EFI_CONVENTIONAL_MEMORY:
> @@ -383,12 +385,13 @@ static efi_status_t efi_mem_init(efi_bootinfo_t
> *efi_bootinfo)
>  			break;
>  		}
> 
> -		if (!(r->flags & MR_F_IO)) {
> -			if (r->start < __phys_offset)
> -				__phys_offset = r->start;
> -			if (r->end > __phys_end)
> -				__phys_end = r->end;
> +		if (!(r.flags & MR_F_IO)) {
> +			if (r.start < __phys_offset)
> +				__phys_offset = r.start;
> +			if (r.end > __phys_end)
> +				__phys_end = r.end;
>  		}
> +		mem_region_add(&r);
>  	}
>  	__phys_end &= PHYS_MASK;
>  	asm_mmu_disable();
> --
> 2.25.1