Early cpio decoder and ACPI table override via initrd making use of it

Early cpio decoder and ACPI table override via initrd making use of it

[Thomas Renninger]

Should still apply on latest Linux tree, I can resubmit if not.
Would be great if hpa can push this into a x86 branch to get
this into linux-next and into Linus' tree with the next merge
window.

Thanks,

    Thomas

Dmesg output of a quick test:
 FACP ACPI table found in initrd [kernel/firmware/acpi/facp.aml][0xf4]
 DSDT ACPI table found in initrd [kernel/firmware/acpi/dsdt.aml][0x333c]
 e820: modified physical RAM map:
...
   modified: [mem 0x00000000f7fec000-0x00000000f7fef42f] ACPI data
...
 Disabling lock debugging due to kernel taint
 ACPI: Override [FACP-OEMFACP ], this is unsafe: tainting kernel
 ACPI: FACP 00000000f7ff0281 Physical table override, new table: ffffffffff4b7000
 ACPI: FACP 00000000f7fec000 000F4 (v03 A M I  OEMFACP  08000713 INTL 20120518)
 ACPI: Override [DSDT-0AAAA001], this is unsafe: tainting kernel
 ACPI: DSDT 00000000f7ff0410 Physical table override, new table: ffffffffff4b70f4
 ACPI: DSDT 00000000f7fec0f4 0333C (v01  0AAAA 0AAAA001 00000001 INTL 20120518)

[PATCH 1/2] lib: Add early cpio decoder

[Thomas Renninger]

From: "H. Peter Anvin" <hpa@linux.intel.com>

Add a simple cpio decoder without library dependencies for the purpose
of extracting components from the initramfs blob for early kernel
uses.  Intended consumers so far are microcode and ACPI override.

Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
CC: Thomas Renninger <trenn@suse.de>
Link: http://lkml.kernel.org/r/201203261651.29640.trenn@suse.de
Link: https://lkml.org/lkml/2012/9/21/190
Signed-off-by: Thomas Renninger <trenn@suse.de>
---
 include/linux/earlycpio.h |   17 +++++
 lib/Makefile              |    2 +-
 lib/earlycpio.c           |  142 +++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 160 insertions(+), 1 deletions(-)
 create mode 100644 include/linux/earlycpio.h
 create mode 100644 lib/earlycpio.c

diff --git a/include/linux/earlycpio.h b/include/linux/earlycpio.h
new file mode 100644
index 0000000..111f46d
--- /dev/null
+++ b/include/linux/earlycpio.h
@@ -0,0 +1,17 @@
+#ifndef _LINUX_EARLYCPIO_H
+#define _LINUX_EARLYCPIO_H
+
+#include <linux/types.h>
+
+#define MAX_CPIO_FILE_NAME 18
+
+struct cpio_data {
+	void *data;
+	size_t size;
+	char name[MAX_CPIO_FILE_NAME];
+};
+
+struct cpio_data find_cpio_data(const char *path, void *data, size_t len,
+				long *offset);
+
+#endif /* _LINUX_EARLYCPIO_H */
diff --git a/lib/Makefile b/lib/Makefile
index 42d283e..0924041 100644
--- a/lib/Makefile
+++ b/lib/Makefile
@@ -12,7 +12,7 @@ lib-y := ctype.o string.o vsprintf.o cmdline.o \
 	 idr.o int_sqrt.o extable.o prio_tree.o \
 	 sha1.o md5.o irq_regs.o reciprocal_div.o argv_split.o \
 	 proportions.o flex_proportions.o prio_heap.o ratelimit.o show_mem.o \
-	 is_single_threaded.o plist.o decompress.o
+	 is_single_threaded.o plist.o decompress.o earlycpio.o
 
 lib-$(CONFIG_MMU) += ioremap.o
 lib-$(CONFIG_SMP) += cpumask.o
diff --git a/lib/earlycpio.c b/lib/earlycpio.c
new file mode 100644
index 0000000..c6f1a20
--- /dev/null
+++ b/lib/earlycpio.c
@@ -0,0 +1,142 @@
+/* ----------------------------------------------------------------------- *
+ *
+ *   Copyright 2012 Intel Corporation; author H. Peter Anvin
+ *
+ *   This file is part of the Linux kernel, and is made available
+ *   under the terms of the GNU General Public License version 2, as
+ *   published by the Free Software Foundation.
+ *
+ *   This program is distributed in the hope it will be useful, but
+ *   WITHOUT ANY WARRANTY; without even the implied warranty of
+ *   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ *   General Public License for more details.
+ *
+ * ----------------------------------------------------------------------- */
+
+/*
+ * earlycpio.c
+ *
+ * Find a specific cpio member; must precede any compressed content.
+ * This is used to locate data items in the initramfs used by the
+ * kernel itself during early boot (before the main initramfs is
+ * decompressed.)  It is the responsibility of the initramfs creator
+ * to ensure that these items are uncompressed at the head of the
+ * blob.  Depending on the boot loader or package tool that may be a
+ * separate file or part of the same file.
+ */
+
+#include <linux/earlycpio.h>
+#include <linux/kernel.h>
+#include <linux/string.h>
+
+enum cpio_fields {
+	C_MAGIC,
+	C_INO,
+	C_MODE,
+	C_UID,
+	C_GID,
+	C_NLINK,
+	C_MTIME,
+	C_FILESIZE,
+	C_MAJ,
+	C_MIN,
+	C_RMAJ,
+	C_RMIN,
+	C_NAMESIZE,
+	C_CHKSUM,
+	C_NFIELDS
+};
+
+/**
+ * cpio_data find_cpio_data - Search for files in an uncompressed cpio
+ * @path:   The directory to search for, including a slash at the end
+ * @data:   Pointer to the the cpio archive or a header inside
+ * @len:    Remaining length of the cpio based on data pointer
+ * @offset: When a matching file is found, this is the offset to the
+ *          beginning of the cpio. It can be used to iterate through
+ *          the cpio to find all files inside of a directory path
+ *
+ * @return: struct cpio_data containing the address, length and
+ *          filename (with the directory path cut off) of the found file.
+ */
+
+struct cpio_data __cpuinit find_cpio_data(const char *path, void *data,
+					  size_t len,  long *offset)
+{
+	const size_t cpio_header_len = 8*C_NFIELDS - 2;
+	struct cpio_data cd = { NULL, 0, "" };
+	const char *p, *dptr, *nptr;
+	unsigned int ch[C_NFIELDS], *chp, v;
+	unsigned char c, x;
+	size_t mypathsize = strlen(path);
+	int i, j;
+
+	p = data;
+
+	while (len > cpio_header_len) {
+		if (!*p) {
+			/* All cpio headers need to be 4-byte aligned */
+			p += 4;
+			len -= 4;
+			continue;
+		}
+
+		j = 6;		/* The magic field is only 6 characters */
+		chp = ch;
+		for (i = C_NFIELDS; i; i--) {
+			v = 0;
+			while (j--) {
+				v <<= 4;
+				c = *p++;
+
+				x = c - '0';
+				if (x < 10) {
+					v += x;
+					continue;
+				}
+
+				x = (c | 0x20) - 'a';
+				if (x < 6) {
+					v += x + 10;
+					continue;
+				}
+
+				goto quit; /* Invalid hexadecimal */
+			}
+			*chp++ = v;
+			j = 8;	/* All other fields are 8 characters */
+		}
+
+		if ((ch[C_MAGIC] - 0x070701) > 1)
+			goto quit; /* Invalid magic */
+
+		len -= cpio_header_len;
+
+		dptr = PTR_ALIGN(p + ch[C_NAMESIZE], 4);
+		nptr = PTR_ALIGN(dptr + ch[C_FILESIZE], 4);
+
+		if (nptr > p + len || dptr < p || nptr < dptr)
+			goto quit; /* Buffer overrun */
+
+		if ((ch[C_MODE] & 0170000) == 0100000 &&
+		    ch[C_NAMESIZE] >= mypathsize &&
+		    !memcmp(p, path, mypathsize)) {
+			*offset = (long)nptr - (long)data;
+			if (ch[C_NAMESIZE] - mypathsize >= MAX_CPIO_FILE_NAME) {
+				pr_warn(
+				"File %s exceeding MAX_CPIO_FILE_NAME [%d]\n",
+				p, MAX_CPIO_FILE_NAME);
+			}
+			strlcpy(cd.name, p + mypathsize, MAX_CPIO_FILE_NAME);
+
+			cd.data = (void *)dptr;
+			cd.size = ch[C_FILESIZE];
+			return cd; /* Found it! */
+		}
+		len -= (nptr - p);
+		p = nptr;
+	}
+
+quit:
+	return cd;
+}
-- 
1.7.6.1

[PATCH 2/2] ACPI: Override arbitrary ACPI tables via initrd for debugging

[Thomas Renninger]

Details can be found in:
Documentation/acpi/initrd_table_override.txt

Signed-off-by: Thomas Renninger <trenn@suse.de>
CC: eric.piel@tremplin-utc.net
CC: vojcek@tlen.pl
CC: Lin Ming <ming.m.lin@intel.com>
CC: lenb@kernel.org
CC: robert.moore@intel.com
CC: hpa@zytor.com
CC: yinghai@kernel.org
---
 Documentation/acpi/initrd_table_override.txt |  122 +++++++++++++++
 arch/x86/kernel/setup.c                      |    4 +
 drivers/acpi/Kconfig                         |    9 +
 drivers/acpi/osl.c                           |  203 ++++++++++++++++++++++++--
 include/linux/acpi.h                         |    4 +
 5 files changed, 331 insertions(+), 11 deletions(-)
 create mode 100644 Documentation/acpi/initrd_table_override.txt

diff --git a/Documentation/acpi/initrd_table_override.txt b/Documentation/acpi/initrd_table_override.txt
new file mode 100644
index 0000000..b550831
--- /dev/null
+++ b/Documentation/acpi/initrd_table_override.txt
@@ -0,0 +1,122 @@
+Overriding ACPI tables via initrd
+=================================
+
+1) Introduction (What is this about)
+2) What is this for
+3) How does it work
+4) References (Where to retrieve userspace tools)
+
+1) What is this about
+---------------------
+
+If ACPI_INITRD_TABLE_OVERRIDE compile option is true, it is possible to
+override nearly any ACPI table provided by the BIOS with an instrumented,
+modified one.
+
+For a full list of ACPI tables that can be overridden, take a look at
+the char *table_sigs[MAX_ACPI_SIGNATURE]; definition in drivers/acpi/osl.c
+All ACPI tables iasl (Intel's ACPI compiler and disassembler) knows should
+be overridable, except:
+   - ACPI_SIG_RSDP (has a signature of 6 bytes)
+   - ACPI_SIG_FACS (does not have an ordinary ACPI table header)
+Both could get implemented as well.
+
+
+2) What is this for
+-------------------
+
+Please keep in mind that this is a debug option.
+ACPI tables should not get overridden for productive use.
+If BIOS ACPI tables are overridden the kernel will get tainted with the
+TAINT_OVERRIDDEN_ACPI_TABLE flag.
+Complain to your platform/BIOS vendor if you find a bug which is that sever
+that a workaround is not accepted in the Linus kernel.
+
+Still, it can and should be enabled in any kernel, because:
+  - There is no functional change with not instrumented initrds
+  - It provides a powerful feature to easily debug and test ACPI BIOS table
+    compatibility with the Linux kernel.
+
+Until now it was only possible to override the DSDT by compiling it into
+the kernel. This is a nightmare when trying to work on ACPI related bugs
+and a lot bugs got stuck because of that.
+Even for people with enough kernel knowledge, building a kernel to try out
+things is very time consuming. Also people may have to browse and modify the
+ACPI interpreter code to find a possible BIOS bug. With this feature, people
+can correct the ACPI tables and try out quickly whether this is the root cause
+that needs to get addressed in the kernel.
+
+This could even ease up testing for BIOS providers who could flush their BIOS
+to test, but overriding table via initrd is much easier and quicker.
+For example one could prepare different initrds overriding NUMA tables with
+different affinity settings. Set up a script, let the machine reboot and
+run tests over night and one can get a picture how these settings influence
+the Linux kernel and which values are best.
+
+People can instrument the dynamic ACPI (ASL) code (for example with debug
+statements showing up in syslog when the ACPI code is processed, etc.),
+to better understand BIOS to OS interfaces, to hunt down ACPI BIOS code related
+bugs quickly or to easier develop ACPI based drivers.
+
+Intstrumenting ACPI code in SSDTs is now much easier. Before, one had to copy
+all SSDTs into the DSDT to compile it into the kernel for testing
+(because only DSDT could get overridden). That's what the acpi_no_auto_ssdt
+boot param is for: the BIOS provided SSDTs are ignored and all have to get
+copied into the DSDT, complicated and time consuming.
+
+Much more use cases, depending on which ACPI parts you are working on...
+
+
+3) How does it work
+-------------------
+
+# Extract the machine's ACPI tables:
+cd /tmp
+acpidump >acpidump
+acpixtract -a acpidump
+# Disassemble, modify and recompile them:
+iasl -d *.dat
+# For example add this statement into a _PRT (PCI Routing Table) function
+# of the DSDT:
+Store("HELLO WORLD", debug)
+iasl -sa dsdt.dsl
+# Add the raw ACPI tables to an uncompressed cpio archive.
+# They must be put into a /kernel/firmware/acpi directory inside the
+# cpio archive.
+# The uncompressed cpio archive must be the first.
+# Other, typically compressed cpio archives, must be
+# concatenated on top of the uncompressed one.
+mkdir -p kernel/firmware/acpi
+cp dsdt.aml kernel/firmware/acpi
+# A maximum of: #define ACPI_OVERRIDE_TABLES 10
+# tables are  currently allowed (see osl.c):
+iasl -sa facp.dsl
+iasl -sa ssdt1.dsl
+cp facp.aml kernel/firmware/acpi
+cp ssdt1.aml kernel/firmware/acpi
+# Create the uncompressed cpio archive and concatenate the orginal initrd
+# on top:
+find kernel | cpio -H newc --create > /boot/instrumented_initrd
+cat /boot/initrd >>/boot/instrumented_initrd
+# reboot with increased acpi debug level, e.g. boot params:
+acpi.debug_level=0x2 acpi.debug_layer=0xFFFFFFFF
+# and check your syslog:
+[    1.268089] ACPI: PCI Interrupt Routing Table [\_SB_.PCI0._PRT]
+[    1.272091] [ACPI Debug]  String [0x0B] "HELLO WORLD"
+
+iasl is able to disassemble and recompile quite a lot different,
+also static ACPI tables.
+
+4) Where to retrieve userspace tools
+------------------------------------
+
+iasl and acpixtract are part of Intel's ACPICA project:
+http://acpica.org/
+and should be packaged by distributions (for example in the acpica package
+on SUSE).
+
+acpidump can be found in Len Browns pmtools:
+ftp://kernel.org/pub/linux/kernel/people/lenb/acpi/utils/pmtools/acpidump
+This tool is also part of the acpica package on SUSE.
+Alternatively, used ACPI tables can be retrieved via sysfs in latest kernels:
+/sys/firmware/acpi/tables
diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index f4b9b80..6a91058 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -941,6 +941,10 @@ void __init setup_arch(char **cmdline_p)
 
 	reserve_initrd();
 
+#ifdef CONFIG_ACPI_INITRD_TABLE_OVERRIDE
+	acpi_initrd_override((void *)initrd_start, initrd_end - initrd_start);
+#endif
+
 	reserve_crashkernel();
 
 	vsmp_init();
diff --git a/drivers/acpi/Kconfig b/drivers/acpi/Kconfig
index 8099895..a508f77 100644
--- a/drivers/acpi/Kconfig
+++ b/drivers/acpi/Kconfig
@@ -261,6 +261,15 @@ config ACPI_CUSTOM_DSDT
 	bool
 	default ACPI_CUSTOM_DSDT_FILE != ""
 
+config ACPI_INITRD_TABLE_OVERRIDE
+	bool
+	default y
+	help
+	  This option provides functionality to override arbitrary ACPI tables
+	  via initrd. No functional change if no ACPI tables are passed via
+	  initrd, therefore it's safe to say Y.
+	  See Documentation/acpi/initrd_table_override.txt for details
+
 config ACPI_BLACKLIST_YEAR
 	int "Disable ACPI for systems before Jan 1st this year" if X86_32
 	default 0
diff --git a/drivers/acpi/osl.c b/drivers/acpi/osl.c
index 9eaf708..ba7560f 100644
--- a/drivers/acpi/osl.c
+++ b/drivers/acpi/osl.c
@@ -534,6 +534,137 @@ acpi_os_predefined_override(const struct acpi_predefined_names *init_val,
 	return AE_OK;
 }
 
+#ifdef CONFIG_ACPI_INITRD_TABLE_OVERRIDE
+#include <linux/earlycpio.h>
+#include <linux/memblock.h>
+
+#include <asm/e820.h>
+
+static u64 acpi_tables_addr;
+static int all_tables_size;
+
+/* Copied from acpica/tbutils.c:acpi_tb_checksum() */
+u8 __init acpi_table_checksum(u8 *buffer, u32 length)
+{
+	u8 sum = 0;
+	u8 *end = buffer + length;
+
+	while (buffer < end)
+		sum = (u8) (sum + *(buffer++));
+	return sum;
+}
+
+/* All but ACPI_SIG_RSDP and ACPI_SIG_FACS: */
+static const char * const table_sigs[] = {
+	ACPI_SIG_BERT, ACPI_SIG_CPEP, ACPI_SIG_ECDT, ACPI_SIG_EINJ,
+	ACPI_SIG_ERST, ACPI_SIG_HEST, ACPI_SIG_MADT, ACPI_SIG_MSCT,
+	ACPI_SIG_SBST, ACPI_SIG_SLIT, ACPI_SIG_SRAT, ACPI_SIG_ASF,
+	ACPI_SIG_BOOT, ACPI_SIG_DBGP, ACPI_SIG_DMAR, ACPI_SIG_HPET,
+	ACPI_SIG_IBFT, ACPI_SIG_IVRS, ACPI_SIG_MCFG, ACPI_SIG_MCHI,
+	ACPI_SIG_SLIC, ACPI_SIG_SPCR, ACPI_SIG_SPMI, ACPI_SIG_TCPA,
+	ACPI_SIG_UEFI, ACPI_SIG_WAET, ACPI_SIG_WDAT, ACPI_SIG_WDDT,
+	ACPI_SIG_WDRT, ACPI_SIG_DSDT, ACPI_SIG_FADT, ACPI_SIG_PSDT,
+	ACPI_SIG_RSDT, ACPI_SIG_XSDT, ACPI_SIG_SSDT, NULL };
+
+/* Non-fatal errors: Affected tables/files are ignored */
+#define INVALID_TABLE(x, path, name)					\
+	{ pr_err("ACPI OVERRIDE: " x " [%s%s]\n", path, name); continue; }
+
+#define ACPI_HEADER_SIZE sizeof(struct acpi_table_header)
+
+/* Must not increase 10 or needs code modifcation below */
+#define ACPI_OVERRIDE_TABLES 10
+
+void __init acpi_initrd_override(void *data, size_t size)
+{
+	int sig, no, table_nr = 0, total_offset = 0;
+	long offset = 0;
+	struct acpi_table_header *table;
+	char cpio_path[32] = "kernel/firmware/acpi/";
+	struct cpio_data file;
+	struct cpio_data early_initrd_files[ACPI_OVERRIDE_TABLES];
+	char *p;
+
+	if (data == NULL || size == 0)
+		return;
+
+	for (no = 0; no < ACPI_OVERRIDE_TABLES; no++) {
+		file = find_cpio_data(cpio_path, data, size, &offset);
+		if (!file.data)
+			break;
+
+		data += offset;
+		size -= offset;
+
+		if (file.size < sizeof(struct acpi_table_header))
+			INVALID_TABLE("Table smaller than ACPI header",
+				      cpio_path, file.name);
+
+		table = file.data;
+
+		for (sig = 0; table_sigs[sig]; sig++)
+			if (!memcmp(table->signature, table_sigs[sig], 4))
+				break;
+
+		if (!table_sigs[sig])
+			INVALID_TABLE("Unknown signature",
+				      cpio_path, file.name);
+		if (file.size != table->length)
+			INVALID_TABLE("File length does not match table length",
+				      cpio_path, file.name);
+		if (acpi_table_checksum(file.data, table->length))
+			INVALID_TABLE("Bad table checksum",
+				      cpio_path, file.name);
+
+		pr_info("%4.4s ACPI table found in initrd [%s%s][0x%x]\n",
+			table->signature, cpio_path, file.name, table->length);
+
+		all_tables_size += table->length;
+		early_initrd_files[table_nr].data = file.data;
+		early_initrd_files[table_nr].size = file.size;
+		table_nr++;
+	}
+	if (table_nr == 0)
+		return;
+
+	acpi_tables_addr =
+		memblock_find_in_range(0, max_low_pfn_mapped << PAGE_SHIFT,
+				       all_tables_size, PAGE_SIZE);
+	if (!acpi_tables_addr) {
+		WARN_ON(1);
+		return;
+	}
+	/*
+	 * Only calling e820_add_reserve does not work and the
+	 * tables are invalid (memory got used) later.
+	 * memblock_x86_reserve_range works as expected and the tables
+	 * won't get modified. But it's not enough because ioremap will
+	 * complain later (used by acpi_os_map_memory) that the pages
+	 * that should get mapped are not marked "reserved".
+	 * Both memblock_x86_reserve_range and e820_add_region works fine.
+	 */
+	memblock_reserve(acpi_tables_addr, acpi_tables_addr + all_tables_size);
+	e820_add_region(acpi_tables_addr, all_tables_size, E820_ACPI);
+	update_e820();
+	p = early_ioremap(acpi_tables_addr, all_tables_size);
+
+	for (no = 0; no < table_nr; no++) {
+		memcpy(p + total_offset, early_initrd_files[no].data,
+		       early_initrd_files[no].size);
+		total_offset += early_initrd_files[no].size;
+	}
+	early_iounmap(p, all_tables_size);
+}
+#endif /* CONFIG_ACPI_INITRD_TABLE_OVERRIDE */
+
+static void acpi_table_taint(struct acpi_table_header *table)
+{
+	pr_warn(PREFIX
+		"Override [%4.4s-%8.8s], this is unsafe: tainting kernel\n",
+		table->signature, table->oem_table_id);
+	add_taint(TAINT_OVERRIDDEN_ACPI_TABLE);
+}
+
 acpi_status
 acpi_os_table_override(struct acpi_table_header * existing_table,
 		       struct acpi_table_header ** new_table)
@@ -547,24 +678,74 @@ acpi_os_table_override(struct acpi_table_header * existing_table,
 	if (strncmp(existing_table->signature, "DSDT", 4) == 0)
 		*new_table = (struct acpi_table_header *)AmlCode;
 #endif
-	if (*new_table != NULL) {
-		printk(KERN_WARNING PREFIX "Override [%4.4s-%8.8s], "
-			   "this is unsafe: tainting kernel\n",
-		       existing_table->signature,
-		       existing_table->oem_table_id);
-		add_taint(TAINT_OVERRIDDEN_ACPI_TABLE);
-	}
+	if (*new_table != NULL)
+		acpi_table_taint(existing_table);
 	return AE_OK;
 }
 
 acpi_status
 acpi_os_physical_table_override(struct acpi_table_header *existing_table,
-				acpi_physical_address * new_address,
-				u32 *new_table_length)
+				acpi_physical_address *address,
+				u32 *table_length)
 {
-	return AE_SUPPORT;
-}
+#ifndef CONFIG_ACPI_INITRD_TABLE_OVERRIDE
+	*table_length = 0;
+	*address = 0;
+	return AE_OK;
+#else
+	int table_offset = 0;
+	struct acpi_table_header *table;
+
+	*table_length = 0;
+	*address = 0;
+
+	if (!acpi_tables_addr)
+		return AE_OK;
+
+	do {
+		if (table_offset + ACPI_HEADER_SIZE > all_tables_size) {
+			WARN_ON(1);
+			return AE_OK;
+		}
+
+		table = acpi_os_map_memory(acpi_tables_addr + table_offset,
+					   ACPI_HEADER_SIZE);
 
+		if (table_offset + table->length > all_tables_size) {
+			acpi_os_unmap_memory(table, ACPI_HEADER_SIZE);
+			WARN_ON(1);
+			return AE_OK;
+		}
+
+		table_offset += table->length;
+
+		if (memcmp(existing_table->signature, table->signature, 4)) {
+			acpi_os_unmap_memory(table,
+				     ACPI_HEADER_SIZE);
+			continue;
+		}
+
+		/* Only override tables with matching oem id */
+		if (memcmp(table->oem_table_id, existing_table->oem_table_id,
+			   ACPI_OEM_TABLE_ID_SIZE)) {
+			acpi_os_unmap_memory(table,
+				     ACPI_HEADER_SIZE);
+			continue;
+		}
+
+		table_offset -= table->length;
+		*table_length = table->length;
+		acpi_os_unmap_memory(table, ACPI_HEADER_SIZE);
+		*address = acpi_tables_addr + table_offset;
+		add_taint(TAINT_OVERRIDDEN_ACPI_TABLE);
+		break;
+	} while (table_offset + ACPI_HEADER_SIZE < all_tables_size);
+
+	if (*address != 0)
+		acpi_table_taint(existing_table);
+	return AE_OK;
+#endif
+}
 
 static irqreturn_t acpi_irq(int irq, void *dev_id)
 {
diff --git a/include/linux/acpi.h b/include/linux/acpi.h
index 4f2a762..87e2c9e 100644
--- a/include/linux/acpi.h
+++ b/include/linux/acpi.h
@@ -76,6 +76,10 @@ typedef int (*acpi_table_handler) (struct acpi_table_header *table);
 
 typedef int (*acpi_table_entry_handler) (struct acpi_subtable_header *header, const unsigned long end);
 
+#ifdef CONFIG_ACPI_INITRD_TABLE_OVERRIDE
+void __init acpi_initrd_override(void *data, size_t size);
+#endif
+
 char * __acpi_map_table (unsigned long phys_addr, unsigned long size);
 void __acpi_unmap_table(char *map, unsigned long size);
 int early_acpi_boot_init(void);
-- 
1.7.6.1

Re: [PATCH 2/2] ACPI: Override arbitrary ACPI tables via initrd for debugging

[Yinghai Lu]

On Fri, Sep 21, 2012 at 6:28 AM, Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org> wrote:
> Details can be found in:
> Documentation/acpi/initrd_table_override.txt
>
> Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
> CC: eric.piel-VkQ1JFuSMpfAbQlEx87xDw@public.gmane.org
> CC: vojcek-wYtBgQxc//8@public.gmane.org
> CC: Lin Ming <ming.m.lin-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
> CC: lenb-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org
> CC: robert.moore-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org
> CC: hpa-YMNOUZJC4hwAvxtiuMwx3w@public.gmane.org
> CC: yinghai-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org
> ---
>  Documentation/acpi/initrd_table_override.txt |  122 +++++++++++++++
>  arch/x86/kernel/setup.c                      |    4 +
>  drivers/acpi/Kconfig                         |    9 +
>  drivers/acpi/osl.c                           |  203 ++++++++++++++++++++++++--
>  include/linux/acpi.h                         |    4 +
>  5 files changed, 331 insertions(+), 11 deletions(-)
>  create mode 100644 Documentation/acpi/initrd_table_override.txt
>
> diff --git a/Documentation/acpi/initrd_table_override.txt b/Documentation/acpi/initrd_table_override.txt
> new file mode 100644
> index 0000000..b550831
> --- /dev/null
> +++ b/Documentation/acpi/initrd_table_override.txt
> @@ -0,0 +1,122 @@
> +Overriding ACPI tables via initrd
> +=================================
> +
> +1) Introduction (What is this about)
> +2) What is this for
> +3) How does it work
> +4) References (Where to retrieve userspace tools)
> +
> +1) What is this about
> +---------------------
> +
> +If ACPI_INITRD_TABLE_OVERRIDE compile option is true, it is possible to
> +override nearly any ACPI table provided by the BIOS with an instrumented,
> +modified one.
> +
> +For a full list of ACPI tables that can be overridden, take a look at
> +the char *table_sigs[MAX_ACPI_SIGNATURE]; definition in drivers/acpi/osl.c
> +All ACPI tables iasl (Intel's ACPI compiler and disassembler) knows should
> +be overridable, except:
> +   - ACPI_SIG_RSDP (has a signature of 6 bytes)
> +   - ACPI_SIG_FACS (does not have an ordinary ACPI table header)
> +Both could get implemented as well.
> +
> +
> +2) What is this for
> +-------------------
> +
> +Please keep in mind that this is a debug option.
> +ACPI tables should not get overridden for productive use.
> +If BIOS ACPI tables are overridden the kernel will get tainted with the
> +TAINT_OVERRIDDEN_ACPI_TABLE flag.
> +Complain to your platform/BIOS vendor if you find a bug which is that sever
> +that a workaround is not accepted in the Linus kernel.
> +
> +Still, it can and should be enabled in any kernel, because:
> +  - There is no functional change with not instrumented initrds
> +  - It provides a powerful feature to easily debug and test ACPI BIOS table
> +    compatibility with the Linux kernel.
> +
> +Until now it was only possible to override the DSDT by compiling it into
> +the kernel. This is a nightmare when trying to work on ACPI related bugs
> +and a lot bugs got stuck because of that.
> +Even for people with enough kernel knowledge, building a kernel to try out
> +things is very time consuming. Also people may have to browse and modify the
> +ACPI interpreter code to find a possible BIOS bug. With this feature, people
> +can correct the ACPI tables and try out quickly whether this is the root cause
> +that needs to get addressed in the kernel.
> +
> +This could even ease up testing for BIOS providers who could flush their BIOS
> +to test, but overriding table via initrd is much easier and quicker.
> +For example one could prepare different initrds overriding NUMA tables with
> +different affinity settings. Set up a script, let the machine reboot and
> +run tests over night and one can get a picture how these settings influence
> +the Linux kernel and which values are best.
> +
> +People can instrument the dynamic ACPI (ASL) code (for example with debug
> +statements showing up in syslog when the ACPI code is processed, etc.),
> +to better understand BIOS to OS interfaces, to hunt down ACPI BIOS code related
> +bugs quickly or to easier develop ACPI based drivers.
> +
> +Intstrumenting ACPI code in SSDTs is now much easier. Before, one had to copy
> +all SSDTs into the DSDT to compile it into the kernel for testing
> +(because only DSDT could get overridden). That's what the acpi_no_auto_ssdt
> +boot param is for: the BIOS provided SSDTs are ignored and all have to get
> +copied into the DSDT, complicated and time consuming.
> +
> +Much more use cases, depending on which ACPI parts you are working on...
> +
> +
> +3) How does it work
> +-------------------
> +
> +# Extract the machine's ACPI tables:
> +cd /tmp
> +acpidump >acpidump
> +acpixtract -a acpidump
> +# Disassemble, modify and recompile them:
> +iasl -d *.dat
> +# For example add this statement into a _PRT (PCI Routing Table) function
> +# of the DSDT:
> +Store("HELLO WORLD", debug)
> +iasl -sa dsdt.dsl
> +# Add the raw ACPI tables to an uncompressed cpio archive.
> +# They must be put into a /kernel/firmware/acpi directory inside the
> +# cpio archive.
> +# The uncompressed cpio archive must be the first.
> +# Other, typically compressed cpio archives, must be
> +# concatenated on top of the uncompressed one.
> +mkdir -p kernel/firmware/acpi
> +cp dsdt.aml kernel/firmware/acpi
> +# A maximum of: #define ACPI_OVERRIDE_TABLES 10
> +# tables are  currently allowed (see osl.c):
> +iasl -sa facp.dsl
> +iasl -sa ssdt1.dsl
> +cp facp.aml kernel/firmware/acpi
> +cp ssdt1.aml kernel/firmware/acpi
> +# Create the uncompressed cpio archive and concatenate the orginal initrd
> +# on top:
> +find kernel | cpio -H newc --create > /boot/instrumented_initrd
> +cat /boot/initrd >>/boot/instrumented_initrd
> +# reboot with increased acpi debug level, e.g. boot params:
> +acpi.debug_level=0x2 acpi.debug_layer=0xFFFFFFFF
> +# and check your syslog:
> +[    1.268089] ACPI: PCI Interrupt Routing Table [\_SB_.PCI0._PRT]
> +[    1.272091] [ACPI Debug]  String [0x0B] "HELLO WORLD"
> +
> +iasl is able to disassemble and recompile quite a lot different,
> +also static ACPI tables.
> +
> +4) Where to retrieve userspace tools
> +------------------------------------
> +
> +iasl and acpixtract are part of Intel's ACPICA project:
> +http://acpica.org/
> +and should be packaged by distributions (for example in the acpica package
> +on SUSE).
> +
> +acpidump can be found in Len Browns pmtools:
> +ftp://kernel.org/pub/linux/kernel/people/lenb/acpi/utils/pmtools/acpidump
> +This tool is also part of the acpica package on SUSE.
> +Alternatively, used ACPI tables can be retrieved via sysfs in latest kernels:
> +/sys/firmware/acpi/tables
> diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
> index f4b9b80..6a91058 100644
> --- a/arch/x86/kernel/setup.c
> +++ b/arch/x86/kernel/setup.c
> @@ -941,6 +941,10 @@ void __init setup_arch(char **cmdline_p)
>
>         reserve_initrd();
>
> +#ifdef CONFIG_ACPI_INITRD_TABLE_OVERRIDE
> +       acpi_initrd_override((void *)initrd_start, initrd_end - initrd_start);
> +#endif
> +

could use
#ifdef CONFIG_ACPI_INITRD_TABLE_OVERRIDE
void acpi_initrd_override(void *data, size_t size);
#else
static inline void acpi_initrd_override(void *data, size_t size)
{
}
#endif

in one header file to avoid MACRO in setup.c

>         reserve_crashkernel();
>
>         vsmp_init();
> diff --git a/drivers/acpi/Kconfig b/drivers/acpi/Kconfig
> index 8099895..a508f77 100644
> --- a/drivers/acpi/Kconfig
> +++ b/drivers/acpi/Kconfig
> @@ -261,6 +261,15 @@ config ACPI_CUSTOM_DSDT
>         bool
>         default ACPI_CUSTOM_DSDT_FILE != ""
>
> +config ACPI_INITRD_TABLE_OVERRIDE
> +       bool
> +       default y
> +       help
> +         This option provides functionality to override arbitrary ACPI tables
> +         via initrd. No functional change if no ACPI tables are passed via
> +         initrd, therefore it's safe to say Y.
> +         See Documentation/acpi/initrd_table_override.txt for details
> +
>  config ACPI_BLACKLIST_YEAR
>         int "Disable ACPI for systems before Jan 1st this year" if X86_32
>         default 0
> diff --git a/drivers/acpi/osl.c b/drivers/acpi/osl.c
> index 9eaf708..ba7560f 100644
> --- a/drivers/acpi/osl.c
> +++ b/drivers/acpi/osl.c
> @@ -534,6 +534,137 @@ acpi_os_predefined_override(const struct acpi_predefined_names *init_val,
>         return AE_OK;
>  }
>
> +#ifdef CONFIG_ACPI_INITRD_TABLE_OVERRIDE
> +#include <linux/earlycpio.h>
> +#include <linux/memblock.h>
> +
> +#include <asm/e820.h>



> +
> +static u64 acpi_tables_addr;
> +static int all_tables_size;
> +
> +/* Copied from acpica/tbutils.c:acpi_tb_checksum() */
> +u8 __init acpi_table_checksum(u8 *buffer, u32 length)
> +{
> +       u8 sum = 0;
> +       u8 *end = buffer + length;
> +
> +       while (buffer < end)
> +               sum = (u8) (sum + *(buffer++));
> +       return sum;
> +}
> +
> +/* All but ACPI_SIG_RSDP and ACPI_SIG_FACS: */
> +static const char * const table_sigs[] = {
> +       ACPI_SIG_BERT, ACPI_SIG_CPEP, ACPI_SIG_ECDT, ACPI_SIG_EINJ,
> +       ACPI_SIG_ERST, ACPI_SIG_HEST, ACPI_SIG_MADT, ACPI_SIG_MSCT,
> +       ACPI_SIG_SBST, ACPI_SIG_SLIT, ACPI_SIG_SRAT, ACPI_SIG_ASF,
> +       ACPI_SIG_BOOT, ACPI_SIG_DBGP, ACPI_SIG_DMAR, ACPI_SIG_HPET,
> +       ACPI_SIG_IBFT, ACPI_SIG_IVRS, ACPI_SIG_MCFG, ACPI_SIG_MCHI,
> +       ACPI_SIG_SLIC, ACPI_SIG_SPCR, ACPI_SIG_SPMI, ACPI_SIG_TCPA,
> +       ACPI_SIG_UEFI, ACPI_SIG_WAET, ACPI_SIG_WDAT, ACPI_SIG_WDDT,
> +       ACPI_SIG_WDRT, ACPI_SIG_DSDT, ACPI_SIG_FADT, ACPI_SIG_PSDT,
> +       ACPI_SIG_RSDT, ACPI_SIG_XSDT, ACPI_SIG_SSDT, NULL };
> +
> +/* Non-fatal errors: Affected tables/files are ignored */
> +#define INVALID_TABLE(x, path, name)                                   \
> +       { pr_err("ACPI OVERRIDE: " x " [%s%s]\n", path, name); continue; }
> +
> +#define ACPI_HEADER_SIZE sizeof(struct acpi_table_header)
> +
> +/* Must not increase 10 or needs code modifcation below */
> +#define ACPI_OVERRIDE_TABLES 10
> +
> +void __init acpi_initrd_override(void *data, size_t size)
> +{
> +       int sig, no, table_nr = 0, total_offset = 0;
> +       long offset = 0;
> +       struct acpi_table_header *table;
> +       char cpio_path[32] = "kernel/firmware/acpi/";
> +       struct cpio_data file;
> +       struct cpio_data early_initrd_files[ACPI_OVERRIDE_TABLES];
> +       char *p;
> +
> +       if (data == NULL || size == 0)
> +               return;
> +
> +       for (no = 0; no < ACPI_OVERRIDE_TABLES; no++) {
> +               file = find_cpio_data(cpio_path, data, size, &offset);
> +               if (!file.data)
> +                       break;
> +
> +               data += offset;
> +               size -= offset;
> +
> +               if (file.size < sizeof(struct acpi_table_header))
> +                       INVALID_TABLE("Table smaller than ACPI header",
> +                                     cpio_path, file.name);
> +
> +               table = file.data;
> +
> +               for (sig = 0; table_sigs[sig]; sig++)
> +                       if (!memcmp(table->signature, table_sigs[sig], 4))
> +                               break;
> +
> +               if (!table_sigs[sig])
> +                       INVALID_TABLE("Unknown signature",
> +                                     cpio_path, file.name);
> +               if (file.size != table->length)
> +                       INVALID_TABLE("File length does not match table length",
> +                                     cpio_path, file.name);
> +               if (acpi_table_checksum(file.data, table->length))
> +                       INVALID_TABLE("Bad table checksum",
> +                                     cpio_path, file.name);
> +
> +               pr_info("%4.4s ACPI table found in initrd [%s%s][0x%x]\n",
> +                       table->signature, cpio_path, file.name, table->length);
> +
> +               all_tables_size += table->length;
> +               early_initrd_files[table_nr].data = file.data;
> +               early_initrd_files[table_nr].size = file.size;
> +               table_nr++;
> +       }
> +       if (table_nr == 0)
> +               return;
> +
> +       acpi_tables_addr =
> +               memblock_find_in_range(0, max_low_pfn_mapped << PAGE_SHIFT,
> +                                      all_tables_size, PAGE_SIZE);
> +       if (!acpi_tables_addr) {
> +               WARN_ON(1);
> +               return;
> +       }
> +       /*
> +        * Only calling e820_add_reserve does not work and the
> +        * tables are invalid (memory got used) later.
> +        * memblock_x86_reserve_range works as expected and the tables
> +        * won't get modified. But it's not enough because ioremap will
> +        * complain later (used by acpi_os_map_memory) that the pages
> +        * that should get mapped are not marked "reserved".
> +        * Both memblock_x86_reserve_range and e820_add_region works fine.
> +        */
> +       memblock_reserve(acpi_tables_addr, acpi_tables_addr + all_tables_size);
> +       e820_add_region(acpi_tables_addr, all_tables_size, E820_ACPI);
> +       update_e820();

need to move those arch related to arch/x86

> +       p = early_ioremap(acpi_tables_addr, all_tables_size);
> +
> +       for (no = 0; no < table_nr; no++) {
> +               memcpy(p + total_offset, early_initrd_files[no].data,
> +                      early_initrd_files[no].size);
> +               total_offset += early_initrd_files[no].size;
> +       }

You may use one loop function, and it could take one call back.
callback 1 will get item and size.
callback 2 will do the copy...

so you can remove hard limit of ACPI_OVERRIDE_TABLES.

> +       early_iounmap(p, all_tables_size);
> +}
> +#endif /* CONFIG_ACPI_INITRD_TABLE_OVERRIDE */
> +
> +static void acpi_table_taint(struct acpi_table_header *table)
> +{
> +       pr_warn(PREFIX
> +               "Override [%4.4s-%8.8s], this is unsafe: tainting kernel\n",
> +               table->signature, table->oem_table_id);
> +       add_taint(TAINT_OVERRIDDEN_ACPI_TABLE);
> +}
> +

can you split acpi_table_taint split change to another patch?

>  acpi_status
>  acpi_os_table_override(struct acpi_table_header * existing_table,
>                        struct acpi_table_header ** new_table)
> @@ -547,24 +678,74 @@ acpi_os_table_override(struct acpi_table_header * existing_table,
>         if (strncmp(existing_table->signature, "DSDT", 4) == 0)
>                 *new_table = (struct acpi_table_header *)AmlCode;
>  #endif
> -       if (*new_table != NULL) {
> -               printk(KERN_WARNING PREFIX "Override [%4.4s-%8.8s], "
> -                          "this is unsafe: tainting kernel\n",
> -                      existing_table->signature,
> -                      existing_table->oem_table_id);
> -               add_taint(TAINT_OVERRIDDEN_ACPI_TABLE);
> -       }
> +       if (*new_table != NULL)
> +               acpi_table_taint(existing_table);
>         return AE_OK;
>  }
>
>  acpi_status
>  acpi_os_physical_table_override(struct acpi_table_header *existing_table,
> -                               acpi_physical_address * new_address,
> -                               u32 *new_table_length)
> +                               acpi_physical_address *address,
> +                               u32 *table_length)
>  {
> -       return AE_SUPPORT;
> -}
> +#ifndef CONFIG_ACPI_INITRD_TABLE_OVERRIDE
> +       *table_length = 0;
> +       *address = 0;
> +       return AE_OK;
> +#else

also could hide macro in header file.

> +       int table_offset = 0;
> +       struct acpi_table_header *table;
> +
> +       *table_length = 0;
> +       *address = 0;
> +
> +       if (!acpi_tables_addr)
> +               return AE_OK;
> +
> +       do {
> +               if (table_offset + ACPI_HEADER_SIZE > all_tables_size) {
> +                       WARN_ON(1);
> +                       return AE_OK;
> +               }
> +
> +               table = acpi_os_map_memory(acpi_tables_addr + table_offset,
> +                                          ACPI_HEADER_SIZE);
>
> +               if (table_offset + table->length > all_tables_size) {
> +                       acpi_os_unmap_memory(table, ACPI_HEADER_SIZE);
> +                       WARN_ON(1);
> +                       return AE_OK;
> +               }
> +
> +               table_offset += table->length;
> +
> +               if (memcmp(existing_table->signature, table->signature, 4)) {
> +                       acpi_os_unmap_memory(table,
> +                                    ACPI_HEADER_SIZE);
> +                       continue;
> +               }
> +
> +               /* Only override tables with matching oem id */
> +               if (memcmp(table->oem_table_id, existing_table->oem_table_id,
> +                          ACPI_OEM_TABLE_ID_SIZE)) {
> +                       acpi_os_unmap_memory(table,
> +                                    ACPI_HEADER_SIZE);
> +                       continue;
> +               }
> +
> +               table_offset -= table->length;
> +               *table_length = table->length;
> +               acpi_os_unmap_memory(table, ACPI_HEADER_SIZE);
> +               *address = acpi_tables_addr + table_offset;
> +               add_taint(TAINT_OVERRIDDEN_ACPI_TABLE);
> +               break;
> +       } while (table_offset + ACPI_HEADER_SIZE < all_tables_size);
> +
> +       if (*address != 0)
> +               acpi_table_taint(existing_table);
> +       return AE_OK;
> +#endif
> +}

could be split to another patch too.


>
>  static irqreturn_t acpi_irq(int irq, void *dev_id)
>  {
> diff --git a/include/linux/acpi.h b/include/linux/acpi.h
> index 4f2a762..87e2c9e 100644
> --- a/include/linux/acpi.h
> +++ b/include/linux/acpi.h
> @@ -76,6 +76,10 @@ typedef int (*acpi_table_handler) (struct acpi_table_header *table);
>
>  typedef int (*acpi_table_entry_handler) (struct acpi_subtable_header *header, const unsigned long end);
>
> +#ifdef CONFIG_ACPI_INITRD_TABLE_OVERRIDE
> +void __init acpi_initrd_override(void *data, size_t size);

could drop __init declaration in header file.

> +#endif
> +
>  char * __acpi_map_table (unsigned long phys_addr, unsigned long size);
>  void __acpi_unmap_table(char *map, unsigned long size);
>  int early_acpi_boot_init(void);
> --
> 1.7.6.1
>

Re: [PATCH 2/2] ACPI: Override arbitrary ACPI tables via initrd for debugging

[Thomas Renninger]

...
> > +       /*
> > +        * Only calling e820_add_reserve does not work and the
> > +        * tables are invalid (memory got used) later.
> > +        * memblock_x86_reserve_range works as expected and the tables
> > +        * won't get modified. But it's not enough because ioremap will
> > +        * complain later (used by acpi_os_map_memory) that the pages
> > +        * that should get mapped are not marked "reserved".
> > +        * Both memblock_x86_reserve_range and e820_add_region works fine.
> > +        */
> > +       memblock_reserve(acpi_tables_addr, acpi_tables_addr + all_tables_size);
> > +       e820_add_region(acpi_tables_addr, all_tables_size, E820_ACPI);
> > +       update_e820();
> 
> need to move those arch related to arch/x86

I agree it should get moved out.
I have split this into a separate patch.

> > +       p = early_ioremap(acpi_tables_addr, all_tables_size);
> > +
> > +       for (no = 0; no < table_nr; no++) {
> > +               memcpy(p + total_offset, early_initrd_files[no].data,
> > +                      early_initrd_files[no].size);
> > +               total_offset += early_initrd_files[no].size;
> > +       }
> 
> You may use one loop function, and it could take one call back.
> callback 1 will get item and size.
> callback 2 will do the copy...
> 
> so you can remove hard limit of ACPI_OVERRIDE_TABLES.
I do not fully understand this one.
Currently I have 3 steps:
  1) iterate over all tables and
      - remember address and size of each
      - sum up total size
  2) memblock reserve total size
  3) copy each table into the memblock reserved area

I cannot see how I could get around the limit easily.
Also the restriction is not a big deal, 10 tables is
a lot. It can also be increased without much bad side-effects,
because all xy[ACPI_OVERRIDE_TABLES] arrays are not global.

I'll sent a split up patchset...

> > +       early_iounmap(p, all_tables_size);
> > +}
> > +#endif /* CONFIG_ACPI_INITRD_TABLE_OVERRIDE */
> > +
> > +static void acpi_table_taint(struct acpi_table_header *table)
> > +{
> > +       pr_warn(PREFIX
> > +               "Override [%4.4s-%8.8s], this is unsafe: tainting kernel\n",
> > +               table->signature, table->oem_table_id);
> > +       add_taint(TAINT_OVERRIDDEN_ACPI_TABLE);
> > +}
> > +
> 
> can you split acpi_table_taint split change to another patch?
Yep.

> >  acpi_status
> >  acpi_os_table_override(struct acpi_table_header * existing_table,
> >                        struct acpi_table_header ** new_table)
> > @@ -547,24 +678,74 @@ acpi_os_table_override(struct acpi_table_header * existing_table,
> >         if (strncmp(existing_table->signature, "DSDT", 4) == 0)
> >                 *new_table = (struct acpi_table_header *)AmlCode;
> >  #endif
> > -       if (*new_table != NULL) {
> > -               printk(KERN_WARNING PREFIX "Override [%4.4s-%8.8s], "
> > -                          "this is unsafe: tainting kernel\n",
> > -                      existing_table->signature,
> > -                      existing_table->oem_table_id);
> > -               add_taint(TAINT_OVERRIDDEN_ACPI_TABLE);

One taint too much the one below is enough...

> > -       }
> > +       if (*new_table != NULL)
> > +               acpi_table_taint(existing_table);
> >         return AE_OK;
> >  }
> >
> >  acpi_status
> >  acpi_os_physical_table_override(struct acpi_table_header *existing_table,
> > -                               acpi_physical_address * new_address,
> > -                               u32 *new_table_length)
> > +                               acpi_physical_address *address,
> > +                               u32 *table_length)
> >  {
> > -       return AE_SUPPORT;
> > -}
> > +#ifndef CONFIG_ACPI_INITRD_TABLE_OVERRIDE
> > +       *table_length = 0;
> > +       *address = 0;
> > +       return AE_OK;
> > +#else
> 
> also could hide macro in header file.
No, that's not possible.
This would be acpica, multi OS headers, I doubt they want
to have Linux specific macros in there...
 
I addressed all the rest and will sent out split up patches.

Thanks for your review!

   Thomas

Re: [PATCH 2/2] ACPI: Override arbitrary ACPI tables via initrd for debugging

[Len Brown]

This isn't a NAK, but I'm at best, "like warm" on this.

I'm not convinced it is a good thing to have - enabled by default -
the ability for users to easily over-ride ACPI tables.

I am concerned this will result in users hacking their BIOS, and making themselves
unsupportable by both their HW and SW suppliers.

"But it is just a _debug_ capability", one counters, "we'd never have
to support somebody who actually uses it."

Even if you toss support (and security) out the window, there is a more insidious
problem.  When such a user latches onto a workaround that tickles their
itch, they are satisfied, and they have zero incentive to get
either their BIOS or Linux fixed for the benefit of other users.

Today we have 2 methods to override AML:
ACPI_CUSTOM_METHOD (default n) allows you to scribble on AML
on the current running system.
ACPI_CUSTOM_DSDT (default n) allows you to override the entire DSDT/SSDT,
but requires you to build that DSDT into the custom kernel itself.

Developers running on their own systems are not complaining about these.

But what if you want to debug something on a remote system
with a distro binary kernel, you say?  The user doesn't know how
to build a kernel, and the distro is too busy to do so.
Some distros ship the initrd hack to  address this problem,
even though it has been repeatedly rejected upstream.
But curiously, even larger distros do NOT ship that hack
and somehow they have survived the last decade of Linux/ACPI
deployment without it.  How is that possible?

Yes, convenience sounds like an improvement over inconvenience.
Yes, generality to override any table sounds like a good thing
over the limitation to override just AML tables.
But does that make it a good idea?

specific comments in-line below...


On 09/21/2012 09:28 AM, Thomas Renninger wrote:
> Details can be found in:
> Documentation/acpi/initrd_table_override.txt
> 
> Signed-off-by: Thomas Renninger <trenn@suse.de>
> CC: eric.piel@tremplin-utc.net
> CC: vojcek@tlen.pl
> CC: Lin Ming <ming.m.lin@intel.com>
> CC: lenb@kernel.org
> CC: robert.moore@intel.com
> CC: hpa@zytor.com
> CC: yinghai@kernel.org
> ---
>  Documentation/acpi/initrd_table_override.txt |  122 +++++++++++++++
>  arch/x86/kernel/setup.c                      |    4 +
>  drivers/acpi/Kconfig                         |    9 +
>  drivers/acpi/osl.c                           |  203 ++++++++++++++++++++++++--
>  include/linux/acpi.h                         |    4 +
>  5 files changed, 331 insertions(+), 11 deletions(-)
>  create mode 100644 Documentation/acpi/initrd_table_override.txt
> 
> diff --git a/Documentation/acpi/initrd_table_override.txt b/Documentation/acpi/initrd_table_override.txt
> new file mode 100644
> index 0000000..b550831
> --- /dev/null
> +++ b/Documentation/acpi/initrd_table_override.txt
> @@ -0,0 +1,122 @@
> +Overriding ACPI tables via initrd
> +=================================
> +
> +1) Introduction (What is this about)
> +2) What is this for
> +3) How does it work
> +4) References (Where to retrieve userspace tools)
> +
> +1) What is this about
> +---------------------
> +
> +If ACPI_INITRD_TABLE_OVERRIDE compile option is true, it is possible to

If "the" ACPI_...

> +override nearly any ACPI table provided by the BIOS with an instrumented,
> +modified one.
> +
> +For a full list of ACPI tables that can be overridden, take a look at
> +the char *table_sigs[MAX_ACPI_SIGNATURE]; definition in drivers/acpi/osl.c
> +All ACPI tables iasl (Intel's ACPI compiler and disassembler) knows should
> +be overridable, except:
> +   - ACPI_SIG_RSDP (has a signature of 6 bytes)
> +   - ACPI_SIG_FACS (does not have an ordinary ACPI table header)
> +Both could get implemented as well.
> +
> +
> +2) What is this for
> +-------------------
> +
> +Please keep in mind that this is a debug option.
> +ACPI tables should not get overridden for productive use.
> +If BIOS ACPI tables are overridden the kernel will get tainted with the
> +TAINT_OVERRIDDEN_ACPI_TABLE flag.
> +Complain to your platform/BIOS vendor if you find a bug which is that sever

"bug, which is so severe"

> +that a workaround is not accepted in the Linus kernel.

"Linux kernel"

> +
> +Still, it can and should be enabled in any kernel, because:
> +  - There is no functional change with not instrumented initrds
> +  - It provides a powerful feature to easily debug and test ACPI BIOS table
> +    compatibility with the Linux kernel.
> +
> +Until now it was only possible to override the DSDT by compiling it into
> +the kernel. This is a nightmare when trying to work on ACPI related bugs
> +and a lot bugs got stuck because of that.
> +
> +Even for people with enough kernel knowledge, building a kernel to try out
> +things is very time consuming. Also people may have to browse and modify the
> +ACPI interpreter code to find a possible BIOS bug. With this feature, people
> +can correct the ACPI tables and try out quickly whether this is the root cause
> +that needs to get addressed in the kernel.
> +
> +This could even ease up testing for BIOS providers who could flush their BIOS
> +to test, but overriding table via initrd is much easier and quicker.
> +For example one could prepare different initrds overriding NUMA tables with
> +different affinity settings. Set up a script, let the machine reboot and
> +run tests over night and one can get a picture how these settings influence
> +the Linux kernel and which values are best.
> +
> +People can instrument the dynamic ACPI (ASL) code (for example with debug
> +statements showing up in syslog when the ACPI code is processed, etc.),
> +to better understand BIOS to OS interfaces, to hunt down ACPI BIOS code related
> +bugs quickly or to easier develop ACPI based drivers.
> +
> +Intstrumenting ACPI code in SSDTs is now much easier. Before, one had to copy
> +all SSDTs into the DSDT to compile it into the kernel for testing
> +(because only DSDT could get overridden). That's what the acpi_no_auto_ssdt
> +boot param is for: the BIOS provided SSDTs are ignored and all have to get
> +copied into the DSDT, complicated and time consuming.
> +
> +Much more use cases, depending on which ACPI parts you are working on...

The above 6 paragraphs are appropriate for the mailing list,
but don't belong in the README file.  Presumably this capability
is present in the kernel being run, so you don't have to sell it
over alternative debug methods.

> +
> +
> +3) How does it work
> +-------------------
> +
> +# Extract the machine's ACPI tables:
> +cd /tmp
> +acpidump >acpidump
> +acpixtract -a acpidump
> +# Disassemble, modify and recompile them:
> +iasl -d *.dat
> +# For example add this statement into a _PRT (PCI Routing Table) function
> +# of the DSDT:
> +Store("HELLO WORLD", debug)
> +iasl -sa dsdt.dsl
> +# Add the raw ACPI tables to an uncompressed cpio archive.
> +# They must be put into a /kernel/firmware/acpi directory inside the
> +# cpio archive.
> +# The uncompressed cpio archive must be the first.
> +# Other, typically compressed cpio archives, must be
> +# concatenated on top of the uncompressed one.
> +mkdir -p kernel/firmware/acpi
> +cp dsdt.aml kernel/firmware/acpi
> +# A maximum of: #define ACPI_OVERRIDE_TABLES 10
> +# tables are  currently allowed (see osl.c):
> +iasl -sa facp.dsl
> +iasl -sa ssdt1.dsl
> +cp facp.aml kernel/firmware/acpi
> +cp ssdt1.aml kernel/firmware/acpi
> +# Create the uncompressed cpio archive and concatenate the orginal initrd

"original"

> +# on top:
> +find kernel | cpio -H newc --create > /boot/instrumented_initrd
> +cat /boot/initrd >>/boot/instrumented_initrd
> +# reboot with increased acpi debug level, e.g. boot params:
> +acpi.debug_level=0x2 acpi.debug_layer=0xFFFFFFFF
> +# and check your syslog:
> +[    1.268089] ACPI: PCI Interrupt Routing Table [\_SB_.PCI0._PRT]
> +[    1.272091] [ACPI Debug]  String [0x0B] "HELLO WORLD"
> +
> +iasl is able to disassemble and recompile quite a lot different,
> +also static ACPI tables.
> +
> +4) Where to retrieve userspace tools
> +------------------------------------
> +
> +iasl and acpixtract are part of Intel's ACPICA project:
> +http://acpica.org/
> +and should be packaged by distributions (for example in the acpica package
> +on SUSE).
> +
> +acpidump can be found in Len Browns pmtools:
> +ftp://kernel.org/pub/linux/kernel/people/lenb/acpi/utils/pmtools/acpidump
> +This tool is also part of the acpica package on SUSE.
> +Alternatively, used ACPI tables can be retrieved via sysfs in latest kernels:
> +/sys/firmware/acpi/tables
> diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
> index f4b9b80..6a91058 100644
> --- a/arch/x86/kernel/setup.c
> +++ b/arch/x86/kernel/setup.c
> @@ -941,6 +941,10 @@ void __init setup_arch(char **cmdline_p)
>  
>  	reserve_initrd();
>  
> +#ifdef CONFIG_ACPI_INITRD_TABLE_OVERRIDE
> +	acpi_initrd_override((void *)initrd_start, initrd_end - initrd_start);
> +#endif
> +
>  	reserve_crashkernel();
>  
>  	vsmp_init();
> diff --git a/drivers/acpi/Kconfig b/drivers/acpi/Kconfig
> index 8099895..a508f77 100644
> --- a/drivers/acpi/Kconfig
> +++ b/drivers/acpi/Kconfig
> @@ -261,6 +261,15 @@ config ACPI_CUSTOM_DSDT
>  	bool
>  	default ACPI_CUSTOM_DSDT_FILE != ""
>  
> +config ACPI_INITRD_TABLE_OVERRIDE
> +	bool
> +	default y
> +	help
> +	  This option provides functionality to override arbitrary ACPI tables
> +	  via initrd. No functional change if no ACPI tables are passed via
> +	  initrd, therefore it's safe to say Y.
> +	  See Documentation/acpi/initrd_table_override.txt for details
> +
>  config ACPI_BLACKLIST_YEAR
>  	int "Disable ACPI for systems before Jan 1st this year" if X86_32
>  	default 0
> diff --git a/drivers/acpi/osl.c b/drivers/acpi/osl.c
> index 9eaf708..ba7560f 100644
> --- a/drivers/acpi/osl.c
> +++ b/drivers/acpi/osl.c
> @@ -534,6 +534,137 @@ acpi_os_predefined_override(const struct acpi_predefined_names *init_val,
>  	return AE_OK;
>  }
>  
> +#ifdef CONFIG_ACPI_INITRD_TABLE_OVERRIDE
> +#include <linux/earlycpio.h>
> +#include <linux/memblock.h>
> +
> +#include <asm/e820.h>
> +
> +static u64 acpi_tables_addr;
> +static int all_tables_size;
> +
> +/* Copied from acpica/tbutils.c:acpi_tb_checksum() */
> +u8 __init acpi_table_checksum(u8 *buffer, u32 length)
> +{
> +	u8 sum = 0;
> +	u8 *end = buffer + length;
> +
> +	while (buffer < end)
> +		sum = (u8) (sum + *(buffer++));
> +	return sum;
> +}
> +
> +/* All but ACPI_SIG_RSDP and ACPI_SIG_FACS: */
> +static const char * const table_sigs[] = {
> +	ACPI_SIG_BERT, ACPI_SIG_CPEP, ACPI_SIG_ECDT, ACPI_SIG_EINJ,
> +	ACPI_SIG_ERST, ACPI_SIG_HEST, ACPI_SIG_MADT, ACPI_SIG_MSCT,
> +	ACPI_SIG_SBST, ACPI_SIG_SLIT, ACPI_SIG_SRAT, ACPI_SIG_ASF,
> +	ACPI_SIG_BOOT, ACPI_SIG_DBGP, ACPI_SIG_DMAR, ACPI_SIG_HPET,
> +	ACPI_SIG_IBFT, ACPI_SIG_IVRS, ACPI_SIG_MCFG, ACPI_SIG_MCHI,
> +	ACPI_SIG_SLIC, ACPI_SIG_SPCR, ACPI_SIG_SPMI, ACPI_SIG_TCPA,
> +	ACPI_SIG_UEFI, ACPI_SIG_WAET, ACPI_SIG_WDAT, ACPI_SIG_WDDT,
> +	ACPI_SIG_WDRT, ACPI_SIG_DSDT, ACPI_SIG_FADT, ACPI_SIG_PSDT,
> +	ACPI_SIG_RSDT, ACPI_SIG_XSDT, ACPI_SIG_SSDT, NULL };
> +
> +/* Non-fatal errors: Affected tables/files are ignored */
> +#define INVALID_TABLE(x, path, name)					\
> +	{ pr_err("ACPI OVERRIDE: " x " [%s%s]\n", path, name); continue; }
> +
> +#define ACPI_HEADER_SIZE sizeof(struct acpi_table_header)
> +
> +/* Must not increase 10 or needs code modifcation below */

"modification"

Re: [PATCH 2/2] ACPI: Override arbitrary ACPI tables via initrd for debugging

[Thomas Renninger]

Hi Len,

On Saturday 22 September 2012 17:16:47 Len Brown wrote:
> This isn't a NAK, but I'm at best, "like warm" on this.
We had this discussion already...
And you already had similar patches Signed-off and put in your queue
for Linus in 2008:
http://www.mail-archive.com/linux-acpi-u79uwXL29TY76Z2rM5mHXA@public.gmane.org/msg11907.html

Iirc they have not been accepted because populate_rootfs was too
late or the way the tables have been passed has been (correctly)
considered a bad approach.

This is the clean re-write.
 
> I'm not convinced it is a good thing to have - enabled by default -
> the ability for users to easily over-ride ACPI tables.
> 
> I am concerned this will result in users hacking their BIOS, and making themselves
> unsupportable by both their HW and SW suppliers.
They can do that already by overriding via re-compiling the kernel.
Making it as hard as possible, also means making debugging ACPI problems
on Linux as hard as possible.
There are quite some bugs that got stuck, because people could not compile
a kernel.
And quite some people who have the knowledge and had overridden their DSDTs
via recompiling the kernel wasted some hours/days (compiling a kernel on the
private laptop at home takes time) and they could have invested their time
better.

> "But it is just a _debug_ capability", one counters, "we'd never have
> to support somebody who actually uses it."
Both is right and these are the answers to most of your concerns.
 
> Even if you toss support (and security) out the window, there is a more insidious
> problem.  When such a user latches onto a workaround that tickles their
> itch, they are satisfied, and they have zero incentive to get
> either their BIOS or Linux fixed for the benefit of other users.
I expect it exactly the other way round:
It has been made clear to report issues.
Currently these guys are more or less on their own, not finding a workaround
or any solution at all.
Now with some help they will find workarounds, post them on the list and
developers will already get a concrete idea of what is going wrong.
 
> Today we have 2 methods to override AML:
> ACPI_CUSTOM_METHOD (default n) allows you to scribble on AML
> on the current running system.
> ACPI_CUSTOM_DSDT (default n) allows you to override the entire DSDT/SSDT,
> but requires you to build that DSDT into the custom kernel itself.
I know, what is so bad to make developing this stuff on Linux "even easier"?

> Developers running on their own systems are not complaining about these.
> 
> But what if you want to debug something on a remote system
> with a distro binary kernel, you say?  The user doesn't know how
> to build a kernel, and the distro is too busy to do so.
> Some distros ship the initrd hack to  address this problem,
> even though it has been repeatedly rejected upstream.
> But curiously, even larger distros do NOT ship that hack
> and somehow they have survived the last decade of Linux/ACPI
> deployment without it.  How is that possible?
>  
> Yes, convenience sounds like an improvement over inconvenience.
> Yes, generality to override any table sounds like a good thing
> over the limitation to override just AML tables.
> But does that make it a good idea?

I do not want to re-discuss the topic all over again if possible,
here my major points:
  - Even on "Linux supported by vendor" systems it can be convenient
    for example for a distribution to be able to prove a bug to be
    a BIOS and not a kernel bug, by simply booting the fixed BIOS
    and verify the issue to be fixed.
  - Quite some BIOS tables, even from vendors supporting Linux, have
    dozens of warnings and errors in their ACPI tables.
    If it is easier to develop and debug ACPI code on Linux, there
    might be more vendors doing that and providing more robust BIOSes.
  - Unfortunately a lot or most laptop/desktop vendors still do not
    care that much about Linux. Quite some systems are sold and nobody
    even tried to boot Linux on them. The BIOS did only get verified
    on Windows. On such systems it's essential that people can find
    out as easy as possible how and why their BIOS behaves like the
    way it does and then contribute that information to the list to get
    things solved.
  - ...
    
> specific comments in-line below...

Thanks a lot for going through it!
I'll address yours (and Yinghai's) findings and try to get an update
sent out on Monday.

    Thomas

Re: [PATCH 2/2] ACPI: Override arbitrary ACPI tables via initrd for debugging

[Len Brown]

> +config ACPI_INITRD_TABLE_OVERRIDE
> +       bool
> +       default y

Do distros in addition to SuSE concur they want to ship this way?

The last time we tried to make debugging easier we added
ACPI_CUSTOM_METHOD, which allowed root to over-ride an AML method
on a running system.  Distro security-minded people were not amused.

thanks,
-Len Brown, Intel Open Source Technology Center

ps I noticed your reference to acpidump in the README.
That reminded me to push it to the kernel source tree.
Its new home will be tools/power/acpi/

Re: [PATCH 2/2] ACPI: Override arbitrary ACPI tables via initrd for debugging

[Thomas Renninger]

On Sunday 23 September 2012 06:25:40 Len Brown wrote:
> > +config ACPI_INITRD_TABLE_OVERRIDE
> > +       bool
> > +       default y
> 
> Do distros in addition to SuSE concur they want to ship this way?
Whether distros ship this in their enterprise, community or just in
a -debug kernel flavor is up to them.
I cannot see why this cannot be enabled by default on all.
That is what the TAINT flag is for...
 
> The last time we tried to make debugging easier we added
> ACPI_CUSTOM_METHOD, which allowed root to over-ride an AML method
> on a running system.  Distro security-minded people were not amused.
Yep and therefore you have to remove this one from the tools for
ACPI debugging you listed.
The issue is/was, that root can inject code at runtime which is then
executed in kernel environment.
Afaik there are "security" provisions or say setups, which do
hide modprobe/insmod and do not allow root to load any kernel drivers 
or similar.
If one can write the kernel or initrd which gets booted, I guess there
are not much security restrictions anymore you could put on this user...
But thanks for the pointer, I'll go and double check with some
security guys.

> thanks,
> -Len Brown, Intel Open Source Technology Center
> 
> ps I noticed your reference to acpidump in the README.
> That reminded me to push it to the kernel source tree.
> Its new home will be tools/power/acpi
This is the one which I tried to/did adjust to acpica headers?
This sounds like a very good idea. I'll adjust the docs.

pss: Can this tool live there as well:
ftp://ftp.suse.com/pub/people/trenn/sources/ec/ec_access.c
It's the userspace tool for examining EC values (and changes) via
ec_sys debug driver and a corresponding /sys/kernel/debug/.. file.
It's more ore less doing the same what the old thinkpad_acpi driver
could, but offers this to all machines with an EC device.

    Thomas

Re: [PATCH 2/2] ACPI: Override arbitrary ACPI tables via initrd for debugging

[Alan Cox]

> The issue is/was, that root can inject code at runtime which is then
> executed in kernel environment.

Yes there are lots of other ways to do this too. The constraint we use
for it is CAP_SYS_RAWIO. With that capability you can totally do raw
hardware access and the like so requiring it for runtime ACPI updating
and execution is consistent with the security model.

> Afaik there are "security" provisions or say setups, which do
> hide modprobe/insmod and do not allow root to load any kernel drivers 
> or similar.

To do this you have to revoke CAP_SYS_RAWIO.

Alan

[PATCH] ACPI: Only allow users with CAP_SYS_RAWIO rights to overwrite ACPI funcs at runtime

[Thomas Renninger]

ACPI: Only allow users with CAP_SYS_RAWIO rights to overwrite ACPI funcs at runtime

and tab align file_operations callbacks.

Signed-off-by: Thomas Renninger <trenn@suse.de>
---
 drivers/acpi/custom_method.c |   10 ++++++++--
 1 files changed, 8 insertions(+), 2 deletions(-)

diff --git a/drivers/acpi/custom_method.c b/drivers/acpi/custom_method.c
index 5d42c24..9a0294d 100644
--- a/drivers/acpi/custom_method.c
+++ b/drivers/acpi/custom_method.c
@@ -72,9 +72,15 @@ static ssize_t cm_write(struct file *file, const char __user * user_buf,
 	return count;
 }
 
+static int open_check_acc(struct inode *inode, struct file *filp)
+{
+	return capable(CAP_SYS_RAWIO) ? 0 : -EPERM;
+}
+
 static const struct file_operations cm_fops = {
-	.write = cm_write,
-	.llseek = default_llseek,
+	.write	= cm_write,
+	.llseek	= default_llseek,
+	.open	= open_check_acc,
 };
 
 static int __init acpi_custom_method_init(void)

Re: [PATCH 2/2] ACPI: Override arbitrary ACPI tables via initrd for debugging

[Matthew Garrett]

On Sun, Sep 23, 2012 at 12:25:40AM -0400, Len Brown wrote:
> > +config ACPI_INITRD_TABLE_OVERRIDE
> > +       bool
> > +       default y
> 
> Do distros in addition to SuSE concur they want to ship this way?

We certainly don't.

-- 
Matthew Garrett | mjg59@srcf.ucam.org

Re: [PATCH 2/2] ACPI: Override arbitrary ACPI tables via initrd for debugging

[H. Peter Anvin]

On 09/22/2012 08:16 AM, Len Brown wrote:
> This isn't a NAK, but I'm at best, "like warm" on this.
> 
> I'm not convinced it is a good thing to have - enabled by default -
> the ability for users to easily over-ride ACPI tables.

For the record: I'm fine with the implementation from a technical
standpoint now; I'm completely agnostic on the policy.  Since Len is the
ACPI maintainer this is up to him.

	-hpa