Commit 24bbb1fa authored by Michael Holzheu's avatar Michael Holzheu Committed by Linus Torvalds

[PATCH] s390_hypfs filesystem

On zSeries machines there exists an interface which allows the operating
system to retrieve LPAR hypervisor accounting data.  For example, it is
possible to get usage data for physical and virtual cpus.  In order to
provide this information to user space programs, I implemented a new
virtual Linux file system named 's390_hypfs' using the Linux 2.6 libfs
framework.  The name 's390_hypfs' stands for 'S390 Hypervisor Filesystem'.
All the accounting information is put into different virtual files which
can be accessed from user space.  All data is represented as ASCII strings.

When the file system is mounted the accounting information is retrieved and
a file system tree is created with the attribute files containing the cpu
information.  The content of the files remains unchanged until a new update
is made.  An update can be triggered from user space through writing
'something' into a special purpose update file.

We create the following directory structure:

<mount-point>/
        update
        cpus/
                <cpu-id>
                        type
                        mgmtime
                <cpu-id>
                        ...
        hyp/
                type
        systems/
                <lpar-name>
                        cpus/
                                <cpu-id>
                                        type
                                        mgmtime
                                        cputime
                                        onlinetime
                                <cpu-id>
                                        ...
                <lpar-name>
                        cpus/
                                ...

- update: File to trigger update
- cpus/: Directory for all physical cpus
- cpus/<cpu-id>/: Directory for one physical cpu.
- cpus/<cpu-id>/type: Type name of physical zSeries cpu.
- cpus/<cpu-id>/mgmtime: Physical-LPAR-management time in microseconds.
- hyp/: Directory for hypervisor information
- hyp/type: Typ of hypervisor (currently only 'LPAR Hypervisor')
- systems/: Directory for all LPARs
- systems/<lpar-name>/: Directory for one LPAR.
- systems/<lpar-name>/cpus/<cpu-id>/: Directory for the virtual cpus
- systems/<lpar-name>/cpus/<cpu-id>/type: Typ of cpu.
- systems/<lpar-name>/cpus/<cpu-id>/mgmtime:
Accumulated number of microseconds during which a physical
CPU was assigned to the logical cpu and the cpu time was
consumed by the hypervisor and was not provided to
the LPAR (LPAR overhead).

- systems/<lpar-name>/cpus/<cpu-id>/cputime:
Accumulated number of microseconds during which a physical CPU
was assigned to the logical cpu and the cpu time was consumed
by the LPAR.

- systems/<lpar-name>/cpus/<cpu-id>/onlinetime:
Accumulated number of microseconds during which the logical CPU
has been online.

As mount point for the filesystem /sys/hypervisor/s390 is created.

The update process is triggered when writing 'something' into the
'update' file at the top level hypfs directory. You can do this e.g.
with 'echo 1 > update'. During the update the whole directory structure
is deleted and built up again.

Cc: Pekka Enberg <penberg@cs.helsinki.fi>
Cc: Ingo Oeser <ioe-lkml@rameria.de>
Cc: Joern Engel <joern@wohnheim.fh-wedel.de>
Acked-by: default avatarMartin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: default avatarMichael Holzheu <holzheu@de.ibm.com>
Signed-off-by: default avatarAndrew Morton <akpm@osdl.org>
Signed-off-by: default avatarLinus Torvalds <torvalds@osdl.org>
parent b9e122c8
......@@ -446,6 +446,14 @@ config NO_IDLE_HZ_INIT
The HZ timer is switched off in idle by default. That means the
HZ timer is already disabled at boot time.
config S390_HYPFS_FS
bool "s390 hypervisor file system support"
select SYS_HYPERVISOR
default y
help
This is a virtual file system intended to provide accounting
information in an s390 hypervisor environment.
config KEXEC
bool "kexec system call (EXPERIMENTAL)"
depends on EXPERIMENTAL
......
......@@ -76,7 +76,7 @@ LDFLAGS_vmlinux := -e start
head-y := arch/$(ARCH)/kernel/head.o arch/$(ARCH)/kernel/init_task.o
core-y += arch/$(ARCH)/mm/ arch/$(ARCH)/kernel/ arch/$(ARCH)/crypto/ \
arch/$(ARCH)/appldata/
arch/$(ARCH)/appldata/ arch/$(ARCH)/hypfs/
libs-y += arch/$(ARCH)/lib/
drivers-y += drivers/s390/
drivers-$(CONFIG_MATHEMU) += arch/$(ARCH)/math-emu/
......
#
# Makefile for the linux hypfs filesystem routines.
#
obj-$(CONFIG_S390_HYPFS_FS) += s390_hypfs.o
s390_hypfs-objs := inode.o hypfs_diag.o
/*
* fs/hypfs/hypfs.h
* Hypervisor filesystem for Linux on s390.
*
* Copyright (C) IBM Corp. 2006
* Author(s): Michael Holzheu <holzheu@de.ibm.com>
*/
#ifndef _HYPFS_H_
#define _HYPFS_H_
#include <linux/fs.h>
#include <linux/types.h>
#define REG_FILE_MODE 0440
#define UPDATE_FILE_MODE 0220
#define DIR_MODE 0550
extern struct dentry *hypfs_mkdir(struct super_block *sb, struct dentry *parent,
const char *name);
extern struct dentry *hypfs_create_u64(struct super_block *sb,
struct dentry *dir, const char *name,
__u64 value);
extern struct dentry *hypfs_create_str(struct super_block *sb,
struct dentry *dir, const char *name,
char *string);
#endif /* _HYPFS_H_ */
/*
* fs/hypfs/hypfs_diag.c
* Hypervisor filesystem for Linux on s390. Diag 204 and 224
* implementation.
*
* Copyright (C) IBM Corp. 2006
* Author(s): Michael Holzheu <holzheu@de.ibm.com>
*/
#include <linux/types.h>
#include <linux/errno.h>
#include <linux/string.h>
#include <linux/vmalloc.h>
#include <asm/ebcdic.h>
#include "hypfs.h"
#define LPAR_NAME_LEN 8 /* lpar name len in diag 204 data */
#define CPU_NAME_LEN 16 /* type name len of cpus in diag224 name table */
#define TMP_SIZE 64 /* size of temporary buffers */
/* diag 204 subcodes */
enum diag204_sc {
SUBC_STIB4 = 4,
SUBC_RSI = 5,
SUBC_STIB6 = 6,
SUBC_STIB7 = 7
};
/* The two available diag 204 data formats */
enum diag204_format {
INFO_SIMPLE = 0,
INFO_EXT = 0x00010000
};
/* bit is set in flags, when physical cpu info is included in diag 204 data */
#define LPAR_PHYS_FLG 0x80
static char *diag224_cpu_names; /* diag 224 name table */
static enum diag204_sc diag204_store_sc; /* used subcode for store */
static enum diag204_format diag204_info_type; /* used diag 204 data format */
static void *diag204_buf; /* 4K aligned buffer for diag204 data */
static void *diag204_buf_vmalloc; /* vmalloc pointer for diag204 data */
static int diag204_buf_pages; /* number of pages for diag204 data */
/*
* DIAG 204 data structures and member access functions.
*
* Since we have two different diag 204 data formats for old and new s390
* machines, we do not access the structs directly, but use getter functions for
* each struct member instead. This should make the code more readable.
*/
/* Time information block */
struct info_blk_hdr {
__u8 npar;
__u8 flags;
__u16 tslice;
__u16 phys_cpus;
__u16 this_part;
__u64 curtod;
} __attribute__ ((packed));
struct x_info_blk_hdr {
__u8 npar;
__u8 flags;
__u16 tslice;
__u16 phys_cpus;
__u16 this_part;
__u64 curtod1;
__u64 curtod2;
char reserved[40];
} __attribute__ ((packed));
static inline int info_blk_hdr__size(enum diag204_format type)
{
if (type == INFO_SIMPLE)
return sizeof(struct info_blk_hdr);
else /* INFO_EXT */
return sizeof(struct x_info_blk_hdr);
}
static inline __u8 info_blk_hdr__npar(enum diag204_format type, void *hdr)
{
if (type == INFO_SIMPLE)
return ((struct info_blk_hdr *)hdr)->npar;
else /* INFO_EXT */
return ((struct x_info_blk_hdr *)hdr)->npar;
}
static inline __u8 info_blk_hdr__flags(enum diag204_format type, void *hdr)
{
if (type == INFO_SIMPLE)
return ((struct info_blk_hdr *)hdr)->flags;
else /* INFO_EXT */
return ((struct x_info_blk_hdr *)hdr)->flags;
}
static inline __u16 info_blk_hdr__pcpus(enum diag204_format type, void *hdr)
{
if (type == INFO_SIMPLE)
return ((struct info_blk_hdr *)hdr)->phys_cpus;
else /* INFO_EXT */
return ((struct x_info_blk_hdr *)hdr)->phys_cpus;
}
/* Partition header */
struct part_hdr {
__u8 pn;
__u8 cpus;
char reserved[6];
char part_name[LPAR_NAME_LEN];
} __attribute__ ((packed));
struct x_part_hdr {
__u8 pn;
__u8 cpus;
__u8 rcpus;
__u8 pflag;
__u32 mlu;
char part_name[LPAR_NAME_LEN];
char lpc_name[8];
char os_name[8];
__u64 online_cs;
__u64 online_es;
__u8 upid;
char reserved1[3];
__u32 group_mlu;
char group_name[8];
char reserved2[32];
} __attribute__ ((packed));
static inline int part_hdr__size(enum diag204_format type)
{
if (type == INFO_SIMPLE)
return sizeof(struct part_hdr);
else /* INFO_EXT */
return sizeof(struct x_part_hdr);
}
static inline __u8 part_hdr__rcpus(enum diag204_format type, void *hdr)
{
if (type == INFO_SIMPLE)
return ((struct part_hdr *)hdr)->cpus;
else /* INFO_EXT */
return ((struct x_part_hdr *)hdr)->rcpus;
}
static inline void part_hdr__part_name(enum diag204_format type, void *hdr,
char *name)
{
if (type == INFO_SIMPLE)
memcpy(name, ((struct part_hdr *)hdr)->part_name,
LPAR_NAME_LEN);
else /* INFO_EXT */
memcpy(name, ((struct x_part_hdr *)hdr)->part_name,
LPAR_NAME_LEN);
EBCASC(name, LPAR_NAME_LEN);
name[LPAR_NAME_LEN] = 0;
strstrip(name);
}
struct cpu_info {
__u16 cpu_addr;
char reserved1[2];
__u8 ctidx;
__u8 cflag;
__u16 weight;
__u64 acc_time;
__u64 lp_time;
} __attribute__ ((packed));
struct x_cpu_info {
__u16 cpu_addr;
char reserved1[2];
__u8 ctidx;
__u8 cflag;
__u16 weight;
__u64 acc_time;
__u64 lp_time;
__u16 min_weight;
__u16 cur_weight;
__u16 max_weight;
char reseved2[2];
__u64 online_time;
__u64 wait_time;
__u32 pma_weight;
__u32 polar_weight;
char reserved3[40];
} __attribute__ ((packed));
/* CPU info block */
static inline int cpu_info__size(enum diag204_format type)
{
if (type == INFO_SIMPLE)
return sizeof(struct cpu_info);
else /* INFO_EXT */
return sizeof(struct x_cpu_info);
}
static inline __u8 cpu_info__ctidx(enum diag204_format type, void *hdr)
{
if (type == INFO_SIMPLE)
return ((struct cpu_info *)hdr)->ctidx;
else /* INFO_EXT */
return ((struct x_cpu_info *)hdr)->ctidx;
}
static inline __u16 cpu_info__cpu_addr(enum diag204_format type, void *hdr)
{
if (type == INFO_SIMPLE)
return ((struct cpu_info *)hdr)->cpu_addr;
else /* INFO_EXT */
return ((struct x_cpu_info *)hdr)->cpu_addr;
}
static inline __u64 cpu_info__acc_time(enum diag204_format type, void *hdr)
{
if (type == INFO_SIMPLE)
return ((struct cpu_info *)hdr)->acc_time;
else /* INFO_EXT */
return ((struct x_cpu_info *)hdr)->acc_time;
}
static inline __u64 cpu_info__lp_time(enum diag204_format type, void *hdr)
{
if (type == INFO_SIMPLE)
return ((struct cpu_info *)hdr)->lp_time;
else /* INFO_EXT */
return ((struct x_cpu_info *)hdr)->lp_time;
}
static inline __u64 cpu_info__online_time(enum diag204_format type, void *hdr)
{
if (type == INFO_SIMPLE)
return 0; /* online_time not available in simple info */
else /* INFO_EXT */
return ((struct x_cpu_info *)hdr)->online_time;
}
/* Physical header */
struct phys_hdr {
char reserved1[1];
__u8 cpus;
char reserved2[6];
char mgm_name[8];
} __attribute__ ((packed));
struct x_phys_hdr {
char reserved1[1];
__u8 cpus;
char reserved2[6];
char mgm_name[8];
char reserved3[80];
} __attribute__ ((packed));
static inline int phys_hdr__size(enum diag204_format type)
{
if (type == INFO_SIMPLE)
return sizeof(struct phys_hdr);
else /* INFO_EXT */
return sizeof(struct x_phys_hdr);
}
static inline __u8 phys_hdr__cpus(enum diag204_format type, void *hdr)
{
if (type == INFO_SIMPLE)
return ((struct phys_hdr *)hdr)->cpus;
else /* INFO_EXT */
return ((struct x_phys_hdr *)hdr)->cpus;
}
/* Physical CPU info block */
struct phys_cpu {
__u16 cpu_addr;
char reserved1[2];
__u8 ctidx;
char reserved2[3];
__u64 mgm_time;
char reserved3[8];
} __attribute__ ((packed));
struct x_phys_cpu {
__u16 cpu_addr;
char reserved1[2];
__u8 ctidx;
char reserved2[3];
__u64 mgm_time;
char reserved3[80];
} __attribute__ ((packed));
static inline int phys_cpu__size(enum diag204_format type)
{
if (type == INFO_SIMPLE)
return sizeof(struct phys_cpu);
else /* INFO_EXT */
return sizeof(struct x_phys_cpu);
}
static inline __u16 phys_cpu__cpu_addr(enum diag204_format type, void *hdr)
{
if (type == INFO_SIMPLE)
return ((struct phys_cpu *)hdr)->cpu_addr;
else /* INFO_EXT */
return ((struct x_phys_cpu *)hdr)->cpu_addr;
}
static inline __u64 phys_cpu__mgm_time(enum diag204_format type, void *hdr)
{
if (type == INFO_SIMPLE)
return ((struct phys_cpu *)hdr)->mgm_time;
else /* INFO_EXT */
return ((struct x_phys_cpu *)hdr)->mgm_time;
}
static inline __u64 phys_cpu__ctidx(enum diag204_format type, void *hdr)
{
if (type == INFO_SIMPLE)
return ((struct phys_cpu *)hdr)->ctidx;
else /* INFO_EXT */
return ((struct x_phys_cpu *)hdr)->ctidx;
}
/* Diagnose 204 functions */
static int diag204(unsigned long subcode, unsigned long size, void *addr)
{
register unsigned long _subcode asm("0") = subcode;
register unsigned long _size asm("1") = size;
asm volatile (" diag %2,%0,0x204\n"
"0: \n" ".section __ex_table,\"a\"\n"
#ifndef __s390x__
" .align 4\n"
" .long 0b,0b\n"
#else
" .align 8\n"
" .quad 0b,0b\n"
#endif
".previous":"+d" (_subcode), "+d"(_size)
:"d"(addr)
:"memory");
if (_subcode)
return -1;
else
return _size;
}
/*
* For the old diag subcode 4 with simple data format we have to use real
* memory. If we use subcode 6 or 7 with extended data format, we can (and
* should) use vmalloc, since we need a lot of memory in that case. Currently
* up to 93 pages!
*/
static void diag204_free_buffer(void)
{
if (!diag204_buf)
return;
if (diag204_buf_vmalloc) {
vfree(diag204_buf_vmalloc);
diag204_buf_vmalloc = NULL;
} else {
free_pages((unsigned long) diag204_buf, 0);
}
diag204_buf_pages = 0;
diag204_buf = NULL;
}
static void *diag204_alloc_vbuf(int pages)
{
/* The buffer has to be page aligned! */
diag204_buf_vmalloc = vmalloc(PAGE_SIZE * (pages + 1));
if (!diag204_buf_vmalloc)
return ERR_PTR(-ENOMEM);
diag204_buf = (void*)((unsigned long)diag204_buf_vmalloc
& ~0xfffUL) + 0x1000;
diag204_buf_pages = pages;
return diag204_buf;
}
static void *diag204_alloc_rbuf(void)
{
diag204_buf = (void*)__get_free_pages(GFP_KERNEL,0);
if (diag204_buf)
return ERR_PTR(-ENOMEM);
diag204_buf_pages = 1;
return diag204_buf;
}
static void *diag204_get_buffer(enum diag204_format fmt, int *pages)
{
if (diag204_buf) {
*pages = diag204_buf_pages;
return diag204_buf;
}
if (fmt == INFO_SIMPLE) {
*pages = 1;
return diag204_alloc_rbuf();
} else {/* INFO_EXT */
*pages = diag204(SUBC_RSI | INFO_EXT, 0, 0);
if (*pages <= 0)
return ERR_PTR(-ENOSYS);
else
return diag204_alloc_vbuf(*pages);
}
}
/*
* diag204_probe() has to find out, which type of diagnose 204 implementation
* we have on our machine. Currently there are three possible scanarios:
* - subcode 4 + simple data format (only one page)
* - subcode 4-6 + extended data format
* - subcode 4-7 + extended data format
*
* Subcode 5 is used to retrieve the size of the data, provided by subcodes
* 6 and 7. Subcode 7 basically has the same function as subcode 6. In addition
* to subcode 6 it provides also information about secondary cpus.
* In order to get as much information as possible, we first try
* subcode 7, then 6 and if both fail, we use subcode 4.
*/
static int diag204_probe(void)
{
void *buf;
int pages, rc;
buf = diag204_get_buffer(INFO_EXT, &pages);
if (!IS_ERR(buf)) {
if (diag204(SUBC_STIB7 | INFO_EXT, pages, buf) >= 0) {
diag204_store_sc = SUBC_STIB7;
diag204_info_type = INFO_EXT;
goto out;
}
if (diag204(SUBC_STIB6 | INFO_EXT, pages, buf) >= 0) {
diag204_store_sc = SUBC_STIB7;
diag204_info_type = INFO_EXT;
goto out;
}
diag204_free_buffer();
}
/* subcodes 6 and 7 failed, now try subcode 4 */
buf = diag204_get_buffer(INFO_SIMPLE, &pages);
if (IS_ERR(buf)) {
rc = PTR_ERR(buf);
goto fail_alloc;
}
if (diag204(SUBC_STIB4 | INFO_SIMPLE, pages, buf) >= 0) {
diag204_store_sc = SUBC_STIB4;
diag204_info_type = INFO_SIMPLE;
goto out;
} else {
rc = -ENOSYS;
goto fail_store;
}
out:
rc = 0;
fail_store:
diag204_free_buffer();
fail_alloc:
return rc;
}
static void *diag204_store(void)
{
void *buf;
int pages;
buf = diag204_get_buffer(diag204_info_type, &pages);
if (IS_ERR(buf))
goto out;
if (diag204(diag204_store_sc | diag204_info_type, pages, buf) < 0)
return ERR_PTR(-ENOSYS);
out:
return buf;
}
/* Diagnose 224 functions */
static void diag224(void *ptr)
{
asm volatile(" diag %0,%1,0x224\n"
: :"d" (0), "d"(ptr) : "memory");
}
static int diag224_get_name_table(void)
{
/* memory must be below 2GB */
diag224_cpu_names = kmalloc(PAGE_SIZE, GFP_KERNEL | GFP_DMA);
if (!diag224_cpu_names)
return -ENOMEM;
diag224(diag224_cpu_names);
EBCASC(diag224_cpu_names + 16, (*diag224_cpu_names + 1) * 16);
return 0;
}
static void diag224_delete_name_table(void)
{
kfree(diag224_cpu_names);
}
static int diag224_idx2name(int index, char *name)
{
memcpy(name, diag224_cpu_names + ((index + 1) * CPU_NAME_LEN),
CPU_NAME_LEN);
name[CPU_NAME_LEN] = 0;
strstrip(name);
return 0;
}
__init int hypfs_diag_init(void)
{
int rc;
if (diag204_probe()) {
printk(KERN_ERR "hypfs: diag 204 not working.");
return -ENODATA;
}
rc = diag224_get_name_table();
if (rc) {
diag224_delete_name_table();
printk(KERN_ERR "hypfs: could not get name table.\n");
}
return rc;
}
__exit void hypfs_diag_exit(void)
{
diag224_delete_name_table();
diag204_free_buffer();
}
/*
* Functions to create the directory structure
* *******************************************
*/
static int hypfs_create_cpu_files(struct super_block *sb,
struct dentry *cpus_dir, void *cpu_info)
{
struct dentry *cpu_dir;
char buffer[TMP_SIZE];
void *rc;
snprintf(buffer, TMP_SIZE, "%d", cpu_info__cpu_addr(diag204_info_type,
cpu_info));
cpu_dir = hypfs_mkdir(sb, cpus_dir, buffer);
rc = hypfs_create_u64(sb, cpu_dir, "mgmtime",
cpu_info__acc_time(diag204_info_type, cpu_info) -