* Copyright (c) 2008, 2010, Oracle and/or its affiliates. All rights reserved.
* Use is subject to license terms.
*
- * Copyright (c) 2011, Whamcloud, Inc.
+ * Copyright (c) 2011, 2012, Intel Corporation.
*/
/*
* This file is part of Lustre, http://www.lustre.org/
#ifndef __LIBCFS_PRIVATE_H__
#define __LIBCFS_PRIVATE_H__
-/* XXX this layering violation is for nidstrings */
-#include <lnet/types.h>
-
#ifndef DEBUG_SUBSYSTEM
# define DEBUG_SUBSYSTEM S_UNDEFINED
#endif
*
* requires -Wall. Unfortunately this rules out use of likely/unlikely.
*/
-#define LASSERTF(cond, fmt, ...) \
-do { \
- if (cond) \
- ; \
- else { \
- LIBCFS_DEBUG_MSG_DATA_DECL(msgdata, D_EMERG, NULL); \
- libcfs_debug_msg(&msgdata, \
- "ASSERTION( %s ) failed: " fmt, #cond, \
- ## __VA_ARGS__); \
- LBUG(); \
- } \
-} while(0)
+#define LASSERTF(cond, fmt, ...) \
+do { \
+ if (cond) \
+ ; \
+ else { \
+ LIBCFS_DEBUG_MSG_DATA_DECL(__msg_data, D_EMERG, NULL); \
+ libcfs_debug_msg(&__msg_data, \
+ "ASSERTION( %s ) failed: " fmt, #cond, \
+ ## __VA_ARGS__); \
+ lbug_with_loc(&__msg_data); \
+ } \
+} while (0)
#define LASSERT(cond) LASSERTF(cond, "\n")
#else /* !LASSERT_CHECKED */
-#define LASSERTF(cond, fmt, ...) \
-do { \
- if (unlikely(!(cond))) { \
- LIBCFS_DEBUG_MSG_DATA_DECL(msgdata, D_EMERG, NULL); \
- libcfs_debug_msg(&msgdata, \
- "ASSERTION( %s ) failed: " fmt, #cond, \
- ## __VA_ARGS__ ); \
- LBUG(); \
- } \
-} while(0)
+#define LASSERTF(cond, fmt, ...) \
+do { \
+ if (unlikely(!(cond))) { \
+ LIBCFS_DEBUG_MSG_DATA_DECL(__msg_data, D_EMERG, NULL); \
+ libcfs_debug_msg(&__msg_data, \
+ "ASSERTION( %s ) failed: " fmt, #cond, \
+ ## __VA_ARGS__); \
+ lbug_with_loc(&__msg_data); \
+ } \
+} while (0)
#define LASSERT(cond) LASSERTF(cond, "\n")
#endif /* !LASSERT_CHECKED */
# define LASSERTF(cond, ...) ((void)sizeof!!(cond))
#endif /* !LIBCFS_DEBUG */
-#ifdef INVARIANT_CHECK
+#ifdef CONFIG_LUSTRE_DEBUG_EXPENSIVE_CHECK
/**
* This is for more expensive checks that one doesn't want to be enabled all
* the time. LINVRNT() has to be explicitly enabled by --enable-invariants
lbug_with_loc(&msgdata); \
} while(0)
-extern cfs_atomic_t libcfs_kmemory;
+extern atomic_t libcfs_kmemory;
/*
* Memory
*/
# define libcfs_kmem_inc(ptr, size) \
do { \
- cfs_atomic_add(size, &libcfs_kmemory); \
+ atomic_add(size, &libcfs_kmemory); \
} while (0)
# define libcfs_kmem_dec(ptr, size) \
do { \
- cfs_atomic_sub(size, &libcfs_kmemory); \
+ atomic_sub(size, &libcfs_kmemory); \
} while (0)
# define libcfs_kmem_read() \
- cfs_atomic_read(&libcfs_kmemory)
+ atomic_read(&libcfs_kmemory)
#else
# define libcfs_kmem_inc(ptr, size) do {} while (0)
#endif /* LIBCFS_DEBUG */
#ifndef LIBCFS_VMALLOC_SIZE
-#define LIBCFS_VMALLOC_SIZE (2 << CFS_PAGE_SHIFT) /* 2 pages */
+#define LIBCFS_VMALLOC_SIZE (2 << PAGE_CACHE_SHIFT) /* 2 pages */
#endif
#define LIBCFS_ALLOC_PRE(size, mask) \
do { \
- LASSERT(!cfs_in_interrupt() || \
+ LASSERT(!in_interrupt() || \
((size) <= LIBCFS_VMALLOC_SIZE && \
- ((mask) & CFS_ALLOC_ATOMIC)) != 0); \
+ ((mask) & GFP_ATOMIC)) != 0); \
} while (0)
#define LIBCFS_ALLOC_POST(ptr, size) \
CERROR("LNET: %d total bytes allocated by lnet\n", \
libcfs_kmem_read()); \
} else { \
- memset((ptr), 0, (size)); \
libcfs_kmem_inc((ptr), (size)); \
CDEBUG(D_MALLOC, "alloc '" #ptr "': %d at %p (tot %d).\n", \
(int)(size), (ptr), libcfs_kmem_read()); \
/**
* allocate memory with GFP flags @mask
+ * The allocated memory is zeroed-out.
*/
#define LIBCFS_ALLOC_GFP(ptr, size, mask) \
do { \
LIBCFS_ALLOC_PRE((size), (mask)); \
(ptr) = (size) <= LIBCFS_VMALLOC_SIZE ? \
- cfs_alloc((size), (mask)) : cfs_alloc_large(size); \
+ kzalloc((size), (mask)) : vzalloc(size); \
LIBCFS_ALLOC_POST((ptr), (size)); \
} while (0)
* default allocator
*/
#define LIBCFS_ALLOC(ptr, size) \
- LIBCFS_ALLOC_GFP(ptr, size, CFS_ALLOC_IO)
+ LIBCFS_ALLOC_GFP(ptr, size, GFP_NOFS)
/**
* non-sleeping allocator
*/
#define LIBCFS_ALLOC_ATOMIC(ptr, size) \
- LIBCFS_ALLOC_GFP(ptr, size, CFS_ALLOC_ATOMIC)
+ LIBCFS_ALLOC_GFP(ptr, size, GFP_ATOMIC)
/**
* allocate memory for specified CPU partition
* \a cptab != NULL, \a cpt is CPU partition id of \a cptab
* \a cptab == NULL, \a cpt is HW NUMA node id
+ * The allocated memory is zeroed-out.
*/
#define LIBCFS_CPT_ALLOC_GFP(ptr, cptab, cpt, size, mask) \
do { \
LIBCFS_ALLOC_PRE((size), (mask)); \
(ptr) = (size) <= LIBCFS_VMALLOC_SIZE ? \
- cfs_cpt_malloc((cptab), (cpt), (size), (mask)) : \
- cfs_cpt_vmalloc((cptab), (cpt), (size)); \
+ cfs_cpt_malloc((cptab), (cpt), (size), (mask) | __GFP_ZERO) : \
+ cfs_cpt_vzalloc((cptab), (cpt), (size)); \
LIBCFS_ALLOC_POST((ptr), (size)); \
} while (0)
/** default numa allocator */
#define LIBCFS_CPT_ALLOC(ptr, cptab, cpt, size) \
- LIBCFS_CPT_ALLOC_GFP(ptr, cptab, cpt, size, CFS_ALLOC_IO)
-
-#define LIBCFS_FREE(ptr, size) \
-do { \
- int s = (size); \
- if (unlikely((ptr) == NULL)) { \
- CERROR("LIBCFS: free NULL '" #ptr "' (%d bytes) at " \
- "%s:%d\n", s, __FILE__, __LINE__); \
- break; \
- } \
- libcfs_kmem_dec((ptr), s); \
- CDEBUG(D_MALLOC, "kfreed '" #ptr "': %d at %p (tot %d).\n", \
+ LIBCFS_CPT_ALLOC_GFP(ptr, cptab, cpt, size, GFP_NOFS)
+
+#define LIBCFS_FREE(ptr, size) \
+do { \
+ int s = (size); \
+ if (unlikely((ptr) == NULL)) { \
+ CERROR("LIBCFS: free NULL '" #ptr "' (%d bytes) at " \
+ "%s:%d\n", s, __FILE__, __LINE__); \
+ break; \
+ } \
+ libcfs_kmem_dec((ptr), s); \
+ CDEBUG(D_MALLOC, "kfreed '" #ptr "': %d at %p (tot %d).\n", \
s, (ptr), libcfs_kmem_read()); \
- if (unlikely(s > LIBCFS_VMALLOC_SIZE)) \
- cfs_free_large(ptr); \
- else \
- cfs_free(ptr); \
+ if (unlikely(s > LIBCFS_VMALLOC_SIZE)) \
+ vfree(ptr); \
+ else \
+ kfree(ptr); \
} while (0)
/******************************************************************************/
#define ntohs(x) ___ntohs(x)
#endif
-void libcfs_debug_dumpstack(cfs_task_t *tsk);
+void libcfs_debug_dumpstack(struct task_struct *tsk);
void libcfs_run_upcall(char **argv);
void libcfs_run_lbug_upcall(struct libcfs_debug_msg_data *);
void libcfs_debug_dumplog(void);
assert(cond); \
} while (0)
# define LBUG() assert(0)
-# ifdef INVARIANT_CHECK
+# ifdef CONFIG_LUSTRE_DEBUG_EXPENSIVE_CHECK
# define LINVRNT(exp) LASSERT(exp)
# else
# define LINVRNT(exp) ((void)sizeof!!(exp))
# endif /* LIBCFS_DEBUG */
# define KLASSERT(e) ((void)0)
# define printk printf
-# ifdef CRAY_XT3 /* buggy calloc! */
-# define LIBCFS_ALLOC_GFP(ptr, size, mask) \
- do { \
- (ptr) = malloc(size); \
- memset(ptr, 0, size); \
- } while (0)
-# else
-# define LIBCFS_ALLOC_GFP(ptr, size, mask) \
- do { \
+#define LIBCFS_ALLOC_GFP(ptr, size, mask) \
+do { \
(ptr) = calloc(1, size); \
- } while (0)
-# endif
+} while (0)
# define LIBCFS_FREE(ptr, size) do { free(ptr); } while((size) - (size))
# define LIBCFS_ALLOC(ptr, size) \
LIBCFS_ALLOC_GFP(ptr, size, 0)
* build go below this comment. Actual compiler/compiler version
* specific implementations come from the above header files
*/
-#ifdef __GNUC__
#define likely(x) __builtin_expect(!!(x), 1)
#define unlikely(x) __builtin_expect(!!(x), 0)
-#else
-#define likely(x) (!!(x))
-#define unlikely(x) (!!(x))
-#endif
/* !__KERNEL__ */
#endif
+struct cfs_cpt_table;
+
+/*
+ * allocate per-cpu-partition data, returned value is an array of pointers,
+ * variable can be indexed by CPU ID.
+ * cptable != NULL: size of array is number of CPU partitions
+ * cptable == NULL: size of array is number of HW cores
+ */
+void *cfs_percpt_alloc(struct cfs_cpt_table *cptab, unsigned int size);
+/*
+ * destory per-cpu-partition variable
+ */
+void cfs_percpt_free(void *vars);
+int cfs_percpt_number(void *vars);
+void *cfs_percpt_current(void *vars);
+void *cfs_percpt_index(void *vars, int idx);
+
+#define cfs_percpt_for_each(var, i, vars) \
+ for (i = 0; i < cfs_percpt_number(vars) && \
+ ((var) = (vars)[i]) != NULL; i++)
+
+/*
+ * allocate a variable array, returned value is an array of pointers.
+ * Caller can specify length of array by count.
+ */
+void *cfs_array_alloc(int count, unsigned int size);
+void cfs_array_free(void *vars);
+
#define LASSERT_ATOMIC_ENABLED (1)
#if LASSERT_ATOMIC_ENABLED
/** assert value of @a is equal to @v */
#define LASSERT_ATOMIC_EQ(a, v) \
do { \
- LASSERTF(cfs_atomic_read(a) == v, \
- "value: %d\n", cfs_atomic_read((a))); \
+ LASSERTF(atomic_read(a) == v, \
+ "value: %d\n", atomic_read((a))); \
} while (0)
/** assert value of @a is unequal to @v */
#define LASSERT_ATOMIC_NE(a, v) \
do { \
- LASSERTF(cfs_atomic_read(a) != v, \
- "value: %d\n", cfs_atomic_read((a))); \
+ LASSERTF(atomic_read(a) != v, \
+ "value: %d\n", atomic_read((a))); \
} while (0)
/** assert value of @a is little than @v */
#define LASSERT_ATOMIC_LT(a, v) \
do { \
- LASSERTF(cfs_atomic_read(a) < v, \
- "value: %d\n", cfs_atomic_read((a))); \
+ LASSERTF(atomic_read(a) < v, \
+ "value: %d\n", atomic_read((a))); \
} while (0)
/** assert value of @a is little/equal to @v */
#define LASSERT_ATOMIC_LE(a, v) \
do { \
- LASSERTF(cfs_atomic_read(a) <= v, \
- "value: %d\n", cfs_atomic_read((a))); \
+ LASSERTF(atomic_read(a) <= v, \
+ "value: %d\n", atomic_read((a))); \
} while (0)
/** assert value of @a is great than @v */
#define LASSERT_ATOMIC_GT(a, v) \
do { \
- LASSERTF(cfs_atomic_read(a) > v, \
- "value: %d\n", cfs_atomic_read((a))); \
+ LASSERTF(atomic_read(a) > v, \
+ "value: %d\n", atomic_read((a))); \
} while (0)
/** assert value of @a is great/equal to @v */
#define LASSERT_ATOMIC_GE(a, v) \
do { \
- LASSERTF(cfs_atomic_read(a) >= v, \
- "value: %d\n", cfs_atomic_read((a))); \
+ LASSERTF(atomic_read(a) >= v, \
+ "value: %d\n", atomic_read((a))); \
} while (0)
/** assert value of @a is great than @v1 and little than @v2 */
#define LASSERT_ATOMIC_GT_LT(a, v1, v2) \
do { \
- int __v = cfs_atomic_read(a); \
- LASSERTF(__v > v1 && __v < v2, "value: %d\n", __v); \
+ int __v = atomic_read(a); \
+ LASSERTF(__v > v1 && __v < v2, "value: %d\n", __v); \
} while (0)
/** assert value of @a is great than @v1 and little/equal to @v2 */
#define LASSERT_ATOMIC_GT_LE(a, v1, v2) \
do { \
- int __v = cfs_atomic_read(a); \
- LASSERTF(__v > v1 && __v <= v2, "value: %d\n", __v); \
+ int __v = atomic_read(a); \
+ LASSERTF(__v > v1 && __v <= v2, "value: %d\n", __v); \
} while (0)
/** assert value of @a is great/equal to @v1 and little than @v2 */
#define LASSERT_ATOMIC_GE_LT(a, v1, v2) \
do { \
- int __v = cfs_atomic_read(a); \
- LASSERTF(__v >= v1 && __v < v2, "value: %d\n", __v); \
+ int __v = atomic_read(a); \
+ LASSERTF(__v >= v1 && __v < v2, "value: %d\n", __v); \
} while (0)
/** assert value of @a is great/equal to @v1 and little/equal to @v2 */
#define LASSERT_ATOMIC_GE_LE(a, v1, v2) \
do { \
- int __v = cfs_atomic_read(a); \
- LASSERTF(__v >= v1 && __v <= v2, "value: %d\n", __v); \
+ int __v = atomic_read(a); \
+ LASSERTF(__v >= v1 && __v <= v2, "value: %d\n", __v); \
} while (0)
#else /* !LASSERT_ATOMIC_ENABLED */
#define CFS_ALLOC_PTR(ptr) LIBCFS_ALLOC(ptr, sizeof (*(ptr)));
#define CFS_FREE_PTR(ptr) LIBCFS_FREE(ptr, sizeof (*(ptr)));
+/*
+ * percpu partition lock
+ *
+ * There are some use-cases like this in Lustre:
+ * . each CPU partition has it's own private data which is frequently changed,
+ * and mostly by the local CPU partition.
+ * . all CPU partitions share some global data, these data are rarely changed.
+ *
+ * LNet is typical example.
+ * CPU partition lock is designed for this kind of use-cases:
+ * . each CPU partition has it's own private lock
+ * . change on private data just needs to take the private lock
+ * . read on shared data just needs to take _any_ of private locks
+ * . change on shared data needs to take _all_ private locks,
+ * which is slow and should be really rare.
+ */
+
+enum {
+ CFS_PERCPT_LOCK_EX = -1, /* negative */
+};
+
+#ifdef __KERNEL__
+
+struct cfs_percpt_lock {
+ /* cpu-partition-table for this lock */
+ struct cfs_cpt_table *pcl_cptab;
+ /* exclusively locked */
+ unsigned int pcl_locked;
+ /* private lock table */
+ spinlock_t **pcl_locks;
+};
+
+/* return number of private locks */
+#define cfs_percpt_lock_num(pcl) cfs_cpt_number(pcl->pcl_cptab)
+
+#else /* !__KERNEL__ */
+
+# ifdef HAVE_LIBPTHREAD
+
+struct cfs_percpt_lock {
+ pthread_mutex_t pcl_mutex;
+};
+
+# else /* !HAVE_LIBPTHREAD */
+
+struct cfs_percpt_lock {
+ int pcl_lock;
+};
+
+static const struct cfs_percpt_lock CFS_PERCPT_LOCK_MAGIC;
+
+# endif /* HAVE_LIBPTHREAD */
+# define cfs_percpt_lock_num(pcl) 1
+#endif /* __KERNEL__ */
+
+/*
+ * create a cpu-partition lock based on CPU partition table \a cptab,
+ * each private lock has extra \a psize bytes padding data
+ */
+struct cfs_percpt_lock *cfs_percpt_lock_alloc(struct cfs_cpt_table *cptab);
+/* destroy a cpu-partition lock */
+void cfs_percpt_lock_free(struct cfs_percpt_lock *pcl);
+
+/* lock private lock \a index of \a pcl */
+void cfs_percpt_lock(struct cfs_percpt_lock *pcl, int index);
+/* unlock private lock \a index of \a pcl */
+void cfs_percpt_unlock(struct cfs_percpt_lock *pcl, int index);
+/* create percpt (atomic) refcount based on @cptab */
+atomic_t **cfs_percpt_atomic_alloc(struct cfs_cpt_table *cptab, int val);
+/* destroy percpt refcount */
+void cfs_percpt_atomic_free(atomic_t **refs);
+/* return sum of all percpu refs */
+int cfs_percpt_atomic_summary(atomic_t **refs);
+
+
/** Compile-time assertion.
* Check an invariant described by a constant expression at compile time by
* value after conversion...
*
*/
-#define CLASSERT(cond) do {switch(42) {case (cond): case 0: break;}} while (0)
-
-/* support decl needed both by kernel and liblustre */
-int libcfs_isknown_lnd(int type);
-char *libcfs_lnd2modname(int type);
-char *libcfs_lnd2str(int type);
-int libcfs_str2lnd(const char *str);
-char *libcfs_net2str(__u32 net);
-char *libcfs_nid2str(lnet_nid_t nid);
-__u32 libcfs_str2net(const char *str);
-lnet_nid_t libcfs_str2nid(const char *str);
-int libcfs_str2anynid(lnet_nid_t *nid, const char *str);
-char *libcfs_id2str(lnet_process_id_t id);
-void cfs_free_nidlist(cfs_list_t *list);
-int cfs_parse_nidlist(char *str, int len, cfs_list_t *list);
-int cfs_match_nid(lnet_nid_t nid, cfs_list_t *list);
-
-/** \addtogroup lnet_addr
- * @{ */
-/* how an LNET NID encodes net:address */
-/** extract the address part of an lnet_nid_t */
-#define LNET_NIDADDR(nid) ((__u32)((nid) & 0xffffffff))
-/** extract the network part of an lnet_nid_t */
-#define LNET_NIDNET(nid) ((__u32)(((nid) >> 32)) & 0xffffffff)
-/** make an lnet_nid_t from a network part and an address part */
-#define LNET_MKNID(net,addr) ((((__u64)(net))<<32)|((__u64)(addr)))
-/* how net encodes type:number */
-#define LNET_NETNUM(net) ((net) & 0xffff)
-#define LNET_NETTYP(net) (((net) >> 16) & 0xffff)
-#define LNET_MKNET(typ,num) ((((__u32)(typ))<<16)|((__u32)(num)))
-/** @} lnet_addr */
-
-/* max value for numeric network address */
-#define MAX_NUMERIC_VALUE 0xffffffff
+#define CLASSERT(cond) do {switch (1) {case (cond): case 0: break; } } while (0)
/* implication */
#define ergo(a, b) (!(a) || (b))
# define CFS_CURRENT_TIME time(0)
#endif
-/* --------------------------------------------------------------------
- * Light-weight trace
- * Support for temporary event tracing with minimal Heisenberg effect.
- * All stuff about lwt are put in arch/kp30.h
- * -------------------------------------------------------------------- */
-
struct libcfs_device_userstate
{
- int ldu_memhog_pages;
- cfs_page_t *ldu_memhog_root_page;
+ int ldu_memhog_pages;
+ struct page *ldu_memhog_root_page;
};
/* what used to be in portals_lib.h */
#define MKSTR(ptr) ((ptr))? (ptr) : ""
-static inline int cfs_size_round4 (int val)
+static inline size_t cfs_size_round4(size_t val)
{
return (val + 3) & (~0x3);
}
#ifndef HAVE_CFS_SIZE_ROUND
-static inline int cfs_size_round (int val)
+static inline size_t cfs_size_round(size_t val)
{
return (val + 7) & (~0x7);
}
#define HAVE_CFS_SIZE_ROUND
#endif
-static inline int cfs_size_round16(int val)
+static inline size_t cfs_size_round16(size_t val)
{
return (val + 0xf) & (~0xf);
}
-static inline int cfs_size_round32(int val)
+static inline size_t cfs_size_round32(size_t val)
{
return (val + 0x1f) & (~0x1f);
}
-static inline int cfs_size_round0(int val)
+static inline size_t cfs_size_round0(size_t val)
{
if (!val)
return 0;
static inline size_t cfs_round_strlen(char *fset)
{
- return (size_t)cfs_size_round((int)strlen(fset) + 1);
+ return cfs_size_round(strlen(fset) + 1);
+}
+
+/* roundup \a val to power2 */
+static inline size_t cfs_power2_roundup(size_t val)
+{
+ if (val != LOWEST_BIT_SET(val)) { /* not a power of 2 already */
+ do {
+ val &= ~LOWEST_BIT_SET(val);
+ } while (val != LOWEST_BIT_SET(val));
+ /* ...and round up */
+ val <<= 1;
+ }
+ return val;
}
#define LOGL(var,len,ptr) \
ptr += cfs_size_round(len); \
} while (0)
-#define LOGL0(var,len,ptr) \
-do { \
- if (!len) \
- break; \
- memcpy((char *)ptr, (const char *)var, len); \
- *((char *)(ptr) + len) = 0; \
- ptr += cfs_size_round(len + 1); \
-} while (0)
-
/**
* Lustre Network Driver types.
*/
VIBLND = 11,
MXLND = 12,
GNILND = 13,
+ GNIIPLND = 14,
};
+extern struct cfs_psdev_ops libcfs_psdev_ops;
+extern struct miscdevice libcfs_dev;
+extern struct cfs_wi_sched *cfs_sched_rehash;
+
+extern int insert_proc(void);
+extern void remove_proc(void);
+
#endif
git://git.whamcloud.com / fs / lustre-release.git / blobdiff