autoconf.h needs to be handled differently, since it is copied from obj/lib

[libucw.git] / lib / fastbuf.h

diff --git a/lib/fastbuf.h b/lib/fastbuf.h
index c62ed4cd470cf2af754ce50b46ee398cfbffb2a3..1589843d5e3ccde0a701004ee893812b9a37c1af 100644 (file)
--- a/lib/fastbuf.h
+++ b/lib/fastbuf.h
@@ -1,15 +1,22 @@
 /*
 /*

- *     Sherlock Library -- Fast Buffered I/O
+ *     UCW Library -- Fast Buffered I/O

  *
  *

- *     (c) 1997--2000 Martin Mares <mj@ucw.cz>
+ *     (c) 1997--2007 Martin Mares <mj@ucw.cz>
+ *     (c) 2004 Robert Spalek <robert@ucw.cz>
+ *
+ *     This software may be freely distributed and used according to the terms
+ *     of the GNU Lesser General Public License.

  */
 
  */
 

-#ifndef EOF
-#include <stdio.h>
-#endif
+#ifndef _UCW_FASTBUF_H
+#define _UCW_FASTBUF_H
+
+#include <string.h>
+#include <alloca.h>

 
 /*
 
 /*

- *  Generic buffered I/O on a top of buffer swapping functions.
+ *  Generic buffered I/O. You supply hooks to be called for low-level operations
+ *  (swapping of buffers, seeking and closing), we do the rest.

  *
  *  Buffer layout when reading:
  *
  *
  *  Buffer layout when reading:
  *


@@ -25,50 +32,126 @@
  *
  *  When writing:
  *
  *
  *  When writing:
  *

- *  +----------------+---------------------------+
- *  | written data   | free space                |
- *  +----------------+---------------------------+
- *  ^                 ^                           ^
- *  buffer=bstop      bptr                        bufend
+ *  +--------+--------------+--------------------+
+ *  | unused | written data | free space         |
+ *  +--------+--------------+--------------------+
+ *  ^         ^              ^                    ^
+ *  buffer    bstop          bptr                 bufend
+ *
+ *  Dirty tricks:
+ *
+ *    - You can mix reads and writes on the same stream, but you must
+ *     call bflush() in between and remember that the file position
+ *     points after the flushed buffer which is not necessarily the same
+ *     as after the data you've read.
+ *    - The spout/refill hooks can change not only bptr and bstop, but also
+ *     the location of the buffer; fb-mem.c takes advantage of it.
+ *    - In some cases, the user of the bdirect interface can be allowed to modify
+ *     the data in the buffer to avoid unnecessary copying. If the back-end
+ *     allows such modifications, it can set can_overwrite_buffer accordingly:
+ *             *  0 if no modification is allowed,
+ *             *  1 if the user can modify the buffer on the condition that
+ *                  the modifications will be undone before calling the next
+ *                  fastbuf operation
+ *             *  2 if the user is allowed to overwrite the data in the buffer
+ *                  if bdirect_read_commit_modified() is called afterwards.
+ *                  In this case, the back-end must be prepared for trimming
+ *                  of the buffer which is done by the commit function.

  */
 
 struct fastbuf {
  */
 
 struct fastbuf {

+  byte is_fastbuf[0];                  /* Dummy field for checking of type casts */

   byte *bptr, *bstop;                  /* Access pointers */
   byte *buffer, *bufend;               /* Start and end of the buffer */
   byte *bptr, *bstop;                  /* Access pointers */
   byte *buffer, *bufend;               /* Start and end of the buffer */

-  byte *name;                          /* File name for error messages */
-  uns buflen;                          /* Size of the buffer */
-  sh_off_t pos;                                /* Position of buffer start in the file */
-  sh_off_t fdpos;                      /* Current position in the non-buffered file */
-  int fd;                              /* File descriptor, -1 if not a real file */
-  void *lldata;                                /* Data private to access functions below */
-  void *llpos;                         /* ... continued ... */
+  char *name;                          /* File name for error messages */
+  sh_off_t pos;                                /* Position of bstop in the file */

   int (*refill)(struct fastbuf *);     /* Get a buffer with new data */
   void (*spout)(struct fastbuf *);     /* Write buffer data to the file */
   int (*refill)(struct fastbuf *);     /* Get a buffer with new data */
   void (*spout)(struct fastbuf *);     /* Write buffer data to the file */

-  void (*seek)(struct fastbuf *, sh_off_t, int);  /* Slow path for bseek(), buffer already flushed */
+  int (*seek)(struct fastbuf *, sh_off_t, int);  /* Slow path for bseek(), buffer already flushed; returns success */

   void (*close)(struct fastbuf *);     /* Close the stream */
   void (*close)(struct fastbuf *);     /* Close the stream */

+  int (*config)(struct fastbuf *, uns, int);   /* Configure the stream */
+  int can_overwrite_buffer;            /* Can the buffer be altered? (see discussion above) 0=never, 1=temporarily, 2=permanently */

 };
 
 };
 

-/* FastIO on standard files */
+/* FastIO on standard files (specify buffer size 0 to enable mmaping) */
+
+struct fastbuf *bopen(const char *name, uns mode, uns buflen);
+struct fastbuf *bopen_try(const char *name, uns mode, uns buflen);
+struct fastbuf *bopen_tmp(uns buflen);
+struct fastbuf *bfdopen(int fd, uns buflen);
+struct fastbuf *bfdopen_shared(int fd, uns buflen);
+void bfilesync(struct fastbuf *b);

 
 

-struct fastbuf *bopen(byte *name, uns mode, uns buffer);
-struct fastbuf *bfdopen(int fd, uns buffer);
-void bbcopy(struct fastbuf *f, struct fastbuf *t, uns l);
+#define TEMP_FILE_NAME_LEN 256
+void temp_file_name(char *name);

 
 /* FastIO on in-memory streams */
 
 
 /* FastIO on in-memory streams */
 

-struct fastbuf *fbmem_create(unsigned blocksize);      /* Create stream and return its writing fastbuf */
+struct fastbuf *fbmem_create(uns blocksize);           /* Create stream and return its writing fastbuf */

 struct fastbuf *fbmem_clone_read(struct fastbuf *);    /* Create reading fastbuf */
 
 struct fastbuf *fbmem_clone_read(struct fastbuf *);    /* Create reading fastbuf */
 

+/* FastIO on memory mapped files */
+
+struct fastbuf *bopen_mm(const char *name, uns mode);
+
+/* FastI on file descriptors with limit */
+
+struct fastbuf *bopen_limited_fd(int fd, uns bufsize, uns limit);
+
+/* FastIO on static buffers */
+
+void fbbuf_init_read(struct fastbuf *f, byte *buffer, uns size, uns can_overwrite);
+void fbbuf_init_write(struct fastbuf *f, byte *buffer, uns size);
+static inline uns
+fbbuf_count_written(struct fastbuf *f)
+{
+  return f->bptr - f->bstop;
+}
+
+/* FastIO on recyclable growing buffers */
+
+struct fastbuf *fbgrow_create(unsigned basic_size);
+void fbgrow_reset(struct fastbuf *b);                  /* Reset stream and prepare for writing */
+void fbgrow_rewind(struct fastbuf *b);                 /* Prepare for reading */
+
+/* FastO with atomic writes for multi-threaded programs */
+
+struct fb_atomic {
+  struct fastbuf fb;
+  struct fb_atomic_file *af;
+  byte *expected_max_bptr;
+  uns slack_size;
+};
+#define FB_ATOMIC(f) ((struct fb_atomic *)(f)->is_fastbuf)
+
+struct fastbuf *fbatomic_open(const char *name, struct fastbuf *master, uns bufsize, int record_len);
+void fbatomic_internal_write(struct fastbuf *b);
+
+static inline void
+fbatomic_commit(struct fastbuf *b)
+{
+  if (b->bptr >= ((struct fb_atomic *)b)->expected_max_bptr)
+    fbatomic_internal_write(b);
+}
+
+/* Configuring stream parameters */
+
+int bconfig(struct fastbuf *f, uns type, int data);
+
+#define BCONFIG_IS_TEMP_FILE 0
+

 /* Universal functions working on all fastbuf's */
 
 void bclose(struct fastbuf *f);
 void bflush(struct fastbuf *f);
 void bseek(struct fastbuf *f, sh_off_t pos, int whence);
 void bsetpos(struct fastbuf *f, sh_off_t pos);
 /* Universal functions working on all fastbuf's */
 
 void bclose(struct fastbuf *f);
 void bflush(struct fastbuf *f);
 void bseek(struct fastbuf *f, sh_off_t pos, int whence);
 void bsetpos(struct fastbuf *f, sh_off_t pos);

+void brewind(struct fastbuf *f);
+sh_off_t bfilesize(struct fastbuf *f);         // -1 if not seekable

 
 static inline sh_off_t btell(struct fastbuf *f)
 {
 
 static inline sh_off_t btell(struct fastbuf *f)
 {

-  return f->pos + (f->bptr - f->buffer);
+  return f->pos + (f->bptr - f->bstop);

 }
 
 int bgetc_slow(struct fastbuf *f);
 }
 
 int bgetc_slow(struct fastbuf *f);


@@ -83,13 +166,13 @@ static inline int bpeekc(struct fastbuf *f)
   return (f->bptr < f->bstop) ? (int) *f->bptr : bpeekc_slow(f);
 }
 
   return (f->bptr < f->bstop) ? (int) *f->bptr : bpeekc_slow(f);
 }
 

-static inline void bungetc(struct fastbuf *f, byte c)
+static inline void bungetc(struct fastbuf *f)

 {
 {

-  *--f->bptr = c;
+  f->bptr--;

 }
 
 }
 

-void bputc_slow(struct fastbuf *f, byte c);
-static inline void bputc(struct fastbuf *f, byte c)
+void bputc_slow(struct fastbuf *f, uns c);
+static inline void bputc(struct fastbuf *f, uns c)

 {
   if (f->bptr < f->bufend)
     *f->bptr++ = c;
 {
   if (f->bptr < f->bufend)
     *f->bptr++ = c;


@@ -97,222 +180,178 @@ static inline void bputc(struct fastbuf *f, byte c)
     bputc_slow(f, c);
 }
 
     bputc_slow(f, c);
 }
 

-word bgetw_slow(struct fastbuf *f);
-static inline word bgetw(struct fastbuf *f)
+static inline uns
+bavailr(struct fastbuf *f)

 {
 {

-  word w;
-  if (f->bptr + 2 <= f->bstop)
-    {
-      byte *p = f->bptr;
-#ifdef CPU_CAN_DO_UNALIGNED_WORDS
-      w = * ((word *) p);
-#else
-#ifdef CPU_BIG_ENDIAN
-      w = (p[0] << 8) | p[1];
-#else
-      w = (p[1] << 8) | p[0];
-#endif
-#endif
-      f->bptr += 2;
-      return w;
-    }
-  else
-    return bgetw_slow(f);
+  return f->bstop - f->bptr;

 }
 
 }
 

-u32 bgetl_slow(struct fastbuf *f);
-static inline u32 bgetl(struct fastbuf *f)
+static inline uns
+bavailw(struct fastbuf *f)

 {
 {

-  u32 l;
-  if (f->bptr + 4 <= f->bstop)
-    {
-      byte *p = f->bptr;
-#ifdef CPU_CAN_DO_UNALIGNED_LONGS
-      l = * ((u32 *) p);
-#else
-#ifdef CPU_BIG_ENDIAN
-      l = (p[0] << 24) | (p[1] << 16) | (p[2] << 8) | p[3];
-#else
-      l = (p[3] << 24) | (p[2] << 16) | (p[1] << 8) | p[0];
-#endif
-#endif
-      f->bptr += 4;
-      return l;
-    }
-  else
-    return bgetl_slow(f);
+  return f->bufend - f->bptr;

 }
 
 }
 

-u64 bgetq_slow(struct fastbuf *f);
-static inline u64 bgetq(struct fastbuf *f)
+uns bread_slow(struct fastbuf *f, void *b, uns l, uns check);
+static inline uns bread(struct fastbuf *f, void *b, uns l)

 {
 {

-  if (f->bptr + 8 <= f->bstop)
+  if (bavailr(f) >= l)

     {
     {

-      u64 l;
-      memcpy(&l, f->bptr, 8);
-      f->bptr += 8;
+      memcpy(b, f->bptr, l);
+      f->bptr += l;

       return l;
     }
   else
       return l;
     }
   else

-    return bgetq_slow(f);
+    return bread_slow(f, b, l, 0);

 }
 
 }
 

-u64 bget5_slow(struct fastbuf *f);
-static inline u64 bget5(struct fastbuf *f)
+static inline uns breadb(struct fastbuf *f, void *b, uns l)

 {
 {

-  u64 l;
-  if (f->bptr + 5 <= f->bstop)
+  if (bavailr(f) >= l)

     {
     {

-      byte *p = f->bptr;
-#ifdef CPU_BIG_ENDIAN
-      l = ((u64)p[0] << 32) | (u32)((p[1] << 24) | (p[2] << 16) | (p[3] << 8) | p[4]);
-#else
-      l = ((u64)p[4] << 32) | (u32)((p[3] << 24) | (p[2] << 16) | (p[1] << 8) | p[0]);
-#endif
-      f->bptr += 5;
+      memcpy(b, f->bptr, l);
+      f->bptr += l;

       return l;
     }
   else
       return l;
     }
   else

-    return bget5_slow(f);
+    return bread_slow(f, b, l, 1);

 }
 
 }
 

-void bputw_slow(struct fastbuf *f, word w);
-static inline void bputw(struct fastbuf *f, word w)
+void bwrite_slow(struct fastbuf *f, const void *b, uns l);
+static inline void bwrite(struct fastbuf *f, const void *b, uns l)

 {
 {

-  if (f->bptr + 2 <= f->bufend)
+  if (bavailw(f) >= l)

     {
     {

-      byte *p = f->bptr;
-#ifdef CPU_CAN_DO_UNALIGNED_WORDS
-      * ((word *) p) = w;
-#else
-#ifdef CPU_BIG_ENDIAN
-      p[0] = w >> 8U;
-      p[1] = w;
-#else
-      p[1] = w >> 8U;
-      p[0] = w;
-#endif
-#endif
-      f->bptr += 2;
+      memcpy(f->bptr, b, l);
+      f->bptr += l;

     }
   else
     }
   else

-    bputw_slow(f, w);
+    bwrite_slow(f, b, l);

 }
 
 }
 

-void bputl_slow(struct fastbuf *f, u32 l);
-static inline void bputl(struct fastbuf *f, u32 l)
+/*
+ *  Functions for reading of strings:
+ *
+ *     bgets()         reads a line, strip the trailing '\n' and return a pointer
+ *                     to the terminating 0 or NULL on EOF. Dies if the line is too long.
+ *     bgets0()                does the same for 0-terminated strings.
+ *     bgets_nodie()   a variant of bgets() which returns either the length of the
+ *                             string (excluding the terminator) or -1 if the line does not fit
+ *                             in the buffer. In such cases, it returns after reading exactly `l'
+ *                             bytes of input.
+ *     bgets_bb()      a variant of bgets() which allocates the string in a growing buffer
+ *     bgets_mp()      the same, but in a mempool
+ *     bgets_stk()     the same, but on the stack by alloca()
+ */
+
+char *bgets(struct fastbuf *f, char *b, uns l);
+char *bgets0(struct fastbuf *f, char *b, uns l);
+int bgets_nodie(struct fastbuf *f, char *b, uns l);
+
+struct mempool;
+struct bb_t;
+uns bgets_bb(struct fastbuf *f, struct bb_t *b, uns limit);
+char *bgets_mp(struct fastbuf *f, struct mempool *mp);
+
+struct bgets_stk_struct {
+  struct fastbuf *f;
+  byte *old_buf, *cur_buf, *src;
+  uns old_len, cur_len, src_len;
+};
+void bgets_stk_init(struct bgets_stk_struct *s);
+void bgets_stk_step(struct bgets_stk_struct *s);
+#define bgets_stk(fb) ({ struct bgets_stk_struct _s; _s.f = (fb); for (bgets_stk_init(&_s); _s.cur_len; _s.cur_buf = alloca(_s.cur_len), bgets_stk_step(&_s)); _s.cur_buf; })
+
+static inline void
+bputs(struct fastbuf *f, const char *b)

 {
 {

-  if (f->bptr + 4 <= f->bufend)
-    {
-      byte *p = f->bptr;
-#ifdef CPU_CAN_DO_UNALIGNED_LONGS
-      * ((u32 *) p) = l;
-#else
-#ifdef CPU_BIG_ENDIAN
-      p[0] = l >> 24U;
-      p[1] = l >> 16U;
-      p[2] = l >> 8U;
-      p[3] = l;
-#else
-      p[3] = l >> 24U;
-      p[2] = l >> 16U;
-      p[1] = l >> 8U;
-      p[0] = l;
-#endif
-#endif
-      f->bptr += 4;
-    }
-  else
-    bputl_slow(f, l);
+  bwrite(f, b, strlen(b));

 }
 
 }
 

-void bputq_slow(struct fastbuf *f, u64 l);
-static inline void bputq(struct fastbuf *f, u64 l)
+static inline void
+bputs0(struct fastbuf *f, const char *b)

 {
 {

-  if (f->bptr + 8 <= f->bufend)
-    {
-      memcpy(f->bptr, &l, 8);
-      f->bptr += 8;
-    }
-  else
-    bputq_slow(f, l);
+  bwrite(f, b, strlen(b)+1);
+}
+
+static inline void
+bputsn(struct fastbuf *f, const char *b)
+{
+  bputs(f, b);
+  bputc(f, '\n');

 }
 
 }
 

-void bput5_slow(struct fastbuf *f, u64 l);
-static inline void bput5(struct fastbuf *f, u64 l)
+void bbcopy_slow(struct fastbuf *f, struct fastbuf *t, uns l);
+static inline void
+bbcopy(struct fastbuf *f, struct fastbuf *t, uns l)

 {
 {

-  if (f->bptr + 5 <= f->bufend)
+  if (bavailr(f) >= l && bavailw(t) >= l)

     {
     {

-      byte *p = f->bptr;
-      u32 low = l;
-#ifdef CPU_BIG_ENDIAN
-      p[0] = l >> 32U;
-      p[1] = low >> 24U;
-      p[2] = low >> 16U;
-      p[3] = low >> 8U;
-      p[4] = low;
-#else
-      p[4] = l >> 32U;
-      p[3] = low >> 24U;
-      p[2] = low >> 16U;
-      p[1] = low >> 8U;
-      p[0] = low;
-#endif
-      f->bptr += 5;
+      memcpy(t->bptr, f->bptr, l);
+      t->bptr += l;
+      f->bptr += l;

     }
   else
     }
   else

-    bput5_slow(f, l);
+    bbcopy_slow(f, t, l);

 }
 
 }
 

-void bread_slow(struct fastbuf *f, void *b, uns l);
-static inline void bread(struct fastbuf *f, void *b, uns l)
+int bskip_slow(struct fastbuf *f, uns len);
+static inline int bskip(struct fastbuf *f, uns len)

 {
 {

-  if (f->bptr + l <= f->bstop)
+  if (bavailr(f) >= len)

     {
     {

-      memcpy(b, f->bptr, l);
-      f->bptr += l;
+      f->bptr += len;
+      return 1;

     }
   else
     }
   else

-    bread_slow(f, b, l);
+    return bskip_slow(f, len);

 }
 
 }
 

-void bwrite_slow(struct fastbuf *f, void *b, uns l);
-static inline void bwrite(struct fastbuf *f, void *b, uns l)
+/* Direct I/O on buffers */
+
+static inline uns
+bdirect_read_prepare(struct fastbuf *f, byte **buf)

 {
 {

-  if (f->bptr + l <= f->bufend)
+  if (f->bptr == f->bstop && !f->refill(f))

     {
     {

-      memcpy(f->bptr, b, l);
-      f->bptr += l;
+      *buf = NULL;  // This is not needed, but it helps to get rid of spurious warnings
+      return 0;

     }
     }

-  else
-    bwrite_slow(f, b, l);
+  *buf = f->bptr;
+  return bavailr(f);

 }
 
 }
 

-byte *bgets(struct fastbuf *f, byte *b, uns l);        /* Non-std */
+static inline void
+bdirect_read_commit(struct fastbuf *f, byte *pos)
+{
+  f->bptr = pos;
+}

 
 static inline void
 
 static inline void

-bputs(struct fastbuf *f, byte *b)
+bdirect_read_commit_modified(struct fastbuf *f, byte *pos)

 {
 {

-  bwrite(f, b, strlen(b));
+  f->bptr = pos;
+  f->buffer = pos;     /* Avoid seeking backwards in the buffer */
+}
+
+static inline uns
+bdirect_write_prepare(struct fastbuf *f, byte **buf)
+{
+  if (f->bptr == f->bufend)
+    f->spout(f);
+  *buf = f->bptr;
+  return bavailw(f);

 }
 
 static inline void
 }
 
 static inline void

-bputsn(struct fastbuf *f, byte *b)
+bdirect_write_commit(struct fastbuf *f, byte *pos)

 {
 {

-  bputs(f, b);
-  bputc(f, '\n');
+  f->bptr = pos;

 }
 
 }
 

-/* Depending on compile-time configuration, we select the right function for reading/writing of file offsets */
-
-#ifdef SHERLOCK_CONFIG_LARGE_DB
-#define bgeto(f) bget5(f)
-#define bputo(f,l) bput5(f,l)
-#define bgetp(f) bgetq(f)
-#define bputp(f,l) bputq(f,l)
-#else
-#define bgeto(f) bgetl(f)
-#define bputo(f,l) bputl(f,l)
-#define bgetp(f) bgetl(f)
-#define bputp(f,l) bputl(f,l)
+/* Formatted output */
+
+int bprintf(struct fastbuf *b, const char *msg, ...) FORMAT_CHECK(printf,2,3);
+int vbprintf(struct fastbuf *b, const char *msg, va_list args);
+

 #endif
 #endif