/* * Copyright © 1999 NCSA * All rights reserved. * * Programmer: Robb Matzke * Thursday, July 29, 1999 * * Purpose: The POSIX unbuffered file driver using only the HDF5 public * API and with a few optimizations: the lseek() call is made * only when the current file position is unknown or needs to be * changed based on previous I/O through this driver (don't mix * I/O from this driver with I/O from other parts of the * application to the same file). */ #include #include #include #include #include #include #include #ifdef HAVE_UNISTD_H #include #endif #ifdef WIN32 #include #endif #undef MAX #define MAX(X,Y) ((X)>(Y)?(X):(Y)) /* The driver identification number, initialized at runtime */ static hid_t H5FD_SEC2_g = 0; /* File operations */ #define OP_UNKNOWN 0 #define OP_READ 1 #define OP_WRITE 2 /* * The description of a file belonging to this driver. The `eoa' and `eof' * determine the amount of hdf5 address space in use and the high-water mark * of the file (the current size of the underlying Unix file). The `pos' * value is used to eliminate file position updates when they would be a * no-op. Unfortunately we've found systems that use separate file position * indicators for reading and writing so the lseek can only be eliminated if * the current operation is the same as the previous operation. When opening * a file the `eof' will be set to the current file size, `eoa' will be set * to zero, `pos' will be set to H5F_ADDR_UNDEF (as it is when an error * occurs), and `op' will be set to H5F_OP_UNKNOWN. */ typedef struct H5FD_sec2_t { H5FD_t pub; /*public stuff, must be first */ int fd; /*the unix file */ haddr_t eoa; /*end of allocated region */ haddr_t eof; /*end of file; current file size*/ haddr_t pos; /*current file I/O position */ int op; /*last operation */ #ifndef WIN32 /* * On most systems the combination of device and i-node number uniquely * identify a file. */ dev_t device; /*file device number */ ino_t inode; /*file i-node number */ #else /* * On WIN32 the low-order word of a unique identifier associated with the * file and the volume serial number uniquely identify a file. This number * (which, both? -rpm) may change when the system is restarted or when the * file is opened. After a process opens a file, the identifier is * constant until the file is closed. An application can use this * identifier and the volume serial number to determine whether two * handles refer to the same file. */ int fileindexlo; int fileindexhi; #endif } H5FD_sec2_t; /* * This driver supports systems that have the lseek64() function by defining * some macros here so we don't have to have conditional compilations later * throughout the code. * * file_offset_t: The datatype for file offsets, the second argument of * the lseek() or lseek64() call. * * file_seek: The function which adjusts the current file position, * either lseek() or lseek64(). */ #ifdef HAVE_LSEEK64 # define file_offset_t off64_t # define file_seek lseek64 #else # define file_offset_t off_t # define file_seek lseek #endif /* * These macros check for overflow of various quantities. These macros * assume that file_offset_t is signed and haddr_t and size_t are unsigned. * * ADDR_OVERFLOW: Checks whether a file address of type `haddr_t' * is too large to be represented by the second argument * of the file seek function. * * SIZE_OVERFLOW: Checks whether a buffer size of type `hsize_t' is too * large to be represented by the `size_t' type. * * REGION_OVERFLOW: Checks whether an address and size pair describe data * which can be addressed entirely by the second * argument of the file seek function. */ #define MAXADDR (((haddr_t)1<<(8*sizeof(file_offset_t)-1))-1) #define ADDR_OVERFLOW(A) (HADDR_UNDEF==(A) || \ ((A) & ~(haddr_t)MAXADDR)) #define SIZE_OVERFLOW(Z) ((Z) & ~(hsize_t)MAXADDR) #define REGION_OVERFLOW(A,Z) (ADDR_OVERFLOW(A) || SIZE_OVERFLOW(Z) || \ sizeof(file_offset_t)fd = fd; file->eof = sb.st_size; file->pos = HADDR_UNDEF; file->op = OP_UNKNOWN; #ifdef WIN32 filehandle = _get_osfhandle(fd); results = GetFileInformationByHandle(filehandle, &fileinfo); file->fileindexhi = fileinfo.nFileIndexHigh; file->fileindexlo = fileinfo.nFileIndexLow; #else file->device = sb.st_dev; file->inode = sb.st_ino; #endif return (H5FD_t*)file; } /*------------------------------------------------------------------------- * Function: H5FD_sec2_close * * Purpose: Closes a Unix file. * * Return: Success: 0 * * Failure: -1, file not closed. * * Programmer: Robb Matzke * Thursday, July 29, 1999 * * Modifications: * *------------------------------------------------------------------------- */ static herr_t H5FD_sec2_close(H5FD_t *_file) { H5FD_sec2_t *file = (H5FD_sec2_t*)_file; if (H5FD_sec2_flush(_file)<0) return -1; if (close(file->fd)<0) return -1; free(file); return 0; } /*------------------------------------------------------------------------- * Function: H5FD_sec2_cmp * * Purpose: Compares two files belonging to this driver using an * arbitrary (but consistent) ordering. * * Return: Success: A value like strcmp() * * Failure: never fails (arguments were checked by the * caller). * * Programmer: Robb Matzke * Thursday, July 29, 1999 * * Modifications: * *------------------------------------------------------------------------- */ static int H5FD_sec2_cmp(const H5FD_t *_f1, const H5FD_t *_f2) { const H5FD_sec2_t *f1 = (const H5FD_sec2_t*)_f1; const H5FD_sec2_t *f2 = (const H5FD_sec2_t*)_f2; #ifdef WIN32 if (f1->fileindexhi < f2->fileindexhi) return -1; if (f1->fileindexhi > f2->fileindexhi) return 1; if (f1->fileindexlo < f2->fileindexlo) return -1; if (f1->fileindexlo > f2->fileindexlo) return 1; #else if (f1->device < f2->device) return -1; if (f1->device > f2->device) return 1; if (f1->inode < f2->inode) return -1; if (f1->inode > f2->inode) return 1; #endif return 0; } /*------------------------------------------------------------------------- * Function: H5FD_sec2_get_eoa * * Purpose: Gets the end-of-address marker for the file. The EOA marker * is the first address past the last byte allocated in the * format address space. * * Return: Success: The end-of-address marker. * * Failure: HADDR_UNDEF * * Programmer: Robb Matzke * Monday, August 2, 1999 * * Modifications: * *------------------------------------------------------------------------- */ static haddr_t H5FD_sec2_get_eoa(H5FD_t *_file) { H5FD_sec2_t *file = (H5FD_sec2_t*)_file; return file->eoa; } /*------------------------------------------------------------------------- * Function: H5FD_sec2_set_eoa * * Purpose: Set the end-of-address marker for the file. This function is * called shortly after an existing HDF5 file is opened in order * to tell the driver where the end of the HDF5 data is located. * * Return: Success: 0 * * Failure: -1 * * Programmer: Robb Matzke * Thursday, July 29, 1999 * * Modifications: * *------------------------------------------------------------------------- */ static herr_t H5FD_sec2_set_eoa(H5FD_t *_file, haddr_t addr) { H5FD_sec2_t *file = (H5FD_sec2_t*)_file; file->eoa = addr; return 0; } /*------------------------------------------------------------------------- * Function: H5FD_sec2_get_eof * * Purpose: Returns the end-of-file marker, which is the greater of * either the Unix end-of-file or the HDF5 end-of-address * markers. * * Return: Success: End of file address, the first address past * the end of the "file", either the Unix file * or the HDF5 file. * * Failure: HADDR_UNDEF * * Programmer: Robb Matzke * Thursday, July 29, 1999 * * Modifications: * *------------------------------------------------------------------------- */ static haddr_t H5FD_sec2_get_eof(H5FD_t *_file) { H5FD_sec2_t *file = (H5FD_sec2_t*)_file; return MAX(file->eof, file->eoa); } /*------------------------------------------------------------------------- * Function: H5FD_sec2_read * * Purpose: Reads SIZE bytes of data from FILE beginning at address ADDR * into buffer BUF according to data transfer properties in * DXPL_ID. * * Return: Success: Zero. Result is stored in caller-supplied * buffer BUF. * * Failure: -1, Contents of buffer BUF are undefined. * * Programmer: Robb Matzke * Thursday, July 29, 1999 * * Modifications: * *------------------------------------------------------------------------- */ static herr_t H5FD_sec2_read(H5FD_t *_file, hid_t dxpl_id/*unused*/, haddr_t addr, hsize_t size, void *buf/*out*/) { H5FD_sec2_t *file = (H5FD_sec2_t*)_file; ssize_t nbytes; assert(file && file->pub.cls); assert(buf); /* Check for overflow conditions */ if (HADDR_UNDEF==addr) return -1; if (REGION_OVERFLOW(addr, size)) return -1; if (addr+size>file->eoa) return -1; /* Seek to the correct location */ if ((addr!=file->pos || OP_READ!=file->op) && file_seek(file->fd, (file_offset_t)addr, SEEK_SET)<0) { file->pos = HADDR_UNDEF; file->op = OP_UNKNOWN; return -1; } /* * Read data, being careful of interrupted system calls, partial results, * and the end of the file. */ while (size>0) { do nbytes = read(file->fd, buf, size); while (-1==nbytes && EINTR==errno); if (-1==nbytes) { /* error */ file->pos = HADDR_UNDEF; file->op = OP_UNKNOWN; return -1; } if (0==nbytes) { /* end of file but not end of format address space */ memset(buf, 0, size); size = 0; } assert(nbytes>=0); assert((hsize_t)nbytes<=size); size -= (hsize_t)nbytes; addr += (haddr_t)nbytes; buf = (char*)buf + nbytes; } /* Update current position */ file->pos = addr; file->op = OP_READ; return 0; } /*------------------------------------------------------------------------- * Function: H5FD_sec2_write * * Purpose: Writes SIZE bytes of data to FILE beginning at address ADDR * from buffer BUF according to data transfer properties in * DXPL_ID. * * Return: Success: Zero * * Failure: -1 * * Programmer: Robb Matzke * Thursday, July 29, 1999 * * Modifications: * *------------------------------------------------------------------------- */ static herr_t H5FD_sec2_write(H5FD_t *_file, hid_t dxpl_id/*unused*/, haddr_t addr, hsize_t size, const void *buf) { H5FD_sec2_t *file = (H5FD_sec2_t*)_file; ssize_t nbytes; assert(file && file->pub.cls); assert(buf); /* Check for overflow conditions */ if (HADDR_UNDEF==addr) return -1; if (REGION_OVERFLOW(addr, size)) return -1; if (addr+size>file->eoa) return -1; /* Seek to the correct location */ if ((addr!=file->pos || OP_WRITE!=file->op) && file_seek(file->fd, (file_offset_t)addr, SEEK_SET)<0) { file->pos = HADDR_UNDEF; file->op = OP_UNKNOWN; return -1; } /* * Write the data, being careful of interrupted system calls and partial * results */ while (size>0) { do nbytes = write(file->fd, buf, size); while (-1==nbytes && EINTR==errno); if (-1==nbytes) { /* error */ file->pos = HADDR_UNDEF; file->op = OP_UNKNOWN; return -1; } assert(nbytes>0); assert((hsize_t)nbytes<=size); size -= (hsize_t)nbytes; addr += (haddr_t)nbytes; buf = (const char*)buf + nbytes; } /* Update current position and eof */ file->pos = addr; file->op = OP_WRITE; if (file->pos>file->eof) file->eof = file->pos; return 0; } /*------------------------------------------------------------------------- * Function: H5FD_sec2_flush * * Purpose: Makes sure that the true file size is the same (or larger) * than the end-of-address. * * Return: Success: Non-negative * * Failure: Negative * * Programmer: Robb Matzke * Wednesday, August 4, 1999 * * Modifications: * *------------------------------------------------------------------------- */ static herr_t H5FD_sec2_flush(H5FD_t *_file) { H5FD_sec2_t *file = (H5FD_sec2_t*)_file; if (file->eoa>file->eof) { if (-1==file_seek(file->fd, file->eoa-1, SEEK_SET)) return -1; if (write(file->fd, "", 1)!=1) return -1; file->eof = file->eoa; file->pos = file->eoa; file->op = OP_WRITE; } return 0; }