mirror/mariadb
1
0
Fork 0
mirror of https://github.com/MariaDB/server.git synced 2025-01-18 21:12:26 +01:00
Code Issues Projects Releases Packages Wiki Activity Actions
mariadb/mysys/my_lib.c

430 lines
12 KiB
C
Raw Normal View History

WL#2360 Performance schema Part III: mysys instrumentation
2009-12-10 04:19:51 +01:00
/* Copyright (C) 2000 MySQL AB, 2008-2009 Sun Microsystems, Inc
Update copyright Fixed memory leak on shutdown (Affects the embedded version & MyODBC)
2001-12-06 13:10:51 +01:00
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
Many files: Changed header to GPL version 2 only
2006-12-23 20:17:15 +01:00
the Free Software Foundation; version 2 of the License.
Update copyright Fixed memory leak on shutdown (Affects the embedded version & MyODBC)
2001-12-06 13:10:51 +01:00
This program is distributed in the hope that it will be useful,
Import changeset
2000-07-31 21:29:14 +02:00
but WITHOUT ANY WARRANTY; without even the implied warranty of
Update copyright Fixed memory leak on shutdown (Affects the embedded version & MyODBC)
2001-12-06 13:10:51 +01:00
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
Import changeset
2000-07-31 21:29:14 +02:00
/* TODO: check for overun of memory for names. */
#include "mysys_priv.h"
#include <m_string.h>
#include <my_dir.h> /* Structs used by my_dir,includes sys/types */
#include "mysys_err.h"
#if defined(HAVE_DIRENT_H)
# include <dirent.h>
# define NAMLEN(dirent) strlen((dirent)->d_name)
#else
# define dirent direct
# define NAMLEN(dirent) (dirent)->d_namlen
# if defined(HAVE_SYS_NDIR_H)
# include <sys/ndir.h>
# endif
# if defined(HAVE_SYS_DIR_H)
# include <sys/dir.h>
# endif
# if defined(HAVE_NDIR_H)
# include <ndir.h>
# endif
This is the downport of Bug#24509 - 2048 file descriptor limit on windows needs increasing, also WL#3049 - improved Windows I/O The patch replaces the use of the POSIX I/O interfaces in mysys on Windows with the Win32 API calls (CreateFile, WriteFile, etc). The Windows HANDLE for the open file is stored in the my_file_info struct, along with a flag for append mode because the Windows API does not support opening files in append mode in all cases) The default max open files has been increased to 16384 and can be increased further by setting --max-open-files=<value> during the server start. Another major change in this patch that almost all Windows specific file IO code has been moved to a new file my_winfile.c, greatly reducing the amount of code in #ifdef blocks within mysys, thus improving readability. Minor enhancements: - my_(f)stat() is changed to use __stati64 structure with 64 file size and timestamps. It will return correct file size now (C runtime implementation used to report outdated information) - my_lock on Windows is prepared to handle additional timeout parameter - after review : changed __WIN__ to _WIN32 in the new and changed code.
2009-09-11 22:26:35 +02:00
# if defined(_WIN32)
Import changeset
2000-07-31 21:29:14 +02:00
# ifdef __BORLANDC__
# include <dir.h>
# endif
# endif
#endif
Fixes for OS2. Fix bug in isamlog Add argument types to function declarations.
2001-08-22 00:45:07 +02:00
Import changeset
2000-07-31 21:29:14 +02:00
#if defined(THREAD) && defined(HAVE_READDIR_R)
#define READDIR(A,B,C) ((errno=readdir_r(A,B,&C)) != 0 || !C)
#else
#define READDIR(A,B,C) (!(C=readdir(A)))
#endif
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
/*
We are assuming that directory we are reading is either has less than
100 files and so can be read in one initial chunk or has more than 1000
files and so big increment are suitable.
*/
#define ENTRIES_START_SIZE (8192/sizeof(FILEINFO))
#define ENTRIES_INCREMENT (65536/sizeof(FILEINFO))
#define NAMES_START_SIZE 32768
Import changeset
2000-07-31 21:29:14 +02:00
static int comp_names(struct fileinfo *a,struct fileinfo *b);
/* We need this because program don't know with malloc we used */
void my_dirend(MY_DIR *buffer)
{
DBUG_ENTER("my_dirend");
if (buffer)
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
{
delete_dynamic((DYNAMIC_ARRAY*)((char*)buffer +
ALIGN_SIZE(sizeof(MY_DIR))));
free_root((MEM_ROOT*)((char*)buffer + ALIGN_SIZE(sizeof(MY_DIR)) +
ALIGN_SIZE(sizeof(DYNAMIC_ARRAY))), MYF(0));
Bug#34043: Server loops excessively in _checkchunk() when safemalloc is enabled Essentially, the problem is that safemalloc is excruciatingly slow as it checks all allocated blocks for overrun at each memory management primitive, yielding a almost exponential slowdown for the memory management functions (malloc, realloc, free). The overrun check basically consists of verifying some bytes of a block for certain magic keys, which catches some simple forms of overrun. Another minor problem is violation of aliasing rules and that its own internal list of blocks is prone to corruption. Another issue with safemalloc is rather the maintenance cost as the tool has a significant impact on the server code. Given the magnitude of memory debuggers available nowadays, especially those that are provided with the platform malloc implementation, maintenance of a in-house and largely obsolete memory debugger becomes a burden that is not worth the effort due to its slowness and lack of support for detecting more common forms of heap corruption. Since there are third-party tools that can provide the same functionality at a lower or comparable performance cost, the solution is to simply remove safemalloc. Third-party tools can provide the same functionality at a lower or comparable performance cost. The removal of safemalloc also allows a simplification of the malloc wrappers, removing quite a bit of kludge: redefinition of my_malloc, my_free and the removal of the unused second argument of my_free. Since free() always check whether the supplied pointer is null, redudant checks are also removed. Also, this patch adds unit testing for my_malloc and moves my_realloc implementation into the same file as the other memory allocation primitives.
2010-07-08 23:20:08 +02:00
my_free(buffer);
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
}
Import changeset
2000-07-31 21:29:14 +02:00
DBUG_VOID_RETURN;
} /* my_dirend */
/* Compare in sort of filenames */
static int comp_names(struct fileinfo *a, struct fileinfo *b)
{
return (strcmp(a->name,b->name));
} /* comp_names */
This is the downport of Bug#24509 - 2048 file descriptor limit on windows needs increasing, also WL#3049 - improved Windows I/O The patch replaces the use of the POSIX I/O interfaces in mysys on Windows with the Win32 API calls (CreateFile, WriteFile, etc). The Windows HANDLE for the open file is stored in the my_file_info struct, along with a flag for append mode because the Windows API does not support opening files in append mode in all cases) The default max open files has been increased to 16384 and can be increased further by setting --max-open-files=<value> during the server start. Another major change in this patch that almost all Windows specific file IO code has been moved to a new file my_winfile.c, greatly reducing the amount of code in #ifdef blocks within mysys, thus improving readability. Minor enhancements: - my_(f)stat() is changed to use __stati64 structure with 64 file size and timestamps. It will return correct file size now (C runtime implementation used to report outdated information) - my_lock on Windows is prepared to handle additional timeout parameter - after review : changed __WIN__ to _WIN32 in the new and changed code.
2009-09-11 22:26:35 +02:00
#if !defined(_WIN32)
Import changeset
2000-07-31 21:29:14 +02:00
MY_DIR *my_dir(const char *path, myf MyFlags)
{
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
char *buffer;
MY_DIR *result= 0;
FILEINFO finfo;
DYNAMIC_ARRAY *dir_entries_storage;
MEM_ROOT *names_storage;
Import changeset
2000-07-31 21:29:14 +02:00
DIR *dirp;
struct dirent *dp;
char tmp_path[FN_REFLEN+1],*tmp_file;
#ifdef THREAD
char dirent_tmp[sizeof(struct dirent)+_POSIX_PATH_MAX+1];
#endif
DBUG_ENTER("my_dir");
Simple fixes (nothing notable)
2003-03-12 00:40:06 +01:00
DBUG_PRINT("my",("path: '%s' MyFlags: %d",path,MyFlags));
Import changeset
2000-07-31 21:29:14 +02:00
#if defined(THREAD) && !defined(HAVE_READDIR_R)
WL#2360 Performance schema Part III: mysys instrumentation
2009-12-10 04:19:51 +01:00
mysql_mutex_lock(&THR_LOCK_open);
Import changeset
2000-07-31 21:29:14 +02:00
#endif
WL#3817: Simplify string / memory area types and make things more consistent (first part) The following type conversions was done: - Changed byte to uchar - Changed gptr to uchar* - Change my_string to char * - Change my_size_t to size_t - Change size_s to size_t Removed declaration of byte, gptr, my_string, my_size_t and size_s. Following function parameter changes was done: - All string functions in mysys/strings was changed to use size_t instead of uint for string lengths. - All read()/write() functions changed to use size_t (including vio). - All protocoll functions changed to use size_t instead of uint - Functions that used a pointer to a string length was changed to use size_t* - Changed malloc(), free() and related functions from using gptr to use void * as this requires fewer casts in the code and is more in line with how the standard functions work. - Added extra length argument to dirname_part() to return the length of the created string. - Changed (at least) following functions to take uchar* as argument: - db_dump() - my_net_write() - net_write_command() - net_store_data() - DBUG_DUMP() - decimal2bin() & bin2decimal() - Changed my_compress() and my_uncompress() to use size_t. Changed one argument to my_uncompress() from a pointer to a value as we only return one value (makes function easier to use). - Changed type of 'pack_data' argument to packfrm() to avoid casts. - Changed in readfrm() and writefrom(), ha_discover and handler::discover() the type for argument 'frmdata' to uchar** to avoid casts. - Changed most Field functions to use uchar* instead of char* (reduced a lot of casts). - Changed field->val_xxx(xxx, new_ptr) to take const pointers. Other changes: - Removed a lot of not needed casts - Added a few new cast required by other changes - Added some cast to my_multi_malloc() arguments for safety (as string lengths needs to be uint, not size_t). - Fixed all calls to hash-get-key functions to use size_t*. (Needed to be done explicitely as this conflict was often hided by casting the function to hash_get_key). - Changed some buffers to memory regions to uchar* to avoid casts. - Changed some string lengths from uint to size_t. - Changed field->ptr to be uchar* instead of char*. This allowed us to get rid of a lot of casts. - Some changes from true -> TRUE, false -> FALSE, unsigned char -> uchar - Include zlib.h in some files as we needed declaration of crc32() - Changed MY_FILE_ERROR to be (size_t) -1. - Changed many variables to hold the result of my_read() / my_write() to be size_t. This was needed to properly detect errors (which are returned as (size_t) -1). - Removed some very old VMS code - Changed packfrm()/unpackfrm() to not be depending on uint size (portability fix) - Removed windows specific code to restore cursor position as this causes slowdown on windows and we should not mix read() and pread() calls anyway as this is not thread safe. Updated function comment to reflect this. Changed function that depended on original behavior of my_pwrite() to itself restore the cursor position (one such case). - Added some missing checking of return value of malloc(). - Changed definition of MOD_PAD_CHAR_TO_FULL_LENGTH to avoid 'long' overflow. - Changed type of table_def::m_size from my_size_t to ulong to reflect that m_size is the number of elements in the array, not a string/memory length. - Moved THD::max_row_length() to table.cc (as it's not depending on THD). Inlined max_row_length_blob() into this function. - More function comments - Fixed some compiler warnings when compiled without partitions. - Removed setting of LEX_STRING() arguments in declaration (portability fix). - Some trivial indentation/variable name changes. - Some trivial code simplifications: - Replaced some calls to alloc_root + memcpy to use strmake_root()/strdup_root(). - Changed some calls from memdup() to strmake() (Safety fix) - Simpler loops in client-simple.c
2007-05-10 11:59:39 +02:00
dirp = opendir(directory_file_name(tmp_path,(char *) path));
new crash-me results for Interbase.
2001-11-08 16:40:55 +01:00
#if defined(__amiga__)
if ((dirp->dd_fd) < 0) /* Directory doesn't exists */
goto error;
#endif
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
if (dirp == NULL ||
! (buffer= my_malloc(ALIGN_SIZE(sizeof(MY_DIR)) +
ALIGN_SIZE(sizeof(DYNAMIC_ARRAY)) +
sizeof(MEM_ROOT), MyFlags)))
goto error;
dir_entries_storage= (DYNAMIC_ARRAY*)(buffer + ALIGN_SIZE(sizeof(MY_DIR)));
names_storage= (MEM_ROOT*)(buffer + ALIGN_SIZE(sizeof(MY_DIR)) +
ALIGN_SIZE(sizeof(DYNAMIC_ARRAY)));
if (my_init_dynamic_array(dir_entries_storage, sizeof(FILEINFO),
ENTRIES_START_SIZE, ENTRIES_INCREMENT))
{
Bug#34043: Server loops excessively in _checkchunk() when safemalloc is enabled Essentially, the problem is that safemalloc is excruciatingly slow as it checks all allocated blocks for overrun at each memory management primitive, yielding a almost exponential slowdown for the memory management functions (malloc, realloc, free). The overrun check basically consists of verifying some bytes of a block for certain magic keys, which catches some simple forms of overrun. Another minor problem is violation of aliasing rules and that its own internal list of blocks is prone to corruption. Another issue with safemalloc is rather the maintenance cost as the tool has a significant impact on the server code. Given the magnitude of memory debuggers available nowadays, especially those that are provided with the platform malloc implementation, maintenance of a in-house and largely obsolete memory debugger becomes a burden that is not worth the effort due to its slowness and lack of support for detecting more common forms of heap corruption. Since there are third-party tools that can provide the same functionality at a lower or comparable performance cost, the solution is to simply remove safemalloc. Third-party tools can provide the same functionality at a lower or comparable performance cost. The removal of safemalloc also allows a simplification of the malloc wrappers, removing quite a bit of kludge: redefinition of my_malloc, my_free and the removal of the unused second argument of my_free. Since free() always check whether the supplied pointer is null, redudant checks are also removed. Also, this patch adds unit testing for my_malloc and moves my_realloc implementation into the same file as the other memory allocation primitives.
2010-07-08 23:20:08 +02:00
my_free(buffer);
Import changeset
2000-07-31 21:29:14 +02:00
goto error;
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
}
init_alloc_root(names_storage, NAMES_START_SIZE, NAMES_START_SIZE);
/* MY_DIR structure is allocated and completly initialized at this point */
result= (MY_DIR*)buffer;
Import changeset
2000-07-31 21:29:14 +02:00
tmp_file=strend(tmp_path);
#ifdef THREAD
dp= (struct dirent*) dirent_tmp;
#else
dp=0;
#endif
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
while (!(READDIR(dirp,(struct dirent*) dirent_tmp,dp)))
Import changeset
2000-07-31 21:29:14 +02:00
{
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
if (!(finfo.name= strdup_root(names_storage, dp->d_name)))
goto error;
if (MyFlags & MY_WANT_STAT)
Import changeset
2000-07-31 21:29:14 +02:00
{
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
if (!(finfo.mystat= (MY_STAT*)alloc_root(names_storage,
sizeof(MY_STAT))))
goto error;
bzero(finfo.mystat, sizeof(MY_STAT));
Backport of: ---------------------------------------------------------------------- ChangeSet@1.2571, 2008-04-08 12:30:06+02:00, vvaintroub@wva. +122 -0 Bug#32082 : definition of VOID in my_global.h conflicts with Windows SDK headers VOID macro is now removed. Its usage is replaced with void cast. In some cases, where cast does not make much sense (pthread_*, printf, hash_delete, my_seek), cast is ommited.
2009-11-24 14:54:59 +01:00
(void) strmov(tmp_file,dp->d_name);
(void) my_stat(tmp_path, finfo.mystat, MyFlags);
Fix for bug#15851 Unlistable directories yield no info from information_schema to skip directories which are not allowed for read
2006-01-24 12:59:08 +01:00
if (!(finfo.mystat->st_mode & MY_S_IREAD))
continue;
Import changeset
2000-07-31 21:29:14 +02:00
}
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
else
finfo.mystat= NULL;
WL#3817: Simplify string / memory area types and make things more consistent (first part) The following type conversions was done: - Changed byte to uchar - Changed gptr to uchar* - Change my_string to char * - Change my_size_t to size_t - Change size_s to size_t Removed declaration of byte, gptr, my_string, my_size_t and size_s. Following function parameter changes was done: - All string functions in mysys/strings was changed to use size_t instead of uint for string lengths. - All read()/write() functions changed to use size_t (including vio). - All protocoll functions changed to use size_t instead of uint - Functions that used a pointer to a string length was changed to use size_t* - Changed malloc(), free() and related functions from using gptr to use void * as this requires fewer casts in the code and is more in line with how the standard functions work. - Added extra length argument to dirname_part() to return the length of the created string. - Changed (at least) following functions to take uchar* as argument: - db_dump() - my_net_write() - net_write_command() - net_store_data() - DBUG_DUMP() - decimal2bin() & bin2decimal() - Changed my_compress() and my_uncompress() to use size_t. Changed one argument to my_uncompress() from a pointer to a value as we only return one value (makes function easier to use). - Changed type of 'pack_data' argument to packfrm() to avoid casts. - Changed in readfrm() and writefrom(), ha_discover and handler::discover() the type for argument 'frmdata' to uchar** to avoid casts. - Changed most Field functions to use uchar* instead of char* (reduced a lot of casts). - Changed field->val_xxx(xxx, new_ptr) to take const pointers. Other changes: - Removed a lot of not needed casts - Added a few new cast required by other changes - Added some cast to my_multi_malloc() arguments for safety (as string lengths needs to be uint, not size_t). - Fixed all calls to hash-get-key functions to use size_t*. (Needed to be done explicitely as this conflict was often hided by casting the function to hash_get_key). - Changed some buffers to memory regions to uchar* to avoid casts. - Changed some string lengths from uint to size_t. - Changed field->ptr to be uchar* instead of char*. This allowed us to get rid of a lot of casts. - Some changes from true -> TRUE, false -> FALSE, unsigned char -> uchar - Include zlib.h in some files as we needed declaration of crc32() - Changed MY_FILE_ERROR to be (size_t) -1. - Changed many variables to hold the result of my_read() / my_write() to be size_t. This was needed to properly detect errors (which are returned as (size_t) -1). - Removed some very old VMS code - Changed packfrm()/unpackfrm() to not be depending on uint size (portability fix) - Removed windows specific code to restore cursor position as this causes slowdown on windows and we should not mix read() and pread() calls anyway as this is not thread safe. Updated function comment to reflect this. Changed function that depended on original behavior of my_pwrite() to itself restore the cursor position (one such case). - Added some missing checking of return value of malloc(). - Changed definition of MOD_PAD_CHAR_TO_FULL_LENGTH to avoid 'long' overflow. - Changed type of table_def::m_size from my_size_t to ulong to reflect that m_size is the number of elements in the array, not a string/memory length. - Moved THD::max_row_length() to table.cc (as it's not depending on THD). Inlined max_row_length_blob() into this function. - More function comments - Fixed some compiler warnings when compiled without partitions. - Removed setting of LEX_STRING() arguments in declaration (portability fix). - Some trivial indentation/variable name changes. - Some trivial code simplifications: - Replaced some calls to alloc_root + memcpy to use strmake_root()/strdup_root(). - Changed some calls from memdup() to strmake() (Safety fix) - Simpler loops in client-simple.c
2007-05-10 11:59:39 +02:00
if (push_dynamic(dir_entries_storage, (uchar*)&finfo))
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
goto error;
Import changeset
2000-07-31 21:29:14 +02:00
}
(void) closedir(dirp);
#if defined(THREAD) && !defined(HAVE_READDIR_R)
WL#2360 Performance schema Part III: mysys instrumentation
2009-12-10 04:19:51 +01:00
mysql_mutex_unlock(&THR_LOCK_open);
Import changeset
2000-07-31 21:29:14 +02:00
#endif
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
result->dir_entry= (FILEINFO *)dir_entries_storage->buffer;
result->number_off_files= dir_entries_storage->elements;
if (!(MyFlags & MY_DONT_SORT))
Fix for bug #31207: Test "join_nested" shows different strategy on IA64 CPUs / Intel's ICC compile The bug is a combination of two problems: 1. IA64/ICC MySQL binaries use glibc's qsort(), not the one in mysys. 2. The order relation implemented by join_tab_cmp() is not transitive, i.e. it is possible to choose such a, b and c that (a < b) && (b < c) but (c < a). This implies that result of a sort using the relation implemented by join_tab_cmp() depends on the order in which elements are compared, i.e. the result is implementation-specific. Since choose_plan() uses qsort() to pre-sort the join tables using join_tab_cmp() as a compare function, the results of the sorting may vary depending on qsort() implementation. It is neither possible nor important to implement a better ordering algorithm in join_tab_cmp(). Therefore the only way to fix it is to force our own qsort() to be used by renaming it to my_qsort(), so we don't depend on linker to decide that. This patch also "fixes" bug #20530: qsort redefinition violates the standard.
2007-10-17 18:08:58 +02:00
my_qsort((void *) result->dir_entry, result->number_off_files,
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
sizeof(FILEINFO), (qsort_cmp) comp_names);
DBUG_RETURN(result);
Import changeset
2000-07-31 21:29:14 +02:00
error:
#if defined(THREAD) && !defined(HAVE_READDIR_R)
WL#2360 Performance schema Part III: mysys instrumentation
2009-12-10 04:19:51 +01:00
mysql_mutex_unlock(&THR_LOCK_open);
Import changeset
2000-07-31 21:29:14 +02:00
#endif
my_errno=errno;
if (dirp)
(void) closedir(dirp);
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
my_dirend(result);
Fixed bug in UNION Fixed replication bug in load_master_data
2001-09-14 18:50:56 +02:00
if (MyFlags & (MY_FAE | MY_WME))
Import changeset
2000-07-31 21:29:14 +02:00
my_error(EE_DIR,MYF(ME_BELL+ME_WAITTANG),path,my_errno);
DBUG_RETURN((MY_DIR *) NULL);
} /* my_dir */
/*
* Convert from directory name to filename.
* On UNIX, it's simple: just make sure there is a terminating /
* Returns pointer to dst;
*/
WL#3817: Simplify string / memory area types and make things more consistent (first part) The following type conversions was done: - Changed byte to uchar - Changed gptr to uchar* - Change my_string to char * - Change my_size_t to size_t - Change size_s to size_t Removed declaration of byte, gptr, my_string, my_size_t and size_s. Following function parameter changes was done: - All string functions in mysys/strings was changed to use size_t instead of uint for string lengths. - All read()/write() functions changed to use size_t (including vio). - All protocoll functions changed to use size_t instead of uint - Functions that used a pointer to a string length was changed to use size_t* - Changed malloc(), free() and related functions from using gptr to use void * as this requires fewer casts in the code and is more in line with how the standard functions work. - Added extra length argument to dirname_part() to return the length of the created string. - Changed (at least) following functions to take uchar* as argument: - db_dump() - my_net_write() - net_write_command() - net_store_data() - DBUG_DUMP() - decimal2bin() & bin2decimal() - Changed my_compress() and my_uncompress() to use size_t. Changed one argument to my_uncompress() from a pointer to a value as we only return one value (makes function easier to use). - Changed type of 'pack_data' argument to packfrm() to avoid casts. - Changed in readfrm() and writefrom(), ha_discover and handler::discover() the type for argument 'frmdata' to uchar** to avoid casts. - Changed most Field functions to use uchar* instead of char* (reduced a lot of casts). - Changed field->val_xxx(xxx, new_ptr) to take const pointers. Other changes: - Removed a lot of not needed casts - Added a few new cast required by other changes - Added some cast to my_multi_malloc() arguments for safety (as string lengths needs to be uint, not size_t). - Fixed all calls to hash-get-key functions to use size_t*. (Needed to be done explicitely as this conflict was often hided by casting the function to hash_get_key). - Changed some buffers to memory regions to uchar* to avoid casts. - Changed some string lengths from uint to size_t. - Changed field->ptr to be uchar* instead of char*. This allowed us to get rid of a lot of casts. - Some changes from true -> TRUE, false -> FALSE, unsigned char -> uchar - Include zlib.h in some files as we needed declaration of crc32() - Changed MY_FILE_ERROR to be (size_t) -1. - Changed many variables to hold the result of my_read() / my_write() to be size_t. This was needed to properly detect errors (which are returned as (size_t) -1). - Removed some very old VMS code - Changed packfrm()/unpackfrm() to not be depending on uint size (portability fix) - Removed windows specific code to restore cursor position as this causes slowdown on windows and we should not mix read() and pread() calls anyway as this is not thread safe. Updated function comment to reflect this. Changed function that depended on original behavior of my_pwrite() to itself restore the cursor position (one such case). - Added some missing checking of return value of malloc(). - Changed definition of MOD_PAD_CHAR_TO_FULL_LENGTH to avoid 'long' overflow. - Changed type of table_def::m_size from my_size_t to ulong to reflect that m_size is the number of elements in the array, not a string/memory length. - Moved THD::max_row_length() to table.cc (as it's not depending on THD). Inlined max_row_length_blob() into this function. - More function comments - Fixed some compiler warnings when compiled without partitions. - Removed setting of LEX_STRING() arguments in declaration (portability fix). - Some trivial indentation/variable name changes. - Some trivial code simplifications: - Replaced some calls to alloc_root + memcpy to use strmake_root()/strdup_root(). - Changed some calls from memdup() to strmake() (Safety fix) - Simpler loops in client-simple.c
2007-05-10 11:59:39 +02:00
char * directory_file_name (char * dst, const char *src)
Import changeset
2000-07-31 21:29:14 +02:00
{
/* Process as Unix format: just remove test the final slash. */
WL#5486: Remove code for unsupported platforms Remove VMS specific code.
2010-07-15 13:26:38 +02:00
char *end;
Import changeset
2000-07-31 21:29:14 +02:00
if (src[0] == 0)
src= (char*) "."; /* Use empty as current */
end=strmov(dst, src);
if (end[-1] != FN_LIBCHAR)
{
end[0]=FN_LIBCHAR; /* Add last '/' */
end[1]='\0';
}
return dst;
WL#5486: Remove code for unsupported platforms Remove VMS specific code.
2010-07-15 13:26:38 +02:00
}
Import changeset
2000-07-31 21:29:14 +02:00
removing MSDOS defines and code
2006-05-08 21:50:13 +02:00
#else
Import changeset
2000-07-31 21:29:14 +02:00
/*
*****************************************************************************
** Read long filename using windows rutines
*****************************************************************************
*/
Fixes for OS2. Fix bug in isamlog Add argument types to function declarations.
2001-08-22 00:45:07 +02:00
MY_DIR *my_dir(const char *path, myf MyFlags)
Import changeset
2000-07-31 21:29:14 +02:00
{
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
char *buffer;
MY_DIR *result= 0;
FILEINFO finfo;
DYNAMIC_ARRAY *dir_entries_storage;
MEM_ROOT *names_storage;
Import changeset
2000-07-31 21:29:14 +02:00
#ifdef __BORLANDC__
struct ffblk find;
#else
struct _finddata_t find;
#endif
ushort mode;
char tmp_path[FN_REFLEN],*tmp_file,attrib;
Fixed for Ia64 + delayed key creation + a lot of small bug fixes
2000-08-15 19:09:37 +02:00
#ifdef _WIN64
__int64 handle;
#else
Import changeset
2000-07-31 21:29:14 +02:00
long handle;
Fixed for Ia64 + delayed key creation + a lot of small bug fixes
2000-08-15 19:09:37 +02:00
#endif
Import changeset
2000-07-31 21:29:14 +02:00
DBUG_ENTER("my_dir");
DBUG_PRINT("my",("path: '%s' stat: %d MyFlags: %d",path,MyFlags));
/* Put LIB-CHAR as last path-character if not there */
tmp_file=tmp_path;
if (!*path)
*tmp_file++ ='.'; /* From current dir */
Bug#18474 Unlistable directories yield no info from information_schema, part2 - Improved solution by adding an else stetment so that do find next is avoided if erorr occurs, but we still return zero files found instaed of an error
2006-04-11 20:41:05 +02:00
tmp_file= strnmov(tmp_file, path, FN_REFLEN-5);
Import changeset
2000-07-31 21:29:14 +02:00
if (tmp_file[-1] == FN_DEVCHAR)
*tmp_file++= '.'; /* From current dev-dir */
if (tmp_file[-1] != FN_LIBCHAR)
*tmp_file++ =FN_LIBCHAR;
removing MSDOS defines and code
2006-05-08 21:50:13 +02:00
tmp_file[0]='*'; /* Windows needs this !??? */
Import changeset
2000-07-31 21:29:14 +02:00
tmp_file[1]='.';
tmp_file[2]='*';
tmp_file[3]='\0';
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
if (!(buffer= my_malloc(ALIGN_SIZE(sizeof(MY_DIR)) +
ALIGN_SIZE(sizeof(DYNAMIC_ARRAY)) +
sizeof(MEM_ROOT), MyFlags)))
Import changeset
2000-07-31 21:29:14 +02:00
goto error;
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
dir_entries_storage= (DYNAMIC_ARRAY*)(buffer + ALIGN_SIZE(sizeof(MY_DIR)));
names_storage= (MEM_ROOT*)(buffer + ALIGN_SIZE(sizeof(MY_DIR)) +
ALIGN_SIZE(sizeof(DYNAMIC_ARRAY)));
if (my_init_dynamic_array(dir_entries_storage, sizeof(FILEINFO),
ENTRIES_START_SIZE, ENTRIES_INCREMENT))
Import changeset
2000-07-31 21:29:14 +02:00
{
Bug#34043: Server loops excessively in _checkchunk() when safemalloc is enabled Essentially, the problem is that safemalloc is excruciatingly slow as it checks all allocated blocks for overrun at each memory management primitive, yielding a almost exponential slowdown for the memory management functions (malloc, realloc, free). The overrun check basically consists of verifying some bytes of a block for certain magic keys, which catches some simple forms of overrun. Another minor problem is violation of aliasing rules and that its own internal list of blocks is prone to corruption. Another issue with safemalloc is rather the maintenance cost as the tool has a significant impact on the server code. Given the magnitude of memory debuggers available nowadays, especially those that are provided with the platform malloc implementation, maintenance of a in-house and largely obsolete memory debugger becomes a burden that is not worth the effort due to its slowness and lack of support for detecting more common forms of heap corruption. Since there are third-party tools that can provide the same functionality at a lower or comparable performance cost, the solution is to simply remove safemalloc. Third-party tools can provide the same functionality at a lower or comparable performance cost. The removal of safemalloc also allows a simplification of the malloc wrappers, removing quite a bit of kludge: redefinition of my_malloc, my_free and the removal of the unused second argument of my_free. Since free() always check whether the supplied pointer is null, redudant checks are also removed. Also, this patch adds unit testing for my_malloc and moves my_realloc implementation into the same file as the other memory allocation primitives.
2010-07-08 23:20:08 +02:00
my_free(buffer);
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
goto error;
}
init_alloc_root(names_storage, NAMES_START_SIZE, NAMES_START_SIZE);
/* MY_DIR structure is allocated and completly initialized at this point */
result= (MY_DIR*)buffer;
Bug#18474 Unlistable directories yield no info from information_schema, part2 - Make the windows implementation of 'my_dir' behave like the default implementation
2006-03-23 23:09:34 +01:00
#ifdef __BORLANDC__
if ((handle= findfirst(tmp_path,&find,0)) == -1L)
#else
if ((handle=_findfirst(tmp_path,&find)) == -1L)
#endif
{
Bug#18474 Unlistable directories yield no info from information_schema, part2 - Improved solution by adding an else stetment so that do find next is avoided if erorr occurs, but we still return zero files found instaed of an error
2006-04-11 20:41:05 +02:00
DBUG_PRINT("info", ("findfirst returned error, errno: %d", errno));
Bug#18474 Unlistable directories yield no info from information_schema, part2 - Make the windows implementation of 'my_dir' behave like the default implementation
2006-03-23 23:09:34 +01:00
if (errno != EINVAL)
goto error;
/*
Could not read the directory, no read access.
Probably because by "chmod -r".
continue and return zero files in dir
*/
}
Bug#18474 Unlistable directories yield no info from information_schema, part2 - Improved solution by adding an else stetment so that do find next is avoided if erorr occurs, but we still return zero files found instaed of an error
2006-04-11 20:41:05 +02:00
else
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
{
Bug#18474 Unlistable directories yield no info from information_schema, part2 - Improved solution by adding an else stetment so that do find next is avoided if erorr occurs, but we still return zero files found instaed of an error
2006-04-11 20:41:05 +02:00
do
{
Fix for BUG#4375: Windows specific directories are copied during replication. Modified my_dir(). Now this function skips hidden and system files which sometimes are created by Windows. NOTE. The fix is similar to the previuos one (05 July 2004) except for correct setting of the 'attrib' variable value (within the previous fix this variable was left uninitialized when my_dir() was called with My_flags & MY_WANT_STAT == 0).
2005-10-19 16:01:23 +02:00
#ifdef __BORLANDC__
Bug#18474 Unlistable directories yield no info from information_schema, part2 - Improved solution by adding an else stetment so that do find next is avoided if erorr occurs, but we still return zero files found instaed of an error
2006-04-11 20:41:05 +02:00
attrib= find.ff_attrib;
Fix for BUG#4375: Windows specific directories are copied during replication. Modified my_dir(). Now this function skips hidden and system files which sometimes are created by Windows. NOTE. The fix is similar to the previuos one (05 July 2004) except for correct setting of the 'attrib' variable value (within the previous fix this variable was left uninitialized when my_dir() was called with My_flags & MY_WANT_STAT == 0).
2005-10-19 16:01:23 +02:00
#else
Bug#18474 Unlistable directories yield no info from information_schema, part2 - Improved solution by adding an else stetment so that do find next is avoided if erorr occurs, but we still return zero files found instaed of an error
2006-04-11 20:41:05 +02:00
attrib= find.attrib;
/*
Do not show hidden and system files which Windows sometimes create.
Note. Because Borland's findfirst() is called with the third
argument = 0 hidden/system files are excluded from the search.
*/
if (attrib & (_A_HIDDEN | _A_SYSTEM))
continue;
#endif
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
#ifdef __BORLANDC__
Bug#18474 Unlistable directories yield no info from information_schema, part2 - Improved solution by adding an else stetment so that do find next is avoided if erorr occurs, but we still return zero files found instaed of an error
2006-04-11 20:41:05 +02:00
if (!(finfo.name= strdup_root(names_storage, find.ff_name)))
goto error;
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
#else
Bug#18474 Unlistable directories yield no info from information_schema, part2 - Improved solution by adding an else stetment so that do find next is avoided if erorr occurs, but we still return zero files found instaed of an error
2006-04-11 20:41:05 +02:00
if (!(finfo.name= strdup_root(names_storage, find.name)))
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
goto error;
Bug#18474 Unlistable directories yield no info from information_schema, part2 - Improved solution by adding an else stetment so that do find next is avoided if erorr occurs, but we still return zero files found instaed of an error
2006-04-11 20:41:05 +02:00
#endif
if (MyFlags & MY_WANT_STAT)
{
if (!(finfo.mystat= (MY_STAT*)alloc_root(names_storage,
sizeof(MY_STAT))))
goto error;
bzero(finfo.mystat, sizeof(MY_STAT));
Import changeset
2000-07-31 21:29:14 +02:00
#ifdef __BORLANDC__
Bug#18474 Unlistable directories yield no info from information_schema, part2 - Improved solution by adding an else stetment so that do find next is avoided if erorr occurs, but we still return zero files found instaed of an error
2006-04-11 20:41:05 +02:00
finfo.mystat->st_size=find.ff_fsize;
Import changeset
2000-07-31 21:29:14 +02:00
#else
Bug#18474 Unlistable directories yield no info from information_schema, part2 - Improved solution by adding an else stetment so that do find next is avoided if erorr occurs, but we still return zero files found instaed of an error
2006-04-11 20:41:05 +02:00
finfo.mystat->st_size=find.size;
Import changeset
2000-07-31 21:29:14 +02:00
#endif
Bug#18474 Unlistable directories yield no info from information_schema, part2 - Improved solution by adding an else stetment so that do find next is avoided if erorr occurs, but we still return zero files found instaed of an error
2006-04-11 20:41:05 +02:00
mode= MY_S_IREAD;
if (!(attrib & _A_RDONLY))
mode|= MY_S_IWRITE;
if (attrib & _A_SUBDIR)
mode|= MY_S_IFDIR;
finfo.mystat->st_mode= mode;
Import changeset
2000-07-31 21:29:14 +02:00
#ifdef __BORLANDC__
Bug#18474 Unlistable directories yield no info from information_schema, part2 - Improved solution by adding an else stetment so that do find next is avoided if erorr occurs, but we still return zero files found instaed of an error
2006-04-11 20:41:05 +02:00
finfo.mystat->st_mtime= ((uint32) find.ff_ftime);
Import changeset
2000-07-31 21:29:14 +02:00
#else
Bug#18474 Unlistable directories yield no info from information_schema, part2 - Improved solution by adding an else stetment so that do find next is avoided if erorr occurs, but we still return zero files found instaed of an error
2006-04-11 20:41:05 +02:00
finfo.mystat->st_mtime= ((uint32) find.time_write);
Import changeset
2000-07-31 21:29:14 +02:00
#endif
Bug#18474 Unlistable directories yield no info from information_schema, part2 - Improved solution by adding an else stetment so that do find next is avoided if erorr occurs, but we still return zero files found instaed of an error
2006-04-11 20:41:05 +02:00
}
else
finfo.mystat= NULL;
Import changeset
2000-07-31 21:29:14 +02:00
WL#3817: Simplify string / memory area types and make things more consistent (first part) The following type conversions was done: - Changed byte to uchar - Changed gptr to uchar* - Change my_string to char * - Change my_size_t to size_t - Change size_s to size_t Removed declaration of byte, gptr, my_string, my_size_t and size_s. Following function parameter changes was done: - All string functions in mysys/strings was changed to use size_t instead of uint for string lengths. - All read()/write() functions changed to use size_t (including vio). - All protocoll functions changed to use size_t instead of uint - Functions that used a pointer to a string length was changed to use size_t* - Changed malloc(), free() and related functions from using gptr to use void * as this requires fewer casts in the code and is more in line with how the standard functions work. - Added extra length argument to dirname_part() to return the length of the created string. - Changed (at least) following functions to take uchar* as argument: - db_dump() - my_net_write() - net_write_command() - net_store_data() - DBUG_DUMP() - decimal2bin() & bin2decimal() - Changed my_compress() and my_uncompress() to use size_t. Changed one argument to my_uncompress() from a pointer to a value as we only return one value (makes function easier to use). - Changed type of 'pack_data' argument to packfrm() to avoid casts. - Changed in readfrm() and writefrom(), ha_discover and handler::discover() the type for argument 'frmdata' to uchar** to avoid casts. - Changed most Field functions to use uchar* instead of char* (reduced a lot of casts). - Changed field->val_xxx(xxx, new_ptr) to take const pointers. Other changes: - Removed a lot of not needed casts - Added a few new cast required by other changes - Added some cast to my_multi_malloc() arguments for safety (as string lengths needs to be uint, not size_t). - Fixed all calls to hash-get-key functions to use size_t*. (Needed to be done explicitely as this conflict was often hided by casting the function to hash_get_key). - Changed some buffers to memory regions to uchar* to avoid casts. - Changed some string lengths from uint to size_t. - Changed field->ptr to be uchar* instead of char*. This allowed us to get rid of a lot of casts. - Some changes from true -> TRUE, false -> FALSE, unsigned char -> uchar - Include zlib.h in some files as we needed declaration of crc32() - Changed MY_FILE_ERROR to be (size_t) -1. - Changed many variables to hold the result of my_read() / my_write() to be size_t. This was needed to properly detect errors (which are returned as (size_t) -1). - Removed some very old VMS code - Changed packfrm()/unpackfrm() to not be depending on uint size (portability fix) - Removed windows specific code to restore cursor position as this causes slowdown on windows and we should not mix read() and pread() calls anyway as this is not thread safe. Updated function comment to reflect this. Changed function that depended on original behavior of my_pwrite() to itself restore the cursor position (one such case). - Added some missing checking of return value of malloc(). - Changed definition of MOD_PAD_CHAR_TO_FULL_LENGTH to avoid 'long' overflow. - Changed type of table_def::m_size from my_size_t to ulong to reflect that m_size is the number of elements in the array, not a string/memory length. - Moved THD::max_row_length() to table.cc (as it's not depending on THD). Inlined max_row_length_blob() into this function. - More function comments - Fixed some compiler warnings when compiled without partitions. - Removed setting of LEX_STRING() arguments in declaration (portability fix). - Some trivial indentation/variable name changes. - Some trivial code simplifications: - Replaced some calls to alloc_root + memcpy to use strmake_root()/strdup_root(). - Changed some calls from memdup() to strmake() (Safety fix) - Simpler loops in client-simple.c
2007-05-10 11:59:39 +02:00
if (push_dynamic(dir_entries_storage, (uchar*)&finfo))
Bug#18474 Unlistable directories yield no info from information_schema, part2 - Improved solution by adding an else stetment so that do find next is avoided if erorr occurs, but we still return zero files found instaed of an error
2006-04-11 20:41:05 +02:00
goto error;
}
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
#ifdef __BORLANDC__
Bug#18474 Unlistable directories yield no info from information_schema, part2 - Improved solution by adding an else stetment so that do find next is avoided if erorr occurs, but we still return zero files found instaed of an error
2006-04-11 20:41:05 +02:00
while (findnext(&find) == 0);
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
#else
Bug#18474 Unlistable directories yield no info from information_schema, part2 - Improved solution by adding an else stetment so that do find next is avoided if erorr occurs, but we still return zero files found instaed of an error
2006-04-11 20:41:05 +02:00
while (_findnext(handle,&find) == 0);
_findclose(handle);
Import changeset
2000-07-31 21:29:14 +02:00
#endif
Bug#18474 Unlistable directories yield no info from information_schema, part2 - Improved solution by adding an else stetment so that do find next is avoided if erorr occurs, but we still return zero files found instaed of an error
2006-04-11 20:41:05 +02:00
}
Import changeset
2000-07-31 21:29:14 +02:00
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
result->dir_entry= (FILEINFO *)dir_entries_storage->buffer;
result->number_off_files= dir_entries_storage->elements;
Bug#18474 Unlistable directories yield no info from information_schema, part2 - Improved solution by adding an else stetment so that do find next is avoided if erorr occurs, but we still return zero files found instaed of an error
2006-04-11 20:41:05 +02:00
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
if (!(MyFlags & MY_DONT_SORT))
Fix for bug #31207: Test "join_nested" shows different strategy on IA64 CPUs / Intel's ICC compile The bug is a combination of two problems: 1. IA64/ICC MySQL binaries use glibc's qsort(), not the one in mysys. 2. The order relation implemented by join_tab_cmp() is not transitive, i.e. it is possible to choose such a, b and c that (a < b) && (b < c) but (c < a). This implies that result of a sort using the relation implemented by join_tab_cmp() depends on the order in which elements are compared, i.e. the result is implementation-specific. Since choose_plan() uses qsort() to pre-sort the join tables using join_tab_cmp() as a compare function, the results of the sorting may vary depending on qsort() implementation. It is neither possible nor important to implement a better ordering algorithm in join_tab_cmp(). Therefore the only way to fix it is to force our own qsort() to be used by renaming it to my_qsort(), so we don't depend on linker to decide that. This patch also "fixes" bug #20530: qsort redefinition violates the standard.
2007-10-17 18:08:58 +02:00
my_qsort((void *) result->dir_entry, result->number_off_files,
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
sizeof(FILEINFO), (qsort_cmp) comp_names);
BUG#20528 Added missing parenthesis to DBUG_PRINT statement
2006-06-23 17:21:57 +02:00
DBUG_PRINT("exit", ("found %d files", result->number_off_files));
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
DBUG_RETURN(result);
Import changeset
2000-07-31 21:29:14 +02:00
error:
my_errno=errno;
#ifndef __BORLANDC__
if (handle != -1)
_findclose(handle);
#endif
Fix for Bug #1952 "SHOW TABLE STATUS very slow w/large number of tables" Replaced old algorithm which were used in my_dir() and stored all information about directory entries in one chunk of memory with new one which stores file names and MY_STAT structures in separate memroot, so now we don't need to copy this data during reallocation of dir_entry array.
2003-12-12 01:39:29 +01:00
my_dirend(result);
Import changeset
2000-07-31 21:29:14 +02:00
if (MyFlags & MY_FAE+MY_WME)
my_error(EE_DIR,MYF(ME_BELL+ME_WAITTANG),path,errno);
DBUG_RETURN((MY_DIR *) NULL);
} /* my_dir */
This is the downport of Bug#24509 - 2048 file descriptor limit on windows needs increasing, also WL#3049 - improved Windows I/O The patch replaces the use of the POSIX I/O interfaces in mysys on Windows with the Win32 API calls (CreateFile, WriteFile, etc). The Windows HANDLE for the open file is stored in the my_file_info struct, along with a flag for append mode because the Windows API does not support opening files in append mode in all cases) The default max open files has been increased to 16384 and can be increased further by setting --max-open-files=<value> during the server start. Another major change in this patch that almost all Windows specific file IO code has been moved to a new file my_winfile.c, greatly reducing the amount of code in #ifdef blocks within mysys, thus improving readability. Minor enhancements: - my_(f)stat() is changed to use __stati64 structure with 64 file size and timestamps. It will return correct file size now (C runtime implementation used to report outdated information) - my_lock on Windows is prepared to handle additional timeout parameter - after review : changed __WIN__ to _WIN32 in the new and changed code.
2009-09-11 22:26:35 +02:00
#endif /* _WIN32 */
Import changeset
2000-07-31 21:29:14 +02:00
/****************************************************************************
** File status
** Note that MY_STAT is assumed to be same as struct stat
****************************************************************************/
This is the downport of Bug#24509 - 2048 file descriptor limit on windows needs increasing, also WL#3049 - improved Windows I/O The patch replaces the use of the POSIX I/O interfaces in mysys on Windows with the Win32 API calls (CreateFile, WriteFile, etc). The Windows HANDLE for the open file is stored in the my_file_info struct, along with a flag for append mode because the Windows API does not support opening files in append mode in all cases) The default max open files has been increased to 16384 and can be increased further by setting --max-open-files=<value> during the server start. Another major change in this patch that almost all Windows specific file IO code has been moved to a new file my_winfile.c, greatly reducing the amount of code in #ifdef blocks within mysys, thus improving readability. Minor enhancements: - my_(f)stat() is changed to use __stati64 structure with 64 file size and timestamps. It will return correct file size now (C runtime implementation used to report outdated information) - my_lock on Windows is prepared to handle additional timeout parameter - after review : changed __WIN__ to _WIN32 in the new and changed code.
2009-09-11 22:26:35 +02:00
int my_fstat(File Filedes, MY_STAT *stat_area,
Fix of UNION code Added heap_delete_table Added HA_EXTRA_PREPARE_FOR_DELETE Added and use my_dup() for faster open of tables. Removed not working no-mix-table-type
2001-08-19 13:43:51 +02:00
myf MyFlags __attribute__((unused)))
Import changeset
2000-07-31 21:29:14 +02:00
{
DBUG_ENTER("my_fstat");
Merge ua141d10.elisa.omakaista.fi:/home/my/bk/mysql-4.1-main into ua141d10.elisa.omakaista.fi:/home/my/bk/mysql-5.0-marvel
2007-03-28 19:46:42 +02:00
DBUG_PRINT("my",("fd: %d MyFlags: %d", Filedes, MyFlags));
This is the downport of Bug#24509 - 2048 file descriptor limit on windows needs increasing, also WL#3049 - improved Windows I/O The patch replaces the use of the POSIX I/O interfaces in mysys on Windows with the Win32 API calls (CreateFile, WriteFile, etc). The Windows HANDLE for the open file is stored in the my_file_info struct, along with a flag for append mode because the Windows API does not support opening files in append mode in all cases) The default max open files has been increased to 16384 and can be increased further by setting --max-open-files=<value> during the server start. Another major change in this patch that almost all Windows specific file IO code has been moved to a new file my_winfile.c, greatly reducing the amount of code in #ifdef blocks within mysys, thus improving readability. Minor enhancements: - my_(f)stat() is changed to use __stati64 structure with 64 file size and timestamps. It will return correct file size now (C runtime implementation used to report outdated information) - my_lock on Windows is prepared to handle additional timeout parameter - after review : changed __WIN__ to _WIN32 in the new and changed code.
2009-09-11 22:26:35 +02:00
#ifdef _WIN32
DBUG_RETURN(my_win_fstat(Filedes, stat_area));
#else
Import changeset
2000-07-31 21:29:14 +02:00
DBUG_RETURN(fstat(Filedes, (struct stat *) stat_area));
This is the downport of Bug#24509 - 2048 file descriptor limit on windows needs increasing, also WL#3049 - improved Windows I/O The patch replaces the use of the POSIX I/O interfaces in mysys on Windows with the Win32 API calls (CreateFile, WriteFile, etc). The Windows HANDLE for the open file is stored in the my_file_info struct, along with a flag for append mode because the Windows API does not support opening files in append mode in all cases) The default max open files has been increased to 16384 and can be increased further by setting --max-open-files=<value> during the server start. Another major change in this patch that almost all Windows specific file IO code has been moved to a new file my_winfile.c, greatly reducing the amount of code in #ifdef blocks within mysys, thus improving readability. Minor enhancements: - my_(f)stat() is changed to use __stati64 structure with 64 file size and timestamps. It will return correct file size now (C runtime implementation used to report outdated information) - my_lock on Windows is prepared to handle additional timeout parameter - after review : changed __WIN__ to _WIN32 in the new and changed code.
2009-09-11 22:26:35 +02:00
#endif
Import changeset
2000-07-31 21:29:14 +02:00
}
Fix of UNION code Added heap_delete_table Added HA_EXTRA_PREPARE_FOR_DELETE Added and use my_dup() for faster open of tables. Removed not working no-mix-table-type
2001-08-19 13:43:51 +02:00
Import changeset
2000-07-31 21:29:14 +02:00
MY_STAT *my_stat(const char *path, MY_STAT *stat_area, myf my_flags)
{
int m_used;
DBUG_ENTER("my_stat");
Merge ua141d10.elisa.omakaista.fi:/home/my/bk/mysql-4.1-main into ua141d10.elisa.omakaista.fi:/home/my/bk/mysql-5.0-marvel
2007-03-28 19:46:42 +02:00
DBUG_PRINT("my", ("path: '%s' stat_area: 0x%lx MyFlags: %d", path,
Remove compiler warnings (Mostly in DBUG_PRINT() and unused arguments) Fixed bug in query cache when used with traceing (--with-debug) Fixed memory leak in mysqldump Removed warnings from mysqltest scripts (replaced -- with #)
2006-11-20 21:42:06 +01:00
(long) stat_area, my_flags));
Import changeset
2000-07-31 21:29:14 +02:00
if ((m_used= (stat_area == NULL)))
This is the downport of Bug#24509 - 2048 file descriptor limit on windows needs increasing, also WL#3049 - improved Windows I/O The patch replaces the use of the POSIX I/O interfaces in mysys on Windows with the Win32 API calls (CreateFile, WriteFile, etc). The Windows HANDLE for the open file is stored in the my_file_info struct, along with a flag for append mode because the Windows API does not support opening files in append mode in all cases) The default max open files has been increased to 16384 and can be increased further by setting --max-open-files=<value> during the server start. Another major change in this patch that almost all Windows specific file IO code has been moved to a new file my_winfile.c, greatly reducing the amount of code in #ifdef blocks within mysys, thus improving readability. Minor enhancements: - my_(f)stat() is changed to use __stati64 structure with 64 file size and timestamps. It will return correct file size now (C runtime implementation used to report outdated information) - my_lock on Windows is prepared to handle additional timeout parameter - after review : changed __WIN__ to _WIN32 in the new and changed code.
2009-09-11 22:26:35 +02:00
if (!(stat_area= (MY_STAT *) my_malloc(sizeof(MY_STAT), my_flags)))
Import changeset
2000-07-31 21:29:14 +02:00
goto error;
This is the downport of Bug#24509 - 2048 file descriptor limit on windows needs increasing, also WL#3049 - improved Windows I/O The patch replaces the use of the POSIX I/O interfaces in mysys on Windows with the Win32 API calls (CreateFile, WriteFile, etc). The Windows HANDLE for the open file is stored in the my_file_info struct, along with a flag for append mode because the Windows API does not support opening files in append mode in all cases) The default max open files has been increased to 16384 and can be increased further by setting --max-open-files=<value> during the server start. Another major change in this patch that almost all Windows specific file IO code has been moved to a new file my_winfile.c, greatly reducing the amount of code in #ifdef blocks within mysys, thus improving readability. Minor enhancements: - my_(f)stat() is changed to use __stati64 structure with 64 file size and timestamps. It will return correct file size now (C runtime implementation used to report outdated information) - my_lock on Windows is prepared to handle additional timeout parameter - after review : changed __WIN__ to _WIN32 in the new and changed code.
2009-09-11 22:26:35 +02:00
#ifndef _WIN32
if (! stat((char *) path, (struct stat *) stat_area) )
DBUG_RETURN(stat_area);
#else
if (! my_win_stat(path, stat_area) )
DBUG_RETURN(stat_area);
#endif
Fixed a possible memory leak on MacOSX when using the shared libmysql.so library (Bug #2061) mysql_server_init() now returns error code if something went wrong (Bug #2062) Don't use my_fopen() when reading symlink information as this may cause problems when a lot of files are opened. Free thread keys with pthread_key_delete() instead of relying on automatic free. (Bug #2062) Fixed bug in UNION statement with alias '*'. (Bug #1249) Fixed a bug in DELETE ... ORDER BY ... LIMIT where the rows where not deleted in the proper order. (Bug #1024). FOUND_ROWS() could return incorrect number of rows after a query with an impossible WHERE condition. HOW DATABASES doesn't anymore show .sym files (on windows) that doesn't point to a valid directory. (Bug #1385)
2003-12-11 05:24:08 +01:00
DBUG_PRINT("error",("Got errno: %d from stat", errno));
my_errno= errno;
Import changeset
2000-07-31 21:29:14 +02:00
if (m_used) /* Free if new area */
Bug#34043: Server loops excessively in _checkchunk() when safemalloc is enabled Essentially, the problem is that safemalloc is excruciatingly slow as it checks all allocated blocks for overrun at each memory management primitive, yielding a almost exponential slowdown for the memory management functions (malloc, realloc, free). The overrun check basically consists of verifying some bytes of a block for certain magic keys, which catches some simple forms of overrun. Another minor problem is violation of aliasing rules and that its own internal list of blocks is prone to corruption. Another issue with safemalloc is rather the maintenance cost as the tool has a significant impact on the server code. Given the magnitude of memory debuggers available nowadays, especially those that are provided with the platform malloc implementation, maintenance of a in-house and largely obsolete memory debugger becomes a burden that is not worth the effort due to its slowness and lack of support for detecting more common forms of heap corruption. Since there are third-party tools that can provide the same functionality at a lower or comparable performance cost, the solution is to simply remove safemalloc. Third-party tools can provide the same functionality at a lower or comparable performance cost. The removal of safemalloc also allows a simplification of the malloc wrappers, removing quite a bit of kludge: redefinition of my_malloc, my_free and the removal of the unused second argument of my_free. Since free() always check whether the supplied pointer is null, redudant checks are also removed. Also, this patch adds unit testing for my_malloc and moves my_realloc implementation into the same file as the other memory allocation primitives.
2010-07-08 23:20:08 +02:00
my_free(stat_area);
Import changeset
2000-07-31 21:29:14 +02:00
error:
if (my_flags & (MY_FAE+MY_WME))
{
my_error(EE_STAT, MYF(ME_BELL+ME_WAITTANG),path,my_errno);
DBUG_RETURN((MY_STAT *) NULL);
}
DBUG_RETURN((MY_STAT *) NULL);
} /* my_stat */
Reference in a new issue Copy permalink