1
0
mirror of git://sourceware.org/git/lvm2.git synced 2025-01-16 03:25:10 +03:00
lvm2/lib/config/config.c

2564 lines
72 KiB
C
Raw Normal View History

/*
* Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.
2018-04-20 10:43:50 -05:00
* Copyright (C) 2004-2011 Red Hat, Inc. All rights reserved.
*
2004-03-30 19:35:44 +00:00
* This file is part of LVM2.
*
* This copyrighted material is made available to anyone wishing to use,
* modify, copy, or redistribute it subject to the terms and conditions
* of the GNU Lesser General Public License v.2.1.
2004-03-30 19:35:44 +00:00
*
* You should have received a copy of the GNU Lesser General Public License
2004-03-30 19:35:44 +00:00
* along with this program; if not, write to the Free Software Foundation,
* Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "lib/misc/lib.h"
#include "lib/config/config.h"
#include "lib/misc/crc.h"
#include "lib/device/device.h"
#include "lib/datastruct/str_list.h"
#include "lib/commands/toolcontext.h"
#include "lib/misc/lvm-file.h"
#include "lib/mm/memlock.h"
#include "lib/label/label.h"
Place the first PE at 1 MiB for all defaults . When using default settings, this commit should change nothing. The first PE continues to be placed at 1 MiB resulting in a metadata area size of 1020 KiB (for 4K page sizes; slightly smaller for larger page sizes.) . When default_data_alignment is disabled in lvm.conf, align pe_start at 1 MiB, based on a default metadata area size that adapts to the page size. Previously, disabling this option would result in mda_size that was too small for common use, and produced a 64 KiB aligned pe_start. . Customized pe_start and mda_size values continue to be set as before in lvm.conf and command line. . Remove the configure option for setting default_data_alignment at build time. . Improve alignment related option descriptions. . Add section about alignment to pvcreate man page. Previously, DEFAULT_PVMETADATASIZE was 255 sectors. However, the fact that the config setting named "default_data_alignment" has a default value of 1 (MiB) meant that DEFAULT_PVMETADATASIZE was having no effect. The metadata area size is the space between the start of the metadata area (page size offset from the start of the device) and the first PE (1 MiB by default due to default_data_alignment 1.) The result is a 1020 KiB metadata area on machines with 4KiB page size (1024 KiB - 4 KiB), and smaller on machines with larger page size. If default_data_alignment was set to 0 (disabled), then DEFAULT_PVMETADATASIZE 255 would take effect, and produce a metadata area that was 188 KiB and pe_start of 192 KiB. This was too small for common use. This is fixed by making the default metadata area size a computed value that matches the value produced by default_data_alignment.
2018-11-13 15:00:11 -06:00
#include "lib/metadata/metadata.h"
2002-11-18 14:01:16 +00:00
#include <sys/stat.h>
#include <sys/mman.h>
#include <unistd.h>
#include <fcntl.h>
#include <assert.h>
config: use config checks and add support for creating trees from config definition (config_def_create_tree fn) Configuration checking is initiated during config load/processing (_process_config fn) which is part of the command context creation/refresh. This patch also defines 5 types of trees that could be created from the configuration definition (config_settings.h), the cfg_def_tree_t: - CFG_DEF_TREE_CURRENT that denotes a tree of all the configuration nodes that are explicitly defined in lvm.conf/--config - CFG_DEF_TREE_MISSING that denotes a tree of all missing configuration nodes for which default valus are used since they're not explicitly used in lvm.conf/--config - CFG_DEF_TREE_DEFAULT that denotes a tree of all possible configuration nodes with default values assigned, no matter what the actual lvm.conf/--config is - CFG_DEF_TREE_NEW that denotes a tree of all new configuration nodes that appeared in given version - CFG_DEF_TREE_COMPLETE that denotes a tree of the whole configuration tree that is used in LVM2 (a combination of CFG_DEF_TREE_CURRENT + CFG_DEF_TREE_MISSING). This is not implemented yet, it will be added later... The function that creates the definition tree of given type: struct dm_config_tree *config_def_create_tree(struct config_def_tree_spec *spec); Where the "spec" specifies the tree type to be created: struct config_def_tree_spec { cfg_def_tree_t type; /* tree type */ uint16_t version; /* tree at this LVM2 version */ int ignoreadvanced; /* do not include advanced configs */ int ignoreunsupported; /* do not include unsupported configs */ }; This tree can be passed to already existing functions that write the tree on output (like we already do with cmd->cft). There is a new lvm.conf section called "config" with two new options: - config/checks which enables/disables checking (enabled by default) - config/abort_on_errors which enables/disables aborts on any type of mismatch found in the config (disabled by default)
2013-03-05 17:36:10 +01:00
#include <ctype.h>
#include <math.h>
#include <float.h>
static const char *_config_source_names[] = {
[CONFIG_UNDEFINED] = "undefined",
[CONFIG_FILE] = "file",
[CONFIG_MERGED_FILES] = "merged files",
[CONFIG_STRING] = "string",
config: differentiate command and metadata profiles and consolidate profile handling code - When defining configuration source, the code now uses separate CONFIG_PROFILE_COMMAND and CONFIG_PROFILE_METADATA markers (before, it was just CONFIG_PROFILE that did not make the difference between the two). This helps when checking the configuration if it contains correct set of options which are all in either command-profilable or metadata-profilable group without mixing these groups together - so it's a firm distinction. The "command profile" can't contain "metadata profile" and vice versa! This is strictly checked and if the settings are mixed, such profile is rejected and it's not used. So in the end, the CONFIG_PROFILE_COMMAND set of options and CONFIG_PROFILE_METADATA are mutually exclusive sets. - Marking configuration with one or the other marker will also determine the way these configuration sources are positioned in the configuration cascade which is now: CONFIG_STRING -> CONFIG_PROFILE_COMMAND -> CONFIG_PROFILE_METADATA -> CONFIG_FILE/CONFIG_MERGED_FILES - Marking configuration with one or the other marker will also make it possible to issue a command context refresh (will be probably a part of a future patch) if needed for settings in global profile set. For settings in metadata profile set this is impossible since we can't refresh cmd context in the middle of reading VG/LV metadata and for each VG/LV separately because each VG/LV can have a different metadata profile assinged and it's not possible to change these settings at this level. - When command profile is incorrect, it's rejected *and also* the command exits immediately - the profile *must* be correct for the command that was run with a profile to be executed. Before this patch, when the profile was found incorrect, there was just the warning message and the command continued without profile applied. But it's more correct to exit immediately in this case. - When metadata profile is incorrect, we reject it during command runtime (as we know the profile name from metadata and not early from command line as it is in case of command profiles) and we *do continue* with the command as we're in the middle of operation. Also, the metadata profile is applied directly and on the fly on find_config_tree_* fn call and even if the metadata profile is found incorrect, we still need to return the non-profiled value as found in the other configuration provided or default value. To exit immediately even in this case, we'd need to refactor existing find_config_tree_* fns so they can return error. Currently, these fns return only config values (which end up with default values in the end if the config is not found). - To check the profile validity before use to be sure it's correct, one can use : lvm dumpconfig --commandprofile/--metadataprofile ProfileName --validate (the --commandprofile/--metadataprofile for dumpconfig will come as part of the subsequent patch) - This patch also adds a reference to --commandprofile and --metadataprofile in the cmd help string (which was missing before for the --profile for some commands). We do not mention --profile now as people should use --commandprofile or --metadataprofile directly. However, the --profile is still supported for backward compatibility and it's translated as: --profile == --metadataprofile for lvcreate, vgcreate, lvchange and vgchange (as these commands are able to attach profile to metadata) --profile == --commandprofile for all the other commands (--metadataprofile is not allowed there as it makes no sense) - This patch also contains some cleanups to make the code handling the profiles more readable...
2014-05-20 14:13:10 +02:00
[CONFIG_PROFILE_COMMAND] = "command profile",
[CONFIG_PROFILE_METADATA] = "metadata profile",
[CONFIG_FILE_SPECIAL] = "special purpose"
};
struct config_file {
off_t st_size;
char *filename;
int exists;
int keep_open;
struct device *dev;
};
struct config_source {
config_source_t type;
struct timespec timestamp;
union {
struct config_file *file;
struct config_file *profile;
} source;
struct cft_check_handle *check_handle;
};
/*
* Map each ID to respective definition of the configuration item.
*/
static struct cfg_def_item _cfg_def_items[CFG_COUNT + 1] = {
2017-07-19 16:17:30 +02:00
#define cfg_section(id, name, parent, flags, since_version, deprecated_since_version, deprecation_comment, comment) {id, parent, name, CFG_TYPE_SECTION, {0}, (flags), since_version, {0}, deprecated_since_version, deprecation_comment, comment},
#define cfg(id, name, parent, flags, type, default_value, since_version, unconfigured_value, deprecated_since_version, deprecation_comment, comment) {id, parent, name, type, {.v_##type = (default_value)}, (flags), since_version, {.v_UNCONFIGURED = (unconfigured_value)}, deprecated_since_version, deprecation_comment, comment},
#define cfg_runtime(id, name, parent, flags, type, since_version, deprecated_since_version, deprecation_comment, comment) {id, parent, name, type, {.fn_##type = get_default_##id}, (flags) | CFG_DEFAULT_RUN_TIME, since_version, {.fn_UNCONFIGURED = get_default_unconfigured_##id}, deprecated_since_version, (deprecation_comment), comment},
#define cfg_array(id, name, parent, flags, types, default_value, since_version, unconfigured_value, deprecated_since_version, deprecation_comment, comment) {id, parent, name, CFG_TYPE_ARRAY | (types), {.v_CFG_TYPE_STRING = (default_value)}, (flags), (since_version), {.v_UNCONFIGURED = (unconfigured_value)}, deprecated_since_version, deprecation_comment, comment},
#define cfg_array_runtime(id, name, parent, flags, types, since_version, deprecated_since_version, deprecation_comment, comment) {id, parent, name, CFG_TYPE_ARRAY | (types), {.fn_CFG_TYPE_STRING = get_default_##id}, (flags) | CFG_DEFAULT_RUN_TIME, (since_version), {.fn_UNCONFIGURED = get_default_unconfigured_##id}, deprecated_since_version, deprecation_comment, comment},
#include "lib/config/config_settings.h"
#undef cfg_section
#undef cfg
config: add CFG_DEFAULT_RUN_TIME for config options with runtime defaults Previously, we declared a default value as undefined ("NULL") for settings which require runtime context to be set first (e.g. settings for paths that rely on SYSTEM_DIR environment variable or they depend on any other setting in some way). If we want to output default values as they are really used in runtime, we should make it possible to define a default value as function which is evaluated, not just providing a firm constant value as it was before. This patch defines simple prototypes for such functions. Also, there's new helper macros "cfg_runtime" and "cfg_array_runtime" - they provide exactly the same functionality as the original "cfg" and "cfg_array" macros when defining the configuration settings in config_settings.h, but they don't set the constant default value. Instead, they automatically link the configuration setting definition with one of these functions: typedef int (*t_fn_CFG_TYPE_BOOL) (struct cmd_context *cmd, struct profile *profile); typedef int (*t_fn_CFG_TYPE_INT) (struct cmd_context *cmd, struct profile *profile); typedef float (*t_fn_CFG_TYPE_FLOAT) (struct cmd_context *cmd, struct profile *profile); typedef const char* (*t_fn_CFG_TYPE_STRING) (struct cmd_context *cmd, struct profile *profile); typedef const char* (*t_fn_CFG_TYPE_ARRAY) (struct cmd_context *cmd, struct profile *profile); (The new macros actually set the CFG_DEFAULT_RUNTIME flag properly and set the default value link to the function accordingly). Then such configuration setting requires a function of selected type to be defined. This function has a predefined name: get_default_<id> ...where the <id> is the id of the setting as defined in config_settings.h. For example "backup_archive_dir_CFG" if defined as a setting with default value evaluated in runtime with "cfg_runtime" will automatically have "get_default_backup_archive_dir_CFG" function linked to this setting to get the default value.
2014-03-03 12:34:11 +01:00
#undef cfg_runtime
#undef cfg_array
config: add CFG_DEFAULT_RUN_TIME for config options with runtime defaults Previously, we declared a default value as undefined ("NULL") for settings which require runtime context to be set first (e.g. settings for paths that rely on SYSTEM_DIR environment variable or they depend on any other setting in some way). If we want to output default values as they are really used in runtime, we should make it possible to define a default value as function which is evaluated, not just providing a firm constant value as it was before. This patch defines simple prototypes for such functions. Also, there's new helper macros "cfg_runtime" and "cfg_array_runtime" - they provide exactly the same functionality as the original "cfg" and "cfg_array" macros when defining the configuration settings in config_settings.h, but they don't set the constant default value. Instead, they automatically link the configuration setting definition with one of these functions: typedef int (*t_fn_CFG_TYPE_BOOL) (struct cmd_context *cmd, struct profile *profile); typedef int (*t_fn_CFG_TYPE_INT) (struct cmd_context *cmd, struct profile *profile); typedef float (*t_fn_CFG_TYPE_FLOAT) (struct cmd_context *cmd, struct profile *profile); typedef const char* (*t_fn_CFG_TYPE_STRING) (struct cmd_context *cmd, struct profile *profile); typedef const char* (*t_fn_CFG_TYPE_ARRAY) (struct cmd_context *cmd, struct profile *profile); (The new macros actually set the CFG_DEFAULT_RUNTIME flag properly and set the default value link to the function accordingly). Then such configuration setting requires a function of selected type to be defined. This function has a predefined name: get_default_<id> ...where the <id> is the id of the setting as defined in config_settings.h. For example "backup_archive_dir_CFG" if defined as a setting with default value evaluated in runtime with "cfg_runtime" will automatically have "get_default_backup_archive_dir_CFG" function linked to this setting to get the default value.
2014-03-03 12:34:11 +01:00
#undef cfg_array_runtime
};
config_source_t config_get_source_type(struct dm_config_tree *cft)
{
struct config_source *cs = dm_config_get_custom(cft);
return cs ? cs->type : CONFIG_UNDEFINED;
}
config: differentiate command and metadata profiles and consolidate profile handling code - When defining configuration source, the code now uses separate CONFIG_PROFILE_COMMAND and CONFIG_PROFILE_METADATA markers (before, it was just CONFIG_PROFILE that did not make the difference between the two). This helps when checking the configuration if it contains correct set of options which are all in either command-profilable or metadata-profilable group without mixing these groups together - so it's a firm distinction. The "command profile" can't contain "metadata profile" and vice versa! This is strictly checked and if the settings are mixed, such profile is rejected and it's not used. So in the end, the CONFIG_PROFILE_COMMAND set of options and CONFIG_PROFILE_METADATA are mutually exclusive sets. - Marking configuration with one or the other marker will also determine the way these configuration sources are positioned in the configuration cascade which is now: CONFIG_STRING -> CONFIG_PROFILE_COMMAND -> CONFIG_PROFILE_METADATA -> CONFIG_FILE/CONFIG_MERGED_FILES - Marking configuration with one or the other marker will also make it possible to issue a command context refresh (will be probably a part of a future patch) if needed for settings in global profile set. For settings in metadata profile set this is impossible since we can't refresh cmd context in the middle of reading VG/LV metadata and for each VG/LV separately because each VG/LV can have a different metadata profile assinged and it's not possible to change these settings at this level. - When command profile is incorrect, it's rejected *and also* the command exits immediately - the profile *must* be correct for the command that was run with a profile to be executed. Before this patch, when the profile was found incorrect, there was just the warning message and the command continued without profile applied. But it's more correct to exit immediately in this case. - When metadata profile is incorrect, we reject it during command runtime (as we know the profile name from metadata and not early from command line as it is in case of command profiles) and we *do continue* with the command as we're in the middle of operation. Also, the metadata profile is applied directly and on the fly on find_config_tree_* fn call and even if the metadata profile is found incorrect, we still need to return the non-profiled value as found in the other configuration provided or default value. To exit immediately even in this case, we'd need to refactor existing find_config_tree_* fns so they can return error. Currently, these fns return only config values (which end up with default values in the end if the config is not found). - To check the profile validity before use to be sure it's correct, one can use : lvm dumpconfig --commandprofile/--metadataprofile ProfileName --validate (the --commandprofile/--metadataprofile for dumpconfig will come as part of the subsequent patch) - This patch also adds a reference to --commandprofile and --metadataprofile in the cmd help string (which was missing before for the --profile for some commands). We do not mention --profile now as people should use --commandprofile or --metadataprofile directly. However, the --profile is still supported for backward compatibility and it's translated as: --profile == --metadataprofile for lvcreate, vgcreate, lvchange and vgchange (as these commands are able to attach profile to metadata) --profile == --commandprofile for all the other commands (--metadataprofile is not allowed there as it makes no sense) - This patch also contains some cleanups to make the code handling the profiles more readable...
2014-05-20 14:13:10 +02:00
static inline int _is_profile_based_config_source(config_source_t source)
{
return (source == CONFIG_PROFILE_COMMAND) ||
(source == CONFIG_PROFILE_METADATA);
}
static inline int _is_file_based_config_source(config_source_t source)
{
return (source == CONFIG_FILE) ||
(source == CONFIG_FILE_SPECIAL) ||
_is_profile_based_config_source(source);
}
/*
* public interface
*/
struct dm_config_tree *config_open(config_source_t source,
const char *filename,
int keep_open)
2002-01-07 10:23:52 +00:00
{
struct dm_config_tree *cft = dm_config_create();
struct config_source *cs;
struct config_file *cf;
if (!cft)
return NULL;
if (!(cs = dm_pool_zalloc(cft->mem, sizeof(struct config_source)))) {
log_error("Failed to allocate config source.");
goto fail;
}
2002-01-07 10:23:52 +00:00
config: differentiate command and metadata profiles and consolidate profile handling code - When defining configuration source, the code now uses separate CONFIG_PROFILE_COMMAND and CONFIG_PROFILE_METADATA markers (before, it was just CONFIG_PROFILE that did not make the difference between the two). This helps when checking the configuration if it contains correct set of options which are all in either command-profilable or metadata-profilable group without mixing these groups together - so it's a firm distinction. The "command profile" can't contain "metadata profile" and vice versa! This is strictly checked and if the settings are mixed, such profile is rejected and it's not used. So in the end, the CONFIG_PROFILE_COMMAND set of options and CONFIG_PROFILE_METADATA are mutually exclusive sets. - Marking configuration with one or the other marker will also determine the way these configuration sources are positioned in the configuration cascade which is now: CONFIG_STRING -> CONFIG_PROFILE_COMMAND -> CONFIG_PROFILE_METADATA -> CONFIG_FILE/CONFIG_MERGED_FILES - Marking configuration with one or the other marker will also make it possible to issue a command context refresh (will be probably a part of a future patch) if needed for settings in global profile set. For settings in metadata profile set this is impossible since we can't refresh cmd context in the middle of reading VG/LV metadata and for each VG/LV separately because each VG/LV can have a different metadata profile assinged and it's not possible to change these settings at this level. - When command profile is incorrect, it's rejected *and also* the command exits immediately - the profile *must* be correct for the command that was run with a profile to be executed. Before this patch, when the profile was found incorrect, there was just the warning message and the command continued without profile applied. But it's more correct to exit immediately in this case. - When metadata profile is incorrect, we reject it during command runtime (as we know the profile name from metadata and not early from command line as it is in case of command profiles) and we *do continue* with the command as we're in the middle of operation. Also, the metadata profile is applied directly and on the fly on find_config_tree_* fn call and even if the metadata profile is found incorrect, we still need to return the non-profiled value as found in the other configuration provided or default value. To exit immediately even in this case, we'd need to refactor existing find_config_tree_* fns so they can return error. Currently, these fns return only config values (which end up with default values in the end if the config is not found). - To check the profile validity before use to be sure it's correct, one can use : lvm dumpconfig --commandprofile/--metadataprofile ProfileName --validate (the --commandprofile/--metadataprofile for dumpconfig will come as part of the subsequent patch) - This patch also adds a reference to --commandprofile and --metadataprofile in the cmd help string (which was missing before for the --profile for some commands). We do not mention --profile now as people should use --commandprofile or --metadataprofile directly. However, the --profile is still supported for backward compatibility and it's translated as: --profile == --metadataprofile for lvcreate, vgcreate, lvchange and vgchange (as these commands are able to attach profile to metadata) --profile == --commandprofile for all the other commands (--metadataprofile is not allowed there as it makes no sense) - This patch also contains some cleanups to make the code handling the profiles more readable...
2014-05-20 14:13:10 +02:00
if (_is_file_based_config_source(source)) {
if (!(cf = dm_pool_zalloc(cft->mem, sizeof(struct config_file)))) {
log_error("Failed to allocate config file.");
goto fail;
}
cf->keep_open = keep_open;
if (filename &&
!(cf->filename = dm_pool_strdup(cft->mem, filename))) {
log_error("Failed to duplicate filename.");
goto fail;
}
cs->source.file = cf;
}
cs->type = source;
dm_config_set_custom(cft, cs);
return cft;
fail:
dm_config_destroy(cft);
return NULL;
}
/*
* Doesn't populate filename if the file is empty.
*/
int config_file_check(struct dm_config_tree *cft, const char **filename, struct stat *info)
{
struct config_source *cs = dm_config_get_custom(cft);
struct config_file *cf;
struct stat _info;
config: differentiate command and metadata profiles and consolidate profile handling code - When defining configuration source, the code now uses separate CONFIG_PROFILE_COMMAND and CONFIG_PROFILE_METADATA markers (before, it was just CONFIG_PROFILE that did not make the difference between the two). This helps when checking the configuration if it contains correct set of options which are all in either command-profilable or metadata-profilable group without mixing these groups together - so it's a firm distinction. The "command profile" can't contain "metadata profile" and vice versa! This is strictly checked and if the settings are mixed, such profile is rejected and it's not used. So in the end, the CONFIG_PROFILE_COMMAND set of options and CONFIG_PROFILE_METADATA are mutually exclusive sets. - Marking configuration with one or the other marker will also determine the way these configuration sources are positioned in the configuration cascade which is now: CONFIG_STRING -> CONFIG_PROFILE_COMMAND -> CONFIG_PROFILE_METADATA -> CONFIG_FILE/CONFIG_MERGED_FILES - Marking configuration with one or the other marker will also make it possible to issue a command context refresh (will be probably a part of a future patch) if needed for settings in global profile set. For settings in metadata profile set this is impossible since we can't refresh cmd context in the middle of reading VG/LV metadata and for each VG/LV separately because each VG/LV can have a different metadata profile assinged and it's not possible to change these settings at this level. - When command profile is incorrect, it's rejected *and also* the command exits immediately - the profile *must* be correct for the command that was run with a profile to be executed. Before this patch, when the profile was found incorrect, there was just the warning message and the command continued without profile applied. But it's more correct to exit immediately in this case. - When metadata profile is incorrect, we reject it during command runtime (as we know the profile name from metadata and not early from command line as it is in case of command profiles) and we *do continue* with the command as we're in the middle of operation. Also, the metadata profile is applied directly and on the fly on find_config_tree_* fn call and even if the metadata profile is found incorrect, we still need to return the non-profiled value as found in the other configuration provided or default value. To exit immediately even in this case, we'd need to refactor existing find_config_tree_* fns so they can return error. Currently, these fns return only config values (which end up with default values in the end if the config is not found). - To check the profile validity before use to be sure it's correct, one can use : lvm dumpconfig --commandprofile/--metadataprofile ProfileName --validate (the --commandprofile/--metadataprofile for dumpconfig will come as part of the subsequent patch) - This patch also adds a reference to --commandprofile and --metadataprofile in the cmd help string (which was missing before for the --profile for some commands). We do not mention --profile now as people should use --commandprofile or --metadataprofile directly. However, the --profile is still supported for backward compatibility and it's translated as: --profile == --metadataprofile for lvcreate, vgcreate, lvchange and vgchange (as these commands are able to attach profile to metadata) --profile == --commandprofile for all the other commands (--metadataprofile is not allowed there as it makes no sense) - This patch also contains some cleanups to make the code handling the profiles more readable...
2014-05-20 14:13:10 +02:00
if (!_is_file_based_config_source(cs->type)) {
log_error(INTERNAL_ERROR "config_file_check: expected file, special file or "
"profile config source, found %s config source.",
_config_source_names[cs->type]);
return 0;
}
if (!info)
info = &_info;
cf = cs->source.file;
if (stat(cf->filename, info)) {
log_sys_error("stat", cf->filename);
cf->exists = 0;
return 0;
}
if (!S_ISREG(info->st_mode)) {
log_error("%s is not a regular file", cf->filename);
cf->exists = 0;
return 0;
}
lvm_stat_ctim(&cs->timestamp, info);
cf->exists = 1;
cf->st_size = info->st_size;
if (info->st_size == 0)
log_verbose("%s is empty", cf->filename);
else if (filename)
*filename = cf->filename;
return 1;
}
/*
* Return 1 if config files ought to be reloaded
*/
int config_file_changed(struct dm_config_tree *cft)
{
struct config_source *cs = dm_config_get_custom(cft);
struct config_file *cf;
struct stat info;
struct timespec ts;
if (cs->type != CONFIG_FILE) {
log_error(INTERNAL_ERROR "config_file_changed: expected file config source, "
"found %s config source.", _config_source_names[cs->type]);
return 0;
}
cf = cs->source.file;
if (!cf->filename)
return 0;
if (stat(cf->filename, &info) == -1) {
/* Ignore a deleted config file: still use original data */
if (errno == ENOENT) {
if (!cf->exists)
return 0;
log_very_verbose("Config file %s has disappeared!",
cf->filename);
goto reload;
}
log_sys_error("stat", cf->filename);
log_error("Failed to reload configuration files");
return 0;
}
if (!S_ISREG(info.st_mode)) {
log_error("Configuration file %s is not a regular file",
cf->filename);
goto reload;
}
/* Unchanged? */
lvm_stat_ctim(&ts, &info);
if ((timespeccmp(&cs->timestamp, &ts, ==)) &&
cf->st_size == info.st_size)
return 0;
reload:
log_verbose("Detected config file change to %s", cf->filename);
return 1;
}
void config_destroy(struct dm_config_tree *cft)
{
struct config_source *cs;
struct config_file *cf;
if (!cft)
return;
cs = dm_config_get_custom(cft);
config: differentiate command and metadata profiles and consolidate profile handling code - When defining configuration source, the code now uses separate CONFIG_PROFILE_COMMAND and CONFIG_PROFILE_METADATA markers (before, it was just CONFIG_PROFILE that did not make the difference between the two). This helps when checking the configuration if it contains correct set of options which are all in either command-profilable or metadata-profilable group without mixing these groups together - so it's a firm distinction. The "command profile" can't contain "metadata profile" and vice versa! This is strictly checked and if the settings are mixed, such profile is rejected and it's not used. So in the end, the CONFIG_PROFILE_COMMAND set of options and CONFIG_PROFILE_METADATA are mutually exclusive sets. - Marking configuration with one or the other marker will also determine the way these configuration sources are positioned in the configuration cascade which is now: CONFIG_STRING -> CONFIG_PROFILE_COMMAND -> CONFIG_PROFILE_METADATA -> CONFIG_FILE/CONFIG_MERGED_FILES - Marking configuration with one or the other marker will also make it possible to issue a command context refresh (will be probably a part of a future patch) if needed for settings in global profile set. For settings in metadata profile set this is impossible since we can't refresh cmd context in the middle of reading VG/LV metadata and for each VG/LV separately because each VG/LV can have a different metadata profile assinged and it's not possible to change these settings at this level. - When command profile is incorrect, it's rejected *and also* the command exits immediately - the profile *must* be correct for the command that was run with a profile to be executed. Before this patch, when the profile was found incorrect, there was just the warning message and the command continued without profile applied. But it's more correct to exit immediately in this case. - When metadata profile is incorrect, we reject it during command runtime (as we know the profile name from metadata and not early from command line as it is in case of command profiles) and we *do continue* with the command as we're in the middle of operation. Also, the metadata profile is applied directly and on the fly on find_config_tree_* fn call and even if the metadata profile is found incorrect, we still need to return the non-profiled value as found in the other configuration provided or default value. To exit immediately even in this case, we'd need to refactor existing find_config_tree_* fns so they can return error. Currently, these fns return only config values (which end up with default values in the end if the config is not found). - To check the profile validity before use to be sure it's correct, one can use : lvm dumpconfig --commandprofile/--metadataprofile ProfileName --validate (the --commandprofile/--metadataprofile for dumpconfig will come as part of the subsequent patch) - This patch also adds a reference to --commandprofile and --metadataprofile in the cmd help string (which was missing before for the --profile for some commands). We do not mention --profile now as people should use --commandprofile or --metadataprofile directly. However, the --profile is still supported for backward compatibility and it's translated as: --profile == --metadataprofile for lvcreate, vgcreate, lvchange and vgchange (as these commands are able to attach profile to metadata) --profile == --commandprofile for all the other commands (--metadataprofile is not allowed there as it makes no sense) - This patch also contains some cleanups to make the code handling the profiles more readable...
2014-05-20 14:13:10 +02:00
if (_is_file_based_config_source(cs->type)) {
cf = cs->source.file;
if (cf && cf->dev)
if (!dev_close(cf->dev))
stack;
}
2002-01-07 10:23:52 +00:00
dm_config_destroy(cft);
}
struct dm_config_tree *config_file_open_and_read(const char *config_file,
config_source_t source,
struct cmd_context *cmd)
{
struct dm_config_tree *cft;
struct stat info;
if (!(cft = config_open(source, config_file, 0))) {
log_error("config_tree allocation failed");
return NULL;
}
/* Is there a config file? */
if (stat(config_file, &info) == -1) {
/* Profile file must be present! */
config: differentiate command and metadata profiles and consolidate profile handling code - When defining configuration source, the code now uses separate CONFIG_PROFILE_COMMAND and CONFIG_PROFILE_METADATA markers (before, it was just CONFIG_PROFILE that did not make the difference between the two). This helps when checking the configuration if it contains correct set of options which are all in either command-profilable or metadata-profilable group without mixing these groups together - so it's a firm distinction. The "command profile" can't contain "metadata profile" and vice versa! This is strictly checked and if the settings are mixed, such profile is rejected and it's not used. So in the end, the CONFIG_PROFILE_COMMAND set of options and CONFIG_PROFILE_METADATA are mutually exclusive sets. - Marking configuration with one or the other marker will also determine the way these configuration sources are positioned in the configuration cascade which is now: CONFIG_STRING -> CONFIG_PROFILE_COMMAND -> CONFIG_PROFILE_METADATA -> CONFIG_FILE/CONFIG_MERGED_FILES - Marking configuration with one or the other marker will also make it possible to issue a command context refresh (will be probably a part of a future patch) if needed for settings in global profile set. For settings in metadata profile set this is impossible since we can't refresh cmd context in the middle of reading VG/LV metadata and for each VG/LV separately because each VG/LV can have a different metadata profile assinged and it's not possible to change these settings at this level. - When command profile is incorrect, it's rejected *and also* the command exits immediately - the profile *must* be correct for the command that was run with a profile to be executed. Before this patch, when the profile was found incorrect, there was just the warning message and the command continued without profile applied. But it's more correct to exit immediately in this case. - When metadata profile is incorrect, we reject it during command runtime (as we know the profile name from metadata and not early from command line as it is in case of command profiles) and we *do continue* with the command as we're in the middle of operation. Also, the metadata profile is applied directly and on the fly on find_config_tree_* fn call and even if the metadata profile is found incorrect, we still need to return the non-profiled value as found in the other configuration provided or default value. To exit immediately even in this case, we'd need to refactor existing find_config_tree_* fns so they can return error. Currently, these fns return only config values (which end up with default values in the end if the config is not found). - To check the profile validity before use to be sure it's correct, one can use : lvm dumpconfig --commandprofile/--metadataprofile ProfileName --validate (the --commandprofile/--metadataprofile for dumpconfig will come as part of the subsequent patch) - This patch also adds a reference to --commandprofile and --metadataprofile in the cmd help string (which was missing before for the --profile for some commands). We do not mention --profile now as people should use --commandprofile or --metadataprofile directly. However, the --profile is still supported for backward compatibility and it's translated as: --profile == --metadataprofile for lvcreate, vgcreate, lvchange and vgchange (as these commands are able to attach profile to metadata) --profile == --commandprofile for all the other commands (--metadataprofile is not allowed there as it makes no sense) - This patch also contains some cleanups to make the code handling the profiles more readable...
2014-05-20 14:13:10 +02:00
if (errno == ENOENT && (!_is_profile_based_config_source(source)))
return cft;
log_sys_error("stat", config_file);
goto bad;
}
log_very_verbose("Loading config file: %s", config_file);
2018-04-20 10:43:50 -05:00
if (!config_file_read(cft)) {
log_error("Failed to load config file %s", config_file);
goto bad;
}
return cft;
bad:
config_destroy(cft);
return NULL;
}
struct dm_config_tree *get_config_tree_by_source(struct cmd_context *cmd,
config_source_t source)
{
struct dm_config_tree *cft = cmd->cft;
struct config_source *cs;
while (cft) {
cs = dm_config_get_custom(cft);
if (cs && cs->type == source)
return cft;
cft = cft->cascade;
}
return NULL;
}
/*
* Returns config tree if it was removed.
*/
struct dm_config_tree *remove_config_tree_by_source(struct cmd_context *cmd,
config_source_t source)
{
struct dm_config_tree *previous_cft = NULL;
struct dm_config_tree *cft = cmd->cft;
struct config_source *cs;
while (cft) {
cs = dm_config_get_custom(cft);
if (cs && (cs->type == source)) {
if (previous_cft) {
previous_cft->cascade = cft->cascade;
cmd->cft = previous_cft;
} else
cmd->cft = cft->cascade;
cft->cascade = NULL;
break;
}
previous_cft = cft;
cft = cft->cascade;
}
return cft;
}
struct cft_check_handle *get_config_tree_check_handle(struct cmd_context *cmd,
struct dm_config_tree *cft)
{
struct config_source *cs;
if (!(cs = dm_config_get_custom(cft)))
return NULL;
if (cs->check_handle)
goto out;
/*
* Attach config check handle to all config types but
* CONFIG_FILE_SPECIAL - this one uses its own check
* methods and the cft_check_handle is not applicable here.
*/
if (cs->type != CONFIG_FILE_SPECIAL) {
if (!(cs->check_handle = dm_pool_zalloc(cft->mem, sizeof(*cs->check_handle)))) {
log_error("Failed to allocate configuration check handle.");
return NULL;
}
cs->check_handle->cft = cft;
cs->check_handle->cmd = cmd;
}
out:
return cs->check_handle;
}
int override_config_tree_from_string(struct cmd_context *cmd,
const char *config_settings)
{
struct dm_config_tree *cft_new;
struct config_source *cs = dm_config_get_custom(cmd->cft);
/*
* Follow this sequence:
config: differentiate command and metadata profiles and consolidate profile handling code - When defining configuration source, the code now uses separate CONFIG_PROFILE_COMMAND and CONFIG_PROFILE_METADATA markers (before, it was just CONFIG_PROFILE that did not make the difference between the two). This helps when checking the configuration if it contains correct set of options which are all in either command-profilable or metadata-profilable group without mixing these groups together - so it's a firm distinction. The "command profile" can't contain "metadata profile" and vice versa! This is strictly checked and if the settings are mixed, such profile is rejected and it's not used. So in the end, the CONFIG_PROFILE_COMMAND set of options and CONFIG_PROFILE_METADATA are mutually exclusive sets. - Marking configuration with one or the other marker will also determine the way these configuration sources are positioned in the configuration cascade which is now: CONFIG_STRING -> CONFIG_PROFILE_COMMAND -> CONFIG_PROFILE_METADATA -> CONFIG_FILE/CONFIG_MERGED_FILES - Marking configuration with one or the other marker will also make it possible to issue a command context refresh (will be probably a part of a future patch) if needed for settings in global profile set. For settings in metadata profile set this is impossible since we can't refresh cmd context in the middle of reading VG/LV metadata and for each VG/LV separately because each VG/LV can have a different metadata profile assinged and it's not possible to change these settings at this level. - When command profile is incorrect, it's rejected *and also* the command exits immediately - the profile *must* be correct for the command that was run with a profile to be executed. Before this patch, when the profile was found incorrect, there was just the warning message and the command continued without profile applied. But it's more correct to exit immediately in this case. - When metadata profile is incorrect, we reject it during command runtime (as we know the profile name from metadata and not early from command line as it is in case of command profiles) and we *do continue* with the command as we're in the middle of operation. Also, the metadata profile is applied directly and on the fly on find_config_tree_* fn call and even if the metadata profile is found incorrect, we still need to return the non-profiled value as found in the other configuration provided or default value. To exit immediately even in this case, we'd need to refactor existing find_config_tree_* fns so they can return error. Currently, these fns return only config values (which end up with default values in the end if the config is not found). - To check the profile validity before use to be sure it's correct, one can use : lvm dumpconfig --commandprofile/--metadataprofile ProfileName --validate (the --commandprofile/--metadataprofile for dumpconfig will come as part of the subsequent patch) - This patch also adds a reference to --commandprofile and --metadataprofile in the cmd help string (which was missing before for the --profile for some commands). We do not mention --profile now as people should use --commandprofile or --metadataprofile directly. However, the --profile is still supported for backward compatibility and it's translated as: --profile == --metadataprofile for lvcreate, vgcreate, lvchange and vgchange (as these commands are able to attach profile to metadata) --profile == --commandprofile for all the other commands (--metadataprofile is not allowed there as it makes no sense) - This patch also contains some cleanups to make the code handling the profiles more readable...
2014-05-20 14:13:10 +02:00
* CONFIG_STRING -> CONFIG_PROFILE_COMMAND -> CONFIG_PROFILE_METADATA -> CONFIG_FILE/CONFIG_MERGED_FILES
*/
if (cs->type == CONFIG_STRING) {
log_error(INTERNAL_ERROR "override_config_tree_from_string: "
"config cascade already contains a string config.");
return 0;
}
if (!(cft_new = dm_config_from_string(config_settings))) {
log_error("Failed to set overridden configuration entries.");
return 0;
}
if (cmd->is_interactive &&
!config_force_check(cmd, CONFIG_STRING, cft_new)) {
log_error("Ignoring invalid configuration string.");
dm_config_destroy(cft_new);
return 0;
}
if (!(cs = dm_pool_zalloc(cft_new->mem, sizeof(struct config_source)))) {
log_error("Failed to allocate config source.");
dm_config_destroy(cft_new);
return 0;
}
cs->type = CONFIG_STRING;
dm_config_set_custom(cft_new, cs);
cmd->cft = dm_config_insert_cascaded_tree(cft_new, cmd->cft);
return 1;
}
config: differentiate command and metadata profiles and consolidate profile handling code - When defining configuration source, the code now uses separate CONFIG_PROFILE_COMMAND and CONFIG_PROFILE_METADATA markers (before, it was just CONFIG_PROFILE that did not make the difference between the two). This helps when checking the configuration if it contains correct set of options which are all in either command-profilable or metadata-profilable group without mixing these groups together - so it's a firm distinction. The "command profile" can't contain "metadata profile" and vice versa! This is strictly checked and if the settings are mixed, such profile is rejected and it's not used. So in the end, the CONFIG_PROFILE_COMMAND set of options and CONFIG_PROFILE_METADATA are mutually exclusive sets. - Marking configuration with one or the other marker will also determine the way these configuration sources are positioned in the configuration cascade which is now: CONFIG_STRING -> CONFIG_PROFILE_COMMAND -> CONFIG_PROFILE_METADATA -> CONFIG_FILE/CONFIG_MERGED_FILES - Marking configuration with one or the other marker will also make it possible to issue a command context refresh (will be probably a part of a future patch) if needed for settings in global profile set. For settings in metadata profile set this is impossible since we can't refresh cmd context in the middle of reading VG/LV metadata and for each VG/LV separately because each VG/LV can have a different metadata profile assinged and it's not possible to change these settings at this level. - When command profile is incorrect, it's rejected *and also* the command exits immediately - the profile *must* be correct for the command that was run with a profile to be executed. Before this patch, when the profile was found incorrect, there was just the warning message and the command continued without profile applied. But it's more correct to exit immediately in this case. - When metadata profile is incorrect, we reject it during command runtime (as we know the profile name from metadata and not early from command line as it is in case of command profiles) and we *do continue* with the command as we're in the middle of operation. Also, the metadata profile is applied directly and on the fly on find_config_tree_* fn call and even if the metadata profile is found incorrect, we still need to return the non-profiled value as found in the other configuration provided or default value. To exit immediately even in this case, we'd need to refactor existing find_config_tree_* fns so they can return error. Currently, these fns return only config values (which end up with default values in the end if the config is not found). - To check the profile validity before use to be sure it's correct, one can use : lvm dumpconfig --commandprofile/--metadataprofile ProfileName --validate (the --commandprofile/--metadataprofile for dumpconfig will come as part of the subsequent patch) - This patch also adds a reference to --commandprofile and --metadataprofile in the cmd help string (which was missing before for the --profile for some commands). We do not mention --profile now as people should use --commandprofile or --metadataprofile directly. However, the --profile is still supported for backward compatibility and it's translated as: --profile == --metadataprofile for lvcreate, vgcreate, lvchange and vgchange (as these commands are able to attach profile to metadata) --profile == --commandprofile for all the other commands (--metadataprofile is not allowed there as it makes no sense) - This patch also contains some cleanups to make the code handling the profiles more readable...
2014-05-20 14:13:10 +02:00
static int _override_config_tree_from_command_profile(struct cmd_context *cmd,
struct profile *profile)
{
config: differentiate command and metadata profiles and consolidate profile handling code - When defining configuration source, the code now uses separate CONFIG_PROFILE_COMMAND and CONFIG_PROFILE_METADATA markers (before, it was just CONFIG_PROFILE that did not make the difference between the two). This helps when checking the configuration if it contains correct set of options which are all in either command-profilable or metadata-profilable group without mixing these groups together - so it's a firm distinction. The "command profile" can't contain "metadata profile" and vice versa! This is strictly checked and if the settings are mixed, such profile is rejected and it's not used. So in the end, the CONFIG_PROFILE_COMMAND set of options and CONFIG_PROFILE_METADATA are mutually exclusive sets. - Marking configuration with one or the other marker will also determine the way these configuration sources are positioned in the configuration cascade which is now: CONFIG_STRING -> CONFIG_PROFILE_COMMAND -> CONFIG_PROFILE_METADATA -> CONFIG_FILE/CONFIG_MERGED_FILES - Marking configuration with one or the other marker will also make it possible to issue a command context refresh (will be probably a part of a future patch) if needed for settings in global profile set. For settings in metadata profile set this is impossible since we can't refresh cmd context in the middle of reading VG/LV metadata and for each VG/LV separately because each VG/LV can have a different metadata profile assinged and it's not possible to change these settings at this level. - When command profile is incorrect, it's rejected *and also* the command exits immediately - the profile *must* be correct for the command that was run with a profile to be executed. Before this patch, when the profile was found incorrect, there was just the warning message and the command continued without profile applied. But it's more correct to exit immediately in this case. - When metadata profile is incorrect, we reject it during command runtime (as we know the profile name from metadata and not early from command line as it is in case of command profiles) and we *do continue* with the command as we're in the middle of operation. Also, the metadata profile is applied directly and on the fly on find_config_tree_* fn call and even if the metadata profile is found incorrect, we still need to return the non-profiled value as found in the other configuration provided or default value. To exit immediately even in this case, we'd need to refactor existing find_config_tree_* fns so they can return error. Currently, these fns return only config values (which end up with default values in the end if the config is not found). - To check the profile validity before use to be sure it's correct, one can use : lvm dumpconfig --commandprofile/--metadataprofile ProfileName --validate (the --commandprofile/--metadataprofile for dumpconfig will come as part of the subsequent patch) - This patch also adds a reference to --commandprofile and --metadataprofile in the cmd help string (which was missing before for the --profile for some commands). We do not mention --profile now as people should use --commandprofile or --metadataprofile directly. However, the --profile is still supported for backward compatibility and it's translated as: --profile == --metadataprofile for lvcreate, vgcreate, lvchange and vgchange (as these commands are able to attach profile to metadata) --profile == --commandprofile for all the other commands (--metadataprofile is not allowed there as it makes no sense) - This patch also contains some cleanups to make the code handling the profiles more readable...
2014-05-20 14:13:10 +02:00
struct dm_config_tree *cft = cmd->cft, *cft_previous = NULL;
struct config_source *cs = dm_config_get_custom(cft);
config: differentiate command and metadata profiles and consolidate profile handling code - When defining configuration source, the code now uses separate CONFIG_PROFILE_COMMAND and CONFIG_PROFILE_METADATA markers (before, it was just CONFIG_PROFILE that did not make the difference between the two). This helps when checking the configuration if it contains correct set of options which are all in either command-profilable or metadata-profilable group without mixing these groups together - so it's a firm distinction. The "command profile" can't contain "metadata profile" and vice versa! This is strictly checked and if the settings are mixed, such profile is rejected and it's not used. So in the end, the CONFIG_PROFILE_COMMAND set of options and CONFIG_PROFILE_METADATA are mutually exclusive sets. - Marking configuration with one or the other marker will also determine the way these configuration sources are positioned in the configuration cascade which is now: CONFIG_STRING -> CONFIG_PROFILE_COMMAND -> CONFIG_PROFILE_METADATA -> CONFIG_FILE/CONFIG_MERGED_FILES - Marking configuration with one or the other marker will also make it possible to issue a command context refresh (will be probably a part of a future patch) if needed for settings in global profile set. For settings in metadata profile set this is impossible since we can't refresh cmd context in the middle of reading VG/LV metadata and for each VG/LV separately because each VG/LV can have a different metadata profile assinged and it's not possible to change these settings at this level. - When command profile is incorrect, it's rejected *and also* the command exits immediately - the profile *must* be correct for the command that was run with a profile to be executed. Before this patch, when the profile was found incorrect, there was just the warning message and the command continued without profile applied. But it's more correct to exit immediately in this case. - When metadata profile is incorrect, we reject it during command runtime (as we know the profile name from metadata and not early from command line as it is in case of command profiles) and we *do continue* with the command as we're in the middle of operation. Also, the metadata profile is applied directly and on the fly on find_config_tree_* fn call and even if the metadata profile is found incorrect, we still need to return the non-profiled value as found in the other configuration provided or default value. To exit immediately even in this case, we'd need to refactor existing find_config_tree_* fns so they can return error. Currently, these fns return only config values (which end up with default values in the end if the config is not found). - To check the profile validity before use to be sure it's correct, one can use : lvm dumpconfig --commandprofile/--metadataprofile ProfileName --validate (the --commandprofile/--metadataprofile for dumpconfig will come as part of the subsequent patch) - This patch also adds a reference to --commandprofile and --metadataprofile in the cmd help string (which was missing before for the --profile for some commands). We do not mention --profile now as people should use --commandprofile or --metadataprofile directly. However, the --profile is still supported for backward compatibility and it's translated as: --profile == --metadataprofile for lvcreate, vgcreate, lvchange and vgchange (as these commands are able to attach profile to metadata) --profile == --commandprofile for all the other commands (--metadataprofile is not allowed there as it makes no sense) - This patch also contains some cleanups to make the code handling the profiles more readable...
2014-05-20 14:13:10 +02:00
if (cs->type == CONFIG_STRING) {
cft_previous = cft;
cft = cft->cascade;
cs = dm_config_get_custom(cft);
}
config: differentiate command and metadata profiles and consolidate profile handling code - When defining configuration source, the code now uses separate CONFIG_PROFILE_COMMAND and CONFIG_PROFILE_METADATA markers (before, it was just CONFIG_PROFILE that did not make the difference between the two). This helps when checking the configuration if it contains correct set of options which are all in either command-profilable or metadata-profilable group without mixing these groups together - so it's a firm distinction. The "command profile" can't contain "metadata profile" and vice versa! This is strictly checked and if the settings are mixed, such profile is rejected and it's not used. So in the end, the CONFIG_PROFILE_COMMAND set of options and CONFIG_PROFILE_METADATA are mutually exclusive sets. - Marking configuration with one or the other marker will also determine the way these configuration sources are positioned in the configuration cascade which is now: CONFIG_STRING -> CONFIG_PROFILE_COMMAND -> CONFIG_PROFILE_METADATA -> CONFIG_FILE/CONFIG_MERGED_FILES - Marking configuration with one or the other marker will also make it possible to issue a command context refresh (will be probably a part of a future patch) if needed for settings in global profile set. For settings in metadata profile set this is impossible since we can't refresh cmd context in the middle of reading VG/LV metadata and for each VG/LV separately because each VG/LV can have a different metadata profile assinged and it's not possible to change these settings at this level. - When command profile is incorrect, it's rejected *and also* the command exits immediately - the profile *must* be correct for the command that was run with a profile to be executed. Before this patch, when the profile was found incorrect, there was just the warning message and the command continued without profile applied. But it's more correct to exit immediately in this case. - When metadata profile is incorrect, we reject it during command runtime (as we know the profile name from metadata and not early from command line as it is in case of command profiles) and we *do continue* with the command as we're in the middle of operation. Also, the metadata profile is applied directly and on the fly on find_config_tree_* fn call and even if the metadata profile is found incorrect, we still need to return the non-profiled value as found in the other configuration provided or default value. To exit immediately even in this case, we'd need to refactor existing find_config_tree_* fns so they can return error. Currently, these fns return only config values (which end up with default values in the end if the config is not found). - To check the profile validity before use to be sure it's correct, one can use : lvm dumpconfig --commandprofile/--metadataprofile ProfileName --validate (the --commandprofile/--metadataprofile for dumpconfig will come as part of the subsequent patch) - This patch also adds a reference to --commandprofile and --metadataprofile in the cmd help string (which was missing before for the --profile for some commands). We do not mention --profile now as people should use --commandprofile or --metadataprofile directly. However, the --profile is still supported for backward compatibility and it's translated as: --profile == --metadataprofile for lvcreate, vgcreate, lvchange and vgchange (as these commands are able to attach profile to metadata) --profile == --commandprofile for all the other commands (--metadataprofile is not allowed there as it makes no sense) - This patch also contains some cleanups to make the code handling the profiles more readable...
2014-05-20 14:13:10 +02:00
if (cs->type == CONFIG_PROFILE_COMMAND) {
log_error(INTERNAL_ERROR "_override_config_tree_from_command_profile: "
"config cascade already contains a command profile config.");
return 0;
}
if (cft_previous)
dm_config_insert_cascaded_tree(cft_previous, profile->cft);
else
cmd->cft = profile->cft;
dm_config_insert_cascaded_tree(profile->cft, cft);
return 1;
}
static int _override_config_tree_from_metadata_profile(struct cmd_context *cmd,
struct profile *profile)
{
struct dm_config_tree *cft = cmd->cft, *cft_previous = NULL;
struct config_source *cs = dm_config_get_custom(cft);
if (cs->type == CONFIG_STRING) {
config: differentiate command and metadata profiles and consolidate profile handling code - When defining configuration source, the code now uses separate CONFIG_PROFILE_COMMAND and CONFIG_PROFILE_METADATA markers (before, it was just CONFIG_PROFILE that did not make the difference between the two). This helps when checking the configuration if it contains correct set of options which are all in either command-profilable or metadata-profilable group without mixing these groups together - so it's a firm distinction. The "command profile" can't contain "metadata profile" and vice versa! This is strictly checked and if the settings are mixed, such profile is rejected and it's not used. So in the end, the CONFIG_PROFILE_COMMAND set of options and CONFIG_PROFILE_METADATA are mutually exclusive sets. - Marking configuration with one or the other marker will also determine the way these configuration sources are positioned in the configuration cascade which is now: CONFIG_STRING -> CONFIG_PROFILE_COMMAND -> CONFIG_PROFILE_METADATA -> CONFIG_FILE/CONFIG_MERGED_FILES - Marking configuration with one or the other marker will also make it possible to issue a command context refresh (will be probably a part of a future patch) if needed for settings in global profile set. For settings in metadata profile set this is impossible since we can't refresh cmd context in the middle of reading VG/LV metadata and for each VG/LV separately because each VG/LV can have a different metadata profile assinged and it's not possible to change these settings at this level. - When command profile is incorrect, it's rejected *and also* the command exits immediately - the profile *must* be correct for the command that was run with a profile to be executed. Before this patch, when the profile was found incorrect, there was just the warning message and the command continued without profile applied. But it's more correct to exit immediately in this case. - When metadata profile is incorrect, we reject it during command runtime (as we know the profile name from metadata and not early from command line as it is in case of command profiles) and we *do continue* with the command as we're in the middle of operation. Also, the metadata profile is applied directly and on the fly on find_config_tree_* fn call and even if the metadata profile is found incorrect, we still need to return the non-profiled value as found in the other configuration provided or default value. To exit immediately even in this case, we'd need to refactor existing find_config_tree_* fns so they can return error. Currently, these fns return only config values (which end up with default values in the end if the config is not found). - To check the profile validity before use to be sure it's correct, one can use : lvm dumpconfig --commandprofile/--metadataprofile ProfileName --validate (the --commandprofile/--metadataprofile for dumpconfig will come as part of the subsequent patch) - This patch also adds a reference to --commandprofile and --metadataprofile in the cmd help string (which was missing before for the --profile for some commands). We do not mention --profile now as people should use --commandprofile or --metadataprofile directly. However, the --profile is still supported for backward compatibility and it's translated as: --profile == --metadataprofile for lvcreate, vgcreate, lvchange and vgchange (as these commands are able to attach profile to metadata) --profile == --commandprofile for all the other commands (--metadataprofile is not allowed there as it makes no sense) - This patch also contains some cleanups to make the code handling the profiles more readable...
2014-05-20 14:13:10 +02:00
cft_previous = cft;
cft = cft->cascade;
}
if (cs->type == CONFIG_PROFILE_COMMAND) {
cft_previous = cft;
cft = cft->cascade;
}
config: differentiate command and metadata profiles and consolidate profile handling code - When defining configuration source, the code now uses separate CONFIG_PROFILE_COMMAND and CONFIG_PROFILE_METADATA markers (before, it was just CONFIG_PROFILE that did not make the difference between the two). This helps when checking the configuration if it contains correct set of options which are all in either command-profilable or metadata-profilable group without mixing these groups together - so it's a firm distinction. The "command profile" can't contain "metadata profile" and vice versa! This is strictly checked and if the settings are mixed, such profile is rejected and it's not used. So in the end, the CONFIG_PROFILE_COMMAND set of options and CONFIG_PROFILE_METADATA are mutually exclusive sets. - Marking configuration with one or the other marker will also determine the way these configuration sources are positioned in the configuration cascade which is now: CONFIG_STRING -> CONFIG_PROFILE_COMMAND -> CONFIG_PROFILE_METADATA -> CONFIG_FILE/CONFIG_MERGED_FILES - Marking configuration with one or the other marker will also make it possible to issue a command context refresh (will be probably a part of a future patch) if needed for settings in global profile set. For settings in metadata profile set this is impossible since we can't refresh cmd context in the middle of reading VG/LV metadata and for each VG/LV separately because each VG/LV can have a different metadata profile assinged and it's not possible to change these settings at this level. - When command profile is incorrect, it's rejected *and also* the command exits immediately - the profile *must* be correct for the command that was run with a profile to be executed. Before this patch, when the profile was found incorrect, there was just the warning message and the command continued without profile applied. But it's more correct to exit immediately in this case. - When metadata profile is incorrect, we reject it during command runtime (as we know the profile name from metadata and not early from command line as it is in case of command profiles) and we *do continue* with the command as we're in the middle of operation. Also, the metadata profile is applied directly and on the fly on find_config_tree_* fn call and even if the metadata profile is found incorrect, we still need to return the non-profiled value as found in the other configuration provided or default value. To exit immediately even in this case, we'd need to refactor existing find_config_tree_* fns so they can return error. Currently, these fns return only config values (which end up with default values in the end if the config is not found). - To check the profile validity before use to be sure it's correct, one can use : lvm dumpconfig --commandprofile/--metadataprofile ProfileName --validate (the --commandprofile/--metadataprofile for dumpconfig will come as part of the subsequent patch) - This patch also adds a reference to --commandprofile and --metadataprofile in the cmd help string (which was missing before for the --profile for some commands). We do not mention --profile now as people should use --commandprofile or --metadataprofile directly. However, the --profile is still supported for backward compatibility and it's translated as: --profile == --metadataprofile for lvcreate, vgcreate, lvchange and vgchange (as these commands are able to attach profile to metadata) --profile == --commandprofile for all the other commands (--metadataprofile is not allowed there as it makes no sense) - This patch also contains some cleanups to make the code handling the profiles more readable...
2014-05-20 14:13:10 +02:00
cs = dm_config_get_custom(cft);
if (cs->type == CONFIG_PROFILE_METADATA) {
log_error(INTERNAL_ERROR "_override_config_tree_from_metadata_profile: "
"config cascade already contains a metadata profile config.");
return 0;
}
if (cft_previous)
dm_config_insert_cascaded_tree(cft_previous, profile->cft);
else
cmd->cft = profile->cft;
config: differentiate command and metadata profiles and consolidate profile handling code - When defining configuration source, the code now uses separate CONFIG_PROFILE_COMMAND and CONFIG_PROFILE_METADATA markers (before, it was just CONFIG_PROFILE that did not make the difference between the two). This helps when checking the configuration if it contains correct set of options which are all in either command-profilable or metadata-profilable group without mixing these groups together - so it's a firm distinction. The "command profile" can't contain "metadata profile" and vice versa! This is strictly checked and if the settings are mixed, such profile is rejected and it's not used. So in the end, the CONFIG_PROFILE_COMMAND set of options and CONFIG_PROFILE_METADATA are mutually exclusive sets. - Marking configuration with one or the other marker will also determine the way these configuration sources are positioned in the configuration cascade which is now: CONFIG_STRING -> CONFIG_PROFILE_COMMAND -> CONFIG_PROFILE_METADATA -> CONFIG_FILE/CONFIG_MERGED_FILES - Marking configuration with one or the other marker will also make it possible to issue a command context refresh (will be probably a part of a future patch) if needed for settings in global profile set. For settings in metadata profile set this is impossible since we can't refresh cmd context in the middle of reading VG/LV metadata and for each VG/LV separately because each VG/LV can have a different metadata profile assinged and it's not possible to change these settings at this level. - When command profile is incorrect, it's rejected *and also* the command exits immediately - the profile *must* be correct for the command that was run with a profile to be executed. Before this patch, when the profile was found incorrect, there was just the warning message and the command continued without profile applied. But it's more correct to exit immediately in this case. - When metadata profile is incorrect, we reject it during command runtime (as we know the profile name from metadata and not early from command line as it is in case of command profiles) and we *do continue* with the command as we're in the middle of operation. Also, the metadata profile is applied directly and on the fly on find_config_tree_* fn call and even if the metadata profile is found incorrect, we still need to return the non-profiled value as found in the other configuration provided or default value. To exit immediately even in this case, we'd need to refactor existing find_config_tree_* fns so they can return error. Currently, these fns return only config values (which end up with default values in the end if the config is not found). - To check the profile validity before use to be sure it's correct, one can use : lvm dumpconfig --commandprofile/--metadataprofile ProfileName --validate (the --commandprofile/--metadataprofile for dumpconfig will come as part of the subsequent patch) - This patch also adds a reference to --commandprofile and --metadataprofile in the cmd help string (which was missing before for the --profile for some commands). We do not mention --profile now as people should use --commandprofile or --metadataprofile directly. However, the --profile is still supported for backward compatibility and it's translated as: --profile == --metadataprofile for lvcreate, vgcreate, lvchange and vgchange (as these commands are able to attach profile to metadata) --profile == --commandprofile for all the other commands (--metadataprofile is not allowed there as it makes no sense) - This patch also contains some cleanups to make the code handling the profiles more readable...
2014-05-20 14:13:10 +02:00
dm_config_insert_cascaded_tree(profile->cft, cft);
return 1;
}
config: differentiate command and metadata profiles and consolidate profile handling code - When defining configuration source, the code now uses separate CONFIG_PROFILE_COMMAND and CONFIG_PROFILE_METADATA markers (before, it was just CONFIG_PROFILE that did not make the difference between the two). This helps when checking the configuration if it contains correct set of options which are all in either command-profilable or metadata-profilable group without mixing these groups together - so it's a firm distinction. The "command profile" can't contain "metadata profile" and vice versa! This is strictly checked and if the settings are mixed, such profile is rejected and it's not used. So in the end, the CONFIG_PROFILE_COMMAND set of options and CONFIG_PROFILE_METADATA are mutually exclusive sets. - Marking configuration with one or the other marker will also determine the way these configuration sources are positioned in the configuration cascade which is now: CONFIG_STRING -> CONFIG_PROFILE_COMMAND -> CONFIG_PROFILE_METADATA -> CONFIG_FILE/CONFIG_MERGED_FILES - Marking configuration with one or the other marker will also make it possible to issue a command context refresh (will be probably a part of a future patch) if needed for settings in global profile set. For settings in metadata profile set this is impossible since we can't refresh cmd context in the middle of reading VG/LV metadata and for each VG/LV separately because each VG/LV can have a different metadata profile assinged and it's not possible to change these settings at this level. - When command profile is incorrect, it's rejected *and also* the command exits immediately - the profile *must* be correct for the command that was run with a profile to be executed. Before this patch, when the profile was found incorrect, there was just the warning message and the command continued without profile applied. But it's more correct to exit immediately in this case. - When metadata profile is incorrect, we reject it during command runtime (as we know the profile name from metadata and not early from command line as it is in case of command profiles) and we *do continue* with the command as we're in the middle of operation. Also, the metadata profile is applied directly and on the fly on find_config_tree_* fn call and even if the metadata profile is found incorrect, we still need to return the non-profiled value as found in the other configuration provided or default value. To exit immediately even in this case, we'd need to refactor existing find_config_tree_* fns so they can return error. Currently, these fns return only config values (which end up with default values in the end if the config is not found). - To check the profile validity before use to be sure it's correct, one can use : lvm dumpconfig --commandprofile/--metadataprofile ProfileName --validate (the --commandprofile/--metadataprofile for dumpconfig will come as part of the subsequent patch) - This patch also adds a reference to --commandprofile and --metadataprofile in the cmd help string (which was missing before for the --profile for some commands). We do not mention --profile now as people should use --commandprofile or --metadataprofile directly. However, the --profile is still supported for backward compatibility and it's translated as: --profile == --metadataprofile for lvcreate, vgcreate, lvchange and vgchange (as these commands are able to attach profile to metadata) --profile == --commandprofile for all the other commands (--metadataprofile is not allowed there as it makes no sense) - This patch also contains some cleanups to make the code handling the profiles more readable...
2014-05-20 14:13:10 +02:00
int override_config_tree_from_profile(struct cmd_context *cmd,
struct profile *profile)
{
/*
* Follow this sequence:
* CONFIG_STRING -> CONFIG_PROFILE_COMMAND -> CONFIG_PROFILE_METADATA -> CONFIG_FILE/CONFIG_MERGED_FILES
*/
if (!profile->cft && !load_profile(cmd, profile))
return_0;
if (profile->source == CONFIG_PROFILE_COMMAND)
return _override_config_tree_from_command_profile(cmd, profile);
if (profile->source == CONFIG_PROFILE_METADATA)
config: differentiate command and metadata profiles and consolidate profile handling code - When defining configuration source, the code now uses separate CONFIG_PROFILE_COMMAND and CONFIG_PROFILE_METADATA markers (before, it was just CONFIG_PROFILE that did not make the difference between the two). This helps when checking the configuration if it contains correct set of options which are all in either command-profilable or metadata-profilable group without mixing these groups together - so it's a firm distinction. The "command profile" can't contain "metadata profile" and vice versa! This is strictly checked and if the settings are mixed, such profile is rejected and it's not used. So in the end, the CONFIG_PROFILE_COMMAND set of options and CONFIG_PROFILE_METADATA are mutually exclusive sets. - Marking configuration with one or the other marker will also determine the way these configuration sources are positioned in the configuration cascade which is now: CONFIG_STRING -> CONFIG_PROFILE_COMMAND -> CONFIG_PROFILE_METADATA -> CONFIG_FILE/CONFIG_MERGED_FILES - Marking configuration with one or the other marker will also make it possible to issue a command context refresh (will be probably a part of a future patch) if needed for settings in global profile set. For settings in metadata profile set this is impossible since we can't refresh cmd context in the middle of reading VG/LV metadata and for each VG/LV separately because each VG/LV can have a different metadata profile assinged and it's not possible to change these settings at this level. - When command profile is incorrect, it's rejected *and also* the command exits immediately - the profile *must* be correct for the command that was run with a profile to be executed. Before this patch, when the profile was found incorrect, there was just the warning message and the command continued without profile applied. But it's more correct to exit immediately in this case. - When metadata profile is incorrect, we reject it during command runtime (as we know the profile name from metadata and not early from command line as it is in case of command profiles) and we *do continue* with the command as we're in the middle of operation. Also, the metadata profile is applied directly and on the fly on find_config_tree_* fn call and even if the metadata profile is found incorrect, we still need to return the non-profiled value as found in the other configuration provided or default value. To exit immediately even in this case, we'd need to refactor existing find_config_tree_* fns so they can return error. Currently, these fns return only config values (which end up with default values in the end if the config is not found). - To check the profile validity before use to be sure it's correct, one can use : lvm dumpconfig --commandprofile/--metadataprofile ProfileName --validate (the --commandprofile/--metadataprofile for dumpconfig will come as part of the subsequent patch) - This patch also adds a reference to --commandprofile and --metadataprofile in the cmd help string (which was missing before for the --profile for some commands). We do not mention --profile now as people should use --commandprofile or --metadataprofile directly. However, the --profile is still supported for backward compatibility and it's translated as: --profile == --metadataprofile for lvcreate, vgcreate, lvchange and vgchange (as these commands are able to attach profile to metadata) --profile == --commandprofile for all the other commands (--metadataprofile is not allowed there as it makes no sense) - This patch also contains some cleanups to make the code handling the profiles more readable...
2014-05-20 14:13:10 +02:00
return _override_config_tree_from_metadata_profile(cmd, profile);
log_error(INTERNAL_ERROR "override_config_tree_from_profile: incorrect profile source type");
return 0;
}
/*
* When checksum_only is set, the checksum of buffer is only matched
* and function avoids parsing of mda into config tree which
* remains unmodified and should not be used.
*/
2018-04-20 10:43:50 -05:00
int config_file_read_fd(struct dm_config_tree *cft, struct device *dev, dev_io_reason_t reason,
off_t offset, size_t size, off_t offset2, size_t size2,
checksum_fn_t checksum_fn, uint32_t checksum,
2018-04-20 10:43:50 -05:00
int checksum_only, int no_dup_node_check)
{
char namebuf[NAME_LEN + 1] __attribute__((aligned(8)));
int namelen = 0;
int bad_name = 0;
2018-04-20 10:43:50 -05:00
char *fb, *fe;
2002-11-18 14:01:16 +00:00
int r = 0;
int sz, use_plain_read = 1;
2018-04-20 10:43:50 -05:00
char *buf = NULL;
struct config_source *cs = dm_config_get_custom(cft);
size_t rsize;
config: differentiate command and metadata profiles and consolidate profile handling code - When defining configuration source, the code now uses separate CONFIG_PROFILE_COMMAND and CONFIG_PROFILE_METADATA markers (before, it was just CONFIG_PROFILE that did not make the difference between the two). This helps when checking the configuration if it contains correct set of options which are all in either command-profilable or metadata-profilable group without mixing these groups together - so it's a firm distinction. The "command profile" can't contain "metadata profile" and vice versa! This is strictly checked and if the settings are mixed, such profile is rejected and it's not used. So in the end, the CONFIG_PROFILE_COMMAND set of options and CONFIG_PROFILE_METADATA are mutually exclusive sets. - Marking configuration with one or the other marker will also determine the way these configuration sources are positioned in the configuration cascade which is now: CONFIG_STRING -> CONFIG_PROFILE_COMMAND -> CONFIG_PROFILE_METADATA -> CONFIG_FILE/CONFIG_MERGED_FILES - Marking configuration with one or the other marker will also make it possible to issue a command context refresh (will be probably a part of a future patch) if needed for settings in global profile set. For settings in metadata profile set this is impossible since we can't refresh cmd context in the middle of reading VG/LV metadata and for each VG/LV separately because each VG/LV can have a different metadata profile assinged and it's not possible to change these settings at this level. - When command profile is incorrect, it's rejected *and also* the command exits immediately - the profile *must* be correct for the command that was run with a profile to be executed. Before this patch, when the profile was found incorrect, there was just the warning message and the command continued without profile applied. But it's more correct to exit immediately in this case. - When metadata profile is incorrect, we reject it during command runtime (as we know the profile name from metadata and not early from command line as it is in case of command profiles) and we *do continue* with the command as we're in the middle of operation. Also, the metadata profile is applied directly and on the fly on find_config_tree_* fn call and even if the metadata profile is found incorrect, we still need to return the non-profiled value as found in the other configuration provided or default value. To exit immediately even in this case, we'd need to refactor existing find_config_tree_* fns so they can return error. Currently, these fns return only config values (which end up with default values in the end if the config is not found). - To check the profile validity before use to be sure it's correct, one can use : lvm dumpconfig --commandprofile/--metadataprofile ProfileName --validate (the --commandprofile/--metadataprofile for dumpconfig will come as part of the subsequent patch) - This patch also adds a reference to --commandprofile and --metadataprofile in the cmd help string (which was missing before for the --profile for some commands). We do not mention --profile now as people should use --commandprofile or --metadataprofile directly. However, the --profile is still supported for backward compatibility and it's translated as: --profile == --metadataprofile for lvcreate, vgcreate, lvchange and vgchange (as these commands are able to attach profile to metadata) --profile == --commandprofile for all the other commands (--metadataprofile is not allowed there as it makes no sense) - This patch also contains some cleanups to make the code handling the profiles more readable...
2014-05-20 14:13:10 +02:00
if (!_is_file_based_config_source(cs->type)) {
log_error(INTERNAL_ERROR "config_file_read_fd: expected file, special file "
"or profile config source, found %s config source.",
_config_source_names[cs->type]);
2018-04-20 10:43:50 -05:00
return 0;
}
/* Only use plain read with regular files */
2018-04-20 10:43:50 -05:00
if (!(dev->flags & DEV_REGULAR) || size2)
use_plain_read = 0;
/* Ensure there is extra '\0' after end of buffer since we pass
* buffer to funtions like strtoll() */
if (!(buf = zalloc(size + size2 + 1))) {
log_error("Failed to allocate circular buffer.");
return 0;
}
if (use_plain_read) {
/* Note: also used for lvm.conf to read all settings */
for (rsize = 0; rsize < size; rsize += sz) {
do {
sz = read(dev_fd(dev), buf + rsize, size - rsize);
} while ((sz < 0) && ((errno == EINTR) || (errno == EAGAIN)));
if (sz < 0) {
log_sys_error("read", dev_name(dev));
goto out;
}
2002-11-18 14:01:16 +00:00
}
} else {
if (!dev_read_bytes(dev, offset, size, buf))
2018-04-20 10:43:50 -05:00
goto out;
if (size2) {
if (!dev_read_bytes(dev, offset2, size2, buf + size))
goto out;
2018-04-20 10:43:50 -05:00
}
}
fb = buf;
if (!(dev->flags & DEV_REGULAR)) {
memcpy(namebuf, buf, NAME_LEN);
while (namebuf[namelen] && !isspace(namebuf[namelen]) && namebuf[namelen] != '{' && namelen