linux/fs/dlm/config.c

988 lines
23 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-only
/******************************************************************************
*******************************************************************************
**
** Copyright (C) Sistina Software, Inc. 1997-2003 All rights reserved.
** Copyright (C) 2004-2011 Red Hat, Inc. All rights reserved.
**
**
*******************************************************************************
******************************************************************************/
#include <linux/kernel.h>
#include <linux/init.h>
#include <linux/configfs.h>
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h percpu.h is included by sched.h and module.h and thus ends up being included when building most .c files. percpu.h includes slab.h which in turn includes gfp.h making everything defined by the two files universally available and complicating inclusion dependencies. percpu.h -> slab.h dependency is about to be removed. Prepare for this change by updating users of gfp and slab facilities include those headers directly instead of assuming availability. As this conversion needs to touch large number of source files, the following script is used as the basis of conversion. http://userweb.kernel.org/~tj/misc/slabh-sweep.py The script does the followings. * Scan files for gfp and slab usages and update includes such that only the necessary includes are there. ie. if only gfp is used, gfp.h, if slab is used, slab.h. * When the script inserts a new include, it looks at the include blocks and try to put the new include such that its order conforms to its surrounding. It's put in the include block which contains core kernel includes, in the same order that the rest are ordered - alphabetical, Christmas tree, rev-Xmas-tree or at the end if there doesn't seem to be any matching order. * If the script can't find a place to put a new include (mostly because the file doesn't have fitting include block), it prints out an error message indicating which .h file needs to be added to the file. The conversion was done in the following steps. 1. The initial automatic conversion of all .c files updated slightly over 4000 files, deleting around 700 includes and adding ~480 gfp.h and ~3000 slab.h inclusions. The script emitted errors for ~400 files. 2. Each error was manually checked. Some didn't need the inclusion, some needed manual addition while adding it to implementation .h or embedding .c file was more appropriate for others. This step added inclusions to around 150 files. 3. The script was run again and the output was compared to the edits from #2 to make sure no file was left behind. 4. Several build tests were done and a couple of problems were fixed. e.g. lib/decompress_*.c used malloc/free() wrappers around slab APIs requiring slab.h to be added manually. 5. The script was run on all .h files but without automatically editing them as sprinkling gfp.h and slab.h inclusions around .h files could easily lead to inclusion dependency hell. Most gfp.h inclusion directives were ignored as stuff from gfp.h was usually wildly available and often used in preprocessor macros. Each slab.h inclusion directive was examined and added manually as necessary. 6. percpu.h was updated not to include slab.h. 7. Build test were done on the following configurations and failures were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my distributed build env didn't work with gcov compiles) and a few more options had to be turned off depending on archs to make things build (like ipr on powerpc/64 which failed due to missing writeq). * x86 and x86_64 UP and SMP allmodconfig and a custom test config. * powerpc and powerpc64 SMP allmodconfig * sparc and sparc64 SMP allmodconfig * ia64 SMP allmodconfig * s390 SMP allmodconfig * alpha SMP allmodconfig * um on x86_64 SMP allmodconfig 8. percpu.h modifications were reverted so that it could be applied as a separate patch and serve as bisection point. Given the fact that I had only a couple of failures from tests on step 6, I'm fairly confident about the coverage of this conversion patch. If there is a breakage, it's likely to be something in one of the arch headers which should be easily discoverable easily on most builds of the specific arch. Signed-off-by: Tejun Heo <tj@kernel.org> Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 16:04:11 +08:00
#include <linux/slab.h>
#include <linux/in.h>
#include <linux/in6.h>
#include <linux/dlmconstants.h>
#include <net/ipv6.h>
#include <net/sock.h>
#include "config.h"
#include "midcomms.h"
#include "lowcomms.h"
/*
* /config/dlm/<cluster>/spaces/<space>/nodes/<node>/nodeid
* /config/dlm/<cluster>/spaces/<space>/nodes/<node>/weight
* /config/dlm/<cluster>/comms/<comm>/nodeid
* /config/dlm/<cluster>/comms/<comm>/local
* /config/dlm/<cluster>/comms/<comm>/addr (write only)
* /config/dlm/<cluster>/comms/<comm>/addr_list (read only)
* The <cluster> level is useless, but I haven't figured out how to avoid it.
*/
static struct config_group *space_list;
static struct config_group *comm_list;
static struct dlm_comm *local_comm;
static uint32_t dlm_comm_count;
struct dlm_clusters;
struct dlm_cluster;
struct dlm_spaces;
struct dlm_space;
struct dlm_comms;
struct dlm_comm;
struct dlm_nodes;
struct dlm_node;
static struct config_group *make_cluster(struct config_group *, const char *);
static void drop_cluster(struct config_group *, struct config_item *);
static void release_cluster(struct config_item *);
static struct config_group *make_space(struct config_group *, const char *);
static void drop_space(struct config_group *, struct config_item *);
static void release_space(struct config_item *);
static struct config_item *make_comm(struct config_group *, const char *);
static void drop_comm(struct config_group *, struct config_item *);
static void release_comm(struct config_item *);
static struct config_item *make_node(struct config_group *, const char *);
static void drop_node(struct config_group *, struct config_item *);
static void release_node(struct config_item *);
static struct configfs_attribute *comm_attrs[];
static struct configfs_attribute *node_attrs[];
struct dlm_cluster {
struct config_group group;
unsigned int cl_tcp_port;
unsigned int cl_buffer_size;
unsigned int cl_rsbtbl_size;
unsigned int cl_recover_timer;
unsigned int cl_toss_secs;
unsigned int cl_scan_secs;
unsigned int cl_log_debug;
unsigned int cl_log_info;
unsigned int cl_protocol;
unsigned int cl_mark;
#ifdef CONFIG_DLM_DEPRECATED_API
unsigned int cl_timewarn_cs;
#endif
unsigned int cl_new_rsb_count;
unsigned int cl_recover_callbacks;
char cl_cluster_name[DLM_LOCKSPACE_LEN];
fs: dlm: fix memory leak when fenced I got some kmemleak report when a node was fenced. The user space tool dlm_controld will therefore run some rmdir() in dlm configfs which was triggering some memleaks. This patch stores the sps and cms attributes which stores some handling for subdirectories of the configfs cluster entry and free them if they get released as the parent directory gets freed. unreferenced object 0xffff88810d9e3e00 (size 192): comm "dlm_controld", pid 342, jiffies 4294698126 (age 55438.801s) hex dump (first 32 bytes): 00 00 00 00 00 00 00 00 73 70 61 63 65 73 00 00 ........spaces.. 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ backtrace: [<00000000db8b640b>] make_cluster+0x5d/0x360 [<000000006a571db4>] configfs_mkdir+0x274/0x730 [<00000000b094501c>] vfs_mkdir+0x27e/0x340 [<0000000058b0adaf>] do_mkdirat+0xff/0x1b0 [<00000000d1ffd156>] do_syscall_64+0x40/0x80 [<00000000ab1408c8>] entry_SYSCALL_64_after_hwframe+0x44/0xae unreferenced object 0xffff88810d9e3a00 (size 192): comm "dlm_controld", pid 342, jiffies 4294698126 (age 55438.801s) hex dump (first 32 bytes): 00 00 00 00 00 00 00 00 63 6f 6d 6d 73 00 00 00 ........comms... 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ backtrace: [<00000000a7ef6ad2>] make_cluster+0x82/0x360 [<000000006a571db4>] configfs_mkdir+0x274/0x730 [<00000000b094501c>] vfs_mkdir+0x27e/0x340 [<0000000058b0adaf>] do_mkdirat+0xff/0x1b0 [<00000000d1ffd156>] do_syscall_64+0x40/0x80 [<00000000ab1408c8>] entry_SYSCALL_64_after_hwframe+0x44/0xae Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-06-02 21:45:16 +08:00
struct dlm_spaces *sps;
struct dlm_comms *cms;
};
static struct dlm_cluster *config_item_to_cluster(struct config_item *i)
{
return i ? container_of(to_config_group(i), struct dlm_cluster, group) :
NULL;
}
enum {
CLUSTER_ATTR_TCP_PORT = 0,
CLUSTER_ATTR_BUFFER_SIZE,
CLUSTER_ATTR_RSBTBL_SIZE,
CLUSTER_ATTR_RECOVER_TIMER,
CLUSTER_ATTR_TOSS_SECS,
CLUSTER_ATTR_SCAN_SECS,
CLUSTER_ATTR_LOG_DEBUG,
CLUSTER_ATTR_LOG_INFO,
CLUSTER_ATTR_PROTOCOL,
CLUSTER_ATTR_MARK,
#ifdef CONFIG_DLM_DEPRECATED_API
CLUSTER_ATTR_TIMEWARN_CS,
#endif
CLUSTER_ATTR_NEW_RSB_COUNT,
CLUSTER_ATTR_RECOVER_CALLBACKS,
CLUSTER_ATTR_CLUSTER_NAME,
};
static ssize_t cluster_cluster_name_show(struct config_item *item, char *buf)
{
struct dlm_cluster *cl = config_item_to_cluster(item);
return sprintf(buf, "%s\n", cl->cl_cluster_name);
}
static ssize_t cluster_cluster_name_store(struct config_item *item,
const char *buf, size_t len)
{
struct dlm_cluster *cl = config_item_to_cluster(item);
strlcpy(dlm_config.ci_cluster_name, buf,
sizeof(dlm_config.ci_cluster_name));
strlcpy(cl->cl_cluster_name, buf, sizeof(cl->cl_cluster_name));
return len;
}
CONFIGFS_ATTR(cluster_, cluster_name);
static ssize_t cluster_set(struct dlm_cluster *cl, unsigned int *cl_field,
int *info_field, int (*check_cb)(unsigned int x),
const char *buf, size_t len)
{
unsigned int x;
int rc;
if (!capable(CAP_SYS_ADMIN))
return -EPERM;
rc = kstrtouint(buf, 0, &x);
if (rc)
return rc;
if (check_cb) {
rc = check_cb(x);
if (rc)
return rc;
}
*cl_field = x;
*info_field = x;
return len;
}
#define CLUSTER_ATTR(name, check_cb) \
static ssize_t cluster_##name##_store(struct config_item *item, \
const char *buf, size_t len) \
{ \
struct dlm_cluster *cl = config_item_to_cluster(item); \
return cluster_set(cl, &cl->cl_##name, &dlm_config.ci_##name, \
check_cb, buf, len); \
} \
static ssize_t cluster_##name##_show(struct config_item *item, char *buf) \
{ \
struct dlm_cluster *cl = config_item_to_cluster(item); \
return snprintf(buf, PAGE_SIZE, "%u\n", cl->cl_##name); \
} \
CONFIGFS_ATTR(cluster_, name);
static int dlm_check_protocol_and_dlm_running(unsigned int x)
{
switch (x) {
case 0:
/* TCP */
break;
case 1:
/* SCTP */
break;
default:
return -EINVAL;
}
if (dlm_allow_conn)
return -EBUSY;
return 0;
}
static int dlm_check_zero_and_dlm_running(unsigned int x)
{
if (!x)
return -EINVAL;
if (dlm_allow_conn)
return -EBUSY;
return 0;
}
static int dlm_check_zero(unsigned int x)
{
if (!x)
return -EINVAL;
return 0;
}
static int dlm_check_buffer_size(unsigned int x)
{
if (x < DLM_MAX_SOCKET_BUFSIZE)
return -EINVAL;
return 0;
}
CLUSTER_ATTR(tcp_port, dlm_check_zero_and_dlm_running);
CLUSTER_ATTR(buffer_size, dlm_check_buffer_size);
CLUSTER_ATTR(rsbtbl_size, dlm_check_zero);
CLUSTER_ATTR(recover_timer, dlm_check_zero);
CLUSTER_ATTR(toss_secs, dlm_check_zero);
CLUSTER_ATTR(scan_secs, dlm_check_zero);
CLUSTER_ATTR(log_debug, NULL);
CLUSTER_ATTR(log_info, NULL);
CLUSTER_ATTR(protocol, dlm_check_protocol_and_dlm_running);
CLUSTER_ATTR(mark, NULL);
#ifdef CONFIG_DLM_DEPRECATED_API
CLUSTER_ATTR(timewarn_cs, dlm_check_zero);
#endif
CLUSTER_ATTR(new_rsb_count, NULL);
CLUSTER_ATTR(recover_callbacks, NULL);
static struct configfs_attribute *cluster_attrs[] = {
[CLUSTER_ATTR_TCP_PORT] = &cluster_attr_tcp_port,
[CLUSTER_ATTR_BUFFER_SIZE] = &cluster_attr_buffer_size,
[CLUSTER_ATTR_RSBTBL_SIZE] = &cluster_attr_rsbtbl_size,
[CLUSTER_ATTR_RECOVER_TIMER] = &cluster_attr_recover_timer,
[CLUSTER_ATTR_TOSS_SECS] = &cluster_attr_toss_secs,
[CLUSTER_ATTR_SCAN_SECS] = &cluster_attr_scan_secs,
[CLUSTER_ATTR_LOG_DEBUG] = &cluster_attr_log_debug,
[CLUSTER_ATTR_LOG_INFO] = &cluster_attr_log_info,
[CLUSTER_ATTR_PROTOCOL] = &cluster_attr_protocol,
[CLUSTER_ATTR_MARK] = &cluster_attr_mark,
#ifdef CONFIG_DLM_DEPRECATED_API
[CLUSTER_ATTR_TIMEWARN_CS] = &cluster_attr_timewarn_cs,
#endif
[CLUSTER_ATTR_NEW_RSB_COUNT] = &cluster_attr_new_rsb_count,
[CLUSTER_ATTR_RECOVER_CALLBACKS] = &cluster_attr_recover_callbacks,
[CLUSTER_ATTR_CLUSTER_NAME] = &cluster_attr_cluster_name,
NULL,
};
enum {
COMM_ATTR_NODEID = 0,
COMM_ATTR_LOCAL,
COMM_ATTR_ADDR,
COMM_ATTR_ADDR_LIST,
COMM_ATTR_MARK,
};
enum {
NODE_ATTR_NODEID = 0,
NODE_ATTR_WEIGHT,
};
struct dlm_clusters {
struct configfs_subsystem subsys;
};
struct dlm_spaces {
struct config_group ss_group;
};
struct dlm_space {
struct config_group group;
struct list_head members;
struct mutex members_lock;
int members_count;
struct dlm_nodes *nds;
};
struct dlm_comms {
struct config_group cs_group;
};
struct dlm_comm {
struct config_item item;
int seq;
int nodeid;
int local;
int addr_count;
unsigned int mark;
struct sockaddr_storage *addr[DLM_MAX_ADDR_COUNT];
};
struct dlm_nodes {
struct config_group ns_group;
};
struct dlm_node {
struct config_item item;
struct list_head list; /* space->members */
int nodeid;
int weight;
int new;
int comm_seq; /* copy of cm->seq when nd->nodeid is set */
};
static struct configfs_group_operations clusters_ops = {
.make_group = make_cluster,
.drop_item = drop_cluster,
};
static struct configfs_item_operations cluster_ops = {
.release = release_cluster,
};
static struct configfs_group_operations spaces_ops = {
.make_group = make_space,
.drop_item = drop_space,
};
static struct configfs_item_operations space_ops = {
.release = release_space,
};
static struct configfs_group_operations comms_ops = {
.make_item = make_comm,
.drop_item = drop_comm,
};
static struct configfs_item_operations comm_ops = {
.release = release_comm,
};
static struct configfs_group_operations nodes_ops = {
.make_item = make_node,
.drop_item = drop_node,
};
static struct configfs_item_operations node_ops = {
.release = release_node,
};
static const struct config_item_type clusters_type = {
.ct_group_ops = &clusters_ops,
.ct_owner = THIS_MODULE,
};
static const struct config_item_type cluster_type = {
.ct_item_ops = &cluster_ops,
.ct_attrs = cluster_attrs,
.ct_owner = THIS_MODULE,
};
static const struct config_item_type spaces_type = {
.ct_group_ops = &spaces_ops,
.ct_owner = THIS_MODULE,
};
static const struct config_item_type space_type = {
.ct_item_ops = &space_ops,
.ct_owner = THIS_MODULE,
};
static const struct config_item_type comms_type = {
.ct_group_ops = &comms_ops,
.ct_owner = THIS_MODULE,
};
static const struct config_item_type comm_type = {
.ct_item_ops = &comm_ops,
.ct_attrs = comm_attrs,
.ct_owner = THIS_MODULE,
};
static const struct config_item_type nodes_type = {
.ct_group_ops = &nodes_ops,
.ct_owner = THIS_MODULE,
};
static const struct config_item_type node_type = {
.ct_item_ops = &node_ops,
.ct_attrs = node_attrs,
.ct_owner = THIS_MODULE,
};
static struct dlm_space *config_item_to_space(struct config_item *i)
{
return i ? container_of(to_config_group(i), struct dlm_space, group) :
NULL;
}
static struct dlm_comm *config_item_to_comm(struct config_item *i)
{
return i ? container_of(i, struct dlm_comm, item) : NULL;
}
static struct dlm_node *config_item_to_node(struct config_item *i)
{
return i ? container_of(i, struct dlm_node, item) : NULL;
}
static struct config_group *make_cluster(struct config_group *g,
const char *name)
{
struct dlm_cluster *cl = NULL;
struct dlm_spaces *sps = NULL;
struct dlm_comms *cms = NULL;
cl = kzalloc(sizeof(struct dlm_cluster), GFP_NOFS);
sps = kzalloc(sizeof(struct dlm_spaces), GFP_NOFS);
cms = kzalloc(sizeof(struct dlm_comms), GFP_NOFS);
if (!cl || !sps || !cms)
goto fail;
fs: dlm: fix memory leak when fenced I got some kmemleak report when a node was fenced. The user space tool dlm_controld will therefore run some rmdir() in dlm configfs which was triggering some memleaks. This patch stores the sps and cms attributes which stores some handling for subdirectories of the configfs cluster entry and free them if they get released as the parent directory gets freed. unreferenced object 0xffff88810d9e3e00 (size 192): comm "dlm_controld", pid 342, jiffies 4294698126 (age 55438.801s) hex dump (first 32 bytes): 00 00 00 00 00 00 00 00 73 70 61 63 65 73 00 00 ........spaces.. 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ backtrace: [<00000000db8b640b>] make_cluster+0x5d/0x360 [<000000006a571db4>] configfs_mkdir+0x274/0x730 [<00000000b094501c>] vfs_mkdir+0x27e/0x340 [<0000000058b0adaf>] do_mkdirat+0xff/0x1b0 [<00000000d1ffd156>] do_syscall_64+0x40/0x80 [<00000000ab1408c8>] entry_SYSCALL_64_after_hwframe+0x44/0xae unreferenced object 0xffff88810d9e3a00 (size 192): comm "dlm_controld", pid 342, jiffies 4294698126 (age 55438.801s) hex dump (first 32 bytes): 00 00 00 00 00 00 00 00 63 6f 6d 6d 73 00 00 00 ........comms... 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ backtrace: [<00000000a7ef6ad2>] make_cluster+0x82/0x360 [<000000006a571db4>] configfs_mkdir+0x274/0x730 [<00000000b094501c>] vfs_mkdir+0x27e/0x340 [<0000000058b0adaf>] do_mkdirat+0xff/0x1b0 [<00000000d1ffd156>] do_syscall_64+0x40/0x80 [<00000000ab1408c8>] entry_SYSCALL_64_after_hwframe+0x44/0xae Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-06-02 21:45:16 +08:00
cl->sps = sps;
cl->cms = cms;
config_group_init_type_name(&cl->group, name, &cluster_type);
config_group_init_type_name(&sps->ss_group, "spaces", &spaces_type);
config_group_init_type_name(&cms->cs_group, "comms", &comms_type);
configfs_add_default_group(&sps->ss_group, &cl->group);
configfs_add_default_group(&cms->cs_group, &cl->group);
cl->cl_tcp_port = dlm_config.ci_tcp_port;
cl->cl_buffer_size = dlm_config.ci_buffer_size;
cl->cl_rsbtbl_size = dlm_config.ci_rsbtbl_size;
cl->cl_recover_timer = dlm_config.ci_recover_timer;
cl->cl_toss_secs = dlm_config.ci_toss_secs;
cl->cl_scan_secs = dlm_config.ci_scan_secs;
cl->cl_log_debug = dlm_config.ci_log_debug;
cl->cl_log_info = dlm_config.ci_log_info;
cl->cl_protocol = dlm_config.ci_protocol;
#ifdef CONFIG_DLM_DEPRECATED_API
cl->cl_timewarn_cs = dlm_config.ci_timewarn_cs;
#endif
cl->cl_new_rsb_count = dlm_config.ci_new_rsb_count;
cl->cl_recover_callbacks = dlm_config.ci_recover_callbacks;
memcpy(cl->cl_cluster_name, dlm_config.ci_cluster_name,
DLM_LOCKSPACE_LEN);
space_list = &sps->ss_group;
comm_list = &cms->cs_group;
return &cl->group;
fail:
kfree(cl);
kfree(sps);
kfree(cms);
return ERR_PTR(-ENOMEM);
}
static void drop_cluster(struct config_group *g, struct config_item *i)
{
struct dlm_cluster *cl = config_item_to_cluster(i);
configfs_remove_default_groups(&cl->group);
space_list = NULL;
comm_list = NULL;
config_item_put(i);
}
static void release_cluster(struct config_item *i)
{
struct dlm_cluster *cl = config_item_to_cluster(i);
fs: dlm: fix memory leak when fenced I got some kmemleak report when a node was fenced. The user space tool dlm_controld will therefore run some rmdir() in dlm configfs which was triggering some memleaks. This patch stores the sps and cms attributes which stores some handling for subdirectories of the configfs cluster entry and free them if they get released as the parent directory gets freed. unreferenced object 0xffff88810d9e3e00 (size 192): comm "dlm_controld", pid 342, jiffies 4294698126 (age 55438.801s) hex dump (first 32 bytes): 00 00 00 00 00 00 00 00 73 70 61 63 65 73 00 00 ........spaces.. 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ backtrace: [<00000000db8b640b>] make_cluster+0x5d/0x360 [<000000006a571db4>] configfs_mkdir+0x274/0x730 [<00000000b094501c>] vfs_mkdir+0x27e/0x340 [<0000000058b0adaf>] do_mkdirat+0xff/0x1b0 [<00000000d1ffd156>] do_syscall_64+0x40/0x80 [<00000000ab1408c8>] entry_SYSCALL_64_after_hwframe+0x44/0xae unreferenced object 0xffff88810d9e3a00 (size 192): comm "dlm_controld", pid 342, jiffies 4294698126 (age 55438.801s) hex dump (first 32 bytes): 00 00 00 00 00 00 00 00 63 6f 6d 6d 73 00 00 00 ........comms... 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ backtrace: [<00000000a7ef6ad2>] make_cluster+0x82/0x360 [<000000006a571db4>] configfs_mkdir+0x274/0x730 [<00000000b094501c>] vfs_mkdir+0x27e/0x340 [<0000000058b0adaf>] do_mkdirat+0xff/0x1b0 [<00000000d1ffd156>] do_syscall_64+0x40/0x80 [<00000000ab1408c8>] entry_SYSCALL_64_after_hwframe+0x44/0xae Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-06-02 21:45:16 +08:00
kfree(cl->sps);
kfree(cl->cms);
kfree(cl);
}
static struct config_group *make_space(struct config_group *g, const char *name)
{
struct dlm_space *sp = NULL;
struct dlm_nodes *nds = NULL;
sp = kzalloc(sizeof(struct dlm_space), GFP_NOFS);
nds = kzalloc(sizeof(struct dlm_nodes), GFP_NOFS);
if (!sp || !nds)
goto fail;
config_group_init_type_name(&sp->group, name, &space_type);
config_group_init_type_name(&nds->ns_group, "nodes", &nodes_type);
configfs_add_default_group(&nds->ns_group, &sp->group);
INIT_LIST_HEAD(&sp->members);
mutex_init(&sp->members_lock);
sp->members_count = 0;
sp->nds = nds;
return &sp->group;
fail:
kfree(sp);
kfree(nds);
return ERR_PTR(-ENOMEM);
}
static void drop_space(struct config_group *g, struct config_item *i)
{
struct dlm_space *sp = config_item_to_space(i);
/* assert list_empty(&sp->members) */
configfs_remove_default_groups(&sp->group);
config_item_put(i);
}
static void release_space(struct config_item *i)
{
struct dlm_space *sp = config_item_to_space(i);
kfree(sp->nds);
kfree(sp);
}
static struct config_item *make_comm(struct config_group *g, const char *name)
{
struct dlm_comm *cm;
cm = kzalloc(sizeof(struct dlm_comm), GFP_NOFS);
if (!cm)
return ERR_PTR(-ENOMEM);
config_item_init_type_name(&cm->item, name, &comm_type);
cm->seq = dlm_comm_count++;
if (!cm->seq)
cm->seq = dlm_comm_count++;
cm->nodeid = -1;
cm->local = 0;
cm->addr_count = 0;
cm->mark = 0;
return &cm->item;
}
static void drop_comm(struct config_group *g, struct config_item *i)
{
struct dlm_comm *cm = config_item_to_comm(i);
if (local_comm == cm)
local_comm = NULL;
dlm_midcomms_close(cm->nodeid);
while (cm->addr_count--)
kfree(cm->addr[cm->addr_count]);
config_item_put(i);
}
static void release_comm(struct config_item *i)
{
struct dlm_comm *cm = config_item_to_comm(i);
kfree(cm);
}
static struct config_item *make_node(struct config_group *g, const char *name)
{
struct dlm_space *sp = config_item_to_space(g->cg_item.ci_parent);
struct dlm_node *nd;
nd = kzalloc(sizeof(struct dlm_node), GFP_NOFS);
if (!nd)
return ERR_PTR(-ENOMEM);
config_item_init_type_name(&nd->item, name, &node_type);
nd->nodeid = -1;
nd->weight = 1; /* default weight of 1 if none is set */
nd->new = 1; /* set to 0 once it's been read by dlm_nodeid_list() */
mutex_lock(&sp->members_lock);
list_add(&nd->list, &sp->members);
sp->members_count++;
mutex_unlock(&sp->members_lock);
return &nd->item;
}
static void drop_node(struct config_group *g, struct config_item *i)
{
struct dlm_space *sp = config_item_to_space(g->cg_item.ci_parent);
struct dlm_node *nd = config_item_to_node(i);
mutex_lock(&sp->members_lock);
list_del(&nd->list);
sp->members_count--;
mutex_unlock(&sp->members_lock);
config_item_put(i);
}
static void release_node(struct config_item *i)
{
struct dlm_node *nd = config_item_to_node(i);
kfree(nd);
}
static struct dlm_clusters clusters_root = {
.subsys = {
.su_group = {
.cg_item = {
.ci_namebuf = "dlm",
.ci_type = &clusters_type,
},
},
},
};
int __init dlm_config_init(void)
{
config_group_init(&clusters_root.subsys.su_group);
mutex_init(&clusters_root.subsys.su_mutex);
return configfs_register_subsystem(&clusters_root.subsys);
}
void dlm_config_exit(void)
{
configfs_unregister_subsystem(&clusters_root.subsys);
}
/*
* Functions for user space to read/write attributes
*/
static ssize_t comm_nodeid_show(struct config_item *item, char *buf)
{
return sprintf(buf, "%d\n", config_item_to_comm(item)->nodeid);
}
static ssize_t comm_nodeid_store(struct config_item *item, const char *buf,
size_t len)
{
int rc = kstrtoint(buf, 0, &config_item_to_comm(item)->nodeid);
if (rc)
return rc;
return len;
}
static ssize_t comm_local_show(struct config_item *item, char *buf)
{
return sprintf(buf, "%d\n", config_item_to_comm(item)->local);
}
static ssize_t comm_local_store(struct config_item *item, const char *buf,
size_t len)
{
struct dlm_comm *cm = config_item_to_comm(item);
int rc = kstrtoint(buf, 0, &cm->local);
if (rc)
return rc;
if (cm->local && !local_comm)
local_comm = cm;
return len;
}
static ssize_t comm_addr_store(struct config_item *item, const char *buf,
size_t len)
{
struct dlm_comm *cm = config_item_to_comm(item);
struct sockaddr_storage *addr;
int rv;
if (len != sizeof(struct sockaddr_storage))
return -EINVAL;
if (cm->addr_count >= DLM_MAX_ADDR_COUNT)
return -ENOSPC;
addr = kzalloc(sizeof(*addr), GFP_NOFS);
if (!addr)
return -ENOMEM;
memcpy(addr, buf, len);
rv = dlm_lowcomms_addr(cm->nodeid, addr, len);
if (rv) {
kfree(addr);
return rv;
}
cm->addr[cm->addr_count++] = addr;
return len;
}
static ssize_t comm_addr_list_show(struct config_item *item, char *buf)
{
struct dlm_comm *cm = config_item_to_comm(item);
ssize_t s;
ssize_t allowance;
int i;
struct sockaddr_storage *addr;
struct sockaddr_in *addr_in;
struct sockaddr_in6 *addr_in6;
/* Taken from ip6_addr_string() defined in lib/vsprintf.c */
char buf0[sizeof("AF_INET6 xxxx:xxxx:xxxx:xxxx:xxxx:xxxx:255.255.255.255\n")];
/* Derived from SIMPLE_ATTR_SIZE of fs/configfs/file.c */
allowance = 4096;
buf[0] = '\0';
for (i = 0; i < cm->addr_count; i++) {
addr = cm->addr[i];
switch(addr->ss_family) {
case AF_INET:
addr_in = (struct sockaddr_in *)addr;
s = sprintf(buf0, "AF_INET %pI4\n", &addr_in->sin_addr.s_addr);
break;
case AF_INET6:
addr_in6 = (struct sockaddr_in6 *)addr;
s = sprintf(buf0, "AF_INET6 %pI6\n", &addr_in6->sin6_addr);
break;
default:
s = sprintf(buf0, "%s\n", "<UNKNOWN>");
break;
}
allowance -= s;
if (allowance >= 0)
strcat(buf, buf0);
else {
allowance += s;
break;
}
}
return 4096 - allowance;
}
static ssize_t comm_mark_show(struct config_item *item, char *buf)
{
return sprintf(buf, "%u\n", config_item_to_comm(item)->mark);
}
static ssize_t comm_mark_store(struct config_item *item, const char *buf,
size_t len)
{
fs: dlm: fix mark setting deadlock This patch fixes an deadlock issue when dlm_lowcomms_close() is called. When dlm_lowcomms_close() is called the clusters_root.subsys.su_mutex is held to remove configfs items. At this time we flushing (e.g. cancel_work_sync()) the workers of send and recv workqueue. Due the fact that we accessing configfs items (mark values), these workers will lock clusters_root.subsys.su_mutex as well which are already hold by dlm_lowcomms_close() and ends in a deadlock situation. [67170.703046] ====================================================== [67170.703965] WARNING: possible circular locking dependency detected [67170.704758] 5.11.0-rc4+ #22 Tainted: G W [67170.705433] ------------------------------------------------------ [67170.706228] dlm_controld/280 is trying to acquire lock: [67170.706915] ffff9f2f475a6948 ((wq_completion)dlm_recv){+.+.}-{0:0}, at: __flush_work+0x203/0x4c0 [67170.708026] but task is already holding lock: [67170.708758] ffffffffa132f878 (&clusters_root.subsys.su_mutex){+.+.}-{3:3}, at: configfs_rmdir+0x29b/0x310 [67170.710016] which lock already depends on the new lock. The new behaviour adds the mark value to the node address configuration which doesn't require to held the clusters_root.subsys.su_mutex by accessing mark values in a separate datastructure. However the mark values can be set now only after a node address was set which is the case when the user is using dlm_controld. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-03-02 06:05:09 +08:00
struct dlm_comm *comm;
unsigned int mark;
int rc;
rc = kstrtouint(buf, 0, &mark);
if (rc)
return rc;
fs: dlm: fix mark setting deadlock This patch fixes an deadlock issue when dlm_lowcomms_close() is called. When dlm_lowcomms_close() is called the clusters_root.subsys.su_mutex is held to remove configfs items. At this time we flushing (e.g. cancel_work_sync()) the workers of send and recv workqueue. Due the fact that we accessing configfs items (mark values), these workers will lock clusters_root.subsys.su_mutex as well which are already hold by dlm_lowcomms_close() and ends in a deadlock situation. [67170.703046] ====================================================== [67170.703965] WARNING: possible circular locking dependency detected [67170.704758] 5.11.0-rc4+ #22 Tainted: G W [67170.705433] ------------------------------------------------------ [67170.706228] dlm_controld/280 is trying to acquire lock: [67170.706915] ffff9f2f475a6948 ((wq_completion)dlm_recv){+.+.}-{0:0}, at: __flush_work+0x203/0x4c0 [67170.708026] but task is already holding lock: [67170.708758] ffffffffa132f878 (&clusters_root.subsys.su_mutex){+.+.}-{3:3}, at: configfs_rmdir+0x29b/0x310 [67170.710016] which lock already depends on the new lock. The new behaviour adds the mark value to the node address configuration which doesn't require to held the clusters_root.subsys.su_mutex by accessing mark values in a separate datastructure. However the mark values can be set now only after a node address was set which is the case when the user is using dlm_controld. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-03-02 06:05:09 +08:00
if (mark == 0)
mark = dlm_config.ci_mark;
comm = config_item_to_comm(item);
rc = dlm_lowcomms_nodes_set_mark(comm->nodeid, mark);
if (rc)
return rc;
comm->mark = mark;
return len;
}
CONFIGFS_ATTR(comm_, nodeid);
CONFIGFS_ATTR(comm_, local);
CONFIGFS_ATTR(comm_, mark);
CONFIGFS_ATTR_WO(comm_, addr);
CONFIGFS_ATTR_RO(comm_, addr_list);
static struct configfs_attribute *comm_attrs[] = {
[COMM_ATTR_NODEID] = &comm_attr_nodeid,
[COMM_ATTR_LOCAL] = &comm_attr_local,
[COMM_ATTR_ADDR] = &comm_attr_addr,
[COMM_ATTR_ADDR_LIST] = &comm_attr_addr_list,
[COMM_ATTR_MARK] = &comm_attr_mark,
NULL,
};
static ssize_t node_nodeid_show(struct config_item *item, char *buf)
{
return sprintf(buf, "%d\n", config_item_to_node(item)->nodeid);
}
static ssize_t node_nodeid_store(struct config_item *item, const char *buf,
size_t len)
{
struct dlm_node *nd = config_item_to_node(item);
uint32_t seq = 0;
int rc = kstrtoint(buf, 0, &nd->nodeid);
if (rc)
return rc;
dlm_comm_seq(nd->nodeid, &seq);
nd->comm_seq = seq;
return len;
}
static ssize_t node_weight_show(struct config_item *item, char *buf)
{
return sprintf(buf, "%d\n", config_item_to_node(item)->weight);
}
static ssize_t node_weight_store(struct config_item *item, const char *buf,
size_t len)
{
int rc = kstrtoint(buf, 0, &config_item_to_node(item)->weight);
if (rc)
return rc;
return len;
}
CONFIGFS_ATTR(node_, nodeid);
CONFIGFS_ATTR(node_, weight);
static struct configfs_attribute *node_attrs[] = {
[NODE_ATTR_NODEID] = &node_attr_nodeid,
[NODE_ATTR_WEIGHT] = &node_attr_weight,
NULL,
};
/*
* Functions for the dlm to get the info that's been configured
*/
static struct dlm_space *get_space(char *name)
{
struct config_item *i;
if (!space_list)
return NULL;
mutex_lock(&space_list->cg_subsys->su_mutex);
i = config_group_find_item(space_list, name);
mutex_unlock(&space_list->cg_subsys->su_mutex);
return config_item_to_space(i);
}
static void put_space(struct dlm_space *sp)
{
config_item_put(&sp->group.cg_item);
}
static struct dlm_comm *get_comm(int nodeid)
{
struct config_item *i;
struct dlm_comm *cm = NULL;
int found = 0;
if (!comm_list)
return NULL;
mutex_lock(&clusters_root.subsys.su_mutex);
list_for_each_entry(i, &comm_list->cg_children, ci_entry) {
cm = config_item_to_comm(i);
if (cm->nodeid != nodeid)
continue;
found = 1;
config_item_get(i);
break;
}
mutex_unlock(&clusters_root.subsys.su_mutex);
if (!found)
cm = NULL;
return cm;
}
static void put_comm(struct dlm_comm *cm)
{
config_item_put(&cm->item);
}
/* caller must free mem */
int dlm_config_nodes(char *lsname, struct dlm_config_node **nodes_out,
int *count_out)
{
struct dlm_space *sp;
struct dlm_node *nd;
struct dlm_config_node *nodes, *node;
int rv, count;
sp = get_space(lsname);
if (!sp)
return -EEXIST;
mutex_lock(&sp->members_lock);
if (!sp->members_count) {
rv = -EINVAL;
printk(KERN_ERR "dlm: zero members_count\n");
goto out;
}
count = sp->members_count;
nodes = kcalloc(count, sizeof(struct dlm_config_node), GFP_NOFS);
if (!nodes) {
rv = -ENOMEM;
goto out;
}
node = nodes;
list_for_each_entry(nd, &sp->members, list) {
node->nodeid = nd->nodeid;
node->weight = nd->weight;
node->new = nd->new;
node->comm_seq = nd->comm_seq;
node++;
nd->new = 0;
}
*count_out = count;
*nodes_out = nodes;
rv = 0;
out:
mutex_unlock(&sp->members_lock);
put_space(sp);
return rv;
}
int dlm_comm_seq(int nodeid, uint32_t *seq)
{
struct dlm_comm *cm = get_comm(nodeid);
if (!cm)
return -EEXIST;
*seq = cm->seq;
put_comm(cm);
return 0;
}
int dlm_our_nodeid(void)
{
return local_comm ? local_comm->nodeid : 0;
}
/* num 0 is first addr, num 1 is second addr */
int dlm_our_addr(struct sockaddr_storage *addr, int num)
{
if (!local_comm)
return -1;
if (num + 1 > local_comm->addr_count)
return -1;
memcpy(addr, local_comm->addr[num], sizeof(*addr));
return 0;
}
/* Config file defaults */
#define DEFAULT_TCP_PORT 21064
#define DEFAULT_RSBTBL_SIZE 1024
#define DEFAULT_RECOVER_TIMER 5
#define DEFAULT_TOSS_SECS 10
#define DEFAULT_SCAN_SECS 5
#define DEFAULT_LOG_DEBUG 0
#define DEFAULT_LOG_INFO 1
#define DEFAULT_PROTOCOL DLM_PROTO_TCP
#define DEFAULT_MARK 0
#ifdef CONFIG_DLM_DEPRECATED_API
#define DEFAULT_TIMEWARN_CS 500 /* 5 sec = 500 centiseconds */
#endif
#define DEFAULT_NEW_RSB_COUNT 128
#define DEFAULT_RECOVER_CALLBACKS 0
#define DEFAULT_CLUSTER_NAME ""
struct dlm_config_info dlm_config = {
.ci_tcp_port = DEFAULT_TCP_PORT,
.ci_buffer_size = DLM_MAX_SOCKET_BUFSIZE,
.ci_rsbtbl_size = DEFAULT_RSBTBL_SIZE,
.ci_recover_timer = DEFAULT_RECOVER_TIMER,
.ci_toss_secs = DEFAULT_TOSS_SECS,
.ci_scan_secs = DEFAULT_SCAN_SECS,
.ci_log_debug = DEFAULT_LOG_DEBUG,
.ci_log_info = DEFAULT_LOG_INFO,
.ci_protocol = DEFAULT_PROTOCOL,
.ci_mark = DEFAULT_MARK,
#ifdef CONFIG_DLM_DEPRECATED_API
.ci_timewarn_cs = DEFAULT_TIMEWARN_CS,
#endif
.ci_new_rsb_count = DEFAULT_NEW_RSB_COUNT,
.ci_recover_callbacks = DEFAULT_RECOVER_CALLBACKS,
.ci_cluster_name = DEFAULT_CLUSTER_NAME
};