sr-dev April 2022

sr-dev@lists.kamailio.org

17 participants
283 discussions

git:master:9a8974a8: htable: check if entries are initialized before sync'ing to db

by Daniel-Constantin Mierla

Module: kamailio Branch: master Commit: 9a8974a80b8ce5ffc9166e81f43f8581777cd48f URL: https://github.com/kamailio/kamailio/commit/9a8974a80b8ce5ffc9166e81f43f858… Author: Daniel-Constantin Mierla <miconda(a)gmail.com> Committer: Daniel-Constantin Mierla <miconda(a)gmail.com> Date: 2022-04-13T17:05:18+02:00 htable: check if entries are initialized before sync'ing to db --- Modified: src/modules/htable/ht_db.c --- Diff: https://github.com/kamailio/kamailio/commit/9a8974a80b8ce5ffc9166e81f43f858… Patch: https://github.com/kamailio/kamailio/commit/9a8974a80b8ce5ffc9166e81f43f858… --- diff --git a/src/modules/htable/ht_db.c b/src/modules/htable/ht_db.c index 5b9a5942ed..cbce085f6c 100644 --- a/src/modules/htable/ht_db.c +++ b/src/modules/htable/ht_db.c @@ -551,6 +551,11 @@ int ht_db_save_table(ht_t *ht, str *dbtable) time_t now; int ncols; + if(ht==NULL || ht->entries==NULL) { + LM_DBG("hash table not initialized yet\n"); + return -1; + } + if(ht_db_con==NULL) { LM_ERR("no db connection\n");

2 years

git:master:b1fa63b1: htable: track when records were successfully loaded from database

by Daniel-Constantin Mierla

Module: kamailio Branch: master Commit: b1fa63b1547fc5af6cc1e97db42ea76db1bf927b URL: https://github.com/kamailio/kamailio/commit/b1fa63b1547fc5af6cc1e97db42ea76… Author: Daniel-Constantin Mierla <miconda(a)gmail.com> Committer: Daniel-Constantin Mierla <miconda(a)gmail.com> Date: 2022-04-13T13:11:30+02:00 htable: track when records were successfully loaded from database - write back only in such case --- Modified: src/modules/htable/ht_api.c Modified: src/modules/htable/ht_api.h Modified: src/modules/htable/ht_db.c Modified: src/modules/htable/htable.c --- Diff: https://github.com/kamailio/kamailio/commit/b1fa63b1547fc5af6cc1e97db42ea76… Patch: https://github.com/kamailio/kamailio/commit/b1fa63b1547fc5af6cc1e97db42ea76… --- diff --git a/src/modules/htable/ht_api.c b/src/modules/htable/ht_api.c index 00023e7153..f212600e21 100644 --- a/src/modules/htable/ht_api.c +++ b/src/modules/htable/ht_api.c @@ -1099,7 +1099,7 @@ int ht_db_sync_tables(void) ht = _ht_root; while(ht) { - if(ht->dbtable.len>0 && ht->dbmode!=0 && ht->ncols==0) + if(ht->dbtable.len>0 && ht->dbmode!=0 && ht->dbload!=0 && ht->ncols==0) { LM_DBG("sync db table [%.*s] from ht [%.*s]\n", ht->dbtable.len, ht->dbtable.s, diff --git a/src/modules/htable/ht_api.h b/src/modules/htable/ht_api.h index 4cee835325..b3d0436ea3 100644 --- a/src/modules/htable/ht_api.h +++ b/src/modules/htable/ht_api.h @@ -63,6 +63,7 @@ typedef struct _ht unsigned int htexpire; str dbtable; int dbmode; + int dbload; int ncols; str scols[HT_MAX_COLS]; char pack[4]; diff --git a/src/modules/htable/ht_db.c b/src/modules/htable/ht_db.c index 5cdecb1814..5b9a5942ed 100644 --- a/src/modules/htable/ht_db.c +++ b/src/modules/htable/ht_db.c @@ -211,6 +211,7 @@ int ht_db_load_table(ht_t *ht, str *dbtable, int mode) LM_ERR("failed to use_table\n"); return -1; } + ht->dbload = 0; if(ht->ncols>0) { db_ord = &ht->scols[0]; for(c=0; c<ht->ncols; c++) { @@ -527,6 +528,7 @@ int ht_db_load_table(ht_t *ht, str *dbtable, int mode) ht_dbf.free_result(ht_db_con, db_res); LM_DBG("loaded %d values in hash table\n", cnt); + ht->dbload = 1; return 0; error: diff --git a/src/modules/htable/htable.c b/src/modules/htable/htable.c index f7a9cb389a..fea9786298 100644 --- a/src/modules/htable/htable.c +++ b/src/modules/htable/htable.c @@ -2053,6 +2053,8 @@ static void htable_rpc_reload(rpc_t* rpc, void* c) ht_slot_unlock(ht, i); nht.entries[i].first = first; } + ht->dbload = 1; + /* free old entries */ for(i=0; i<nht.htsize; i++) {

2 years

git:master:d542cb89: htable: reset entries field if hash table initialization fails

by Daniel-Constantin Mierla

Module: kamailio Branch: master Commit: d542cb89afb1a34cba35cbcf0f40315581fbb767 URL: https://github.com/kamailio/kamailio/commit/d542cb89afb1a34cba35cbcf0f40315… Author: Daniel-Constantin Mierla <miconda(a)gmail.com> Committer: Daniel-Constantin Mierla <miconda(a)gmail.com> Date: 2022-04-13T12:52:15+02:00 htable: reset entries field if hash table initialization fails --- Modified: src/modules/htable/ht_api.c --- Diff: https://github.com/kamailio/kamailio/commit/d542cb89afb1a34cba35cbcf0f40315… Patch: https://github.com/kamailio/kamailio/commit/d542cb89afb1a34cba35cbcf0f40315… --- diff --git a/src/modules/htable/ht_api.c b/src/modules/htable/ht_api.c index daf088d1f8..00023e7153 100644 --- a/src/modules/htable/ht_api.c +++ b/src/modules/htable/ht_api.c @@ -395,7 +395,6 @@ int ht_init_tables(void) if(ht->entries==NULL) { LM_ERR("no more shared memory for [%.*s]\n", ht->name.len, ht->name.s); - shm_free(ht); return -1; } memset(ht->entries, 0, ht->htsize*sizeof(ht_entry_t)); @@ -413,7 +412,7 @@ int ht_init_tables(void) i--; } shm_free(ht->entries); - shm_free(ht); + ht->entries = NULL; return -1; }

2 years

The recursive lock of tm module is not safe?

by progrom

Hi： See lock_hash in h_table.c: Flowing lines of code do not guarantee atomicity and I think it is not safe: -      atomic_get(&_tm_table->entries[i].locker_pid) != mypid -      lock(&_tm_table->entries[i].mutex) -      atomic_set(&_tm_table->entries[i].locker_pid, mypid) Would you mind solving my confusion?

2 years

Questions about dlg_search

by progrom

Hi： See dlg_new_dialog in dlg_handlers.c. After dlg_search, the dlg has locked and referenced, but there is no dlg_hash_release/dlg_release in some places ?

2 years

How to add CLI parameter in kamailio.cfg-Json formatting

by Jeetha P J

Hi all I want the syslog output in json structure. When I run 'kamailio --log-engine=json:M' I am getting this but only for that command.I want all syslog out in the json format. Any one please help me to achieve this Thanks & Regards Jeetha P J -- _This message contains information that may be privileged or confidential and is the property of the Quantiphi Inc and/or its affiliates_. It is intended only for the person to whom it is addressed. _If you are not the intended recipient, any review, dissemination, distribution, copying, storage or other use of all or any portion of this message is strictly prohibited. If you received this message in error, please immediately notify the sender by reply e-mail and delete this message in its *entirety*___

2 years

[kamailio/kamailio] dead lock of dialog module (Issue #3084)

by progromm

### Description See `internal_get_dlg`: ```C static inline struct dlg_cell* internal_get_dlg(unsigned int h_entry, str* callid, str* ftag, str* ttag, unsigned int* dir, int mode) { struct dlg_cell* dlg; struct dlg_entry* d_entry; d_entry = &(d_table->entries[h_entry]); dlg_lock(d_table, d_entry); for (dlg = d_entry->first ; dlg ; dlg = dlg->next) { /* Check callid / fromtag / totag */ if (match_dialog(dlg, callid, ftag, ttag, dir) == 1) { ref_dlg_unsafe(dlg, 1); if (likely(mode == 0)) { dlg_unlock(d_table, d_entry); } LM_DBG("dialog callid='%.*s' found on entry %u, dir=%d\n", callid->len, callid->s, h_entry, *dir); return dlg; } } if (likely(mode == 0)) { dlg_unlock(d_table, d_entry); } LM_DBG("no dialog callid='%.*s' found\n", callid->len, callid->s); return 0; } ``` Should judge whether `mode` is 0 before `dlg_lock` . -- Reply to this email directly or view it on GitHub: https://github.com/kamailio/kamailio/issues/3084 You are receiving this because you are subscribed to this thread. Message ID: <kamailio/kamailio/issues/3084(a)github.com>

2 years

(no subject)

by 王斌

2 years

Memleak and deadlock abount dlg_search?

by 王斌

Hi： See dlg_new_dialog in dlg_handlers.c. After dlg_search, the dlg has locked and referenced, but there is no dlg_hash_release/dlg_release in some places ? Best Regards!

2 years

[kamailio/kamailio] the recursive lock of tm module is not safe? (Issue #3083)

by progromm

### Description See `lock_hash` in `h_table.c`: ```C void lock_hash(int i) { int mypid; mypid = my_pid(); if (likely(atomic_get(&_tm_table->entries[i].locker_pid) != mypid)) { lock(&_tm_table->entries[i].mutex); atomic_set(&_tm_table->entries[i].locker_pid, mypid); } else { /* locked within the same process that called us*/ _tm_table->entries[i].rec_lock_level++; } } ``` Flowing lines of code do not guarantee atomicity and I think it is not safe: * atomic_get(&_tm_table->entries[i].locker_pid) != mypid * lock(&_tm_table->entries[i].mutex) * atomic_set(&_tm_table->entries[i].locker_pid, mypid) Would you mind solving my confusion? -- Reply to this email directly or view it on GitHub: https://github.com/kamailio/kamailio/issues/3083 You are receiving this because you are subscribed to this thread. Message ID: <kamailio/kamailio/issues/3083(a)github.com>

2 years

← Newer
1
...
17
18
19
20
21
22
23
...
29
Older →

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

sr-dev April 2022