Hmm, maybe it will be correctly?
void worker_loop(int id)
{
dmq_worker_t* worker;
dmq_job_t* current_job;
peer_reponse_t peer_response;
int ret_value;
int not_parsed;
dmq_node_t *dmq_node = NULL;
worker = &workers[id];
for(;;) {
Are you sure it is the right fix? Is there no possibility that someone else is
parsing the From header before clonning (e.g., from config file)?
---
Reply to this email directly or view it on GitHub:
https://github.com/kamailio/kamailio/pull/59#issuecomment-72432180___
Merged #59.
---
Reply to this email directly or view it on GitHub:
https://github.com/kamailio/kamailio/pull/59#event-228303023___
sr-dev mailing list
sr-dev@lists.sip-router.org
http://lists.sip-router.org/cgi-bin/mailman/listinfo/sr-dev
Having investigated further, the code which calls parse_from_header() is new -
introduced for the purpose of letting the callback function know about the
sending node - and I had wrongly assumed we were already parsing the From
header prior to cloning. So apologies - I would say this is OK to me
Thanks for clarifying, Daniel.
As the From header is parsed prior to cloning (and therefore included in the
clone), it is pointless to parse it again and have to manually clean it up
later - so this part of the code needs changing anyway.
I have added a commit to master relating to the above -
The headers parsed after cloning can be detected and cleaned up, the code
should be like:
```
for( hdr=req->headers ; hdr ; hdr=hdr->next ) {
if ( hdr->parsed && hdr_allocs_parse(hdr) &&
(hdr->parsed<(void*)t->uas.request ||
hdr->parsed>=(void*)
The problem may be earlier in worker_loop():
95 if (parse_from_header(current_job->msg) <
0) {
96 LM_ERR("bad sip message or missing
From hdr\n");
97 } else {
98
Yes, i observed leak in dmq_worker process:
fm_status: count= 3648 size= 1887776 bytes from :
parser/parse_from.c: parse_from_header(63)
fm_status: count= 3648 size=418016 bytes from :
parser/parse_addr_spec.c: parse_to_param(281)
---
Reply to this email directly or view it on GitHub:
Have you observed a leak here? According to the documentation for
sip_msg_shm_clone() [1]:
"org_msg is cloned along with most of its headers and lumps into one shm memory
block (so that a shm_free() on the result will free everything)."
1.
http://rpm.kamailio.org/doxygen/sip-router/branch/mast
You can view, comment on, or merge this pull request online at:
https://github.com/kamailio/kamailio/pull/59
-- Commit Summary --
* dmq: fix memory leak
-- File Changes --
M modules/dmq/worker.c (4)
-- Patch Links --
https://github.com/kamailio/kamailio/pull/59.patch
https://github.
10 matches
Mail list logo
