openldap-bugs

openldap-bugs@openldap.org

1 participants
20967 discussions

Re: (ITS#8016) silence on wrong envflags for lmdb-backend
by hyc＠symas.com 05 Jan '15

05 Jan '15

leo(a)yuriev.ru wrote: > This is a multi-part message in MIME format. > --------------080401010006020303000400 > Content-Type: text/plain; charset=windows-1251; format=flowed > Content-Transfer-Encoding: 7bit > > Please review and merge. > > Leonid. > > --- > > The attached files is derived from OpenLDAP Software. All of the > modifications > to OpenLDAP Software represented in the following patch(es) were > developed by > Peter-Service LLC, Moscow, Russia. Peter-Service LLC has not assigned > rights > and/or interest in this work to any party. I, Leonid Yuriev am > authorized by > Peter-Service LLC, my employer, to release this work under the following > terms. > > Peter-Service LLC hereby places the following modifications to OpenLDAP > Software > (and only these modifications) into the public domain. Hence, these > modifications may be freely used and/or redistributed for any purpose > with or without attribution and/or other notice. > > > > --------------080401010006020303000400 > Content-Type: text/x-patch; > name="its8016.patch" > Content-Transfer-Encoding: 7bit > Content-Disposition: attachment; > filename="its8016.patch" > > commit 36ecfd44559d3b89a7dd5bcbfb6a22d8aeb8142b > Author: Leo Yuriev <leo(a)yuriev.ru> > Date: 2015-01-04 09:05:36 +0300 > > fix: report about wrong envflags in lmdb-backend instead of silently ignore them. > > diff --git a/servers/slapd/back-mdb/config.c b/servers/slapd/back-mdb/config.c > index bb9be73..68009b6 100644 > --- a/servers/slapd/back-mdb/config.c > +++ b/servers/slapd/back-mdb/config.c > @@ -620,8 +620,10 @@ mdb_cf_gen( ConfigArgs *c ) > } > mdb->mi_dbenv_flags |= mdb_envflags[j].mask; > } else { > - /* unknown keyword */ > - rc = 1; > + fprintf( stderr, "%s: " > + "unknown keyword \"%s\" in \"envflags\".\n", > + c->log, c->argv[i] ); > + return 1; > } > } > } > > --------------080401010006020303000400-- Use of fprintf(stderr) for config errors is deprecated. New code should only be using c->cr_msg. -- -- Howard Chu CTO, Symas Corp. http://www.symas.com Director, Highland Sun http://highlandsun.com/hyc/ Chief Architect, OpenLDAP http://www.openldap.org/project/

1 0

Re: (ITS#8016) silence on wrong envflags for lmdb-backend
by leo＠yuriev.ru 04 Jan '15

04 Jan '15

This is a multi-part message in MIME format. --------------080401010006020303000400 Content-Type: text/plain; charset=windows-1251; format=flowed Content-Transfer-Encoding: 7bit Please review and merge. Leonid. --- The attached files is derived from OpenLDAP Software. All of the modifications to OpenLDAP Software represented in the following patch(es) were developed by Peter-Service LLC, Moscow, Russia. Peter-Service LLC has not assigned rights and/or interest in this work to any party. I, Leonid Yuriev am authorized by Peter-Service LLC, my employer, to release this work under the following terms. Peter-Service LLC hereby places the following modifications to OpenLDAP Software (and only these modifications) into the public domain. Hence, these modifications may be freely used and/or redistributed for any purpose with or without attribution and/or other notice. --------------080401010006020303000400 Content-Type: text/x-patch; name="its8016.patch" Content-Transfer-Encoding: 7bit Content-Disposition: attachment; filename="its8016.patch" commit 36ecfd44559d3b89a7dd5bcbfb6a22d8aeb8142b Author: Leo Yuriev <leo(a)yuriev.ru> Date: 2015-01-04 09:05:36 +0300 fix: report about wrong envflags in lmdb-backend instead of silently ignore them. diff --git a/servers/slapd/back-mdb/config.c b/servers/slapd/back-mdb/config.c index bb9be73..68009b6 100644 --- a/servers/slapd/back-mdb/config.c +++ b/servers/slapd/back-mdb/config.c @@ -620,8 +620,10 @@ mdb_cf_gen( ConfigArgs *c ) } mdb->mi_dbenv_flags |= mdb_envflags[j].mask; } else { - /* unknown keyword */ - rc = 1; + fprintf( stderr, "%s: " + "unknown keyword \"%s\" in \"envflags\".\n", + c->log, c->argv[i] ); + return 1; } } } --------------080401010006020303000400--

1 0

(ITS#8016) silence on wrong envflags for lmdb-backend
by leo＠yuriev.ru 04 Jan '15

04 Jan '15

Full_Name: Leonid Yuriev Version: 2.4-HEAD OS: RHEL7 URL: ftp://ftp.openldap.org/incoming/ Submission from: (NULL) (31.130.36.33) No any error reported by lmdb-backend when wrong options was specified in envflags. For example slapd.conf: envflags This is a rebus codestyle and the technical debt Patch will be available shortly.

1 0

Re: (ITS#8011) few mistakes in lmdb-backend
by hyc＠symas.com 03 Jan '15

03 Jan '15

Леонид Юрьев wrote: > Updated. > Please pickup. > diff --git a/servers/slapd/back-mdb/dn2id.c b/servers/slapd/back-mdb/dn2id.c > index 41c4758..0890bce 100644 > --- a/servers/slapd/back-mdb/dn2id.c > +++ b/servers/slapd/back-mdb/dn2id.c > @@ -746,7 +746,8 @@ mdb_idscopes( > /* remember our chain of parents */ > id2.mid = id; > id2.mval = data; > - mdb_id2l_insert( isc->sctmp, &id2 ); > + rc = mdb_id2l_insert( isc->sctmp, &id2 ); > + assert(rc == 0); > } > ptr = data.mv_data; > ptr += data.mv_size - sizeof(ID); Unnecessary, the total number of parents is constrained by the max length of a DN which is 8192 bytes. A maximum depth DN (max number of shortest possible RDNs) will always fit in sctmp. I will commit the remainder of the patch. -- -- Howard Chu CTO, Symas Corp. http://www.symas.com Director, Highland Sun http://highlandsun.com/hyc/ Chief Architect, OpenLDAP http://www.openldap.org/project/

1 0

Re: (ITS#8012) SIGSEGV while disconnect/abandon
by hyc＠symas.com 03 Jan '15

03 Jan '15

Leonid Yuriev wrote: > Patch update: > > Detaching a syncops record from op-list CONDITIONALLY, only when it > was freed by syncprov_free_syncop. > The syncprov_drop_psearch() and syncprov_drop_psearch() now returns > a flag, which is nonzero if the given syncops was freed. Thanks, applied to master. > > Leonid. > > 29.12.2014 13:50, Leonid Yuriev ïèøåò: >> Please review attached patch and merge-in. >> >> Leonid. >> >> --- >> >> The attached files is derived from OpenLDAP Software. All of the >> modifications >> to OpenLDAP Software represented in the following patch(es) were >> developed by >> Peter-Service LLC, Moscow, Russia. Peter-Service LLC has not assigned >> rights >> and/or interest in this work to any party. I, Leonid Yuriev am >> authorized by >> Peter-Service LLC, my employer, to release this work under the >> following terms. >> >> Peter-Service LLC hereby places the following modifications to >> OpenLDAP Software >> (and only these modifications) into the public domain. Hence, these >> modifications may be freely used and/or redistributed for any purpose >> with or without attribution and/or other notice. >> > -- -- Howard Chu CTO, Symas Corp. http://www.symas.com Director, Highland Sun http://highlandsun.com/hyc/ Chief Architect, OpenLDAP http://www.openldap.org/project/

1 0

Re: (ITS#8013) SIGSEGV in test_filter()
by hyc＠symas.com 03 Jan '15

03 Jan '15

hyc(a)symas.com wrote: > Leonid Yuriev wrote: >> >> 04.01.2015 02:47, Howard Chu wrote: >>> The patch is wrong. syncprov_matchops can be called twice for the same >>> operation; toggling the flag where you've placed it means the filter >>> will be incorrect on the 2nd invocation as well as on all subsequent >>> invocations. >> >> Ok, but found another case. >> The ss->s_op->ors_filter may be updated immediately after the s_mutex >> released. > > Seems like the simpler fix then is to keep the mutex held until after > test_filter finishes. Fixed in master. -- -- Howard Chu CTO, Symas Corp. http://www.symas.com Director, Highland Sun http://highlandsun.com/hyc/ Chief Architect, OpenLDAP http://www.openldap.org/project/

1 0

Re: (ITS#8013) SIGSEGV in test_filter()
by hyc＠symas.com 03 Jan '15

03 Jan '15

Leonid Yuriev wrote: > > 04.01.2015 02:47, Howard Chu wrote: >> The patch is wrong. syncprov_matchops can be called twice for the same >> operation; toggling the flag where you've placed it means the filter >> will be incorrect on the 2nd invocation as well as on all subsequent >> invocations. > > Ok, but found another case. > The ss->s_op->ors_filter may be updated immediately after the s_mutex > released. Seems like the simpler fix then is to keep the mutex held until after test_filter finishes. > > This is a simple change for reproducing the bug. > > diff --git a/servers/slapd/overlays/syncprov.c > b/servers/slapd/overlays/syncprov.c > index 5ef2eaa..f579faa 100644 > --- a/servers/slapd/overlays/syncprov.c > +++ b/servers/slapd/overlays/syncprov.c > @@ -1295,6 +1295,7 @@ syncprov_matchops( Operation *op, opcookie *opc, > int saveit ) > } > > if ( fc.fscope ) { > + Filter *paranoia; > ldap_pvt_thread_mutex_lock( &ss->s_mutex ); > op2 = *ss->s_op; > oh = *op->o_hdr; > @@ -1304,6 +1305,7 @@ syncprov_matchops( Operation *op, opcookie *opc, > int saveit ) > op2.o_hdr = &oh; > op2.o_extra = op->o_extra; > op2.o_callback = NULL; > + paranoia = ss->s_op->ors_filter; > if (ss->s_flags & PS_FIX_FILTER) { > /* Skip the AND/GE clause that we stuck on in front. We > would lose deletes/mods that happen during the refresh > @@ -1311,7 +1313,11 @@ syncprov_matchops( Operation *op, opcookie *opc, > int saveit ) > op2.ors_filter = ss->s_op->ors_filter->f_and->f_next; > } > ldap_pvt_thread_mutex_unlock( &ss->s_mutex ); > + assert (paranoia == ss->s_op->ors_filter || paranoia == > ss->s_op->ors_filter->f_and->f_next); > + usleep(1000); > + assert (paranoia == ss->s_op->ors_filter || paranoia == > ss->s_op->ors_filter->f_and->f_next); > rc = test_filter( &op2, e, op2.ors_filter ); > + assert (paranoia == ss->s_op->ors_filter || paranoia == > ss->s_op->ors_filter->f_and->f_next); > } > > Debug( LDAP_DEBUG_TRACE, "syncprov_matchops: sid %03x fscope > %d rc %d\n", > > Updated patch attached. > Please review and merge. > > Leonid. > > --- > > The attached files is derived from OpenLDAP Software. All of the > modifications > to OpenLDAP Software represented in the following patch(es) were > developed by > Peter-Service LLC, Moscow, Russia. Peter-Service LLC has not assigned > rights > and/or interest in this work to any party. I, Leonid Yuriev am > authorized by > Peter-Service LLC, my employer, to release this work under the following > terms. > > Peter-Service LLC hereby places the following modifications to OpenLDAP > Software > (and only these modifications) into the public domain. Hence, these > modifications may be freely used and/or redistributed for any purpose > with or without attribution and/or other notice. > -- -- Howard Chu CTO, Symas Corp. http://www.symas.com Director, Highland Sun http://highlandsun.com/hyc/ Chief Architect, OpenLDAP http://www.openldap.org/project/

1 0

Re: (ITS#8014) minor memleak in slapadd
by hyc＠symas.com 03 Jan '15

03 Jan '15

leo(a)yuriev.ru wrote: > Full_Name: Leonid Yuriev > Version: 2.4-HEAD > OS: RHEL7 > URL: ftp://ftp.openldap.org/incoming/ > Submission from: (NULL) (31.130.36.33) > > > Minor memleak was found by valgrind tool. > Patch will be available shortly. Thanks for the report. Fixed in master. > > Leonid. > > -- -- Howard Chu CTO, Symas Corp. http://www.symas.com Director, Highland Sun http://highlandsun.com/hyc/ Chief Architect, OpenLDAP http://www.openldap.org/project/

1 0

Re: (ITS#8013) SIGSEGV in test_filter()
by leo＠yuriev.ru 03 Jan '15

03 Jan '15

This is a multi-part message in MIME format. --------------070009030106040804000907 Content-Type: text/plain; charset=windows-1251; format=flowed Content-Transfer-Encoding: 7bit 04.01.2015 02:47, Howard Chu wrote: > The patch is wrong. syncprov_matchops can be called twice for the same > operation; toggling the flag where you've placed it means the filter > will be incorrect on the 2nd invocation as well as on all subsequent > invocations. Ok, but found another case. The ss->s_op->ors_filter may be updated immediately after the s_mutex released. This is a simple change for reproducing the bug. diff --git a/servers/slapd/overlays/syncprov.c b/servers/slapd/overlays/syncprov.c index 5ef2eaa..f579faa 100644 --- a/servers/slapd/overlays/syncprov.c +++ b/servers/slapd/overlays/syncprov.c @@ -1295,6 +1295,7 @@ syncprov_matchops( Operation *op, opcookie *opc, int saveit ) } if ( fc.fscope ) { + Filter *paranoia; ldap_pvt_thread_mutex_lock( &ss->s_mutex ); op2 = *ss->s_op; oh = *op->o_hdr; @@ -1304,6 +1305,7 @@ syncprov_matchops( Operation *op, opcookie *opc, int saveit ) op2.o_hdr = &oh; op2.o_extra = op->o_extra; op2.o_callback = NULL; + paranoia = ss->s_op->ors_filter; if (ss->s_flags & PS_FIX_FILTER) { /* Skip the AND/GE clause that we stuck on in front. We would lose deletes/mods that happen during the refresh @@ -1311,7 +1313,11 @@ syncprov_matchops( Operation *op, opcookie *opc, int saveit ) op2.ors_filter = ss->s_op->ors_filter->f_and->f_next; } ldap_pvt_thread_mutex_unlock( &ss->s_mutex ); + assert (paranoia == ss->s_op->ors_filter || paranoia == ss->s_op->ors_filter->f_and->f_next); + usleep(1000); + assert (paranoia == ss->s_op->ors_filter || paranoia == ss->s_op->ors_filter->f_and->f_next); rc = test_filter( &op2, e, op2.ors_filter ); + assert (paranoia == ss->s_op->ors_filter || paranoia == ss->s_op->ors_filter->f_and->f_next); } Debug( LDAP_DEBUG_TRACE, "syncprov_matchops: sid %03x fscope %d rc %d\n", Updated patch attached. Please review and merge. Leonid. --- The attached files is derived from OpenLDAP Software. All of the modifications to OpenLDAP Software represented in the following patch(es) were developed by Peter-Service LLC, Moscow, Russia. Peter-Service LLC has not assigned rights and/or interest in this work to any party. I, Leonid Yuriev am authorized by Peter-Service LLC, my employer, to release this work under the following terms. Peter-Service LLC hereby places the following modifications to OpenLDAP Software (and only these modifications) into the public domain. Hence, these modifications may be freely used and/or redistributed for any purpose with or without attribution and/or other notice. --------------070009030106040804000907 Content-Type: text/x-patch; name="its8013.patch" Content-Transfer-Encoding: 7bit Content-Disposition: attachment; filename="its8013.patch" commit 5014e0b526e68c21bcfd9a9dd43d154d621f7476 Author: Leo Yuriev <leo(a)yuriev.ru> Date: 2015-01-01 16:44:50 +0300 ITS#8013 fix rare SIGSEGV in test_filter(). Filter may be updated/fixed and freed in other thread immediately after release the s_mutex, while the test_filter() running. Also now specifically damaged filter fields when it freed. This will enable faster detection of such errors (fail-fast by SIGSEGV). diff --git a/servers/slapd/filter.c b/servers/slapd/filter.c index b859f73..f440336 100644 --- a/servers/slapd/filter.c +++ b/servers/slapd/filter.c @@ -566,6 +566,8 @@ filter_free_x( Operation *op, Filter *f, int freeme ) break; } + f->f_next = (void*) -1l; + f->f_un.f_un_complex = (void*) -1l; if ( freeme ) { op->o_tmpfree( f, op->o_tmpmemctx ); } diff --git a/servers/slapd/overlays/syncprov.c b/servers/slapd/overlays/syncprov.c index 5ef2eaa..2369edb 100644 --- a/servers/slapd/overlays/syncprov.c +++ b/servers/slapd/overlays/syncprov.c @@ -1304,14 +1304,17 @@ syncprov_matchops( Operation *op, opcookie *opc, int saveit ) op2.o_hdr = &oh; op2.o_extra = op->o_extra; op2.o_callback = NULL; - if (ss->s_flags & PS_FIX_FILTER) { + if ( ss->s_flags & PS_FIX_FILTER ) { /* Skip the AND/GE clause that we stuck on in front. We would lose deletes/mods that happen during the refresh phase otherwise (ITS#6555) */ - op2.ors_filter = ss->s_op->ors_filter->f_and->f_next; + op2.ors_filter = filter_dup( op2.ors_filter->f_and->f_next, NULL ); + } else { + op2.ors_filter = filter_dup( op2.ors_filter, NULL ); } ldap_pvt_thread_mutex_unlock( &ss->s_mutex ); rc = test_filter( &op2, e, op2.ors_filter ); + filter_free( op2.ors_filter ); } Debug( LDAP_DEBUG_TRACE, "syncprov_matchops: sid %03x fscope %d rc %d\n", @@ -2237,6 +2240,7 @@ syncprov_detach_op( Operation *op, syncops *so, slap_overinst *on ) /* Skip the AND/GE clause that we stuck on in front */ if ( so->s_flags & PS_FIX_FILTER ) { + assert(op2->ors_filter->f_choice == LDAP_FILTER_AND); op2->ors_filter = op->ors_filter->f_and->f_next; so->s_flags ^= PS_FIX_FILTER; } else { @@ -2390,7 +2394,6 @@ syncprov_search_response( Operation *op, SlapReply *rs ) ldap_pvt_thread_mutex_unlock( &op->o_conn->c_mutex ); /* syncprov_ab_cleanup will free this syncop */ return SLAPD_ABANDON; - } else { ldap_pvt_thread_mutex_lock( &ss->ss_so->s_mutex ); /* Turn off the refreshing flag */ --------------070009030106040804000907--

1 0

Re: (ITS#8013) SIGSEGV in test_filter()
by hyc＠symas.com 03 Jan '15

03 Jan '15

Leonid Yuriev wrote: > Please review attached patch and merge. The patch is wrong. syncprov_matchops can be called twice for the same operation; toggling the flag where you've placed it means the filter will be incorrect on the 2nd invocation as well as on all subsequent invocations. > > Leonid. > > --- > > The attached files is derived from OpenLDAP Software. All of the > modifications > to OpenLDAP Software represented in the following patch(es) were > developed by > Peter-Service LLC, Moscow, Russia. Peter-Service LLC has not assigned > rights > and/or interest in this work to any party. I, Leonid Yuriev am > authorized by > Peter-Service LLC, my employer, to release this work under the following > terms. > > Peter-Service LLC hereby places the following modifications to OpenLDAP > Software > (and only these modifications) into the public domain. Hence, these > modifications may be freely used and/or redistributed for any purpose > with or without attribution and/or other notice. -- -- Howard Chu CTO, Symas Corp. http://www.symas.com Director, Highland Sun http://highlandsun.com/hyc/ Chief Architect, OpenLDAP http://www.openldap.org/project/

1 0

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

openldap-bugs