openldap-bugs July 2009

openldap-bugs@openldap.org

49 participants
166 discussions

(ITS#6200) slapd crashes under load w/ syncrepl
by imf＠u.washington.edu 08 Jul '09

08 Jul '09

Full_Name: Tracy Stenvik Version: 2.4.16 (+db-4.7.25pl4) OS: Red Hat Enterprise Linux Server release 5.3 (Tikanga) URL: ftp://ftp.openldap.org/incoming/ Submission from: (NULL) (140.142.111.41) In evaluating 2.4.16 (to upgrade our 2.3.39-based directory service) we have encountered some serious issues. A similated production-level load that 2.3.39 (with slurpd running) handles easily causes 2.4.16 (and 2.4.15) to crash within 5 minutes to 24 hours. Without replication active, 2.4.16 survived 24 hours under load. As soon as replication was activated however, the server crashed within 5 minutes. We've tried several syncrepl variants (as described below) but none appear stable. In analyzing the various crash dumps, most of the trouble appears to lie in the caching code. It has a nasty habit of returing nulls when nulls are not expected. Our simulated load consists of an arbitrary mix of adds/modifies/searches to various objects in the directory. This load is defined as "N" instances of the "Stressor" test, as discussed on the "GDS Stress Test Findings" page at, http://wiki.cac.washington.edu/display/infra/GDS+Stress+Test+Findings Following, you will find a series of crash dumps, along with a few comments. Also included are sections the various slapd.confs we were testing with. Please contact me if you require additional information. delta syncrepl crash -------------------- (gdb) where #0 0xb7b7abab in strncmp () from /lib/libc.so.6 #1 0x08143ea4 in bdb_rdn_cmp (v_e1=0xf069be8, v_e2=0xf73cb48) at cache.c:301 #2 0x081d00ee in avl_insert (root=0x8622088, data=0xf069be8, fcmp=0x8143e55 <bdb_rdn_cmp>, fdup=0x81d10a0 <avl_dup_error>) at avl.c:95 #3 0x08144692 in hdb_cache_find_parent (op=0xf95ea78, txn=0xe7def98, id=343637, res=0x14b3ade0) at cache.c:593 #4 0x08144dc9 in hdb_cache_find_id (op=0xf95ea78, tid=0xe7def98, id=343637, eip=0x14b3ade0, flag=0, lock=0x14b3adc4) at cache.c:905 #5 0x0810b6a0 in hdb_search (op=0xf95ea78, rs=0x14bfc0e8) at search.c:705 #6 0x080e9b02 in overlay_op_walk (op=0xf95ea78, rs=0x14bfc0e8, which=op_search, oi=0x85aeed8, on=0x0) at backover.c:669 #7 0x080e9cb7 in over_op_func (op=0xf95ea78, rs=0x14bfc0e8, which=op_search) at backover.c:721 #8 0x080e9d3b in over_op_search (op=0xf95ea78, rs=0x14bfc0e8) at backover.c:743 #9 0x0806f0b3 in fe_op_search (op=0xf95ea78, rs=0x14bfc0e8) at search.c:366 #10 0x0806ea53 in do_search (op=0xf95ea78, rs=0x14bfc0e8) at search.c:217 #11 0x0806b959 in connection_operation (ctx=0x14bfc1d0, arg_v=0xf95ea78) at connection.c:1097 #12 0x0806be33 in connection_read_thread (ctx=0x14bfc1d0, argv=0x23) at connection.c:1223 #13 0x081d74a1 in ldap_int_thread_pool_wrapper (xpool=0x8561178) at tpool.c:663 #14 0xb7e4d49b in start_thread () from /lib/libpthread.so.0 #15 0xb7bdc42e in clone () from /lib/libc.so.6 (gdb) p *(EntryInfo *)v_e1 $2 = {bei_parent = 0x8622058, bei_id = 343637, bei_lockpad = 0, bei_state = 132, bei_finders = 0, bei_nrdn = {bv_len = 45, bv_val = 0xf7af1f8 "serialNumber=40442595174b5ebf53167252b5cbcc34"}, bei_rdn = {bv_len = 45, bv_val = 0xf36d700 "serialNumber=40442595174B5EBF53167252B5CBCC34"}, bei_modrdns = 0, bei_ckids = 0, bei_dkids = 0, bei_e = 0x0, bei_kids = 0x0, bei_kids_mutex = {__data = {__lock = 1, __count = 0, __owner = 32550, __kind = 0, __nusers = 1, {__spins = 0, __list = {__next = 0x0}}}, __size = "\001\000\000\000\000\000\000\000&\177\000\000\000\000\000\000\001\000\000\000\000\000\000", __align = 1}, bei_lrunext = 0x8978ad0, bei_lruprev = 0x0} (gdb) p *(EntryInfo *)v_e2 $3 = {bei_parent = 0x0, bei_id = 410257, bei_lockpad = 0, bei_state = 0, bei_finders = 0, bei_nrdn = {bv_len = 45, bv_val = 0x0}, bei_rdn = { bv_len = 45, bv_val = 0x0}, bei_modrdns = 0, bei_ckids = 0, bei_dkids = 0, bei_e = 0x0, bei_kids = 0x0, bei_kids_mutex = {__data = {__lock = 0, __count = 0, __owner = 0, __kind = 0, __nusers = 0, {__spins = 0, __list = {__next = 0x0}}}, __size = '\0' <repeats 23 times>, __align = 0}, bei_lrunext = 0x8978ad0, bei_lruprev = 0x0} In bdb_rdn_cmp() you're comparing e1->bei_nrdn.bv_val to e2->bei_nrdn.bv_val but the latter is null even though e2->bei_nrdn.bv_len is not zero. standard syncrepl crash ----------------------- (gdb) where #0 0xb7d9f300 in pthread_mutex_lock () from /lib/libpthread.so.0 #1 0x081d8645 in ldap_pvt_thread_mutex_lock (mutex=0x34) at thr_posix.c:296 #2 0x08145bc0 in bdb_cache_delete_internal (cache=0xa25cc04, e=0x147e6d38, decr=0) at cache.c:1372 #3 0x08144b08 in hdb_cache_lru_purge (bdb=0xa25cbc8) at cache.c:776 #4 0x081452d1 in hdb_cache_find_id (op=0x11923ca8, tid=0x1316aa48, id=280869, eip=0x2bffcbfc, flag=0, lock=0x2bffcc54) at cache.c:1053 #5 0x0814b2f7 in hdb_dn2entry (op=0x11923ca8, tid=0x1316aa48, dn=0xa25e3bc, e=0x2bffcc64, matched=0, lock=0x2bffcc54) at dn2entry.c:67 #6 0x08151845 in hdb_entry_get (op=0x11923ca8, ndn=0xa25e3bc, oc=0xa2549f0, at=0xa23ebb0, rw=0, ent=0x2bffceac) at id2entry.c:354 #7 0x080e936d in overlay_entry_get_ov (op=0x11923ca8, dn=0xa25e3bc, oc=0xa2549f0, ad=0xa23ebb0, rw=0, e=0x2bffceac, on=0x0) at backover.c:364 #8 0x080e941f in over_entry_get_rw (op=0x11923ca8, dn=0xa25e3bc, oc=0xa2549f0, ad=0xa23ebb0, rw=0, e=0x2bffceac) at backover.c:396 #9 0x0807cfc3 in be_entry_get_rw (op=0x11923ca8, ndn=0xa25e3bc, oc=0xa2549f0, at=0xa23ebb0, rw=0, e=0x2bffceac) at backend.c:1346 #10 0x0807d1a2 in fe_acl_group (op=0x11923ca8, target=0x0, gr_ndn=0xa25e3bc, op_ndn=0x11923d44, group_oc=0xa2549f0, group_at=0xa23ebb0) at backend.c:1402 #11 0x0807d8c3 in backend_group (op=0x11923ca8, target=0x0, gr_ndn=0xa25e3bc, op_ndn=0x11923d44, group_oc=0xa2549f0, group_at=0xa23ebb0) at backend.c:1597 #12 0x080d32e2 in limits_get (op=0x11923ca8, limit=0x11923cdc) at limits.c:117 #13 0x080d588d in limits_check (op=0x11923ca8, rs=0x2bffe0e8) at limits.c:1106 #14 0x0806f095 in fe_op_search (op=0x11923ca8, rs=0x2bffe0e8) at search.c:364 #15 0x0806ea53 in do_search (op=0x11923ca8, rs=0x2bffe0e8) at search.c:217 #16 0x0806b959 in connection_operation (ctx=0x2bffe1d0, arg_v=0x11923ca8) at connection.c:1097 #17 0x0806be33 in connection_read_thread (ctx=0x2bffe1d0, argv=0x1d) at connection.c:1223 #18 0x081d74a1 in ldap_int_thread_pool_wrapper (xpool=0xa211178) at tpool.c:663 #19 0xb7d9d49b in start_thread () from /lib/libpthread.so.0 #20 0xb7b2c42e in clone () from /lib/libc.so.6 (gdb) p *e->bei_parent Cannot access memory at address 0x0 In bdb_cache_delete_internal(), e->bei_parent is null. 2-way multimaster syncrepl crash -------------------------------- (gdb) where #0 0x08149c1a in hdb_delete (op=0xc01caa8, rs=0x2ecfc0e8) at delete.c:217 #1 0x080e9b02 in overlay_op_walk (op=0xc01caa8, rs=0x2ecfc0e8, which=op_delete, oi=0x88b4550, on=0x0) at backover.c:669 #2 0x080e9cb7 in over_op_func (op=0xc01caa8, rs=0x2ecfc0e8, which=op_delete) at backover.c:721 #3 0x080e9de5 in over_op_delete (op=0xc01caa8, rs=0x2ecfc0e8) at backover.c:773 #4 0x0808ae88 in fe_op_delete (op=0xc01caa8, rs=0x2ecfc0e8) at delete.c:174 #5 0x0808abdb in do_delete (op=0xc01caa8, rs=0x2ecfc0e8) at delete.c:95 #6 0x0806b959 in connection_operation (ctx=0x2ecfc1d0, arg_v=0xc01caa8) at connection.c:1097 #7 0x0806be33 in connection_read_thread (ctx=0x2ecfc1d0, argv=0x22) at connection.c:1223 #8 0x081d74a1 in ldap_int_thread_pool_wrapper (xpool=0x8867178) at tpool.c:663 #9 0xb7e4749b in start_thread () from /lib/libpthread.so.0 #10 0xb7bd642e in clone () from /lib/libc.so.6 (gdb) p eip $1 = (EntryInfo *) 0x0 In hdb_delete(), eip is null. 2-way multimaster syncrepl crash -------------------------------- (gdb) where #0 0x080956db in slap_acl_get (a=0x8b9ba80, count=0x37f5d824, op=0x8c5d948, e=0x37f5dad0, desc=0x8b9b9f0, val=0x0, matches=0x37f5d1c4, state=0x37f5d97c) at acl.c:599 #1 0x080949b9 in slap_access_allowed (op=0x8c5d948, e=0x37f5dad0, desc=0x8b9b9f0, val=0x0, access=ACL_WDEL, state=0x37f5d97c, maskp=0x37f5d8e8) at acl.c:236 #2 0x08094cd5 in fe_access_allowed (op=0x8c5d948, e=0x37f5dad0, desc=0x8b9b9f0, val=0x0, access=ACL_WDEL, state=0x37f5d97c, maskp=0x37f5d8e8) at acl.c:360 #3 0x0809503c in access_allowed_mask (op=0x8c5d948, e=0x37f5dad0, desc=0x8b9b9f0, val=0x0, access=ACL_WDEL, state=0x37f5d97c, maskp=0x0) at acl.c:462 #4 0x08098b8f in acl_check_modlist (op=0x8c5d948, e=0x37f5dad0, mlist=0x34afa7c8) at acl.c:1986 #5 0x0810373a in hdb_modify_internal (op=0x8c5d948, tid=0x961aa10, modlist=0x34afa7c8, e=0x37f5dad0, text=0x37f5f0fc, textbuf=0x37f5db18 "", textlen=256) at modify.c:88 #6 0x08105292 in hdb_modify (op=0x8c5d948, rs=0x37f5f0e8) at modify.c:590 #7 0x080e9b02 in overlay_op_walk (op=0x8c5d948, rs=0x37f5f0e8, which=op_modify, oi=0x8bbb550, on=0x0) at backover.c:669 #8 0x080e9cb7 in over_op_func (op=0x8c5d948, rs=0x37f5f0e8, which=op_modify) at backover.c:721 #9 0x080e9d7f in over_op_modify (op=0x8c5d948, rs=0x37f5f0e8) at backover.c:755 #10 0x08088a89 in fe_op_modify (op=0x8c5d948, rs=0x37f5f0e8) at modify.c:301 #11 0x080884c8 in do_modify (op=0x8c5d948, rs=0x37f5f0e8) at modify.c:175 #12 0x0806b959 in connection_operation (ctx=0x37f5f1d0, arg_v=0x8c5d948) at connection.c:1097 #13 0x0806be33 in connection_read_thread (ctx=0x37f5f1d0, argv=0x22) at connection.c:1223 #14 0x081d74a1 in ldap_int_thread_pool_wrapper (xpool=0x8b6e178) at tpool.c:663 #15 0xb7e6749b in start_thread () from /lib/libpthread.so.0 #16 0xb7bf642e in clone () from /lib/libc.so.6 (gdb) p e->e_nname $3 = {bv_len = 77, bv_val = 0x0} In slap_acl_get(), e->e_nname.bv_val (e->e_ndn) is null. delta syncrepl slapd.conf ------------------------- serverID 1 ldap://rufus01.u.washington.edu:290 serverID 2 ldap://rufus04.u.washington.edu:290 gentlehup on idletimeout 600 sizelimit 20000 sortvals member # ===== Replication db ==== database hdb suffix "cn=log" directory /data/au20aa_log rootdn "cn=Manager,dc=washington,dc=edu" #index objectClass eq #index entryUUID,entryCSN eq index default eq index entryCSN,objectClass,reqEnd,reqResult,reqStart overlay syncprov syncprov-reloadhint true syncprov-nopresent true # ===== Primary db === database hdb suffix "dc=washington,dc=edu" rootdn "cn=Manager,dc=washington,dc=edu" include /usr/local/openldap-2.4.16/etc/openldap/private/slapd-pw.conf directory /data/au20aa index entryCSN eq index entryUUID eq checkpoint 10000 15 cachesize 90000 idlcachesize 180000 overlay syncprov #syncprov-sessionlog 100 syncprov-checkpoint 1000 60 # accesslog overlay definitions for primary database overlay accesslog logdb cn=log logops writes logsuccess true # scan the accesslog DB every day, and purge entries older than 7 days logpurge 07+00:00 01+00:00 standard syncrepl slapd.conf ---------------------------- serverID 1 ldap://rufus01.u.washington.edu:290 serverID 2 ldap://rufus04.u.washington.edu:290 gentlehup on idletimeout 600 sizelimit 20000 sortvals member # ===== Primary db ===== database hdb suffix "dc=washington,dc=edu" rootdn "cn=Manager,dc=washington,dc=edu" include /usr/local/openldap-2.4.16/etc/openldap/private/slapd-pw.conf directory /data/au20aa index entryCSN eq index entryUUID eq checkpoint 10000 15 cachesize 90000 idlcachesize 180000 overlay syncprov #syncprov-sessionlog 100 syncprov-checkpoint 1000 60 2-way multimaster syncrepl slapd.conf ------------------------------------- serverID 1 ldap://rufus01.u.washington.edu:290 serverID 2 ldap://rufus04.u.washington.edu:290 gentlehup on idletimeout 600 sizelimit 20000 sortvals member # ===== Primary db ===== database hdb suffix "dc=washington,dc=edu" rootdn "cn=Manager,dc=washington,dc=edu" include /usr/local/openldap-2.4.16/etc/openldap/private/slapd-pw.conf directory /data/au20aa index entryCSN eq index entryUUID eq checkpoint 10000 15 cachesize 90000 idlcachesize 180000 overlay syncprov #syncprov-sessionlog 100 syncprov-checkpoint 1000 60 ### ### ----------- Replication ### syncrepl rid=101 provider=ldap://rufus01.u.washington.edu:290 searchbase="dc=washington,dc=edu" bindmethod=sasl saslmech=external starttls=yes tls_cert=/usr/local/openldap-2.4.16/etc/openldap/rufus04.crt tls_key=/usr/local/openldap-2.4.16/etc/openldap/private/rufus04.key filter="(objectClass=*)" attrs="*,+" schemachecking=off scope=sub type=refreshAndPersist retry="60 +" syncrepl rid=102 provider=ldap://rufus04.u.washington.edu:290 searchbase="dc=washington,dc=edu" bindmethod=sasl saslmech=external starttls=yes tls_cert=/usr/local/openldap-2.4.16/etc/openldap/rufus04.crt tls_key=/usr/local/openldap-2.4.16/etc/openldap/private/rufus04.key filter="(objectClass=*)" attrs="*,+" schemachecking=off scope=sub type=refreshAndPersist retry="60 +" mirrormode yes

1 0

(ITS#6199) [JLDAP] SASL integrity & confidentiality
by ioplex＠gmail.com 07 Jul '09

07 Jul '09

Full_Name: Michael B Allen Version: JLDAP OS: Linux URL: http://www.ioplex.com/~miallen/openldap-jldap-saslwrap.diff Submission from: (NULL) (71.187.189.103) It would be fabulous if SASL integrity and confidentiality worked in JLDAP. I just tried and failed to add support for it. I could be just misguided but it seems that because the ASN.1 routines encoded and decoded BER directly to and from the socket element by element, there is a fairly low-level issue with wrapping and unwrapping entire message buffers. But I was unable to pinpoint the precise cause. See the URL for my crude patch to test SASL wrap / unwrap. Of course you'll need a SASL client and mechanism that supports at least integrity. Since com.novell.ldap.Connection.writeMessage encodes the entire message before writing it to the socket, the SASL wrap works. However, reading data from the input stream has a strange off-by-one problem. The first byte is missing. But since the first 4 bytes are the big endian SASL buffer size, I can hack to code to skip it and it is unwrapped properly (although it looks like a referral so I think perhaps my search() test program isn't quite right either). It then hangs in LDAPSearchResults.hasMore: at java.lang.Object.wait(Object.java:474) at com.novell.ldap.MessageAgent.getLDAPMessage(MessageAgent.java:399) - locked <0x89152240> (a com.novell.ldap.MessageVector) at com.novell.ldap.LDAPMessageQueue.getResponse(LDAPMessageQueue.java:175) at com.novell.ldap.LDAPMessageQueue.getResponse(LDAPMessageQueue.java:126) at com.novell.ldap.LDAPSearchResults.getBatchOfResults(LDAPSearchResults.java:338) at com.novell.ldap.LDAPSearchResults.resetVectors(LDAPSearchResults.java:213) at com.novell.ldap.LDAPSearchResults.hasMore(LDAPSearchResults.java:175) at org.ietf.ldap.LDAPSearchResults.hasMore(LDAPSearchResults.java:89) Unfortunately this looks like a fairly low-level architectual issue so I'll have to fall-back to JNDI. But hopefully this will help someone who wants to add this critical functionality in the future. Using the builtin integrity and confidentiality offered by mechanisms like GSSAPI is pretty important IMO. At least it's quite a bit more graceful than using TLS.

1 0

Re: (ITS#6195) Replications with 2.3.43 master to 2.4.16 slave is failing
by quanah＠zimbra.com 07 Jul '09

07 Jul '09

--On Thursday, July 02, 2009 10:49 PM +0000 quanah(a)zimbra.com wrote: > --On Thursday, July 02, 2009 3:40 PM -0700 Quanah Gibson-Mount > <quanah(a)zimbra.com> wrote: Fixed with the CSN normalization commit. --Quanah -- Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc -------------------- Zimbra :: the leader in open source messaging and collaboration

1 0

Re: (ITS#6194) Patch - Enhancement - provide LDIF support as libldif
by rmeggins＠redhat.com 07 Jul '09

07 Jul '09

Howard Chu wrote: > rmeggins(a)redhat.com wrote: >> Full_Name: Rich Megginson >> Version: CVS HEAD (2.4.16+) >> OS: Fedora >> URL: >> ftp://ftp.openldap.org/incoming/openldap-2.4.16-libldif-20090702.patch >> Submission from: (NULL) (76.113.59.19) >> >> >> This patch allows liblutil to provide the LDIF reading and writing >> functions as >> public APIs in a libldif, and exposes ldif.h to the public API. > > See also ITS#4033; ldif.h only parses LDIF into chunks/records, it's > still up to the caller to turn these records into something > meaningful. In fact libldif used to be a library of its own, but in > RE22 we made the decision to merge it into liblutil because its > functionality is so limited. I don't expect we'll reverse this > decision unless/until it gets enough parsing added to it to make it > more generally useful. I think it's very useful as it is - the mozldap libldif (which is almost identical in functionality to my proposed openldap libldif) is used in several different applications in the 389 directory server suite. What more does it need? Some sort of higher level "entry" and "attribute" abstraction? > >> This patch file is derived from OpenLDAP Software. All of the >> modifications to OpenLDAP Software represented in the following >> patch(es) were developed by Red Hat, Inc.. Red Hat, Inc. has not >> assigned rights and/or interest in this work to any party. I, Richard >> Megginson am authorized by Red Hat, Inc., my employer, to release this >> work under the following terms. >> >> Red Hat, Inc. hereby place 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. >

1 0

Re: (ITS#6194) Patch - Enhancement - provide LDIF support as libldif
by hyc＠symas.com 07 Jul '09

07 Jul '09

rmeggins(a)redhat.com wrote: > Full_Name: Rich Megginson > Version: CVS HEAD (2.4.16+) > OS: Fedora > URL: ftp://ftp.openldap.org/incoming/openldap-2.4.16-libldif-20090702.patch > Submission from: (NULL) (76.113.59.19) > > > This patch allows liblutil to provide the LDIF reading and writing functions as > public APIs in a libldif, and exposes ldif.h to the public API. See also ITS#4033; ldif.h only parses LDIF into chunks/records, it's still up to the caller to turn these records into something meaningful. In fact libldif used to be a library of its own, but in RE22 we made the decision to merge it into liblutil because its functionality is so limited. I don't expect we'll reverse this decision unless/until it gets enough parsing added to it to make it more generally useful. > This patch file is derived from OpenLDAP Software. All of the > modifications to OpenLDAP Software represented in the following > patch(es) were developed by Red Hat, Inc.. Red Hat, Inc. has not > assigned rights and/or interest in this work to any party. I, Richard > Megginson am authorized by Red Hat, Inc., my employer, to release this > work under the following terms. > > Red Hat, Inc. hereby place 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#5340) REP_ENTRY_MODIFIABLE bug in dynlist
by h.b.furuseth＠usit.uio.no 07 Jul '09

07 Jul '09

I've been trying to factor this MUSTRELEASE/MUSTBEFREED stuff out to some macros. Easier said than done, I ended up with the macro monstrosity below. Do I put that in slap.h? Or a separate slap-response.h file? Generally the idea is that most code which modifies rs->sr_<entry/matched/ref/ctrls> (as opposed to initializing) should use these macros, which obey sr_flag and dispose of the old value if needed. There won't be that many callers actually, mostly overlays. There is overlay code which need to be rearranged a bit or which can't use this, and there is overlay code I don't understand at all and don't dare touch. Oh, and this seems to be wrong for how sr_ctrls are set up, which seems to be allocated differently some places. I'm just including that for reference for now, won't commit it yet in any case. Finally, code which aborts normal processing of some response and instead send an error often needs to be fixed - it often doesn't flush SlapReply data, and the data to be flushed and maybe freeed can be lost because it's overwritten. Generally I think code which modifies sr_type (or in some cases calls a function which will do so) should be responsible for flushing old SlapReply data first. #define REP_ENTRY_MUSTFLUSH (REP_ENTRY_MUSTBEFREED|REP_ENTRY_MUSTRELEASE) #define REP_ENTRY_MASK (REP_ENTRY_MODIFIABLE |REP_ENTRY_MUSTFLUSH) #define REP_MATCHED_MUSTFLUSH REP_MATCHED_MUSTBEFREED #define REP_REF_MUSTFLUSH REP_REF_MUSTBEFREED #define REP_CTRLS_MUSTFLUSH REP_CTRLS_MUSTBEFREED /* Macros updating freeable members of SlapReply, mostly for overlays. */ /* * Ensure rs->sr_entry is modifiable, by duplicating it if necessary. * Obey sr_flags. Set REP_ENTRY_<MODIFIABLE, and MUSTBEFREED if duplicated>. * Do <postmodActions> if the entry was replaced. */ #define RS_MODIFIABLE_ENTRY(op, rs, on, postmodActions) do { \ SlapReply *const rmeRS = (rs); \ if ( rmeRS->sr_flags & REP_ENTRY_MODIFIABLE ) break; \ RS_REPLACE_ENTRY( op, rmeRS, on, entry_dup( rmeRS->sr_entry ), 0 ); \ rmeRS->sr_flags |= REP_ENTRY_MODIFIABLE | REP_ENTRY_MUSTBEFREED; \ postmodActions; \ } while (0) /* Reset rs->sr_<member> if REP_<member>_MUSTBEFREED/MUSTRELEASE says so. */ #define RS_FLUSH_ENTRY(op, rs, on) RS_FLUSH(ENTRY, op, rs, on, 1, {}, {}) #define RS_FLUSH_CTRLS(op, rs) RS_FLUSH(CTRLS, op, rs, -, 1, {}, {}) #define RS_FLUSH_MATCHED( rs) RS_FLUSH(MATCHED, -, rs, -, 1, {}, {}) #define RS_FLUSH_REF( rs) RS_FLUSH(REF, -, rs, -, 1, {}, {}) /* Flush and set rs->sr_<member>. Set REP_<member>_MUSTBEFREED if <isDup>. */ #define RS_REPLACE_ENTRY(op,rs,on,v,isDup) RS_REPLACE(ENTRY,op,rs,on, v,isDup) #define RS_REPLACE_CTRLS(op,rs,v,isDup) RS_REPLACE(CTRLS, op,rs,-, v,isDup) #define RS_REPLACE_MATCHED( rs,v,isDup) RS_REPLACE(MATCHED, -,rs,-, v,isDup) #define RS_REPLACE_REF( rs,v,isDup) RS_REPLACE(REF, -,rs,-, v,isDup) /* Helpers for the above */ #define SR_ENTRY sr_entry #define SR_CTRLS sr_ctrls #define SR_MATCHED sr_matched #define SR_REF sr_ref typedef Entry *SR_ENTRY_t; typedef LDAPControl **SR_CTRLS_t; typedef const char *SR_MATCHED_t; typedef BerVarray SR_REF_t; #define RS_FREE_ENTRY(op, rs, on)\ if ((rs)->sr_flags & REP_ENTRY_MUSTBEFREED) { \ assert (!(rs)->sr_flags & REP_ENTRY_MUSTRELEASE); \ entry_free((rs)->sr_entry); \ } \ else if (on == NULL) be_entry_release_rw(op, (rs)->sr_entry, 0);\ else overlay_entry_release_ov(op, (rs)->sr_entry, 0, on) #define RS_FREE_CTRLS(op, rs, on) slap_free_ctrls( op, (rs)->sr_ctrls ) #define RS_FREE_MATCHED(op, rs, on) ch_free( (char *) (rs)->sr_matched ) #define RS_FREE_REF(op, rs, on) ber_bvarray_free( (rs)->sr_ref ) #define RS_FLUSH(MEMBER, op, rs, on, reset, preActions, postActions) do { \ SlapReply *const rfRS = (rs); \ preActions; \ if ( rfRS->sr_flags & REP_##MEMBER##_MUSTFLUSH ) { \ if ( rfRS->SR_##MEMBER ) { \ RS_FREE_##MEMBER( op, rfRS, on ); \ if ( reset ) rfRS->SR_##MEMBER = NULL; \ } \ } \ rfRS->sr_flags &= ~REP_##MEMBER##_MASK; \ postActions; \ } while (0) #define RS_REPLACE(MEMBER, op, rs, on, val, isDup) \ RS_FLUSH( MEMBER, op, rs, on, 0, \ SR_##MEMBER##_t const rfNewval = (val), \ if ( isDup ) rfRS->sr_flags |= REP_##MEMBER##_MUSTBEFREED; \ rfRS->SR_##MEMBER = rfNewval ) /* Warn if something re#defines the following, which would break RS_REPLACE */ #define ENTRY ENTRY #define CTRLS CTRLS #define MATCHED MATCHED #define REF REF

1 0

Re: (ITS#5892) slapadd doesn't fail on invalid include statements
by hyc＠symas.com 07 Jul '09

07 Jul '09

aredridel(a)nbtsc.org wrote: > Full_Name: Aria Stewart > Version: 2.4.13 > OS: PLD Linux (Th) > URL: ftp://ftp.openldap.org/incoming/ > Submission from: (NULL) (2001:470:8935:7:21b:fcff:fe77:1757) > > > An ldif file with include statements in the form of: > > include: /filename > > rather than > > include: file:///filename > > silently fail. Slapadd stops at that point, but does not print an error message. Oops, the library has provided an error message for this instance since 2.4.5beta (September 2007) but slapadd wasn't enabling the display of that message by default. This is now fixed in HEAD. -- -- 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#6101) Uninitialized SlapReply components
by h.b.furuseth＠usit.uio.no 07 Jul '09

07 Jul '09

I wrote: > We need to check > grep -nH -e 'SlapReply[<tab> ]*[a-zA-Z]' *.c */*.c > for other cases. Not for this ITS. Just moved sru_search first in the union. It's not always clear to me which code is responsible for initializing SlapReply anyway. Fixed two other simple cases: - result.c:send_ldap_disconnect() should reset rs->sr_rspdata. - sasl.c: should (re)set rs->sr_sasldata before doing send_ldap_sasl(). -- Hallvard

1 0

Re: (ITS#5836) hung writers
by quanah＠zimbra.com 07 Jul '09

07 Jul '09

--On Tuesday, July 07, 2009 4:41 PM +0000 quanah(a)zimbra.com wrote: > Also, please confirm what BDB version you are using, including if it has > all patches applied and compile options for BDB, and which backend you > are using. Hi Jorge, Please also confirm that you have writetimeout configured in your slapd configuration. --Quanah -- Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc -------------------- Zimbra :: the leader in open source messaging and collaboration

1 0

Re: (ITS#5836) hung writers
by quanah＠zimbra.com 07 Jul '09

07 Jul '09

--On Tuesday, July 07, 2009 12:51 PM +0000 jorge.perez(a)adaptia.net wrote: Hello, > > The problem still reproduces. Threads hung on the same > ptread_cond_wait.call. This time the assert has not failed. > > Thread 18 (Thread 1390426432 (LWP 17780)): > ---Type <return> to continue, or q <return> to quit--- ># 0=C2=A0 0x00002b5a8515b1c6 in pthread_cond_wait@@GLIBC_2.3.2 () from > /lib64/libpthread.so.0 ># 1=C2=A0 0x00000000004525a4 in send_ldap_ber (op=3D<value optimized out>, > ber=3D0x52e01150) at result.c:156 ># 2=C2=A0 0x000000000045449e in slap_send_search_entry ># (op=3D0x2b5a849138f4, > rs=3D0x52e02da0) at result.c:1259 > Hi Jorge, We're going to need more information from you than you are providing. Please make sure you: (a) build openldap without optimization (-O0) (b) are gdbing the process, not a core file (c) Send a full backtrace of everything: thr apply all bt full Also, please confirm what BDB version you are using, including if it has all patches applied and compile options for BDB, and which backend you are using. Thanks! --Quanah -- Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc -------------------- Zimbra :: the leader in open source messaging and collaboration

1 0

← Newer
1
...
9
10
11
12
13
14
15
16
17
Older →

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

openldap-bugs July 2009