ch_malloc of 0 bytes failed

[Edgar Fuß <ef@math.uni-bonn.de>]

This afternoon, one of our slapds died with "ch_malloc of 0 bytes failed".
Is there any known issue with 2.4.15 that can lead to this behaviour?
In case it matters, the server in question operates as a syncrepl consumer with updateref and the chain overlay.
The last messages (all within the same second) are:
do_syncrep2: cookie=rid=001,csn=20091008134919.158418Z#000000#000#000000 
syncrepl_entry: rid=001 LDAP_RES_SEARCH_ENTRY(LDAP_SYNC_MODIFY) 
syncrepl_entry: rid=001 be_search (0) 
syncrepl_entry: rid=001 uid=<censored>
slap_queue_csn: queing 0xfd1d980 20091008134919.158418Z#000000#000#000000 
slap_graduate_commit_csn: removing 0xfd1f780 20091008134919.158418Z#000000#000#000000 
syncrepl_entry: rid=001 be_modrdn (0) 
slap_queue_csn: queing 0xfd1d980 20091008134919.158418Z#000000#000#000000 
ch_malloc of 0 bytes failed

Restarting slapd simply worked.

Thanks for any help.