View Issue Details

IDProjectCategoryView StatusLast Update
0001586GNUnetfile-sharing servicepublic2011-09-19 10:45
ReporterChristian Grothoff Assigned ToChristian Grothoff  
PrioritynormalSeveritymajorReproducibilityalways
Status closedResolutionfixed 
Summary0001586: FS transmits garbage? Syscall param socketcall.sendto(msg) points to uninitialised byte(s)
DescriptionStart hostlist peer (with fs under valgrind with content), start second peer. On P2P core connect I get:

Jul 22 13:57:55 transport-8925 WARNING Received `PING' message not destined for me!
==8923== Syscall param socketcall.sendto(msg) points to uninitialised byte(s)
==8923== at 0x5FF15B5: send (in /lib/libc-2.7.so)
==8923== by 0x587FC2C: GNUNET_NETWORK_socket_send (network.c:536)
==8923== by 0x586F2DF: transmit_ready (connection.c:1516)
==8923== by 0x58868F1: run_ready (scheduler.c:604)
==8923== by 0x5886CB5: GNUNET_SCHEDULER_run (scheduler.c:731)
==8923== by 0x588D648: GNUNET_SERVICE_run (service.c:1568)
==8923== by 0x408CC0: main (gnunet-service-fs.c:3808)
==8923== Address 0x75d86c8 is 16 bytes inside a block of size 32,864 alloc'd
==8923== at 0x4C22741: realloc (vg_replace_malloc.c:429)
==8923== by 0x5869839: GNUNET_xrealloc_ (common_allocation.c:131)
==8923== by 0x586F49E: GNUNET_CONNECTION_notify_transmit_ready (connection.c:1595)
==8923== by 0x5869295: GNUNET_CLIENT_notify_transmit_ready (client.c:880)
==8923== by 0x56577B7: trigger_next_request (core_api.c:366)
==8923== by 0x5658BB2: GNUNET_CORE_notify_transmit_ready (core_api.c:946)
==8923== by 0x4038F6: consider_migration (gnunet-service-fs.c:895)
==8923== by 0x4042B4: peer_connect_handler (gnunet-service-fs.c:1215)
==8923== by 0x565793A: main_notify_handler (core_api.c:424)
==8923== by 0x58689E3: receive_task (client.c:543)
==8923== by 0x58868F1: run_ready (scheduler.c:604)
==8923== by 0x5886CB5: GNUNET_SCHEDULER_run (scheduler.c:731)
Jul 22 13:57:56 core-8924 ERROR Assertion failed at gnunet-service-core.c:3661.
Additional InformationI suspect the cause is this specific call from fs to transmit data in "consider_migration", which somehow (not obvious from code why) fails to initialize some data in the buffer and then the core service complains about it.

Still need to look at the exact core assertion failure, maybe that'll help explain things.
TagsNo tags attached.

Activities

Christian Grothoff

2010-10-20 16:20

manager   ~0004133

Fixed at least one of those. Was a 'reserved' field not being set to 0. We should add checks on the receiving end for these...

Issue History

Date Modified Username Field Change
2010-07-22 14:01 Christian Grothoff New Issue
2010-07-22 14:01 Christian Grothoff Status new => assigned
2010-07-22 14:01 Christian Grothoff Assigned To => Christian Grothoff
2010-10-20 16:20 Christian Grothoff Note Added: 0004133
2010-10-20 16:20 Christian Grothoff Status assigned => resolved
2010-10-20 16:20 Christian Grothoff Fixed in Version => Git master
2010-10-20 16:20 Christian Grothoff Resolution open => fixed
2010-10-21 15:47 Christian Grothoff Status resolved => closed
2011-09-19 10:45 Christian Grothoff Fixed in Version Git master => 0.9.0pre3
2011-09-19 10:45 Christian Grothoff Target Version => 0.9.0pre3