View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0003872 | GNUnet | cadet service | public | 2015-06-30 09:38 | 2018-06-07 00:25 |
| Reporter | Christian Grothoff | Assigned To | Bart Polot | ||
| Priority | normal | Severity | major | Reproducibility | have not tried |
| Status | closed | Resolution | fixed | ||
| Platform | i7 | OS | Debian GNU/Linux | OS Version | squeeze |
| Product Version | Git master | ||||
| Target Version | 0.11.0pre66 | Fixed in Version | 0.11.0pre66 | ||
| Summary | 0003872: valgrind issue in ax_debug (tunnel.c:4388) | ||||
| Description | ==946== Use of uninitialised value of size 8 ==946== at 0x50A39A5: GNUNET_STRINGS_data_to_string (strings.c:898) ==946== by 0x5072E4D: GNUNET_CRYPTO_hash_to_enc (crypto_hash.c:71) ==946== by 0x50588B4: GNUNET_h2s (common_logging.c:1055) ==946== by 0x4111E6: ax_debug (gnunet-service-cadet_tunnel.c:4388) ==946== by 0x41145F: GCT_debug (gnunet-service-cadet_tunnel.c:4433) ==946== by 0x409670: send_kx (gnunet-service-cadet_tunnel.c:2147) ==946== by 0x410C80: GCT_send_ax_kx (gnunet-service-cadet_tunnel.c:4228) ==946== by 0x4090D1: ax_kx_resend (gnunet-service-cadet_tunnel.c:2042) ==946== by 0x5091D4B: run_ready (scheduler.c:587) ==946== by 0x5092656: GNUNET_SCHEDULER_run (scheduler.c:867) ==946== by 0x50A0E6A: GNUNET_SERVICE_run (service.c:1503) ==946== by 0x43328B: main (gnunet-service-cadet.c:174) ==946== | ||||
| Steps To Reproduce | Just running a peer with SVN 36026. | ||||
| Tags | No tags attached. | ||||
|
|
Could you confirm that the line (gnunet-service-cadet_tunnel.c:4388) is: LOG2 (level, "TTT DHRs\t %s\n", GNUNET_h2s ((struct GNUNET_HashCode *) &pub)); pub is a local variable, no heap should be involved... |
|
|
Yes, but 'pub' is a 32-byte public key, and you're CASTING it to a 64-byte hash for no good reason. So you're accessing uninitialized memory on the stack. This kind of casting must stop. You should be using GNUNET_i2s. |
|
|
The reason was that h2s shows more characters than i2s. I'll change that. |
|
|
That's still a terrible reason to cast a 32-byte value to a 64-byte value (or to cast at all). Also, there is GNUNET_i2s_full. |
|
|
Still no fix for this one? |
|
|
It's fixed, upload pending when I have access to my laptop! |
|
|
Should be fixed after the last commits. |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2015-06-30 09:38 | Christian Grothoff | New Issue | |
| 2015-06-30 09:38 | Christian Grothoff | Status | new => assigned |
| 2015-06-30 09:38 | Christian Grothoff | Assigned To | => Bart Polot |
| 2015-06-30 11:29 | Bart Polot | Note Added: 0009377 | |
| 2015-06-30 11:31 | Christian Grothoff | Note Added: 0009378 | |
| 2015-06-30 11:35 | Bart Polot | Note Added: 0009380 | |
| 2015-06-30 11:37 | Christian Grothoff | Note Added: 0009382 | |
| 2015-07-15 07:36 | Christian Grothoff | Note Added: 0009441 | |
| 2015-07-15 13:06 | Bart Polot | Note Added: 0009445 | |
| 2015-07-15 14:08 | Bart Polot | Note Added: 0009447 | |
| 2015-07-15 14:08 | Bart Polot | Status | assigned => resolved |
| 2015-07-15 14:08 | Bart Polot | Fixed in Version | => Git master |
| 2015-07-15 14:08 | Bart Polot | Resolution | open => fixed |
| 2015-08-03 15:18 | Christian Grothoff | Fixed in Version | Git master => 0.11.0pre66 |
| 2018-06-07 00:25 | Christian Grothoff | Status | resolved => closed |
