1017835 - DataChannelConnection::SendPacket should use array delete

Status: RESOLVED FIXED

(Core :: WebRTC: Networking, defect)

Description

[Eric Rahm [:erahm]]

The buffer[1] passed to |DataChannelConnection::SendPacket| is allocated with an array new, it should be freed with a corresponding array delete[2].


[1] http://dxr.mozilla.org/mozilla-central/source/netwerk/sctp/datachannel/DataChannel.cpp#716
[2] http://dxr.mozilla.org/mozilla-central/source/netwerk/sctp/datachannel/DataChannel.cpp#686

Comment 1

[Randell Jesup [:jesup] (needinfo me)]

Attachment: use array delete for SCTP/DTLS SendPacket data

Comment 2

[Jan-Ivar Bruaroey [:jib] (needinfo? me)]

Comment on attachment 8432263 [details] [diff] [review]
use array delete for SCTP/DTLS SendPacket data

Review of attachment 8432263 [details] [diff] [review]:
-----------------------------------------------------------------

Comments optional.

::: netwerk/sctp/datachannel/DataChannel.cpp
@@ +678,4 @@
>  }
>  
>  int
>  DataChannelConnection::SendPacket(const unsigned char *data, size_t len, bool release)

I find const here a little weird when transferring ownership.

Also, how about DataChannelConnection::SendPacket(unsigned char data [], size_t len, bool release) to hint that an array is expected?

Comment 3

[Randell Jesup [:jesup] (needinfo me)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/f04ad92a05ea
Forgot to add the r=jib...

Comment 4

[Randell Jesup [:jesup] (needinfo me)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/c34082c15a99

Comment 5

[Eric Rescorla (:ekr)]

Comment on attachment 8432263 [details] [diff] [review]
use array delete for SCTP/DTLS SendPacket data

Review of attachment 8432263 [details] [diff] [review]:
-----------------------------------------------------------------

::: netwerk/sctp/datachannel/DataChannel.cpp
@@ +678,4 @@
>  }
>  
>  int
>  DataChannelConnection::SendPacket(const unsigned char *data, size_t len, bool release)

If you're going to transfer ownership, wouldn't it make more sense to use a smart pointer instead to delineate this clearly?

Comment 6

[Phil Ringnalda (:philor)]

https://hg.mozilla.org/mozilla-central/rev/c34082c15a99