| Age | Commit message (Collapse) | Author |
|---|
|
Added support for ip level socket option RECVOPTS.
OTP-14831
|
|
Added support for ip level socket option FREEBIND.
Note that there is an option available on FreeBSD called
IP_BINDANY, which seems to have similar properties (FREEBIND
is *not* available on FreeBSD). There are some restrictions
for this option though (which is not mentioned in the Linux
man page).
OTP-14831
|
|
Added support for socket level socket option TIMESTAMP.
OTP-14831
|
|
Added support for socket level socket option RCVLOWAT and SNDLOWAT.
These are both a little strange, at least on Linux. See the man
pages for more info. For instance, sndlowat cannot be set on Linux.
OTP-14831
|
|
Added support for socket level socket option RCVTIMEO and SNDTIMEO.
These are both a little strange, at least on linux. See the man
pages for more info.
OTP-14831
|
|
Added support for the SCTP option INITMSG.
OTP-14831
|
|
Added support for the SCTP option MAXSEG.
OTP-14831
|
|
Added support for the SCTP option RTOINFO.
We have the same questions for this option as for ASSOCINFO.
Maybe the assoc field shall be made 'out' only (that is, it will
be ignored for setopt). The assoc id used will be that which
is stored in the descriptor (how do we get it to begin with?).
Questions, questions...
OTP-14831
|
|
Added support for the SCTP option ASSOCINFO.
This option is a bit tricky. As the underlying structure (sctp_assocparams)
contains the assoc_id, it begs the question what happens if this option
is fetched for:
* The own assoc (which means that we might have the assoc id in the
descriptor and can initiate that part of the struct accordningly).
* Another assoc: From assoc A asks for info with assoc_id set to
that of assoc B.
* The "owning" endpoint.
* Another endpoint (an endpoint to which the assoc does not belong).
So, if the user calls socket:[getopt|setopt] for an association socket,
shall we require that the assoc_id field is set to -1? Or not set at all
and therefor filled in automatically by the nif-code?
And, if the user calls socket:[getopt|setopt] for an endpoint socket,
shall we require that the assoc_id field is set to a valid id? Or shall
it not be allowed?
Questions, questions...
OTP-14831
|
|
Added support for the SCTP option DISABLE_FRAGMENTS.
OTP-14831
|
|
Added support for the SCTP option EVENTS.
OTP-14831
|
|
Added support for the VPv6 socket option MTU.
OTP-14831.
|
|
Its now possible to have multiple (simultaneous) acceptor
processes for the same listening socket.
OTP-14831
|
|
Added support for socket level socket option REUSEPORT.
OTP-14831
|
|
Added support for socket level socket option BINDTODEVICE.
OTP-14831
|
|
Added support for the IP option RECVTTL.
OTP-14831
|
|
Added support for the IP option NODEFRAG.
OTP-14831
|
|
Added support for the IP option MULTICAST_ALL.
OTP-14831
|
|
Added support for the IP option MTU_DISCOVER.
OTP-14831
|
|
Added support for the IP option MTU.
OTP-14831
|
|
Added support for the IP option MULTICAST_IF.
OTP-14831
|
|
Added support for the IP option MULTICAST_TTL.
OTP-14831
|
|
Added support for the IP option MULTICAST_LOOP.
OTP-14831
|
|
Added support for socket level socket option oobinline (both get
and set).
OTP-14831
|
|
The socket option (level socket) peek_off is now supported.
OTP-14831
|
|
The socket option (level socket) debug is now supported.
To actually set this option, the user must have the proper
"authority" (CAP_NET_ADMIN capability or an effective user ID of 0).
OTP-14831
|
|
The socket (level socket) option acceptconn is now supported
(for getopt).
OTP-14831
|
|
Make it possible to *get* the socket options domain and
protocol (in addition to type).
OTP-14831
|
|
|
|
Fixed handling of recvfrom (used by dgram sockets).
Had forgot to do select(read) when we got block from
the call to recvfrom. Argh!
Also updated the (simple) test server and client to to
be able to use udp (dgram+udp).
OTP-14831
|
|
|
|
Fixed handling of closed in the recv function.
We still need to properly handle when we get
0 bytes of data for other types ock sockets
then stream (its valid for dgram for instance).
OTP-14831
|
|
Added some more features o the simple socket test server
and client.
OTP-14831
|
|
|
|
|
|
* siri/cuddle:
[logger] Skip app level test if app cannot be loaded
Terminate nodes after failed test cases in erl_distribution_SUITE
Skip performance test on build types other than 'opt'
|
|
|
|
|
|
* raimo/receive-TOS-TCLASS/ERIERL-187/OTP-15145:
Write testcases for recvtos and friends
Fix term buffer overflow bug
Fix documentation due to feedback
Implement socket option recvtos and friends
|
|
Use os:type() and os:version() to predict if the individual options
are supposed to be supported. We'll see if this holds.
|
|
A hanging node from this suite can fail later tests, so terminate all
nodes() after each test case.
|
|
* maint:
kernel: Adjust disk_log tests
kernel: Adjust global testcase
|
|
* hasse/kernel/adjust_tests:
kernel: Adjust disk_log tests
kernel: Adjust global testcase
|
|
|
|
|
|
|
|
|
|
|
|
The encoding option was introduced in commit
270d909696a753af022df72a404c73f2895b4a02, to allow report callbacks to
format according to a given encoding. There was, however, no
connection between this encoding option, and the encoding of the
device to which the logger handler was writing.
Since a formatter is defined to return unicode:chardata(), and in
order to avoid mismatch with the encoding of the device, the encoding
option is now removed from the formatter. The handler itself must make
sure that it does not write illegal data to its device.
|
|
|
