| Age | Commit message (Collapse) | Author |
|---|
|
|
|
* lukas/erts/fix-seq_trace-reset_trace/OTP-15490:
erts: Fix seq_trace:reset_trace dirty gc bug
erts: Use sys_memcpy in copy_one_frag
|
|
When seq_trace:reset_trace could be called while a
process was doing a dirty GC. This triggered a race
where all signals was moved to the internal signal
queue during the GC which in turn caused the a heap
overrun problem.
This fix makes it so that the main and msgq lock are
taken before the clear. This will make sure that we
are allowed to do the clear.
|
|
|
|
* siri/logger/std-recreate-file-if-rotated/OTP-15469:
[logger] Re-create log file if deleted
|
|
into maint
* siri/logger/unexpected-exit-from-simple/ERL-788/OTP-15466:
[logger] Unlink simple handler before terminating
|
|
This is an update to logger_std_h, which makes it play well with tools
like logrotate.
|
|
* maint-21:
Updated OTP version
Prepare release
|
|
|
|
This avoids an unexpected EXIT message with reason normal beeing
printed from logger_server when the simple handler is replaced. It
will still be printed if it exists unexpectedly.
|
|
* siri/logger/ets-read-concurrency/ERL-782/OTP-15453:
[logger] Add read_concurrency to options for logger ets table
|
|
* siri/logger/refactor-built-in-handlers:
[logger] Write asynchronously to disk log
[logger] Add statistics counter for drops due to message bursts
[logger] Further refactor built-in handlers
[logger] Move common handler code to logger_h_common
|
|
|
|
'sverker/kernel/net_kernel-failed-accept_pending/ERL-781/OTP-15438' into maint
* sverker/kernel/net_kernel-failed-accept_pending/ERL-781/OTP-15438:
kernel: Fix bug in net_kernel for connection attempt from self
|
|
net_kernel crashed due to invalid return from handle_info.
Could only happen if erts_internal:new_connection fails
due to non-atom Node or if Node is the own node name.
|
|
Fix the default implementation of address_please
|
|
This commit updates logger_disk_log_h to write most log events
asynchronlously (disk_log:balog/2), but synchronously
(disk_log:blog/2) on every 20th event in the same way as logger_std_h
does. This is for efficiency reasons.
|
|
|
|
|
|
There was a lot of duplicated code in logger_std_h and
logger_disk_log_h. Most of this is now moved to logger_h_common, which
now also serves as the gen_server callback.
|
|
Forward set/update indicator to handler callback changing_config
OTP-15364
|
|
* igor/tcp-nopush-ERL-698/OTP-15357:
"cork" tcp socket around file:sendfile
Add nopush TCP socket option
|
|
|
|
RaimoNiskanen/raimo/getifaddrs-netns/ERIERL-189/OTP-15121
Implement {netns,NS} option for inet:getifaddrs/1 and friends
|
|
This translates to TCP_CORK on Linux and TCP_NOPUSH on
BSD.
In effect, this acts as super-Nagle: no partial TCP segments
are sent out until this option is turned off. Once turned off,
all accumulated unsent data is sent out immediately. The latter
is *not* the case on OSX, hence the implementation ignores
"nopush" on OSX to reduce confusion.
|
|
This function is called when a logger API function for fetching
handler configuration is called. The point is to allow the handler to
remove internal data fields that it might have stored in the
configuration database, before returning the handler configuration to
the caller.
An example of such internal data are the 'handler_pid' and 'mode_tab'
fields that logger_std_h and logger_disk_log_h store in their
configuration maps.
|
|
The new parameter to this function, SetOrUpdate, indicates how
unspecified configuration data fields shall be set. The rule is that
if SetOrUpdate equals set, then default values shall be used, and if
SetOrUpdate equals update, then existing configuration values shall be
used. Consequently, these examples now apply to logger_std_h and
logger_disk_log_h:
set_handler_config(default, config, #{sync_mode_qlen => 20}) sets the
value of sync_mode_qlen to 20, and resets all other (writable) fields
in the config map to their default values.
update_handler_config(default, config, #{sync_mode_qlen => 20}) sets
the value of sync_mode_qlen to 20, and leaves all other fields in the
config map unchanged.
|
|
This function is similar to set_handler_config/3, but calls the handler
callback changing_config/3 with the first parameter SetOrUpdate=update.
|
|
Also implement the same option for the legacy undocumented functions
inet:getif/1,getiflist/1,ifget/2,ifset/2.
The arity 1 functions had before this change got signatures that
took a socket port that was used to do the needed syscall, so now
the signature was extended to also take an option list with the
only supported option {netns,Namespace}. The Socket argument
variant remains unsupported.
For inet:getifaddrs/1 the documentation file was changed to old
style function name definition so be able to hide the Socket
argument variant that is visible in the type spec.
The arity 2 functions had got an option list as second argument.
This list had to be partitioned into one list for the namespace
option(s) and the other for the rest.
The namespace option list was then fed to the already existing
namespace support for socket opening, which places the socket
in a namespace and hence made all these functions that in
inet_drv.c used getsockopt() work without change.
The functions that used getifaddrs() in inet_drv.c had to be
changed in inet_drv.c to swap namespaces around the
getifaddrs() syscall. This functionality was separated into
a new function call_getifaddrs().
|
|
|
|
Since logger has now knowledge of the handler specific data ('config'
field in the handler configuration map), it can not merge this data
structure with default or old values upon configuration
change. Instead, the handler callback changing_config must do
this. Earlier, this callback did not reflect if the configuration
change was initiated by a call to set_handler_config or
update_handler_config, so the handler did not know if unspecified
fields should get default values or the values from the existing
configuration.
To overcome this problem, the new parameter SetOrUpdate is added to
this callback. If SetOrUpdate equals set, then default values should
be used. If SetOrUpdate equals update, then existing configuration
values should be used.
|
|
* sverker/erts/robustify-dist-entry-states/OTP-15297:
erts: Refactor port dist_entry & conn_id into PRTSD
Remove ugly fail case macros
Consolidate distribution entry state transitions
erts: Fix bug in undocumented system_flag(scheduling_statistics)
|
|
* Make connection_id part of the distribution handle as {ConnId, DistEntry}
in order for BIFs to verify correct connection.
* Make distribution handle opaque to net_kernel.
* Remove some unsafe lockless reads of DistEntry.flags
* Change state ERTS_DE_STATE_EXITING to be more of an internal state that
prevents erts from enqueue, encode or schedule new data to be sent. Otherwise
it should behave like ERTS_DE_STATE_CONNECTED.
|
|
* 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
|
|
When a handler process is terminated due to overload, it reads its
configuration from the configuration database, so it can be restarted
with the same configuration after a small delay. This was earlier done
in a different process, which was spawned off from the terminate
function. This caused a race condition, where in some cases, the
configuration was already removed before it could be read.
The reason for spawning off a process, is to avoid a deadlock due to
the call to logger:remove_handler/1.
This commit moves the call to logger:get_handler_config/1 back to the
handler process - to ensure that the data is still there, but keeps
the call to logger:remove_handler/1 in the spawned off process - to
avoid deadlock.
|
|
|
|
If the report callback function has two arguments, the second argument
is a map with options to limit the size of the log event. To even
allow a better formatting when the event shall be printed on a single
line, the new option single_line is now included in this
argument.
|
|
If a filter fails, it is removed by a call to
logger:remove_filter/1,2. If the log event that caused the failure was
issued on the logger process itself, this would earlier cause a
deadlock due to a gen_server:call to self(). This is now prevented.
|
|
|
|
|
|
|
|
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.
|
|
|
|
* maint-21:
Updated OTP version
Update release notes
Update version numbers
kernel: Fix missing abort_connection in net_kernel
Prevent inconsistent node lists
Fix an endless rescheduling loop when a process is executing process_info(self(), ...)
|
|
|
|
|
|
Implement socket options recvtclass, recvtos, recvttl and pktoptions.
Document the implemented socket options, new types and message formats.
The options recvtclass, recvtos and recvttl are boolean options that
when activated (true) for a socket will cause ancillary data to be
received through recvmsg(). That is for packet oriented sockets
(UDP and SCTP).
The required options for this feature were recvtclass and recvtos,
and recvttl was only added to test that the ancillary data parsing
handled multiple data items in one message correctly.
These options does not work on Windows since ancillary data
is not handled by the Winsock2 API.
For stream sockets (TCP) there is no clear connection between
a received packet and what is returned when reading data from
the socket, so recvmsg() is not useful. It is possible to get
the same ancillary data through a getsockopt() call with
the IPv6 socket option IPV6_PKTOPTIONS, on Linux named
IPV6_2292PKTOPTIONS after the now obsoleted RFC where it originated.
(unfortunately RFC 3542 that obsoletes it explicitly undefines
this way to get packet ancillary data from a stream socket)
Linux also has got a way to get packet ancillary data for IPv4
TCP sockets through a getsockopt() call with IP_PKTOPTIONS,
which appears to be Linux specific.
This implementation uses a flag field in the inet_drv.c socket
internal data that records if any setsockopt() call with recvtclass,
recvtos or recvttl (IPV6_RECVTCLASS, IP_RECVTOS or IP_RECVTTL)
has been activated. If so recvmsg() is used instead of recvfrom().
Ancillary data is delivered to the application by a new return
tuple format from gen_udp:recv/2,3 containing a list of
ancillary data tuples [{tclass,TCLASS} | {tos,TOS} | {ttl,TTL}],
as returned by recvmsg(). For a socket in active mode a new
message format, containing the ancillary data list, delivers
the data in the same way.
For gen_sctp the ancillary data is delivered in the same way,
except that the gen_sctp return tuple format already contained
an ancillary data list so there are just more possible elements
when using these socket options. Note that the active mode
message format has got an extra tuple level for the ancillary
data compared to what is now implemented gen_udp.
The gen_sctp active mode format was considered to be the odd one
- now all tuples containing ancillary data are flat,
except for gen_sctp active mode.
Note that testing has not shown that Linux SCTP sockets deliver
any ancillary data for these socket options, so it is probably
not implemented yet. Remains to be seen what FreeBSD does...
For gen_tcp inet:getopts([pktoptions]) will deliver the latest
received ancillary data for any activated socket option recvtclass,
recvtos or recvttl, on platforms where IP_PKTOPTIONS is defined
for an IPv4 socket, or where IPV6_PKTOPTIONS or IPV6_2292PKTOPTIONS
is defined for an IPv6 socket. It will be delivered as a
list of ancillary data items in the same way as for gen_udp
(and gen_sctp).
On some platforms, e.g the BSD:s, when you activate IP_RECVTOS
you get ancillary data tagged IP_RECVTOS with the TOS value,
but on Linux you get ancillary data tagged IP_TOS with the
TOS value. Linux follows the style of RFC 2292, and the BSD:s
use an older notion. For RFC 2292 that defines the IP_PKTOPTIONS
socket option it is more logical to tag the items with the
tag that is the item's, than with the tag that defines that you
want the item. Therefore this implementation translates all
BSD style ancillary data tags to the corresponding Linux style
data tags, so the application will only see the tags 'tclass',
'tos' and 'ttl' on all platforms.
|
|
Add logger:set_application_level/2
|
|
When printing a map with ~p in a log message, the formatter's
single_line option in some cases introduced unwanted commas after
'=>', for example:
This is now corrected. When single_line==true, width 0 is added to all
~p and ~P control sequences in the format string.
|
|
OTP-15146
|
