| Age | Commit message (Collapse) | Author |
|---|
|
* tuncer/ei-breakage:
ei: fix breakage reported by Sergei Golovan
Revert "musl: fix gethostbyname_r/gethostbyaddr_ selection"
OTP-13328
|
|
The block count is an unsigned 2 byte integer. The common behaviour of
tftp clients and servers is to overflow the block count and wrap the
counter to 0 in order to support files larger than 32M. Modify the tftp
implementation to do the same rollover.
Interoperability was tested by transferring a 1.4G file using the HPA
tftp client.
|
|
|
|
* dgud/mnesia/del_schema_hangs/OTP-13284:
mnesia: let loader check if tablelock is needed
mnesia: Avoid deadlock possibility in mnesia:del_table_copy schema
|
|
Allow IPv6 nodes to register with and query epmd. On systems with
IPv6 support:
* epmd listens on both the IPv4 and IPv6 ANY or loopback sockets
* the epmd cli client connects to epmd over the IPv6 loopback
* distributed nodes started with "-proto_dist inet6_tcp" will register
with epmd over IPv6
To work on IPv6 capable systems that have IPv6 support disabled,
epmd ignores errors opening the socket if the protocol is not
supported. Similarly, the epmd client will fall back to IPv4 if the IPv6
socket is not available.
Update the minimum supported version of Windows to Windows Vista to
support IPv6.
|
|
|
|
* bjorn/compiler/beam_bsm/OTP-13309:
Eliminate crash because of unsafe delaying of sub-binary creation
|
|
* nybek/speed_up_supervisor_count_children:
Speed up supervisor:count_children/1; simple_one_for_one
Add supervisor:get_callback_module/1
OTP-13290
|
|
The current implementation is roughly O(N*M) where N is the number of items to be removed, and M is the number of items in the map. This does not include the cost of `maps:from_list` or `maps:to_list`. This leads to pretty horrifying execution times on large maps regardless of how many or few keys are to be removed.
The new implementation is O(N) where N is the number of items to be removed. For each N there's the cost of removing a key from a map, and but in practice that turns out to be a vast improvement for all map sizes I tested
The new maps:take/2 implementation similarly builds a list of keys and values by iterating only the list of desired keys, and then hands it off to maps:from_list. This turned out to be faster than N maps:put calls.
|
|
move_table_copy needs the lock that was set previously in del_table_copy.
This doesn't work on old nodes, so bump protocol version and check it.
Remove old protocol conversion code, which have been around since OTP-R15.
Checking if lock is needed requires rpc communication via mnesia_gvar
ets table to be backwards compatible.
|
|
The argument of byte_size() is a bitstring().
The code in erl_bif_types that finds cases where comparisons always
return true or false is corrected when it comes to maps and bit
strings.
|
|
Each service process maintains a dictionary of peers, mapping an
application alias to a {pid(), #diameter_caps{}} list of connected
peers. These lists are potentially large, peers were appended to the end
of the list for no particular reason, and these long lists were
constructed/deconstructed when filtering them for pick_peer callbacks.
Many simultaneous outgoing request could then slow the VM to a crawl,
with many scheduled processes mired in list manipulation.
The pseudo-dicts are now replaced by plain ets tables. The reason for
them was (once upon a time) to have an interface interchangeable with a
plain dict for debugging purposes, but strict swapablity hasn't been the
case for some time now, and in practice a swap has never taken place.
Additional tables mapping Origin-Host/Realm have also been introduced,
to minimize the size of the peers lists when peers are filtered on
host/realm. For example, a filter like
{any, [{all, [realm, host]}, realm]}
is probably a very common case: preferring a Destination-Realm/Host
match before falling back on Destination-Realm alone. This is now more
efficiently (but not equivalently) expressed as
{first, [{all, [realm, host]}, realm]}
to stop the search when the best match is made, and extracts peers from
host/realm tables instead of searching through the list of all peers
supporting the application in question. The code to try and start with a
lookup isn't exhaustive, and the 'any' filter is still as inefficient as
previously.
|
|
See commit 862af31d.
|
|
OTP-13311
* sverk/ecc-fixes:
Ensure testing ssl with supported ciphers only
Only use supported EC curves in crypto tests
Check the result of EC_GROUP_new_curve_* calls
|
|
When searching for a certificate's issuer in the `CertDB`, verify the signature
against the original DER certificate from the handshake instead of a re-encoding
of the parsed certificate. This avoids false negatives due to differences
between DER encoding implementations of OTP and other platforms.
|
|
|
|
* ia/ssl/handshake-fragmentation/OTP-13306:
ssl: Big handshake messages needs to be fragmented on TLS record level
|
|
[James Fish:]
Dialyzer always asserts that files and directories passed in its
options exist. Therefore it is not possible to remove a beam/module
from a PLT when the beam file no longer exists. Dialyzer should not to
check files exist on disk when removing from the PLT.
|
|
|
|
* anders/diameter/appup/OTP-13283:
Add missing appup after 17.5.6.8 merge
|
|
The following code would fail to compile:
decode(<<Code/integer, Bin/binary>>) ->
<<C1/integer, B1/binary>> = Bin,
case C1 of
X when X =:= 1 orelse X =:= 2 ->
Bin2 = <<>>;
_ ->
Bin2 = B1
end,
case Code of
1 -> decode(Bin2);
_ -> Bin2
end.
The error message would be:
t: function decode/1+28:
Internal consistency check failed - please report this bug.
Instruction: return
Error: {match_context,{x,0}}:
The beam_bsm pass would delay the creation of a sub-binary when it was
unsafe to do so. The culprit was the btb_follow_branch/3 function that
for performance reasons cached labels that had already been checked.
The problem was the safety of a label also depends on the contents
of the registers. Therefore, the key for caching needs to be both
the label and the register contents.
Reported-by: José Valim
|
|
A previous commit introduced a change that exposed dead code in this
module and caused dialyzer warnings. This dead code was taken out.
While at it, performed some code refactoring in the handling of the
bs_get_integer primop.
|
|
|
|
del_table_copy grabs a write lock in a new process in prepare_op/3 to
change 'where_to_read' when a table copy is updated.
When del_table_copy(schema, Node) is called all copies located on Node
are deleted, and thus many locks are taken. Since this was done outside
of the schema-transaction, mnesia's deadlock prevention algorithms
was sidestepped and a deadlock could occur.
Fix by always grabbing write-locks for all changed tabs early and in the same
transaction, this might slow done the operation some but it must be done
and it also cleans up the code.
|
|
|
|
* ia/ssl/validator:
ssl: Only start a new session validator if the old one has finished its work
|
|
* tnt-dev/fix-function-index-sort:
Don't sort function index when it's not necessary
OTP-13302
|
|
* legoscia/tls-dist-listen-ip:
TLS distribution: bind erts socket to localhost
OTP-13300
|
|
* legoscia/patch-3:
Fix documentation of net_kernel:allow/1
OTP-13299
|
|
When handling a gen_tcp socket, it's handy to be able to give it its proper type
and have dialyzer be able to validate it, rather than falling back on using
port(). The gen_udp equivalent is already exported.
|
|
* rj/release-handler-noproc-fix:
Fix noproc crash during release installation
OTP-13291
|
|
Speed up supervisor:count_children/1 for simple_one_for_one
supervisors. This is achieved by avoiding looping through all the
child process and verifying that each one is alive.
For a supervisor with 100,000 'temporary' children the count-time will
drop from approx 25ms to about 0.005ms.
For a supervisor with 100,000 'permanent' or 'transient' children the
count-time will drop from approx 30ms to about 0.005ms.
This avoids having the supervisor block for an extended period while
the count takes place. Under normal circumstances the accuracy of the
result should also improve since the duration is too short for many
processes to die during the count.
|
|
This function is used by release_handler during upgrade. This was
earlier implemented in the release_handler, but it required a copy og
the definition of the supervisor's internal state, which caused
problems when this state was updated.
|
|
|
|
When cover instruments binary comprehensions it's generating a
{block, ...} abstract code term inside a {bc, ...} term that is causing
the evaluation to fail at runtime. Removing the block statement
eliminates the error.
The template of a bit string comprehension cannot have a counter since
it is not allowed to be a block.
|
|
* rickard/monotonic-time-improvements/OTP-13222:
Use nano second time unit in tracing
|
|
|
|
Generalise much of inet_tls_dist, so that inet6_tls_dist can reuse it.
|
|
* margnus1/bs_unit_fix:
hipe: Fix signed compares of unsigned sizes
beam: Fix overflow bug in i_bs_add_jId
hipe: Add tests for bad bit syntax float sizes
Add a case testing the handling of guards involving binaries
Add some more binary syntax construction tests
hipe: Guard against enormous numbers in ranges
hipe: Fix constructing huge binaries
hipe: Fix binary constructions failing with badarith
Add missing corner-case to bs_construct_SUITE
hipe: Allow unsigned args in hipe_rtl_arith
hipe: test unit size match in bs_put_binary_all
hipe: test unit size match in bs_append
Fix hipe_rtl_binary_construct:floorlog2/1
OTP-13272
|
|
* aronisstav/dialyzer-missing-callback-info:
Fix inadvertent deletion of callback info
Eliminate ugly case statements
Remove dead code related to missing behaviour info
OTP-13287
|
|
* palas/maint:
inets: Add PATCH method to client and server
OTP-13286
|
|
* legoscia/tls-dist-connect-options:
ssl_dist_SUITE: don't use deprecated functions
TLS distribution: support inet_dist_connect_options
OTP-13285
|
|
* rickard/monotonic-time-improvements/OTP-13222:
Introduce time management in native APIs
Introduce time warp safe replacement for safe_fixed option
Introduce time warp safe trace timestamp formats
|
|
* lrascao/stdlib/beam_lib-race-condition/OTP-13278:
Check for already started beam_lib crypto server
|
|
|
|
When release_handler_1:get_supervised_procs/0 does a recursive walk of
the supervision tree, it calls sys:get_status/1 on supervisors, to check
if they are suspended or running.
This fixes a race condition where a list of supervisor pids is gathered,
one of them (legitimately) exits before release_handler can examine it,
then sys:get_status/1 is called with a dead pid, causing an exit(noproc)
See: http://erlang.org/pipermail/erlang-questions/2015-August/085712.html
(To recreate this problem for testing, I added a timer:sleep into the
release_handler_1 code, and killed a supervisor during
get_supervised_procs being called).
|
|
|
|
* ia/maint/inets/mod_alias/OTP-13248:
inets: Prepare for release
inets: Traverse all aliases looking for the longest match
inets: Use re instead of inets_regexp
# Conflicts:
# lib/inets/vsn.mk
|
|
In commit b8c6a119.
|
|
* gomoripeti/dbg/read_ahead/OTP-13279:
Use file read buffering in dbg:trace_client
|
