| Age | Commit message (Collapse) | Author |
|---|
|
At start of the VM a poll-set that the schedulers
will check is created where fds that have triggered
many (at the moment, many means 10) times without
being deselected inbetween. In this scheduler specific
poll-set fds do not use ONESHOT, which means that the
number of syscalls goes down dramatically for such fds.
This pollset is introduced in order to handle fds that
are used by the erlang distribution and that never
change their state from {active, true}.
This pollset only handles ready_input events,
ready_output is still handled by the poll threads.
During overload, polling the scheduler poll-set is done
on a 10ms timer.
|
|
OTP-15474
|
|
OTP-15473
|
|
The previous implementation uses a round-trip in the poll-set
to simulate a yield of the port context. With the poll thread
implementation this is no longer a good idea as it generated a
lot more work for the system. So this commit changes the
implementation to use a timer instead.
OTP-15471
|
|
By optimizing driver_set_timer for 0 we can use it
instead of select to do a yield in the driver.
Use full for delay_send in the inet driver.
OTP-15472
|
|
The most common case for any socket is to have zero or one
timer, so we optimize for the one case. The only case when
we have more than one timer is when the multi accept feature
is used.
|
|
|
|
OTP-15450
|
|
OTP-15450
|
|
Before this change, it was always the aux thread that was woken
to handle thread progress events scheduled to happen when all
schedulers were going to sleep. This was ok in the pre-OTP-21
implementation when the aux thread just slept on a tse. Now that
it sleeps in the fallback pollset this uses too much cpu so
instead we wake the thread that is doing the request if it is
a managed thread, or else we wake scheduler 1.
|
|
The poll thread does a lot of waking up and then going
back to sleep. A large part of the waking up is managing
thread progress and a large part of that was using thread
specific data to get the thread progress data pointer.
With this refactor the tpd is passed to each of the functions
which greatly decreases the number of ethr_get_tsd calls
which in turn halves the CPU usage of the poller thread in
certain scenarios.
|
|
|
|
|
|
maint-21
* ingela/maint/ssl/packet-options-to-tls-sender/OTP-15348:
ssl: TLS sender process needs to get updates of the socket option packet
|
|
* ingela/maint/ssl/close-alert-ERL-738/OTP-15337:
ssl: ERL-738 - Correct alert handling with new TLS sender process
|
|
maint-21
* hans/eldap/prevent_EXIT_at_close/OTP-15342/ERIERL-242:
eldap: Fix race at socket close
|
|
* john/erts/fix-unc-paths-windows/OTP-15333/ERL-737:
erts: Fix UNC path handling on Windows
erts: Fix a compiler warning
|
|
maint-21
* bjorn/compiler/fix-binary-matching/ERL-689/OTP-15335:
Fix rare bug in binary matching (again)
|
|
* raimo/BSD-sockopt-pktoptions/ERIERL-187/OTP-14297:
Fix bug for sockopt pktoptions on BSD
|
|
* john/erts/fix-prim_file-error-leak/OTP-15318:
erts: Fix memory leak on file read errors
|
|
This is unlikely to be the last of the path problems seen after
OTP 21, and I'm starting to regret my decision to unconditionally
use long paths. The idea to hit all long-path problems all the time
was good in theory as it makes such bugs far more visible, but
there just aren't enough people who test pre-release versions on
Windows, making this the world's slowest game of whack-a-mole.
|
|
If the socket option is set to {packet, 1|2|3|4} sender process needs to
add a packet length header. If packet is changed with ssl:setopts/2 this needs
to be communicated to tls_sender.
|
|
This would've been a bug if the value was used.
|
|
With the new TLS sender process, solving ERL-622, TLS ALERTs sent in
the connection state must be encrypted and sent by the TLS sender
process. This to make sure that the correct encryption state is used
to encode the ALERTS. Care must also be taken to ensure a graceful
close down behavior both for normal shutdown and downgrading from TLS
to TCP.
The original TR ERL-738 is verified by cowboy tests, and close down
behavior by our tests. However we alas have not been able to yet
create a minimal test case for the originating problem.
Also it seems it has become less likely that we run in to the TCP
delivery problem, that is the guarantee is only on transport level,
not application level. Keep work around function in ssl_test_lib but
we can have better test as long as we do not get to much wobbling
tests.
|
|
|
|
The macros for the BSD style option names had accidentally
wound up outside the option parsing loop, causing unclear
behaviour and Valgrind errors.
|
|
2e40d8d1c51a attempted fix a bug in binary matching, but it only
fixed the bug for the minimized test case.
This commit removes the previous fix and fixes the bug in a more
effective way. See the comments in the new code in `sys_core_bsm`
for an explanation.
This commit restores the optimizations in string.erl and dets_v9.erl
that the previous fix disabled.
I have not found any code where this commit will disable optimizations
when they are actually safe. There are some changes to the code
in ssl_cipher.erl in that some bs_start_match2 instruction did not
reuse the binary register for the match context, but the delayed
sub binary optimizations was never applied to the code in the first
place.
https://bugs.erlang.org/browse/ERL-689
|
|
|
|
|
|
|
|
* ingela/ftp/tftp/app.src-fix:
ftp, tftp: Version should not be hardcoded in <app>.app.src
|
|
|
|
* hans/crypto/fix_bad_merge:
crypto: Fix accidental merge
|
|
Commit 8c47a7657d03777a03a6838c7ec937b6347d07f8 accidently
reverted the changes in lib/crypto/src/crypto.erl correctly
introduced in e8de0736005e91afd992e49f434e08c940eddfa0
This fix re-introduces the correct changes.
|
|
|
|
* ingela/ssl/openssl-test-interop:
ssl: Improve interop checks
|
|
|
|
* maint-19:
Updated OTP version
Update release notes
Update version numbers
erts: Fix memory leak when sending to terminating port
|
|
* ingela/ERL-622/fix:
ssl: Move link to correct process
|
|
The link should be between the connection process and the tls_sender
process. But the start of the tls_sender process needs to be done
by the process that also starts the connection process in order to
correctly create the opaque #ssl_socket{}.
|
|
|
|
|
|
|
|
maint-19
* sverker/erts/19/memory-leak-terminating-port/OTP-14609:
erts: Fix memory leak when sending to terminating port
|
|
Cherry-picked from 7c5fcd3f2701cbb614930682ac52ff75b9c26e6c.
Error: Leak_DefinitelyLost
erts_alloc:230 (-> 0x52E54D) [erl_alloc.h]
port_task_alloc:154 (-> 0x52F3CA) [erl_port_task.c]
erts_port_task_alloc_p2p_sig_data:212 (-> 0x52F5D3) [erl_port_task.c]
erts_port_output:2147 (-> 0x4F6057) [io.c]
erts_port_command:4126 (-> 0x4FA10E) [io.c]
do_send:2200 (-> 0x4E4C64) [bif.c]
erl_send:2494 (-> 0x4E5E09) [bif.c]
process_main:1730 (-> 0x43ADA5) [beam_emu.c]
|
|
* raimo/receive-TOS-TCLASS/ERIERL-187/OTP-15145:
Elaborate the disclaimer for 'pktoptions'
Improve platform filter
Fix endianness bug for CMSG parsing
|
|
|
|
|
|
|
|
* hans/crypto/cuddle_docs:
crypto: Re-work the 'Description' section
crypto: Add warnings in RefMan and User's Guide for experimental RSA opts
|
