Azat Khuzhin [Sun, 5 Jul 2020 08:45:56 +0000 (11:45 +0300)]
test: fix UB in evbuffer/empty_reference_prepend_buffer
UBSAN reports:
test/regress_buffer.c:2360:2: runtime error: null pointer passed as argument 1, which is declared to never be null
/usr/include/string.h:140:33: note: nonnull attribute specified here
Azat Khuzhin [Thu, 25 Jun 2020 18:01:53 +0000 (21:01 +0300)]
test: detect test failures if atexit handler calls _exit(!0) (sanitizers)
tinytest uses another way of detecting test failures, it uses pipe
between child and parent, and if the test function in child returns OK
it writes OK flag into pipe, and reads it in parent.
However sanitizers uses atexit handlers to detect leaks, and this will
not detect failures in case of exit() will be called from the atexit
handlers, fix this by checking status after waitpid().
Enji Cooper [Thu, 25 Jun 2020 15:08:10 +0000 (08:08 -0700)]
Make all classes Entry, Struct, etc) new-style classes
This allows the object methods to be properly inherited and called via
`super(..)`, addressing breakage with python 2.x, introduced in cfcc093606e747a5d250787012bac0b149c60d6d.
Fixes #1042.
Signed-off-by: Enji Cooper <yaneurabeya@gmail.com>
Azat Khuzhin [Thu, 25 Jun 2020 07:18:43 +0000 (10:18 +0300)]
evdns: do not pass NULL to memcpy() in evdns_server_request_format_response()
In case of OPT pseudo-RR `class` field is treated as a requestor's UDP
payload size, and class will have 512 (DNS_MAX_UDP_SIZE), and data is
NULL:
(gdb) p *item
$4 = {
next = 0x0,
name = 0x602000000130 "",
type = 41,
class = 512,
ttl = 0,
is_name = 0 '\000',
datalen = 0,
data = 0x0
}
And UBSAN will reports:
../evdns.c:2493:5: runtime error: null pointer passed as argument 2, which is declared to never be null
#0 0x7ffff70b65bb in evdns_server_request_format_response ../evdns.c:2493
#1 0x7ffff70b706b in evdns_server_request_respond ../evdns.c:2529
#2 0x5555557975ab in regress_dns_server_cb ../test/regress_testutils.c:263
#3 0x7ffff70a8489 in request_parse ../evdns.c:1576
#4 0x7ffff70aa445 in server_udp_port_read ../evdns.c:1726
#5 0x7ffff70ac5cc in server_port_ready_callback ../evdns.c:1849
#6 0x7ffff6d3054c in event_persist_closure ../event.c:1645
#7 0x7ffff6d311cd in event_process_active_single_queue ../event.c:1704
#8 0x7ffff6d33258 in event_process_active ../event.c:1805
#9 0x7ffff6d361b5 in event_base_loop ../event.c:2047
#10 0x7ffff6d334ac in event_base_dispatch ../event.c:1839
#11 0x555555739df2 in test_edns ../test/regress_dns.c:2639
#12 0x5555557b9e96 in testcase_run_bare_ ../test/tinytest.c:173
#13 0x5555557ba8f0 in testcase_run_one ../test/tinytest.c:333
#14 0x5555557bc9a0 in tinytest_main ../test/tinytest.c:527
#15 0x555555787faa in main ../test/regress_main.c:528
#16 0x7ffff606c001 in __libc_start_main (/usr/lib/libc.so.6+0x27001)
#17 0x55555569436d in _start (/src/le/libevent/.cmake-debug/bin/regress+0x14036d)
Azat Khuzhin [Thu, 25 Jun 2020 07:15:01 +0000 (10:15 +0300)]
buffer: do not pass NULL to memcpy() from evbuffer_pullup()
UBSAN reports:
evbuffer/remove_buffer_with_empty3: ../buffer.c:1443:3: runtime error: null pointer passed as argument 2, which is declared to never be null
#0 0x7ffff6cd0410 in evbuffer_pullup ../buffer.c:1443
#1 0x5555556d68b9 in test_evbuffer_remove_buffer_with_empty3 ../test/regress_buffer.c:408
#2 0x5555557b95ee in testcase_run_bare_ ../test/tinytest.c:173
#3 0x5555557ba048 in testcase_run_one ../test/tinytest.c:333
#4 0x5555557bc0f8 in tinytest_main ../test/tinytest.c:527
#5 0x555555787702 in main ../test/regress_main.c:528
#6 0x7ffff606c001 in __libc_start_main (/usr/lib/libc.so.6+0x27001)
#7 0x55555569436d in _start (/src/le/libevent/.cmake-debug/bin/regress+0x14036d)
Azat Khuzhin [Thu, 25 Jun 2020 05:45:34 +0000 (08:45 +0300)]
http: fix undefined-shift in EVUTIL_IS*_ helpers
evutil.c:2559:1: runtime error: left shift of 1 by 31 places cannot be represented in type 'int'
#0 0x4f2be0 in EVUTIL_ISXDIGIT_ libevent/evutil.c:2559:1
#1 0x4bd689 in regname_ok libevent/http.c:4838:7
#2 0x4bc16b in parse_authority libevent/http.c:4958:9
#3 0x4bb8b5 in evhttp_uri_parse_with_flags libevent/http.c:5103:7
#4 0x4bb762 in evhttp_uri_parse libevent/http.c:5050:9
#5 0x4b8f41 in evhttp_parse_query_impl libevent/http.c:3505:9
#6 0x4b8ed7 in evhttp_parse_query libevent/http.c:3569:9
fanquake [Mon, 22 Jun 2020 11:40:06 +0000 (19:40 +0800)]
wepoll: use lower-cased windows headers
This matches their use throughout the rest of the project, and fixes the
build when you are cross-compiling on Linux using mingw-w64. i.e:
CC epoll.lo
CC wepoll.lo
CC signal.lo
wepoll.c:138:10: fatal error: WS2tcpip.h: No such file or directory
#include <WS2tcpip.h>
^~~~~~~~~~~~
compilation terminated.
make[2]: *** [Makefile:1918: wepoll.lo] Error 1
Paul Osborne [Thu, 18 Jun 2020 15:47:14 +0000 (15:47 +0000)]
cmake: avoid problems from use of CMAKE_USE_PTHREADS_INIT
In some CMake integrations, portions of the toolchain may end up
defining CMAKE_USE_PTHREADS_INIT even when EVENT__DISABLE_THREAD_SUPPORT
is set for libevent. Modify the build to not rely on this side effect
of find_package(Threads, ...) [which could be done elsewhere] but instead
to use EVENT__HAVE_PTHREADS which is defined only on the desired codepath.
Without this change, affected builds fail as a result of event_pthreads source
files being built but with build defines which cause the build to fail.
Signed-off-by: Paul Osborne <paul.osborne@smartthings.com>
seleznevae [Sun, 31 May 2020 16:59:49 +0000 (19:59 +0300)]
evdns: Add support for setting maximum UDP DNS message size.
Added new option `edns-udp-size` for evdns_base which allows
to control maximum allowed size of UDP DNS messages. This
maximum size is passed to the DNS server via edns mechanism.
Azat Khuzhin [Mon, 25 May 2020 08:25:18 +0000 (11:25 +0300)]
http: Merge branch 'http-max_connections-pr-592'
@jcoffland:
"When the max connection limit is enabled and the limit is reached, the
server will respond immediately with 503 Service Unavailable. This can
be used to prevent servers from running out of file descriptors. This is
better than request limiting because clients may make more than one
request over a single connection. Blocking a request does not
necessarily close the connection and free up a socket."
* http-max_connections-pr-592:
test: cover evhttp max connections
Added evhttp max simultaneous connection limiting
Azat Khuzhin [Mon, 25 May 2020 00:13:00 +0000 (03:13 +0300)]
Merge branch 'evdns-tcp-pr-1004'
@seleznevae:
"Added support for DNS requests via TCP. By default, requests are done
via UDP. In case truncated response is received new attempt is done
via TCP connection. Added 2 new macros DNS_QUERY_USEVC and
DNS_QUERY_IGNTC to force all requests to be done via TCP and to disable
switch to TCP in case of truncated responses.
Also added possibility for DNS server to listen and receive requests on
TCP port. Current implementation of TCP support in DNS server seems
rather preliminary and maybe changes after discussion and code review.
Fallback to TCP in case of truncated DNS requests is done automatically.
To imitate the old behaviour macros DNS_QUERY_IGNTC should be used. To
force all DNS requests to be done via TCP one should use the flag
DNS_QUERY_USEVC. Names DNS_QUERY_IGNTC, DNS_QUERY_USEVC were chosen to
imitate similar flags in c-ares and glibc."
Ok, interfaces looks good, merging to avoid stalling it for too long.
* evdns-tcp-pr-1004:
evdns: fix coding style issues
evdns: fix trailing whitespaces
evdns: bufferevent_setcb before bufferevent_free is redundant
evdns: Implement dns requests via tcp
Nick Grifka [Thu, 23 Apr 2020 02:44:45 +0000 (19:44 -0700)]
Add wepoll support to light up the epoll backend on Windows
libevent is lacking a scalable backend on Windows. Let's leverage the wepoll
library until Windows comes up with an epoll/kqueue compete user mode API.
- All regress tests pass for standard wepoll
- These 2 tests fail intermittently for changelist wepoll, so disabling
changelist wepoll for now
http/cancel_inactive_server
http/stream_in
- verify target on Windows runs tests for both wepoll and win32 backends
- wepoll backend preferred over win32 backend
- wepoll version 1.5.6
v2: cleaner backend abstraction. Disallow wepoll on MinGW/Cygwin.
v3: Add wepoll.h to dist
v4: Make sure wepoll source files are excluded from cygwin/mingw builds
v5: Keep win32 as default backend on windows.
v6: Include wepoll in mingw builds. Verified that regress tests pass w/ WEPOLL backend.
v7: Enable wepoll on mingw when building with cmake
v8: Add wepoll testrunner for autotools test target
v2 (by azat):
- drop package installations
- use local .abi-check over $HOME/abi-check for build dir
- drop regex check, simply use default values (too complex otherwise)
- use sub-shell to avoid cd back
- add missing quotes
- make the style uniq across the whole file (no tabs for indent, copy-paste?)
- drop `set -x`, use `bash -x abi-check.sh` over
- drop EVENT_ABI_CHECK
- use /usr/bin/env bash as shebang
- use `find | xargs` over `cp $(grep -v)`
- adjust markdown syntax in abi-check/README.md
- adjust link to the publicly available documentation
v3 (by azat):
- docker image
- git check-ignore
- make -j8
- allow to change defaults
Azat Khuzhin [Tue, 5 May 2020 11:57:59 +0000 (14:57 +0300)]
ci: change cache key for windows/mingw
Should fix the following current failures:
- 2020-05-04T22:32:02.9490248Z C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\MSBuild\Microsoft\VC\v160\Microsoft.CppCommon.targets(231,5): error MSB6006: "cmd.exe" exited with code 3. [D:\a\libevent\libevent\build\regress.vcxproj]
- cmake : /usr/bin/sh: /C/hostedtoolcache/windows/Python/3.7.6/x64/python.exe: No such file or directory
P.S. I guess python has another path, but on my fork it is the same and
it passes.
Azat Khuzhin [Mon, 4 May 2020 21:21:18 +0000 (00:21 +0300)]
Fix EV_CLOSED detection/reporting (epoll only)
- EV_CLOSED is EPOLLRDHUP in epoll
- EPOLLRDHUP reported w/o EPOLLHUP if the socket closed with shutdown(SHUT_WR)
- EPOLLRDHUP reported w/ EPOLLHUP if the socket closed with close()
so in this case epoll backend will detect this event as error
(EV_READ|EV_WRITE), since the epoll_ctl() will return EPOLLRDHUP with
EPOLLHUP set, but this is not correct, let's fix this.
Azat Khuzhin [Mon, 4 May 2020 21:05:49 +0000 (00:05 +0300)]
Merge branch 'EV_CLOSED-and-EV_ET-fixes'
* EV_CLOSED-and-EV_ET-fixes:
Avoid triggering wrong events with EV_ET set
epoll: handle EV_ET for EV_CLOSED too
test: cover EV_CLOSED with lots of possible scenarious
test: rename simpleclose to simpleclose_rw (since it works via write/read)
Azat Khuzhin [Wed, 8 Apr 2020 23:16:15 +0000 (02:16 +0300)]
Avoid triggering wrong events with EV_ET set
For the event at least something except EV_ET should be set, so checking
ev->ev_events with "triggered" events is wrong, because EV_ET is always
passed (see epoll), since it will be filtered out if it is not set in
event.
Azat Khuzhin [Sat, 28 Mar 2020 13:02:39 +0000 (16:02 +0300)]
Merge branch 'event_rpcgen.py-cleanup'
* event_rpcgen.py-cleanup:
event_rpcgen.py: fix arguments-differ
event_rpcgen.py: fix attribute-defined-outside-init
event_rpcgen: suppress some warnings to make pylint clean
Don't accumulate arguments in `Entry.GetTranslation`
Fix improper string concatenations in lists
Fix warnings regarding unused variables
Don't override the `type` built-in
Call `super` to call methods from the parent class
Address `no-self-use` issues reported by pylint
Run the code through the black formatter
Reformat strings to template
Add `argparse` support
Precompile regular expressions
Use bools instead of ints values where possible
Rename all global variables to match the PEP8 spec
Handle file pointers with context suite patterns
Iterate over `tokens` with a for instead of while
Fix indentation for `RpcGenError`
Don't override `file` built-in
Resolve variable name issues per PEP8
Sort imports per PEP8
Enji Cooper [Fri, 27 Mar 2020 23:27:56 +0000 (16:27 -0700)]
Call `super` to call methods from the parent class
Not only does it make it clearer to the reader what the intent is behind
the code, it fixes MRO evaluation of classes if the parent-child
relationship is nonlinear.
Signed-off-by: Enji Cooper <yaneurabeya@gmail.com>
Enji Cooper [Fri, 27 Mar 2020 23:20:24 +0000 (16:20 -0700)]
Address `no-self-use` issues reported by pylint
Sprinkle around `@staticmethod` where need be and raise
`NotImplementedError` from `Entry.GetInitializer` instead of doing
similar with `assert`. `NotImplementedError` will provide a hint to
static analyzers that there are abstract methods which need to be
implemented, if they aren't, in subclasses.
Signed-off-by: Enji Cooper <yaneurabeya@gmail.com>
Enji Cooper [Fri, 27 Mar 2020 23:04:37 +0000 (16:04 -0700)]
Run the code through the black formatter
This resolves a number of flake8 issues as well as pylint issues. Not
all of the changes mesh between flake8/pylint though, so there are some
false positives, but it's better than before.
Signed-off-by: Enji Cooper <yaneurabeya@gmail.com>