GitHub/libevent
1
0
Fork 0
mirror of https://github.com/libevent/libevent.git synced 2025-01-09 00:56:20 +08:00
Code Issues Packages Projects Releases Wiki Activity
libevent/test/Makefile.nmake

80 lines
2.5 KiB
Makefile
Raw Normal View History

attempt to support OpenSSL in Makefile.nmake
2012-06-08 19:47:23 -07:00
# WATCH OUT! This makefile is a work in progress. -*- makefile -*-
Add nmake files to build with MSVC. Right now, they just make static libraries and unit tests. They probably set lots of options wrong. svn:r1507
2009-11-05 20:40:11 +00:00
attempt to support OpenSSL in Makefile.nmake
2012-06-08 19:47:23 -07:00
!IFDEF OPENSSL_DIR
SSL_CFLAGS=/I$(OPENSSL_DIR)\include /DEVENT__HAVE_OPENSSL
SSL_OBJS=regress_ssl.obj
Whoops. It is gdi.lib, not gdi32.lib. (github issue #61)
2013-05-11 22:21:30 -04:00
SSL_LIBS=..\libevent_openssl.lib $(OPENSSL_DIR)\lib\libeay32.lib $(OPENSSL_DIR)\lib\ssleay32.lib gdi32.lib User32.lib
attempt to support OpenSSL in Makefile.nmake
2012-06-08 19:47:23 -07:00
!ELSE
SSL_CFLAGS=
SSL_OBJS=
SSL_LIBS=
!ENDIF
Fix so that old nmake project still builds.
2014-01-14 17:53:47 +01:00
CFLAGS=/I.. /I../WIN32-Code /I../WIN32-Code/nmake /I../include /I../compat /DHAVE_CONFIG_H /DTINYTEST_LOCAL $(SSL_CFLAGS)
Add nmake files to build with MSVC. Right now, they just make static libraries and unit tests. They probably set lots of options wrong. svn:r1507
2009-11-05 20:40:11 +00:00
CFLAGS=$(CFLAGS) /Ox /W3 /wd4996 /nologo
REGRESS_OBJS=regress.obj regress_buffer.obj regress_http.obj regress_dns.obj \
move dns utility functions into a separate file so that we can use them for http testing
2010-01-14 16:53:25 -08:00
regress_testutils.obj \
Add nmake files to build with MSVC. Right now, they just make static libraries and unit tests. They probably set lots of options wrong. svn:r1507
2009-11-05 20:40:11 +00:00
regress_rpc.obj regress.gen.obj \
regress_et.obj regress_bufferevent.obj \
regress_listener.obj regress_util.obj tinytest.obj \
Rename regress_pthread.c to regress_thread.c
2010-08-17 14:44:12 -04:00
regress_main.obj regress_minheap.obj regress_iocp.obj \
Add regress_finalize to makefile.nmake
2013-05-01 13:37:57 -04:00
regress_thread.obj regress_finalize.obj $(SSL_OBJS)
Add nmake files to build with MSVC. Right now, they just make static libraries and unit tests. They probably set lots of options wrong. svn:r1507
2009-11-05 20:40:11 +00:00
Implemented EV_CLOSED event for epoll backend (EPOLLRDHUP). - Added new EV_CLOSED event - detects premature connection close by clients without the necessity of reading all the pending data. Does not depend on EV_READ and/or EV_WRITE. - Added new EV_FEATURE_EARLY_CLOSED feature for epoll. Must be supported for listening to EV_CLOSED event. - Added new regression test: test-closed.c - All regression tests passed (test/regress and test/test.sh) - strace output of test-closed using EV_CLOSED: socketpair(PF_LOCAL, SOCK_STREAM, 0, [6, 7]) = 0 sendto(6, "test string\0", 12, 0, NULL, 0) = 12 shutdown(6, SHUT_WR) = 0 epoll_ctl(3, EPOLL_CTL_ADD, 7, {EPOLLRDHUP, {u32=7, u64=7}}) = 0 epoll_wait(3, {{EPOLLRDHUP, {u32=7, u64=7}}}, 32, 3000) = 1 epoll_ctl(3, EPOLL_CTL_MOD, 7, {EPOLLRDHUP, {u32=7, u64=7}}) = 0 fstat(1, {st_mode=S_IFCHR|0620, st_rdev=makedev(136, 4), ...}) mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYM... write(1, "closed_cb: detected connection close "..., 45) = 45
2014-01-17 23:20:42 -02:00
OTHER_OBJS=test-init.obj test-eof.obj test-closed.obj test-weof.obj test-time.obj \
Possible fix to 100% cpu usage with epoll and openssl I'm running a fairly simple bit of test code using libevent2 with epoll and openssl bufferevents and I've run into a 100% cpu usage problem. Looking into it 100% usage was caused by epoll_wait constantly returning write events on the openssl socket when it shouldn't really have been looking for write events at all (N_ACTIVE_CALLBACKS() was returning 0 also). Looking a bit deeper eventbuffer_openssl socket seems to be requesting that the EV_WRITE event be removed when it should, but the event isn't actually being removed from epoll. Continuing to follow this I think I've found a bug in event_changelist_del. For evpoll event_del calls event_changelist_del which caches the change which is then actioned later when evpoll_dispatch is called. In event_changlist_del there is a check so that if the currently changed action is an add then the cached action is changed to a no-op rather than a delete (which makes sense). The problem arises if there are more than two add or delete operations between calls to dispatch, in this case it's possible that the delete is turned into a no-op when it shouldn't have been. For example starting with the event on, a delete followed by an add and then another delete results in a no-op when it should have been a delete (I added a fair bit of debug output that seems to confirm this behaviour). I've applied a small change that checks the original old_event stored with the change and only converts the delete to a no-op if the event isn't on in old_event. This seems to have fixed my problem.
2010-07-19 13:44:56 +12:00
bench.obj bench_cascade.obj bench_http.obj bench_httpclient.obj \
a program to print out the error strings for winsock errors
2013-02-07 17:06:49 -08:00
test-changelist.obj \
print-winsock-errors.obj
Add nmake files to build with MSVC. Right now, they just make static libraries and unit tests. They probably set lots of options wrong. svn:r1507
2009-11-05 20:40:11 +00:00
PROGRAMS=regress.exe \
Implemented EV_CLOSED event for epoll backend (EPOLLRDHUP). - Added new EV_CLOSED event - detects premature connection close by clients without the necessity of reading all the pending data. Does not depend on EV_READ and/or EV_WRITE. - Added new EV_FEATURE_EARLY_CLOSED feature for epoll. Must be supported for listening to EV_CLOSED event. - Added new regression test: test-closed.c - All regression tests passed (test/regress and test/test.sh) - strace output of test-closed using EV_CLOSED: socketpair(PF_LOCAL, SOCK_STREAM, 0, [6, 7]) = 0 sendto(6, "test string\0", 12, 0, NULL, 0) = 12 shutdown(6, SHUT_WR) = 0 epoll_ctl(3, EPOLL_CTL_ADD, 7, {EPOLLRDHUP, {u32=7, u64=7}}) = 0 epoll_wait(3, {{EPOLLRDHUP, {u32=7, u64=7}}}, 32, 3000) = 1 epoll_ctl(3, EPOLL_CTL_MOD, 7, {EPOLLRDHUP, {u32=7, u64=7}}) = 0 fstat(1, {st_mode=S_IFCHR|0620, st_rdev=makedev(136, 4), ...}) mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYM... write(1, "closed_cb: detected connection close "..., 45) = 45
2014-01-17 23:20:42 -02:00
test-init.exe test-eof.exe test-closed.exe test-weof.exe test-time.exe \
a program to print out the error strings for winsock errors
2013-02-07 17:06:49 -08:00
test-changelist.exe \
print-winsock-errors.exe
Add nmake files to build with MSVC. Right now, they just make static libraries and unit tests. They probably set lots of options wrong. svn:r1507
2009-11-05 20:40:11 +00:00
# Disabled for now:
# bench.exe bench_cascade.exe bench_http.exe bench_httpclient.exe
Fix windows and msvc build
2010-01-23 20:07:05 -05:00
LIBS=..\libevent.lib ws2_32.lib shell32.lib advapi32.lib
Add nmake files to build with MSVC. Right now, they just make static libraries and unit tests. They probably set lots of options wrong. svn:r1507
2009-11-05 20:40:11 +00:00
all: $(PROGRAMS)
regress.exe: $(REGRESS_OBJS)
attempt to support OpenSSL in Makefile.nmake
2012-06-08 19:47:23 -07:00
$(CC) $(CFLAGS) $(LIBS) $(SSL_LIBS) $(REGRESS_OBJS)
Add nmake files to build with MSVC. Right now, they just make static libraries and unit tests. They probably set lots of options wrong. svn:r1507
2009-11-05 20:40:11 +00:00
test-init.exe: test-init.obj
$(CC) $(CFLAGS) $(LIBS) test-init.obj
test-eof.exe: test-eof.obj
$(CC) $(CFLAGS) $(LIBS) test-eof.obj
Implemented EV_CLOSED event for epoll backend (EPOLLRDHUP). - Added new EV_CLOSED event - detects premature connection close by clients without the necessity of reading all the pending data. Does not depend on EV_READ and/or EV_WRITE. - Added new EV_FEATURE_EARLY_CLOSED feature for epoll. Must be supported for listening to EV_CLOSED event. - Added new regression test: test-closed.c - All regression tests passed (test/regress and test/test.sh) - strace output of test-closed using EV_CLOSED: socketpair(PF_LOCAL, SOCK_STREAM, 0, [6, 7]) = 0 sendto(6, "test string\0", 12, 0, NULL, 0) = 12 shutdown(6, SHUT_WR) = 0 epoll_ctl(3, EPOLL_CTL_ADD, 7, {EPOLLRDHUP, {u32=7, u64=7}}) = 0 epoll_wait(3, {{EPOLLRDHUP, {u32=7, u64=7}}}, 32, 3000) = 1 epoll_ctl(3, EPOLL_CTL_MOD, 7, {EPOLLRDHUP, {u32=7, u64=7}}) = 0 fstat(1, {st_mode=S_IFCHR|0620, st_rdev=makedev(136, 4), ...}) mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYM... write(1, "closed_cb: detected connection close "..., 45) = 45
2014-01-17 23:20:42 -02:00
test-closed.exe: test-closed.obj
$(CC) $(CFLAGS) $(LIBS) test-closed.obj
Possible fix to 100% cpu usage with epoll and openssl I'm running a fairly simple bit of test code using libevent2 with epoll and openssl bufferevents and I've run into a 100% cpu usage problem. Looking into it 100% usage was caused by epoll_wait constantly returning write events on the openssl socket when it shouldn't really have been looking for write events at all (N_ACTIVE_CALLBACKS() was returning 0 also). Looking a bit deeper eventbuffer_openssl socket seems to be requesting that the EV_WRITE event be removed when it should, but the event isn't actually being removed from epoll. Continuing to follow this I think I've found a bug in event_changelist_del. For evpoll event_del calls event_changelist_del which caches the change which is then actioned later when evpoll_dispatch is called. In event_changlist_del there is a check so that if the currently changed action is an add then the cached action is changed to a no-op rather than a delete (which makes sense). The problem arises if there are more than two add or delete operations between calls to dispatch, in this case it's possible that the delete is turned into a no-op when it shouldn't have been. For example starting with the event on, a delete followed by an add and then another delete results in a no-op when it should have been a delete (I added a fair bit of debug output that seems to confirm this behaviour). I've applied a small change that checks the original old_event stored with the change and only converts the delete to a no-op if the event isn't on in old_event. This seems to have fixed my problem.
2010-07-19 13:44:56 +12:00
test-changelist.exe: test-changelist.obj
$(CC) $(CFLAGS) $(LIBS) test-changelist.obj
Add nmake files to build with MSVC. Right now, they just make static libraries and unit tests. They probably set lots of options wrong. svn:r1507
2009-11-05 20:40:11 +00:00
test-weof.exe: test-weof.obj
$(CC) $(CFLAGS) $(LIBS) test-weof.obj
test-time.exe: test-time.obj
$(CC) $(CFLAGS) $(LIBS) test-time.obj
a program to print out the error strings for winsock errors
2013-02-07 17:06:49 -08:00
print-winsock-errors.exe: print-winsock-errors.obj
$(CC) $(CFLAGS) $(LIBS) print-winsock-errors.obj
Add nmake files to build with MSVC. Right now, they just make static libraries and unit tests. They probably set lots of options wrong. svn:r1507
2009-11-05 20:40:11 +00:00
bench.exe: bench.obj
$(CC) $(CFLAGS) $(LIBS) bench.obj
bench_cascade.exe: bench_cascade.obj
$(CC) $(CFLAGS) $(LIBS) bench_cascade.obj
bench_http.exe: bench_http.obj
$(CC) $(CFLAGS) $(LIBS) bench_http.obj
bench_httpclient.exe: bench_httpclient.obj
$(CC) $(CFLAGS) $(LIBS) bench_httpclient.obj
Don't break when building tests from git without python installed
2011-06-14 14:37:49 -04:00
regress.gen.c regress.gen.h: regress.rpc ../event_rpcgen.py
echo // > regress.gen.c
echo #define NO_PYTHON_EXISTS > regress.gen.h
-python ..\event_rpcgen.py regress.rpc
Add nmake files to build with MSVC. Right now, they just make static libraries and unit tests. They probably set lots of options wrong. svn:r1507
2009-11-05 20:40:11 +00:00
clean:
-del $(REGRESS_OBJS)
-del $(OTHER_OBJS)
remove all exes on "make clean", not just regress.exe
2013-02-07 17:14:18 -08:00
-del $(PROGRAMS)
Reference in New Issue Copy Permalink