summaryrefslogtreecommitdiff
path: root/regress/sys/kern
AgeCommit message (Collapse)Author
2013-01-17Improve socket splicing tests:Alexander Bluhm
- One more UDP test. - Enforce that kernel returns EFBIG where appropriate. - Better comments in test files.
2013-01-15The socket splicing tests allow the kernel to return EFBIG in caseAlexander Bluhm
the splicing maximum has been reached.
2013-01-08Add performance test for socket splicing. Two tcpbench processesAlexander Bluhm
are used as server and client. Between them is a relay process that can either copy or splice the data. Benchmarking is done for TCP and UDP and for copy and splice. All tests are executed on localhost per default. Alternatively the relay process can be started on a remote machine. Depending on hardware and environment socket splicing has increased my throughput by 30% or up to 150% in this test.
2013-01-08Autoflush the log file of the child process.Alexander Bluhm
2013-01-03Reengineer the socket splicing regression tests:Alexander Bluhm
- Move the tests from splice to new sosplice directory for consistent naming. - Split the API tests and the TCP splicing tests into separate directories. - Create some tests for the upcoming UDP splicing. - Tests can be run in obj directories now. - The API tests can run both on a local and on a remote machine now. - Fix the forking TCP tests which splice and read or write simultaneously. - Bunch of little fixes for races in the tests. - Deduplicate code, move checks into common functions.
2013-01-07missing argument to printfMiod Vallat
2013-01-05Fix test descriptions.Alexander Bluhm
2013-01-05Add more tests for UDP socket splicing and enhance existing tests.Alexander Bluhm
2013-01-03Reengineer the socket splicing regression tests:Alexander Bluhm
- Move the tests from splice to new sosplice directory for consistent naming. - Split the API tests and the TCP splicing tests into separate directories. - Create some tests for the upcoming UDP splicing. - Tests can be run in obj directories now. - The API tests can run both on a local and on a remote machine now. - Fix the forking TCP tests which splice and read or write simultaneously. - Bunch of little fixes for races in the tests. - Deduplicate code, move checks into common functions.
2013-01-04Fix another race in socket splicing test.Alexander Bluhm
2013-01-03Reengineer the socket splicing regression tests:Alexander Bluhm
- Move the tests from splice to new sosplice directory for consistent naming. - Split the API tests and the TCP splicing tests into separate directories. - Create some tests for the upcoming UDP splicing. - Tests can be run in obj directories now. - The API tests can run both on a local and on a remote machine now. - Fix the forking TCP tests which splice and read or write simultaneously. - Bunch of little fixes for races in the tests. - Deduplicate code, move checks into common functions.
2013-01-03The socket splicing tests have been moved from directory splice toAlexander Bluhm
sosplice.
2013-01-03Add subdir sosplice.Alexander Bluhm
2013-01-03Reengineer the socket splicing regression tests:Alexander Bluhm
- Move the tests from splice to new sosplice directory for consistent naming. - Split the API tests and the TCP splicing tests into separate directories. - Create some tests for the upcoming UDP splicing. - Tests can be run in obj directories now. - The API tests can run both on a local and on a remote machine now. - Fix the forking TCP tests which splice and read or write simultaneously. - Bunch of little fixes for races in the tests. - Deduplicate code, move checks into common functions.
2013-01-02tyop in commentJasper Lievisse Adriaanse
2013-01-02regress test for pty TTYHOG bug; OK miod@Todd C. Miller
2012-07-09This test is not reliable in echo mode.Alexander Bluhm
2012-07-09Modern Perl deferes signal handlers between opcodes. To send TCPAlexander Bluhm
resets at the indended moments into the spliced stream, terminate the client or server process with the alarm default action.
2012-07-09Add a socket splicing test that sends a reset from the client whileAlexander Bluhm
the socket buffers are filled.
2012-07-08Add a test for kevent(EV_DELETE) screwing with POSIX file locksPhilip Guenthe
2012-06-29Fix year in public domain declaration. Also, include <stddef.h> forMatthew Dempsky
POSIX compatibility so the test doesn't depend on <signal.h> defining NULL.
2012-06-29Clear out the sigset_t again before calling sigpending() to make sureMatthew Dempsky
sigpending() actually sets it and it wasn't just still set from calling sigprocmask().
2012-06-29add a regress test for sigpending(); add licensing notice to stackjmp.cMatthew Dempsky
2012-06-23add (currently failing) stackjmp regress testMatthew Dempsky
2012-04-12rfork RIPPhilip Guenthe
2012-03-05If the main thread exits, the process should still be able to getPhilip Guenthe
signals
2011-11-26If sigaltstack() fails, report which thread correctlyPhilip Guenthe
Pointed out by Steffen Daode Nurpmeso (sdaoden at googlemail.com)
2011-11-22Add a test for forked processes inheriting their parent's signal stack.Joshua Elsasser
ok guenther@
2011-11-06Regress tests for negative offsets, wrapping offsets, and the lack ofPhilip Guenthe
interaction of pwrite/pwritev with O_APPEND. Based on a patch from Alexander Polakov (polachok at gmail.com) Further tests with deraadt@
2011-11-05Verify this fails with ESPIPE on ttysPhilip Guenthe
2011-08-29Test that reverse splicing timeout does not affect forward splicing.Alexander Bluhm
2011-08-28Minor cleanup.Alexander Bluhm
2011-08-28Allow splice make regress to run with and without obj directory.Alexander Bluhm
2011-08-21Add one more regression test for socket splicing with idle timeout.Alexander Bluhm
2011-07-23TCP out-of-band data is not reliable. Allow socket splicingAlexander Bluhm
regression tests to tolerate missing urgent bytes.
2011-07-07Add tests for kqueue EVFILT_SIGNAL and that they can't be passed overPhilip Guenthe
sockets
2011-07-07braces rock!Philip Guenthe
2011-07-07Suppress known failuresPhilip Guenthe
2011-07-06Add a regression test for file descriptor passing with UNIX sockets.Matthew Dempsky
ok claudio@
2011-07-06Test handling of non-NUL terminated sun_path values as well as garbagePhilip Guenthe
in the end of the sockaddr_un. Done with claudio@
2011-07-06Switch to a device that isn't deprecated and return more info onPhilip Guenthe
read failure
2011-07-06Add regress for errno returns of failed fchdir()sPhilip Guenthe
2011-07-05Hook unixsock upClaudio Jeker
2011-07-05Regress test for the path length of unix sockets.Claudio Jeker
2011-07-04Add new socket splicing tests for the idle timeout feature. NoteAlexander Bluhm
that you need p5-BSD-Socket-Splice version 0.03 to run the regression tests.
2011-04-22Make the socket splicing regression tests architecture independentAlexander Bluhm
by using the perl module BSD::Socket::Splice from ports. Tested on i386 and sparc64.
2011-03-13Make regression tests more reliable.Alexander Bluhm
2011-03-13Test that reading from a spliced socket does not return eof if thereAlexander Bluhm
is still data in the receive socket buffer.
2011-03-12Add more regression tests for socket splicing. These find theAlexander Bluhm
kernel bugs that have slipped through.
2011-03-08Add regression test that checks ENOTSOCK error for non-socket drain.Alexander Bluhm