summaryrefslogtreecommitdiff
path: root/regress/sys
AgeCommit message (Collapse)Author
2018-04-10Check that a file descriptor's plegde flag is correctly propagated.Alexander Bluhm
The test covers the system calls dup, dup2, dup3, open /dev/fd, and file descriptor passing. The fstat(1) output is analysed.
2018-04-10Add subdir filedescriptor.Alexander Bluhm
2018-04-10Check that a file descriptor's plegde flag is correctly propagated.Alexander Bluhm
The test covers the system calls dup, dup2, dup3, open /dev/fd, and file descriptor passing. The fstat(1) output is analysed.
2018-04-09Add regress test for stack pivot mitigationmortimer
ok @deraadt
2018-02-27Do not expect an eui64 link-local source address during duplicateAlexander Bluhm
address detection. The remote side may have soii, so match ICMP6 packets from any fe80::/64 address.
2018-02-05typosanton
2018-01-23Add a smarter test that opens a file, unlinks it, and remounts theAlexander Bluhm
file system read-only. After unmounting this should result in a dirty file system that is not marked clean. Disable both tests with dangling vnodes for now as they fail.
2018-01-16Enable traceroute test for incoming reply-to rule.Alexander Bluhm
2018-01-15Add tests that unmount a file system while it is busy. On a vndAlexander Bluhm
device "cp -r" and "rm -rf" are working in the background while "umount -f" or "mount -ur" are executed. Then "fsck -f" is used to verify that the file system is clean. Currently this triggers a bug quite reliably. After the mount point has been changed to read-only, the file system check shows that not everything was flushed and synced properly.
2018-01-15Enable traceroute test for incoming route-to rule.Alexander Bluhm
2018-01-14Fix after recent mutex change.Martin Pieuchot
2018-01-11Add tests for incoming route-to and outgoing reply-to rules. DisableAlexander Bluhm
the cases that currently fail.
2017-12-26As the divert-to pf state is not connected, it is not linked to theAlexander Bluhm
incpb and so not removed when the socket is closed. Packets matching reply and reply-to states are linked to the sockets immediately. So these states must go away when the socket is closed. Check this with the output of "pfctl -s states" after running the Perl scripts.
2017-12-21Add some comments what the tests are doing.Alexander Bluhm
2017-12-20Skip this regress test if no remote test machine is specified asAlexander Bluhm
environment variable.
2017-12-18When setting KTRACE in the environment, client and server writeAlexander Bluhm
ktrace files. This simplifies debugging as system calls can be inspected with kdump.
2017-12-16hookup execpromiseFlorian Obser
2017-12-16execpromise regressFlorian Obser
2017-12-15The pledge(2) regress tests fail since the second parameter hasAlexander Bluhm
been converted from paths to execpromises. Remove the test fragments for the non-existing paths feature. OK semarie@ florian@
2017-11-30Add regress to ensure that execve bypasses pledgepath checks (for now)Bob Beck
2017-11-28copyrightBob Beck
2017-11-27decriptor relative openat() and chmodBob Beck
2017-11-27readlink and friendsBob Beck
2017-11-27symlink and a couple more...Bob Beck
2017-11-27access and chflagsBob Beck
2017-11-27Add the start of regression tests for upcoming pledgepath stuff.Bob Beck
still some more to add, and we won't link into regress until later ok deraadt@
2017-11-27lenght->length, mostly in commentsStuart Henderson
2017-11-24Do not run sysctl and pfctl by != during make clean and make obj.Alexander Bluhm
requested by miod@
2017-11-20Skip the test and print a warning if pf does not process packetsAlexander Bluhm
on additional loopback devices.
2017-11-20Now pledge("inet") allows socket option IPV6_V6ONLY. Adapt test.Alexander Bluhm
2017-11-17Let packets jump between routing domains until the ttl is used up.Alexander Bluhm
Use tcpdump to check that the ttl decrements to one.
2017-11-17Link loopback test to build.Alexander Bluhm
2017-11-17Set up two loopback interfaces in different routing domains. TryAlexander Bluhm
to ping existing and non existing addresses in these domains. Also test pinging to different rdomains via pf. Create routing loops and check that the packet vanishes.
2017-11-15Makefile cleanup, avoid DNS lookups, add copyright.Alexander Bluhm
2017-11-15Do not unconfigure vnd1, the test did not configure it.Alexander Bluhm
2017-11-14Add more checks helping to set up the test environment.Alexander Bluhm
2017-11-08Wait for the splicing syscall by grepping it in the relay log. ThisAlexander Bluhm
ensures that scapy's SYN+ACK packet hits the TCP stack when it hurts.
2017-11-07Test more variations of TCP packets in the scapy splicing tests.Alexander Bluhm
Delay connect and SYN+ACK to splice data into a connecting socket.
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.
2017-11-05Fix some comments and dependencies in this regress test.Alexander Bluhm
2017-11-05Insert a rule into pf regress anchor that passes all packets withoutAlexander Bluhm
state. The default pf rules interfere with this scapy test.
2017-10-27Fix the glob to find all scapy test scripts.Alexander Bluhm
2017-10-27I am trying to find a bug that seems to be triggered by a combinationAlexander Bluhm
of TCP timing and socket splicing. Add tests that use Scapy scripts as client and server. That allows to drop, delay and retransmit single TCP packets to execute various code paths in the remote OpenBSD stack.
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.
2017-10-18Move the test implementation into the Makefile. Then we have onlyAlexander Bluhm
one place to look at. Make prints the commands it executes to assist debugging and analysis. If the test fails, there are simple regress targets that can be run to reproduce the error easily.
2017-10-12Use a regex(3) to find the new pledge violation string in the dmesg.Alexander Bluhm
2017-10-09The divert-packet socket option IP_DIVERTFL, IPPROTO_DIVERT_RESP,Alexander Bluhm
IPPROTO_DIVERT_INIT feature is gone. Remove the corresponding tests.
2017-09-16The copyin(9) family of functions should return EFAULT for all memoryPhilip Guenther
protection issues. So make amd64 do that instead of passing through uvm_fault()'s return value, which lead to other possible errors (particularly EACCES) being returned to userspace. Clean up a vestige on arm64 too. ok bluhn@ deraadt@
2017-09-15Clean up pf rules and states on remote machine so that they do notAlexander Bluhm
interfere with later tests.
2017-09-07Copy IPv6 fragment test cases to IPv4 and adapt to the different Alexander Bluhm
behavior. All tests are run once with pf fragment reassembly and once with pf disabled and IP input reassembly. Three sub tests are currently disabled as the IP stack behaves differently than pf.