[ANNOUNCE] The Linux Test project ltp-20060717 Released

From: Michael Reed
Date: Mon Jul 17 2006 - 18:40:40 EST

The Linux Test Project test suite <http://www.linuxtestproject.org> has been released. The latest version of the testsuite contains 2900+ tests for the Linux OS. Our web site also contains other information such as:
- A Linux test tools matrix
- Technical papers
- How To's on Linux testing
- Code coverage analysis tool.

Release Highlights:

* Code Cleanups by Will Newton, Jacky Malcles, Andy Echols, Amit K. Amora, Mike Frysinger, Mitsuru Chinen, and Liang Shuang

We encourage the community to post results to ltp-results@xxxxxxxxxxxx, and patches, new tests, or comments/questions to ltp-list@xxxxxxxxxxxx

See ChangeLog Below

- The tarball default-tests.tar.gz is a replacement for
testcases/pounder21/default-tests.tar.gz. This new pounder config
enables the magic sysrq key when pounder starts.

- A patch submitted by Derek Wong to reduce the memory requirements of
pounder's ramsnake test.

- A patch submitted by Will Newton that allows for compatibility changes
gcc 2.95.2 in th following files:

lib.c, lib64.c, test.c, test64.c, test_func.c, test_func64.c, tools.c

- Fixed ColdFire Makefile mistake in the syscall and syscalls/mmap directory

- Added a note for uClinux users in the top level Makefile

- A fix for failures in fcntl27 and fcntl28 for bugs 21614 and 23235.

- A fix submitted for make_tree.c by Jacky Malcles that fixes this
testcase by setting envp
- A fix submitted to Jacky Malcles that fixes read_checkzero.c. The lseek
function allows the file offset to be set beyond the end of the existing
end-of-file of the file. If data is later written at this point,
subsequent reads of the data is in the gap returns bytes of zeros until
data is actually written into the gap.

- In the testcase semget05.c the value of MAXIDS was changed for the specific
machine by reading the system limit for SEMMNI - The maximum number of
semaphore sets. This is a fix for bug 24745

- A fix submitted by Amit K. Amora that initializes the alarm received
code and allows the test to pass more than just once on 2.6.17-rc6

- A fix was submitted by Andy Echols for pan.c to fix an infinite
loop problem that occurs in pan if runltp tries to run a test
that isn't present.

- A fix was submitted to cast TEST_RETURN to gid_t to avoid implicit casts
which tend to cause problems with the testcase setregid03.c,

- A patch submitted by Jacky Malcles that fixes the problem where
i0_getevents() return value is not checked and may return 0 if
no events are available and may generate a SIGSEGV in the testcase

- Backed out the _USC_LIB change for several Makefiles because it was
breaking on the PowerPc platform on Fedora Core

- Added code to ignore looking for PID_MAX on powerpc, s390, and i386 to
fix build problems on newer kernel versions on the following files:

sysctl05.c, setpriority04.c sysfs01.c, sysfs02.c, sysfs03.c, sysfs04.c,
sysfs06. getdents01.c, sysctl03.c getsid02.c, sysctl01.c,wait402.c

- TCP.c was changed to delete broken whitespace and also the call for
accept(2) takes a socklen_t, not an int

- Changes were added to the following files to use memset() instead of

tlibio.c, write_log.c, doio.c, iogen.c, fsstress.c, fsx-linux.c,
pthcli.c, pthserv.c, pth_str01.c, pth_str03.c, recvmsg01.c, sendmsg01.c,
crash01.c, crash02.c, pingpong6.c, test_getname.c, fancy_timed_loop.c,
infinite_loop.c, run.c, timed_loop.c, snake.c, rpc1.c, pipeio.c, mc_recv.c

- Changes were added to the following files to use memcpy() instead of
serverCommunication.cpp, member.c, rpc1.c,pipeio.c,mc_recv.c

- A change submitted by Liang Shuang that fixes su01_su for the arm

- A series of patches created by Mitsuru Chinen that created some addtional
network stress tests.
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/