Git Log


Commit Date  
[a8e24b] (HEADmaster) by Cyril Hrubis Cyril Hrubis

Remove repo, add README pointing to GitHub repo

See https://github.com/linux-test-project/ltp

Signed-off-by: Cyril Hrubis <metan@ucw.cz>

2013-12-11 14:16:08 Tree
[1eb440] by Cyril Hrubis Cyril Hrubis

Fix 743f2fcd5744744d649496d8f0524cf5ca7d3859.

The gethost is called from multicast testcases too, which I failed to
find, apologies everyone.

This commit moves the gethost to muticast directory and renames it to
mc_gethost and fixes all usages accordingly.

Signed-off-by: Cyril Hrubis <chrubis@suse.cz>

2013-11-14 15:27:04 Tree
[b185f2] by Jan Stancek Jan Stancek , pushed by Cyril Hrubis Cyril Hrubis

add test for tst_checkpoint_signal_child timing out

This is simple test for patch which adds timeout to
tst_checkpoint_signal_child().

The test is expected to end with:

Failed to open fifo 'tst_checkpoint_fifo' at
tst_checkpoint_no_child.c:43: errno=ETIMEDOUT(110): Connection timed out

Signed-off-by: Jan Stancek <jstancek@redhat.com>
Signed-off-by: Cyril Hrubis <chrubis@suse.cz>

2013-11-14 14:47:33 Tree
[cb80f4] by Stanislav Kholmanskikh Stanislav Kholmanskikh , pushed by Cyril Hrubis Cyril Hrubis

tst_checkpoint_signal_child: implemented timeout

If a child exits before opening TST_CHECKPOINT_FIFO
for reading, tst_checkpoint_signal_child() issued from the parent
will block forever.

To handle such situations added timeout logic to tst_checkpoint_signal_child();

Signed-off-by: Stanislav Kholmanskikh <stanislav.kholmanskikh@oracle.com>
Reviewed-by: Jan Stancek <jstancek@redhat.com>
Acked-by: Cyril Hrubis <chrubis@suse.cz>

2013-11-14 14:22:24 Tree
[7b2002] by Cyril Hrubis Cyril Hrubis

lib: tst_res: Add more errnos.

Add all known Linux errnos.

Signed-off-by: Cyril Hrubis <chrubis@suse.cz>

2013-11-14 14:07:22 Tree
[9b4860] by Siva Padhy Siva Padhy , pushed by Cyril Hrubis Cyril Hrubis

BindMounts: fix to properly unmount during cleanup

The fs_bind testcases creates the mount directories for testing.
When the testcase exits, those directories were not properly unmounted.
The reason of the issues are:

1. Some directories were kept mounted, even after the cleanup() routine, as the unmount code is missing.
2. Some "umount" command execution fails, as they are dependent on some other mounted directories.

Error Report
------------
rbind/test02 1 TPASS : rbind/test02: PASSED
/kvt.new/ltp/testscripts/test_fs_bind.sh 0 TWARN : rbind/test02 did not properly clean up "/tmp/ltp-XXXX9qdYqG/sandbox"

rbind/test38 1 TPASS : rbind/test38: PASSED
/kvt.new/ltp/testscripts/test_fs_bind.sh 0 TWARN : rbind/test38 did not properly clean up "/tmp/ltp-XXXX9qdYqG/sandbox"

This patch add/reorder the unmount code to fix the above TWARN messages.

Modified-by: Jan Stancek <jstancek@redhat.com>
Signed-off-by: Siva <siva.padhy@lge.com>
Tested-by: Jan Stancek <jstancek@redhat.com>

2013-11-13 16:37:39 Tree
[119cf4] by Stanislav Kholmanskikh Stanislav Kholmanskikh , pushed by Cyril Hrubis Cyril Hrubis

msgctl{08, 09, 10, 11}: cleanup

* Moved common functions, definitions to libmsgctl.h, libmsgctl.c.
* Changed tst_* in the children with printf + exit (or return).
* Some cleanup.

Signed-off-by: Stanislav Kholmanskikh <stanislav.kholmanskikh@oracle.com>

2013-11-13 15:28:26 Tree
[b6e703] by Stanislav Kholmanskikh Stanislav Kholmanskikh , pushed by Cyril Hrubis Cyril Hrubis

syscalls/ipc/Makefile.inc: changed LIB to LIBIPC

Both lib.mk and testcases/kernel/syscalls/ipc/Makefile.inc set
LIB. It confuses make when both failes are included into a leaf Makefile.

Renamed LIB to LIBIPC in Makefile.inc.

The idea is by Cyril Hrubis <chrubis@suse.cz>

Signed-off-by: Stanislav Kholmanskikh <stanislav.kholmanskikh@oracle.com>

2013-11-13 13:55:07 Tree
[09bd93] by Stanislav Kholmanskikh Stanislav Kholmanskikh , pushed by Cyril Hrubis Cyril Hrubis

setregid02: invalid user value fix

Commit 705926d096f7a64c9d53efeb789288e25457d68f
introduced a bug that the value for inval_user was treated
as '-1' when passed to 16-bit version of setregid().

Fixed it the same way as it's done for setreuid test cases.

Reported-by: Honggyu Kim <hong.gyu.kim@lge.com>
Signed-off-by: Stanislav Kholmanskikh <stanislav.kholmanskikh@oracle.com>

2013-11-13 12:45:45 Tree
[743f2f] by Simon Xu Simon Xu , pushed by Cyril Hrubis Cyril Hrubis

move gethost from tools to sendfile01

Move tools/gethost.c to
testcases/network/tcp_cmds/sendfile/sendfile01_gethost.c
since sendfile01 is the only test that requires gethost.

Signed-off-by: Simon Xu <xu.simon@oracle.com>
Acked-by: Cyril Hrubis <chrubis@suse.cz>

2013-11-12 12:36:58 Tree
[a4f5e2] by Simon Xu Simon Xu , pushed by Cyril Hrubis Cyril Hrubis

xinetd_tests.sh: remove RECORD and eliminate "|| RC=$?"

1) In xinetd.conf, `RECORD' is not a valid value for entry `log_on_failure'
(I couldn't find a distro in which it's valid), and it makes xinetd
successfully starts and promptly exit with error message:

"Bad log_on_failure flag: RECORD [file=/etc/xinetd.conf] [line=6]",

but the test will pass because telnet service is not started and that's
what it expect.

2) Eliminate '|| RC=$?' because because it may not be excecuted and the
original value in RC can mess things up. Also remove the RC=0
initializations that are now unnecessary.

Signed-off-by: Simon Xu <xu.simon@oracle.com>

2013-11-11 17:05:33 Tree
[4716af] by Simon Xu Simon Xu , pushed by Cyril Hrubis Cyril Hrubis

sendfile01: rename SF_Server to sendfile01_server and add it into Makefile

1) Rename SF_Server to sendfile01_server to avoid collisions
2) The test always fails because SF_Server is not installed, add
sendfile01_server into sendfile01 makefile to install it.

Signed-off-by: Simon Xu <xu.simon@oracle.com>

2013-11-11 17:02:04 Tree
[bfd986] by Simon Xu Simon Xu , pushed by Cyril Hrubis Cyril Hrubis

mc_cmds: Multicast ping being disabled should not fail the test

Signed-off-by: Simon Xu <xu.simon@oracle.com>

2013-11-11 16:49:21 Tree
[021494] by Simon Xu Simon Xu , pushed by Cyril Hrubis Cyril Hrubis

rdist01: fix mistakes

Fix the following mistakes in the script:
1) the way it gets dirname and basename of files is broken
2) ${FILES} and ${HOSTS} get substituted when generating distfile
3) the way it calculates sum of files on RHOST is broken
4) there is no $DIRECTORIES variable

Signed-off-by: Simon Xu <xu.simon@oracle.com>

2013-11-11 16:46:50 Tree
[87f593] by Simon Xu Simon Xu , pushed by Cyril Hrubis Cyril Hrubis

tcpdump01: fix the way to identify network interface

The script tries to identify network interface that can communicate with RHOST
by searching for the IP of RHOST in ifconfig output on localhost, which doesn't
make any sense when RHOST is not localhost. This patch changes the script to
indetify the network interface by ping RHOST.

Signed-off-by: Simon Xu <xu.simon@oracle.com>

2013-11-07 14:02:16 Tree
[b87a83] by Cyril Hrubis Cyril Hrubis

fcntl16: Fix incorrect flock strucure

The fcntl16 testcase is being compiled twice, once normal, and the other
time with the _FILE_OFFSET_BITS=64 define set.

When compiled with 64bit file functions (as fcntl16_64) all file
functions will automatically use their 64bit variants, e.g. struct
flock64" instead of "struct flock".

So, when in the code e.g.
fcntl(fd, F_SETLK, thislock)
is called, this will automatically be translated into
fcntl64(fd, F_SETLK64, thislock)
and "thislock" needs to be of type "struct flock64".

We need to make sure that thislock is of the correct type, with the
start and len fields being of type "off_t" which then will be translated
into "loff_t" during compilation when _FILE_OFFSET_BITS=64 is set.

In orded to do that we make use of struct flock instead of the incorrect
version that was defined in the test source.

This fixes the fcntl16_64 testcase on hppa64-linux-gnu with 32bit
userspace running on 64bit kernel.

Signed-off-by: Helge Deller <deller@gmx.de>
Signed-off-by: Cyril Hrubis <chrubis@suse.cz>

2013-11-07 11:11:26 Tree
[d97f51] by Jeff Layton Jeff Layton , pushed by Jan Stancek Jan Stancek

diotest3: fix -f argument

The '-f' argument to diotest3 is currently ignored. During options
processing it's copied to the buffer, then subsequently overwritten
by the default in a later sprintf() call.

Move the setting of the default prior to the options processing.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Jan Stancek <jstancek@redhat.com>

2013-11-06 21:27:07 Tree
[840aa5] by Helge Deller Helge Deller , pushed by Jan Stancek Jan Stancek

sendfile08: fix buf comparison

The sendfile08 testcase uses the sendfile() syscall
to write the strings "hello" and "world" to a file descriptor
and then reads them back in through another file descriptor.

The length of the strings written are determined by strlen(),
so that in total 10 bytes are written.

When reading the strings back in, again 10 bytes are read, but
the problem is, that those are read back into a *local* string
buffer array (buf[]) which hasn't been initialized.

Afterwards,
strcmp(buf, "helloworld")
is called to compare input and output.
This strcmp() is wrong, because basically "helloworld\0"
(with the trailing string zero terminator) is compared to
"helloworld" (likely without a trailing zero in the buf[] array).

So, if buf[] hasn't been initialized, it's not guaranteed
that the byte in buf[10] equals to '\0'.

This problem was found on hppa64-linux-gnu architecture, most
likely, because it's a stack-grows-up architecture.

Fix this by using strncmp().

Signed-off-by: Helge Deller <deller@gmx.de>
Signed-off-by: Jan Stancek <jstancek@redhat.com>

2013-11-06 21:10:08 Tree
[e533f0] by Stanislav Kholmanskikh Stanislav Kholmanskikh , pushed by Cyril Hrubis Cyril Hrubis

fsync02: SIGFPE fix

It may happen that stat_buf.f_frsize is greater than
BLOCKSIZE (for example, if we use "non-default" FS block sizes or NFS) and
because on (BLOCKSIZE / stat_buf.f_frsize) is 0.

Moved stat_buf.f_frsize to the dividend.

Signed-off-by: Stanislav Kholmanskikh <stanislav.kholmanskikh@oracle.com>

2013-11-06 12:56:32 Tree
[c7b23c] by Wang, Xiaoguang/王 小光 Wang, Xiaoguang/王 小光 , pushed by Cyril Hrubis Cyril Hrubis

access(2): add ELOOP, ENOTDIR, EROFS error number test

access/access05.c: add ELOOP, ENOTDIR error number test

new access/access06.c: add EROFS error number test

Signed-off-by: Xiaoguang Wang <wangxg.fnst@cn.fujitsu.com>

2013-11-06 12:36:23 Tree
[90d190] by Stanislav Kholmanskikh Stanislav Kholmanskikh , pushed by Cyril Hrubis Cyril Hrubis

prot_hsymlinks: use SAFE_TOUCH

This test case does not need to keep the file descriptors
open after doing creat() on them. So we use SAFE_TOUCH instead
of SAFE_CREAT.

Otherwise on NFS it outputs:

TWARN : tst_rmdir: rmobj(/tmpdir/ltp-vdIWJBmrkz/provfaUSg) failed: remove(/tmpdir/ltp-vdIWJBmrkz/provfaUSg/tmp_root/hsym) failed; errno=66: Directory not empty

or:
TWARN : tst_rmdir: rmobj(/mnt/proXZmq0T) failed: unlink(/mnt/proXZmq0T/root/.nfs00000000000008470000015e) failed; errno=16: Device or resource busy

The same idea as for 8c200cb8e843724afb49fa6617fceec09ac826a5.

Signed-off-by: Stanislav Kholmanskikh <stanislav.kholmanskikh@oracle.com>

2013-11-06 10:59:28 Tree
[abf28b] by Stanislav Kholmanskikh Stanislav Kholmanskikh , pushed by Cyril Hrubis Cyril Hrubis

Implemented SAFE_TOUCH macro

Signed-off-by: Stanislav Kholmanskikh <stanislav.kholmanskikh@oracle.com>

2013-11-06 10:59:28 Tree
[36531d] by Simon Xu Simon Xu , pushed by Cyril Hrubis Cyril Hrubis

rwho01: fix getting hostnames from command output and $COUNT

Fix the way the script gets hostnames from rwho/ruptime command output,
and the mistake that $COUNT is used instead of $TST_COUNT in do_test().

Signed-off-by: Simon Xu <xu.simon@oracle.com>

2013-11-06 10:53:59 Tree
[ba3bf0] by Honggyu Kim Honggyu Kim , pushed by Wanlong Gao Wanlong Gao

acct01: add a check routine for acct implementation

A check routine for acct implementation is added at the beginning of the test.
check_acct_in_kernel function calls acct(NULL) and checks if errno is ENOSYS.
It means that acct system call is not implemented in kernel.
The other ENOSYS check routines in setup are removed.

This patch prevents from unintentional calling acct again in cleanup, which
makes TCONF to TBROK and helps configuration problems stay as TCONF.

Previous test output: (becomes TBROK in the end)
acct01 1 TCONF : BSD process accounting is not configured in this kernel
acct01 2 TCONF : Remaining cases not appropriate for configuration
acct01 3 TBROK : acct(NULL) failed: errno=ENOSYS(38): Function not implemented

Fixed test output: (stays as TCONF)
acct01 1 TCONF : BSD process accounting is not configured in this kernel

Signed-off-by: Honggyu Kim <hong.gyu.kim@lge.com>
Reviewed-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2013-11-06 02:26:38 Tree
[aba5c5] by Alexey Kodanev Alexey Kodanev , pushed by Wanlong Gao Wanlong Gao

device-drivers: pci: remove bus_slot file on exit

Signed-off-by: Alexey Kodanev <alexey.kodanev@oracle.com>
Reviewed-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2013-11-06 00:45:39 Tree
Older >