Work at SourceForge, help us to make it a better place! We have an immediate need for a Support Technician in our San Francisco or Denver office.

Close

Git Log


Commit Date  
[dff530] (20120903) by shubham shubham

LTP 20120903

2012-09-03 06:24:13 Tree
[8c48cf] by Zhouping Liu Zhouping Liu , pushed by Wanlong Gao Wanlong Gao

tunable/min_free_kbytes: add 32-bit system limit

in 32-bit system, one process only allocate 3Gb memory at
most, even though there are enough memory and setting
overcommit_memory=1, the porcess always exit when it arrives
in 3Gb memory.

the patch use '__WORDSIZE' macro to avoid the incorrect 'TFAIL'
without the patch:
---- snip ----
min_free_kbytes 0 TINFO : set overcommit_memory to 1
min_free_kbytes 0 TINFO : set min_free_kbytes to 3794
memfree is 15478420 kB before eatup mem
memfree is 12333408 kB after eatup mem
min_free_kbytes 3 TFAIL : child unexpectedly failed: 0
min_free_kbytes 0 TINFO : set min_free_kbytes to 7588
---- snip ----

after applying the patch:
---- snip ----
min_free_kbytes 0 TINFO : set overcommit_memory to 1
min_free_kbytes 0 TINFO : set min_free_kbytes to 3794
memfree is 15478420 kB before eatup mem
memfree is 12333408 kB after eatup mem
min_free_kbytes 0 TINFO : Child can't allocate >3Gb memory in 32bit system
min_free_kbytes 0 TINFO : set min_free_kbytes to 7588
---- snip ----

Reported-by: Shuang Qiu <shuang.qiu@oracle.com>
Signed-off-by: Zhouping Liu <zliu@redhat.com>
Reviewed-by: Caspar Zhang <caspar@casparzhang.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2012-08-31 02:50:57 Tree
[56696a] by Zhouping Liu Zhouping Liu , pushed by Wanlong Gao Wanlong Gao

tunable/min_free_kbytes: remove map_count args

as the type of total_mem is 'unsigned long', inside
32-bit system, it will overflow in this sentence
map_count = total_mem * KB / MAP_SIZE

e.g: if total_mem > 4194304, 'total_mem*KB' will > ULONG_MAX.

after re-cehck the code, I found map_count was not necessary,
and removing it can fix the overflow issue.

Reported-by: Shuang Qiu <shuang.qiu@oracle.com>
Signed-off-by: Zhouping Liu <zliu@redhat.com>
Reviewed-by: Caspar Zhang <caspar@casparzhang.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2012-08-31 02:50:57 Tree
[a5c561] by Zhouping Liu Zhouping Liu , pushed by Wanlong Gao Wanlong Gao

tunable: reduce the min_free_kbytes tunable

Setting min_free_kbytes too high will cause system hangs,
especially in i386 arch, using less than 5% of total memory
can avoid it, so choose %5 of free memory or 2% of total memory.
Thanks Shuang pointed out it.

also, in order to enlarge the coverage, the patch adds a new
case item: 2 * default tunable

Signed-off-by: Zhouping Liu <zliu@redhat.com>
Signed-off-by: Shuang Qiu <shuang.qiu@oracle.com>
Reviewed-by: Caspar Zhang <caspar@casparzhang.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2012-08-31 02:50:57 Tree
[9dc832] by Wanlong Gao Wanlong Gao

posix:sigpause:3-1: further cleanup

make variables and functions static.

Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2012-08-29 03:08:38 Tree
[5318e4] by Wanlong Gao Wanlong Gao

posix:sigpause:4-1: code cleanup

cleanup coding style.

Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2012-08-29 03:05:52 Tree
[0265dc] by Wanlong Gao Wanlong Gao

posix:sigpause:2-1: code cleanup

cleanup coding style.

Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2012-08-29 03:05:40 Tree
[ea6031] by Wanlong Gao Wanlong Gao

posix:sigpause:1-2: code cleanup

cleanup coding style.

Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2012-08-29 03:05:34 Tree
[1241f4] by Wanlong Gao Wanlong Gao

posix:sigpause:1-1: code cleanup

cleanup coding style.

Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2012-08-29 03:05:19 Tree
[a1ee86] by DAN LI DAN LI , pushed by Wanlong Gao Wanlong Gao

sigpause/3-1.c Fix hung

If sem is set as INTHREAD finally,the loop will be endless.

Signed-off-by: DAN LI <li.dan@cn.fujitsu.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2012-08-29 03:03:07 Tree
[5d2207] by DAN LI DAN LI , pushed by Wanlong Gao Wanlong Gao

sigpause/3-1.c Clean up

Clean up sigpause/3-1.c.

Signed-off-by: DAN LI <li.dan@cn.fujitsu.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2012-08-29 03:03:06 Tree
[9d911a] by Jan Stancek Jan Stancek , pushed by Wanlong Gao Wanlong Gao

pthread_cond_signal: don't count pthread_cond_signals

Don't count how many pthread_cond_signals it took,
if any fails SIGALRM will trigger testcase failure.

Signed-off-by: Jan Stancek <jstancek@redhat.com>
Signed-off-by: Kang Kai <kai.kang@windriver.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2012-08-24 01:38:49 Tree
[37bfee] by Jan Stancek Jan Stancek , pushed by Wanlong Gao Wanlong Gao

proc01: add option to limit max read bytes

Add option -m to set upper limit of read megabytes from
single proc file. When this limit is reached testcase
will continue with next file.

Reason for this new option is to allow running this testcase
as sanity test, that doesn't need go through full content of
files such as "pagemap" (which can take hours).

Signed-off-by: Jan Stancek <jstancek@redhat.com>
Reviewed-by: Caspar Zhang <caspar@casparzhang.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2012-08-23 03:39:14 Tree
[2cb658] by DAN LI DAN LI , pushed by Wanlong Gao Wanlong Gao

pthread_cond_brodcast/1-2.c hung when mem left is not enough

pthread_cond_brodcast/1-2.c try to create as many childrens(thread or process)
as possible,however,if current free memory is exhausted before all childrens are
started,the case will timeout and hung since it's stated with program "t0".

The patch adjust count of childrens according to judgement of size of current free
memory.

Signed-off-by: DAN LI <li.dan@cn.fujitsu.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2012-08-23 01:33:06 Tree
[4f733f] by Kang Kai Kang Kai , pushed by Wanlong Gao Wanlong Gao

pthread_detach/4-3: Fix typo

Fix typo.

Signed-off-by: Kang Kai <kai.kang@windriver.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2012-08-21 09:22:40 Tree
[29e6de] by Kang Kai Kang Kai , pushed by Wanlong Gao Wanlong Gao

timer_settime/5-3: fix test hung

This case hangs on sugarbay(intel x86_64) platform.

The root cause is that the case calls timer_settime with time which
already took place and then calls sleep. When the timer expires and
sends signal the sleep will be break.

But on sugarbay the timer always expires before the case enters sleep,
then it always sleeps NUMTESTS * LONGSLEEPTIME seconds. So it exceeds
the time limit set by auto-run test shell script, and reports that it
hangs.

Decrease LONGSLEEPTIME value to avoid this issue.

Signed-off-by: Kang Kai <kai.kang@windriver.com>
Reviewed-by: Jan Stancek <jstancek@redhat.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2012-08-21 03:57:55 Tree
[f72ad7] by Jan Stancek Jan Stancek , pushed by Wanlong Gao Wanlong Gao

ioctl: check if tty device nodes are usable

ioctl testcases fail on s390x because it doesn't
support virtual consoles.

Trying to use such device node fails:
ioctl01_02 0 TINFO : Testing ioctl01 with /dev/tty0
ioctl01 1 TBROK : Couldn't open /dev/tty0: errno=ENXIO(6): No
such device or address

This patch tries to get terminal line settings for each tty,
that would be used in test. If that fails, tty is skipped.

Signed-off-by: Jan Stancek <jstancek@redhat.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2012-08-18 23:10:36 Tree
[07e750] by Caspar Zhang Caspar Zhang , pushed by Wanlong Gao Wanlong Gao

syscalls/numa: check syscall availability

some NUMA related syscalls: mbind, get_mempolicy are not supported on
some arches (e.g. i386), this patch checks the syscalls' availability in
setup() function by using the macro syscall() implemented in
linux_syscall_numbers.h, if the syscall is not implemented on the
system, a TCONF message will be given like this:

get_mempolicy01 1 TCONF : syscall __NR_get_mempolicy01 not supported on your arch

Signed-off-by: Caspar Zhang <caspar@casparzhang.com>
Reviewed-by: Zhouping Liu <zliu@redhat.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2012-08-10 09:14:47 Tree
[38c075] by Caspar Zhang Caspar Zhang , pushed by Wanlong Gao Wanlong Gao

mm: add common Makefile include: libmem.mk

libmem.mk is used as a common Makefile used by tests under mem/ dir to
avoid duplicated libmem-related lines.

Signed-off-by: Caspar Zhang <caspar@casparzhang.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2012-08-10 09:14:47 Tree
[a98ac1] by Caspar Zhang Caspar Zhang , pushed by Wanlong Gao Wanlong Gao

mm: use new numa_helper

This patch makes the tests in mem/ dir use numa_helper in libkerntest.

Signed-off-by: Caspar Zhang <caspar@casparzhang.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2012-08-10 09:14:47 Tree
[6e77d4] by Caspar Zhang Caspar Zhang , pushed by Wanlong Gao Wanlong Gao

numa_helper: move to libkerntest

libnuma_helper could be used by testcases both under syscalls/ and mem/,
it's better to add a new common lib in their parent dir, so I created
libkerntest under testcases/kernel/lib/, and put numa_helper to it. In
the future, we can add more APIs to it.

Signed-off-by: Caspar Zhang <caspar@casparzhang.com>
Tested-by: Zhouping Liu <zliu@redhat.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2012-08-10 09:14:47 Tree
[d534a4] by Jan Stancek Jan Stancek , pushed by Wanlong Gao Wanlong Gao

update testcases to use flag for get_allowed_nodes()

Interface of get_allowed_nodes() now requires a flag,
update all testcases using this interface.

Also fix some checkpatch warnings.

Signed-off-by: Jan Stancek <jstancek@redhat.com>
Signed-off-by: Caspar Zhang <caspar@casparzhang.com>
Reviewed-by: Zhouping Liu <zliu@redhat.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2012-08-10 09:14:47 Tree
[78de65] by Jan Stancek Jan Stancek , pushed by Wanlong Gao Wanlong Gao

numa_helper: add flag to get_allowed_nodes()

add flag, so testcases can request nodes with available
memory/cpus or both.

Also add small debug function.

Signed-off-by: Jan Stancek <jstancek@redhat.com>
Signed-off-by: Caspar Zhang <caspar@casparzhang.com>
Reviewed-by: Zhouping liu <zliu@redhat.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2012-08-10 09:14:47 Tree
[cbdd1d] by Marios Makris Marios Makris , pushed by Wanlong Gao Wanlong Gao

V3 /lib/tst_tmpdir Implemented a simple testcase.

Simple testcase to check the correct functionality of the tst_tmpdir.c file.
Also removed the Unit test blocks from the tst_tmdir.c.

Signed-off-by: Marios Makris <marios.makris@gmail.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2012-08-10 00:57:25 Tree
[c128c4] by Marios Makris Marios Makris , pushed by Wanlong Gao Wanlong Gao

/lib/tst_tmpdir.c removed the TDIRECTORY part

After recent question to the list, it seems that the TDIRECTORY PART is not
used by anyone, therefore it is removed.

Signed-off-by: Marios Makris <marios.makris@gmail.com>
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>

2012-08-10 00:57:24 Tree
Older >