From: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
mkswap refuses files of size 10*(page size).
On systems with 8192 page size swapon02, swapon03 tests fail because
they try to use files of size 41920 ( 81920).
Modified swapon02, swapon03 so they operate on files of size
10*(page
From: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
lib/tst_fill_file.c: added tst_fill_file function
mkswap refuses files of size 10*(page size).
On systems with 8192 page size swapon02, swapon03 tests fail because
they try to use files of size 41920 ( 81920).
Modified swapon02
From: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
Signed-off-by: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
---
include/test.h|9 -
lib/tst_run_cmd.c | 20 ++--
2 files changed, 26 insertions(+), 3 deletions(-)
diff --git a/include
From: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
Signed-off-by: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
---
include/test.h|9 -
lib/tst_run_cmd.c | 20 ++--
2 files changed, 26 insertions(+), 3 deletions(-)
diff --git a/include
From: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
Modified lib/tst_module.c:
* for updated tst_run_cmd specification
* to hide gcc warning initialization discards qualifiers from pointer target
type
Signed-off-by: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
---
lib
From: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
Signed-off-by: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
---
include/test.h | 10 +++
lib/tst_fill_file.c | 67 +++
2 files changed, 77 insertions(+), 0
From: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
mkswap refuses files of size 10*(page size).
On systems with 8192 page size swapon02, swapon03 tests fail because
they try to use files of size 41920 ( 81920).
Modified swapon1, swapon02, swapon03 so they operate on files of size
From: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
* kernel test for max_map_count_sysctl is:
/* Too many mappings? */
if (mm-map_count sysctl_max_map_count)
return -ENOMEM;
so in LTP test map_count should be greater than max_map_count by 1
* only [vsyscall
From: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
Signed-off-by: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
---
include/test.h|9 -
lib/tst_run_cmd.c | 50 --
2 files changed, 56 insertions(+), 3
Hi!
Hi, Cyril.
-if (!pid)
+if (!pid) {
+/* redirecting stdout and stderr if needed */
+if (stdout_path != NULL) {
+child_stdout = open(stdout_path,
+OPEN_FLAGS, OPEN_MODE);
+
+if
From: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
* now it uses tst_kvercmp (to support kernels = 3)
* some distros doesn't allow creation of quota files in
directories with SELinux file_t type. Now we explicitly
change file type of the test directory to tmp_t
Signed-off
Hi!
This is a little fix to run quota_remount_test01.sh with SELinux enabled.
Some distros (CentOS 6.x and, likely, RHEL 6) prevent from creation of quota
files in 'file_t' directories, some - don't care.
I tried this fix on CentOS 5.x/6.x and Oracle Linux, Debian 7. Seems, It's ok,
and
On 08/05/2013 03:33 PM, chru...@suse.cz wrote:
Hi!
What about passing file descriptors to function (-1 means no
redirection) and create a second function on the top of it that
opens the files, does the checks and the calls the tst_run_cmd?
So you propose creating two functions: one will
On 08/05/2013 07:41 PM, chru...@suse.cz wrote:
Hi!
Hi, Cyril.
* kernel test for max_map_count_sysctl is:
/* Too many mappings? */
if (mm-map_count sysctl_max_map_count)
return -ENOMEM;
Hmm, that looks like we allow the map_count to became one greater than
max_map_count, is
On 08/05/2013 07:16 AM, Zhouping Liu wrote:
On 07/30/2013 09:47 PM, Stanislav Kholmanskikh wrote:
From: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
* kernel test for max_map_count_sysctl is:
/* Too many mappings? */
if (mm-map_count sysctl_max_map_count)
return
From: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
Modified lib/tst_module.c:
* for updated tst_run_cmd specification
* to hide gcc warning initialization discards qualifiers from pointer target
type
Signed-off-by: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
---
lib
Now there are two functions:
tst_run_cmd() - redirects cmd's stdout/stderr to file descriptors
tst_run_cmd_rdr_file() - a wrapper around tst_run_cmd(), redirects its output
to files.
--
Get your SQL database under
From: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
For tst_run_cmd added support for stdout/stderr redirection to
a file descriptor.
Added function tst_run_cmd_rdr_file which allows to redirect
tst_run_cmd output to a file.
Signed-off-by: Stanislav kholmanskikh stanislav.kholmansk
From: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
Signed-off-by: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
---
lib/tst_module.c |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/lib/tst_module.c b/lib/tst_module.c
index 832da27..c0f2c8f
From: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
Added function tst_run_cmd_fds which allows redirection of
stdout/stderr to file descriptors.
Now function tst_run_cmd is a wrapper to tst_run_cmd_fds.
Signed-off-by: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
On 08/05/2013 07:45 PM, chru...@suse.cz wrote:
Hi!
* now it uses tst_kvercmp (to support kernels = 3)
* some distros doesn't allow creation of quota files in
directories with SELinux file_t type. Now we explicitly
change file type of the test directory to tmp_t
I can't check that the
From: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
Signed-off-by: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
---
include/test.h|4 ++--
lib/tst_module.c |4 ++--
lib/tst_run_cmd.c |4 ++--
3 files changed, 6 insertions(+), 6 deletions(-)
diff --git
Signed-off-by: Stanislav kholmanskikh stanislav.kholmansk...@oracle.com
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
include/test.h|4 ++--
lib/tst_module.c |4 ++--
lib/tst_run_cmd.c |6 +++---
3 files changed, 7 insertions(+), 7 deletions(-)
diff
sizes.
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
testcases/kernel/syscalls/swapon/Makefile|5 ++
testcases/kernel/syscalls/swapon/libswapon.c | 48 +
testcases/kernel/syscalls/swapon/libswapon.h | 34
testcases/kernel/syscalls
Hi!
I think all the style issues are fixed.
--
Get 100% visibility into Java/.NET code with AppDynamics Lite!
It's a free troubleshooting tool designed for production.
Get down to code-level detail for bottlenecks, with
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
include/test.h | 10 +++
lib/tst_fill_file.c | 67 +++
2 files changed, 77 insertions(+), 0 deletions(-)
create mode 100644 lib/tst_fill_file.c
diff --git
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
include/ltp_signal.h | 17
.../kernel/syscalls/rt_sigaction/rt_sigaction02.c | 21 +--
.../kernel/syscalls/rt_sigaction/rt_sigaction03.c |2 +
3
On 08/06/2013 10:52 AM, Stanislav Kholmanskikh wrote:
On 08/05/2013 07:41 PM, chru...@suse.cz wrote:
Hi!
Hi, Cyril.
* kernel test for max_map_count_sysctl is:
/* Too many mappings? */
if (mm-map_count sysctl_max_map_count)
return -ENOMEM;
Hmm, that looks like we allow
On 08/14/2013 07:56 AM, Zhouping Liu wrote:
On 08/13/2013 09:29 PM, Stanislav Kholmanskikh wrote:
On 08/06/2013 10:52 AM, Stanislav Kholmanskikh wrote:
On 08/05/2013 07:41 PM, chru...@suse.cz wrote:
Hi!
Hi, Cyril.
* kernel test for max_map_count_sysctl is:
/* Too many mappings
sizes.
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
testcases/kernel/syscalls/swapon/Makefile|5 ++
testcases/kernel/syscalls/swapon/libswapon.c | 48 +
testcases/kernel/syscalls/swapon/libswapon.h | 34
testcases/kernel/syscalls
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
include/test.h | 10 +++
lib/tst_fill_file.c | 71 +++
2 files changed, 81 insertions(+), 0 deletions(-)
create mode 100644 lib/tst_fill_file.c
diff --git
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
include/ltp_signal.h | 17 +
.../kernel/syscalls/rt_sigaction/rt_sigaction02.c | 20 +---
.../kernel/syscalls/rt_sigaction/rt_sigaction03.c |1 +
3
One of parameters to setgroups() syscall is (gid_t *) pointer.
If TST_USE_COMPAT16_VSYSCALL is defined a pointer to GID_T is passed
instead (and sizeof(GID_T) sizeof(gid_t)). It's not safe and
can result in unaligned access (and SIGBUS) on several platforms.
Signed-off-by: Stanislav Kholmanskikh
In accordance to man migrate_pages(), mbind() the bit mask size
should be rounded to next multiple of sizeof(unsigned long).
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
.../syscalls/migrate_pages/migrate_pages01.c |5 -
.../syscalls/migrate_pages
On 08/21/2013 01:17 PM, chru...@suse.cz wrote:
Hi!
Hi.
---
.../syscalls/migrate_pages/migrate_pages01.c |5 -
.../syscalls/migrate_pages/migrate_pages02.c |5 -
2 files changed, 8 insertions(+), 2 deletions(-)
diff --git
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
include/test.h |6 ++
1 files changed, 6 insertions(+), 0 deletions(-)
diff --git a/include/test.h b/include/test.h
index d9eba31..a517ff9 100644
--- a/include/test.h
+++ b/include/test.h
@@ -105,6 +105,12
Now nodemask_size is rounded up to the next multiple
of sizeof(nodemask_t).
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
testcases/kernel/lib/numa_helper.c |6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/testcases/kernel/lib
On 08/21/2013 04:29 PM, Jan Stancek wrote:
- Original Message -
From: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
To: ltp-list@lists.sourceforge.net
Cc: vasily isaenko vasily.isae...@oracle.com, jstan...@redhat.com
Sent: Wednesday, 21 August, 2013 1:54:58 PM
Subject
On 08/21/2013 06:28 PM, Jan Stancek wrote:
- Original Message -
From: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
To: Jan Stancek jstan...@redhat.com
Cc: ltp-list@lists.sourceforge.net, vasily isaenko
vasily.isae...@oracle.com
Sent: Wednesday, 21 August, 2013 3:22:04
In accordance to man migrate_pages(), mbind() the bit mask size
should be rounded to the next multiple of sizeof(unsigned long).
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
.../syscalls/migrate_pages/migrate_pages01.c |4 ++--
.../syscalls/migrate_pages
Now nodemask_size is rounded up to the next
multiple of sizeof(unsigned long).
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
testcases/kernel/lib/numa_helper.c |8
1 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/testcases/kernel/lib
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
include/test.h |6 ++
1 files changed, 6 insertions(+), 0 deletions(-)
diff --git a/include/test.h b/include/test.h
index d9eba31..a517ff9 100644
--- a/include/test.h
+++ b/include/test.h
@@ -105,6 +105,12
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
include/test.h |6 ++
1 files changed, 6 insertions(+), 0 deletions(-)
diff --git a/include/test.h b/include/test.h
index d9eba31..a517ff9 100644
--- a/include/test.h
+++ b/include/test.h
@@ -105,6 +105,12
On 08/23/2013 11:57 AM, Jan Stancek wrote:
- Original Message -
From: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
To: ltp-list@lists.sourceforge.net
Cc: vasily isaenko vasily.isae...@oracle.com, jstan...@redhat.com
Sent: Thursday, 22 August, 2013 9:41:47 AM
Subject
) on several platforms.
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
testcases/kernel/syscalls/setgroups/compat_16.h | 22
+++-
testcases/kernel/syscalls/setgroups/setgroups04.c |6 -
2 files changed, 25 insertions(+), 3 deletions
On 08/23/2013 03:39 PM, chru...@suse.cz wrote:
Hi!
Thank you for this point.
At first sight it seems that this testcase is very specific for i386
(maybe for other architectures also).
In i386 I found two variants of setgroups syscall (from
arch/x86/syscalls/syscall_32.tbl):
81
On 08/26/2013 03:30 PM, chru...@suse.cz wrote:
Hi!
I looked at arch/* from linux sources and prepared this table:
Architecture | __NR_setgroups32 defined | __NR_setgroups defined | glibc
for setgroups() invokes |
alpha| | + |
arm | + | + |
avr32 |
On 08/26/2013 06:16 PM, chru...@suse.cz wrote:
Hi!
But we can't rely on the fact that group ids of a user invoking the
testcases are fit into __old_gid_t.
And the purpose of current GETGROUPS definition is to handle such cases.
A situation may happen where root user is a member of a group
getgroups01_16 and getgroups03_16 did not use 16-bit
version of getgroups() syscall. Fixed this in the same manner as
syscalls/setgroups.
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
testcases/kernel/syscalls/getgroups/compat_16.h | 78
Now 16-bit testcases are executed only if platform
has corresponding 16-bit version of setgroups() syscall.
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
testcases/kernel/syscalls/setgroups/compat_16.h | 11 +--
1 files changed, 9 insertions(+), 2
On 08/21/2013 04:19 PM, Dave Kleikamp wrote:
On 08/21/2013 01:50 AM, Jan Stancek wrote:
- Original Message -
From: Dave Kleikamp dave.kleik...@oracle.com
To: ltp-list@lists.sourceforge.net
Sent: Tuesday, 20 August, 2013 10:34:39 PM
Subject: [LTP] [PATCH] ltp: handle missing
On 08/27/2013 03:22 PM, chru...@suse.cz wrote:
Hi!
But we can't rely on the fact that group ids of a user invoking the
testcases are fit into __old_gid_t.
And the purpose of current GETGROUPS definition is to handle such cases.
A situation may happen where root user is a member of a group
In accordance to man migrate_pages(), mbind() the bit mask size
should be rounded to next multiple of sizeof(unsigned long).
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
.../syscalls/migrate_pages/migrate_pages01.c |4 ++--
.../syscalls/migrate_pages
This is a V4 series of patches.
Changes since V3:
* Macros ALIGN, __ALIGN_MASK renamed to LTP_ALIGN, __LTP_ALIGN_MASK
* Both macros are placed in compiler.h (instead of test.h)
--
Introducing Performance Central, a
Now nodemask_size is rounded up to the next
multiple of sizeof(unsigned long).
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
testcases/kernel/lib/numa_helper.c |9 +
1 files changed, 5 insertions(+), 4 deletions(-)
diff --git a/testcases/kernel/lib
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
include/compiler.h |6 ++
1 files changed, 6 insertions(+), 0 deletions(-)
diff --git a/include/compiler.h b/include/compiler.h
index ec714fe..4b41f65 100644
--- a/include/compiler.h
+++ b/include/compiler.h
On 08/27/2013 04:16 PM, chru...@suse.cz wrote:
Hi!
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
testcases/kernel/syscalls/getgroups/compat_16.h | 78
+
testcases/kernel/syscalls/getgroups/getgroups01.c | 22 +++---
testcases
On 08/28/2013 05:07 PM, Dave Kleikamp wrote:
Make map_address a volatile pointer so that it is dereferenced in
read_mem() after read_lock is set.
Also read_lock must protect the verbose printf that reads from map_address
as well.
Signed-off-by: Dave Kleikamp dave.kleik...@oracle.com
diff
syscalls/setgroups: moved compat_16.h to ../utils. This header will be
common for all 16-bit syscalls definitions.
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
testcases/kernel/syscalls/setgroups/compat_16.h | 85 --
testcases/kernel/syscalls
If we create/modify/touch file syscalls/utils/compat_16.h and execute
'make' in any of syscalls directories which includes compat_16.mk, then
nothing will happen, because this approach:
%.c: $(COMPAT_16_H)
is not working.
Fixed this.
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk
getgroups01_16 and getgroups03_16 did not use 16-bit
versions of getgroups() syscall. Fixed this in the same manner like
syscalls/setgroups.
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
testcases/kernel/syscalls/getgroups/getgroups01.c | 22
this.
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
testcases/kernel/syscalls/utils/compat_16.mk |6 --
1 files changed, 4 insertions(+), 2 deletions(-)
diff --git a/testcases/kernel/syscalls/utils/compat_16.mk
b/testcases/kernel/syscalls/utils
On 09/02/2013 07:23 PM, chru...@suse.cz wrote:
We can even simplify this macros. At least the innter part of SETGROUPS
and GETGROUPS could be created by a single macro:
COMPAT16_SYSCALL(setgroups, gitsetsize, list16)
#define COMPAT16_SYSCALL(sys_name, ...) \
# ifdef
On 09/03/2013 02:11 PM, chru...@suse.cz wrote:
Hi!
Sorry, but is it possible to use #ifdef or #if inside a #define macro?
Right it does not, sorry. I tend to forget how dumb it the C preprocessor is.
This doesn't work:
#define MYDEF(NN) \
# ifdef TST_USE_COMPAT16_SYSCALL
return
syscalls/setgroups: moved compat_16.h to ../utils. This header will be
common for all 16-bit syscalls definitions.
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
testcases/kernel/syscalls/setgroups/compat_16.h | 85 -
testcases/kernel/syscalls
getgroups01_16 and getgroups03_16 did not use 16-bit
versions of getgroups() syscall. Fixed this in the same manner like
syscalls/setgroups.
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
testcases/kernel/syscalls/getgroups/getgroups01.c | 22
syscalls/setgroups: moved compat_16.h to ../utils. This header will be
common for all 16-bit syscalls definitions.
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
testcases/kernel/syscalls/setgroups/compat_16.h | 85 -
testcases/kernel/syscalls
Hi!
Changes since V3:
[PATCH 1/3]: no changes
[PATCH 2/3]:
* Added Oracle copyright string to syscalls/utils/compat_16.h
* Fixed comment next to LTP_CREATE_SYSCALL
* Used alternative variant: passing cleanup callback parameter to all
caps functions. I think It looks more compact.
[PATCH
If we create/modify/touch file syscalls/utils/compat_16.h and execute
'make' in any of syscalls directories which includes compat_16.mk, then
nothing will happen, because this approach:
%.c: $(COMPAT_16_H)
is not working.
Fixed this.
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk
Hi!
This is part 1 of fixes for 16-bit syscalls.
--
How ServiceNow helps IT people transform IT departments:
1. Consolidate legacy IT systems to a single system of record for IT
2. Standardize and globalize service
Fixed 16-bit versions of testcases:
* getgid
* getuid
* setgid
* setuid
the way like setgroups.
Testcase getuid02 in fact tests geteuid() so renamed it to geteuid02.
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
runtest/ltplite |2
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
testcases/kernel/syscalls/getegid/compat_16.h | 48 -
testcases/kernel/syscalls/getegid/getegid01.c |3 +-
testcases/kernel/syscalls/getegid/getegid02.c | 32 +++--
testcases
Hello!
LTP employs FORK_OR_VFORK macro which is defined as:
#ifdef UCLINUX
#define FORK_OR_VFORK vfork
#define MAP_PRIVATE_EXCEPT_UCLINUX 0
#else
#define FORK_OR_VFORK fork
#define MAP_PRIVATE_EXCEPT_UCLINUX MAP_PRIVATE
#endif
So, on UCLINUX it is a
On 09/12/2013 05:41 PM, chru...@suse.cz wrote:
Hi!
This is part 1 of fixes for 16-bit syscalls.
Pushed, thanks.
Note that I've cleaned up the test sources and fixed setuid04 to use
standard temp directory instread of messing up CWD.
Thank you.
I'm proceeding with setfs{uid,gid},
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
testcases/kernel/syscalls/setfsgid/setfsgid01.c | 97
testcases/kernel/syscalls/setfsgid/setfsgid02.c | 95
testcases/kernel/syscalls/setfsgid/setfsgid03.c | 118 ++-
testcases
On 09/16/2013 07:18 PM, chru...@suse.cz wrote:
Hi!
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
testcases/kernel/syscalls/setfsgid/setfsgid01.c | 97
testcases/kernel/syscalls/setfsgid/setfsgid02.c | 95
testcases/kernel
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
testcases/kernel/syscalls/getgroups/getgroups01.c |2 +-
testcases/kernel/syscalls/getgroups/getgroups03.c |2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/testcases/kernel/syscalls/getgroups
On 09/19/2013 04:47 PM, Stanislav Kholmanskikh wrote:
Hi!
During looking into shell testcases I came across a situation where
* main shell script is a bash script (ie it has #!/bin/bash in the header)
* for each testcase number main shell script executes a separate (child)
shell-script
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
.../cpuset_base_ops_testset.sh | 12 ++--
.../cpuset_exclusive_test/cpuset_exclusive_test.sh |6 +-
.../kernel/controllers/cpuset/cpuset_funcs.sh |2 +-
.../cpuset_hierarchy_test
Hi!
I have very limited access to real hardware with 4 NUMA nodes. So most
of these fixes were checked with numa=fake=4 kernel cmdline.
It would be great if somebody verified it on real hardware.
And:
1. What do you think about idea of assigning to each cpuset controller test
a separate
Fixed lack of '#include config.h', perform some cleanup
and set nr_cpus, nr_mems based on environment.
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
.../cpuset_load_balance_test/cpuset_cpu_hog.c | 16
.../cpuset_load_balance_test.sh
to a value of 2 is not sufficient. Because there
is no guarantee that these hugepages will be allocated on the specified
node
* cleanup and several typos are fixed
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
.../cpuset/cpuset_memory_test/cpuset_memory_test.c | 31
For unification number of nodes is set based on environment.
I've not found a better way to guarantee dropping of caches than
just increasing sleep timeout twice (to 20 seconds).
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
.../cpuset_memory_spread_testset.sh
At least since kernel 3.0.0 writing '0-' to
$CPUSET/N/{cpus, mems} files doesn't raise an error.
Reading content of these files after writing gives '0'.
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
.../cpuset_base_ops_testset.sh | 16
Because of lack of '#include config.h' cpuinfo.c file
was compiled incorrectly (i.e. final lib libcpu_set.a didn't contain
the functions/variables from cpuinfo.c)
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
.../kernel/controllers/cpuset/cpuset_lib/cpuinfo.c |1
stack anon=5 dirty=5 N2=4 N3=1
So modified a check in test 10 and removed unnessesary checks for default
memory policy from tests 14, 15, 16.
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
.../cpuset/cpuset_syscall_test/Makefile|1
On 09/26/2013 04:06 PM, chru...@suse.cz wrote:
Hi!
For unification number of nodes is set based on environment.
I've not found a better way to guarantee dropping of caches than
just increasing sleep timeout twice (to 20 seconds).
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
tools/apicmds/.gitignore |1 +
tools/apicmds/Makefile|4 +-
tools/apicmds/ltpapicmd.c | 81 -
3 files changed, 83 insertions(+), 3 deletions(-)
diff --git
resources
* setting nr_hugepages to a value of 2 is not sufficient. Because there
is no guarantee that these hugepages will be allocated on the specified
node
* cleanup and several typos are fixed
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
.../cpuset
Renamed cpuset01 to cpuset12.
A separate name (from cpuset01 to cpuset11) is assigned to each cpuset
shell testcase (which was invoked earlier by run_cpuset_test.sh).
This will simplify results analysis.
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
runtest
On 09/26/2013 06:06 PM, chru...@suse.cz wrote:
Hi!
If I remove this sleep and perform two sequential runs I get this error:
root@debian7-x64:/opt/ltp/testcases/bin# ./cpuset_memory_spread_testset.sh
100+0 records in
100+0 records out
104857600 bytes (105 MB) copied, 0.0595533 s, 1.8 GB/s
On 09/30/2013 12:45 PM, Stanislav Kholmanskikh wrote:
On 09/26/2013 06:06 PM, chru...@suse.cz wrote:
Hi!
If I remove this sleep and perform two sequential runs I get this error:
root@debian7-x64:/opt/ltp/testcases/bin# ./cpuset_memory_spread_testset.sh
100+0 records in
100+0 records out
For unification number of nodes is set based on environment.
In accordance to Documentation/sysctl/vm.txt to reliably drop caches
we should invoke 'sync' before 'echo N /proc/sys/vm/drop_caches'.
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
On 09/30/2013 05:04 PM, chru...@suse.cz wrote:
Hi!
Hmmm, Mems_allowed* content of /proc/$test_pid/status before sending
SIGUSR1 to $test_pid is:
Mems_allowed:,0001
Mems_allowed_list:0
cpuset111 TFAIL : hog the memory on the unexpected
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
tools/apicmds/.gitignore |1 +
tools/apicmds/Makefile|4 +-
tools/apicmds/ltpapicmd.c | 83 +++--
3 files changed, 83 insertions(+), 5 deletions(-)
diff --git
On 09/30/2013 08:34 PM, chru...@suse.cz wrote:
Hi!
Hi!
You don't have to copy the argv[3] string, you are allowed to modify it
by the C standard.
Sent V2 version.
Well the whole ltpapicmd.c code is a bit hairy. I would personaly like
to see the main being split into functions but that may
On 09/30/2013 06:17 PM, chru...@suse.cz wrote:
Hi!
Renamed cpuset01 to cpuset12.
I slightly preffere to keep the test ids constant if possible. Note that
there are databases of test results used for comparsions and changing
the test ids would confuse such systems.
I agree.
So in this case
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
tools/apicmds/.gitignore |1 +
tools/apicmds/Makefile|4 +-
tools/apicmds/ltpapicmd.c | 77 +++--
3 files changed, 77 insertions(+), 5 deletions(-)
diff --git
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
tools/apicmds/ltpapicmd.c | 339 +
1 files changed, 189 insertions(+), 150 deletions(-)
diff --git a/tools/apicmds/ltpapicmd.c b/tools/apicmds/ltpapicmd.c
index f74ffca
A separate name (from controller_cpuset01 to controller_cpuset11) is
assigned to each cpuset shell testcase (which was invoked earlier
by run_cpuset_test.sh).
This will simplify results analysis.
Signed-off-by: Stanislav Kholmanskikh stanislav.kholmansk...@oracle.com
---
runtest/controllers
1 - 100 of 427 matches
Mail list logo
