From: Subrata <ris...@li...> - 2010-07-03 18:15:59
|
The branch, master, has been updated via 848f37617b3aee594408d399253d8e196095ca01 (commit) from e3bc3758d3e4d17a0e04009ef873df53819b9ec6 (commit) - Log ----------------------------------------------------------------- commit 848f37617b3aee594408d399253d8e196095ca01 Author: Subrata Modak <sub...@su...> Date: Sat Jul 3 23:45:03 2010 +0530 Changes till July 3rd 2010, but will be release as 20100630 ;-). Signed-of-by: Subrata Modak <su...@li...>, ----------------------------------------------------------------------- Summary of changes: ChangeLog | 422 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 files changed, 422 insertions(+), 0 deletions(-) diff --git a/ChangeLog b/ChangeLog index dcba787..9b43401 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,3 +1,425 @@ +LTP-20100630 + +commit e3bc3758d3e4d17a0e04009ef873df53819b9ec6 +Author: Subrata Modak <sub...@su...> +Date: Sat Jul 3 23:32:27 2010 +0530 + + [PATCH v4] Remove files which does not have a dmesg entry + + The recently introduced option: + + -K DMESG_LOG_DIR + Log Kernel messages generated for each test cases inside this directory, + + when used will generate huge no. of files inside the DMESG_DIR. Many of them will be ZERO size as we do not expect KERNEL MESSAGE to be generated against all the LTP TE + Signed-off-by: Subrata Modak <su...@li...>, + +commit 9e365e8e209904086b18420438b4792851c219b8 +Author: Subrata Modak <sub...@su...> +Date: Sat Jul 3 23:13:41 2010 +0530 + + semctl01 and shmctl01 will hang during uClinux test: ltp test case shmctl01 and semctl01 can not execute. When press ctrl and c, it can not exist. + -- + root:/> shmctl01 + + ^Cshmctl01 0 INFO : received unexpected signal 2 + ^Cshmctl01 0 INFO : received unexpected signal 2 + -- + The reason is the child will unlink the pipe if pipename isn't NULLï¼which may make the parent waitting forever. Signed-off-by: Bob Liu <ll...@gm...>. + +commit 65f7070cb3f67eba009e68413187727f508623e1 +Author: Subrata Modak <sub...@su...> +Date: Sat Jul 3 23:10:38 2010 +0530 + + From: gowrishankar <gow...@li...> + + In a scenario where a thread other than high priority noise thread is scheduled after the barrier, a priority inversion will not occur, defeating the purpose of test. W + Signed-off-by: Gowrishankar<gow...@in...> + Tested-by: Gowrishankar<gow...@in...> + Acked-by: Darren Hart <dv...@us...> + +commit 785ef40630157d66067b1577a238184e0170e383 +Author: Subrata Modak <sub...@su...> +Date: Sat Jul 3 23:10:14 2010 +0530 + + From: gowrishankar <gow...@li...> + + testpi-4 defines THREAD_STOP which is not used anywhere in the program. + + Signed-off-by: Gowrishankar <gow...@in...> + Tested-by: Gowrishankar <gow...@in...> + Acked-by: Darren Hart <dv...@us...> + +Author: Subrata Modak <sub...@su...> +Date: Sat Jul 3 23:09:55 2010 +0530 + + From: gowrishankar <gow...@li...> + + Instead of lock/unlock of mutex to burn cpu cycles, we can use busy_work_ms + from librttest which is more explicit (burn cycles, not provide some + synchronization). + + Also this patch removes the id increment since it is not used. + + Signed-off-by: Gowrishankar <gow...@in...> + Tested-by: Gowrishankar <gow...@in...> + Acked-by: Darren Hart <dv...@us...> + +commit a9e2238be689fd8e4d25bc7eefca1cb11535ce7a +Author: Subrata Modak <sub...@su...> +Date: Sat Jul 3 23:09:32 2010 +0530 + + From: gowrishankar <gow...@li...> + + Since main() sets the cpu affinity, it is not necessary for the + spawned threads to do so, as the threads inherit the cpu affinity. + + Signed-off-by: Gowrishankar <gow...@in...> + Tested-by: Gowrishankar <gow...@in...> + Acked-by: Darren Hart <dv...@us...> + +commit 5305c4a10cf26799c81e6f0362750988457d3002 +Author: Subrata Modak <sub...@su...> +Date: Sat Jul 3 23:09:02 2010 +0530 + + From: gowrishankar <gow...@li...> + + This patch adds the description to the test testpi-4. + Signed-off-by: Gowrishankar <gow...@in...> + Tested-by: Gowrishankar <gowrishankar.m@in.ibm. + +commit a21dab25b7ef183d049b31b0d3fed1d48c16dc2b +Author: Subrata Modak <sub...@su...> +Date: Sat Jul 3 23:06:38 2010 +0530 + + From: gowrishankar <gow...@li...> + + testpi-4 is initializing mutex step by step which can be done + by reusing the init_pi_mutex() funtion available in librttest. + + Signed-off-by: Gowrishankar <gow...@in...> + Tested-by: Gowrishankar <gow...@in...> + Acked-by: Darren Hart <dv...@us...> + +commit b4f57c7d26b58f3d80b46340f10a397d21957866 +Author: Subrata Modak <sub...@su...> +Date: Sat Jul 3 23:06:16 2010 +0530 + + From: gowrishankar <gow...@li...> + + This patch cleans up the testpi-4 so as to follow the + C coding style standards. + + Signed-off-by: Gowrishankar <gow...@in...> + Tested-by: Gowrishankar <gow...@in...> + Acked-by: Darren Hart <dv...@us...> + +commit 76a720a381f351b33eb203c7c519e6f44bbe7219 +Author: Subrata Modak <sub...@su...> +Date: Sat Jul 3 21:08:18 2010 +0530 + + While executing the ltp on an nfs root mounted system, I faced a problem with some of the ltp test cases. These tests either did not close their file descriptors or did + +commit 13adcb0c4496e75bdcdbc02410b46b9fbc6120a8 +Author: Subrata Modak <sub...@su...> +Date: Fri Jul 2 22:47:57 2010 +0530 + + ima_measure doesn't build correctly. It needs config.h to determine whether openssl headers are present, and verbose needs to be global. Broken by commit ef77253961f90 + Signed-off-by: Stephen D. Smalley <sd...@ty...>, + Acked-by: Garrett Cooper <yan...@gm...>, + +commit 04b500a71b8a410adf5b92978c9d93d659912c20 +Author: Subrata Modak <sub...@su...> +Date: Thu Jul 1 22:33:51 2010 +0530 + + fix hugetlb tests: I want to suggest a patch for hugetlb tests. There are "-c" options in this "${LTPROOT}/runtest/hugetlb" file: + ------------ + hugeshmat01 hugeshmat01 -i 5 + hugeshmat02 hugeshmat02 -i 5 + hugeshmat03 hugeshmat03 -i 5 + hugeshmctl01 hugeshmctl01 -i 5 -c 3 + hugeshmctl02 hugeshmctl02 -i 5 -c 3 + hugeshmctl03 hugeshmctl03 -i 5 -c 3 + hugeshmdt01 hugeshmdt01 -I 5 -c10 + hugeshmget01 hugeshmget01 -c 10 -i 10 + hugeshmget02 hugeshmget02 -c 10 -i 10 + hugeshmget03 hugeshmget03 -c 10 -i 10 + hugeshmget05 hugeshmget05 -c 10 -i 10 + ------------ + But I think that these "-c" options have not important meanings. I think that the test for several times may not be necessary. For example, the purpose of "hugeshmget02 + Signed-off-by: Tomonori Mitani <mi...@ry...>, + +commit 0591a5dcc4c5d8b1cb282e4afd163f70b62350ce +Author: Subrata Modak <sub...@su...> +Date: Thu Jul 1 22:29:20 2010 +0530 + + [PATCH] realtime: rearrange the pthread attr functions order in librttest, + From: gowrishankar <gow...@li...> + + In the recent pthread libraries, order of calling pthread_attr_setschedpolicy and pthread_attr_setschedparam plays important role in a way that, former should always be + Signed-off-by: Gowrishankar <gow...@in...>, + Tested-by: Gowrishankar <gow...@in...>, + Acked-by: Will Schmidt <wil...@vn...>, + Acked-by: Darren Hart <dv...@us...>, + +commit 66c15baaa8be08c5baae5a0c819df3b6a8c834dd +Author: Subrata Modak <sub...@su...> +Date: Thu Jul 1 22:27:54 2010 +0530 + + [PATCH] realtime: fix min and max calculations in matrix multiplication test. + From: gowrishankar <gow...@li...> + + In matrix multiplication test, the min and max calculations are done on stats_container_t data with its index field never set to the last element position. Due to this + Signed-off-by: Gowrishankar <gow...@in...>, + Tested-by: Gowrishankar <gow...@in...>, + Acked-by: Will Schmidt <wil...@vn...>, + Acked-by: Darren Hart <dv...@us...>, + +commit 95e8b61433efcefb83699aaf859e92ad2b576373 +Author: Subrata Modak <sub...@su...> +Date: Mon Jun 14 17:59:09 2010 +0530 + + Fix issue where growfiles tests still not creating temp files in directory specified by TMPDIR as expected. This is an updated patch which changes the arguments for gf1 + Scripts which use growfiles with command-line file names are not performing the tests in the expected target directory. When file names are provided on the growfiles co + + gf01 growfiles -B 800000 -W gf01 -b -e 1 -u -i 0 -L 20 -w -C 1 -l -I r -T 10 glseek20 glseek20.2 -d $TMPDIR + + The -d $TMPDIR construct is almost always used in the scripts to point to the filesystem under test. However, the presence of the glseek20 glseek20.2 command line param + + -f specifies the 'auto_file' which is the base file name (gf is used if it's missing) + -d specifies the 'auto_dir' to be used ('.' is used if it's missing) + -S specifies the number of 'auto seq' files to be used. These files are of the form: "%s/%s%d" using auto_dir, auto_file, file_number. + Signed-off-by: Henry Yei <hy...@mv...>, + +commit 73c785ff2eb296091d06d3b7184bddf9bec8ec7e +Author: Subrata Modak <sub...@su...> +Date: Mon Jun 14 17:56:43 2010 +0530 + + I want to suggest a patch for #2 test of "mail_tests.sh". The failure of #2 is as following: + ------------ + mail_tests::mail02 0 TINFO : Test #2: mail u...@bad-domain will result in failure + mail_tests::mail02 0 TINFO : Test #2: to deliver the mail. Mailer daemon should + mail_tests::mail02 0 TINFO : Test #2: report this failure. + mail_tests::mail02 2 TFAIL : Test #2: No new mail for root. Reason: No mail for root + ------------ + I use RHEL5.4, RHEL4.8 systems and x86, x86_64, ia64 architectures. "#2" test failed in all combinations. At "#2" test, there is "No mail for root" in "tst_mail.res" fi +: +commit 82d3ded3632c79c62e231f3f8d7a632ce3da4e40 +Author: Subrata Modak <sub...@su...> +Date: Mon Jun 14 17:54:59 2010 +0530 + + echo test is failing with the below errors as we didn't initialized the 'addrlen' parameter of getsockname. + Creating socket ..... + echoes 1 TBROK : getsockname call failed (pid=7086): Invalid argument + echo: Invalid argument + So, attached the patch which initializes the 'addrlen' parameter. Signed-off-by: Munipradeep Beerakam <mpr...@li...>. Please correct if anything is wrong + +commit a9ff0b53d9e45e89ea4ec794fc87bbf67620e2c4 +Author: Subrata Modak <sub...@su...> +Date: Mon Jun 14 17:52:10 2010 +0530 + + We need to add the below entry: + {"read", "/proc/powerpc/rtas/error_log", EINVAL}, + to the "Mapping known_issues[]" array, as the new kernels are having /proc/powerpc directory along with /proc/ppc64. Attached the patch to this mail. Signed-off-by: Mun + Please correct me if any thing is wrong. + +commit f2e36d93c856782c4292ad0c21729c35776c1e31 +Author: Subrata Modak <sub...@su...> +Date: Wed Jun 9 23:45:26 2010 +0530 + + TOMOYO Linux update. Update test programs for TOMOYO LSM module. This patch updates test programs for TOMOYO 2.3 which will be included into Linux kernel 2.6.36. Signed + +commit 7f20a3add6c7f2166dbab5a712713ae8b5943448 +Author: Subrata Modak <sub...@su...> +Date: Wed Jun 2 18:16:46 2010 +0530 + + Fix ext4_subdir_limit fix: We get the following error results since the if condition in ext4_subdir_limit_tset.sh does not work fine. The patch fixes this issue. + [root@bsd086 ltp]# ./runltp -f fs_ext4 + <snip> + Ext4 subdir limit test + ext4-subdir-limit 0 TINFO : Num of dirs to create: 65537, Dir name len: long name, mkdir: No space left on device + ext4-subdir-limit 2 TFAIL : failed to create directories - 19118 + <snip> + ext4-subdir-limit 0 TINFO : Num of dirs to create: 65537, Dir name len: long name, mkdir: No space left on device + ext4-subdir-limit 4 TFAIL : failed to create directories - 19592 + Signed-off-by: Akira Fujita <a-f...@rs...>, + +commit 76ec975d298e3890a4ab2f9471940e164de0c643 +Author: Subrata Modak <sub...@su...> +Date: Wed Jun 2 18:15:19 2010 +0530 + + Fix for UTIMENSAT test fails hugely: This patch installs the trap for cleanup_test which removes the sudoers file only if it was installed by utimensat script, and also + +commit 91f337f9ee05b028af8506c7e83dc1b8c23a3b32 +Author: Subrata Modak <sub...@su...> +Date: Fri May 28 12:57:36 2010 +0530 + fix io_xxxx Makefiles: + Makefiles for io_xxx syscall tests contains line: + LDFLAGS += $(AIO_LIBS) + Which leads to wrong order of flags passed to the linker and compilation failure on certain linux systems. Changing these lines to: + LDLIBS += $(AIO_LIBS) + Fixes this problem. Patch is attached. + Signed-off-by: Cyril Hrubis ch...@su... + +commit 6a19cb6ea8f0b160438288f87a96f57d8e63aad2 +Author: Subrata Modak <sub...@su...> +Date: Fri May 28 12:53:59 2010 +0530 + + It seems like file runtest/fs_perms_simple doesn't made it into ltp-dev git tree (also it seems to be included in the patch). + see: + http://ltp.git.sourceforge.net/git/gitweb.cgi?p=ltp/ltp-dev.git;a=commit;h=ca758b73afff8200df5e1a24e9abe8a6468c9e06 + -- + Cyril Hrubis + ch...@su... + +commit 4e94765d753c0ea91b4745ffac30df427a4d9202 +Author: Subrata Modak <sub...@su...> +Date: Fri May 28 12:50:38 2010 +0530 + + Timer related syscalls for LTP test: Signed-off-by: Suzuki K P <su...@in...>, + Fix timer_t argument for timer related direct syscalls. The tests timer_getoverrun and timer_gettime should be moved to testcases/kernel/timers directory. If that is do + /* Description */ + There is a mismatch in the "timer_t" type defined by the Linux kernel and the GNU Libc. + Linux kernel defines it as an int for all architectures. + include/linux/types.h : + typedef __kernel_timer_t timer_t; + arch/powerpc/include/asm/posix_types.h : + typedef int __kernel_timer_t; + Now, the glibc provides a wrapper to the timer related system calls. While providing the wrapper, Glibc stores some internal data associated with the timer. So it defin + /usr/include/time.h : + /* Timer ID returned by `timer_create'. */ + typedef __timer_t timer_t; + in /usr/include/bits/types.h : + /* Timer ID returned by `timer_create'. */ + __STD_TYPE __TIMER_T_TYPE __timer_t; + in /usr/include/bits/typesizes.h + So that leads to the definiton of timer_t as void* by glibc. And this cannot be changed. But the glibc uses the "proper" typed argument for the syscalls it issues by de + +commit 82f8e8fdfe892266d7346792b71783ac23bfa1a2 +Author: Subrata Modak <sub...@su...> +Date: Fri May 28 12:45:16 2010 +0530 + + Following set of patches applies functional changes in pi-tests for better reliability in recent kernels and their order is: + [PATCH 1/6] realtime: reuse init_pi_mutex from librttest + [PATCH 2/6] realtime: add test descriptions + [PATCH 3/6] realtime: threads need not set cpu affinity while main does + [PATCH 4/6] realtime: use busy_work_ms from librttest to burn cpu cycles + [PATCH 5/6] realtime: remove thread_stop which is not used anywhere + [PATCH 6/6] realtime: synchronization fixes for high priority noise thread + + All need to be applied in their order on top or code clean up patch series sent previously. + Signed-off-by: Gowrishankar <gow...@in...>, + Tested-by: Gowrishankar <gow...@in...>, + Acked-by: Darren Hart <dv...@us...>, + +commit e1c39ef4f1d61c05f176d063d1afa6f704bc6740 +Author: Subrata Modak <sub...@su...> +Date: Fri May 28 12:36:06 2010 +0530 + + realtime: code cleanup on pi-tests to adapt C coding standard: Following bulk of patches applies the coding style in pi-tests to follow the C coding standard. All are t + [PATCH 1/5] realtime: code cleanup on pi-test 0 + [PATCH 2/5] realtime: code cleanup on pi-test 1 + [PATCH 3/5] realtime: code cleanup on pi-test 2 + [PATCH 4/5] realtime: code cleanup on pi-test 5 + [PATCH 5/5] realtime: code cleanup on pi-test 6 + Signed-off-by: Gowrishankar <gow...@in...>, + Tested-by: Gowrishankar <gow...@in...>, + Acked-by: John Stultz <joh...@us...>, + Acked-by: Darren Hart <dv...@us...>, + +commit c81df6233bca8b3ad601a3a7dd6c02319a32a76b +Author: Subrata Modak <sub...@su...> +Date: Fri May 28 12:25:20 2010 +0530 + + Following patch adds "return NULL;" to the function returning void* in math float tests. This change silences compiler warnings. Signed-off-by: Cyril Hrubis chrubis@sus + +commit 4aafc2802f0c2c05e4891cea1f721f414b41ac67 +Author: Subrata Modak <sub...@su...> +Date: Fri May 28 12:22:02 2010 +0530 + + The mq_timedreceive01.c test defines _XOPEN_SOURCE 500 on the beginning but this causes missing prototypes for seteuid() and mq_timedreceive() as they need _XOPEN_SOURC + +commit e75668f99ade62e18e34891b2c522c8d97ee982b +Author: Subrata Modak <sub...@su...> +Date: Tue May 18 01:34:04 2010 +0530 + + ftruncate04.c: cleanup mandatory locking check. The attached patch tries to cleanup the `mandatory locking' check in the ftruncate04.c testcase. The currently test is m + +commit 6e9c735833c19ad5031a4116b2e74ee77ff53424 +Author: Subrata Modak <sub...@su...> +Date: Tue May 18 01:31:35 2010 +0530 + + New: open("a/", O_NOFOLLOW) fails with ELOOP if "a" is a symbolic link to a directory. Add O_NOFOLLOW open test with trailing slash. In a situation where symlink (say ' + +commit ca758b73afff8200df5e1a24e9abe8a6468c9e06 +Author: Subrata Modak <sub...@su...> +Date: Tue May 18 01:28:14 2010 +0530 + fix fs_perms tests. Signed-off-by: Cyril Hrubis <ch...@su...>. + +commit 440fa0f6d857d4a15802b5ecdfdf6057e59975bf +Author: Subrata Modak <sub...@su...> +Date: Wed May 12 22:07:14 2010 +0530 + + Add an Option '-K' to Create dmesg logs for each test: + Recently while running LTP , i found that the system had generated many kernel messages worth investigating. However, there was a loose gap to find out which test exact + 1) Absolute path(s) as specified by the user, else + 2) Inside $LTPROOT/output/ + Comments welcome. + Signed-off-by: Subrata Modak <su...@li...> + +commit bd9d440be1da8020c93ac92e497e3a16e66e9fa3 +Author: Subrata Modak <sub...@su...> +Date: Wed May 12 22:03:21 2010 +0530 + + Add an Option '-K' to Create dmesg logs for each test: + Recently while running LTP , i found that the system had generated many kernel messages worth investigating. However, there was a loose gap to find out which test exact + 1) Absolute path(s) as specified by the user, else + 2) Inside $LTPROOT/output/ + Comments welcome. + Signed-off-by: Subrata Modak <su...@li...>, + +commit 0ce709315c3f2493d8ff97d574641cb91f1b1fa6 +Author: Subrata Modak <sub...@su...> +Date: Sun May 9 19:44:04 2010 +0530 + + NetNS test fixes: By the way, these tests appear to start up sshd for no reason other than possibly checking that basic sockets work in the netns. Given the frailty and + NetNS test fixes (v2): + This patch fixes a couple of netns test issues that cause invalid failures relating to exiting with a status variable that doesn't exist. It also fixes an instance of + Changes in v2: + - Initialize status=0 at the top of child_1.sh to prevent overshadowing a (meaningless) sshd failure, + Signed-off-by: Dan Smith <da...@us...>, + Acked-by: Serge E. Hallyn <se...@us...>, + +commit 3aabc944423945e4857372fe726584f7889cbb8e +Author: Subrata Modak <sub...@su...> +Date: Fri May 7 14:28:18 2010 +0530 + + Most of these are belated cleanup after the move to using /opt/ltp. But come on, replacing 'return' with tst_exit(), are you just trying to mess with my head? + Changelog: + may 4: address Garrett's feedback + 1. single return 0 in print_caps.c + 2. use $TMP if defined for location of caps_fifo + 3. use tst_brkm in place of tst_resm. + may 5: address Garrett's comment: + don't add 1 to null pointer and then check for 1 + Signed-off-by: Serge E. Hallyn <se...@us...>, + Acked-by: Garrett Cooper <yan...@gm...>, +commit c98956ba8f40b4c9004ff5767db2b2d87ab46e41 +Author: Subrata Modak <sub...@su...> +Date: Fri May 7 14:21:00 2010 +0530 + + The MAX_CLOCKS supported by the kernel is 16. But the kernel may not have all the clock sources defined in the range 0-15. Hence, to test whether kernel returns error f + Signed-off-by: Suzuki K P <su...@in...>, + Acked-by: Garrett Cooper <yan...@gm...>, + +commit f0be8b0c4489defeb6eaf6715576524a48b8998a +Author: Subrata Modak <sub...@su...> +Date: Fri May 7 10:33:00 2010 +0530 + + We should to check the return value of trigger_eventfd_overflow() in overflow_poll_test(). + Signed-off-by: Shi Weihua <sh...@cn...>, + Acked-by: Garrett Cooper <yan...@gm...>, + LTP-20100430 Commit: 84f2e35f109a9fe6591af83be0ddbda4a5f7c78c Date: Thu, 29 Apr 2010 19:02:46 +0530 hooks/post-receive -- ltp |