Uploaded image for project: 'Lustre'
  1. Lustre
  2. LU-1580

POSIX: pathconf.32: pathconf("/dev/pts/0", _PC_NAME_MAX) did not give correct results

Details

    • Bug
    • Resolution: Fixed
    • Blocker
    • None
    • Lustre 2.3.0
    • None
    • 3
    • 5807

    Description

      This issue was created by maloo for Chris Gearing <chris@whamcloud.com>

      This issue relates to the following test suite run: https://maloo.whamcloud.com/test_sets/e8bc176e-c0a4-11e1-a5ca-52540035b04c.

      The sub-test test_1 failed with the following error:

      Run POSIX testsuite on /mnt/lustre failed

      Info required for matching: posix 1

      Attachments

        Issue Links

          Activity

            [LU-1580] POSIX: pathconf.32: pathconf("/dev/pts/0", _PC_NAME_MAX) did not give correct results
            yujian Jian Yu added a comment -

            I just verified again that the same failure also occurred on ext4 filesystem. After updating the LSB-VSX POSIX test suite to disable the terminal file test in pathconf.32, Lustre POSIX compliance testing passed on T.pathconf test set (baseline is ext4):

            Lustre branch: master
            Lustre build: http://build.whamcloud.com/job/lustre-master/766/
            https://maloo.whamcloud.com/test_sets/3df07240-e088-11e1-a388-52540035b04c

            yujian Jian Yu added a comment - I just verified again that the same failure also occurred on ext4 filesystem. After updating the LSB-VSX POSIX test suite to disable the terminal file test in pathconf.32, Lustre POSIX compliance testing passed on T.pathconf test set (baseline is ext4): Lustre branch: master Lustre build: http://build.whamcloud.com/job/lustre-master/766/ https://maloo.whamcloud.com/test_sets/3df07240-e088-11e1-a388-52540035b04c
            pjones Peter Jones added a comment -

            Actually, Yujian is working on a fix.

            pjones Peter Jones added a comment - Actually, Yujian is working on a fix.

            This item needs to be added to the release notes.

            jlevi Jodi Levi (Inactive) added a comment - This item needs to be added to the release notes.
            yujian Jian Yu added a comment -

            The unlink.6 and write.15 failures are LU-380. The readdir.6 failure is LU-379. This ticket is used to track pathconf.32 failure.

            I found that while formatting a partition/disk as ext4 filesystem and run POSIX test suite on it to generate baseline, the pathconf.32 test also failed. However, if a loop device is used to be formatted as ext4, the test passed. So, the issue is related to the baseline, not Lustre.

            yujian Jian Yu added a comment - The unlink.6 and write.15 failures are LU-380 . The readdir.6 failure is LU-379 . This ticket is used to track pathconf.32 failure. I found that while formatting a partition/disk as ext4 filesystem and run POSIX test suite on it to generate baseline, the pathconf.32 test also failed. However, if a loop device is used to be formatted as ext4, the test passed. So, the issue is related to the baseline, not Lustre.

            Since Lustre never pass all the POSIX test cases. We need to check whether above failures are new ones or known issues.

            yong.fan nasf (Inactive) added a comment - Since Lustre never pass all the POSIX test cases. We need to check whether above failures are new ones or known issues.

            Yu Jian,
            Can you please take a look at this one?

            jlevi Jodi Levi (Inactive) added a comment - Yu Jian, Can you please take a look at this one?
            mdiep Minh Diep added a comment - - edited

            I have reproduced this manually. This seems like Posix issue
            https://maloo.whamcloud.com/test_sets/2ed77eb4-d077-11e1-aa18-52540035b04c

            mdiep Minh Diep added a comment - - edited I have reproduced this manually. This seems like Posix issue https://maloo.whamcloud.com/test_sets/2ed77eb4-d077-11e1-aa18-52540035b04c
            mdiep Minh Diep added a comment -

            I'll try this manually to see if it makes a difference. This is not due to testsuite but posix detect some issue.

            mdiep Minh Diep added a comment - I'll try this manually to see if it makes a difference. This is not due to testsuite but posix detect some issue.
            sarah Sarah Liu added a comment -

            another failure: https://maloo.whamcloud.com/test_sets/aeb9eac8-c57c-11e1-ab7d-52540035b04c
            From the test report there are 4 failed tests, 3 of them are known issues listed in LU-380 and LU-379 while the pathconf.32 is a new one.

            14:21:04:Test Name: pathconf.32 Failed
            14:21:04:
            14:21:04: Test Description:
            14:21:04:When path does not refer to a directory, then a call to pathconf(path,
            14:21:04:_PC_NAME_MAX) either returns a value of -1 and sets errno to [EINVAL],
            14:21:04:or returns a valid value for

            {NAME_MAX}

            .
            14:21:04:Posix Ref: Component PATHCONF Assertion 5.7.1.4-52(A)
            14:21:04:
            14:21:04: Test Strategy:
            14:21:04:FOR a regular file, a FIFO (when not testing XNFS) and a terminal file
            14:21:04:(when not testing XNFS)
            14:21:04: OBTAIN the system value associated with pathconf() and _PC_NAME_MAX
            14:21:04: VERIFY that pathconf() returned successfully or pathconf() returned
            14:21:04: -1 and errno was set to EINVAL
            14:21:04:
            14:21:04: Test Information:
            14:21:04:pathconf("/dev/pts/0", _PC_NAME_MAX) did not give correct results
            14:21:04:expected: return value == -1, errno == 22 (EINVAL)
            14:21:04: OR: return value != -1
            14:21:04:observed: return value == -1, errno == 2 (ENOENT)

            sarah Sarah Liu added a comment - another failure: https://maloo.whamcloud.com/test_sets/aeb9eac8-c57c-11e1-ab7d-52540035b04c From the test report there are 4 failed tests, 3 of them are known issues listed in LU-380 and LU-379 while the pathconf.32 is a new one. 14:21:04:Test Name: pathconf.32 Failed 14:21:04: 14:21:04: Test Description: 14:21:04:When path does not refer to a directory, then a call to pathconf(path, 14:21:04:_PC_NAME_MAX) either returns a value of -1 and sets errno to [EINVAL] , 14:21:04:or returns a valid value for {NAME_MAX} . 14:21:04:Posix Ref: Component PATHCONF Assertion 5.7.1.4-52(A) 14:21:04: 14:21:04: Test Strategy: 14:21:04:FOR a regular file, a FIFO (when not testing XNFS) and a terminal file 14:21:04:(when not testing XNFS) 14:21:04: OBTAIN the system value associated with pathconf() and _PC_NAME_MAX 14:21:04: VERIFY that pathconf() returned successfully or pathconf() returned 14:21:04: -1 and errno was set to EINVAL 14:21:04: 14:21:04: Test Information: 14:21:04:pathconf("/dev/pts/0", _PC_NAME_MAX) did not give correct results 14:21:04:expected: return value == -1, errno == 22 (EINVAL) 14:21:04: OR: return value != -1 14:21:04:observed: return value == -1, errno == 2 (ENOENT)

            People

              yujian Jian Yu
              maloo Maloo
              Votes:
              0 Vote for this issue
              Watchers:
              5 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: