Uploaded image for project: 'MariaDB Server'
  1. MariaDB Server
  2. MDEV-13196

Remove handler::store_lock()

Details

    Description

      There's potential deadlock between "thread killing another thread that has open InnoDB HANDLER or delayed insert thread performing insert into InnoDB table" and "InnoDB service threads like srv_monitor_thread that calls thd_get_error_context_description".

      I doubt MariaDB is affected, because it does try_lock in thd_get_error_context_description, see 6e9a48b67fceab17089ca4cd1406e302386a601b.

      But anyway, since InnoDB doesn't use thr_lock anymore, it is probably a good idea to get rid of ha_innobase::store_lock() by moving some code to external_lock. Note: compared to store_lock external_lock can be called for temporary tables too.

      Attachments

        Issue Links

          relates to

          Bug - A problem which impairs or prevents the functions of the product. MDEV-12070 Introduce thd_query_safe() from MySQL 5.7

          • Major - Major loss of function.
          • Closed

          Bug - A problem which impairs or prevents the functions of the product. MDEV-11896 thd_get_error_context_description race condition

          • Major - Major loss of function.
          • Closed

          Bug - A problem which impairs or prevents the functions of the product. MDEV-13983 Mariadb becomes unresponsive

          • Major - Major loss of function.
          • Closed

          Bug - A problem which impairs or prevents the functions of the product. MDEV-16467 MariaDB crashes because of "long semaphore wait"after migrating from 10.1 to 10.3

          • Critical - Crashes, loss of data, severe memory leak.
          • Closed

          Activity

            marko Marko Mäkelä added a comment -

            As far as I can tell, MDEV-11896 changed the potential deadlock to a real deadlock. It was semi-accidentally fixed later, as noted in MDEV-13983.

            I think that it would be very welcome to remove ha_innobase::store_lock() and to simplify the locking logic. I wonder why that was not done when MDL was introduced.

            marko Marko Mäkelä added a comment - As far as I can tell, MDEV-11896 changed the potential deadlock to a real deadlock. It was semi-accidentally fixed later, as noted in MDEV-13983 . I think that it would be very welcome to remove ha_innobase::store_lock() and to simplify the locking logic. I wonder why that was not done when MDL was introduced.

            People

              serg Sergei Golubchik
              svoj Sergey Vojtovich
              Votes:
              0 Vote for this issue
              Watchers:
              4 Start watching this issue

              Dates

                Created:
                Updated:

                Git Integration

                  Error rendering 'com.xiplink.jira.git.jira_git_plugin:git-issue-webpanel'. Please contact your Jira administrators.