HACK: Add a default for my_timer_cycles()
This one-line change (plus moving some prototypes) is a workaround for https://bugs.debian.org/1029163
A proper fix should revisit all this 15-20 year old timer code for all architectures, my patch instead aims at ensuring not causing any changes/regressions for other architectures.
Merge request reports
Activity
requested review from @otto
Thanks Adrian!
Could you please consider submitting this upstream at https://github.com/mariadb/server as well so we get upstream feedback and have a path to get this fixed everywhere?
I know the title says 'hack' but still it shows what works (I assume you tested this and tests passed).
mentioned in commit 74e711fe
Test build at https://launchpad.net/
mysql-ubuntu/+archive/ubuntu/mariadb-10.11/+build/25531056/+files/buildlog_ubuntu-lunar-armhf.mariadb_1%3A10.11.1-2ubuntu23.04.1~1674979070.4c184598.dev.otto_BUILDING.txt.gz confirms that this patch fixed Bug#1029163.I also tested the patch from https://patch-diff.githubusercontent.com/raw/MariaDB/server/pull/2448.patch and seems to work based on test build at https://launchpad.net/
mysql-ubuntu/+archive/ubuntu/mariadb-10.11/+build/25533120/+files/buildlog_ubuntu-lunar-armhf.mariadb_1%3A10.11.1-2ubuntu23.04.1~1675043290.397d3975.dev.otto_BUILDING.txt.gzI think I will go with the upstream patch as it is more complete and using it in Debian will increase the likelihood of it getting validated and merged upstream. Your initial fix was however very valuable as it sparked upstream to dive deeper into the issue. Thanks!
mentioned in commit 340b190a
Closed via 340b190a