On Wed, Aug 28, 2024 at 5:02 AM Kinga Stefaniuk <kinga.stefaniuk@xxxxxxxxx> wrote: > > Patch is caused by changes in mdadm - now mdmonitor is relying on udev > events instead of pooling changes on /proc/mdstat file. This change > gave possibility to send events to the userspace. > Now, more events will be read by mdmonitor, but mechanism of getting > details of the event will remain the same - all devices have to be > scanned on new event and their parameters read. It can be changed by > adding environmental data, which will fully describe this uevent, > as Kuai suggested. This change needs to be analyzed and planned, > because it requires a lot of work in md and design changes in > mdmonitor - to stop scanning all of the devices on new event and read > uevent details directly from udev. Mariusz will take care of this topic, > and follow up on this later. > Paul, the test suite which was used is internal and checks if mdmonitor > has noticed expected events. For now there is no plan for adding test, > which checking events to mdadm. > > Kinga Stefaniuk (1): > md: generate CHANGE uevents for md device It appears this patch doesn't apply on the md-6.12 branch. Please rebase and resend the patch. Thanks, Song > > drivers/md/md.c | 148 +++++++++++++++++++++++++++----------------- > drivers/md/md.h | 3 +- > drivers/md/raid10.c | 2 +- > drivers/md/raid5.c | 2 +- > 4 files changed, 94 insertions(+), 61 deletions(-) > > -- > 2.43.0 >
