Hi, Mianquan, On Wed, Feb 25, 2026 at 07:36:20PM +0800, hongmianquan wrote: > Use a hashtable to store cpr fds to reduce the time > consumption of `cpr_find_fd` in scenarios with a large > number of fds, which relies on the migration support for > GHashTable implemented earlier.
A few things to mention. Firstly, please remember to copy reviewers on the CPR side when you change something for CPR, especially the ABI: $ ./scripts/get_maintainer.pl -f migration/cpr.c Peter Xu <[email protected]> (reviewer:CheckPoint and Re...) Fabiano Rosas <[email protected]> (reviewer:CheckPoint and Re...) Mark Kanda <[email protected]> (reviewer:CheckPoint and Re...) Ben Chaney <[email protected]> (reviewer:CheckPoint and Re...) [email protected] (open list:All patches CC here) Then, please attach a cover letter if you have more than one patch in a series, including information to introduce your series. That should include some explanation on why the lookup can be slow and how slow it is. It should also cover how it improves after the change. Regarding to "change it to hash" side of thing, it looks fine, but could you explain why do you need to migrate a hash? Can you still migrate as a list, then reconstruct that into a hash? It's fine if we want to provide an API to migrate a hash, but I want to make sure it's required. Also I want to see if it's a must to change the ABI. Thanks, -- Peter Xu
