https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=211361
--- Comment #9 from Nathan Whitehorn <nwhiteh...@freebsd.org> --- I would like to see this done in the kernel preferred IO alignment reporting rather than the installer. This limits the amount of magic in the installer and fixes similar issues in the rest of the system. There are three options to solve this problem, in general: 1. The kernel makes up a number. The installer, gpart, and sade align to that number. It may become stale if new, weird hardware appears. 2. The installer makes up a number and aligns to it. sade is the installer, so it follows along. gpart doesn't and creates misaligned partitions by default on systems that currently have problems (as do other disk tools, like graid). sade and gpart end up with different behaviors. The number may become stale in the same way. 3. The status quo. The kernel tries to figure out the right number, but it is too small on some hardware. The installer, gpart, and sade align to that number. It is already wrong in some cases. #1 seems strictly better than #2 here. The numbers in both cases may become stale, but you don't have to make the same fix in a bunch of places in #1 or end up in a situation where gpart and the curses version of gpart (sade) have different behavior. It also makes ZFS and graid and whatever follow along instead of picking too-small IO chunks by default on affected systems. The ZFS issue in particular is probably at least as serious as the alignment of UFS boot partitions. -- You are receiving this mail because: You are the assignee for the bug. _______________________________________________ freebsd-bugs@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-bugs To unsubscribe, send any mail to "freebsd-bugs-unsubscr...@freebsd.org"
