Github messages for voidlinux
 help / color / mirror / Atom feed
From: q66 <q66@users.noreply.github.com>
To: ml@inbox.vuxu.org
Subject: Re: [PR PATCH] [Merged]: drop slub_debug/page_poison from bootloaders
Date: Fri, 05 Mar 2021 00:33:43 +0100	[thread overview]
Message-ID: <20210304233343.BC8wBZae6FWC1wit_weaw8LOwEA3popKU-x9mKQxnmM@z> (raw)
In-Reply-To: <gh-mailinglist-notifications-41a7ca26-5023-4802-975b-f1789d68868e-void-packages-29142@inbox.vuxu.org>

[-- Attachment #1: Type: text/plain, Size: 1933 bytes --]

There's a merged pull request on the void-packages repository

drop slub_debug/page_poison from bootloaders
https://github.com/void-linux/void-packages/pull/29142

Description:
These are kernel hardening options introduced into Void in 2016. On most modern hardware they have a small performance hit, on older hardware they can have a significant performance hit. In kernel 5.3, new options were introduced, `init_on_alloc` and `init_on_free`. There are also kernel options to make these default.

The new defaults have been toggled in kernels 5.4, 5.10 and 5.11 in bd0d33eec0bd774d8cc62e32689b6fbfda517179. I explicitly did not enable `init_on_free` since that has a much bigger performance penalty, even on current-day hardware.

The only "drawback" of these changes is that people using old kernels will not get these options anymore. OTOH, the new behavior is much better in other ways (it has less of an impact, it's enabled in kernels rather than default configs so any changes to it will propagate to all users unconditionally, etc). And also, there isn't any other mainstream distribution that was using these options in the first place, for a reason, as far as I can tell.

People on kernels 4.4, 4.9, 4.14 and 4.19 who really care can update it themselves. But since older kernels tend to be associated with older hardware, they'll probably want to keep the options off in the first place.

What troubles me more is that even if I merge this PR, it will not propagate to existing users who have modified their `/etc/default/grub`, so they will need to change those files themselves. The new default kernel options are not incompatible with the old options, but people will likely want to remove them anyway. So we should probably put information about this in the documentation, too.

[ci skip]

Anyway, waiting with merging this until the affected kernels have been rebuilt (i.e. for their next version bump).

      parent reply	other threads:[~2021-03-04 23:33 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-02-28 23:14 [PR PATCH] " q66
2021-03-01  8:42 ` fosslinux
2021-03-04 23:30 ` [PR PATCH] [Updated] " q66
2021-03-04 23:33 ` q66 [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20210304233343.BC8wBZae6FWC1wit_weaw8LOwEA3popKU-x9mKQxnmM@z \
    --to=q66@users.noreply.github.com \
    --cc=ml@inbox.vuxu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).