Bug #17236

Enable the init_on_alloc=1 and init_on_free=1 Linux options

Added by intrigeri 2019-11-15 09:26:15 . Updated 2020-01-06 07:56:55 .

Status:
Resolved
Priority:
Normal
Assignee:
Category:
Target version:
Tails_4.2
Start date:
Due date:
% Done:

0%

Feature Branch:
feature/17236-heap-variable-initialization+force-all-tests
Type of work:
Code
Blueprint:

Starter:
Affected tool:
Deliverable for:

Description

We should normally have done this when upgrading to Linux 5.3, but Kees’ blog post was not published yet back then.

Subtasks

Related issues

Related to Tails - Bug #17124: Install Linux 5.3 from sid Resolved
Related to Tails - Bug #17117: Upgrade to Linux 5.3 Resolved
Blocks Tails - Feature #16209: Core work: Foundations Team Confirmed

History

#1 Updated by intrigeri 2019-11-15 09:26:38

  • related to Bug #17124: Install Linux 5.3 from sid added

#2 Updated by intrigeri 2019-11-15 09:26:49

  • related to Bug #17177: Greeter has no option to 'Show Passphrase' anymore added

#3 Updated by intrigeri 2019-11-15 09:26:54

  • related to Bug #17117: Upgrade to Linux 5.3 added

#4 Updated by intrigeri 2019-11-15 09:27:00

  • related to deleted (Bug #17177: Greeter has no option to 'Show Passphrase' anymore)

#5 Updated by intrigeri 2019-11-15 09:27:09

#6 Updated by intrigeri 2019-11-16 08:26:05

  • Status changed from Confirmed to In Progress

Applied in changeset commit:tails|6aec55afe525b2a941f12a95dd5c32b8222b83f8.

#7 Updated by intrigeri 2019-11-16 08:27:07

  • Feature Branch set to feature/17236-heap-variable-initialization+force-all-tests

#8 Updated by intrigeri 2019-11-16 18:10:36

No performance nor robustness regression spotted on my local Jenkins.

#9 Updated by intrigeri 2019-11-17 05:09:54

The first 2 test suite runs on lizard each expose at least one occurrence of “Remote shell seems to be down” (smells bad to me), but no performance regression.

#10 Updated by intrigeri 2019-11-27 09:04:39

  • Target version changed from Tails_4.1 to Tails_4.2

This can totally wait a bit and it’s now a bit too late in the 4.1 cycle for me to feel comfortable including this.

#11 Updated by intrigeri 2019-12-28 15:16:52

  • Subject changed from Consider enabling the init_on_alloc=1 and init_on_free=1 Linux options to Enable the init_on_alloc=1 and init_on_free=1 Linux options
  • Status changed from In Progress to Needs Validation
  • Assignee deleted (intrigeri)
  • Type of work changed from Test to Code

The 7-25% performance hit with init_on_free=1 is not noticeable on Jenkins: recent test suite runs on this branch took about the same time as devel. I think that’s because we already have page poisoning and SLUB poisoning enabled, so the actual impact of init_on_free=1 is much more limited than it would otherwise be, as https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=6471384af2a6530696fc0203bafe4de41a23c9ef says:

> If either SLUB poisoning or page poisoning is enabled, those options take
> precedence over init_on_alloc and init_on_free: initialization is only
> applied to unpoisoned allocations.

Jenkins test suite results are not more brittle than on the devel branch.

#12 Updated by intrigeri 2020-01-05 19:01:27

  • Status changed from Needs Validation to In Progress

Applied in changeset commit:tails|360a8abc3ae2c2c8ffa0cf93ec5c617c77dbacb2.

#13 Updated by intrigeri 2020-01-06 07:56:55

  • Status changed from In Progress to Resolved

Someone (segfault, I guess :) merged a rebased version of this branch into stable ⇒ closing.