| commit | a0e8fc3cfb9c75fe2ac931ec8a994d526ec6ae44 | [log] [tgz] |
|---|---|---|
| author | Adrian Danis <adanis@fuchsia.infra.roller.google.com> | Wed Jun 12 02:42:49 2024 +0000 |
| committer | Copybara-Service <copybara-worker@google.com> | Tue Jun 11 19:44:20 2024 -0700 |
| tree | 992084a59b5ee4d2258edfea0336fab0c73e4dc2 | |
| parent | 18a3cfccc59e1d16ee7e47bb70ae1601518c8682 [diff] |
[roll] Roll fuchsia [kernel][vm] Fix DirtyPage allocations Fixes two related latent bugs in the DirtyPagesLocked zero page allocation logic. First bug is theoretically benign, where even if all pages are allocated by the call to pmm_alloc_pages, since the count is not updated we still allocate each page individually, ultimately causing double the required allocations. Second bug occurs if the passed in allocation list already has enough items to satisfy all the zero pages, since the number of zero pages is pre truncated this will cause us to completely skip the the process of then actually inserting and marking the pages dirty. These two bugs cannot manifest today as pager backed VMOs are, also incorrectly, not marked as being able to have delayed allocations. As the allocations cannot delay, the second set of individual page allocations will never need to generate a page request, and so we will never have to retry the function with too many items already in the alloc_list. Given this a test cannot be written today to validate this bug. However, the existing tests will fail with the current code if pager backed VMOs gain PMM_ALLOC_FLAG_CAN_WAIT, and so if/when that happens that will count as a regression test. Original-Reviewed-on: https://fuchsia-review.googlesource.com/c/fuchsia/+/1064112 Original-Revision: 0eabe588d7537a72b8dc9608249fd740bc8e89fe GitOrigin-RevId: d70ea235df9ff56532d841f0b7683945f37a1f4d Change-Id: I59702929596520fa222bb5a74c88f02c2e2c57f8
This repository contains Fuchsia's Global Integration manifest files.
All changes should be made to the internal version of this repository. Our infrastructure automatically updates this version when the internal one changes.
Currently all changes must be made by a Google employee. Non-Google employees wishing to make a change can ask for assistance via the IRC channel #fuchsia on Freenode.
First install Jiri.
Next run:
$ jiri init $ jiri import minimal https://fuchsia.googlesource.com/integration $ jiri update
Third party projects should have their own subdirectory in ./third_party.