Page MenuHomeFreeBSD
Differential D31958

vm_phys: do not ignore phys_avail[] segments that do not fit completely into vm_phys segments
ClosedPublic
Actions

Authored by kib on Sep 14 2021, 12:34 PM.
Tags
None
Referenced Files
Unknown Object (File)
Thu, Jan 23, 9:36 PM
Unknown Object (File)
Thu, Jan 23, 9:28 AM
Unknown Object (File)
Tue, Jan 14, 1:17 AM
Unknown Object (File)
Dec 8 2024, 8:57 AM
Unknown Object (File)
Dec 6 2024, 1:45 AM
Unknown Object (File)
Dec 4 2024, 9:52 PM
Unknown Object (File)
Nov 19 2024, 3:41 AM
Unknown Object (File)
Nov 19 2024, 3:34 AM
View All 87 Files
Subscribers
cy
madpilot

Details

Reviewers
alc
markj
andrew
madpilot
Commits
rG181bfb42fd01: vm_phys: do not ignore phys_avail[] segments that do not fit completely into…
Summary

If phys_avail[] segment only intersect with some vm_phys segment, add pages from it to the free list that belong to the given vm_phys_seg, instead of dropping them.

The vm_phys segments are generally result of subdivision of phys_avail segments, for instance DMA32 or LOWMEM boundaries split them. On amd64, after UEFI in-place kernel activation (copy_staging disable) was enabled, we typically have a large phys_avail[] segment below 4G which crosses LOWMEM (1M) boundary. With the current way of requiring phys_avail[] fully fit into vm_phys_seg, this memory was ignored.

Diff Detail

Repository
rG FreeBSD src repository
Lint
Lint Skipped
Unit
Tests Skipped

Event Timeline

kib requested review of this revision.Sep 14 2021, 12:34 PM
kib created this revision.
madpilot accepted this revision.Sep 14 2021, 1:42 PM

I confirm that with this patch applied my system is now able to see all its RAM.

Thank you!

This revision is now accepted and ready to land.Sep 14 2021, 1:42 PM
markj added inline comments.Sep 14 2021, 2:49 PM
sys/vm/vm_page.c
778

This comment no longer holds.

kib updated this revision to Diff 95152.Sep 14 2021, 3:01 PM
kib marked an inline comment as done.

Update comment.

This revision now requires review to proceed.Sep 14 2021, 3:01 PM
cy added a subscriber: cy.Sep 14 2021, 5:45 PM
alc added inline comments.Sep 15 2021, 6:44 AM
sys/vm/vm_page.c
777–778
kib updated this revision to Diff 95189.Sep 15 2021, 7:30 AM
kib marked an inline comment as done.

Reword comment, by Alan.

markj added inline comments.Sep 15 2021, 1:38 PM
sys/vm/vm_page.c
803

Isn't it possible for some of the excluded range of the phys_seg to be present in a different phys_avail[] region?

kib added inline comments.Sep 15 2021, 1:43 PM
sys/vm/vm_page.c
803

Do you mean that this break statement should be removed?

803

And the condition seg->start< phys_avail[i] replaced by seg->end < phys_avail[i] perhaps

kib updated this revision to Diff 95194.Sep 15 2021, 2:00 PM

Walk over all phys_avail[] segments which overlap with our vm_phys_seg when gathering the vm_phys_seg' pages

markj accepted this revision.Sep 15 2021, 8:02 PM
This revision is now accepted and ready to land.Sep 15 2021, 8:02 PM
Closed by commit rG181bfb42fd01: vm_phys: do not ignore phys_avail[] segments that do not fit completely into… (authored by kib). · Explain WhySep 16 2021, 5:11 PM
This revision was automatically updated to reflect the committed changes.
kib added a commit: rG181bfb42fd01: vm_phys: do not ignore phys_avail[] segments that do not fit completely into….
alc added inline comments.Sep 17 2021, 6:12 PM
sys/vm/vm_page.c
783

Why not test for equality here? It seems pointless to allow execution to continue if they are equal.

787

Shouldn't this be startp - seg->start?

Revision Contents
Changeset List

PathSize
sys/
vm/
12 lines

Diff 95149

View Options

sys/vm/vm_page.c

Loading...