Page MenuHomeFreeBSD
Differential D32006

Convert consumers to vm_page_alloc_noobj_contig()
ClosedPublic
Actions

Authored by markj on Sep 17 2021, 2:17 PM.
Tags
None
Referenced Files
Unknown Object (File)
Sat, Nov 23, 11:20 PM
Unknown Object (File)
Fri, Nov 22, 12:30 PM
Unknown Object (File)
Thu, Nov 14, 5:03 AM
Unknown Object (File)
Sat, Nov 9, 8:38 PM
Unknown Object (File)
Thu, Nov 7, 4:59 AM
Unknown Object (File)
Tue, Nov 5, 10:22 AM
Unknown Object (File)
Oct 23 2024, 6:52 AM
Unknown Object (File)
Oct 17 2024, 2:24 PM
View All Files
Subscribers
andrew
emaste
• hselasky
imp
jhibbits
linuxkpi

Details

Reviewers
alc
kib
• hselasky
Commits
rG84c3922243a7: Convert consumers to vm_page_alloc_noobj_contig()
Summary

Remove now-unneeded page zeroing. No functional change intended.

Diff Detail

Repository
rG FreeBSD src repository
Lint
Lint Not Applicable
Unit
Tests Not Applicable

Event Timeline

markj created this revision.Sep 17 2021, 2:17 PM
Herald added subscribers: linuxkpi, emaste, jhibbits and 2 others. · View Herald TranscriptSep 17 2021, 2:17 PM
markj requested review of this revision.Sep 17 2021, 2:17 PM
Harbormaster completed remote builds in B41568: Diff 95291.Sep 17 2021, 2:17 PM
markj added a parent revision: D32005: Introduce vm_page_alloc_noobj_contig().Sep 17 2021, 2:17 PM
• hselasky accepted this revision.Sep 17 2021, 2:30 PM
• hselasky added a subscriber: • hselasky.

Looks good, assuming zero handling is preserved.

This revision is now accepted and ready to land.Sep 17 2021, 2:30 PM
alc added inline comments.Sep 17 2021, 5:57 PM
sys/amd64/amd64/mp_machdep.c
337

I observe that you eliminated VM_ALLOC_NOBUSY here but not elsewhere. Why?

markj added inline comments.Sep 17 2021, 6:22 PM
sys/amd64/amd64/mp_machdep.c
337

No good reason. I think it makes sense to simply drop the VM_ALLOC_NOBUSY flag from all callers, as it is already implied by VM_ALLOC_NOOBJ.

alc added inline comments.Sep 17 2021, 6:47 PM
sys/amd64/amd64/mp_machdep.c
337

I agree.

markj updated this revision to Diff 95315.Sep 17 2021, 7:08 PM

Drop unused VM_ALLOC_NOBUSY from consumers.

This revision now requires review to proceed.Sep 17 2021, 7:08 PM
Harbormaster completed remote builds in B41574: Diff 95315.Sep 17 2021, 7:08 PM
markj added a child revision: D32033: vm_page: Stop handling VM_ALLOC_NOOBJ in vm_page_alloc_domain_after().Oct 1 2021, 9:56 PM
markj updated this revision to Diff 96189.Oct 4 2021, 12:03 AM

Drop uses of VM_ALLOC_NORMAL from vm_page_alloc_noobj_contig() and
vm_page_alloc_noobj_contig_domain() callers.

Harbormaster completed remote builds in B41938: Diff 96189.Oct 4 2021, 12:03 AM
kib accepted this revision.Oct 4 2021, 1:47 AM
kib added inline comments.
sys/arm/nvidia/drm2/tegra_bo.c
105–106

The variable becomes somewhat pointless. You can directly pass _WIRED|_ALLOC_ZERO to vm_page_alloc_noobj_contig(), and vm_page_reclaim_contig() is fine with 0 instead pflags.

This revision is now accepted and ready to land.Oct 4 2021, 1:47 AM
markj updated this revision to Diff 96192.Oct 4 2021, 1:59 AM
markj marked an inline comment as done.

Drop an unneeded local variable.

This revision now requires review to proceed.Oct 4 2021, 1:59 AM
Harbormaster completed remote builds in B41941: Diff 96192.Oct 4 2021, 1:59 AM
alc accepted this revision.Oct 17 2021, 1:04 AM
alc added inline comments.
sys/compat/linuxkpi/common/src/linux_page.c
107–109

Is PAGE_SIZE actually correct? I would guess that Linux guarantees that an order n allocation is aligned to an order n address. That said, at the moment, this is probably not an issue because of the way that vm_phys does contig allocations.

sys/powerpc/aim/mmu_radix.c
3576

Does VM_ALLOC_ZERO actually make sense here?

This revision is now accepted and ready to land.Oct 17 2021, 1:04 AM
markj added inline comments.Oct 17 2021, 7:43 PM
sys/powerpc/aim/mmu_radix.c
3576

It doesn't seem so. I think both of these XXX comments can be resolved by ORing malloc2vm_flags(flags) into req.

markj added inline comments.Oct 17 2021, 9:14 PM
sys/compat/linuxkpi/common/src/linux_page.c
107–109

I can't find any Linux documentation or code that definitively answers this. Looking at the alloc_pages() implementation, it does seem to return self-aligned runs in practice, but I can't tell if this is a guarantee of the interface. alloc_contig_pages() has a comment stating that self-alignment is guaranteed if the size of the run is a power of 2, but this function is not used much.

markj updated this revision to Diff 97021.Oct 17 2021, 9:15 PM

Use malloc2vm_flags() in radix_pgd_import().

This revision now requires review to proceed.Oct 17 2021, 9:15 PM
Harbormaster completed remote builds in B42205: Diff 97021.Oct 17 2021, 9:15 PM
This revision was not accepted when it landed; it landed in state Needs Review.Oct 27 2021, 3:00 PM
Closed by commit rG84c3922243a7: Convert consumers to vm_page_alloc_noobj_contig() (authored by markj). · Explain Why
This revision was automatically updated to reflect the committed changes.
markj added a commit: rG84c3922243a7: Convert consumers to vm_page_alloc_noobj_contig().

Revision Contents
Changeset List

PathSize
sys/
amd64/
amd64/
3 lines
28 lines
arm/
nvidia/
drm2/
12 lines
compat/
linuxkpi/
common/
src/
17 lines
dev/
drm2/
ttm/
11 lines
6 lines
kern/
7 lines
powerpc/
aim/
8 lines
9 lines
riscv/
riscv/
2 lines
vm/
19 lines
28 lines

Diff 97551

View Options

sys/amd64/amd64/mp_machdep.c

Loading...
View Options

sys/amd64/amd64/pmap.c

Loading...
View Options

sys/arm/nvidia/drm2/tegra_bo.c

Loading...
View Options

sys/compat/linuxkpi/common/src/linux_page.c

Loading...
View Options

sys/dev/drm2/ttm/ttm_bo.c

Loading...
View Options

sys/dev/drm2/ttm/ttm_page_alloc.c

Loading...
View Options

sys/kern/uipc_ktls.c

Loading...
View Options

sys/powerpc/aim/mmu_radix.c

Loading...
View Options

sys/powerpc/aim/slb.c

Loading...
View Options

sys/riscv/riscv/pmap.c

Loading...
View Options

sys/vm/uma_core.c

Loading...
View Options

sys/vm/vm_page.c

Loading...