vm_radix_init: use initializer
ClosedPublic
Actions

Authored by dougm on Jul 11 2023, 10:08 AM.

Details

Reviewers

alc
andrew
manu
kib
markj

Commits

rG3e04ae433fdc: vm_radix_init: use initializer

Summary

Several vm_radix tries are not initialized with vm_radix_init. That works, for now, since static initialization zeroes the root field anyway, but if initialization changes, these tries will fail. Add missing initializer calls.

Test Plan

A kernel that checked for proper initialization was built and run on amd64 and all the vm_radix tries that were used were verified as properly initialized.

Diff Detail

Lint

Lint Skipped

Unit

Tests Skipped

Event Timeline

dougm created this revision.Jul 11 2023, 10:08 AM

Herald added a reviewer: andrew. · View Herald TranscriptJul 11 2023, 10:08 AM

Herald added a reviewer: andrew. · View Herald Transcript

Herald added a reviewer: manu. · View Herald Transcript

Herald added subscribers: riscv, emaste. · View Herald Transcript

dougm requested review of this revision.Jul 11 2023, 10:08 AM

dougm added a reviewer: kib.Jul 13 2023, 6:33 PM

dougm added a reviewer: markj.Jul 13 2023, 11:08 PM

markj accepted this revision.Jul 14 2023, 12:48 AM

This revision is now accepted and ready to land.Jul 14 2023, 12:48 AM

kib accepted this revision.Jul 14 2023, 3:41 AM

alc added inline comments.Jul 14 2023, 4:32 AM

sys/vm/vm_object.c
297	If this were an ordinary object, then this would happen before the call to _vm_object_allocate().

Move lines after written and verbal direction from alc.

This revision now requires review to proceed.Jul 14 2023, 5:00 AM

alc added inline comments.Jul 14 2023, 5:10 AM

sys/arm64/arm64/pmap.c
921–927	Ugh, in reviewing this, I observed that neither arm64 nor riscv ever TAILQ_INIT() the pm_pvchunk field. I believe that the only reason that they don't crash is that their first operations on the field are TAILQ_FIRST() and TAILQ_INSERT_HEAD().