Clean up a couple of MD warts in vm_fault_populate():
ClosedPublic
Actions

Authored by jah on Mar 27 2021, 3:17 AM.

Details

Reviewers

kib
markj
bdragon

Group Reviewers

PowerPC
MIPS

Commits

rG8dc8feb53da0: Clean up a couple of MD warts in vm_fault_populate():

Summary

--Eliminate a big ifdef that encompassed all currently-supported
architectures except mips and powerpc32. This applied to the case
in which we've allocated a superpage but the pager-populated range
is insufficient for a superpage mapping. For platforms that don't
support superpages the check should be inexpensive as we shouldn't
get a superpage in the first place. Make the normal-page fallback
logic identical for all platforms and provide a simple implementation
of pmap_ps_enabled() for MIPS and Book-E/AIM32 powerpc.

--Apply the logic for handling pmap_enter() failure if a superpage
mapping can't be supported due to higher-level protection policy.
Use KERN_PROTECTION_FAILURE instead of KERN_FAILURE for this case,
and note Intel PKU on amd64 as the first example of such protection
policy.

Diff Detail

Repository

rS FreeBSD src repository - subversion

Lint

Lint Passed

Unit

No Test Coverage

Build Status

Buildable 38119
Build 35008: arc lint + arc unit

Event Timeline

jah created this revision.Mar 27 2021, 3:17 AM

Herald added subscribers: jhibbits, imp. · View Herald TranscriptMar 27 2021, 3:17 AM

jah requested review of this revision.Mar 27 2021, 3:17 AM

Harbormaster completed remote builds in B38103: Diff 86376.Mar 27 2021, 3:17 AM

jah added reviewers: kib, markj, PowerPC, MIPS.Mar 27 2021, 3:19 AM

jah added inline comments.Mar 27 2021, 3:23 AM

sys/powerpc/aim/mmu_oea.c
1128	For PowerPC I believe pmap_nomethod() dispatch should work for the purpose of this change without needing these added functions. In fact it seems that it already is for 64-bit Book E MMUs. But it seemed less fragile to provide the real dispatch entries since we already allocate them anyway.

kib added inline comments.Mar 27 2021, 11:14 AM

sys/amd64/amd64/pmap.c
7150	This is fine and makes things consistent with pmap_enter_largepage(), but now I see that there are some issues with KERN_PROTECTION_FAILURE. Since they are mine and pre-existing, I propose D29442.

markj mentioned this in D29442: vm_fault: handle KERN_PROTECTION_FAILURE.Mar 27 2021, 2:58 PM

markj accepted this revision.Mar 27 2021, 3:07 PM

This revision is now accepted and ready to land.Mar 27 2021, 3:07 PM

kib added inline comments.Mar 27 2021, 6:46 PM

sys/vm/vm_fault.c
560	I would say 'lower-level', since PKRU is somewhere below pmap. Or remove the 'higher-level' words.
576–577	IMO it would be cleaner to assert rv value right after pmap_enter(). Something like `MPASS(rv == KERN_SUCCESS \|\| (psind > 0 && rv == KERN_PROTECTION_FAILURE);`