cache: add symlink support to lockless lookup

Reviewed by: kib (previous version)
Tested by: pho (previous version)
Differential Revision: https://reviews.freebsd.org/D27488

vfs: extend vfs:namei:lookup:return probe with nameidata

vfs: add NDFREE_NOTHING and convert several NDFREE_PNBUF callers

Check the comment above the routine for reasoning.

vfs: add FAILIFEXISTS flag

Both FreeBSD and Linux mkdir -p walk the tree up ignoring any EEXIST on
the way and both are used a lot when building respective kernels.

This poses a problem as spurious

vfs: add FAILIFEXISTS flag

Both FreeBSD and Linux mkdir -p walk the tree up ignoring any EEXIST on
the way and both are used a lot when building respective kernels.

This poses a problem as spurious locking avoidably interferes with
concurrent operations like getdirentries on affected directories.

Work around the problem by adding FAILIFEXISTS flag. In case of lockless
lookup this manages to avoid any work to begin with, there is no speed
up for the locked case but perhaps this can be augmented later on.

For simplicity the only supported semantics are as used by mkdir.

Reviewed by: kib (previous version)
Differential Revision: https://reviews.freebsd.org/D27789

show more ...

vfs: add cleanup on error missed in r368375

Noted by: jrtc27

vfs: factor buffer allocation/copyin out of namei

Remove the 'wantparent' variable, unused since r145004.

Reviewed by: kib
MFC after: 2 weeks
Sponsored by: NetApp, Inc.
Sponsored by: Klara, Inc.
Differential Revision: https://reviews.freebsd.org/D2

Remove the 'wantparent' variable, unused since r145004.

Reviewed by: kib
MFC after: 2 weeks
Sponsored by: NetApp, Inc.
Sponsored by: Klara, Inc.
Differential Revision: https://reviews.freebsd.org/D27193

show more ...

vfs: change nt_zone into a malloc type

Elements are small in size and allocated for short periods.

vfs: add NAMEI_DBG_HADSTARTDIR handling lost in rewrite

Noted by: rpokala

vfs: add NDREINIT to facilitate repeated namei calls

struct nameidata mixes caller arguments, internal state and output, which
can be quite error prone.

Recent addition of valdiating ni_resflags un

vfs: add NDREINIT to facilitate repeated namei calls

struct nameidata mixes caller arguments, internal state and output, which
can be quite error prone.

Recent addition of valdiating ni_resflags uncovered a caller which could
repeatedly call namei, effectively operating on partially populated state.

Add bare minimium validation this does not happen. The real fix would
decouple aforementioned state.

Reported by: pho
Tested by: pho (different variant)

show more ...

cache: add missing NIRES_ABS handling

Do not allow to use O_BENEATH as an oracle.

Specifically, if lookup() returned any error and the topping directory
was not latched, which means that (non-existent) path did not returned
to the toppi

Do not allow to use O_BENEATH as an oracle.

Specifically, if lookup() returned any error and the topping directory
was not latched, which means that (non-existent) path did not returned
to the topping location, give ENOTCAPABLE a priority over the lookup()
error.

PR: 249960
Reviewed by: emaste, ngie
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D26695

show more ...

Add O_RESOLVE_BENEATH and AT_RESOLVE_BENEATH to mimic Linux' RESOLVE_BENEATH.

It is like O_BENEATH, but disables to walk out of the subtree rooted
in the starting directory. O_BENEATH does not care

Add O_RESOLVE_BENEATH and AT_RESOLVE_BENEATH to mimic Linux' RESOLVE_BENEATH.

It is like O_BENEATH, but disables to walk out of the subtree rooted
in the starting directory. O_BENEATH does not care if path walks out
if it returned.

Requested by: Dan Gohman <dev@sunfishcode.online>
PR: 248335
Reviewed by: markj
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D25886

show more ...

Change O_BENEATH to handle relative paths same as absolute.

Do not care if path walks out of the topping directory if it returns back.

Requested and reviewed by: markj
Tested by: pho
Sponsored by:

Change O_BENEATH to handle relative paths same as absolute.

Do not care if path walks out of the topping directory if it returns back.

Requested and reviewed by: markj
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D25886

show more ...

Only clear latch for BENEATH when we walk out of the startdir,

not unconditionally on any dotdot component.

Reviewed by: markj
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 1 week

Only clear latch for BENEATH when we walk out of the startdir,

not unconditionally on any dotdot component.

Reviewed by: markj
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D25886

show more ...

Add NIRES_STRICTREL.

Stop abusing internal namei flag NI_LCF_STRICTRELATIVE as indicator of
cap-restricted lookup. Add designated returned flag NIRES_STRICTREL
to inform kern_openat() that lookup w

Add NIRES_STRICTREL.

Stop abusing internal namei flag NI_LCF_STRICTRELATIVE as indicator of
cap-restricted lookup. Add designated returned flag NIRES_STRICTREL
to inform kern_openat() that lookup was restricted.

Reviewed by: markj
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D25886

show more ...

lookup: Track last lookup component if it is directory.

This makes open("/a/../a", O_BENEATH) with cwd == "/a" work.

Reviewed by: markj
Reported and tested by: pho
Sponsored by: The FreeBSD Foundat

lookup: Track last lookup component if it is directory.

This makes open("/a/../a", O_BENEATH) with cwd == "/a" work.

Reviewed by: markj
Reported and tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D25886

show more ...

Improve comment above nameicap_check_dotdot().

Explain why tracker is needed at all.

Reviewed by: markj
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision:

Improve comment above nameicap_check_dotdot().

Explain why tracker is needed at all.

Reviewed by: markj
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D25886

show more ...

Add a comment to clarify when and why cached names are deleted
during pathname lookup.

Reviewed by: kib
MFC after: 3 days
Sponsored by: Netflix

MFH

Sponsored by: Rubicon Communications, LLC (netgate.com)

vfs: validate ndp state after the lookup

The intent is to remove known-to-be-nops NDFREE calls after many lookups.

vfs: convert nameiop into an enum

While here change the field size from long to int and move it into the
gap next to cn_flags.

Shrinks struct componentname from 64 to 56 bytes on amd64.

vfs: assert that HASBUF is only set with SAVENAME or SAVESTART

as requested by the caller. The intent is to eradicate the mostly
spurious NDFREE_PNBUF calls.

vfs: add a work around for vp_crossmp bug to realpath

The actual bug is not yet addressed as it will get much easier after other
problems are addressed (most notably rename contract).

The only affe

vfs: add a work around for vp_crossmp bug to realpath

The actual bug is not yet addressed as it will get much easier after other
problems are addressed (most notably rename contract).

The only affected in-tree consumer is realpath. Everyone else happens to be
performing lookups within a mount point, having a side effect of ni_dvp being
set to mount point's root vnode in the worst case.

Reported by: pho

show more ...

Merge ^/head r364264 through r364278.