d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
255,880 Commits 72 Branches 135 Tags 3.2 GiB
e8a5a1ad71
Commit Graph

428 Commits

Author SHA1 Message Date
Mateusz Guzik
8ddea0b127 cache: just assign ni_resflags = NIRES_ABS 
It is guaranteed to be 0 on entry.
 2021-01-08 13:57:10 +00:00
Mateusz Guzik
fee405e057 cache: stop checkpointing cn_flags 
They are only modified, if ever, for the last component.
 2021-01-07 23:29:52 +00:00
Mateusz Guzik
ac7715471c cache: stop checkpointing cn_nameptr 
For aborts cn_nameptr is the same as cn_pnbuf. For partial results
the same cn_nameptr is to be used.
 2021-01-07 23:29:38 +00:00
Mateusz Guzik
0f1fc3a31f cache: stop manipulating pathlen 
It is a copy-pasto from regular lookup. Add debug to ensure the result
is the same.
 2021-01-07 23:26:53 +00:00
Mateusz Guzik
f2b794e1e9 cache: unengrish the comment in previous commit 
Reported by:	rpokala, brd
 2021-01-06 23:46:05 +00:00
Mateusz Guzik
deabdc6868 cache: stop pre-checking seqc when starting the lookup 
Tested by:	pho
 2021-01-06 07:28:07 +00:00
Mateusz Guzik
71a6a0b545 cache: skip checking for spurious slashes if possible 
Tested by:	pho
 2021-01-06 07:28:06 +00:00
Mateusz Guzik
33f3e81df5 cache: combine fast path enabled status into one flag 
Tested by:	pho
 2021-01-06 07:28:06 +00:00
Mateusz Guzik
dbbbc07cc3 cache: split handling of 0 and non-0 error codes 
Tested by:	pho
 2021-01-06 07:07:24 +01:00
Mateusz Guzik
a1a8f8ada1 cache: deinline state handling 
The intent is to reduce branchfest when finishing the lookup.

Tested by:	pho
 2021-01-06 07:05:22 +01:00
Mateusz Guzik
05803be000 cache: stop setting cn_nameptr on entry as matches cn_pnbuf already 
While here tidy up other asserts.
 2021-01-06 07:03:41 +01:00
Mateusz Guzik
3814bea00a cache: drop the now spurious doomed check when crossing a mount point 2021-01-03 21:22:16 +00:00
Mateusz Guzik
82397d7919 vfs: denote vnode being a mount point with VIRF_MOUNTPOINT 
Reviewed by:	kib (previous version)
Differential Revision:	https://reviews.freebsd.org/D27794
 2021-01-03 06:50:06 +00:00
Mateusz Guzik
51bf55fa6c cache: stop checkpointing cn_namelen 
The variable is recomputed by regular lookup from the get go.
 2021-01-03 06:50:06 +00:00
Mateusz Guzik
7220a10b5b cache: predict on no spurious slashes in cache_fpl_handle_root 
This is a step towards speculatively not handling them.
 2021-01-03 06:50:06 +00:00
Mateusz Guzik
30a2fc91fa cache: postpone NAME_MAX check as it may be unnecessary 2021-01-03 06:50:06 +00:00
Mateusz Guzik
eca899bd5d cache: remove spurious null check in sdt probe 2021-01-03 06:50:06 +00:00
Mateusz Guzik
1365b5f86f cache: fold NCF_WHITE check into the rest 
Tested by:	pho
 2021-01-01 00:10:43 +00:00
Mateusz Guzik
d7c62d98c9 cache: call cache_fplookup_modifying in neg 
Tested by:	pho
 2021-01-01 00:10:43 +00:00
Mateusz Guzik
6fe7de1a25 cache: refactor cache_fpl_handle_root to fit the rest of the code better 
Tested by:	pho
 2021-01-01 00:10:43 +00:00
Mateusz Guzik
e17e01bd0e cache: refactor dot handling 
Tested by:	pho
 2021-01-01 00:10:43 +00:00
Mateusz Guzik
4651db56c7 cache: remove a branch from mount point checking 
Tested by:	pho
 2021-01-01 00:10:42 +00:00
Mateusz Guzik
0b5bd1afd8 cache: support lockless lookup of degenerate paths 
Tested by:	pho
 2021-01-01 00:10:42 +00:00
Mateusz Guzik
1d6eb97677 cache: save on branching when parsing the path by inserting a sentinel 
Tested by:	pho
 2021-01-01 00:10:42 +00:00
Mateusz Guzik
67297766b5 cache: hoist trailing slash and degenerate path handling out of the loop 
Tested by:	pho
 2021-01-01 00:10:42 +00:00
Mateusz Guzik
0c09f4b0cc cache: work around corner case of dvp == tvp in cache_fplookup_final_modifying 
Fixes a panic where the kernel would unlock an unheld lock coming from
rename looking up "foo/." as the source.

Reported by:	markj (syzkaller)
 2020-12-28 21:38:20 +00:00
Mateusz Guzik
4ab7d9f484 cache: reduce engrish in previous commit 2020-12-28 02:05:30 +00:00
Mateusz Guzik
0714f921cd cache: save on some branching in common case mount point traversal 2020-12-28 01:53:28 +00:00
Mateusz Guzik
002e18eb7f 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
 2020-12-28 01:53:27 +00:00
Mateusz Guzik
ff97bc034f cache: simplify lockless dot lookups 2020-12-28 01:53:27 +00:00
Mateusz Guzik
abd7ded451 cache: modification and last entry filling support in lockless lookup v2 
The previous patch failed to set the ISDOTDOT flag when appropriate,
which in turn fail to properly handle degenerate lookups.

While here sprinkle some extra assertions.

Tested by:	pho (previous version)
 2020-12-27 21:03:18 +00:00
Mateusz Guzik
623daa69f9 cache: assert internal flags are not passed by namei 2020-12-27 19:49:24 +00:00
Mateusz Guzik
a1fc1f10c6 Revert "cache: modification and last entry filling support in lockless lookup" 
This reverts commit 6dbb07ed68.

Some ports unreliably fail to build with rmdir getting ENOTEMPTY.
 2020-12-27 19:02:29 +00:00
Mateusz Guzik
6dbb07ed68 cache: modification and last entry filling support in lockless lookup 
Tested by:	pho (previous version)
 2020-12-27 17:22:25 +00:00
Mateusz Guzik
906a73e791 cache: fix up cache_hold_vnode comment 2020-12-23 07:24:29 +00:00
Mateusz Guzik
8ab96e265d cache: fix ups bad predicts 
- last level fallback normally sees CREATE; the code should be optimized to not
get there for said case
- fast path commonly fails with ENOENT
 2020-12-13 21:29:39 +00:00
Ryan Libby
d3bbf8af68 cache_fplookup: quiet gcc -Wreturn-type 
Reviewed by:	markj, mjg
Sponsored by:	Dell EMC Isilon
Differential Revision:	https://reviews.freebsd.org/D27555
 2020-12-11 22:51:44 +00:00
Mateusz Guzik
f6dd1aefb7 vfs: group mount per-cpu vars into one struct 
While here move frequently read stuff into the same cacheline.

This shrinks struct mount by 64 bytes.

Tested by:	pho
 2020-11-09 23:02:13 +00:00
Mateusz Guzik
4bfebc8d2c cache: add cache_vop_mkdir and rename cache_rename to cache_vop_rename 2020-10-30 10:46:35 +00:00
Mateusz Guzik
d681c51d36 cache: add missing NIRES_ABS handling 2020-10-26 18:01:18 +00:00
Mateusz Guzik
eb65cde4f5 cache: assorted typo fixes 2020-10-24 13:31:40 +00:00
Mateusz Guzik
029cfccc71 cache: add the missing NC_NOMAKEENTRY and NC_KEEPPOSENTRY to lockless lookup 
They are de facto ignored.
 2020-10-24 13:31:25 +00:00
Mateusz Guzik
acb41008f3 cache: batch updates to numcache in case of mass removal 2020-10-24 01:14:52 +00:00
Mateusz Guzik
208cb7c4b6 cache: refactor alloc/free 
This in particular centralizes manipulation of numcache.
 2020-10-24 01:14:17 +00:00
Mateusz Guzik
1d44405690 cache: fold branch prediction into cache_ncp_canuse 2020-10-24 01:13:47 +00:00
Mateusz Guzik
c13d7d1f98 cache: fix some typos 2020-10-24 01:13:16 +00:00
Mateusz Guzik
f878526f20 cache: drop write-only vars 2020-10-24 01:13:02 +00:00
Mateusz Guzik
3862838921 cache: reduce memory waste in struct namecache 
The previous scheme for calculating the total size was doing sizeof
on the struct and then adding the wanted space for the buffer.

nc_name is at offset 58 while sizeof(struct namecache) is 64.
With CACHE_PATH_CUTOFF of 39 bytes and 1 byte of padding we were
allocating 104 bytes for the entry and never accounting for the 6
byte padding, wasting that space.
 2020-10-23 15:56:22 +00:00
Mateusz Guzik
c7520caa4f vfs: prevent avoidable evictions on mkdir of existing directories 
mkdir -p /foo/bar/baz will mkdir each path component and ignore EEXIST.

The NOCACHE lookup will make the namecache unnecessarily evict the existing entry,
and then fallback to the fs lookup routine eventually leading namei to return an
error as the directory is already there.

For invocations like mkdir -p /usr/obj/usr/src/sys/GENERIC/modules this triggers
fallbacks to the slowpath for concurrently executing lookups.

Tested by:	pho
Discussed with:	kib
 2020-10-22 19:28:12 +00:00
Mateusz Guzik
54f09403a3 cache: assert the created entry does not point to itself 2020-10-22 19:22:34 +00:00