Skip to content

reworked minimal inode_lock breaking series

[email protected] linux-fsdevel lkml 1 version In review

Get series with b4
b4 am [email protected]

Versions newest first

  1. v1 Latest14 patches 3 repliesIn review

Patches in v1 No applies recorded yet

#Subject and filesSizeRepliesApplied
1/14fs: icache begin inode_lock lock breakingfs/inode.c+17 -10 replies
2/14fs: icache lock i_countarch/powerpc/platforms/cell/spufs/file.c drivers/staging/pohmelfs/inode.c fs/9p/vfs_inode.c +57 more+159 -1100 replies
3/14fs: icache lock inodes icache statefs/drop_caches.c fs/fs-writeback.c fs/inode.c +2 more+176 -530 replies
4/14fs: icache unmount code cleanupfs/inode.c fs/notify/inode_mark.c include/linux/fsnotify_backend.h+17 -150 replies
5/14fs: icache lock s_inodes listfs/drop_caches.c fs/fs-writeback.c fs/inode.c +3 more+75 -70 replies
6/14fs: icache lock inode hashfs/inode.c+75 -130 replies
7/14fs: icache lock lru/writeback listsfs/fs-writeback.c fs/inode.c fs/internal.h +2 more+114 -90 replies
8/14fs: icache make nr_inodes and nr_unused atomicfs/fs-writeback.c fs/inode.c include/linux/fs.h +1 more+54 -200 replies
9/14fs: inode atomic last_ino, iunique lockfs/inode.c+33 -110 replies
10/14fs: icache remove inode_lockDocumentation/filesystems/Locking Documentation/filesystems/porting Documentation/filesystems/vfs.txt +13 more+59 -1540 replies
11/14fs: icache factor hash lock into functionsfs/inode.c+24 -140 replies
12/14fs: icache lazy inode lrufs/9p/vfs_inode.c fs/affs/inode.c fs/afs/dir.c +43 more+116 -1450 replies
13/14fs: icache split IO and LRU listsfs/fs-writeback.c fs/inode.c fs/internal.h +4 more+74 -593 replies
14/14fs: icache split writeback and lru locksfs/inode.c+15 -110 replies

Reviews trailers per patch

No Reviewed-by, Acked-by or Tested-by trailers for v1 yet.

Read the v1 thread

Cover letter [patch 00/14] reworked minimal inode_lock breaking series

I am still very much against the locking design of Dave's patchset, but especially the way it is broken out. I could fathom accepting some of the i_lock width reduction if they are properly broken out and justified and documented why they are safe. But lumping this huge series of stuff before inode_lock is even lifted is not the right way to go about it.

I still maintain that i_lock for icache state lock is the way to go at least until we get quite a way down the track. I think a lot of people can't see it maybe because my patchset wasn't broken out terribly well.

So I am posting here just the initial lock breaking part, reworked. In particular, i_lock coverage is established before adding broken out data structure locks, but it also cleans things up and attempts not to move stuff around that isn't strictly required.

I still wait until everything is covered before touching inode_lock, but I have explained how it is actually possible to start removing some of inode_lock even before then, which I think is a good demonstration of the power of having full i_lock coverage.

The steps here should be relatively easy to follow and verify (I hope), and they lead quite easily to the actual scalability improvement steps. So please don't get sidetracked on the temporary trylock ugliness!

Thanks,
Nick