Subject: [PATCHSET] sysfs: sysfs rework, take #2 Hello, all. This is respin of syswork rework patchset in gregkh patch series. Changes are... * ida implemented and used to allocate sysfs_dirent inode numbers instead of using pointer values which caused problems on certain configurations including 32bit userland on 64bit kernel. * reordered patches a bit such that minor updates added later are now in the front of the patchset. * 09-sysfs-flatten-and-fix-sysfs_rename_dir-error_handling added. * 12-sysfs-add-sysfs_dirent-s_name now updates syfs_rename_dir() such that the renaming is reflected in sd->s_name. This fixes netif rename problem. * lockdep annotation updated to handle sysfs nodes deleting other nodes properlyand merged into 17-sysfs-implement-sysfs_dirent- active-references-and-immediate-disconnect * patches which fix fallouts of kill-attr-owner conversion were added 19-sysfs-kill-unnecessary-attribute-owner. Source tree is scanned with grep and brace matching perl script and more fallouts have been converted. * reimplement-sysfs_drop_dentry() patches are completely redone. This was necessary because there was no way to look up the dentry matching a sysfs_dirent for shadow directories, not without getting the hands all dirty with dcache internals. Please read descriptions of patch #20 and 21 for more info. This fixes two races conditions around sysfs dentry/inode reclamation which triggers BUG_ON(). All the known regrssions are fixed. I think I've got all the attr->owner but haven't verified with cross compiling yet, just allyesconfig on x86-64 and i386. I'll try to setup some cross compile environments tomorrow and follow up if I can find more fallouts. The patches are against Linus's master branch of the day - b9099ff63c75216d6ca10bce5a1abcd9293c27e6 Thanks. -- tejun - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
- Follow-Ups:
- Re: [PATCHSET] sysfs: sysfs rework, take #2
- From: Cornelia Huck <[email protected]>
- Re: [PATCHSET] sysfs: sysfs rework, take #2
- From: Alan Stern <[email protected]>
- [PATCH 19/21] sysfs: kill unnecessary attribute->owner
- From: Tejun Heo <[email protected]>
- [PATCH 18/21] sysfs: kill attribute file orphaning
- From: Tejun Heo <[email protected]>
- [PATCH 06/21] sysfs: make sysfs_put() ignore NULL sd
- From: Tejun Heo <[email protected]>
- [PATCH 08/21] sysfs: flatten cleanup paths in sysfs_add_link() and create_dir()
- From: Tejun Heo <[email protected]>
- [PATCH 07/21] sysfs: fix error handling in binattr write()
- From: Tejun Heo <[email protected]>
- [PATCH 03/21] ida: implement idr based id allocator
- From: Tejun Heo <[email protected]>
- [PATCH 04/21] sysfs: move release_sysfs_dirent() to dir.c
- From: Tejun Heo <[email protected]>
- [PATCH 09/21] sysfs: flatten and fix sysfs_rename_dir() error handling
- From: Tejun Heo <[email protected]>
- [PATCH 14/21] sysfs: implement kobj_sysfs_assoc_lock
- From: Tejun Heo <[email protected]>
- [PATCH 10/21] sysfs: consolidate sysfs_dirent creation functions
- From: Tejun Heo <[email protected]>
- [PATCH 05/21] sysfs: allocate inode number using ida
- From: Tejun Heo <[email protected]>
- [PATCH 11/21] sysfs: add sysfs_dirent->s_parent
- From: Tejun Heo <[email protected]>
- [PATCH 12/21] sysfs: add sysfs_dirent->s_name
- From: Tejun Heo <[email protected]>
- [PATCH 13/21] sysfs: make sysfs_dirent->s_element a union
- From: Tejun Heo <[email protected]>
- [PATCH 15/21] sysfs: reimplement symlink using sysfs_dirent tree
- From: Tejun Heo <[email protected]>
- [PATCH 21/21] sysfs: reimplement syfs_drop_dentry()
- From: Tejun Heo <[email protected]>
- [PATCH 17/21] sysfs: implement sysfs_dirent active reference and immediate disconnect
- From: Tejun Heo <[email protected]>
- [PATCH 16/21] sysfs: implement bin_buffer
- From: Tejun Heo <[email protected]>
- [PATCH 20/21] sysfs: separate out sysfs_attach_dentry()
- From: Tejun Heo <[email protected]>
- [PATCH 02/21] idr: separate out idr_mark_full()
- From: Tejun Heo <[email protected]>
- [PATCH 01/21] idr: fix obscure bug in allocation path
- From: Tejun Heo <[email protected]>
- Re: [PATCHSET] sysfs: sysfs rework, take #2
- Prev by Date: can a kmalloc be both GFP_ATOMIC and GFP_KERNEL at the same time?
- Next by Date: [PATCH 07/21] sysfs: fix error handling in binattr write()
- Previous by thread: can a kmalloc be both GFP_ATOMIC and GFP_KERNEL at the same time?
- Next by thread: [PATCH 01/21] idr: fix obscure bug in allocation path
- Index(es):