[patch][rfc] fs: shrink struct dentry

Nick Piggin — 2008-12-01T08:33:43

Hi, Comments? Thanks, Nick -- struct dentry is one of the most critical structures in the kernel. So it's sad to see it going neglected. With CONFIG_PROFILING turned on (which is probably the common case at least for distros and kernel developers), sizeof(struct dcache) == 208 here (64-bit). This gives 19 objects per slab. I packed d_mounted into a hole, and took another 4 bytes off the inline name length to take the padding out from the end of the structure. This shinks it to 200 bytes. I could have gone the other way and increased the length to 40, but I'm aiming for a magic number, read on... I then got rid of the d_cookie pointer. This shrinks it to 192 bytes. Rant: why was this ever a good idea? The cookie system should increase its hash size or use a tree or something if lookups are a problem. Also the "fast dcookie lookups" in oprofile should be moved into the dcookie code -- how can oprofile possibly care about the dcookie_mutex? It gets dropped after get_dcookie() returns so it can't be providin

Developing Tux3 with UML

Daniel Phillips — 2008-11-30T04:48:32

Hi, Running the Tux3 kernel code under UML is easy, fast and safe. Here is a recipe where you can be up and running in about ten minutes from a standing start: # get a kernel tree wget http://kernel.org/pub/linux/kernel/v2.6/linux-2.6.26.5.tar.bz2 tar -xjf linux-2.6.26.5.tar.bz2 cd linux-2.6.26.5 # get the latest tux3 patch and patch the kernel wget http://tux3.org/patches/tux3-2.6.26.5 patch

[PATCH] proc: move inode comment text file to source file

Randy Dunlap — 2008-11-29T21:12:57

From: Randy Dunlap oracle.com> Move fs/proc/inode-alloc.txt comment into a source file so that people will realize that it exists and can update it as needed. Signed-off-by: Randy Dunlap oracle.com> --- fs/proc/generic.c | 15 +++++++++++++++ fs/proc/inode-alloc.txt | 14 -------------- 2 files changed, 15 insertions(+), 14 deletions(-) --- linux-next-20081125.orig/fs/proc/generic.c +++ linux-next-20081125/fs/proc/generic.c < at >< at > -307,6 +307,21 < at >< at > static DEFINE_SPINLOCK(proc_inum_lock); /* * Return an inode number between PROC_DYNAMIC_FIRST and * 0xffffffff, or zero on failure. + * + * Current inode allocations in the proc-fs (hex-numbers): + * + * 00000000reserved + * 00000001-00000fffstatic entries(goners) + * 001root-ino + * + * 00001000-00001fffunused + * 0001xxxx-7fffxxxxpid-dir entries for pid 1-7fff + * 80000000-efffffffunused + * f0000000-ffffffffdynamic entries + * + * Goal: + *Once we split the thing into several virtual filesystems, + *

[rfc] lockdep: check fs reclaim recursion

Nick Piggin — 2008-11-28T12:05:48

Hi, After yesterday noticing some code in mm/filemap.c accidentally perform a __GFP_FS allocation when it should not have been, I thought it might be a good idea to try to catch this kind of thing with lockdep. I coded up a little idea that seems to work. Unfortunately the system has to actually be in __GFP_FS page reclaim, then take the lock, before it will mark it. But at least that might still be some orders of magnitude more common (and more debuggable) than an actual deadlock condition, so we have some improvement I hope. I guess we could do the same thing with __GFP_IO and even GFP_NOIO locks too, but I don't know how expensive it is to add these annotations to lockdep. Filesystems will have the most locks and fiddly code paths... Also, I've just stolen a process flag for this first attempt, but actually I should be using a CONFIG_LOCKDEP specific flag for this (rather than add stuff to pf flags). I've also not added the last few bits for reader/writer lock support, or added all the user output stu

[PATCH] add a FMODE flag to make XFS invisible I/O less hacky

Christoph Hellwig — 2008-11-28T11:24:02

XFS has a mode called invisble I/O that doesn't update any of the timestamps. It's used for HSM-style applications and exposed through the nasty open by handle ioctl. Instead of doing directly assignment of file operations that set an internal flag for it add a new FMODE_INVISIBLE flag that we can check in the normal file operations. Signed-off-by: Christoph Hellwig lst.de>

[rfc git patch] union directory

Miklos Szeredi — 2008-11-28T11:01:32

I've been doing some small fixing/cleanup work on the union directory patches by Jan, and just noticed there's a thread about the union mounts on LKML, so I thought publicizing won't hurt. It's still a work in progress, notably the readdir code currently only works on a few specific filesystem types. Git tree is here: git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/vfs.git union-dir Folded patch below. Comments are welcome. Thanks, Miklos commit bedcd5f45131e7d1fc77e5ce2f9064ef2b4b1abe Author: Jan Blunck suse.de> Date: Fri Nov 28 11:40:25 2008 +0100 union-mount: Simple union-mount readdir implementation This is a very simple union mount readdir implementation. It modifies the readdir routine to merge the entries of union mounted directories and eliminate duplicates while walking the union stack. FIXME: This patch needs to be reworked! At the moment this only works for ext2 and tmpfs. All kind of index directories that return d_off > i_size

you have won 3,500,000 GBP:jamesbill_claimsdesk< at >msn.com

Patrick Marvin — 2008-11-28T02:53:23

-- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo< at >vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html

Questions about exportfs

David Howells — 2008-11-27T14:51:40

Hi Christoph, Can you answer some questions about exportfs please? (1) In linux/exportfs.h, the description of the encode_fh() op says that "encode_fh() should return the number of bytes stored or a negative error code such as %-ENOSPC". This, however, contradicts export_encode_fh(), which returns a member of enum fid_type instead. I presume the docs are wrong, or at least, unclear. (2) Why does exportfs_encode_fh() return the type of the FID? Why not embed it in the fid struct? It seems that the value has to be handed back to the decode routine. (3) Is fat_encode_fh() allowed to return a value not in the fid_type enum? David -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo< at >vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html

[patch 1/2] mm: pagecache allocation gfp fixes

Nick Piggin — 2008-11-27T09:34:01

Frustratingly, gfp_t is really divided into two classes of flags. One are the context dependent ones (can we sleep? can we enter filesystem? block subsystem? should we use some extra reserves, etc.). The other ones are the type of memory required and depend on how the algorithm is implemented rather than the point at which the memory is allocated (highmem? dma memory? etc). Some of functions which allocate a page and add it to page cache take a gfp_t, but sometimes those functions or their callers aren't really doing the right thing: when allocating pagecache page, the memory type should be mapping_gfp_mask(mapping). When allocating radix tree nodes, the memory type should be kernel mapped (not highmem) memory. The gfp_t argument should only really be needed for context dependent options. This patch doesn't really solve that tangle in a nice way, but it does attempt to fix a couple of bugs. find_or_create_page changes its radix-tree allocation to only include the main context dependent flags in order so t

Newbie Question -- error: implicit declaration of function 'iget'

Qingye Jiang (John — 2008-11-27T04:03:32

Hi all, I am a newbie on Linux file systems. I am currently learning Linux file systems with Steve French's "Linux File Systems in 21 Days". The URL for his tutorials is: http://svn.samba.org/samba/ftp/cifs-cvs/samplefs.tar.gz http://svn.samba.org/samba/ftp/cifs-cvs/ols2007-fs-tutorial-smf.odp I have finished the day1 example, and get stuck at the day2 example. When compiling the day2 example, I get the following error message: qyjohn< at >qyjohn-laptop:/usr/src/linux-source-2.6.27$ sudo make M=~/samplefs/day2 [sudo] password for qyjohn: CC [M] /home/qyjohn/samplefs/day2/super.o /home/qyjohn/samplefs/day2/super.c: In function 'samplefs_fill_super': /home/qyjohn/samplefs/day2/super.c:112: error: implicit declaration of function 'iget' /home/qyjohn/samplefs/day2/super.c:112: warning: assignment makes pointer from integer without a cast make[1]: *** [/home/qyjohn/samplefs/day2/super.o] Error 1 make: *** [_module_/home/qyjohn/samplefs/day2] Error 2 I looked at linux/fs.h, and found that the declaration of the __

[PATCH 1/2] relatime: Make atime updates more useful

Matthew Garrett — 2008-11-26T19:54:57

Allow atime to be updated once per day even with relatime enabled. This lets utilities like tmpreaper (which deletes files based on last access time) continue working. Signed-off-by: Matthew Garrett redhat.com> --- Updated version of Ingo's patch from last year - this section is identical. commit 2c145e187600ca961715fa82ae3ae7919d744bc9 Author: Matthew Garrett redhat.com> Date: Wed Nov 26 17:44:07 2008 +0000 Make relatime smarter Allow atime to be updated once per day even with relatime. This lets utilities like tmpreaper (which delete files based on last access time) continue working. diff --git a/fs/inode.c b/fs/inode.c index 0487ddb..348fa16 100644 --- a/fs/inode.c +++ b/fs/inode.c < at >< at > -1179,6 +1179,41 < at >< at > sector_t bmap(struct inode * inode, sector_t block) } EXPORT_SYMBOL(bmap); +/* + * Relative atime updates frequency (default: 1 day): + */ +int relatime_interval __read_mostly = 24*60*60; + +/* + * With relative atime, only update atime if the + * previous ati

[PATCH] 9p: convert d_iname references to d_name.name

Wu Fengguang — 2008-11-26T13:42:54

d_iname is rubbish for long file names. Use d_name.name in printks instead. Signed-off-by: Wu Fengguang linux.intel.com> --- diff --git a/fs/9p/fid.c b/fs/9p/fid.c index 3031e32..2a983d4 100644 --- a/fs/9p/fid.c +++ b/fs/9p/fid.c < at >< at > -45,7 +45,7 < at >< at > int v9fs_fid_add(struct dentry *dentry, struct p9_fid *fid) struct v9fs_dentry *dent; P9_DPRINTK(P9_DEBUG_VFS, "fid %d dentry %s\n", -fid->fid, dentry->d_iname); +fid->fid, dentry->d_name.name); dent = dentry->d_fsdata; if (!dent) { < at >< at > -79,7 +79,7 < at >< at > static struct p9_fid *v9fs_fid_find(struct dentry *dentry, u32 uid, int any) struct p9_fid *fid, *ret; P9_DPRINTK(P9_DEBUG_VFS, " dentry: %s (%p) uid %d any %d\n", -dentry->d_iname, dentry, uid, any); +dentry->d_name.name, dentry, uid, any); dent = (struct v9fs_dentry *) dentry->d_fsdata; ret = NULL; if (dent) { diff --git a/fs/9p/vfs_dentry.c b/fs/9p/vfs_dentry.c index f9534f1..9fcfbaf 100644 --- a/fs/9p/vfs_dentry.c +++ b/fs/9p/vfs_dentry.c < at >< at > -52,7 +52,7 < at >< at > static int v

[PATCH] vfs: initialize d_iname to null for long file names

Wu Fengguang — 2008-11-26T13:41:05

dentry->d_iname is currently uninitialized for long file names. Init it to null string to avoid bad behaviors on careless references to it. Signed-off-by: Wu Fengguang linux.intel.com> --- diff --git a/fs/dcache.c b/fs/dcache.c index a1d86c7..29d1768 100644 --- a/fs/dcache.c +++ b/fs/dcache.c < at >< at > -929,6 +929,7 < at >< at > struct dentry *d_alloc(struct dentry * parent, const struct qstr *name) kmem_cache_free(dentry_cache, dentry); return NULL; } +dentry->d_iname[0] = '\0'; } else { dname = dentry->d_iname; } -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo< at >vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html

[PATCH] fs: use menuconfig to control the Misc. filesystems menu

Randy Dunlap — 2008-11-25T21:51:43

Subject: [PATCH] fs: use menuconfig to control the Misc. filesystems menu To: linux-fsdevel< at >vger.kernel.org Cc: akpm linux-foundation.org>,Al Viro zeniv.linux.org.uk> Bcc: =-=-=-=-=-=-=-=-=# Don't remove this line #=-=-=-=-=-=-=-=-=- From: Randy Dunlap oracle.com> Have one option to control Miscellaneous filesystems. This makes it easy to disable all of them at one time. Signed-off-by: Randy Dunlap oracle.com> --- Was this already attempted? If so, was there any problem with it? fs/Kconfig | 17 +++++++++++++++-- 1 file changed, 15 insertions(+), 2 deletions(-) --- linux-next-20081125.orig/fs/Kconfig +++ linux-next-20081125/fs/Kconfig < at >< at > -758,7 +758,20 < at >< at > config CONFIGFS_FS endmenu -menu "Miscellaneous filesystems" +menuconfig MISC_FILESYSTEMS +bool "Miscellaneous filesystems" +default y +---help--- + Say Y here to get to see options for various miscellaneous + filesystems, such as filesystems that came from other + operat

[PATCH] do_mpage_readpage(): remove first_logical_block parameter

Franck Bui-Huu — 2008-11-25T20:29:11

From: Franck Bui-Huu gmail.com> This patch makes this function more readable IMHO by getting rid of its parameter 'first_logical_block'. This parameter was previously used to remember the original block number in the file (ie the page index) that 'map_bh' was pointing at first. Indeed this was needed since 'map_bh->page' was modified even if 'map_bh' maps the whole current page (IOW when get_block() is not used). This had the bad effect to make the state of 'map_bh' inconsistent too. This patch changes 'map_bh->page' only when get_block() is called hence removes the need of the 'first_logical_block' argument. The value stored in 'logical_block_parameter' is now recalculated each time do_mpage_readpage() but it shouldn't matter since the operation is trivial. Signed-off-by: Franck Bui-Huu gmail.com> --- Andrew, I made this when trying to understand this function. IMHO it makes the code more readable. I still don't know why 'map_bh->page' was setup every time do_mpage_readpage(

[resend take 3 0/4] Distributed storage.

Evgeniy Polyakov — 2008-11-25T14:37:46

Hello. Distributed storage is a feature-rich network block device. Supported features: * Kernel-side client and server. No need for any special tools for data processing (like special userspace applications) except for configuration. * Bullet-proof memory allocations via memory pools for all temporary objects (transaction and so on). All clients structures are allocated as single transaction from the memory pool and except this there is no allocation overhead. Network adds own though. Server uses memory pools too, but number of allocations is higher (bio, transaction and pages for IO). * Zero-copy sending (except header) if supported by device using sendpage(). * Failover recovery in case of broken link (reconnection if remote node is down). * Full transaction support (resending of the failed transactions on timeout of after reconnect to failed node). * Dynamically resizeable pool of threads used for data receiving and crypto processing. * Initial autoconfigur

[PATCH fwd] poll: allow f_op->poll to sleep

Miklos Szeredi — 2008-11-22T08:58:33

[Fwd to Andrew, CC fsdevel & Arjan] Andrew, can you please apply this? I don't think it should be sneaked in through the fuse tree (even though the immediate requirement is from the poll implementation in fuse). Thanks, Miklos ---- From: Tejun Heo gmail.com> f_op->poll is the only vfs operation which is not allowed to sleep. It's because poll and select implementation used task state to synchronize against wake ups, which doesn't have to be the case anymore as wait/wake interface can now use custom wake up functions. The non-sleep restriction can be a bit tricky because ->poll is not called from an atomic context and the result of accidentally sleeping in ->poll only shows up as temporary busy looping when the timing is right or rather wrong. This patch converts poll/select to use custom wake up function and use separate triggered variable to synchronize against wake up events. The only added overhead is an extra function call during wake up and negligible. This patch removes the one non-sleep

PG_updatodate vs BH_updatodate

Francis Moreau — 2008-11-21T22:03:27

Hello, I'm wondering if there's any difference between these 2 flags except that one is a page flag and the other one is a buffer head flag. But in my understanding these 2 flags means that the page contains uptodate data. So the buffer head flag seems useless since it could be calculated from bh->b_page->flags. Thanks for any hints Francis -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo< at >vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html

tracing at filesystem level

2008-11-21T15:53:40

Hello, i`d like to be able to trace filesystem access at early boot time, i.e to see what files being opened/closed on early boot (and later on). one possible way to do it is using nfs-root, so we can trace it at the network or server level - but how can this be done without using network filesystem ? i came across tracefs (http://www.filesystems.org/docs/tracefs-fast04/tracefs.pdf) which looks very promising, but it seems it`s not actively maintained. isn`t there a standard way to do that with recent kernels ? i searched for a while but didn`t find something appropriate.... regards Roland ps: this is also very interesting for intrusion detection - think of virtual machine`s filesystem activity being watched trough serial console (i.e. with nothing running in userspace and without hackers ability to disable it) ____________________________________________________________________ Psssst! Schon vom neuen WEB.DE MultiMessenger gehört? Der kann`s mit allen: http://www.produkte.web.de/messenger/?did=312

UBIFS updates for 2.6.28

Artem Bityutskiy — 2008-11-21T17:19:22

Hi, here are the UBIFS fixes we have scheduled for 2.6.28. We are planning to send them to Linus a little bit later. I send them for your review. The "UBIFS: allow for gaps when dirtying the LPT" patch solves the problem reported by Deepak Saxena. The "UBIFS: remove printk" fixes a little bit annoying message from UBIFS which comes every time we re-mount the FS. The "UBIFS: do not print scary memory allocation warnings", "UBIFS: do not allocate too much", and "UBIFS: pre-allocate bulk-read buffer" fix UBIFS memory allocation problems when bulk-read is enabled. Adrian Hunter (1): UBIFS: allow for gaps when dirtying the LPT Artem Bityutskiy (6): UBIFS: remove printk MAINTAINERS: change UBI/UBIFS git tree URLs UBIFS: fix compilation warnings UBIFS: do not print scary memory allocation warnings UBIFS: do not allocate too much UBIFS: pre-allocate bulk-read buffer Harvey Harrison (1): UBIFS: endian handling fixes and annotations MAINTAINERS

[PATCH 0/2] HFS+ journal improvement - take 6

Warren Turkal — 2008-11-21T04:01:25

This is a patchset to change the way that the HFS+ filesystem detects whether a volume has a journal or not. The code currently mounts an HFS+ volume read-only by default when a journal is detected. One can force a read/write mount by giving the "force" mount option. The current code has this behavior since there is no support for the HFS+ journal. My problem is that the detection of the journal could be better. The current code tests the attribute bit in the volume header that indicates there is a journal. If that bit is set, the code assumes that there is a journal. Unfortunately, this is not enough to really determine if there is a journal or not. When that bit is set, one must also examine the journal info block field of the volume header. If this field is 0, there is no journal, and the volume can be mounted read/write. Thanks, wt Warren Turkal (2): Identify journal info block in volume header. Fix journal detection on HFS+. fs/hfsplus/hfsplus_raw.h | 2 +- fs/hfsplus/super.c | 10

gmane.linux.file-systems