Add a helper to read the AGI header and perform basic verification.
Based on hunks from a larger patch from Dave Chinner.
(First sent on Juli 23rd)
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
When an I/O error occurs during an intermediate commit on a rolling
transaction, xfs_trans_commit() will free the transaction structure
and the related ticket. However, the duplicate transaction that
gets used as the transaction continues still contains a pointer
to the ticket. Hence when the duplicate transaction is cancelled
and freed, we free the ticket a second time.
Add reference counting to the ticket so that we hold an extra
reference to the ticket over the transaction commit. We drop the
extra reference once we have checked that the transaction commit
did not return an error, thus avoiding a double free on commit
error.
Credit to Nick Piggin for tripping over the problem.
SGI-PV: 989741
Signed-off-by: Dave Chinner <david@fromorbit.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
To make sure we free the security data inodes need to be freed using the
proper VFS helper (which we also need to export for this). We mark these
inodes bad so we can skip the flush path for them.
SGI-PV: 987246
SGI-Modid: xfs-linux-melb:xfs-kern:32398a
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: David Chinner <david@fromorbit.com>
xfs_bulkstat only wants the dinode, offset and buffer from a given inode
number. Instead of using xfs_itobp on a fake inode which is complicated
and currently leads to leaks of the security data just use xfs_inotobp
which is designed to do exactly the kind of lookup xfs_bulkstat wants. The
only thing that's missing in xfs_inotobp is a flags paramter that let's us
pass down XFS_IMAP_BULKSTAT, but that can easily added.
SGI-PV: 987246
SGI-Modid: xfs-linux-melb:xfs-kern:32397a
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: David Chinner <david@fromorbit.com>
Change all the remaining AIL API functions that are passed struct
xfs_mount pointers to pass pointers directly to the struct xfs_ail being
used. With this conversion, all external access to the AIL is via the
struct xfs_ail. Hence the operation and referencing of the AIL is almost
entirely independent of the xfs_mount that is using it - it is now much
more tightly tied to the log and the items it is tracking in the log than
it is tied to the xfs_mount.
SGI-PV: 988143
SGI-Modid: xfs-linux-melb:xfs-kern:32353a
Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Bring the ail lock inside the struct xfs_ail. This means the AIL can be
entirely manipulated via the struct xfs_ail rather than needing both the
struct xfs_mount and the struct xfs_ail.
SGI-PV: 988143
SGI-Modid: xfs-linux-melb:xfs-kern:32350a
Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
When copying lsn's from the log item to the inode or dquot flush lsn, we
currently grab the AIL lock. We do this because the LSN is a 64 bit
quantity and it needs to be read atomically. The lock is used to guarantee
atomicity for 32 bit platforms.
Make the LSN copying a small function, and make the function used
conditional on BITS_PER_LONG so that 64 bit machines don't need to take
the AIL lock in these places.
SGI-PV: 988143
SGI-Modid: xfs-linux-melb:xfs-kern:32349a
Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Now that the deleted inodes list is unused, kill it. This also removes the
i_reclaim list head from the xfs_inode, shrinking it by two pointers.
SGI-PV: 988142
SGI-Modid: xfs-linux-melb:xfs-kern:32334a
Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
To avoid issues with different lifecycles of XFS and Linux inodes, embedd
the linux inode inside the XFS inode. This means that the linux inode has
the same lifecycle as the XFS inode, even when it has been released by the
OS. XFS inodes don't live much longer than this (a short stint in reclaim
at most), so there isn't significant memory usage penalties here.
Version 3 o kill xfs_icount()
Version 2 o remove unused commented out code from xfs_iget(). o kill
useless cast in VFS_I()
SGI-PV: 988141
SGI-Modid: xfs-linux-melb:xfs-kern:32323a
Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
structures.
Always use the generic xfs_btree_block type instead of the short / long
structures. Add XFS_BTREE_SBLOCK_LEN / XFS_BTREE_LBLOCK_LEN defines for
the length of a short / long form block. The rationale for this is that we
will grow more btree block header variants to support CRCs and other RAS
information, and always accessing them through the same datatype with
unions for the short / long form pointers makes implementing this much
easier.
SGI-PV: 988146
SGI-Modid: xfs-linux-melb:xfs-kern:32300a
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Donald Douwsma <donaldd@sgi.com>
Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Replace the generic record / key / ptr addressing macros that use cpp
token pasting with simpler macros that do the job for just one given btree
type. The new macros lose the cur argument and thus can be used outside
the core btree code, but also gain an xfs_mount * argument to allow for
checking the CRC flag in the near future. Note that many of these macros
aren't actually used in the kernel code, but only in userspace (mostly in
xfs_repair).
SGI-PV: 988146
SGI-Modid: xfs-linux-melb:xfs-kern:32295a
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Donald Douwsma <donaldd@sgi.com>
Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Clean up the way the maximum and minimum records for the btree blocks are
calculated. For the alloc and inobt btrees all the values are
pre-calculated in xfs_mount_common, and we switch the current loop around
the ugly generic macros that use cpp token pasting to generate type names
to two small helpers in normal C code. For the bmbt and bmdr trees these
helpers also exist, but can be called during runtime, too. Here we also
kill various macros dealing with them and inline the logic into the
get_minrecs / get_maxrecs / get_dmaxrecs methods in xfs_bmap_btree.c.
Note that all these new helpers take an xfs_mount * argument which will be
needed to determine the size of a btree block once we add support for
extended btree blocks with CRCs and other RAS information.
SGI-PV: 988146
SGI-Modid: xfs-linux-melb:xfs-kern:32292a
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Donald Douwsma <donaldd@sgi.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
xfs_iflush_all() walks the m_inodes list to find inodes that need
reclaiming. We already have such a list - the m_del_inodes list. Replace
xfs_iflush_all() with a call to xfs_finish_reclaim_all() and clean up
xfs_finish_reclaim_all() to handle the different flush modes now needed.
Originally based on a patch from Christoph Hellwig.
Version 3 o rediff against new linux-2.6/xfs_sync.c code
Version 2 o revert xfs_syncsub() inode reclaim behaviour back to original
code o xfs_quiesce_fs() should use XFS_IFLUSH_DELWRI_ELSE_ASYNC, not
XFS_IFLUSH_ASYNC, to prevent change of behaviour.
SGI-PV: 988139
SGI-Modid: xfs-linux-melb:xfs-kern:32284a
Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
It's possible to have outstanding xfs_ioend_t's queued when the file size
is zero. This can happen in the direct I/O path when a direct I/O write
fails due to ENOSPC. In this case the xfs_ioend_t will still be queued (ie
xfs_end_io_direct() does not know that the I/O failed so can't force the
xfs_ioend_t to be flushed synchronously).
When we truncate a file on unlink we don't know to wait for these
xfs_ioend_ts and we can have a use-after-free situation if the inode is
reclaimed before the xfs_ioend_t is finally processed.
As was suggested by Dave Chinner lets wait for all I/Os to complete when
truncating the file size to zero.
SGI-PV: 981668
SGI-Modid: xfs-linux-melb:xfs-kern:32216a
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Make the existing bmap btree tracing generic so that it applies to all
btree types.
Some fragments lifted from a patch by Dave Chinner.
SGI-PV: 985583
SGI-Modid: xfs-linux-melb:xfs-kern:32187a
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: Bill O'Donnell <billodo@sgi.com>
Signed-off-by: David Chinner <david@fromorbit.com>
To avoid having to initialise some fields of the XFS inode on every
allocation, we can use the slab init-once feature to initialise them. All
we have to guarantee is that when we free the inode, all it's entries are
in the initial state. Add asserts where possible to ensure debug kernels
check this initial state before freeing and after allocation.
SGI-PV: 981498
SGI-Modid: xfs-linux-melb:xfs-kern:31925a
Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Yet another bug was found in xfs_iext_irec_compact_full() and while the
source of the bug was found it wasn't an easy task to track it down
because the conditions are very difficult to reproduce.
A HUGE thank-you goes to Russell Cattelan and Eric Sandeen for their
significant effort in tracking down the source of this corruption.
xfs_iext_irec_compact_full() and xfs_iext_irec_compact_pages() are almost
identical - they both compact indirect extent lists by moving extents from
subsequent buffers into earlier ones. xfs_iext_irec_compact_pages() only
moves extents if all of the extents in the next buffer will fit into the
empty space in the buffer before it. xfs_iext_irec_compact_full() will go
a step further and move part of the next buffer if all the extents wont
fit. It will then shift the remaining extents in the next buffer up to the
start of the buffer. The bug here was that we did not update er_extoff and
this caused extent list corruption.
It does not appear that this extra functionality gains us much. Calling
xfs_iext_irec_compact_pages() instead will do a good enough job at
compacting the indirect list and will be quicker too.
For the case in xfs_iext_indirect_to_direct() the total number of extents
in the indirect list will fit into one buffer so we will never need the
extra functionality of xfs_iext_irec_compact_full() there.
Also xfs_iext_irec_compact_pages() doesn't need to do a memmove() (the
buffers will never overlap) so we don't want the performance hit that can
incur.
SGI-PV: 987159
SGI-Modid: xfs-linux-melb:xfs-kern:32166a
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
If we don't move all the records from the next buffer into the current
buffer then we need to update the er_extoff field of the next buffer as we
shift the remaining records to the start of the buffer.
SGI-PV: 987159
SGI-Modid: xfs-linux-melb:xfs-kern:32165a
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: Russell Cattelan <cattelan@thebarn.com>
The patches that are intended to introduce copy-on-write credentials for 2.6.28
require abstraction of access to some fields of the task structure,
particularly for the case of one task accessing another's credentials where RCU
will have to be observed.
Introduced here are trivial no-op versions of the desired accessors for current
and other tasks so that other subsystems can start to be converted over more
easily.
Wrappers are introduced into a new header (linux/cred.h) for UID/GID,
EUID/EGID, SUID/SGID, FSUID/FSGID, cap_effective and current's subscribed
user_struct. These wrappers are macros because the ordering between header
files mitigates against making them inline functions.
linux/cred.h is #included from linux/sched.h.
Further, XFS is modified such that it no longer defines and uses parameterised
versions of current_fs[ug]id(), thus getting rid of the namespace collision
otherwise incurred.
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
Use KM_NOFS to prevent recursion back into the filesystem which can cause
deadlocks.
In the case of xfs_iread() we hold the lock on the inode cluster buffer
while allocating memory for the trace buffers. If we recurse back into XFS
to flush data that may require a transaction to allocate extents which
needs log space. This can deadlock with the xfsaild thread which can't
push the tail of the log because it is trying to get the inode cluster
buffer lock.
SGI-PV: 981498
SGI-Modid: xfs-linux-melb:xfs-kern:31838a
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: David Chinner <david@fromorbit.com>
In xfs_ialloc we just want to set all timestamps to the current time. We
don't need to mark the inode dirty like xfs_ichgtime does, and we don't
need nor want the opimizations in xfs_ichgtime that I will introduce in
the next patch.
So just opencode the timestamp update in xfs_ialloc, and remove the new
unused XFS_ICHGTIME_ACC case in xfs_ichgtime.
SGI-PV: 981498
SGI-Modid: xfs-linux-melb:xfs-kern:31825a
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Use the new completion flush code to implement the inode flush lock.
Removes one of the final users of semaphores in the XFS code base.
SGI-PV: 981498
SGI-Modid: xfs-linux-melb:xfs-kern:31817a
Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Sanitize setting up the Linux indode.
Setting up the xfs_inode <-> inode link is opencoded in xfs_iget_core now
because that's the only place it needs to be done, xfs_initialize_vnode is
renamed to xfs_setup_inode and loses all superflous paramaters. The check
for I_NEW is removed because it always is true and the di_mode check moves
into xfs_iget_core because it's only needed there.
xfs_set_inodeops and xfs_revalidate_inode are merged into xfs_setup_inode
and the whole things is moved into xfs_iops.c where it belongs.
SGI-PV: 981498
SGI-Modid: xfs-linux-melb:xfs-kern:31782a
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
All remaining bhv_vnode_t instance are in code that's more or less Linux
specific. (Well, for xfs_acl.c that could be argued, but that code is on
the removal list, too). So just do an s/bhv_vnode_t/struct inode/ over the
whole tree. We can clean up variable naming and some useless helpers
later.
SGI-PV: 981498
SGI-Modid: xfs-linux-melb:xfs-kern:31781a
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
In various places we can just move a VFS_I call into the argument list of
called functions/macros instead of having a local bhv_vnode_t.
SGI-PV: 981498
SGI-Modid: xfs-linux-melb:xfs-kern:31776a
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
If we allow incore extent tree allocations to recurse into the
filesystem under memory pressure, new delayed allocations through
xfs_iomap_write_delay() can deadlock on themselves if memory
reclaim tries to write back dirty pages from that inode.
It will deadlock in xfs_iomap_write_allocate() trying to take the
ilock we already hold. This can also show up as complex ABBA deadlocks
when multiple threads are triggering memory reclaim when trying to
allocate extents.
The main cause of this is the fact that delayed allocation is not done in
a transaction, so KM_NOFS is not automatically added to the allocations to
prevent this recursion.
Mark all allocations done for the incore inode extent tree as KM_NOFS to
ensure they never recurse back into the filesystem.
Version 2: o KM_NOFS implies KM_SLEEP, so just use KM_NOFS
SGI-PV: 981498
SGI-Modid: xfs-linux-melb:xfs-kern:31726a
Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
This function is used to compact the indirect extent list by moving
extents from one page to the previous to fill them up. After we move some
extents to an earlier page we need to shuffle the remaining extents to the
start of the page. The actual bug here is the second argument to memmove()
needs to index past the extents, that were copied to the previous page,
and move the remaining extents. For pages that are already full (ie
ext_avail == 0) the compaction code has no net effect so don't do it.
SGI-PV: 983337
SGI-Modid: xfs-linux-melb:xfs-kern:31332a
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
During a forced shutdown a xfs inode can be destroyed before log I/O
involving that inode is complete. We need to wait for the inode to be
unpinned before tearing it down. Version 2 cleans up the code a bit by
relying on xfs_iflush() to do the unpinning and forced shutdown check.
SGI-PV: 981240
SGI-Modid: xfs-linux-melb:xfs-kern:31326a
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: David Chinner <dgc@sgi.com>
xfs_igrow_start just expands to xfs_zero_eof with two asserts that are
useless in the context of the only caller and some rather confusing
comments.
xfs_igrow_finish is just a few lines of code decorated again with useless
asserts and confusing comments.
Just kill those two and merge them into xfs_setattr.
SGI-PV: 981498
SGI-Modid: xfs-linux-melb:xfs-kern:31186a
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
kmem_free() function takes (ptr, size) arguments but doesn't actually use
second one.
This patch removes size argument from all callsites.
SGI-PV: 981498
SGI-Modid: xfs-linux-melb:xfs-kern:31050a
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
We only need to allocate space for the number of inodes in the cluster
when writing back inodes, not every byte in the inode cluster. This
reduces the amount of memory needing to be allocated to 256 bytes instead
of 64k.
SGI-PV: 981949
SGI-Modid: xfs-linux-melb:xfs-kern:31182a
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
writeback
If we allow memory reclaim to wait on the pages under writeback in inode
cluster writeback we could deadlock because we are currently holding the
ILOCK on the initial writeback inode which is needed in data I/O
completion to change the file size or do unwritten extent conversion
before the pages are taken out of writeback state.
SGI-PV: 981091
SGI-Modid: xfs-linux-melb:xfs-kern:31015a
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
The writer field is not needed for non_DEBU builds so remove it. While
we're at i also clean up the interface for is locked asserts to go through
and xfs_iget.c helper with an interface like the xfs_ilock routines to
isolated the XFS codebase from mrlock internals. That way we can kill
mrlock_t entirely once rw_semaphores grow an islocked facility. Also
remove unused flags to the ilock family of functions.
SGI-PV: 976035
SGI-Modid: xfs-linux-melb:xfs-kern:30902a
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
On success, we still need to join the inode to the current transaction in
xfs_itruncate_finish(). Fixes regression from error handling changes.
SGI-PV: 980084
SGI-Modid: xfs-linux-melb:xfs-kern:30845a
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
xfs_iflush_fork() never returns an error. Mark it void and clean up the
code calling it that checks for errors.
SGI-PV: 980084
SGI-Modid: xfs-linux-melb:xfs-kern:30827a
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
xfs_bawrite() can return immediate error status on async writes. Unlike
xfsbdstrat() we don't ever check the error on the buffer after the call,
so we currently do not catch errors at all here. Ensure we catch and
propagate or warn to the syslog about up-front async write errors.
SGI-PV: 980084
SGI-Modid: xfs-linux-melb:xfs-kern:30824a
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
xfs_trans_commit() can return errors when there are problems in the
transaction subsystem. They are indicative that the entire transaction may
be incomplete, and hence the error should be propagated as there is a good
possibility that there is something fatally wrong in the filesystem. Catch
and propagate or warn about commit errors in the places where they are
currently ignored.
SGI-PV: 980084
SGI-Modid: xfs-linux-melb:xfs-kern:30795a
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Remove open coded checks for the whether the inode is clean and replace
them with an inlined function.
SGI-PV: 977461
SGI-Modid: xfs-linux-melb:xfs-kern:30503a
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Remove the xfs_icluster structure and replace with a radix tree lookup.
We don't need to keep a list of inodes in each cluster around anymore as
we can look them up quickly when we need to. The only time we need to do
this now is during inode writeback.
Factor the inode cluster writeback code out of xfs_iflush and convert it
to use radix_tree_gang_lookup() instead of walking a list of inodes built
when we first read in the inodes.
This remove 3 pointers from each xfs_inode structure and the xfs_icluster
structure per inode cluster. Hence we reduce the cache footprint of the
xfs_inodes by between 5-10% depending on cluster sparseness.
To be truly efficient we need a radix_tree_gang_lookup_range() call to
stop searching once we are past the end of the cluster instead of trying
to find a full cluster's worth of inodes.
Before (ia64):
$ cat /sys/slab/xfs_inode/object_size 536
After:
$ cat /sys/slab/xfs_inode/object_size 512
SGI-PV: 977460
SGI-Modid: xfs-linux-melb:xfs-kern:30502a
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
When pdflush is writing back inodes, it can get stuck on inode cluster
buffers that are currently under I/O. This occurs when we write data to
multiple inodes in the same inode cluster at the same time.
Effectively, delayed allocation marks the inode dirty during the data
writeback. Hence if the inode cluster was flushed during the writeback of
the first inode, the writeback of the second inode will block waiting for
the inode cluster write to complete before writing it again for the newly
dirtied inode.
Basically, we want to avoid this from happening so we don't block pdflush
and slow down all of writeback. Hence we introduce a non-blocking async
inode flush flag that pdflush uses. If this flag is set, we use
non-blocking operations (e.g. try locks) whereever we can to avoid
blocking or extra I/O being issued.
SGI-PV: 970925
SGI-Modid: xfs-linux-melb:xfs-kern:30501a
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
The only difference between the functions is one passes an inode for the
lookup, the other passes an inode number. However, they don't do the same
validity checking or set all the same state on the buffer that is returned
yet they should.
Factor the functions into a common implementation.
SGI-PV: 970925
SGI-Modid: xfs-linux-melb:xfs-kern:30500a
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Remove macro-to-small-function indirection from xfs_sb.h, and remove some
which are completely unused.
SGI-PV: 976035
SGI-Modid: xfs-linux-melb:xfs-kern:30528a
Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: Donald Douwsma <donaldd@sgi.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Now that all direct caller of xfs_iaccess are gone we can kill xfs_iaccess
and xfs_access and just use generic_permission with a check_acl callback.
This is required for the per-mount read-only patchset in -mm to work
properly with XFS.
SGI-PV: 971186
SGI-Modid: xfs-linux-melb:xfs-kern:30370a
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Currently XFS_IFORK_* and XFS_DFORK* are implemented by means of
XFS_CFORK* macros. But given that XFS_IFORK_* operates on an xfs_inode
that embedds and xfs_icdinode_core and XFS_DFORK_* operates on an
xfs_dinode that embedds a xfs_dinode_core one will have to do endian
swapping while the other doesn't. Instead of having the current mess with
the CFORK macros that have byteswapping and non-byteswapping version
(which are inconsistantly named while we're at it) just define each family
of the macros to stand by itself and simplify the whole matter.
A few direct references to the CFORK variants were cleaned up to use IFORK
or DFORK to make this possible.
SGI-PV: 971186
SGI-Modid: xfs-linux-melb:xfs-kern:30163a
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Tim Shimmin <tes@sgi.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
SGI-PV: 971186
SGI-Modid: xfs-linux-melb:xfs-kern:30098a
Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Use XFS_IS_REALTIME_INODE in more places, and #define it to 0 if
CONFIG_XFS_RT is off. This should be safe because mount checks in
xfs_rtmount_init:
so if we get mounted w/o CONFIG_XFS_RT, no realtime inodes should be
encountered after that.
Defining XFS_IS_REALTIME_INODE to 0 saves a bit of stack space,
presumeably gcc can optimize around the various "if (0)" type checks:
xfs_alloc_file_space -8 xfs_bmap_adjacent -16 xfs_bmapi -8
xfs_bmap_rtalloc -16 xfs_bunmapi -28 xfs_free_file_space -64 xfs_imap +8
<-- ? hmm. xfs_iomap_write_direct -12 xfs_qm_dqusage_adjust -4
xfs_qm_vop_chown_reserve -4
SGI-PV: 971186
SGI-Modid: xfs-linux-melb:xfs-kern:30014a
Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
The log force added in xfs_iget_core() has been a performance issue since
it was introduced for tight loops that allocate then unlink a single file.
under heavy writeback, this can introduce unnecessary latency due tothe
log I/o getting stuck behind bulk data writes.
Fix this latency problem by avoinding the need for the log force by moving
the place we mark linux inode dirty to the transaction commit rather than
on transaction completion.
This also closes a potential hole in the sync code where a linux inode is
not dirty between the time it is modified and the time the log buffer has
been written to disk.
SGI-PV: 972753
SGI-Modid: xfs-linux-melb:xfs-kern:30007a
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
We need xfs_bulkstat() to report inode stat for inodes with link count
zero but reference count non zero.
The fix here:
http://oss.sgi.com/archives/xfs/2007-09/msg00266.html
changed this behavior and made xfs_bulkstat() to filter all unlinked
inodes including those that are not destroyed yet but held by reference.
The attached patch returns back to the original behavior by marking the
on-disk inode buffer "dirty" when di_mode is cleared (at that time both
inode link and reference counter are zero).
SGI-PV: 972004
SGI-Modid: xfs-linux-melb:xfs-kern:29914a
Signed-off-by: Vlad Apostolov <vapo@sgi.com>
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
xfs_iocore_t is a structure embedded in xfs_inode. Except for one field it
just duplicates fields already in xfs_inode, and there is nothing this
abstraction buys us on XFS/Linux. This patch removes it and shrinks source
and binary size of xfs aswell as shrinking the size of xfs_inode by 60/44
bytes in debug/non-debug builds.
SGI-PV: 970852
SGI-Modid: xfs-linux-melb:xfs-kern:29754a
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
Currently there is an indirection called ioops in the XFS data I/O path.
Various functions are called by functions pointers, but there is no
coherence in what this is for, and of course for XFS itself it's entirely
unused. This patch removes it instead and significantly reduces source and
binary size of XFS while making maintaince easier.
SGI-PV: 970841
SGI-Modid: xfs-linux-melb:xfs-kern:29737a
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Tim Shimmin <tes@sgi.com>
Simplify vnode tracing calls by embedding function name & return addr in
the calling macro.
Also do a lot of vnode->inode renaming for consistency, while we're at it.
SGI-PV: 970335
SGI-Modid: xfs-linux-melb:xfs-kern:29650a
Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
On last close of a file we purge blocks beyond eof. The same code is used
when we truncate the file size down. In this case we need to wait for any
pending I/Os for dirty pages beyond the new eof. For the last close case
we are not changing the file size and therefore do not need to wait for
any I/Os to complete. This fixes a performance bottleneck where writes
into the page cache and cache flushes can become mutually exclusive.
SGI-PV: 964002
SGI-Modid: xfs-linux-melb:xfs-kern:30220a
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: Peter Leckie <pleckie@sgi.com>
In the following scenario xfs_bulkstat() returns incorrect stale inode
state:
1. File_A is created and its inode synced to disk. 2. File_A is unlinked
and doesn't exist anymore. 3. Filesystem sync is invoked. 4. File_B is
created. File_B happens to reclaim File_A's inode. 5. xfs_bulkstat() is
called and detects File_B but reports the
incorrect File_A inode state.
Explanation for the incorrect inode state is that inodes are not
immediately synced on file create for performance reasons. This leaves the
on-disk inode buffer uninitialized (or with old state from a previous
generation inode) and this is what xfs_bulkstat() would report.
The patch marks the on-disk inode buffer "dirty" on unlink. When the inode
is reclaimed (by a new file create), xfs_bulkstat() would filter this
inode by the "dirty" mark. Once the inode is flushed to disk, the on-disk
buffer "dirty" mark is automatically removed and a following
xfs_bulkstat() would return the correct inode state.
Marking the on-disk inode buffer "dirty" on unlink is achieved by setting
the on-disk di_nlink field to 0. Note that the in-core di_nlink has
already been set to 0 and a corresponding transaction logged by
xfs_droplink(). This is an exception from the rule that any on-disk inode
buffer changes has to be followed by a disk write (inode flush).
Synchronizing the in-core to on-disk di_nlink values in advance (before
the actual inode flush to disk) should be fine in this case because the
inode is already unlinked and it would never change its di_nlink again for
this inode generation.
SGI-PV: 970842
SGI-Modid: xfs-linux-melb:xfs-kern:29757a
Signed-off-by: Vlad Apostolov <vapo@sgi.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Mark Goodwin <markgw@sgi.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
All flags are added to xfs_mount's m_flag instead. Note that the 32bit
inode flag was duplicated in both of them, but only cleared in the mount
when it was not nessecary due to the filesystem beeing small enough. Two
flags are still required here - one to indicate the mount option setting,
and one to indicate if it applies or not.
SGI-PV: 969608
SGI-Modid: xfs-linux-melb:xfs-kern:29507a
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
Also remove the now dead behavior code.
SGI-PV: 969608
SGI-Modid: xfs-linux-melb:xfs-kern:29505a
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
struct bhv_vnode is on it's way out, so move the trace buffer to the XFS
inode. Note that this makes the tracing macros rather misnamed, but this
kind of fallout will be fixed up incrementally later on.
SGI-PV: 969608
SGI-Modid: xfs-linux-melb:xfs-kern:29498a
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
struct bhv_vnode is on it's way out, so move the I/O count to the XFS
inode.
SGI-PV: 969608
SGI-Modid: xfs-linux-melb:xfs-kern:29497a
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
We can easily get at the vfsp through the super_block but it will soon be
gone anyway.
SGI-PV: 969608
SGI-Modid: xfs-linux-melb:xfs-kern:29494a
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
One of the perpetual scaling problems XFS has is indexing it's incore
inodes. We currently uses hashes and the default hash sizes chosen can
only ever be a tradeoff between memory consumption and the maximum
realistic size of the cache.
As a result, anyone who has millions of inodes cached on a filesystem
needs to tunes the size of the cache via the ihashsize mount option to
allow decent scalability with inode cache operations.
A further problem is the separate inode cluster hash, whose size is based
on the ihashsize but is smaller, and so under certain conditions (sparse
cluster cache population) this can become a limitation long before the
inode hash is causing issues.
The following patchset removes the inode hash and cluster hash and
replaces them with radix trees to avoid the scalability limitations of the
hashes. It also reduces the size of the inodes by 3 pointers....
SGI-PV: 969561
SGI-Modid: xfs-linux-melb:xfs-kern:29481a
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Tim Shimmin <tes@sgi.com>
Biggest bit is duplicating the dinode structure so we have one annotated for
native endianess and one for disk endianess. The other significant change
is that xfs_xlate_dinode_core is split into one helper per direction to
allow for proper annotations, everything else is trivial.
As a sidenode splitting out the incore dinode means we can move it into
xfs_inode.h in a later patch and severely improving on the include hell in
xfs.
SGI-PV: 968563
SGI-Modid: xfs-linux-melb:xfs-kern:29476a
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
Generally we try not to directly include linux header files in core xfs
code; xfs_linux.h is the spot for that.
SGI-PV: 968563
SGI-Modid: xfs-linux-melb:xfs-kern:29326a
Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
currently xfs_bmbt_rec_t is used both for ondisk extents as well as
host-endian ones. This patch adds a new xfs_bmbt_rec_host_t for the native
endian ones and cleans up the fallout. There have been various endianess
issues in the tracing / debug printf code that are fixed by this patch.
SGI-PV: 968563
SGI-Modid: xfs-linux-melb:xfs-kern:29318a
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
Remove the hardcoded "fnames" for tracing, and just embed them in tracing
macros via __FUNCTION__. Kills a lot of #ifdefs too.
SGI-PV: 967353
SGI-Modid: xfs-linux-melb:xfs-kern:29099a
Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
Avoid using a special "zero inode" as the parent of the quota inode as
this can confuse the filestreams code into thinking the quota inode has a
parent. We do not want the quota inode to follow filestreams allocation
rules, so pass a NULL as the parent inode and detect this condition when
doing stream associations.
SGI-PV: 964469
SGI-Modid: xfs-linux-melb:xfs-kern:29098a
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
In media spaces, video is often stored in a frame-per-file format. When
dealing with uncompressed realtime HD video streams in this format, it is
crucial that files do not get fragmented and that multiple files a placed
contiguously on disk.
When multiple streams are being ingested and played out at the same time,
it is critical that the filesystem does not cross the streams and
interleave them together as this creates seek and readahead cache miss
latency and prevents both ingest and playout from meeting frame rate
targets.
This patch set creates a "stream of files" concept into the allocator to
place all the data from a single stream contiguously on disk so that RAID
array readahead can be used effectively. Each additional stream gets
placed in different allocation groups within the filesystem, thereby
ensuring that we don't cross any streams. When an AG fills up, we select a
new AG for the stream that is not in use.
The core of the functionality is the stream tracking - each inode that we
create in a directory needs to be associated with the directories' stream.
Hence every time we create a file, we look up the directories' stream
object and associate the new file with that object.
Once we have a stream object for a file, we use the AG that the stream
object point to for allocations. If we can't allocate in that AG (e.g. it
is full) we move the entire stream to another AG. Other inodes in the same
stream are moved to the new AG on their next allocation (i.e. lazy
update).
Stream objects are kept in a cache and hold a reference on the inode.
Hence the inode cannot be reclaimed while there is an outstanding stream
reference. This means that on unlink we need to remove the stream
association and we also need to flush all the associations on certain
events that want to reclaim all unreferenced inodes (e.g. filesystem
freeze).
SGI-PV: 964469
SGI-Modid: xfs-linux-melb:xfs-kern:29096a
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Barry Naujok <bnaujok@sgi.com>
Signed-off-by: Donald Douwsma <donaldd@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Tim Shimmin <tes@sgi.com>
Signed-off-by: Vlad Apostolov <vapo@sgi.com>
The problem that has been addressed is that of synchronising updates of
the file size with writes that extend a file. Without the fix the update
of a file's size, as a result of a write beyond eof, is independent of
when the cached data is flushed to disk. Often the file size update would
be written to the filesystem log before the data is flushed to disk. When
a system crashes between these two events and the filesystem log is
replayed on mount the file's size will be set but since the contents never
made it to disk the file is full of holes. If some of the cached data was
flushed to disk then it may just be a section of the file at the end that
has holes.
There are existing fixes to help alleviate this problem, particularly in
the case where a file has been truncated, that force cached data to be
flushed to disk when the file is closed. If the system crashes while the
file(s) are still open then this flushing will never occur.
The fix that we have implemented is to introduce a second file size,
called the in-memory file size, that represents the current file size as
viewed by the user. The existing file size, called the on-disk file size,
is the one that get's written to the filesystem log and we only update it
when it is safe to do so. When we write to a file beyond eof we only
update the in- memory file size in the write operation. Later when the I/O
operation, that flushes the cached data to disk completes, an I/O
completion routine will update the on-disk file size. The on-disk file
size will be updated to the maximum offset of the I/O or to the value of
the in-memory file size if the I/O includes eof.
SGI-PV: 958522
SGI-Modid: xfs-linux-melb:xfs-kern:28322a
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
This patch handles error return values in fs_flush_pages and
fs_flushinval_pages. It changes the prototype of fs_flushinval_pages so we
can propogate the errors and handle them at higher layers. I also modified
xfs_itruncate_start so that it could propogate the error further.
SGI-PV: 961990
SGI-Modid: xfs-linux-melb:xfs-kern:28231a
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: Stewart Smith <stewart@flamingspork.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
xfs_mac.h and xfs_cap.h provide definitions and macros that aren't used
anywhere in XFS at all. They are left-overs from "to be implement at some
point in the future" functionality that Irix XFS has. If this
functionality ever goes into Linux, it will be provided at a different
layer, most likely through the security hooks in the kernel so we will
never need this functionality in XFS.
Patch provided by Eric Sandeen (sandeen@sandeen.net).
SGI-PV: 960895
SGI-Modid: xfs-linux-melb:xfs-kern:28036a
Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
The firstblock argument to xfs_bmap_finish is not used by that function.
Remove it and cleanup the code a bit.
Patch provided by Eric Sandeen.
SGI-PV: 960196
SGI-Modid: xfs-linux-melb:xfs-kern:28034a
Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
The problem is the two callers of xfs_iozero() are rounding out the range
to be zeroed to the end of a fsb and in some cases this extends past the
new eof. The call to commit_write() in xfs_iozero() will cause the Linux
inode's file size to be set too high.
SGI-PV: 960788
SGI-Modid: xfs-linux-melb:xfs-kern:28013a
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
gcc-4.1 and more recent aggressively inline static functions which
increases XFS stack usage by ~15% in critical paths. Prevent this from
occurring by adding noinline to the STATIC definition.
Also uninline some functions that are too large to be inlined and were
causing problems with CONFIG_FORCED_INLINING=y.
Finally, clean up all the different users of inline, __inline and
__inline__ and put them under one STATIC_INLINE macro. For debug kernels
the STATIC_INLINE macro uninlines those functions.
SGI-PV: 957159
SGI-Modid: xfs-linux-melb:xfs-kern:27585a
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: David Chatterton <chatz@sgi.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
The previous fixes for the use after free in xfs_iunpin left a nasty log
deadlock when xfslogd unpinned the inode and dropped the last reference to
the inode. the ->clear_inode() method can issue transactions, and if the
log was full, the transaction could push on the log and get stuck trying
to push the inode it was currently unpinning.
To fix this, we provide xfs_iunpin a guarantee that it will always have a
valid xfs_inode <-> linux inode link or a particular flag will be set on
the inode. We then use log forces during lookup to ensure transactions are
completed before we recycle the inode. This ensures that xfs_iunpin will
never use the linux inode after it is being freed, and any lookup on an
inode on the reclaim list will wait until it is safe to attach a new linux
inode to the xfs inode.
SGI-PV: 956832
SGI-Modid: xfs-linux-melb:xfs-kern:27359a
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Shailendra Tripathi <stripathi@agami.com>
Signed-off-by: Takenori Nagano <t-nagano@ah.jp.nec.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
The previous attempts to fix the linux inode use-after-free in xfs_iunpin
simply made the problem harder to hit. We actually need complete exclusion
between xfs_reclaim and xfs_iunpin, as well as ensuring that the i_flags
are consistent during both of these functions. Introduce a new spinlock
for exclusion and the i_flags, and fix up xfs_iunpin to use igrab before
marking the inode dirty.
SGI-PV: 952967
SGI-Modid: xfs-linux-melb:xfs-kern:26964a
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
is check if semaphore is actually locked, which can be trivially done in
portable way. Code gets more reabable, while we are at it...
SGI-PV: 953915
SGI-Modid: xfs-linux-melb:xfs-kern:26274a
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Nathan Scott <nathans@sgi.com>
transaction completion from marking the inode dirty while it is being
cleaned up on it's way out of the system.
SGI-PV: 952967
SGI-Modid: xfs-linux-melb:xfs-kern:26040a
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Nathan Scott <nathans@sgi.com>
the range spanned by modifications to the in-core extent map. Add
XFS_BUNMAPI() and XFS_SWAP_EXTENTS() macros that call xfs_bunmapi() and
xfs_swap_extents() via the ioops vector. Change all calls that may modify
the in-core extent map for the data fork to go through the ioops vector.
This allows a cache of extent map data to be kept in sync.
SGI-PV: 947615
SGI-Modid: xfs-linux-melb:xfs-kern:209226a
Signed-off-by: Olaf Weber <olaf@sgi.com>
Signed-off-by: Nathan Scott <nathans@sgi.com>
Christoph Hellwig