- reshuffle various conditionals for data vs attr fork to make the code
more readable
- do fine-grainded goto-based error handling
- exit early from conditionals instead of keeping a long else branch around
- allow kmem_alloc to fail
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Eric Sandeen <sandeen@sandeen.net> Reviewed-by: Felix Blyakher <felixb@sgi.com> Signed-off-by: Felix Blyakher <felixb@sgi.com>
Olaf Weber [Fri, 17 Apr 2009 21:12:45 +0000 (16:12 -0500)]
xfs: add more checks to superblock validation
There had been reports where xfs filesystem was randomly
corrupted with fsfuzzer, and xfs failed to handle it
gracefully. This patch fixes couple of reported problem
by providing additional checks in the superblock
validation routine.
Signed-off-by: Olaf Weber <olaf@sgi.com> Reviewed-by: Josef 'Jeff' Sipek <jeffpc@josefsipek.net> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Felix Blyakher <felixb@sgi.com>
The problem here is that xfs_file_last_byte() does not acquire the
inode lock and can therefore race with another thread that is modifying
the extext list. While xfs_bmap_last_offset() is trying to lookup
what was the last extent some extents were merged and the extent list
shrunk so the index we lookup is now beyond the end of the extent list
and potentially in a freed buffer.
Signed-off-by: Lachlan McIlroy <lmcilroy@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Felix Blyakher <felixb@sgi.com> Signed-off-by: Felix Blyakher <felixb@sgi.com>
Dave Chinner [Mon, 6 Apr 2009 16:49:12 +0000 (18:49 +0200)]
xfs: remove xfs_flush_space
The only thing we need to do now when we get an ENOSPC condition during delayed
allocation reservation is flush all the other inodes with delalloc blocks on
them and retry without EOF preallocation. Remove the unneeded mess that is
xfs_flush_space() and just call xfs_flush_inodes() directly from
xfs_iomap_write_delay().
Also, change the location of the retry label to avoid trying to do EOF
preallocation because we don't want to do that at ENOSPC. This enables us to
remove the BMAPI_SYNC flag as it is no longer used.
Signed-off-by: Dave Chinner <david@fromorbit.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Dave Chinner [Mon, 6 Apr 2009 16:48:30 +0000 (18:48 +0200)]
xfs: flush delayed allcoation blocks on ENOSPC in create
If we are creating lots of small files, we can fail to get
a reservation for inode create earlier than we should due to
EOF preallocation done during delayed allocation reservation.
Hence on the first reservation ENOSPC failure flush all the
delayed allocation blocks out of the system and retry.
This fixes the last commonly triggered spurious ENOSPC issue
that has been reported.
Signed-off-by: Dave Chinner <david@fromorbit.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Dave Chinner [Mon, 6 Apr 2009 16:47:27 +0000 (18:47 +0200)]
xfs: block callers of xfs_flush_inodes() correctly
xfs_flush_inodes() currently uses a magic timeout to wait for
some inodes to be flushed before returning. This isn't
really reliable but used to be the best that could be done
due to deadlock potential of waiting for the entire flush.
Now the inode flush is safe to execute while we hold page
and inode locks, we can wait for all the inodes to flush
synchronously. Convert the wait mechanism to a completion
to do this efficiently. This should remove all remaining
spurious ENOSPC errors from the delayed allocation reservation
path.
This is extracted almost line for line from a larger patch
from Mikulas Patocka.
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Dave Chinner [Mon, 6 Apr 2009 16:45:44 +0000 (18:45 +0200)]
xfs: make inode flush at ENOSPC synchronous
When we are writing to a single file and hit ENOSPC, we trigger a background
flush of the inode and try again. Because we hold page locks and the iolock,
the flush won't proceed until after we release these locks. This occurs once
we've given up and ENOSPC has been reported. Hence if this one is the only
dirty inode in the system, we'll get an ENOSPC prematurely.
To fix this, remove the async flush from the allocation routines and move
it to the top of the write path where we can do a synchronous flush
and retry the write again. Only retry once as a second ENOSPC indicates
that we really are ENOSPC.
This avoids a page cache deadlock when trying to do this flush synchronously
in the allocation layer that was identified by Mikulas Patocka.
Signed-off-by: Dave Chinner <david@fromorbit.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Dave Chinner [Mon, 6 Apr 2009 16:44:54 +0000 (18:44 +0200)]
xfs: use xfs_sync_inodes() for device flushing
Currently xfs_device_flush calls sync_blockdev() which is
a no-op for XFS as all it's metadata is held in a different
address to the one sync_blockdev() works on.
Call xfs_sync_inodes() instead to flush all the delayed
allocation blocks out. To do this as efficiently as possible,
do it via two passes - one to do an async flush of all the
dirty blocks and a second to wait for all the IO to complete.
This requires some modification to the xfs-sync_inodes_ag()
flush code to do efficiently.
Signed-off-by: Dave Chinner <david@fromorbit.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Dave Chinner [Mon, 6 Apr 2009 16:42:59 +0000 (18:42 +0200)]
xfs: inform the xfsaild of the push target before sleeping
When trying to reserve log space, we find the amount of space
we need, then go to sleep waiting for space. When we are
woken, we try to push the tail of the log forward to make
sure we have space available.
Unfortunately, this means that if there is not space available, and
everyone who needs space goes to sleep there is no-one left to push
the tail of the log to make space available. Once we have a thread
waiting for space to become available, the others queue up behind
it in a FIFO, and none of them push the tail of the log.
This can result in everyone going to sleep in xlog_grant_log_space()
if the first sleeper races with the last I/O that moves the tail
of the log forward. With no further I/O tomove the tail of the log,
there is nothing to wake the sleepers and hence all transactions
just stop.
Fix this by making sure the xfsaild will create enough space for the
transaction that is about to sleep by moving the push target far
enough forwards to ensure that that the curent proceeees will have
enough space available when it is woken. That is, we push the
AIL before we go to sleep.
Because we've inserted the log ticket into the queue before we've
pushed and gone to sleep, subsequent transactions will wait behind
this one. Hence we are guaranteed to have space available when we
are woken.
Signed-off-by: Dave Chinner <david@fromorbit.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Dave Chinner [Mon, 6 Apr 2009 16:42:11 +0000 (18:42 +0200)]
xfs: prevent unwritten extent conversion from blocking I/O completion
Unwritten extent conversion can recurse back into the filesystem due
to memory allocation. Memory reclaim requires I/O completions to be
processed to allow the callers to make progress. If the I/O
completion workqueue thread is doing the recursion, then we have a
deadlock situation.
Move unwritten extent completion into it's own workqueue so it
doesn't block I/O completions for normal delayed allocation or
overwrite data.
Signed-off-by: Dave Chinner <david@fromorbit.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Dave Chinner [Mon, 6 Apr 2009 16:40:17 +0000 (18:40 +0200)]
xfs: fix double free of inode
If we fail to initialise the VFS inode in inode_init_always(),
it will call ->delete_inode internally resulting in the inode being
freed. Hence we need to delay the call to inode_init_always()
until after the XFS inode is sufficient set up to handle a
call to ->delete_inode, and then if that fails do not touch
the inode again at all as it has been freed.
Signed-off-by: Dave Chinner <david@fromorbit.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
Dave Chinner [Mon, 6 Apr 2009 16:39:27 +0000 (18:39 +0200)]
xfs: validate log feature fields correctly
If the large log sector size feature bit is set in the
superblock by accident (say disk corruption), the then
fields that are now considered valid are not checked on
production kernels. The checks are present as ASSERT
statements so cause a panic on a debug kernel.
Change this so that the fields are validity checked if
the feature bit is set and abort the log mount if the
fields do not contain valid values.
Reported-by: Eric Sesterhenn <snakebyte@gmx.de> Signed-off-by: Dave Chinner <david@fromorbit.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6:
wireless: remove duplicated .ndo_set_mac_address
netfilter: xtables: fix IPv6 dependency in the cluster match
tg3: Add GRO support.
niu: Add GRO support.
ucc_geth: Fix use-after-of_node_put() in ucc_geth_probe().
gianfar: Fix use-after-of_node_put() in gfar_of_init().
kernel: remove HIPQUAD()
netpoll: store local and remote ip in net-endian
netfilter: fix endian bug in conntrack printks
dmascc: fix incomplete conversion to network_device_ops
gso: Fix support for linear packets
skbuff.h: fix missing kernel-doc
ni5010: convert to net_device_ops
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6:
sparc64: Fix reset hangs on Niagara systems.
cpumask: use mm_cpumask() wrapper: sparc
cpumask: remove dangerous CPU_MASK_ALL_PTR, &CPU_MASK_ALL.: sparc
cpumask: remove the now-obsoleted pcibus_to_cpumask(): sparc
cpumask: remove cpu_coregroup_map: sparc
cpumask: prepare for iterators to only go to nr_cpu_ids/nr_cpumask_bits.: sparc
cpumask: prepare for iterators to only go to nr_cpu_ids/nr_cpumask_bits.: sparc64
cpumask: Use accessors code.: sparc64
cpumask: Use accessors code: sparc
cpumask: arch_send_call_function_ipi_mask: sparc
cpumask: Use smp_call_function_many(): sparc64
* git://git.kernel.org/pub/scm/linux/kernel/git/rusty/linux-2.6-lguest-and-virtio:
lguest: barrier me harder
lguest: use bool instead of int
lguest: use KVM hypercalls
lguest: wire up pte_update/pte_update_defer
lguest: fix spurious BUG_ON() on invalid guest stack.
virtio: more neatening of virtio_ring macros.
virtio: fix BAD_RING, START_US and END_USE macros
Linus Torvalds [Tue, 31 Mar 2009 00:54:32 +0000 (17:54 -0700)]
Merge branch 'hwmon-for-linus' of git://jdelvare.pck.nerim.net/jdelvare-2.6
* 'hwmon-for-linus' of git://jdelvare.pck.nerim.net/jdelvare-2.6:
hwmon: (fschmd) Add support for the FSC Hades IC
hwmon: (fschmd) Add support for the FSC Syleus IC
i2c-i801: Instantiate FSC hardware montioring chips
dmi: Let dmi_walk() users pass private data
hwmon: Define a standard interface for chassis intrusion detection
Move the pcf8591 driver to hwmon
hwmon: (w83627ehf) Only expose in6 or temp3 on the W83667HG
hwmon: (w83627ehf) Add support for W83667HG
hwmon: (w83627ehf) Invert fan pin variables logic
hwmon: (hdaps) Fix Thinkpad X41 axis inversion
hwmon: (hdaps) Allow inversion of separate axis
hwmon: (ds1621) Clean up documentation
hwmon: (ds1621) Avoid unneeded register access
hwmon: (ds1621) Clean up register access
hwmon: (ds1621) Reorder code statements
Linus Torvalds [Mon, 30 Mar 2009 23:06:04 +0000 (16:06 -0700)]
Merge branch 'proc-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/adobriyan/proc
* 'proc-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/adobriyan/proc:
Revert "proc: revert /proc/uptime to ->read_proc hook"
proc 2/2: remove struct proc_dir_entry::owner
proc 1/2: do PDE usecounting even for ->read_proc, ->write_proc
proc: fix sparse warnings in pagemap_read()
proc: move fs/proc/inode-alloc.txt comment into a source file
Linus Torvalds [Mon, 30 Mar 2009 22:12:14 +0000 (15:12 -0700)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/suspend-2.6
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/suspend-2.6:
PCI PM: Make pci_prepare_to_sleep() disable wake-up if needed
radeonfb: Use __pci_complete_power_transition()
PCI PM: Introduce __pci_[start|complete]_power_transition() (rev. 2)
PCI PM: Restore config spaces of all devices during early resume
PCI PM: Make pci_set_power_state() handle devices with no PM support
PCI PM: Put devices into low power states during late suspend (rev. 2)
PCI PM: Move pci_restore_standard_config to pci-driver.c
PCI PM: Use pci_set_power_state during early resume
PCI PM: Consistently use variable name "error" for pm call return values
kexec: Change kexec jump code ordering
PM: Change hibernation code ordering
PM: Change suspend code ordering
PM: Rework handling of interrupts during suspend-resume
PM: Introduce functions for suspending and resuming device interrupts
Randy Dunlap [Mon, 30 Mar 2009 21:08:44 +0000 (14:08 -0700)]
dma-debug: fix printk formats (i386)
Fix printk format warnings in dma-debug:
lib/dma-debug.c:645: warning: format '%016llx' expects type 'long long unsigned int', but argument 6 has type 'dma_addr_t'
lib/dma-debug.c:662: warning: format '%016llx' expects type 'long long unsigned int', but argument 6 has type 'dma_addr_t'
lib/dma-debug.c:676: warning: format '%016llx' expects type 'long long unsigned int', but argument 6 has type 'dma_addr_t'
lib/dma-debug.c:686: warning: format '%016llx' expects type 'long long unsigned int', but argument 6 has type 'dma_addr_t'
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Fix this build error when REISERFS_FS_POSIX_ACL is not set:
fs/reiserfs/inode.c: In function 'reiserfs_new_inode':
fs/reiserfs/inode.c:1919: warning: passing argument 1 of 'reiserfs_inherit_default_acl' from incompatible pointer type
fs/reiserfs/inode.c:1919: warning: passing argument 2 of 'reiserfs_inherit_default_acl' from incompatible pointer type
fs/reiserfs/inode.c:1919: warning: passing argument 3 of 'reiserfs_inherit_default_acl' from incompatible pointer type
fs/reiserfs/inode.c:1919: error: too many arguments to function 'reiserfs_inherit_default_acl'
due to a missing transaction-handle argument in the non-acl
compatibility function.
Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com> Acked-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Alexey Dobriyan [Wed, 25 Mar 2009 19:48:06 +0000 (22:48 +0300)]
proc 2/2: remove struct proc_dir_entry::owner
Setting ->owner as done currently (pde->owner = THIS_MODULE) is racy
as correctly noted at bug #12454. Someone can lookup entry with NULL
->owner, thus not pinning enything, and release it later resulting
in module refcount underflow.
We can keep ->owner and supply it at registration time like ->proc_fops
and ->data.
But this leaves ->owner as easy-manipulative field (just one C assignment)
and somebody will forget to unpin previous/pin current module when
switching ->owner. ->proc_fops is declared as "const" which should give
some thoughts.
->read_proc/->write_proc were just fixed to not require ->owner for
protection.
rmmod'ed directories will be empty and return "." and ".." -- no harm.
And directories with tricky enough readdir and lookup shouldn't be modular.
We definitely don't want such modular code.
Removing ->owner will also make PDE smaller.
So, let's nuke it.
Kudos to Jeff Layton for reminding about this, let's say, oversight.
Alexey Dobriyan [Fri, 20 Feb 2009 14:04:33 +0000 (17:04 +0300)]
proc 1/2: do PDE usecounting even for ->read_proc, ->write_proc
struct proc_dir_entry::owner is going to be removed. Now it's only necessary
to protect PDEs which are using ->read_proc, ->write_proc hooks.
However, ->owner assignments are racy and make it very easy for someone to switch
->owner on live PDE (as some subsystems do) without fixing refcounts and so on.
http://bugzilla.kernel.org/show_bug.cgi?id=12454
So, ->owner is on death row.
Proxy file operations exist already (proc_file_operations), just bump usecount
when necessary.
fs/proc/task_mmu.c:696:12: warning: cast removes address space of expression
fs/proc/task_mmu.c:696:9: warning: incorrect type in assignment (different address spaces)
fs/proc/task_mmu.c:696:9: expected unsigned long long [noderef] [usertype] <asn:1>*out
fs/proc/task_mmu.c:696:9: got unsigned long long [usertype] *<noident>
fs/proc/task_mmu.c:697:12: warning: cast removes address space of expression
fs/proc/task_mmu.c:697:9: warning: incorrect type in assignment (different address spaces)
fs/proc/task_mmu.c:697:9: expected unsigned long long [noderef] [usertype] <asn:1>*end
fs/proc/task_mmu.c:697:9: got unsigned long long [usertype] *<noident>
fs/proc/task_mmu.c:723:12: warning: cast removes address space of expression
fs/proc/task_mmu.c:723:26: error: subtraction of different types can't work (different address spaces)
fs/proc/task_mmu.c:725:24: error: subtraction of different types can't work (different address spaces)
mm/slob.c:469: error: ‘flags’ undeclared (first use in this function)
mm/slob.c:469: error: (Each undeclared identifier is reported only once
mm/slob.c:469: error: for each function it appears in.)
Cc: Nick Piggin <npiggin@suse.de> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <20090128135457.350751756@chello.nl> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Linus Torvalds [Mon, 30 Mar 2009 20:54:50 +0000 (13:54 -0700)]
Merge branch 'drm-next' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6
* 'drm-next' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6: (53 commits)
drm: detect hdmi monitor by hdmi identifier (v3)
drm: drm_fops.c unlock missing on error path
drm: reorder struct drm_ioctl_desc to save space on 64 bit builds
radeon: add some new pci ids
drm: read EDID extensions from monitor
drm: Use a little stash on the stack to avoid kmalloc in most DRM ioctls.
drm/radeon: add regs required for occlusion queries support
drm/i915: check the return value from the copy from user
drm/radeon: fix logic in r600_page_table_init() to match ati_gart
drm/radeon: r600 ptes are 64-bit, cleanup cleanup function.
drm/radeon: don't call irq changes on r600 suspend/resume
drm/radeon: fix r600 writeback across suspend/resume
drm/radeon: fix r600 writeback setup.
drm: fix warnings about new mappings in info code.
drm/radeon: NULL noise: drivers/gpu/drm/radeon/radeon_*.c
drm/radeon: fix r600 pci mapping calls.
drm/radeon: r6xx/r7xx: fix possible oops in r600_page_table_cleanup()
radeon: call the correct idle function, logic got inverted.
drm/radeon: RS600: fix interrupt handling
drm/r600: fix rptr address along lines of previous fixes to radeon.
...
Linus Torvalds [Mon, 30 Mar 2009 20:41:00 +0000 (13:41 -0700)]
Merge branch 'iommu-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
* 'iommu-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (60 commits)
dma-debug: make memory range checks more consistent
dma-debug: warn of unmapping an invalid dma address
dma-debug: fix dma_debug_add_bus() definition for !CONFIG_DMA_API_DEBUG
dma-debug/x86: register pci bus for dma-debug leak detection
dma-debug: add a check dma memory leaks
dma-debug: add checks for kernel text and rodata
dma-debug: print stacktrace of mapping path on unmap error
dma-debug: Documentation update
dma-debug: x86 architecture bindings
dma-debug: add function to dump dma mappings
dma-debug: add checks for sync_single_sg_*
dma-debug: add checks for sync_single_range_*
dma-debug: add checks for sync_single_*
dma-debug: add checking for [alloc|free]_coherent
dma-debug: add add checking for map/unmap_sg
dma-debug: add checking for map/unmap_page/single
dma-debug: add core checking functions
dma-debug: add debugfs interface
dma-debug: add kernel command line parameters
dma-debug: add initialization code
...
Fix trivial conflicts due to whitespace changes in arch/x86/kernel/pci-nommu.c
PCI PM: Make pci_prepare_to_sleep() disable wake-up if needed
If the device is not supposed to wake up the system, ie. when
device_may_wakeup(&dev->dev) returns 'false', pci_prepare_to_sleep()
should pass 'false' to pci_enable_wake() so that it calls the
platform to disable the wake-up capability of the device.
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl> Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
The radeonfb driver needs to program the device's PMCSR directly due
to some quirky hardware it has to handle (see
http://bugzilla.kernel.org/show_bug.cgi?id=12846 for details) and
after doing that it needs to call the platform (usually ACPI) to
finish the power transition of the device. Currently it uses
pci_set_power_state() for this purpose, however making a specific
assumption about the internal behavior of this function, which has
changed recently so that this assumption is no longer satisfied.
For this reason, introduce __pci_complete_power_transition() that may
be called by the radeonfb driver to complete the power transition of
the device. For symmetry, introduce __pci_start_power_transition().
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl> Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
PCI PM: Restore config spaces of all devices during early resume
At present the configuration spaces of PCI devices that have no
drivers or no PM support in the drivers (either legacy or through a
pm object) are not saved during suspend and, consequently, they are
not restored during resume. This generally may lead to the state of
the system being slightly inconsistent after the resume, so it's
better to save and restore the configuration spaces of these devices
as well.
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl> Acked-by: Ingo Molnar <mingo@elte.hu> Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
PCI PM: Make pci_set_power_state() handle devices with no PM support
There is a problem with PCI devices without any PM support (either
native or through the platform) that pci_set_power_state() always
returns error code for them, even if they are being put into D0.
However, such devices are always in D0, so pci_set_power_state()
should return success when attempting to put such a device into D0.
It also should update the current_state field for these devices as
appropriate. This modification is necessary so that the standard
configuration registers of these devices are successfully restored by
pci_restore_standard_config() during the "early" phase of resume.
In addition, pci_set_power_state() should check the value of
current_state before calling the platform to change the power state
of the device to avoid doing that unnecessarily.
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl> Acked-by: Ingo Molnar <mingo@elte.hu> Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
PCI PM: Put devices into low power states during late suspend (rev. 2)
Once we have allowed timer interrupts to be enabled during the late
phase of suspending devices, we are now able to use the generic
pci_set_power_state() to put PCI devices into low power states at
that time. We can also use some related platform callbacks, like the
ones preparing devices for wake-up, during the late suspend.
Doing this will allow us to avoid the race condition where a device
using shared interrupts is put into a low power state with interrupts
enabled and then an interrupt (for another device) comes in and
confuses its driver. At the same time, devices that don't support
the native PCI PM or that require some additional, platform-specific
operations to be carried out to put them into low power states will
be handled as appropriate.
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl> Acked-by: Ingo Molnar <mingo@elte.hu> Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
PCI PM: Use pci_set_power_state during early resume
Once we have allowed timer interrupts to be enabled during the early
phase of resuming devices, we are now able to use the generic
pci_set_power_state() to put PCI devices into D0 at that time. Then,
the platform-specific PM code will have a chance to handle devices
that don't implement the native PCI PM or that require some
additional, platform-specific operations to be carried out to power
them up. Also, by doing this we can simplify the code quite a bit.
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl> Acked-by: Ingo Molnar <mingo@elte.hu> Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Frans Pop [Mon, 16 Mar 2009 21:39:56 +0000 (22:39 +0100)]
PCI PM: Consistently use variable name "error" for pm call return values
I noticed two functions use a variable "i" to store the return value of PM
function calls while the rest of the file uses "error". As "i" normally
indicates a counter of some sort it seems better to keep this consistent.
Signed-off-by: Frans Pop <elendil@planet.nl> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Change the ordering of the hibernation core code so that the platform
"prepare" callbacks are executed and the nonboot CPUs are disabled
after calling device drivers' "late suspend" methods.
This change (along with the previous analogous change of the suspend
core code) will allow us to rework the PCI PM core so that the power
state of devices is changed in the "late" phase of suspend (and
analogously in the "early" phase of resume), which in turn will allow
us to avoid the race condition where a device using shared interrupts
is put into a low power state with interrupts enabled and then an
interrupt (for another device) comes in and confuses its driver.
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl> Acked-by: Ingo Molnar <mingo@elte.hu>
Change the ordering of the suspend core code so that the platform
"prepare" callback is executed and the nonboot CPUs are disabled
after calling device drivers' "late suspend" methods.
This change will allow us to rework the PCI PM core so that the power
state of devices is changed in the "late" phase of suspend (and
analogously in the "early" phase of resume), which in turn will allow
us to avoid the race condition where a device using shared interrupts
is put into a low power state with interrupts enabled and then an
interrupt (for another device) comes in and confuses its driver.
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl> Acked-by: Ingo Molnar <mingo@elte.hu>
PM: Rework handling of interrupts during suspend-resume
Use the functions introduced in by the previous patch,
suspend_device_irqs(), resume_device_irqs() and check_wakeup_irqs(),
to rework the handling of interrupts during suspend (hibernation) and
resume. Namely, interrupts will only be disabled on the CPU right
before suspending sysdevs, while device drivers will be prevented
from receiving interrupts, with the help of the new helper function,
before their "late" suspend callbacks run (and analogously during
resume).
In addition, since the device interrups are now disabled before the
CPU has turned all interrupts off and the CPU will ACK the interrupts
setting the IRQ_PENDING bit for them, check in sysdev_suspend() if
any wake-up interrupts are pending and abort suspend if that's the
case.
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl> Acked-by: Ingo Molnar <mingo@elte.hu>
PM: Introduce functions for suspending and resuming device interrupts
Introduce helper functions allowing us to prevent device drivers from
getting any interrupts (without disabling interrupts on the CPU)
during suspend (or hibernation) and to make them start to receive
interrupts again during the subsequent resume. These functions make it
possible to keep timer interrupts enabled while the "late" suspend and
"early" resume callbacks provided by device drivers are being
executed. In turn, this allows device drivers' "late" suspend and
"early" resume callbacks to sleep, execute ACPI callbacks etc.
The functions introduced here will be used to rework the handling of
interrupts during suspend (hibernation) and resume. Namely,
interrupts will only be disabled on the CPU right before suspending
sysdevs, while device drivers will be prevented from receiving
interrupts, with the help of the new helper function, before their
"late" suspend callbacks run (and analogously during resume).
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl> Acked-by: Ingo Molnar <mingo@elte.hu>
Detect various FSC hwmon IC's based on DMI tables and then let
the i2c-i801 driver instantiate the i2c client devices. Note that
some of the info in the added table is indentical for all rows, still
this is kept in the table to keep the code general and thus (hopefully)
easily extensible in the future.
Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Jean Delvare <khali@linux-fr.org>
Jean Delvare [Mon, 30 Mar 2009 19:46:44 +0000 (21:46 +0200)]
dmi: Let dmi_walk() users pass private data
At the moment, dmi_walk() lacks flexibility, users can't pass data to
the callback function. Add a pointer for private data to make this
function more flexible.
Signed-off-by: Jean Delvare <khali@linux-fr.org> Cc: Hans de Goede <hdegoede@redhat.com> Cc: Matthew Garrett <mjg@redhat.com> Cc: Roland Dreier <rolandd@cisco.com>
Jean Delvare [Mon, 30 Mar 2009 19:46:44 +0000 (21:46 +0200)]
hwmon: Define a standard interface for chassis intrusion detection
Define a standard interface for the chassis intrusion detection feature
some hardware monitoring chips have. Some drivers have custom sysfs
entries for it, but a standard interface would allow integration with
user-space (namely libsensors.)
Signed-off-by: Jean Delvare <khali@linux-fr.org> Acked-by: Hans de Goede <j.w.r.degoede@hhs.nl> Acked-by: Matt Roberds <mattroberds@cox.net>
Jean Delvare [Mon, 30 Mar 2009 19:46:43 +0000 (21:46 +0200)]
Move the pcf8591 driver to hwmon
Directory drivers/i2c/chips is going away, so drivers there must find
new homes. For the pcf8591 driver, the best choice seems to be the
hwmon subsystem. While the Philips PCF8591 device isn't a typical
hardware monitoring chip, its DAC interface is compatible with the
hwmon one, so it fits somewhat.
If a better subsystem is ever created for ADC/DAC chips, the driver
could be moved there.
Signed-off-by: Jean Delvare <khali@linux-fr.org> Cc: Aurelien Jarno <aurelien@aurel32.net>
Gong Jun [Mon, 30 Mar 2009 19:46:43 +0000 (21:46 +0200)]
hwmon: (w83627ehf) Only expose in6 or temp3 on the W83667HG
The pin for in6 and temp3 is shared on the W83667HG, so only one of
these features can be supported on any given system. Let the driver
select which one depending on the temp3 disabled bit.
Signed-off-by: Gong Jun <JGong@nuvoton.com> Signed-off-by: Jean Delvare <khali@linux-fr.org>
Gong Jun [Mon, 30 Mar 2009 19:46:42 +0000 (21:46 +0200)]
hwmon: (w83627ehf) Add support for W83667HG
Add initial support for the Nuvoton W83667HG chip to the w83627ehf
driver. It has been tested on ASUS P5QL PRO by Gong Jun.
At the moment there is still a usability issue which is that only in6
or temp3 can be present on the W83667HG, so the driver shouldn't
expose both. This will be addressed later.
Signed-off-by: Gong Jun <JGong@nuvoton.com> Acked-by: David Hubbard <david.c.hubbard@gmail.com> Signed-off-by: Jean Delvare <khali@linux-fr.org>
Jean Delvare [Mon, 30 Mar 2009 19:46:41 +0000 (21:46 +0200)]
hwmon: (ds1621) Clean up documentation
* The alarms sysfs file is deprecated, and individual alarm files are
self-explanatory.
* The driver doesn't implement high-reslution temperature readings so
don't document that.
Signed-off-by: Jean Delvare <khali@linux-fr.org> Cc: Aurelien Jarno <aurelien@aurel32.net>
Jean Delvare [Mon, 30 Mar 2009 19:46:40 +0000 (21:46 +0200)]
hwmon: (ds1621) Avoid unneeded register access
Register access over SMBus isn't cheap, so avoid register access where
possible:
* Only write back the configuration register if it changed.
* Don't refresh the register cache when we don't have to.
Signed-off-by: Jean Delvare <khali@linux-fr.org> Cc: Aurelien Jarno <aurelien@aurel32.net>
Jean Delvare [Mon, 30 Mar 2009 19:46:40 +0000 (21:46 +0200)]
hwmon: (ds1621) Clean up register access
Fix a few oddities in how the ds1621 driver accesses the registers:
* We don't need a wrapper to access the configuration register.
* Check for error before calling swab16. Error checking isn't
complete yet, but that's a start.
* Device-specific read functions should never be called during
detection, as by definition we don't know what device we are talking
to at that point.
* Likewise, don't assume that register reads succeed during detection.
Signed-off-by: Jean Delvare <khali@linux-fr.org> Cc: Aurelien Jarno <aurelien@aurel32.net>
Linus Torvalds [Mon, 30 Mar 2009 19:29:21 +0000 (12:29 -0700)]
Merge branch 'reiserfs-updates' from Jeff Mahoney
* reiserfs-updates: (35 commits)
reiserfs: rename [cn]_* variables
reiserfs: rename p_._ variables
reiserfs: rename p_s_tb to tb
reiserfs: rename p_s_inode to inode
reiserfs: rename p_s_bh to bh
reiserfs: rename p_s_sb to sb
reiserfs: strip trailing whitespace
reiserfs: cleanup path functions
reiserfs: factor out buffer_info initialization
reiserfs: add atomic addition of selinux attributes during inode creation
reiserfs: use generic readdir for operations across all xattrs
reiserfs: journaled xattrs
reiserfs: use generic xattr handlers
reiserfs: remove i_has_xattr_dir
reiserfs: make per-inode xattr locking more fine grained
reiserfs: eliminate per-super xattr lock
reiserfs: simplify xattr internal file lookups/opens
reiserfs: Clean up xattrs when REISERFS_FS_XATTR is unset
reiserfs: remove IS_PRIVATE helpers
reiserfs: remove link detection code
...
Fixed up conflicts manually due to:
- quota name cleanups vs variable naming changes:
fs/reiserfs/inode.c
fs/reiserfs/namei.c
fs/reiserfs/stree.c
fs/reiserfs/xattr.c
- exported include header cleanups
include/linux/reiserfs_fs.h
Jeff Mahoney [Mon, 30 Mar 2009 18:02:50 +0000 (14:02 -0400)]
reiserfs: rename [cn]_* variables
This patch renames n_, c_, etc variables to something more sane. This
is the sixth in a series of patches to rip out some of the awful
variable naming in reiserfs.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Jeff Mahoney [Mon, 30 Mar 2009 18:02:49 +0000 (14:02 -0400)]
reiserfs: rename p_._ variables
This patch is a simple s/p_._//g to the reiserfs code. This is the
fifth in a series of patches to rip out some of the awful variable
naming in reiserfs.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Jeff Mahoney [Mon, 30 Mar 2009 18:02:48 +0000 (14:02 -0400)]
reiserfs: rename p_s_tb to tb
This patch is a simple s/p_s_tb/tb/g to the reiserfs code. This is the
fourth in a series of patches to rip out some of the awful variable
naming in reiserfs.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Jeff Mahoney [Mon, 30 Mar 2009 18:02:47 +0000 (14:02 -0400)]
reiserfs: rename p_s_inode to inode
This patch is a simple s/p_s_inode/inode/g to the reiserfs code. This
is the third in a series of patches to rip out some of the awful
variable naming in reiserfs.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Jeff Mahoney [Mon, 30 Mar 2009 18:02:46 +0000 (14:02 -0400)]
reiserfs: rename p_s_bh to bh
This patch is a simple s/p_s_bh/bh/g to the reiserfs code. This is the
second in a series of patches to rip out some of the awful variable
naming in reiserfs.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Jeff Mahoney [Mon, 30 Mar 2009 18:02:45 +0000 (14:02 -0400)]
reiserfs: rename p_s_sb to sb
This patch is a simple s/p_s_sb/sb/g to the reiserfs code. This is the
first in a series of patches to rip out some of the awful variable
naming in reiserfs.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Jeff Mahoney [Mon, 30 Mar 2009 18:02:41 +0000 (14:02 -0400)]
reiserfs: add atomic addition of selinux attributes during inode creation
Some time ago, some changes were made to make security inode attributes
be atomically written during inode creation. ReiserFS fell behind in
this area, but with the reworking of the xattr code, it's now fairly
easy to add.
The following patch adds the ability for security attributes to be added
automatically during inode creation.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Jeff Mahoney [Mon, 30 Mar 2009 18:02:40 +0000 (14:02 -0400)]
reiserfs: use generic readdir for operations across all xattrs
The current reiserfs xattr implementation open codes reiserfs_readdir
and frees the path before calling the filldir function. Typically, the
filldir function is something that modifies the file system, such as a
chown or an inode deletion that also require reading of an inode
associated with each direntry. Since the file system is modified, the
path retained becomes invalid for the next run. In addition, it runs
backwards in attempt to minimize activity.
This is clearly suboptimal from a code cleanliness perspective as well
as performance-wise.
This patch implements a generic reiserfs_for_each_xattr that uses the
generic readdir and a specific filldir routine that simply populates an
array of dentries and then performs a specific operation on them. When
all files have been operated on, it then calls the operation on the
directory itself.
The result is a noticable code reduction and better performance.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Jeff Mahoney [Mon, 30 Mar 2009 18:02:39 +0000 (14:02 -0400)]
reiserfs: journaled xattrs
Deadlocks are possible in the xattr code between the journal lock and the
xattr sems.
This patch implements journalling for xattr operations. The benefit is
twofold:
* It gets rid of the deadlock possibility by always ensuring that xattr
write operations are initiated inside a transaction.
* It corrects the problem where xattr backing files aren't considered any
differently than normal files, despite the fact they are metadata.
I discussed the added journal load with Chris Mason, and we decided that
since xattrs (versus other journal activity) is fairly rare, the introduction
of larger transactions to support journaled xattrs wouldn't be too big a deal.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Jeff Mahoney [Mon, 30 Mar 2009 18:02:38 +0000 (14:02 -0400)]
reiserfs: use generic xattr handlers
Christoph Hellwig had asked me quite some time ago to port the reiserfs
xattrs to the generic xattr interface.
This patch replaces the reiserfs-specific xattr handling code with the
generic struct xattr_handler.
However, since reiserfs doesn't split the prefix and name when accessing
xattrs, it can't leverage generic_{set,get,list,remove}xattr without
needlessly reconstructing the name on the back end.
Update 7/26/07: Added missing dput() to deletion path.
Update 8/30/07: Added missing mark_inode_dirty when i_mode is used to
represent an ACL and no previous ACL existed.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Jeff Mahoney [Mon, 30 Mar 2009 18:02:36 +0000 (14:02 -0400)]
reiserfs: make per-inode xattr locking more fine grained
The per-inode locking can be made more fine-grained to surround just the
interaction with the filesystem itself. This really only applies to
protecting reads during a write, since concurrent writes are barred with
inode->i_mutex at the vfs level.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The xattr file open/lookup code is needlessly complex. We can use
vfs-level operations to perform the same work, and also simplify the
locking constraints. The locking advantages will be exploited in future
patches.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Jeff Mahoney [Mon, 30 Mar 2009 18:02:33 +0000 (14:02 -0400)]
reiserfs: Clean up xattrs when REISERFS_FS_XATTR is unset
The current reiserfs xattr implementation will not clean up old xattr
files if files are deleted when REISERFS_FS_XATTR is unset. This
results in inaccessible lost files, wasting space.
This patch compiles in basic xattr knowledge, such as how to delete them
and change ownership for quota tracking. If the file system has never
used xattrs, then the operation is quite fast: it returns immediately
when it sees there is no .reiserfs_priv directory.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Jeff Mahoney [Mon, 30 Mar 2009 18:02:32 +0000 (14:02 -0400)]
reiserfs: remove IS_PRIVATE helpers
There are a number of helper functions for marking a reiserfs inode
private that were leftover from reiserfs did its own thing wrt to
private inodes. S_PRIVATE has been in the kernel for some time, so this
patch removes the helpers and uses IS_PRIVATE instead.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Jeff Mahoney [Mon, 30 Mar 2009 18:02:31 +0000 (14:02 -0400)]
reiserfs: remove link detection code
Early in the reiserfs xattr development, there was a plan to use
hardlinks to save disk space for identical xattrs. That code never
materialized and isn't going to, so this patch removes the detection
code.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Jeff Mahoney [Mon, 30 Mar 2009 18:02:27 +0000 (14:02 -0400)]
reiserfs: introduce reiserfs_error()
Although reiserfs can currently handle severe errors such as journal failure,
it cannot handle less severe errors like metadata i/o failure. The following
patch adds a reiserfs_error() function akin to the one in ext3.
Subsequent patches will use this new error handler to handle errors more
gracefully in general.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Jeff Mahoney [Mon, 30 Mar 2009 18:02:26 +0000 (14:02 -0400)]
reiserfs: rearrange journal abort
This patch kills off reiserfs_journal_abort as it is never called, and
combines __reiserfs_journal_abort_{soft,hard} into one function called
reiserfs_abort_journal, which performs the same work. It is silent
as opposed to the old version, since the message was always issued
after a regular 'abort' message.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Jeff Mahoney [Mon, 30 Mar 2009 18:02:25 +0000 (14:02 -0400)]
reiserfs: rework reiserfs_panic
ReiserFS panics can be somewhat inconsistent.
In some cases:
* a unique identifier may be associated with it
* the function name may be included
* the device may be printed separately
This patch aims to make warnings more consistent. reiserfs_warning() prints
the device name, so printing it a second time is not required. The function
name for a warning is always helpful in debugging, so it is now automatically
inserted into the output. Hans has stated that every warning should have
a unique identifier. Some cases lack them, others really shouldn't have them.
reiserfs_warning() now expects an id associated with each message. In the
rare case where one isn't needed, "" will suffice.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Jeff Mahoney [Mon, 30 Mar 2009 18:02:24 +0000 (14:02 -0400)]
reiserfs: add locking around error buffer
The formatting of the error buffer is race prone. It uses static buffers
for both formatting and output. While overwriting the error buffer
can product garbled output, overwriting the format buffer with incompatible
% directives can cause crashes.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Jeff Mahoney [Mon, 30 Mar 2009 18:02:23 +0000 (14:02 -0400)]
reiserfs: eliminate reiserfs_warning from uniqueness functions
uniqueness2type and type2uniquness issue a warning when the value is
unknown. When called from reiserfs_warning, this causes a re-entrancy
problem and deadlocks on the error buffer lock.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Jeff Mahoney [Mon, 30 Mar 2009 18:02:21 +0000 (14:02 -0400)]
reiserfs: rework reiserfs_warning
ReiserFS warnings can be somewhat inconsistent.
In some cases:
* a unique identifier may be associated with it
* the function name may be included
* the device may be printed separately
This patch aims to make warnings more consistent. reiserfs_warning() prints
the device name, so printing it a second time is not required. The function
name for a warning is always helpful in debugging, so it is now automatically
inserted into the output. Hans has stated that every warning should have
a unique identifier. Some cases lack them, others really shouldn't have them.
reiserfs_warning() now expects an id associated with each message. In the
rare case where one isn't needed, "" will suffice.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Jeff Mahoney [Mon, 30 Mar 2009 18:02:20 +0000 (14:02 -0400)]
reiserfs: make some warnings informational
In several places, reiserfs_warning is used when there is no warning, just
a notice. This patch changes some of them to indicate that the message
is merely informational.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Jeff Mahoney [Mon, 30 Mar 2009 18:02:18 +0000 (14:02 -0400)]
reiserfs: use buffer_info for leaf_paste_entries
This patch makes leaf_paste_entries more consistent with respect to the
other leaf operations. Using buffer_info instead of buffer_head
directly allows us to get a superblock pointer for use in error
handling.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Jeff Mahoney [Mon, 30 Mar 2009 18:02:17 +0000 (14:02 -0400)]
reiserfs: audit transaction ids to always be unsigned ints
This patch fixes up the reiserfs code such that transaction ids are
always unsigned ints. In places they can currently be signed ints or
unsigned longs.
The former just causes an annoying clm-2200 warning and may join a
transaction when it should wait.
The latter is just for correctness since the disk format uses a 32-bit
transaction id. There aren't any runtime problems that result from it
not wrapping at the correct location since the value is truncated
correctly even on big endian systems. The 0 value might make it to
disk, but the mount-time checks will bump it to 10 itself.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Jeff Mahoney [Mon, 30 Mar 2009 18:02:16 +0000 (14:02 -0400)]
reiserfs: add support for mount count incrementing
The following patch adds the fields for tracking mount counts and last
fsck timestamps to the superblock. It also increments the mount count
on every read-write mount.
Reiserfsprogs 3.6.21 added support for these fields.
Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>