| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Subvol delete is a special kind of awful where we use the global reserve to
cover the ENOSPC requirements. The problem is once we're done removing
everything we do a btrfs_update_inode(), which by default will try to do the
delayed update stuff which will use it's own reserve. There will be no
space in this reserve and we'll return ENOSPC. So instead use
btrfs_update_inode_fallback() which will just fallback to updating the inode
item in the case of enospc. This is fine because the global reserve covers
the space requirements for this. With this patch I can now delete a subvol
on a problem image Dave Sterba sent me. Thanks,
Reported-by: David Sterba <[email protected]>
Signed-off-by: Josef Bacik <[email protected]>
Signed-off-by: Chris Mason <[email protected]>
|
|
When we created a new snapshot, the mtime and ctime of its parent directory
were not updated. Fix it.
Signed-off-by: Miao Xie <[email protected]>
Signed-off-by: Chris Mason <[email protected]>
|
|
With commit
commit d1270cd91f308c9d22b2804720c36ccd32dbc35e
Author: Arne Jansen <[email protected]>
Date: Tue Sep 13 15:16:43 2011 +0200
Btrfs: put back delayed refs that are too new
I added a window where the delayed_ref's head->ref_mod code can diverge
from the sum of the remaining refs, because we release the head->mutex
in the middle. This leads to btrfs_lookup_extent_info returning wrong
numbers. This patch fixes this by adjusting the head's ref_mod with each
delayed ref we run.
Signed-off-by: Arne Jansen <[email protected]>
Signed-off-by: Chris Mason <[email protected]>
|
|
When we split a leaf, we may end up inserting a new root on top of that
leaf. The reflog code was incorrectly assuming the old root was always
a node. This makes sure we skip over leaves.
Signed-off-by: Chris Mason <[email protected]>
|
|
Arne was complaining about the space cache having mismatching generation
numbers when debugging a deadlock. This is because we can run out of space
in our preallocated range for our space cache if you have a pretty
fragmented amount of space in your pinned space. So just increase the
amount of space we preallocate for space cache so we can be sure to have
enough space. This will only really affect data ranges since their the only
chunks that end up larger than 256MB. Thanks,
Signed-off-by: Josef Bacik <[email protected]>
Signed-off-by: Chris Mason <[email protected]>
|
|
We need a barrir before calling waitqueue_active otherwise we will miss
wakeups. So in places that do atomic_dec(); then atomic_read() use
atomic_dec_return() which imply a memory barrier (see memory-barriers.txt)
and then add an explicit memory barrier everywhere else that need them.
Thanks,
Signed-off-by: Josef Bacik <[email protected]>
|
|
Commit a168650c introduced a waiting mechanism to prevent busy waiting in
btrfs_run_delayed_refs. This can deadlock with btrfs_run_ordered_operations,
where a tree_mod_seq is held while waiting for the io to complete, while
the end_io calls btrfs_run_delayed_refs.
This whole mechanism is unnecessary. If not enough runnable refs are
available to satisfy count, just return as count is more like a guideline
than a strict requirement.
In case we have to run all refs, commit transaction makes sure that no
other threads are working in the transaction anymore, so we just assert
here that no refs are blocked.
Signed-off-by: Arne Jansen <[email protected]>
Signed-off-by: Chris Mason <[email protected]>
|
|
Fix a real bug caught by coccinelle.
fs/btrfs/delayed-inode.c:1013:1-11: second lock on line 1013
Signed-off-by: Fengguang Wu <[email protected]>
|
|
We've been allocating a big array for csums instead of storing them in the
io_tree like we do for buffered reads because previously we were locking the
entire range, so we didn't have an extent state for each sector of the
range. But now that we do the range locking as we map the buffers we can
limit the mapping lenght to sectorsize and use the private part of the
io_tree for our csums. This allows us to avoid an extra memory allocation
for direct reads which could incur latency. Thanks,
Signed-off-by: Josef Bacik <[email protected]>
|
|
When we close devices we add back empty devices for some reason that escapes
me. In the case of a missing dev we don't allocate an rcu_string for it's
name, so check to see if the device has a name and if it doesn't don't
bother strdup()'ing it. Thanks,
Signed-off-by: Josef Bacik <[email protected]>
|
|
If you do the following
mkfs.btrfs /dev/sdb /dev/sdc
rmmod btrfs
dd if=/dev/zero of=/dev/sdb bs=1M count=1
mount -o degraded /dev/sdc /mnt/btrfs-test
the box will panic trying to deref the name for the missing dev since it is
the lower numbered devid. So fix show_devname to not use missing devices.
Thanks,
Signed-off-by: Josef Bacik <[email protected]>
|
|
In iterate_inodes_from_logical() the error result from
extent_from_logical() is patched by mistake. Typically ENOENT is
patched to EINVAL because (-ENOENT & BTRFS_EXTENT_FLAG_TREE_BLOCK)
evaluates to true.
Signed-off-by: Stefan Behrens <[email protected]>
|
|
A deadlock in xfstests 113 was uncovered by commit
d187663ef24cd3d033f0cbf2867e70b36a3a90b8
This is because we would not return EIOCBQUEUED for short AIO reads, instead
we'd wait for the DIO to complete and then return the amount of data we
transferred, which would allow our stuff to unlock the remaning amount. But
with this change this no longer happens, so if we have a short AIO read (for
example if we try to read past EOF), we could leave the section from EOF to
the end of where we tried to read locked. Fixing this is tricky since there
is no clear way to know exactly how much data DIO truly submitted for IO, so
to make this less hard on ourselves and less combersome we need to lock the
extents as we try to map them, and then we unlock any areas we didn't
actually map. This makes us completely safe from deadlocks and reliance on
a particular behavior of the DIO code. This also lays the groundwork for
allowing us to use the normal csum storage method for reads which means we
can remove an allocation. Thanks,
Signed-off-by: Josef Bacik <[email protected]>
|
|
"trans->transid" is cpu endian but we want to store the data as little
endian. "item->ctime.nsec" is only 32 bits, not 64.
Signed-off-by: Dan Carpenter <[email protected]>
|
|
We should release this mutex before returning the error code.
Signed-off-by: Dan Carpenter <[email protected]>
|
|
add_qgroup_rb() never returns NULL, only error pointers.
Signed-off-by: Dan Carpenter <[email protected]>
|
|
These are returning zero when it should be returning a negative error
code.
Signed-off-by: Dan Carpenter <[email protected]>
|
|
This should obviously not be "if (&flag)" but "if (flag)".
Signed-off-by: Stefan Behrens <[email protected]>
|
|
Make use of the same atomic pool as DMA does, and skip a kernel page
mapping which can involve sleep'able operations at allocating a kernel
page table.
Signed-off-by: Hiroshi Doyu <[email protected]>
Signed-off-by: Marek Szyprowski <[email protected]>
|
|
Support atomic allocation in __iommu_get_pages().
Signed-off-by: Hiroshi Doyu <[email protected]>
[moved __atomic_get_pages() under #ifdef CONFIG_ARM_DMA_USE_IOMMU
to avoid unused fuction warning for no-IOMMU case]
Signed-off-by: Marek Szyprowski <[email protected]>
|
|
Check the given range("start", "size") is included in "atomic_pool" or not.
Signed-off-by: Hiroshi Doyu <[email protected]>
Signed-off-by: Marek Szyprowski <[email protected]>
|
|
struct page **pages is necessary to align with non atomic path in
__iommu_get_pages(). atomic_pool() has the intialized **pages instead
of just *page.
Signed-off-by: Hiroshi Doyu <[email protected]>
Signed-off-by: Marek Szyprowski <[email protected]>
|
|
The default 256 KiB coherent pool may be too small for some of the Kirkwood
devices, so increase it to make sure that devices will be able to allocate
their buffers with GFP_ATOMIC flag.
Suggested-by: Josh Coombs <[email protected]>
Signed-off-by: Marek Szyprowski <[email protected]>
Acked-by: Jason Cooper <[email protected]>
|
|
Print a loud warning when system runs out of memory from atomic DMA
coherent pool to let users notice the potential problem.
Reported-by: Aaro Koskinen <[email protected]>
Signed-off-by: Marek Szyprowski <[email protected]>
|
|
Some platforms might require to increase atomic coherent pool to make
sure that their device will be able to allocate all their buffers from
atomic context. This function can be also used to decrease atomic
coherent pool size if coherent allocations are not used for the given
sub-platform.
Suggested-by: Josh Coombs <[email protected]>
Signed-off-by: Marek Szyprowski <[email protected]>
|
|
Contiguous Memory Allocator requires only paging and MMU enabled not
particular CPU architectures, so there is no need for strict dependency
on CPU type. This enables to use CMA on some older ARM v5 systems which
also might need large contiguous blocks for the multimedia processing hw
modules.
Reported-by: Prabhakar Lad <[email protected]>
Signed-off-by: Marek Szyprowski <[email protected]>
Tested-by: Prabhakar Lad <[email protected]>
|
|
Contiguous Memory Allocator requires each of its regions to be aligned
in such a way that it is possible to change migration type for all
pageblocks holding it and then isolate page of largest possible order from
the buddy allocator (which is MAX_ORDER-1). This patch relaxes alignment
requirements by one order, because MAX_ORDER alignment is not really
needed.
Signed-off-by: Marek Szyprowski <[email protected]>
CC: Michal Nazarewicz <[email protected]>
Acked-by: Michal Nazarewicz <[email protected]>
|
|
These codecs seem reporting EPSS but require longer delay for the
proper D3 transition. For example, D3_STOP_CLOCK_OK bit won't be set
correctly even after D3.
In this patch, codec->epss flag is overridden for avoid the
misbehavior.
Signed-off-by: Takashi Iwai <[email protected]>
|
|
EPSS parameter should be static, so we can read it once and remember.
This also allows more easily to override the wrong EPSS capability
reported from a codec by changing the flag in the codec
initialization step.
Signed-off-by: Takashi Iwai <[email protected]>
|
|
Nobody does set_orig_insn(verify => false), and I think nobody will.
Remove this argument. IIUC set_orig_insn(verify => false) was needed
to single-step without xol area.
Signed-off-by: Oleg Nesterov <[email protected]>
Acked-by: Srikar Dronamraju <[email protected]>
|
|
Now that we have uprobe_dup_mmap() we can fold uprobe_reset_state()
into the new hook and remove it. mmput()->uprobe_clear_state() can't
be called before dup_mmap().
Signed-off-by: Oleg Nesterov <[email protected]>
Acked-by: Srikar Dronamraju <[email protected]>
|
|
Add the new MMF_HAS_UPROBES flag. It is set by install_breakpoint()
and it is copied by dup_mmap(), uprobe_pre_sstep_notifier() checks
it to avoid the slow path if the task was never probed. Perhaps it
makes sense to check it in valid_vma(is_register => false) as well.
This needs the new dup_mmap()->uprobe_dup_mmap() hook. We can't use
uprobe_reset_state() or put MMF_HAS_UPROBES into MMF_INIT_MASK, we
need oldmm->mmap_sem to avoid the race with uprobe_register() or
mmap() from another thread.
Currently we never clear this bit, it can be false-positive after
uprobe_unregister() or uprobe_munmap() or if dup_mmap() hits the
probed VM_DONTCOPY vma. But this is fine correctness-wise and has
no effect unless the task hits the non-uprobe breakpoint.
Signed-off-by: Oleg Nesterov <[email protected]>
Acked-by: Srikar Dronamraju <[email protected]>
|
|
-EEXIST from install_breakpoint() no longer makes sense, all
callers should simply treat it as "success". Change the code
to return zero and simplify register_for_each_vma().
Signed-off-by: Oleg Nesterov <[email protected]>
Acked-by: Srikar Dronamraju <[email protected]>
|
|
fatal_signal_pending()
Once install_breakpoint() fails uprobe_mmap() "ignores" all other
uprobes and returns the error.
It was never really needed to to stop after the first error, and
in fact it was always wrong at least in -ENOTSUPP case.
Change uprobe_mmap() to ignore the errors and always return 0.
This is not what we want in the long term, but until we teach
the callers to handle the failure it would be better to remove
the pointless complications. And this doesn't look too bad, the
only "reasonable" error is ENOMEM but in this case the caller
should be oom-killed in the likely case or the system has more
serious problems.
However it makes sense to stop if fatal_signal_pending() == T.
In particular this helps if the task was oom-killed.
Signed-off-by: Oleg Nesterov <[email protected]>
Acked-by: Srikar Dronamraju <[email protected]>
|
|
1. Kill dup_mmap()->uprobe_mmap(), it was only needed to calculate
new_mm->uprobes_state.count removed by the previous patch.
If the forking process has a pending uprobe (int3) in vma, it will
be copied by copy_page_range(), note that it checks vma->anon_vma
so "Don't copy ptes" is not possible after install_breakpoint()
which does anon_vma_prepare().
2. Remove is_swbp_at_addr() and "int count" in uprobe_mmap(). Again,
this was needed for uprobes_state.count.
As a side effect this fixes the bug pointed out by Srikar,
this code lacked the necessary put_uprobe().
3. uprobe_munmap() becomes a nop after the previous patch. Remove the
meaningless code but do not remove the helper, we will need it.
Signed-off-by: Oleg Nesterov <[email protected]>
Acked-by: Srikar Dronamraju <[email protected]>
|
|
uprobes_state->count is only needed to avoid the slow path in
uprobe_pre_sstep_notifier(). It is also checked in uprobe_munmap()
but ironically its only goal to decrement this counter. However,
it is very broken. Just some examples:
- uprobe_mmap() can race with uprobe_unregister() and wrongly
increment the counter if it hits the non-uprobe "int3". Note
that install_breakpoint() checks ->consumers first and returns
-EEXIST if it is NULL.
"atomic_sub() if error" in uprobe_mmap() looks obviously wrong
too.
- uprobe_munmap() can race with uprobe_register() and wrongly
decrement the counter by the same reason.
- Suppose an appication tries to increase the mmapped area via
sys_mremap(). vma_adjust() does uprobe_munmap(whole_vma) first,
this can nullify the counter temporarily and race with another
thread which can hit the bp, the application will be killed by
SIGTRAP.
- Suppose an application mmaps 2 consecutive areas in the same file
and one (or both) of these areas has uprobes. In the likely case
mmap_region()->vma_merge() suceeds. Like above, this leads to
uprobe_munmap/uprobe_mmap from vma_merge()->vma_adjust() but then
mmap_region() does another uprobe_mmap(resulting_vma) and doubles
the counter.
This patch only removes this counter and fixes the compile errors,
then we will try to cleanup the changed code and add something else
instead.
Signed-off-by: Oleg Nesterov <[email protected]>
Acked-by: Srikar Dronamraju <[email protected]>
|
|
by the time we get here (after we pass cleanup_ret) uprobe is always is
set. If it is NULL we leave very early in the code.
Signed-off-by: Sebastian Andrzej Siewior <[email protected]>
Signed-off-by: Oleg Nesterov <[email protected]>
|
|
Since read_opcode() reads from the referenced page and doesnt modify
the page contents nor the page attributes, there is no need to lock
the page.
Signed-off-by: Srikar Dronamraju <[email protected]>
Signed-off-by: Oleg Nesterov <[email protected]>
|
|
Pick up the latest fixes because upcoming uprobes changes will rely on it.
Signed-off-by: Ingo Molnar <[email protected]>
|
|
Fix this compile error:
arch/s390/kernel/machine_kexec.c: In function ‘setup_regs’:
arch/s390/kernel/machine_kexec.c:63:3: error: implicit declaration
of function ‘smp_store_status’ [-Werror=implicit-function-declaration]
Signed-off-by: Heiko Carstens <[email protected]>
Signed-off-by: Martin Schwidefsky <[email protected]>
|
|
For unimplemented ioctls the dasd driver should return -ENOTTY.
Reported-by: Wanlong Gao <[email protected]>
Acked-by: Stefan Weinhuber <[email protected]>
Signed-off-by: Heiko Carstens <[email protected]>
Signed-off-by: Martin Schwidefsky <[email protected]>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/horms/renesas into fixes
* 'armadillo800eva' of git://git.kernel.org/pub/scm/linux/kernel/git/horms/renesas:
ARM: mach-shmobile: armadillo800eva: Enable power button as wakeup source
ARM: mach-shmobile: armadillo800eva: Fix GPIO buttons descriptions
|
|
Fixes for AT91 related to:
- move to sparse IRQ: some drivers were forgotten
- a DTS typo
- the delay for removal of old at91_mci driver
* tag 'at91-fixes' of git://github.com/at91linux/linux-at91:
ARM: at91/feature-removal-schedule: delay at91_mci removal
ARM: at91/dts: remove partial parameter in at91sam9g25ek.dts
ARM: at91/clock: fix PLLA overclock warning
ARM: at91: fix rtc-at91sam9 irq issue due to sparse irq support
ARM: at91: fix system timer irq issue due to sparse irq support
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/rafael/renesas into fixes
* 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/renesas:
ARM: shmobile: sh73a0: fixup RELOC_BASE of intca_irq_pins_desc
|
|
KVM_GET_MSR was missing support for PV EOI,
which is needed for migration.
Signed-off-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Marcelo Tosatti <[email protected]>
|
|
This patch (as1603) adds a NOGET quirk for the Eaton Ellipse MAX UPS
device. (The USB IDs were already present in hid-ids.h, apparently
under a different name.)
Signed-off-by: Alan Stern <[email protected]>
Reported-by: Laurent Bigonville <[email protected]>
CC: <[email protected]>
Signed-off-by: Jiri Kosina <[email protected]>
|
|
Under certain workloads we see the following warnings:
WQ on CPU0, prefer CPU1
WQ on CPU0, prefer CPU2
WQ on CPU0, prefer CPU3
It warns the user that the wq to access a per-cpu buffers runs not on
the same cpu. This happens if the wq is rescheduled on a different cpu
than where the buffer is located. This was probably implemented to
detect performance issues. Not sure if there actually is one as the
buffers are copied to a single buffer anyway which should be the
actual bottleneck.
We wont change WQ implementation. Since a user can do nothing the
warning is pointless. Removing it.
Cc: Andi Kleen <[email protected]>
Signed-off-by: Robert Richter <[email protected]>
|
|
If oprofilefs_ulong_from_user() is called with count equals zero, *val
remains unchanged. Depending on the implementation it might be
uninitialized. Fixing users of oprofilefs_ulong_ from_user().
We missed these s390 changes with:
913050b oprofile: Fix uninitialized memory access when writing to writing to oprofilefs
Cc: [email protected] # 3.3+
Signed-off-by: Robert Richter <[email protected]>
|
|
Delay sd/mmc driver at91_mci.c removal because of tight schedule to
move platform data to new driver atmel-mci.
Signed-off-by: Ludovic Desroches <[email protected]>
Signed-off-by: Nicolas Ferre <[email protected]>
|
|
The following patch makes the microcode update code path
actually invoke the perf_check_microcode() function and
thus potentially renabling SNB PEBS.
By default, CONFIG_MICROCODE_OLD_INTERFACE is
forced to Y in arch/x86/Kconfig. There is no
way to disable this. That means that the code
path used in arch/x86/kernel/microcode_core.c
did not include the call to perf_check_microcode().
Thus, even though the microcode was updated to a
version that fixes the SNB PEBS problem, perf_event
would still return EOPNOTSUPP when enabling precise
sampling.
This patch simply adds a call to perf_check_microcode()
in the call path used when OLD_INTERFACE=y.
Signed-off-by: Stephane Eranian <[email protected]>
Acked-by: Borislav Petkov <[email protected]>
Cc: [email protected]
Cc: [email protected]
Link: http://lkml.kernel.org/r/20120824133434.GA8014@quad
Signed-off-by: Ingo Molnar <[email protected]>
|