lockdep found a AB BC CA lock inversion in retry-based AIO:
1) The task struct's alloc_lock (A) is acquired in process context with
interrupts enabled. An interrupt might arrive and call wake_up() which
grabs the wait queue's q->lock (B).
2) When performing retry-based AIO the AIO core registers
aio_wake_function() as the wake funtion for iocb->ki_wait. It is called
with the wait queue's q->lock (B) held and then tries to add the iocb to
the run list after acquiring the ctx_lock (C).
3) aio_kick_handler() holds the ctx_lock (C) while acquiring the
alloc_lock (A) via lock_task() and unuse_mm(). Lockdep emits a warning
saying that we're trying to connect the irq-safe q->lock to the
irq-unsafe alloc_lock via ctx_lock.
This fixes the inversion by calling unuse_mm() in the AIO kick handing path
after we've released the ctx_lock. As Ben LaHaise pointed out __put_ioctx
could set ctx->mm to NULL, so we must only access ctx->mm while we have the
lock.
Signed-off-by: Zach Brown <zach.brown@oracle.com>
Signed-off-by: Suparna Bhattacharya <suparna@in.ibm.com>
Acked-by: Benjamin LaHaise <bcrl@kvack.org>
Cc: "Chen, Kenneth W" <kenneth.w.chen@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
activate_mm() is not the right thing to be using in use_mm(). It should be
switch_mm().
On normal x86, they're synonymous, but for the Xen patches I'm adding a
hook which assumes that activate_mm is only used the first time a new mm
is used after creation (I have another hook for dealing with dup_mm). I
think this use of activate_mm() is the only place where it could be used
a second time on an mm.
>From a quick look at the other architectures I think this is OK (most
simply implement one in terms of the other), but some are doing some
subtly different stuff between the two.
Acked-by: David Miller <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Remove the ki_retried member from struct kiocb. I think the idea was
bounced around a while back, but Arnaldo pointed out another reason that we
should dig it up when he pointed out that the last cacheline of struct
kiocb only contains 4 bytes. By removing the debugging member, we save
more than the 8 byte on 64 bit machines.
Signed-off-by: Benjamin LaHaise <bcrl@kvack.org>
Acked-by: Ken Chen <kenneth.w.chen@intel.com>
Acked-by: Zach Brown <zach.brown@oracle.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
io_submit_one assigns ki_left = ki_nbytes = iocb->aio_nbytes, then calls
down to aio_setup_iocb, then to aio_setup_single_vector. In there,
ki_nbytes is reassigned to the same value it got two call stack above it.
There is no need to do so.
Signed-off-by: Ken Chen <kenneth.w.chen@intel.com>
Acked-by: Zach Brown <zach.brown@oracle.com>
Cc: Benjamin LaHaise <bcrl@kvack.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Replace all uses of kmem_cache_t with struct kmem_cache.
The patch was generated using the following script:
#!/bin/sh
#
# Replace one string by another in all the kernel sources.
#
set -e
for file in `find * -name "*.c" -o -name "*.h"|xargs grep -l $1`; do
quilt add $file
sed -e "1,\$s/$1/$2/g" $file >/tmp/$$
mv /tmp/$$ $file
quilt refresh
done
The script was run like this
sh replace kmem_cache_t "struct kmem_cache"
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Conflicts:
drivers/infiniband/core/iwcm.c
drivers/net/chelsio/cxgb2.c
drivers/net/wireless/bcm43xx/bcm43xx_main.c
drivers/net/wireless/prism54/islpci_eth.c
drivers/usb/core/hub.h
drivers/usb/input/hid-core.c
net/core/netpoll.c
Fix up merge failures with Linus's head and fix new compilation failures.
Signed-Off-By: David Howells <dhowells@redhat.com>
This patch converts a if () BUG(); construct to BUG_ON();
which occupies less space, uses unlikely and is safer when
BUG() is disabled.
Signed-off-by: Eric Sesterhenn <snakebyte@gmx.de>
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Pass the work_struct pointer to the work function rather than context data.
The work function can use container_of() to work out the data.
For the cases where the container of the work_struct may go away the moment the
pending bit is cleared, it is made possible to defer the release of the
structure by deferring the clearing of the pending bit.
To make this work, an extra flag is introduced into the management side of the
work_struct. This governs auto-release of the structure upon execution.
Ordinarily, the work queue executor would release the work_struct for further
scheduling or deallocation by clearing the pending bit prior to jumping to the
work function. This means that, unless the driver makes some guarantee itself
that the work_struct won't go away, the work function may not access anything
else in the work_struct or its container lest they be deallocated.. This is a
problem if the auxiliary data is taken away (as done by the last patch).
However, if the pending bit is *not* cleared before jumping to the work
function, then the work function *may* access the work_struct and its container
with no problems. But then the work function must itself release the
work_struct by calling work_release().
In most cases, automatic release is fine, so this is the default. Special
initiators exist for the non-auto-release case (ending in _NAR).
Signed-Off-By: David Howells <dhowells@redhat.com>
Separate delayable work items from non-delayable work items be splitting them
into a separate structure (delayed_work), which incorporates a work_struct and
the timer_list removed from work_struct.
The work_struct struct is huge, and this limits it's usefulness. On a 64-bit
architecture it's nearly 100 bytes in size. This reduces that by half for the
non-delayable type of event.
Signed-Off-By: David Howells <dhowells@redhat.com>
aio: use size_t length modifier in pr_debug format arguments
Signed-off-by: Zach Brown <zach.brown@oracle.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This work is initially done by Zach Brown to add support for vectored aio.
These are the core changes for AIO to support
IOCB_CMD_PREADV/IOCB_CMD_PWRITEV.
[akpm@osdl.org: huge build fix]
Signed-off-by: Zach Brown <zach.brown@oracle.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Badari Pulavarty <pbadari@us.ibm.com>
Acked-by: Benjamin LaHaise <bcrl@kvack.org>
Acked-by: James Morris <jmorris@namei.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch vectorizes aio_read() and aio_write() methods to prepare for
collapsing all aio & vectored operations into one interface - which is
aio_read()/aio_write().
Signed-off-by: Badari Pulavarty <pbadari@us.ibm.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Michael Holzheu <HOLZHEU@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
list_splice_init(list, head) does unneeded job if it is known that
list_empty(head) == 1. We can use list_replace_init() instead.
Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Use atomic_inc_not_zero for rcu files instead of special case rcuref.
Signed-off-by: Nick Piggin <npiggin@suse.de>
Cc: "Paul E. McKenney" <paulmck@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Sync iocbs have a life cycle that don't need a kioctx. Their retrying, if
any, is done in the context of their owner who has allocated them on the
stack.
The sole user of a sync iocb's ctx reference was aio_complete() checking for
an elevated iocb ref count that could never happen. No path which grabs an
iocb ref has access to sync iocbs.
If we were to implement sync iocb cancelation it would be done by the owner of
the iocb using its on-stack reference.
Removing this chunk from aio_complete allows us to remove the entire kioctx
instance from mm_struct, reducing its size by a third. On a i386 testing box
the slab size went from 768 to 504 bytes and from 5 to 8 per page.
Signed-off-by: Zach Brown <zach.brown@oracle.com>
Acked-by: Benjamin LaHaise <bcrl@kvack.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
AIO was adding a new context's max requests to the global total before
testing if that resulting total was over the global limit. This let
innocent tasks get their new limit tested along with a racing guilty task
that was crossing the limit. This serializes the _nr accounting with a
spinlock It also switches to using unsigned long for the global totals.
Individual contexts are still limited to an unsigned int's worth of
requests by the syscall interface.
The problem and fix were verified with a simple program that spun creating
and destroying a context while holding on to another long lived context.
Before the patch a task creating a tiny context could get a spurious EAGAIN
if it raced with a task creating a very large context that overran the
limit.
Signed-off-by: Zach Brown <zach.brown@oracle.com>
Cc: Benjamin LaHaise <bcrl@kvack.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Another case of missing call to security_file_permission: aio functions
(namely, io_submit) does not check credentials with security modules.
Below is the simple patch to the problem. It seems that it is enough to
check for rights at the request submission time.
Signed-off-by: Kostik Belousov <kostikbel@gmail.com>
Signed-off-by: Chris Wright <chrisw@osdl.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
lock_kiocb() was introduced to serialize retrying and cancellation. In the
process of doing so it tried to sleep waiting for KIF_LOCKED while holding
the ctx_lock spinlock. Recent fixes have ensured that multiple concurrent
retries won't be attempted for a given iocb. Cancel has other problems and
has no significant in-tree users that have been complaining about it. So
for the immediate future we'll revert sleeping with the lock held and will
address proper cancellation and retry serialization in the future.
Signed-off-by: Zach Brown <zach.brown@oracle.com>
Acked-by: Benjamin LaHaise <bcrl@kvack.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Recently aio_p{read,write} changed to perform retries internally rather
than returning -EIOCBRETRY. This inadvertantly resulted in always calling
aio_{read,write} with ki_left at 0 which would in turn immediately return
0. Harmless, but we can avoid this call by checking in the caller.
Signed-off-by: Zach Brown <zach.brown@oracle.com>
Signed-off-by: Benjamin LaHaise <bcrl@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Only one of the run or kick path is supposed to put an iocb on the run
list. If both of them do it than one of them can end up referencing a
freed iocb. The kick path could delete the task_list item from the wait
queue before getting the ctx_lock and putting the iocb on the run list.
The run path was testing the task_list item outside the lock so that it
could catch ki_retry methods that return -EIOCBRETRY *without* putting the
iocb on a wait queue and promising to call kick_iocb. This unlocked check
could then race with the kick path to cause both to try and put the iocb on
the run list.
The patch stops the run path from testing task_list by requring that any
ki_retry that returns -EIOCBRETRY *must* guarantee that kick_iocb() will be
called in the future. aio_p{read,write}, the only in-tree -EIOCBRETRY
users, are updated.
Signed-off-by: Zach Brown <zach.brown@oracle.com>
Signed-off-by: Benjamin LaHaise <bcrl@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Only one of the run or kick path is supposed to put an iocb on the run
list. If both of them do it than one of them can end up referencing a
freed iocb. The kick patch could set the Kicked bit before acquiring the
ctx_lock and putting the iocb on the run list. The run path, while holding
the ctx_lock, could see this partial kick and mistake it for a kick that
was deferred while it was doing work with the run_list NULLed out. It
would then race with the kick thread to add the iocb to the run list.
This patch moves the kick setting under the ctx_lock so that only one of
the kick or run path queues the iocb on the run list, as intended.
Signed-off-by: Zach Brown <zach.brown@oracle.com>
Signed-off-by: Benjamin LaHaise <bcrl@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Add smp_mb__after_clear_bit() to unlock_kiocb()
AIO's use of wait_on_bit_lock()/wake_up_bit() forgot to add a barrier
between clearing its lock bit and calling wake_up_bit() so wake_up_bit()'s
unlocked waitqueue_active() can race. This puts AIO's use in line with the
others and the comment above wake_up_bit().
Signed-off-by: Zach Brown <zach.brown@oracle.com>
Acked-by: Benjamin LaHaise <bcrl@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Patch to eliminate struct files_struct.file_lock spinlock on the reader side
and use rcu refcounting rcuref_xxx api for the f_count refcounter. The
updates to the fdtable are done by allocating a new fdtable structure and
setting files->fdt to point to the new structure. The fdtable structure is
protected by RCU thereby allowing lock-free lookup. For fd arrays/sets that
are vmalloced, we use keventd to free them since RCU callbacks can't sleep. A
global list of fdtable to be freed is not scalable, so we use a per-cpu list.
If keventd is already handling the current cpu's work, we use a timer to defer
queueing of that work.
Since the last publication, this patch has been re-written to avoid using
explicit memory barriers and use rcu_assign_pointer(), rcu_dereference()
premitives instead. This required that the fd information is kept in a
separate structure (fdtable) and updated atomically.
Signed-off-by: Dipankar Sarma <dipankar@in.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Implement a per-kiocb lock to serialise retry operations and cancel. This
is done using wait_on_bit_lock() on the KIF_LOCKED bit of kiocb->ki_flags.
Also, make the cancellation path lock the kiocb and subsequently release
all references to it if the cancel was successful. This version includes a
fix for the deadlock with __aio_run_iocbs.
Signed-off-by: Benjamin LaHaise <bcrl@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Note that other than few exceptions, most of the current filesystem and/or
drivers do not have aio cancel specifically defined (kiob->ki_cancel field
is mostly NULL). However, sys_io_cancel system call universally sets
return code to -EAGAIN. This gives applications a wrong impression that
this call is implemented but just never works. We have customer inquires
about this issue.
Changed by Benjamin LaHaise to EINVAL instead of ENOSYS
Signed-off-by: S. Wendy Cheng <wcheng@redhat.com>
Acked-by: Benjamin LaHaise <bcrl@kvack.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Normally, activate_mm() is called from exec(), and thus it used to be a
no-op because we use a completely new "MM context" on the host (for
instance, a new process), and so we didn't need to flush any "TLB entries"
(which for us are the set of memory mappings for the host process from the
virtual "RAM" file).
Kernel threads, instead, are usually handled in a different way. So, when
for AIO we call use_mm(), things used to break and so Benjamin implemented
activate_mm(). However, that is only needed for AIO, and could slow down
exec() inside UML, so be smart: detect being called for AIO (via
PF_BORROWED_MM) and do the full flush only in that situation.
Comment also the caller so that people won't go breaking UML without
noticing. I also rely on the caller's locks for testing current->flags.
Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
CC: Benjamin LaHaise <bcrl@kvack.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
In the case of buffered AIO, in the aio retry path (aio_run_iocb), when the
retry method returns EIOCBRETRY the kicked iocb is added to the context run
list but is never queued onto the work queue. The request therefore is
never completed.
This patch fixes that by adding the appropriate call to aio_queue_work in
aio_run_aiocb so that subsequent retries will be handled by the aio worker
thread.
Signed-off-by: Sbastien Dugu <sebastien.dugue@bull.net>
Acked-by: Benjamin LaHaise <benjamin.c.lahaise@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch optimizes io_submit_one to call aio_run_iocb() directly if
ctx->run_list is empty. When the list is empty, the operation of adding to
the list, then call to __aio_run_iocbs() is unnecessary because these
operations are done in one atomic step. ctx->run_list always has only one
element in this case. This optimization speeds up industry standard db
transaction processing benchmark by 0.2%.
Signed-off-by: Ken Chen <kenneth.w.chen@intel.com>
Cc: Benjamin LaHaise <bcrl@kvack.org>
Cc: Suparna Bhattacharya <suparna@in.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Clean up code that was previously used for debug purpose. Remove aio_run,
aio_wakeups, iocb->ki_queued and iocb->ki_kicked. Also clean up unused
variable count in __aio_run_iocbs() and debug code in read_events().
Signed-off-by: Ken Chen <kenneth.w.chen@intel.com>
Cc: Benjamin LaHaise <bcrl@kvack.org>
Cc: Suparna Bhattacharya <suparna@in.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Since the tail pointer in aio_ring structure never wrap ring size more than
once, so a simple compare is sufficient to wrap the index around. This avoid
a more expensive mod operation.
Signed-off-by: Ken Chen <kenneth.w.chen@intel.com>
Cc: Benjamin LaHaise <bcrl@kvack.org>
Cc: Suparna Bhattacharya <suparna@in.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch removes superfluous kiocb member initialization in the AIO
allocation and deallocation path. For example, in really_put_req(),
right before kiocb is returned to slab, 5 variables are reset to NULL.
The same variables will be initialized at the kiocb allocation time,
so why bother reset them knowing that they will be set to valid data
at alloc time? Another example: ki_retry is initialized in __aio_get_req,
but is initialized again in io_submit_one.
Signed-off-by: Ken Chen <kenneth.w.chen@intel.com>
Cc: Benjamin LaHaise <bcrl@kvack.org>
Cc: Suparna Bhattacharya <suparna@in.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch makes some needlessly global code static.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Acked-by: Benjamin LaHaise <bcrl@kvack.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Initial git repository build. I'm not bothering with the full history,
even though we have it. We can create a separate "historical" git
archive of that later if we want to, and in the meantime it's about
3.2GB when imported into git - space that would just make the early
git days unnecessarily complicated, when we don't have a lot of good
infrastructure for it.
Let it rip!
Zach Brown
Jeremy Fitzhardinge
Benjamin LaHaise
Chen, Kenneth W
Christoph Lameter
David Howells
Eric Sesterhenn
Badari Pulavarty
Andreas Mohr
Oleg Nesterov
Oliver Neukum
Nick Piggin
Kostik Belousov
Dipankar Sarma
Benjamin LaHaise
Wendy Cheng
Paolo 'Blaisorblade' Giarrusso
Sbastien Dugu
Adrian Bunk
Linus Torvalds