From: Zach Brown
use_mm() is used in more places than just aio. There's no need to
mention callers when describing the function.
Signed-off-by: Zach Brown
Signed-off-by: Kent Overstreet
---
mm/mmu_context.c | 3 ---
1 file changed, 3 deletions(-)
diff --git a/mm/mmu_context.c
Signed-off-by: Kent Overstreet
---
drivers/usb/gadget/inode.c | 1 +
fs/aio.c | 61 ++
include/linux/aio.h| 61 --
3 files changed, 62 insertions(+), 61 deletions(-)
diff --git a
just allocate and free slots, and we know how to write per cpu
allocators.
So as prep work for that, we convert reqs_active to reqs_available.
Signed-off-by: Kent Overstreet
---
fs/aio.c | 29 ++---
1 file changed, 14 insertions(+), 15 deletions(-)
diff --git a/fs/aio.c
This just converts the ioctx refcount to the new generic dynamic percpu
refcount code.
Signed-off-by: Kent Overstreet
---
fs/aio.c | 30 +-
1 file changed, 13 insertions(+), 17 deletions(-)
diff --git a/fs/aio.c b/fs/aio.c
index 94218b7..0975675 100644
--- a/fs
to be pulling it out from under us as
we're using it.
So, we use an old trick to get rid of this second forced cache miss -
make the data the lock protects be the lock itself, so we grab them both
at once.
Signed-off-by: Kent Overstreet
---
fs/aio.c | 24 +---
1 file chan
.
This also fixes a bug where if copying the event to userspace failed,
we'd lose the event.
Signed-off-by: Kent Overstreet
---
fs/aio.c | 173 +--
1 file changed, 80 insertions(+), 93 deletions(-)
diff --git a/fs/aio.c b/fs/aio
Yay faster kernel compiles
Signed-off-by: Kent Overstreet
---
arch/s390/hypfs/inode.c | 1 +
block/scsi_ioctl.c | 1 +
drivers/char/mem.c | 1 +
drivers/infiniband/hw/ipath/ipath_file_ops.c | 1 +
drivers/infiniband/hw/qib
I expect this'll be good enough in
practice.
Signed-off-by: Kent Overstreet
---
include/linux/percpu-refcount.h | 29 +++
lib/Makefile| 2 +-
lib/percpu-refcount.c | 164
3 files changed, 194 insertions(+), 1 d
e, it was being rounded up to the next
page before so this doesn't change the actual semantics.
Signed-off-by: Kent Overstreet
---
fs/aio.c | 97
1 file changed, 85 insertions(+), 12 deletions(-)
diff --git a/fs/aio.c b/fs
eordering a bit clearer.
Signed-off-by: Kent Overstreet
---
fs/aio.c | 142 ++-
1 file changed, 68 insertions(+), 74 deletions(-)
diff --git a/fs/aio.c b/fs/aio.c
index f2d616e..1ff4d3b 100644
--- a/fs/aio.c
+++ b/fs/aio.c
@@ -58,18 +58,6 @@
removal of the locking
around the unused run list in the submission path.
This has only been compiled.
Signed-off-by: Kent Overstreet
---
fs/aio.c | 348 --
fs/ocfs2/dlmglue.c| 2 +-
fs/read_write.c | 34 +
include/linux
Converting read_events() to prepare_to_wait_exclusive() and
hrtimers simplifies it quite a bit.
Signed-off-by: Kent Overstreet
---
fs/aio.c | 78 +---
1 file changed, 21 insertions(+), 57 deletions(-)
diff --git a/fs/aio.c b/fs/aio.c
other syscalls may still be
submitting kiocbs!
Now, we cancel and wait for outstanding kiocbs in free_ioctx(), after
kioctx->users has dropped to 0 and we know no more iocbs could be
submitted.
Signed-off-by: Kent Overstreet
---
fs/aio.c | 236 ++--
Signed-off-by: Kent Overstreet
---
fs/aio.c | 20 +---
1 file changed, 13 insertions(+), 7 deletions(-)
diff --git a/fs/aio.c b/fs/aio.c
index 1ff4d3b..7dee3aa 100644
--- a/fs/aio.c
+++ b/fs/aio.c
@@ -67,13 +67,6 @@ struct kioctx {
unsigned long user_id
cachelines, so that means we can just rip this code out.
Signed-off-by: Kent Overstreet
---
fs/aio.c| 117 +++-
include/linux/aio.h | 1 -
2 files changed, 15 insertions(+), 103 deletions(-)
diff --git a/fs/aio.c b/fs/aio.c
index 0d5062
But - we can improve this quite a bit if we just change reqs_active to
mean "number of outstanding requests and unreaped completions" - that
means kiocb allocation doesn't have to look at the ringbuffer, which is
a fairly significant win.
Signed-off-by: Kent Overstreet
Cancelling kiocbs requires adding them to a per kioctx linked list,
which is one of the few things we need to take the kioctx lock for in
the fast path. But most kiocbs can't be cancelled - so if we just do
this lazily, we can avoid quite a bit of locking overhead.
Signed-off-by: Kent Overs
tive list. This occurs in two places: kiocb_batch_free(), which
is going away in a later patch, and the error path in io_submit_one.
Signed-off-by: Kent Overstreet
---
fs/aio.c| 85 +
include/linux/aio.h | 4 +--
2 files changed, 35 in
aio_get_req() will fail if we have the maximum number of requests
outstanding, which depending on the application may not be uncommon. So
avoid doing an unnecessary fget().
Signed-off-by: Kent Overstreet
---
fs/aio.c | 22 +-
1 file changed, 9 insertions(+), 13 deletions
Signed-off-by: Kent Overstreet
---
fs/aio.c | 57 -
1 file changed, 24 insertions(+), 33 deletions(-)
diff --git a/fs/aio.c b/fs/aio.c
index 4dcc02f..0ed14e8 100644
--- a/fs/aio.c
+++ b/fs/aio.c
@@ -8,6 +8,8 @@
*
* See ../COPYING
rformance
issues to speak of in the aio code.
Real world benchmarks are still lacking, I've just been focused on
profiles. I'll try and post some actual benchmarks/profiles later.
The patch series is on top of v3.7-rc7, git repo is at
http://evilpiepirate.org/git/linux-bcache.git aio-up
From: Zach Brown
These are handy for measuring the cost of the aio infrastructure with
operations that do very little and complete immediately.
Signed-off-by: Zach Brown
Signed-off-by: Kent Overstreet
---
drivers/char/mem.c | 35 +++
1 file changed, 35
Minor refactoring, to get rid of some duplicated code
Signed-off-by: Kent Overstreet
---
fs/aio.c | 72 +++-
1 file changed, 39 insertions(+), 33 deletions(-)
diff --git a/fs/aio.c b/fs/aio.c
index 91879d4..a993234 100644
--- a/fs
Nothing used the return value, and it probably wasn't possible to use it
safely for the locked versions (aio_complete(), aio_put_req()). Just
kill it.
Acked-by: Zach Brown
Signed-off-by: Kent Overstreet
---
fs/aio.c| 19 +++
include/linux/aio.h | 8 --
call. It only used the AIO retry facility to
get the submitter's mm context for copying the result of a read back to
user space. This is easy to implement with use_mm() and a work struct,
much like kvm does with async_pf_execute() for get_user_pages().
Signed-off-by: Zach Brown
Signed-of
From: Zach Brown
Signed-off-by: Zach Brown
Signed-off-by: Kent Overstreet
---
include/linux/aio.h | 24
1 file changed, 24 deletions(-)
diff --git a/include/linux/aio.h b/include/linux/aio.h
index 31ff6db..b46a09f 100644
--- a/include/linux/aio.h
+++ b/include/linux
On Wed, Nov 28, 2012 at 04:07:53PM -0800, Zach Brown wrote:
> On Wed, Nov 28, 2012 at 08:43:31AM -0800, Kent Overstreet wrote:
> > Minor refactoring, to get rid of some duplicated code
>
> A minor nit:
>
> > spin_lock_irq(&ctx->ctx_lock);
> > -
On Wed, Nov 28, 2012 at 04:24:36PM -0800, Zach Brown wrote:
> > - int i = 0;
> > + DEFINE_WAIT(wait);
> > + struct hrtimer_sleeper t;
> > + size_t i = 0;
>
> Changing i to size_t is kind of surprising. Is that on purpose?
I doubt it matters due to limits on ringbuffer
On Wed, Nov 28, 2012 at 04:17:59PM -0800, Zach Brown wrote:
>
> > struct kioctx {
> > atomic_tusers;
> > - int dead;
> > + atomic_tdead;
>
> Do we want to be paranoid and atomic_set() that to 0 when the ioctx is
> allocated?
I suppose
On Wed, Nov 28, 2012 at 07:46:31PM -0500, Benjamin LaHaise wrote:
> Hi Kent,
>
> On Wed, Nov 28, 2012 at 08:43:36AM -0800, Kent Overstreet wrote:
> > + * now it's safe to cancel any that need to be.
> > + */
> > +static void free_ioctx(struct kioctx *ctx)
>
On Thu, Nov 29, 2012 at 10:45:04AM -0800, Andi Kleen wrote:
> Kent Overstreet writes:
>
> > This implements a refcount with similar semantics to
> > atomic_get()/atomic_dec_and_test(), that starts out as just an atomic_t
> > but dynamically switches to per cpu refc
On Wed, Nov 28, 2012 at 04:03:03PM -0800, Zach Brown wrote:
> On Wed, Nov 28, 2012 at 08:43:24AM -0800, Kent Overstreet wrote:
> > Bunch of performance improvements and cleanups Zach Brown and I have
> > been working on. The code should be pretty solid at this point, though
> &g
On Thu, Nov 29, 2012 at 07:59:53PM +0100, Andi Kleen wrote:
> On Thu, Nov 29, 2012 at 10:57:20AM -0800, Kent Overstreet wrote:
> > On Thu, Nov 29, 2012 at 10:45:04AM -0800, Andi Kleen wrote:
> > > Kent Overstreet writes:
> > >
> > > > This implemen
On Thu, Nov 29, 2012 at 08:20:03PM +0100, Andi Kleen wrote:
> > The trick is that we don't watch for the refcount hitting 0 until we're
> > shutting down - so this only works if you keep track of your initial
> > refcount. As long as we're not shutting down, we know the refcount can't
> > hit 0 bec
On Wed, Nov 28, 2012 at 04:38:16PM -0800, Zach Brown wrote:
> > We can't use cmpxchg() on the ring buffer's head pointer directly, since
> > it's modded to nr_events and would be susceptible to ABA. So instead we
> > maintain a shadow head that uses the full 32 bits, and cmpxchg() that
> > and then
On Thu, Nov 29, 2012 at 02:34:52PM -0500, Benjamin LaHaise wrote:
> On Thu, Nov 29, 2012 at 11:29:25AM -0800, Kent Overstreet wrote:
> > There's some kind of symmetry going on here, and if I'd been awake more
> > in college I could probably say exactly why it works, but
On Thu, Nov 29, 2012 at 12:42:17PM -0800, Andi Kleen wrote:
> Kent Overstreet writes:
>
> > On Thu, Nov 29, 2012 at 02:34:52PM -0500, Benjamin LaHaise wrote:
> >> On Thu, Nov 29, 2012 at 11:29:25AM -0800, Kent Overstreet wrote:
> >> > There's some kind of s
On Thu, Nov 29, 2012 at 09:54:47PM +0100, Andi Kleen wrote:
> > > The regular atomic_t is limited in ways that you are not.
> > > See my original mail.
> >
> > I don't follow, can you explain?
>
> For most cases the reference count is tied to some object, which are
> naturally limited by memory s
sted on irc earlier:
commit 913ff32bbd4de15a87b07a87ac196e978bc29e17
Author: Kent Overstreet
Date: Thu Nov 29 14:12:40 2012 -0800
aio: Make aio_read_evt() more efficient
Previously, aio_read_event() pulled a single completion off the
ringbuffer at a time, locking and
to be pulling it out from under us as we're using
it.
So, we use an old trick to get rid of this second forced cache miss - make
the data the lock protects be the lock itself, so we grab them both at
once.
Signed-off-by: Kent Overstreet
Cc: Zach Brown
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
where rational]
[a...@linux-foundation.org: fix warnings]
[minc...@kernel.org: fix build error due to bio_endio_batch]
[a...@linux-foundation.org: fix tracepoint in batch_complete()]
Signed-off-by: Kent Overstreet
Cc: Zach Brown
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
Cc: Mark Fasheh
Cc: Joel Be
Signed-off-by: Kent Overstreet
Cc: Zach Brown
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
Cc: Mark Fasheh
Cc: Joel Becker
Cc: Rusty Russell
Cc: Jens Axboe
Cc: Asai Thambi S P
Cc: Selvan Mani
Cc: Sam Bradshaw
Cc: Jeff Moyer
Cc: Al Viro
Cc: Benjamin LaHaise
Cc: Theodore Ts'o
Signed-o
cachelines, so that means we can just rip this code out.
Signed-off-by: Kent Overstreet
Cc: Zach Brown
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
Cc: Mark Fasheh
Cc: Joel Becker
Cc: Rusty Russell
Cc: Jens Axboe
Cc: Asai Thambi S P
Cc: Selvan Mani
Cc: Sam Bradshaw
Cc: Jeff Moyer
Cc: A
()
* mtip_handle_tfe()]
Signed-off-by: Kent Overstreet
Signed-off-by: Asai Thambi S P
Cc: Zach Brown
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
Cc: Mark Fasheh
Cc: Joel Becker
Cc: Rusty Russell
Cc: Jens Axboe
Cc: Asai Thambi S P
Cc: Selvan Mani
Cc: Sam Bradshaw
Cc: Jeff Moyer
Cc: Al
ki_key wasn't actually used for anything previously - it was always 0.
Drop it to trim struct kiocb a bit.
Signed-off-by: Kent Overstreet
Cc: Zach Brown
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
Cc: Mark Fasheh
Cc: Joel Becker
Cc: Rusty Russell
Cc: Jens Axboe
Cc: Asai Thambi S P
Cc: S
Faster kernel compiles by way of fewer unnecessary includes.
[a...@linux-foundation.org: fix fallout]
Signed-off-by: Kent Overstreet
Cc: Zach Brown
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
Cc: Mark Fasheh
Cc: Joel Becker
Cc: Rusty Russell
Cc: Jens Axboe
Cc: Asai Thambi S P
Cc: Selvan Mani
But I expect this'll be good enough in practice.
[a...@linux-foundation.org: fix build]
[a...@linux-foundation.org: coding-style tweak]
Signed-off-by: Kent Overstreet
Cc: Zach Brown
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
Cc: Mark Fasheh
Cc: Joel Becker
Cc: Rusty Russell
Cc: Jens Axboe
C
vents wasn't
being used as a hard limit before, it was being rounded up to the next
page before so this doesn't change the actual semantics.
Signed-off-by: Kent Overstreet
Cc: Zach Brown
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
Cc: Mark Fasheh
Cc: Joel Becker
Cc: Rusty Russell
Cc: Je
just allocate and free slots, and we know how to write per cpu allocators.
So as prep work for that, we convert reqs_active to reqs_available.
Signed-off-by: Kent Overstreet
Cc: Zach Brown
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
Cc: Mark Fasheh
Cc: Joel Becker
Cc: Rusty Russell
Cc: Jens
From: Zach Brown
These are handy for measuring the cost of the aio infrastructure with
operations that do very little and complete immediately.
Signed-off-by: Zach Brown
Signed-off-by: Kent Overstreet
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
Cc: Mark Fasheh
Cc: Joel Becker
Cc: Rusty
aio_get_req() will fail if we have the maximum number of requests
outstanding, which depending on the application may not be uncommon. So
avoid doing an unnecessary fget().
Signed-off-by: Kent Overstreet
Cc: Zach Brown
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
Cc: Mark Fasheh
Cc: Joel Becker
Signed-off-by: Kent Overstreet
Cc: Zach Brown
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
Cc: Mark Fasheh
Cc: Joel Becker
Cc: Rusty Russell
Cc: Jens Axboe
Cc: Asai Thambi S P
Cc: Selvan Mani
Cc: Sam Bradshaw
Cc: Jeff Moyer
Cc: Al Viro
Cc: Benjamin LaHaise
Signed-off-by: Andrew Morton
eordering a bit clearer.
Signed-off-by: Kent Overstreet
Cc: Zach Brown
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
Cc: Mark Fasheh
Cc: Joel Becker
Cc: Rusty Russell
Cc: Jens Axboe
Cc: Asai Thambi S P
Cc: Selvan Mani
Cc: Sam Bradshaw
Cc: Jeff Moyer
Cc: Al Viro
Cc: Benjamin LaHaise
Cc: Theodore T
Ouch.
But - we can improve this quite a bit if we just change reqs_active to
mean "number of outstanding requests and unreaped completions" - that
means kiocb allocation doesn't have to look at the ringbuffer, which is a
fairly significant win.
Signed-off-by: Kent Overstreet
Cc: Za
From: Zach Brown
Signed-off-by: Zach Brown
Signed-off-by: Kent Overstreet
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
Cc: Mark Fasheh
Cc: Joel Becker
Cc: Rusty Russell
Cc: Jens Axboe
Cc: Asai Thambi S P
Cc: Selvan Mani
Cc: Sam Bradshaw
Cc: Jeff Moyer
Cc: Al Viro
Cc: Benjamin LaHaise
Cc
instead of using a flag bit switch to using ki_cancel
itself to indicate that a kiocb has been cancelled/completed. This lets
us get rid of ki_flags entirely.
[a...@linux-foundation.org: remove buggy BUG()]
Signed-off-by: Kent Overstreet
Cc: Zach Brown
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
lose the event.
Also convert it to wait_event_interruptible_hrtimeout(), which
simplifies it quite a bit.
Signed-off-by: Kent Overstreet
Cc: Zach Brown
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
Cc: Mark Fasheh
Cc: Joel Becker
Cc: Rusty Russell
Cc: Jens Axboe
Cc: Asai Thambi S P
Cc: Selvan Man
behaviour arising from that too.
I came to the conclusion that if other users end up actually needing the
amount of time remaining, the sanest thing to do would be to create a
version that uses absolute timeouts instead of relative.
[a...@linux-foundation.org: fix description of `timeout' a
us() * 4);
+ BUG_ON(!ctx->req_batch);
+
/* limit the number of system wide aios */
spin_lock(&aio_nr_lock);
if (aio_nr + nr_events > aio_max_nr ||
Benjamin LaHaise (1):
aio: fix kioctx not being freed after cancellation at exit time
Kent Overstreet (27):
call. It only used the AIO retry facility to get the
submitter's mm context for copying the result of a read back to user
space. This is easy to implement with use_mm() and a work struct, much
like kvm does with async_pf_execute() for get_user_pages().
Signed-off-by: Zach Brown
Signed-of
Nothing used the return value, and it probably wasn't possible to use it
safely for the locked versions (aio_complete(), aio_put_req()). Just kill
it.
Signed-off-by: Kent Overstreet
Acked-by: Zach Brown
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
Cc: Mark Fasheh
Cc: Joel Becker
Cc:
Minor refactoring, to get rid of some duplicated code
[a...@linux-foundation.org: fix warning]
Signed-off-by: Kent Overstreet
Cc: Zach Brown
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
Cc: Mark Fasheh
Cc: Joel Becker
Cc: Rusty Russell
Cc: Jens Axboe
Cc: Asai Thambi S P
Cc: Selvan Mani
Cc
oint, other syscalls may still be
submitting kiocbs!
Now, we cancel and wait for outstanding kiocbs in free_ioctx(), after
kioctx->users has dropped to 0 and we know no more iocbs could be
submitted.
Signed-off-by: Kent Overstreet
Cc: Zach Brown
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
Cc: Mark
a...@linux-foundation.org: use fmode_t in aio_run_iocb()]
Signed-off-by: Kent Overstreet
Cc: Zach Brown
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
Cc: Mark Fasheh
Cc: Joel Becker
Cc: Rusty Russell
Cc: Jens Axboe
Cc: Asai Thambi S P
Cc: Selvan Mani
Cc: Sam Bradshaw
Cc: Jeff Moyer
Cc: Al Viro
Cc: Benjam
[a...@linux-foundation.org: make reqs_active __cacheline_aligned_in_smp]
Signed-off-by: Kent Overstreet
Cc: Zach Brown
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
Cc: Mark Fasheh
Cc: Joel Becker
Cc: Rusty Russell
Cc: Jens Axboe
Cc: Asai Thambi S P
Cc: Selvan Mani
Cc: Sam Bradshaw
Cc: Jeff
This just converts the ioctx refcount to the new generic dynamic percpu
refcount code.
Signed-off-by: Kent Overstreet
Cc: Zach Brown
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
Cc: Mark Fasheh
Cc: Joel Becker
Cc: Rusty Russell
Cc: Jens Axboe
Cc: Asai Thambi S P
Cc: Selvan Mani
Cc: Sam
: Benjamin LaHaise
Cc: Kent Overstreet
Cc: Josh Boyer
Cc: Zach Brown
Signed-off-by: Andrew Morton
Signed-off-by: Kent Overstreet
---
fs/aio.c | 15 +++
1 file changed, 3 insertions(+), 12 deletions(-)
diff --git a/fs/aio.c b/fs/aio.c
index 4dbd240..d2c1a82 100644
--- a/fs/aio.c
+++ b/fs
Add a struct batch_complete * argument to bi_end_io; infrastructure to
make use of it comes in the next patch.
Signed-off-by: Kent Overstreet
Cc: Zach Brown
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
Cc: Mark Fasheh
Cc: Joel Becker
Cc: Rusty Russell
Cc: Jens Axboe
Cc: Asai Thambi S P
Cc
tive list. This occurs in two places: kiocb_batch_free(), which
is going away in a later patch, and the error path in io_submit_one.
Signed-off-by: Kent Overstreet
Cc: Zach Brown
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
Cc: Mark Fasheh
Cc: Joel Becker
Cc: Rusty Russell
Cc: Jens Axboe
Cc: Asai
Signed-off-by: Kent Overstreet
Cc: Zach Brown
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
Cc: Mark Fasheh
Cc: Joel Becker
Cc: Rusty Russell
Cc: Jens Axboe
Cc: Asai Thambi S P
Cc: Selvan Mani
Cc: Sam Bradshaw
Cc: Jeff Moyer
Cc: Al Viro
Cc: Benjamin LaHaise
Cc: Theodore Ts'o
Signed-o
removal of the locking around
the unused run list in the submission path.
This has only been compiled.
Signed-off-by: Kent Overstreet
Signed-off-by: Zach Brown
Cc: Zach Brown
Cc: Felipe Balbi
Cc: Greg Kroah-Hartman
Cc: Mark Fasheh
Cc: Joel Becker
Cc: Rusty Russell
Cc: Jens Axboe
Cc: Asai
On Fri, Jan 18, 2013 at 04:37:59PM -0500, Mike Snitzer wrote:
> On Fri, Jan 18 2013 at 4:25pm -0500,
> Darrick J. Wong wrote:
>
> > Since Joe is putting together a testing tree to compare the three caching
> > things, what do you all think of having a(nother) session about ssd caching
> > at
>
On Tue, Feb 05, 2013 at 10:53:00AM -0500, valdis.kletni...@vt.edu wrote:
> On Thu, 31 Jan 2013 16:37:27 -0800, Kent Overstreet said:
> > On Thu, Jan 31, 2013 at 01:59:52PM -0800, Andrew Morton wrote:
> > > Did this get fixed?
>
> > With the patches I sent you, yes -
On Thu, Feb 07, 2013 at 06:57:40AM +0800, Amit Kale wrote:
> > -Original Message-
> > From: Michel Lespinasse [mailto:wal...@google.com]
> > Sent: Friday, February 01, 2013 4:58 PM
> > To: Darrick J. Wong
> > Cc: Amit Kale; linux-bcache; device-mapper developm
On Fri, Feb 08, 2013 at 09:38:11AM -0800, Zach Brown wrote:
> > The draft implementation will look like this. struct bio should have
> > some way to get current status of kiocb that generated bio. So we add
> > a pointer to bool flag.
> >
> > struct bio {
> > bool *cancelled;
> > }
> >
> > in a
On Fri, Feb 08, 2013 at 06:44:08AM -0800, Tejun Heo wrote:
> (cc'ing Andrew)
>
> On Wed, Dec 26, 2012 at 06:00:02PM -0800, Kent Overstreet wrote:
> > This implements a refcount with similar semantics to
> > atomic_get()/atomic_dec_and_test(), that starts out a
On Fri, Feb 08, 2013 at 03:49:02PM +0100, Jens Axboe wrote:
> On Fri, Feb 08 2013, Tejun Heo wrote:
> > (cc'ing Andrew)
> >
> > On Wed, Dec 26, 2012 at 06:00:02PM -0800, Kent Overstreet wrote:
> > > This implements a refcount with similar semantics to
>
I finally started hacking on the dio code, and it's far from done but
it's turning out better than I expected so I thought I'd show off what
I've got so far.
The end result is _vastly_ simpler - direct-io.c is now ~700 lines, vs.
~1300 previously. dio_submit is almost gone, I'm down to 4 things le
On Mon, Feb 11, 2013 at 05:24:13PM -0800, Andi Kleen wrote:
> On Mon, Feb 11, 2013 at 04:53:26PM -0800, Kent Overstreet wrote:
> > I finally started hacking on the dio code, and it's far from done but
> > it's turning out better than I expected so I thought I'd sho
On Tue, Apr 02, 2013 at 12:21:02PM -0400, Sasha Levin wrote:
> Hi all,
>
> It seems that trying to fuzz bcache without any devices triggers a BUG:
>
> That BUG looks very intentional there, any reason for adding it instead of
> exiting
> if there aren't any devices?
The fuzz tester is meant to
On Mon, Jan 28, 2013 at 01:50:42PM -0800, Tejun Heo wrote:
> Hello, Kent.
>
> On Mon, Jan 28, 2013 at 01:45:06PM -0800, Kent Overstreet wrote:
> > Ahh. Bias value sounds... hacky (i.e. harder to convince myself it's
> > correct) but I see what you're getting at.
>
I started screwing aronud just to see how hard a conversion would be and
what it'd look like. I _think_ this is complete, but there's enough
going on I undoubtedly missed something.
Completely untested - builds and that's it. I'm sure it's broken.
Deletes almost 100 lines of code though. I like h
On Tue, Jan 29, 2013 at 11:29:04AM -0800, Tejun Heo wrote:
> Hey, Kent.
>
> On Tue, Jan 29, 2013 at 08:39:42AM -0800, Kent Overstreet wrote:
> > Oh, if this is going to be widely used I should probably have a
> > different implementation for archs that don't have atomic6
On Tue, Jan 29, 2013 at 12:02:18PM -0800, Tejun Heo wrote:
> Hello, Kent.
>
> On Tue, Jan 29, 2013 at 11:51:41AM -0800, Kent Overstreet wrote:
> > > What about overflow? Note that we can have systemetic cases where ref
> > > is gotten on one cpu and put on anoth
On Thu, Jan 31, 2013 at 01:59:52PM -0800, Andrew Morton wrote:
> On Tue, 22 Jan 2013 16:28:18 -0500
> valdis.kletni...@vt.edu wrote:
>
> > On Tue, 22 Jan 2013 21:43:27 +0800, Hillf Danton said:
> > > On Mon, Jan 21, 2013 at 9:24 PM, Valdis Kletnieks
> > > wrote:
> > > > Am seeing a reproducible B
On Fri, Feb 1, 2013 at 4:57 PM, Michel Lespinasse wrote:
> On Fri, Feb 1, 2013 at 4:44 PM, Darrick J. Wong
> wrote:
>> This is a patch to migrate STEC's enhanceio driver out of their github
>> repository and into the staging tree. From their README:
>>
>> "EnhanceIO driver is based on EnhanceIO
Also, where's the end user documentation?
On Fri, Feb 1, 2013 at 5:23 PM, Kent Overstreet wrote:
> On Fri, Feb 1, 2013 at 4:57 PM, Michel Lespinasse wrote:
>> On Fri, Feb 1, 2013 at 4:44 PM, Darrick J. Wong
>> wrote:
>>> This is a patch to migrate STEC's en
On Thu, Jan 24, 2013 at 12:22:21PM -0500, valdis.kletni...@vt.edu wrote:
> On Wed, 23 Jan 2013 20:10:03 +0800, Hillf Danton said:
>
> > Try again?
> > ---
> >
> > --- a/fs/aio.c Tue Jan 22 21:37:54 2013
> > +++ b/fs/aio.c Wed Jan 23 20:06:14 2013
>
> Now seeing this:
>
> [ 2941.495370]
On Thu, Jan 24, 2013 at 01:27:59PM -0800, Andrew Morton wrote:
> On Thu, 24 Jan 2013 13:18:50 -0800
> Kent Overstreet wrote:
>
> > So, Andrew - that "smoosh struct kiocb" patch should just be dropped,
> > even if I fixed that issue clearly the idea is a lot less sa
The batch completion code was trying to be a bit too clever, and skip
checking ctx where it couldn't be NULL - but that broke if a kiocb had
been cancelled. Move the check to kioctx_ring_unlock().
Reported-by: Valdis Kletnieks
Signed-off-by: Kent Overstreet
---
fs/aio.c | 6 --
1
The "aio: kill ki-retry" patch was assuming that we didn't touch struct
kiocb after passing it off to something that would call aio_complete() -
which was wrong. So, revert the refcounting changes.
Signed-off-by: Kent Overstreet
---
fs/aio.c | 6 --
1 file changed, 4 i
fficient since we're synchronizing with aio_complete()
which isn't taking any locks.
Signed-off-by: Kent Overstreet
---
fs/aio.c| 32 ++--
include/linux/aio.h | 11 +++
2 files changed, 33 insertions(+), 10 deletions(-)
diff --git a/fs/
On Thu, Jan 24, 2013 at 01:27:59PM -0800, Andrew Morton wrote:
> Please also take a look at Jan's recent
> http://www.spinics.net/lists/linux-fsdevel/msg61738.html and have a
> think about how this plays with your patchset.
I can't think of any possible interactions - none of my aio stuff messes
w
On Thu, Jan 24, 2013 at 02:25:37PM -0800, Zach Brown wrote:
> > No, I didn't see that bug until after I'd fixed the other three, but as
> > far as I can tell everything's fixed with the patches I'm about to mail
> > out - my test VM has been running for the past two days without errors,
> > it's ki
On Thu, Jan 17, 2013 at 03:39:40AM -0800, Kent Overstreet wrote:
> Suppose I could fill out the bcache version...
>
> On Thu, Jan 17, 2013 at 05:52:00PM +0800, Amit Kale wrote:
> > 11. Error conditions - Handling power failures, intermittent and permanent
> > device failures
On Fri, Jan 18, 2013 at 05:08:37PM +0800, Amit Kale wrote:
> > From: Jason Warr [mailto:ja...@warr.net]
> > On 01/17/2013 11:53 AM, Amit Kale wrote:
> > >>> 9. Performance - Throughput is generally most important. Latency is
> > >> > also one more performance comparison point. Performance under
>
On Thu, Jan 24, 2013 at 04:51:36PM -0800, Tejun Heo wrote:
> (cc'ing percpu / rcu crowd)
>
> Hello, Kent.
>
> On Wed, Dec 26, 2012 at 06:00:02PM -0800, Kent Overstreet wrote:
> > This implements a refcount with similar semantics to
> > atomic_get()/atomic_dec_and
On Fri, Jan 25, 2013 at 03:12:51PM -0800, Andrew Morton wrote:
> On Fri, 25 Jan 2013 21:30:32 +0800
> Hillf Danton wrote:
>
> > On Fri, Jan 25, 2013 at 5:43 AM, Kent Overstreet
> > wrote:
> > > The cancellation changes were fubar - we can't cancel a kiocb
On Thu, Jan 24, 2013 at 06:09:43PM -0800, Tejun Heo wrote:
> Hello, again.
>
> On Thu, Jan 24, 2013 at 06:03:40PM -0800, Tejun Heo wrote:
> > Yeah, if we're aiming to replace refcnts in file and kobj, dynamic
> > alloc may be justified. Hopefully, the accounting necessary to decide
> > whethre to
On Fri, Jan 25, 2013 at 04:45:10PM +1030, Rusty Russell wrote:
> Tejun Heo writes:
> >> It also implements two stage shutdown, as we need it to tear down the
> >> percpu counts. Before dropping the initial refcount, you must call
> >> percpu_ref_kill(); this puts the refcount in "shutting down mod
1 - 100 of 1414 matches
Mail list logo
