(2014/02/06 1:55), James Bottomley wrote:
On Wed, 2014-02-05 at 14:47 +0900, Eiichi Tsukata wrote:
Currently, scsi error handling in scsi_decide_disposition() tries to
unconditionally requeue scsi command when device keeps some error state.
This is because retryable errors are thought to be temp
> diff --git a/block/blk-mq.h b/block/blk-mq.h
> index b771080..88a2366 100644
> --- a/block/blk-mq.h
> +++ b/block/blk-mq.h
> @@ -30,6 +30,8 @@ void blk_mq_init_flush(struct request_queue *q);
> void blk_mq_drain_queue(struct request_queue *q);
> void blk_mq_free_queue(struct request_queue *q);
On Wed, Feb 5, 2014 at 4:41 AM, Christoph Hellwig wrote:
> Switch to using a preallocated flush_rq for blk-mq similar to what's done
> with the old request path. This allows us to set up the request properly
> with a tag from the actually allowed range and ->rq_disk as needed by
> some drivers.
On Wed, 2014-02-05 at 04:39 -0800, Christoph Hellwig wrote:
> If we don't have starved devices we don't need to take the host lock
> to iterate over them. Also split the function up to be more clear.
This looks reasonable and can potentially go separately too.
> Signed-off-by: Christoph Hellwig
On Wed, 2014-02-05 at 04:39 -0800, Christoph Hellwig wrote:
> Just have one level of alloc/free functions that take a host instead
> of two levels for the allocation and different calling conventions
> for the free.
>
> Signed-off-by: Christoph Hellwig
> ---
> drivers/scsi/scsi.c | 77
>
On Wed, 2014-02-05 at 04:39 -0800, Christoph Hellwig wrote:
> Avoid hitting the host-wide free_list lock unless we need to put a command
> back onto the freelist.
This one looks like it can go independently on its own.
> Signed-off-by: Christoph Hellwig
> ---
> drivers/scsi/scsi.c | 13 ++
On Wed, 2014-02-05 at 04:39 -0800, Christoph Hellwig wrote:
> This series contains various optimizations for the SCSI data I/O path.
> They increase the number of IOPS seen with iSCSI or SRP between 2%
> and 3.5% in workloads that previously hit the host_lock hard. While this
> isn't a lot it now
Subsystems that want to register CPU hotplug callbacks, as well as perform
initialization for the CPUs that are already online, often do it as shown
below:
get_online_cpus();
for_each_online_cpu(cpu)
init_cpu(cpu);
register_cpu_notifier(&foobar_cpu_notifie
Subsystems that want to register CPU hotplug callbacks, as well as perform
initialization for the CPUs that are already online, often do it as shown
below:
get_online_cpus();
for_each_online_cpu(cpu)
init_cpu(cpu);
register_cpu_notifier(&foobar_cpu_notifie
Subsystems that want to register CPU hotplug callbacks, as well as perform
initialization for the CPUs that are already online, often do it as shown
below:
get_online_cpus();
for_each_online_cpu(cpu)
init_cpu(cpu);
register_cpu_notifier(&foobar_cpu_notifie
Hi nab, I'm getting back to looking at this patchset, but wanted to just
discuss and understand this one first because all the kref ones are
similar. see below.
On 12/16/2013 12:52 PM, Nicholas A. Bellinger wrote:
> On Fri, 2013-12-13 at 15:58 -0800, Andy Grover wrote:
>> Use kref to handle refere
On 01/29/2014 01:16 AM, Jayamohan Kallickal wrote:
> From: Jayamohan Kallickal
>
> This patchset updates be2iscsi driver to 10.2.125.0.
>
> Regarding comments by Mike Christie on Version 1
>
> Patch 5/7) - converted to common function
> Patch 4/7 - Will withdraw for now as there is lots o
From: Paolo Bonzini
---
This is a commit scheduled for the next v2.6.34 longterm release.
http://git.kernel.org/?p=linux/kernel/git/paulg/longterm-queue-2.6.34.git
If you see a problem with using this for longterm, please comment.
From: Paolo Bonzini
---
This is a commit scheduled for the next v2.6.34 longterm release.
http://git.kernel.org/?p=linux/kernel/git/paulg/longterm-queue-2.6.34.git
If you see a problem with using this for longterm, please comment.
On Wed, 2014-02-05 at 14:47 +0900, Eiichi Tsukata wrote:
> Currently, scsi error handling in scsi_decide_disposition() tries to
> unconditionally requeue scsi command when device keeps some error state.
> This is because retryable errors are thought to be temporary and the scsi
> device will soon r
On 02/03/2014 09:55 PM, Martin K. Petersen wrote:
"David" == David Milburn writes:
David> When enabling DIX T10-DIF-TYPE1-IP protection you can hit the
David> bip_vec full condition which fails to attach the integrity
David> metadata and returns 0 back to bio_integrity_prep()
Looks like Kent
Guten Tag,
wieso zahlen Sie für einen Kredit so hohe Zinsen? Jetzt (noch) haben Sie bei
ausländischen Instituten die besten Chancen.
Also ideal für jeden, der frisches Geld braucht und insbesondere für
Selbständige.
Die Zinsen liegen zwischen 4,5% und 5,5% Zinsen p.a., die Kredite haben
Laufzei
__blk_put_request needs to call into the blk-mq code just like
blk_put_request. As we don't have the queue lock in this case both
end up calling the same function.
Signed-off-by: Christoph Hellwig
---
block/blk-core.c |5 +
1 file changed, 5 insertions(+)
diff --git a/block/blk-core.c
Instead of letting the ULD play games with the prep_fn move back to
the model of a central prep_fn with a callback to the ULD. This
already cleans up and shortens the code by itself, and will be required
to properly support blk-mq in the SCSI midlayer.
Signed-off-by: Christoph Hellwig
---
drive
To behave the same way as the old request path.
Signed-off-by: Christoph Hellwig
---
block/blk-mq.c |2 ++
1 file changed, 2 insertions(+)
diff --git a/block/blk-mq.c b/block/blk-mq.c
index 0b72927..d116282 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -1363,6 +1363,8 @@ struct reques
This will have to be funnelled to blk-mq directly, but skip it for now.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/scsi.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/scsi/scsi.c b/drivers/scsi/scsi.c
index cf5c110..014b642 100644
--- a/drivers/scsi/scsi.
Add support for using the blk-mq code to submit requests to SCSI
drivers. There is very little blk-mq specific code, but that's
partially because important functionality like partial completions
and request requeueing is still missing in blk-mq. I hope to keep
most of the additions for these in t
Factor code common out so that it can be reused independently for the blk-mq
code path.
Split from a bigger patch by Nicholas Bellinger.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/scsi_lib.c | 18 +++---
1 file changed, 11 insertions(+), 7 deletions(-)
diff --git a/drivers
Now that scsi doesn't use it anymore there's no user left.
Signed-off-by: Christoph Hellwig
---
block/blk-core.c | 11 ++-
block/blk-settings.c | 17 -
include/linux/blkdev.h |3 ---
3 files changed, 2 insertions(+), 29 deletions(-)
diff --git a/block/blk
Factor out a helper to set the _blocked values, which we'll reuse for the
blk-mq code path.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/scsi_lib.c | 44 ++--
1 file changed, 26 insertions(+), 18 deletions(-)
diff --git a/drivers/scsi/scsi_lib.c b/
Make sure we only have the logic for requeing commands in one place.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/scsi.c | 36 +---
drivers/scsi/scsi_lib.c |6 --
2 files changed, 17 insertions(+), 25 deletions(-)
diff --git a/drivers/scsi/scsi
Signed-off-by: Christoph Hellwig
---
drivers/scsi/virtio_scsi.c |2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/scsi/virtio_scsi.c b/drivers/scsi/virtio_scsi.c
index d9a6074..59b030c 100644
--- a/drivers/scsi/virtio_scsi.c
+++ b/drivers/scsi/virtio_scsi.c
@@ -680,6 +680,7 @@ stat
Don't bother returning a stale pointer when a bool does the work much
better.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/scsi_lib.c | 34 --
1 file changed, 16 insertions(+), 18 deletions(-)
diff --git a/drivers/scsi/scsi_lib.c b/drivers/scsi/scsi_lib.c
Switch to using a preallocated flush_rq for blk-mq similar to what's done
with the old request path. This allows us to set up the request properly
with a tag from the actually allowed range and ->rq_disk as needed by
some drivers. To make life easier we also switch to dynamic allocation
of ->flus
Make blk-mq handle the dma_drain_size field the same way as the old request
path.
Signed-off-by: Christoph Hellwig
---
block/blk-mq.c | 10 ++
1 file changed, 10 insertions(+)
diff --git a/block/blk-mq.c b/block/blk-mq.c
index 6838fe8..0b72927 100644
--- a/block/blk-mq.c
+++ b/block/b
As explained in my earlier proposal I think we need to move to blk-mq
for the SCSI midlayer quickly, and in a way that keeps all the existing
queueing mechanisms working. This series shows my current progress towards
this goal. It works under medium loads for iSCSI and virtio_scsi and I'm
fairly
This is neede for proper SG_IO operation as well as various uses of
blk_execute_rq from the SCSI midlayer.
Signed-off-by: Christoph Hellwig
---
block/blk-exec.c |4 ++--
block/blk-mq.c | 17 ++---
block/blk-mq.h |2 ++
include/linux/blk-mq.h |1 -
Signed-off-by: Christoph Hellwig
---
drivers/scsi/iscsi_tcp.c |1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/scsi/iscsi_tcp.c b/drivers/scsi/iscsi_tcp.c
index add6d15..44aae3d 100644
--- a/drivers/scsi/iscsi_tcp.c
+++ b/drivers/scsi/iscsi_tcp.c
@@ -957,6 +957,7 @@ static struct sc
Avoid taking the host-wide host_lock to check the per-target queue limit.
Instead we do an atomic_inc_return early on to grab our slot in the queue,
and if nessecary decrement it after finishing all checks.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/scsi_lib.c| 52 ++
If we don't have starved devices we don't need to take the host lock
to iterate over them. Also split the function up to be more clear.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/scsi_lib.c | 31 ---
1 file changed, 12 insertions(+), 19 deletions(-)
diff --
Taken almost entirely from Nicholas Bellinger's scsi-mq conversion.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/virtio_scsi.c | 25 +++--
1 file changed, 7 insertions(+), 18 deletions(-)
diff --git a/drivers/scsi/virtio_scsi.c b/drivers/scsi/virtio_scsi.c
index 16bfd
SCSI devices may only be removed by calling scsi_remove_device().
That function must invoke blk_cleanup_queue() before the final put
of sdev->sdev_gendev. Since blk_cleanup_queue() waits for the
block queue to drain and then tears it down, scsi_request_fn cannot
be active anymore after blk_cleanup_
We don't use the passed in scsi command for anything, so just add a
adapter-wide internal status to go along with the internal scb that
is used unter int_mtx to pass back the return value and get rid of
all the complexities and abuse of the scsi_cmnd structure.
Signed-off-by: Christoph Hellwig
--
Avoid taking the queue_lock to check the per-device queue limit. Instead
we do an atomic_inc_return early on to grab our slot in the queue,
and if nessecary decrement it after finishing all checks.
Unlike the host and target busy counters this doesn't allow us to avoid the
queue_lock in the reque
Avoid taking the host-wide host_lock to check the per-host queue limit.
Instead we do an atomic_inc_return early on to grab our slot in the queue,
and if nessecary decrement it after finishing all checks.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/advansys.c |4 +-
drivers
All users are gone, so we can get rid of these.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/scsi.c | 56 --
include/scsi/scsi_cmnd.h |3 ---
2 files changed, 59 deletions(-)
diff --git a/drivers/scsi/scsi.c b/drivers/scsi/scsi.c
index
Eliminate a get_device() / put_device() pair from scsi_next_command().
Both are atomic operations hence removing these slightly improves
performance.
[hch: slight changes due to different context]
Signed-off-by: Bart Van Assche
Signed-off-by: Christoph Hellwig
---
drivers/scsi/scsi_lib.c |
Signed-off-by: Christoph Hellwig
---
drivers/scsi/scsi_error.c |6 ++
1 file changed, 6 insertions(+)
diff --git a/drivers/scsi/scsi_error.c b/drivers/scsi/scsi_error.c
index 78b004d..f8b54c1 100644
--- a/drivers/scsi/scsi_error.c
+++ b/drivers/scsi/scsi_error.c
@@ -2289,6 +2289,11 @@ sc
This series contains various optimizations for the SCSI data I/O path.
They increase the number of IOPS seen with iSCSI or SRP between 2%
and 3.5% in workloads that previously hit the host_lock hard. While this
isn't a lot it now fully shifts the contention to the queue_lock, which
will get out of
Avoid hitting the host-wide free_list lock unless we need to put a command
back onto the freelist.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/scsi.c | 13 +++--
1 file changed, 7 insertions(+), 6 deletions(-)
diff --git a/drivers/scsi/scsi.c b/drivers/scsi/scsi.c
index ebcea6c.
Seems like these counters are missing any sort of synchronization for
updates, as a over 10 year old comment from me noted. Fix this by
using atomic counters, and while we're at it also make sure they are
in the same cacheline as the _busy counters and not needlessly stored
to in every I/O complet
Avoid a spurious device get/put cycle by using scsi_put_command and folding
scsi_unprep_request into scsi_requeue_command.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/scsi_lib.c | 35 +++
1 file changed, 3 insertions(+), 32 deletions(-)
diff --git a/drive
Prepare for not taking a host-wide lock in the dispatch path by pushing
the lock down into the places that actually need it. Note that this
patch is just a preparation step, as it will actually increase lock
roundtrips and thus decrease performance on its own.
Signed-off-by: Christoph Hellwig
--
Many callers won't need this and we can optimize them away. In addition
the handling in the __-prefixed variants was inconsistant to start with.
Based on an earlier patch from Bart Van Assche.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/scsi.c | 36 -
This allows drivers to specify the size of their per-command private
data in the host template and then get extra memory allocated for
each command instead of needing another allocation in ->queuecommand.
With the current SCSI code that already does multiple allocations for
each command this proba
Just have one level of alloc/free functions that take a host instead
of two levels for the allocation and different calling conventions
for the free.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/scsi.c | 77 +++
1 file changed, 22 insertions
We've run into many issues where the SCSI layer simply does not scale to
keep up with today's hardware, be that in simple single-thread IOPs, or
in lock contention when using multiple LUNs or targets under a single
SCSI host. This proposal tries to draw a path how to fix this properly
and avoids w
--
Hello,
The Project is about the exportation of 100,000 barrels of Light Crude
Oil daily out from Iraq to Turkey through my client's company in Iraq
at the rate of $92.00 a barrel. This amount to $9,200,000 daily. I ask
for your support as a foreigner to handle this business project with my
c
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
On 01/15/2014 09:37 PM, Alan Stern wrote:
> Evidently some wacky USB-ATA bridges don't recognize the SYNCHRONIZE
> CACHE command, as shown in this email thread:
>
> http://marc.info/?t=13897835622&r=1&w=2
>
> The fact that we can't tell them to drain their caches shouldn't
> prevent the
55 matches
Mail list logo
