Before commit 0df21c86bdbf ("scsi: implement .get_budget and .put_budget
for blk-mq"), we run queue after 3ms if queue is idle and SCSI device
queue isn't ready, which is done in handling BLK_STS_RESOURCE. After
commit 0df21c86bdbf is introduced, queue won't be run any more under
this situation.
On Tue, Dec 05, 2017 at 03:19:46PM +0900, Byungchul Park wrote:
> On 12/5/2017 2:46 PM, Byungchul Park wrote:
> > On 12/5/2017 2:30 PM, Matthew Wilcox wrote:
> > > On Mon, Dec 04, 2017 at 02:16:19PM +0900, Byungchul Park wrote:
> > > > For now, wait_for_completion() / complete() works with
Tang Junhui--
Hi
On 12/03/2017 10:11 PM, tang.jun...@zte.com.cn wrote:
> From: Tang Junhui
>
> Hello Mike & Coly
>
> Could you please have a reveiw for this patch?
>
>> From: Tang Junhui
>>
>> In such scenario that there are some flash only
On Tue, Dec 05, 2017 at 01:16:24PM +0800, Ming Lei wrote:
> On Mon, Dec 04, 2017 at 11:48:07PM +, Holger Hoffstätte wrote:
> > On Tue, 05 Dec 2017 06:45:08 +0800, Ming Lei wrote:
> >
> > > On Mon, Dec 04, 2017 at 03:09:20PM +, Bart Van Assche wrote:
> > >> On Sun, 2017-12-03 at 00:31
On 12/5/2017 2:46 PM, Byungchul Park wrote:
On 12/5/2017 2:30 PM, Matthew Wilcox wrote:
On Mon, Dec 04, 2017 at 02:16:19PM +0900, Byungchul Park wrote:
For now, wait_for_completion() / complete() works with lockdep, add
lock_page() / unlock_page() and its family to lockdep support.
Changes
On 12/5/2017 2:30 PM, Matthew Wilcox wrote:
On Mon, Dec 04, 2017 at 02:16:19PM +0900, Byungchul Park wrote:
For now, wait_for_completion() / complete() works with lockdep, add
lock_page() / unlock_page() and its family to lockdep support.
Changes from v1
- Move lockdep_map_cross outside of
On Mon, Dec 04, 2017 at 02:16:19PM +0900, Byungchul Park wrote:
> For now, wait_for_completion() / complete() works with lockdep, add
> lock_page() / unlock_page() and its family to lockdep support.
>
> Changes from v1
> - Move lockdep_map_cross outside of page_ext to make it flexible
> -
On Mon, Dec 04, 2017 at 11:48:07PM +, Holger Hoffstätte wrote:
> On Tue, 05 Dec 2017 06:45:08 +0800, Ming Lei wrote:
>
> > On Mon, Dec 04, 2017 at 03:09:20PM +, Bart Van Assche wrote:
> >> On Sun, 2017-12-03 at 00:31 +0800, Ming Lei wrote:
> >> > Fixes: 0df21c86bdbf ("scsi: implement
On Tue, Dec 05, 2017 at 01:59:51AM +, Bart Van Assche wrote:
> On Tue, 2017-12-05 at 09:15 +0800, Ming Lei wrote:
> > On Mon, Dec 04, 2017 at 04:38:08PM -0800, Bart Van Assche wrote:
> > > Since the next patch will make it possible that scsi_show_rq() gets
> > > called before the CDB pointer
On Mon, Dec 04, 2017 at 10:42:28PM -0500, Martin K. Petersen wrote:
>
> Hi Ming,
>
> > Please cook a patch for fixing the crash issue only, since we need
> > to backport the fix to stable kernel.
>
> I thought you were going to submit a V5 that addressed James' concerns?
>
> --
> Martin K.
On Tue, 2017-12-05 at 09:15 +0800, Ming Lei wrote:
> On Mon, Dec 04, 2017 at 04:38:08PM -0800, Bart Van Assche wrote:
> > Since the next patch will make it possible that scsi_show_rq() gets
> > called before the CDB pointer is changed into a non-NULL value,
> > only show the CDB if the CDB pointer
On Tue, Dec 05, 2017 at 01:13:43AM +, Bart Van Assche wrote:
> On Tue, 2017-12-05 at 09:04 +0800, Ming Lei wrote:
> > Then no reason to revert commit(0df21c86bdbf scsi: implement .get_budget an
> > .put_budget for blk-mq) for one issue which may never happen in reality
> > since
> > this
On Tue, 2017-12-05 at 09:04 +0800, Ming Lei wrote:
> Then no reason to revert commit(0df21c86bdbf scsi: implement .get_budget an
> .put_budget for blk-mq) for one issue which may never happen in reality since
> this reproducer need out-of-tree patch.
Sorry but I disagree completely. You seem to
On Mon, Dec 04, 2017 at 04:38:08PM -0800, Bart Van Assche wrote:
> Since the next patch will make it possible that scsi_show_rq() gets
> called before the CDB pointer is changed into a non-NULL value,
> only show the CDB if the CDB pointer is not NULL. Additionally,
> show the request timeout and
On Tue, Dec 05, 2017 at 12:29:59AM +, Bart Van Assche wrote:
> On Tue, 2017-12-05 at 08:20 +0800, Ming Lei wrote:
> > Also it is a bit odd to see request in hctx->dispatch now, and it can only
> > happen now when scsi_target_queue_ready() returns false, so I guess you
> > apply
> > some
Hello Jens,
While debugging an issue with the SCSI error handler I noticed that commands
that got stuck in that error handler are not shown in debugfs. That is very
annoying for anyone who relies on the information in debugfs for root-causing
such an issue. Hence this patch series that makes sure
Since the next patch will make it possible that scsi_show_rq() gets
called before the CDB pointer is changed into a non-NULL value,
only show the CDB if the CDB pointer is not NULL. Additionally,
show the request timeout and SCSI command flags. This patch also
fixes a bug that was reported by Ming
When debugging e.g. the SCSI timeout handler it is important that
requests that have not yet been started or that already have
completed are also reported through debugfs.
Signed-off-by: Bart Van Assche
Cc: Ming Lei
Cc: Christoph Hellwig
On Tue, 2017-12-05 at 08:20 +0800, Ming Lei wrote:
> Also it is a bit odd to see request in hctx->dispatch now, and it can only
> happen now when scsi_target_queue_ready() returns false, so I guess you apply
> some change on target->can_queue(such as setting it as 1 in srp/ib code
> manually)?
On Mon, Dec 04, 2017 at 11:32:27PM +, Bart Van Assche wrote:
> On Tue, 2017-12-05 at 07:01 +0800, Ming Lei wrote:
> > On Mon, Dec 04, 2017 at 10:48:18PM +, Bart Van Assche wrote:
> > > On Tue, 2017-12-05 at 06:42 +0800, Ming Lei wrote:
> > > > On Mon, Dec 04, 2017 at 09:30:32AM -0800, Bart
On Tue, 05 Dec 2017 06:45:08 +0800, Ming Lei wrote:
> On Mon, Dec 04, 2017 at 03:09:20PM +, Bart Van Assche wrote:
>> On Sun, 2017-12-03 at 00:31 +0800, Ming Lei wrote:
>> > Fixes: 0df21c86bdbf ("scsi: implement .get_budget and .put_budget for
>> > blk-mq")
>>
>> It might be safer to revert
On Tue, 2017-12-05 at 07:01 +0800, Ming Lei wrote:
> On Mon, Dec 04, 2017 at 10:48:18PM +, Bart Van Assche wrote:
> > On Tue, 2017-12-05 at 06:42 +0800, Ming Lei wrote:
> > > On Mon, Dec 04, 2017 at 09:30:32AM -0800, Bart Van Assche wrote:
> > > > * A systematic lockup for SCSI queues with
On Mon, Dec 04, 2017 at 03:12:15PM -0800, Omar Sandoval wrote:
> From: Omar Sandoval
>
> Commit 8cf466602028 ("kyber: fix hang on domain token wait queue") fixed
> a hang caused by leaving wait entries on the domain token wait queue
> after the __sbitmap_queue_get() retry
On 05.12.2017 01:59, Jeff Moyer wrote:
> Kirill Tkhai writes:
>
>> On 05.12.2017 00:52, Tejun Heo wrote:
>>> Hello, Kirill.
>>>
>>> On Tue, Dec 05, 2017 at 12:44:00AM +0300, Kirill Tkhai wrote:
> Can you please explain how this is a fundamental resource which can't
From: Omar Sandoval
Commit 8cf466602028 ("kyber: fix hang on domain token wait queue") fixed
a hang caused by leaving wait entries on the domain token wait queue
after the __sbitmap_queue_get() retry succeeded, making that wait entry
a "dud" which won't in turn wake more entries
On 05.12.2017 02:02, Tejun Heo wrote:
> Hello, Kirill.
>
> On Tue, Dec 05, 2017 at 01:49:42AM +0300, Kirill Tkhai wrote:
>>> If the only reason is kernel memory consumption protection, the only
>>> thing we need to do is making sure that memory used for aio commands
>>> are accounted against
Hello, Kirill.
On Tue, Dec 05, 2017 at 01:49:42AM +0300, Kirill Tkhai wrote:
> > If the only reason is kernel memory consumption protection, the only
> > thing we need to do is making sure that memory used for aio commands
> > are accounted against cgroup kernel memory consumption and
> >
On Mon, Dec 04, 2017 at 10:48:18PM +, Bart Van Assche wrote:
> On Tue, 2017-12-05 at 06:42 +0800, Ming Lei wrote:
> > On Mon, Dec 04, 2017 at 09:30:32AM -0800, Bart Van Assche wrote:
> > > * A systematic lockup for SCSI queues with queue depth 1. The
> > > following test reproduces that bug
Kirill Tkhai writes:
> On 05.12.2017 00:52, Tejun Heo wrote:
>> Hello, Kirill.
>>
>> On Tue, Dec 05, 2017 at 12:44:00AM +0300, Kirill Tkhai wrote:
Can you please explain how this is a fundamental resource which can't
be controlled otherwise?
>>>
>>> Currently,
On Tue, 2017-12-05 at 06:45 +0800, Ming Lei wrote:
> On Mon, Dec 04, 2017 at 03:09:20PM +, Bart Van Assche wrote:
> > On Sun, 2017-12-03 at 00:31 +0800, Ming Lei wrote:
> > > Fixes: 0df21c86bdbf ("scsi: implement .get_budget and .put_budget for
> > > blk-mq")
> >
> > It might be safer to
On 05.12.2017 00:52, Tejun Heo wrote:
> Hello, Kirill.
>
> On Tue, Dec 05, 2017 at 12:44:00AM +0300, Kirill Tkhai wrote:
>>> Can you please explain how this is a fundamental resource which can't
>>> be controlled otherwise?
>>
>> Currently, aio_nr and aio_max_nr are global. In case of containers
On Tue, 2017-12-05 at 06:42 +0800, Ming Lei wrote:
> On Mon, Dec 04, 2017 at 09:30:32AM -0800, Bart Van Assche wrote:
> > * A systematic lockup for SCSI queues with queue depth 1. The
> > following test reproduces that bug systematically:
> > - Change the SRP initiator such that SCSI target
On Mon, Dec 04, 2017 at 03:09:20PM +, Bart Van Assche wrote:
> On Sun, 2017-12-03 at 00:31 +0800, Ming Lei wrote:
> > Fixes: 0df21c86bdbf ("scsi: implement .get_budget and .put_budget for
> > blk-mq")
>
> It might be safer to revert commit 0df21c86bdbf instead of trying to fix all
> issues
On Mon, Dec 04, 2017 at 09:30:32AM -0800, Bart Van Assche wrote:
> Commit 0df21c86bdbf introduced several bugs:
> * A SCSI queue stall for queue depths > 1, addressed by commit
> 88022d7201e9 ("blk-mq: don't handle failure in .get_budget")
This one is committed already.
> * A systematic lockup
Hello, Kirill.
On Tue, Dec 05, 2017 at 12:44:00AM +0300, Kirill Tkhai wrote:
> > Can you please explain how this is a fundamental resource which can't
> > be controlled otherwise?
>
> Currently, aio_nr and aio_max_nr are global. In case of containers this
> means that a single container may
On 05.12.2017 00:35, Jeff Moyer wrote:
> Kirill Tkhai writes:
>
>> Hi, Benjamin,
>>
>> On 04.12.2017 19:52, Benjamin LaHaise wrote:
>>> Hi Kirill,
>>>
>>> On Mon, Dec 04, 2017 at 07:12:51PM +0300, Kirill Tkhai wrote:
Hi,
this patch set introduces accounting
Hello, Tejun,
On 04.12.2017 23:07, Tejun Heo wrote:
> On Mon, Dec 04, 2017 at 07:12:51PM +0300, Kirill Tkhai wrote:
>> this patch set introduces accounting aio_nr and aio_max_nr per blkio cgroup.
>> It may be used to limit number of aio requests, which are available for
>> a cgroup, and could be
Kirill Tkhai writes:
> Hi, Benjamin,
>
> On 04.12.2017 19:52, Benjamin LaHaise wrote:
>> Hi Kirill,
>>
>> On Mon, Dec 04, 2017 at 07:12:51PM +0300, Kirill Tkhai wrote:
>>> Hi,
>>>
>>> this patch set introduces accounting aio_nr and aio_max_nr per blkio cgroup.
>>> It may
Hi, Benjamin,
On 04.12.2017 19:52, Benjamin LaHaise wrote:
> Hi Kirill,
>
> On Mon, Dec 04, 2017 at 07:12:51PM +0300, Kirill Tkhai wrote:
>> Hi,
>>
>> this patch set introduces accounting aio_nr and aio_max_nr per blkio cgroup.
>> It may be used to limit number of aio requests, which are
Hello, Kirill.
On Mon, Dec 04, 2017 at 07:12:51PM +0300, Kirill Tkhai wrote:
> this patch set introduces accounting aio_nr and aio_max_nr per blkio cgroup.
> It may be used to limit number of aio requests, which are available for
> a cgroup, and could be useful for containers.
Can you please
On Mon, Dec 04, 2017 at 05:48:36PM +0800, Xiao Yang wrote:
> On 2017/12/04 17:25, Eryu Guan wrote:
> > On Mon, Dec 04, 2017 at 05:15:17PM +0800, Xiao Yang wrote:
> > > On 2017/12/04 16:29, Eryu Guan wrote:
> > > > On Wed, Nov 29, 2017 at 08:02:26PM +0800, Xiao Yang wrote:
> > > > > If the entire
Commit 0df21c86bdbf introduced several bugs:
* A SCSI queue stall for queue depths > 1, addressed by commit
88022d7201e9 ("blk-mq: don't handle failure in .get_budget")
* A systematic lockup for SCSI queues with queue depth 1. The
following test reproduces that bug systematically:
- Change
Hi Kirill,
On Mon, Dec 04, 2017 at 07:12:51PM +0300, Kirill Tkhai wrote:
> Hi,
>
> this patch set introduces accounting aio_nr and aio_max_nr per blkio cgroup.
> It may be used to limit number of aio requests, which are available for
> a cgroup, and could be useful for containers.
>
> The
The generic version now takes dma_pfn_offset into account, so there is no
more need for an architecture override.
Signed-off-by: Christoph Hellwig
---
arch/arm64/include/asm/dma-mapping.h | 9 -
1 file changed, 9 deletions(-)
diff --git
This makes sure the generic version can be used with architectures /
devices that have a DMA offset in the direct mapping.
Signed-off-by: Christoph Hellwig
---
include/linux/dma-mapping.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git
The current PCI_DMA_BUS_IS_PHYS decided if a dma implementation is bound
by the dma mask in the device because it directly maps to a physical
address range (modulo an offset in the device), or if it is virtualized
by an iommu and can map any address (that includes virtual iommus like
swiotlb).
Hi all,
this small series tries to get rid of the global and misnamed
PCI_DMA_BUS_IS_PHYS flag, and replace it with a setting in each
struct dma_map_ops instance.
On Sun, Dec 03, 2017 at 10:49:23PM +0100, Richard Weinberger wrote:
> Convert the driver to the modern blk-mq framework.
> As byproduct we get rid of our open coded restart logic and let
> blk-mq handle it.
>
> Signed-off-by: Richard Weinberger
> ---
> arch/um/drivers/ubd_kern.c
On Wed, Nov 29, 2017 at 08:18:09PM +0100, Christian Borntraeger wrote:
> Works fine under KVM with virtio-blk, but still hangs during boot in an LPAR.
> FWIW, the system not only has scsi disks via fcp but also DASDs as a boot
> disk.
> Seems that this is the place where the system stops. (see
Next patch will use the values in more files, so let's make them visible
external.
Signed-off-by: Kirill Tkhai
---
fs/aio.c|4 ++--
include/linux/aio.h |4
2 files changed, 6 insertions(+), 2 deletions(-)
diff --git a/fs/aio.c b/fs/aio.c
index
There is no functional changes, only a preparation
for next patches.
Signed-off-by: Kirill Tkhai
---
fs/aio.c | 44
1 file changed, 32 insertions(+), 12 deletions(-)
diff --git a/fs/aio.c b/fs/aio.c
index
Add a file to configure per-cgroup maximum number of available
aio requests.
Allow write the values, which are less then currently allocated
numbers of requests.
Show numbers of currently allocated and maximum available requests.
Signed-off-by: Kirill Tkhai
---
This patch adds accounting of number of requests of allocated aio
contexts per blkio cgroup, and aggregates child cgroups requests
up the hierarhy. This may be used to limit aio requests available
for containers.
By default, newly allocated blkcg::blkg_aio_max_nr is set
to "unlimited" value (see
This adds new members of struct blkcg, which will
be used to account numbers of cgroup's aio requests.
Also, blkcg_root is used to store sysctl variables
aio_nr and aio_max_nr.
Signed-off-by: Kirill Tkhai
---
block/blk-cgroup.c |4
fs/aio.c
Hi,
this patch set introduces accounting aio_nr and aio_max_nr per blkio cgroup.
It may be used to limit number of aio requests, which are available for
a cgroup, and could be useful for containers.
The accounting is hierarchical, and aio contexts, allocated in child cgroup,
are accounted in
On Sun, 2017-12-03 at 00:31 +0800, Ming Lei wrote:
> Fixes: 0df21c86bdbf ("scsi: implement .get_budget and .put_budget for blk-mq")
It might be safer to revert commit 0df21c86bdbf instead of trying to fix all
issues introduced by that commit for kernel version v4.15 ...
Bart.
On 12/01/2017 10:49 PM, Ming Lei wrote:
Unfortunately v2 has been found to exhibit another problem, an Oops in
the systemd shutdown path for device-mapper LVM volumes at least, when
calling 'q->request_fn(q)' if 'q->request_fn' is NULL.
That issue can be fixed easily in V2 and we understand
On Sun, Dec 03, 2017 at 12:16:08PM -0800, Eric Biggers wrote:
> Looks like BLKTRACESETUP doesn't limit the '.buf_nr' parameter, allowing
> anyone
> who can open a block device to cause an extremely large kmalloc. Here's a
> simplified reproducer:
>
There are lots of places which allow people
On Fri, Dec 01, 2017 at 06:04:29PM +0100, Alban Browaeys wrote:
> I initially reported as https://bugzilla.kernel.org/show_bug.cgi?id=198
> 023 .
>
> I have now bisected this issue to commit a6a252e6491443c1c1 "blk-mq-
> sched: decide how to handle flush rq via RQF_FLUSH_SEQ".
>
> This is with
Commit a33801e8b473 ("block, bfq: move debug blkio stats behind
CONFIG_DEBUG_BLK_CGROUP") introduced two batches of confusing ifdefs:
one reported in [1], plus a similar one in another function. This
commit removes both batches, in the way suggested in [1].
[1]
On 2017/12/04 17:25, Eryu Guan wrote:
On Mon, Dec 04, 2017 at 05:15:17PM +0800, Xiao Yang wrote:
On 2017/12/04 16:29, Eryu Guan wrote:
On Wed, Nov 29, 2017 at 08:02:26PM +0800, Xiao Yang wrote:
If the entire block device is formatted with a filesystem and
mounted, running "blockdev
On Mon, Dec 04, 2017 at 09:18:05AM +0100, Dmitry Vyukov wrote:
> On Mon, Dec 4, 2017 at 9:14 AM, Dan Carpenter
> wrote:
> > On Sun, Dec 03, 2017 at 12:16:08PM -0800, Eric Biggers wrote:
> >> Looks like BLKTRACESETUP doesn't limit the '.buf_nr' parameter, allowing
> >>
On Mon, Dec 04, 2017 at 05:15:17PM +0800, Xiao Yang wrote:
> On 2017/12/04 16:29, Eryu Guan wrote:
> > On Wed, Nov 29, 2017 at 08:02:26PM +0800, Xiao Yang wrote:
> > > If the entire block device is formatted with a filesystem and
> > > mounted, running "blockdev --rereadpt" should fail and return
On 2017/12/04 16:29, Eryu Guan wrote:
On Wed, Nov 29, 2017 at 08:02:26PM +0800, Xiao Yang wrote:
If the entire block device is formatted with a filesystem and
mounted, running "blockdev --rereadpt" should fail and return
EBUSY instead of pass.
Signed-off-by: Xiao Yang
On Mon, Dec 04, 2017 at 09:44:55AM +0100, Johannes Thumshirn wrote:
> Ming Lei writes:
>
>
> > I am happy to do that, but recently I am very busy, so it may be done
> > a bit late by me.
> >
> > But anyone should reproduce the issue 100% with V4.15-rc kernel by just
> >
Ming Lei writes:
> I am happy to do that, but recently I am very busy, so it may be done
> a bit late by me.
>
> But anyone should reproduce the issue 100% with V4.15-rc kernel by just
> running the above script, not any specific hardware is required at all,
> so that means
On Mon, Dec 04, 2017 at 09:19:33AM +0100, Johannes Thumshirn wrote:
>
> Hi Ming,
>
> Ming Lei writes:
> > This issue can be triggered by the following script:
> >
> > #!/bin/sh
> > rmmod scsi_debug
> > modprobe scsi_debug max_queue=1
> >
> > DEVICE=`ls -d
>
On Wed, Nov 29, 2017 at 08:02:26PM +0800, Xiao Yang wrote:
> If the entire block device is formatted with a filesystem and
> mounted, running "blockdev --rereadpt" should fail and return
> EBUSY instead of pass.
>
> Signed-off-by: Xiao Yang
As we have blktests[1] now, I
Hi Ming,
Ming Lei writes:
> This issue can be triggered by the following script:
>
> #!/bin/sh
> rmmod scsi_debug
> modprobe scsi_debug max_queue=1
>
> DEVICE=`ls -d
> /sys/bus/pseudo/drivers/scsi_debug/adapter*/host*/target*/*/block/* | head -1
>
On Mon, Dec 4, 2017 at 9:14 AM, Dan Carpenter wrote:
> On Sun, Dec 03, 2017 at 12:16:08PM -0800, Eric Biggers wrote:
>> Looks like BLKTRACESETUP doesn't limit the '.buf_nr' parameter, allowing
>> anyone
>> who can open a block device to cause an extremely large kmalloc.
70 matches
Mail list logo