From: Tang Junhui
In btree_flush_write(), two places need to take a locker to
avoid races:
Firstly, we need take rcu read locker to protect the bucket_hash
traverse, since hlist_for_each_entry_rcu() must be called under
the protection of rcu read locker.
Secondly, we
On Wed, Jan 24, 2018 at 11:49:26AM +0800, Ming Lei wrote:
> On Tue, Jan 23, 2018 at 02:32:06PM -0700, Jens Axboe wrote:
> > On 1/23/18 10:20 AM, Eryu Guan wrote:
> > > Attributes that only implement .seq_ops are read-only, any write to
> > > them should be rejected. But currently kernel would
On Tue, Jan 23, 2018 at 02:32:06PM -0700, Jens Axboe wrote:
> On 1/23/18 10:20 AM, Eryu Guan wrote:
> > Attributes that only implement .seq_ops are read-only, any write to
> > them should be rejected. But currently kernel would crash when
> > writing to such debugfs entries, e.g.
> >
> > chmod +w
On Tue, Jan 23, 2018 at 04:57:34PM +, Bart Van Assche wrote:
> On Wed, 2018-01-24 at 00:37 +0800, Ming Lei wrote:
> > On Tue, Jan 23, 2018 at 04:24:20PM +, Bart Van Assche wrote:
> > > My opinion about this patch is as follows:
> > > * Changing a blk_mq_delay_run_hw_queue() call followed
On Tue, 2018-01-23 at 23:12 +, Bart Van Assche wrote:
> On Tue, 2018-01-23 at 04:45 -0800, Joe Perches wrote:
> > Perhaps the email subject could be improved to describe
> > the new macro and as well, this macro, without a pr_fmt
> > define somewhat above it loses the __func__ output too.
>
>
On Tue, Jan 23, 2018 at 10:01:37PM +, Bart Van Assche wrote:
> On Wed, 2018-01-24 at 00:59 +0800, Ming Lei wrote:
> > How is that enough to fix the IO hang when driver returns STS_RESOURCE
> > and the queue is idle? If you want to follow previous dm-rq's way of
> > call
On Jan 22, 2018, at 8:28 PM, Jens Axboe wrote:
>
> On 1/22/18 8:18 PM, Goldwyn Rodrigues wrote:
>>> that their application was "already broken". I'd hate for a kernel
>>> upgrade to break them.
>>>
>>> I do wish we could make the change, and maybe we can. But it probably
>>>
On Tue, 2018-01-23 at 04:45 -0800, Joe Perches wrote:
> Perhaps the email subject could be improved to describe
> the new macro and as well, this macro, without a pr_fmt
> define somewhat above it loses the __func__ output too.
Hmm ... I thought that the pr_debug() output can be configured to
On Wed, 2018-01-24 at 00:59 +0800, Ming Lei wrote:
> How is that enough to fix the IO hang when driver returns STS_RESOURCE
> and the queue is idle? If you want to follow previous dm-rq's way of
> call blk_mq_delay_run_hw_queue() in .queue_rq(), the same trick need
> to be applied to other drivers
On 1/23/18 10:20 AM, Eryu Guan wrote:
> Attributes that only implement .seq_ops are read-only, any write to
> them should be rejected. But currently kernel would crash when
> writing to such debugfs entries, e.g.
>
> chmod +w /sys/kernel/debug/block//requeue_list
> echo 0 >
> On Jan 23, 2018, at 13:00, Jens Axboe wrote:
>
> On 1/23/18 10:54 AM, David Zarzycki wrote:
>> Hi Jens,
>>
>> The bug still reproduces with this change. How confident are we that
>> kernel objects are properly reference counted while they are
>> throttled?
>
> I would be
On 1/23/18 10:54 AM, David Zarzycki wrote:
> Hi Jens,
>
> The bug still reproduces with this change. How confident are we that
> kernel objects are properly reference counted while they are
> throttled?
I would be surprised if it made a change, thanks for checking. Since
you're not pulling
Hi Jens,
The bug still reproduces with this change. How confident are we that kernel
objects are properly reference counted while they are throttled?
Dave
> On Jan 23, 2018, at 10:34, Jens Axboe wrote:
>
> On 1/23/18 6:48 AM, David Zarzycki wrote:
>>
>>
>>> On Jan 22,
Attributes that only implement .seq_ops are read-only, any write to
them should be rejected. But currently kernel would crash when
writing to such debugfs entries, e.g.
chmod +w /sys/kernel/debug/block//requeue_list
echo 0 > /sys/kernel/debug/block//requeue_list
chmod -w
On Tue, Jan 23, 2018 at 04:54:02PM +, Bart Van Assche wrote:
> On Wed, 2018-01-24 at 00:49 +0800, Ming Lei wrote:
> > On Tue, Jan 23, 2018 at 04:47:11PM +, Bart Van Assche wrote:
> > > On Wed, 2018-01-24 at 00:41 +0800, Ming Lei wrote:
> > > > Could you explain where to call call_rcu()?
On Wed, 2018-01-24 at 00:37 +0800, Ming Lei wrote:
> On Tue, Jan 23, 2018 at 04:24:20PM +, Bart Van Assche wrote:
> > My opinion about this patch is as follows:
> > * Changing a blk_mq_delay_run_hw_queue() call followed by return
> > BLK_STS_DEV_RESOURCE into return BLK_STS_RESOURCE is wrong
On Wed, 2018-01-24 at 00:49 +0800, Ming Lei wrote:
> On Tue, Jan 23, 2018 at 04:47:11PM +, Bart Van Assche wrote:
> > On Wed, 2018-01-24 at 00:41 +0800, Ming Lei wrote:
> > > Could you explain where to call call_rcu()? call_rcu() can't be used in
> > > IO path at all.
> >
> > Can you explain
On Tue, Jan 23, 2018 at 04:47:11PM +, Bart Van Assche wrote:
> On Wed, 2018-01-24 at 00:41 +0800, Ming Lei wrote:
> > Could you explain where to call call_rcu()? call_rcu() can't be used in
> > IO path at all.
>
> Can you explain what makes you think that call_rcu() can't be used in the
>
On Wed, 2018-01-24 at 00:41 +0800, Ming Lei wrote:
> Could you explain where to call call_rcu()? call_rcu() can't be used in
> IO path at all.
Can you explain what makes you think that call_rcu() can't be used in the
I/O path? As you know call_rcu() invokes a function asynchronously. From
On Tue, 2018-01-23 at 10:22 +0100, Mike Snitzer wrote:
> On Thu, Jan 18 2018 at 5:20pm -0500,
> Bart Van Assche wrote:
>
> > On Thu, 2018-01-18 at 17:01 -0500, Mike Snitzer wrote:
> > > And yet Laurence cannot reproduce any such lockups with your test...
> >
> > Hmm ...
On Tue, Jan 23, 2018 at 08:37:26AM -0800, Bart Van Assche wrote:
> On 01/23/18 08:26, Ming Lei wrote:
> > On Tue, Jan 23, 2018 at 08:17:02AM -0800, Bart Van Assche wrote:
> > > On 01/22/18 16:57, Ming Lei wrote:
> > > > Even though RCU lock is held during dispatch, preemption or interrupt
> > > >
On Tue, Jan 23, 2018 at 04:24:20PM +, Bart Van Assche wrote:
> On Wed, 2018-01-24 at 00:16 +0800, Ming Lei wrote:
> > @@ -1280,10 +1282,18 @@ bool blk_mq_dispatch_rq_list(struct request_queue
> > *q, struct list_head *list,
> > * - Some but not all block drivers stop a queue
On 01/23/18 08:26, Ming Lei wrote:
On Tue, Jan 23, 2018 at 08:17:02AM -0800, Bart Van Assche wrote:
On 01/22/18 16:57, Ming Lei wrote:
Even though RCU lock is held during dispatch, preemption or interrupt
can happen too, so it is simply wrong to depend on the timing to make
sure
On Tue, Jan 23, 2018 at 08:17:02AM -0800, Bart Van Assche wrote:
>
>
> On 01/22/18 16:57, Ming Lei wrote:
> > Even though RCU lock is held during dispatch, preemption or interrupt
> > can happen too, so it is simply wrong to depend on the timing to make
> > sure __blk_mq_run_hw_queue() can see
On Wed, 2018-01-24 at 00:16 +0800, Ming Lei wrote:
> @@ -1280,10 +1282,18 @@ bool blk_mq_dispatch_rq_list(struct request_queue *q,
> struct list_head *list,
>* - Some but not all block drivers stop a queue before
>* returning BLK_STS_RESOURCE. Two exceptions are
On Tue, Jan 23 2018 at 11:16am -0500,
Ming Lei wrote:
> This status is returned from driver to block layer if device related
> resource is run out of, but driver can guarantee that IO dispatch will
> be triggered in future when the resource is available.
>
> This patch
This status is returned from driver to block layer if device related
resource is run out of, but driver can guarantee that IO dispatch will
be triggered in future when the resource is available.
This patch converts some drivers to use this return value. Meantime
if driver returns BLK_STS_RESOURCE
On 1/23/18 12:00 AM, NeilBrown wrote:
> On Mon, Jan 15 2018, Wang Sheng-Hui wrote:
>
>> Dear,
>>
>> Sorry to trouble you.
>>
>> I'm confused by the following code in bio.c/bio_split:
>> -
>> struct bio *bio_split(struct bio *bio, int sectors,
>> ...
>> if
On 1/23/18 6:48 AM, David Zarzycki wrote:
>
>
>> On Jan 22, 2018, at 20:20, Jens Axboe wrote:
>>
>> All of these are off the blk-wbt completion path. I suggested earlier to
>> try and disable CONFIG_BLK_WBT to see if it goes away, or at least to
>> see if the pattern changes.
>
On Tue, 2018-01-23 at 09:44 -0500, Mike Snitzer wrote:
> On Tue, Jan 23 2018 at 9:27am -0500,
> Ming Lei wrote:
>
> > Hello Martin,
> >
> > On Tue, Jan 23, 2018 at 08:30:41AM -0500, Martin K. Petersen wrote:
> > >
> > > Ming,
> > >
> > > > + * Block layer and block
On Tue, Jan 23 2018 at 9:27am -0500,
Ming Lei wrote:
> Hello Martin,
>
> On Tue, Jan 23, 2018 at 08:30:41AM -0500, Martin K. Petersen wrote:
> >
> > Ming,
> >
> > > + * Block layer and block driver specific status, which is ususally
> > > returnd
> >
Hello Martin,
On Tue, Jan 23, 2018 at 08:30:41AM -0500, Martin K. Petersen wrote:
>
> Ming,
>
> > + * Block layer and block driver specific status, which is ususally returnd
> ^^^
> > + * from driver to block layer in IO
On Tue, Jan 23, 2018 at 08:30:41AM -0500, Martin K. Petersen wrote:
>
> Ming,
>
> > + * Block layer and block driver specific status, which is ususally returnd
> ^^^
> > + * from driver to block layer in IO path.
>
>
> On Jan 22, 2018, at 20:20, Jens Axboe wrote:
>
> All of these are off the blk-wbt completion path. I suggested earlier to
> try and disable CONFIG_BLK_WBT to see if it goes away, or at least to
> see if the pattern changes.
Hi Jens,
Bingo! Disabling CONFIG_BLK_WBT makes
Ming,
> + * Block layer and block driver specific status, which is ususally returnd
^^^
> + * from driver to block layer in IO path.
Given that the comment blurb is long and the flag not defined until
later, it is not
The trailing semicolon is an empty statement that does no operation.
Removing it since it doesn't do anything.
Signed-off-by: Luis de Bethencourt
---
Hi,
Similar to the other drivers/clocksource/ patch [0].
After fixing the same thing in drivers/staging/rtl8723bs/, Joe
On Tue, Jan 23 2018 at 7:17am -0500,
Ming Lei wrote:
> On Tue, Jan 23, 2018 at 8:15 PM, Mike Snitzer wrote:
> > On Tue, Jan 23 2018 at 5:53am -0500,
> > Ming Lei wrote:
> >
> >> Hi Mike,
> >>
> >> On Tue, Jan 23, 2018 at
On Tue, Jan 23, 2018 at 8:15 PM, Mike Snitzer wrote:
> On Tue, Jan 23 2018 at 5:53am -0500,
> Ming Lei wrote:
>
>> Hi Mike,
>>
>> On Tue, Jan 23, 2018 at 10:22:04AM +0100, Mike Snitzer wrote:
>> > On Thu, Jan 18 2018 at 5:20pm -0500,
>> > Bart Van
On Tue, Jan 23 2018 at 5:53am -0500,
Ming Lei wrote:
> Hi Mike,
>
> On Tue, Jan 23, 2018 at 10:22:04AM +0100, Mike Snitzer wrote:
> > On Thu, Jan 18 2018 at 5:20pm -0500,
> > Bart Van Assche wrote:
> >
> > > On Thu, 2018-01-18 at 17:01 -0500,
Use pr_debug instead of hand craftet macros. This way it is not needed to
re-compile the kernel to enable bsg debug outputs and it's possible to
selectively enable specific prints.
Signed-off-by: Johannes Thumshirn
Cc: Joe Perches
Cc: Bart Van Assche
This status is returned from driver to block layer if device related
resource is run out of, but driver can guarantee that IO dispatch will
be triggered in future when the resource is available.
This patch converts some drivers to use this return value. Meantime
if driver returns BLK_STS_RESOURCE
Hi Mike,
On Tue, Jan 23, 2018 at 10:22:04AM +0100, Mike Snitzer wrote:
> On Thu, Jan 18 2018 at 5:20pm -0500,
> Bart Van Assche wrote:
>
> > On Thu, 2018-01-18 at 17:01 -0500, Mike Snitzer wrote:
> > > And yet Laurence cannot reproduce any such lockups with your test...
On Thu, Jan 18 2018 at 5:20pm -0500,
Bart Van Assche wrote:
> On Thu, 2018-01-18 at 17:01 -0500, Mike Snitzer wrote:
> > And yet Laurence cannot reproduce any such lockups with your test...
>
> Hmm ... maybe I misunderstood Laurence but I don't think that Laurence has
>
On Tue, 23 Jan 2018, Aishwarya Pant wrote:
> Clean up the sysfs documentation such that it is in the same format as
> described in Documentation/ABI/README. Mainly, the patch moves the
> attribute names to the 'What:' field. This might be useful for scripting
> and tracking changes in the ABI.
Clean up the sysfs documentation such that it is in the same format as
described in Documentation/ABI/README. Mainly, the patch moves the
attribute names to the 'What:' field. This might be useful for scripting
and tracking changes in the ABI.
Signed-off-by: Aishwarya Pant
45 matches
Mail list logo