This is a second version of the patche - it changes the condition in
writecache_pop_from_freelist, to avoid wakeup storm if wc->writeback_size
is greater than wc->freelist_low_watermark-wc->freelist_high_watermark.
From: Mikulas Patocka
Fix a bug that low watermark was ignored and high
On Thu, 07 Jun 2018 17:14:01 -0400, Mike Snitzer said:
> Can you please share what you test is? We've gotten lots of reports
> with failure following wake_up but I don't have a canned test to trigger
Just a laptop with 16G of RAM, no clear reproducer - Chrome with a lot of
tabs, a mail reader,
On Thu, Jun 07 2018 at 4:40pm -0400,
valdis.kletni...@vt.edu wrote:
> I've hit this one twice today with pretty much the same traceback.
> The disk has 3 partitions - one for EFI, one for /boot, and then the rest of
> the disk is a cryptluks partition that contains a dozen or so LVM logical
>
I've hit this one twice today with pretty much the same traceback.
The disk has 3 partitions - one for EFI, one for /boot, and then the rest of
the disk is a cryptluks partition that contains a dozen or so LVM logical
volumes.
'git log -- drivers/md' didn't show any obvious suspects since
Fix a bug that low watermark was ignored and high watermark was used as
both low and high.
The patch also verifies that high watermark is greater or equal than low
watermark - if this condition is violated, the driver would consume too
much CPU.
Signed-off-by: Mikulas Patocka
---
On Thu, 7 Jun 2018, Dan Carpenter wrote:
> The option is in terms of msecs so we should be using msecs_to_jiffies()
> to convert it instead of the other way around.
>
> Fixes: 1be8d9c3da01 ("dm: add writecache target")
> Signed-off-by: Dan Carpenter
OK.
Reviewed-by: Mikulas Patocka
>
On Thu, Jun 07, 2018 at 09:24:00AM +0200, Martin Wilck wrote:
> On Wed, 2018-06-06 at 16:56 -0500, Benjamin Marzinski wrote:
> > On Wed, Jun 06, 2018 at 10:25:36PM +0200, Martin Wilck wrote:
> > >
> > > The patch looks good to me, but doesn't this mean that mpathpersist
> > > would now also
This is second version of this patch - it also removes the label
continue_locked, because it is no longer needed. If forgot to refresh the
patch before sending it, so I sent an olded version.
From: Mikulas Patocka
Subject: [patch 2/3 v2] dm-writecache: convert wait queue to wake_up_process
On Thursday 07 June 2018 02:53 AM, Jens Axboe wrote:
> On 6/6/18 1:35 AM, Ladvine D Almeida wrote:
>> On Friday 01 June 2018 09:13 AM, Christoph Hellwig wrote:
>>> On Mon, May 28, 2018 at 02:43:09PM +0100, Ladvine D Almeida wrote:
This patch introduces new variable under bio structure to
On Mon, Jun 04, 2018 at 08:46:28PM -0400, Mike Snitzer wrote:
> On Mon, Jun 04 2018 at 7:24pm -0400,
> Ross Zwisler wrote:
>
> > On Fri, Jun 01, 2018 at 06:04:43PM -0400, Mike Snitzer wrote:
> > > On Tue, May 29 2018 at 3:51pm -0400,
> > > Ross Zwisler wrote:
> > >
> > > > The
The option is in terms of msecs so we should be using msecs_to_jiffies()
to convert it instead of the other way around.
Fixes: 1be8d9c3da01 ("dm: add writecache target")
Signed-off-by: Dan Carpenter
diff --git a/drivers/md/dm-writecache.c b/drivers/md/dm-writecache.c
index
On Wed, 2018-06-06 at 16:56 -0500, Benjamin Marzinski wrote:
> On Wed, Jun 06, 2018 at 10:25:36PM +0200, Martin Wilck wrote:
> >
> > The patch looks good to me, but doesn't this mean that mpathpersist
> > would now also support persistent reservations with the ALL_TG_PT
> > bit
> > set (IIUC, the
12 matches
Mail list logo