On Fri, Mar 24, 2017 at 04:04:32PM -0600, Jens Axboe wrote:
> On 03/24/2017 03:56 PM, Tahsin Erdogan wrote:
> > blkg_conf_prep() currently calls blkg_lookup_create() while holding
> > request queue spinlock. This means allocating memory for struct
> > blkcg_gq has to be made non-blocking. This
On Fri, Mar 24, 2017 at 04:04:32PM -0600, Jens Axboe wrote:
> On 03/24/2017 03:56 PM, Tahsin Erdogan wrote:
> > blkg_conf_prep() currently calls blkg_lookup_create() while holding
> > request queue spinlock. This means allocating memory for struct
> > blkcg_gq has to be made non-blocking. This
On Sun, Mar 26, 2017 at 3:54 AM, Julia Lawall wrote:
> Is an unlock needed before line 848? Or does blkg_lookup_check take care
> of it?
Unlock is not needed. On success, function returns with locks held.
It is documented at line 805:
"... This function returns with RCU
On Sun, Mar 26, 2017 at 3:54 AM, Julia Lawall wrote:
> Is an unlock needed before line 848? Or does blkg_lookup_check take care
> of it?
Unlock is not needed. On success, function returns with locks held.
It is documented at line 805:
"... This function returns with RCU read
* lock and queue
: Re: [PATCH] blkcg: allocate struct blkcg_gq outside request queue
spinlock
CC: kbuild-...@01.org
In-Reply-To: <20170324215627.12831-1-tah...@google.com>
TO: Tahsin Erdogan <tah...@google.com>
CC: Tejun Heo <t...@kernel.org>, Jens Axboe <ax...@kernel.dk>,
linux-bl...@vg
Is an unlock needed before line 848? Or does blkg_lookup_check take care
of it?
julia
-- Forwarded message --
Date: Sun, 26 Mar 2017 07:50:17 +0800
From: kbuild test robot
To: kbu...@01.org
Cc: Julia Lawall
Subject: Re: [PATCH] blkcg: allocate struct blkcg_gq outside request
On 03/24/2017 03:56 PM, Tahsin Erdogan wrote:
> blkg_conf_prep() currently calls blkg_lookup_create() while holding
> request queue spinlock. This means allocating memory for struct
> blkcg_gq has to be made non-blocking. This causes occasional -ENOMEM
> failures in call paths like below:
>
>
On 03/24/2017 03:56 PM, Tahsin Erdogan wrote:
> blkg_conf_prep() currently calls blkg_lookup_create() while holding
> request queue spinlock. This means allocating memory for struct
> blkcg_gq has to be made non-blocking. This causes occasional -ENOMEM
> failures in call paths like below:
>
>
blkg_conf_prep() currently calls blkg_lookup_create() while holding
request queue spinlock. This means allocating memory for struct
blkcg_gq has to be made non-blocking. This causes occasional -ENOMEM
failures in call paths like below:
pcpu_alloc+0x68f/0x710
__alloc_percpu_gfp+0xd/0x10
blkg_conf_prep() currently calls blkg_lookup_create() while holding
request queue spinlock. This means allocating memory for struct
blkcg_gq has to be made non-blocking. This causes occasional -ENOMEM
failures in call paths like below:
pcpu_alloc+0x68f/0x710
__alloc_percpu_gfp+0xd/0x10
Hello,
On Tue, Feb 28, 2017 at 03:51:27PM -0800, Tahsin Erdogan wrote:
> On Tue, Feb 28, 2017 at 2:47 PM, Tejun Heo wrote:
> >> + if (!blkcg_policy_enabled(q, pol)) {
> >> + ret = -EOPNOTSUPP;
> >> + goto fail;
> >
> > Pulling this out of the
Hello,
On Tue, Feb 28, 2017 at 03:51:27PM -0800, Tahsin Erdogan wrote:
> On Tue, Feb 28, 2017 at 2:47 PM, Tejun Heo wrote:
> >> + if (!blkcg_policy_enabled(q, pol)) {
> >> + ret = -EOPNOTSUPP;
> >> + goto fail;
> >
> > Pulling this out of the queue_lock doesn't seem
On Tue, Feb 28, 2017 at 2:47 PM, Tejun Heo wrote:
>> + if (!blkcg_policy_enabled(q, pol)) {
>> + ret = -EOPNOTSUPP;
>> + goto fail;
>
> Pulling this out of the queue_lock doesn't seem safe to me. This
> function may end up calling into callbacks of
On Tue, Feb 28, 2017 at 2:47 PM, Tejun Heo wrote:
>> + if (!blkcg_policy_enabled(q, pol)) {
>> + ret = -EOPNOTSUPP;
>> + goto fail;
>
> Pulling this out of the queue_lock doesn't seem safe to me. This
> function may end up calling into callbacks of disabled policies
Hello,
Overall, the approach looks good to me but please see below.
On Mon, Feb 27, 2017 at 06:49:57PM -0800, Tahsin Erdogan wrote:
> @@ -806,44 +807,99 @@ int blkg_conf_prep(struct blkcg *blkcg, const struct
> blkcg_policy *pol,
> if (!disk)
> return -ENODEV;
> if
Hello,
Overall, the approach looks good to me but please see below.
On Mon, Feb 27, 2017 at 06:49:57PM -0800, Tahsin Erdogan wrote:
> @@ -806,44 +807,99 @@ int blkg_conf_prep(struct blkcg *blkcg, const struct
> blkcg_policy *pol,
> if (!disk)
> return -ENODEV;
> if
blkg_conf_prep() currently calls blkg_lookup_create() while holding
request queue spinlock. This means allocating memory for struct
blkcg_gq has to be made non-blocking. This causes occasional -ENOMEM
failures in call paths like below:
pcpu_alloc+0x68f/0x710
__alloc_percpu_gfp+0xd/0x10
blkg_conf_prep() currently calls blkg_lookup_create() while holding
request queue spinlock. This means allocating memory for struct
blkcg_gq has to be made non-blocking. This causes occasional -ENOMEM
failures in call paths like below:
pcpu_alloc+0x68f/0x710
__alloc_percpu_gfp+0xd/0x10
18 matches
Mail list logo