From: Jagdish Gediya <jvgediya@linux.ibm.com>
To: "Aneesh Kumar K.V" <aneesh.kumar@linux.ibm.com>
Cc: linux-mm@kvack.org, linux-kernel@vger.kernel.org,
akpm@linux-foundation.org, baolin.wang@linux.alibaba.com,
dave.hansen@linux.intel.com, ying.huang@intel.com,
shy828301@gmail.com, weixugc@google.com, gthelen@google.com,
dan.j.williams@intel.com
Subject: Re: [PATCH v3 3/7] drivers/base/node: Add support to write node_states[] via sysfs
Date: Mon, 25 Apr 2022 15:12:44 +0530 [thread overview]
Message-ID: <YmZtFDMJO5Kmt/0u@li-6e1fa1cc-351b-11b2-a85c-b897023bb5f3.ibm.com> (raw)
In-Reply-To: <87o80rgfpy.fsf@linux.ibm.com>
On Sun, Apr 24, 2022 at 11:55:45AM +0530, Aneesh Kumar K.V wrote:
> Jagdish Gediya <jvgediya@linux.ibm.com> writes:
>
> > Current /sys/devices/system/node/* interface doesn't support
> > to write node_states[], however write support is needed in case
> > users want to set them manually e.g. when user want to override
> > default N_DEMOTION_TARGETS found by the kernel.
> >
> > Rename existing _NODE_ATTR to _NODE_ATTR_RO and introduce new
> > _NODE_ATTR_RW which can be used for node_states[] which can
> > be written from sysfs.
> >
> > It may be necessary to validate written values and take action
> > based on them in a state specific way so a callback 'write' is
> > introduced in 'struct node_attr'.
> >
> > A new function demotion_targets_write() is added to validate
> > the input nodes for N_DEMOTION_TARGETS which should be subset
> > of N_MEMORY and to build new demotion list based on new nodes.
> >
> > Signed-off-by: Jagdish Gediya <jvgediya@linux.ibm.com>
> > ---
> > drivers/base/node.c | 62 +++++++++++++++++++++++++++++++++++----------
> > 1 file changed, 49 insertions(+), 13 deletions(-)
> >
> > diff --git a/drivers/base/node.c b/drivers/base/node.c
> > index 6eef22e6413e..e03eedbc421b 100644
> > --- a/drivers/base/node.c
> > +++ b/drivers/base/node.c
> > @@ -20,6 +20,7 @@
> > #include <linux/pm_runtime.h>
> > #include <linux/swap.h>
> > #include <linux/slab.h>
> > +#include <linux/migrate.h>
> >
> > static struct bus_type node_subsys = {
> > .name = "node",
> > @@ -1013,6 +1014,7 @@ void unregister_one_node(int nid)
> > struct node_attr {
> > struct device_attribute attr;
> > enum node_states state;
> > + int (*write)(nodemask_t nodes);
> > };
> >
> > static ssize_t show_node_state(struct device *dev,
> > @@ -1024,23 +1026,57 @@ static ssize_t show_node_state(struct device *dev,
> > nodemask_pr_args(&node_states[na->state]));
> > }
> >
> > -#define _NODE_ATTR(name, state) \
> > - { __ATTR(name, 0444, show_node_state, NULL), state }
> > +static ssize_t store_node_state(struct device *s,
> > + struct device_attribute *attr,
> > + const char *buf, size_t count)
> > +{
> > + nodemask_t nodes;
> > + struct node_attr *na = container_of(attr, struct node_attr, attr);
> > +
> > + if (nodelist_parse(buf, nodes))
> > + return -EINVAL;
> > +
> > + if (na->write) {
> > + if (na->write(nodes))
> > + return -EINVAL;
> > + } else {
> > + node_states[na->state] = nodes;
> > + }
> > +
> > + return count;
> > +}
> > +
> > +static int demotion_targets_write(nodemask_t nodes)
> > +{
> > + if (nodes_subset(nodes, node_states[N_MEMORY])) {
> > + node_states[N_DEMOTION_TARGETS] = nodes;
> > + set_migration_target_nodes();
> > + return 0;
> > + }
>
> Does this require locking to avoid upating
> node_states[N_DEMOTION_TARGETS] while a parallel
> set_migratiotn_target_nodes() is running.
I think locking is needed if set_migratiotn_target_nodes() is called
form here because currently exclusion is provided by memory hotplug
events being single-threaded, but as this path is not hot-plug event,
separate lock is needed.
> > +
> > + return -EINVAL;
> > +}
> > +
> > +#define _NODE_ATTR_RO(name, state) \
> > + { __ATTR(name, 0444, show_node_state, NULL), state, NULL }
> > +
> > +#define _NODE_ATTR_RW(name, state, write_fn) \
> > + { __ATTR(name, 0644, show_node_state, store_node_state), state, write_fn }
> >
> > static struct node_attr node_state_attr[] = {
> > - [N_POSSIBLE] = _NODE_ATTR(possible, N_POSSIBLE),
> > - [N_ONLINE] = _NODE_ATTR(online, N_ONLINE),
> > - [N_NORMAL_MEMORY] = _NODE_ATTR(has_normal_memory, N_NORMAL_MEMORY),
> > + [N_POSSIBLE] = _NODE_ATTR_RO(possible, N_POSSIBLE),
> > + [N_ONLINE] = _NODE_ATTR_RO(online, N_ONLINE),
> > + [N_NORMAL_MEMORY] = _NODE_ATTR_RO(has_normal_memory, N_NORMAL_MEMORY),
> > #ifdef CONFIG_HIGHMEM
> > - [N_HIGH_MEMORY] = _NODE_ATTR(has_high_memory, N_HIGH_MEMORY),
> > + [N_HIGH_MEMORY] = _NODE_ATTR_RO(has_high_memory, N_HIGH_MEMORY),
> > #endif
> > - [N_MEMORY] = _NODE_ATTR(has_memory, N_MEMORY),
> > - [N_CPU] = _NODE_ATTR(has_cpu, N_CPU),
> > - [N_GENERIC_INITIATOR] = _NODE_ATTR(has_generic_initiator,
> > - N_GENERIC_INITIATOR),
> > - [N_DEMOTION_TARGETS] = _NODE_ATTR(demotion_targets,
> > - N_DEMOTION_TARGETS),
> > -
> > + [N_MEMORY] = _NODE_ATTR_RO(has_memory, N_MEMORY),
> > + [N_CPU] = _NODE_ATTR_RO(has_cpu, N_CPU),
> > + [N_GENERIC_INITIATOR] = _NODE_ATTR_RO(has_generic_initiator,
> > + N_GENERIC_INITIATOR),
> > + [N_DEMOTION_TARGETS] = _NODE_ATTR_RW(demotion_targets,
> > + N_DEMOTION_TARGETS,
> > + demotion_targets_write),
> > };
> >
> > static struct attribute *node_state_attrs[] = {
> > --
> > 2.35.1
next prev parent reply other threads:[~2022-04-25 9:43 UTC|newest]
Thread overview: 35+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-04-22 19:55 [PATCH v3 0/7] mm: demotion: Introduce new node state N_DEMOTION_TARGETS Jagdish Gediya
2022-04-22 19:55 ` [PATCH v3 1/7] mm: demotion: Fix demotion targets sharing among sources Jagdish Gediya
2022-04-24 3:25 ` ying.huang
2022-04-25 9:32 ` Jagdish Gediya
2022-04-26 7:26 ` ying.huang
2022-04-22 19:55 ` [PATCH v3 2/7] mm: demotion: Add new node state N_DEMOTION_TARGETS Jagdish Gediya
2022-04-22 20:29 ` Wei Xu
2022-04-22 19:55 ` [PATCH v3 3/7] drivers/base/node: Add support to write node_states[] via sysfs Jagdish Gediya
2022-04-22 20:32 ` Wei Xu
2022-04-24 6:25 ` Aneesh Kumar K.V
2022-04-25 9:42 ` Jagdish Gediya [this message]
2022-04-24 6:29 ` ying.huang
2022-04-22 19:55 ` [PATCH v3 4/7] device-dax/kmem: Set node state as N_DEMOTION_TARGETS Jagdish Gediya
2022-04-22 20:34 ` Wei Xu
2022-04-22 19:55 ` [PATCH v3 5/7] mm: demotion: Build demotion list based on N_DEMOTION_TARGETS Jagdish Gediya
2022-04-22 20:39 ` Wei Xu
2022-04-22 19:55 ` [PATCH v3 6/7] mm: demotion: expose per-node demotion targets via sysfs Jagdish Gediya
2022-04-22 20:47 ` Wei Xu
2022-04-23 7:30 ` kernel test robot
2022-04-23 8:38 ` kernel test robot
2022-04-22 19:55 ` [PATCH v3 7/7] docs: numa: Add documentation for demotion Jagdish Gediya
2022-04-24 3:19 ` [PATCH v3 0/7] mm: demotion: Introduce new node state N_DEMOTION_TARGETS ying.huang
2022-04-25 11:15 ` Jagdish Gediya
2022-04-25 13:57 ` Jonathan Cameron
2022-04-25 14:44 ` Aneesh Kumar K V
2022-04-26 10:43 ` Jonathan Cameron
2022-04-27 1:29 ` ying.huang
2022-04-27 2:57 ` Aneesh Kumar K V
2022-04-27 3:34 ` ying.huang
2022-04-25 14:53 ` Aneesh Kumar K V
2022-04-26 10:37 ` Jonathan Cameron
2022-04-26 7:55 ` ying.huang
2022-04-26 9:07 ` Aneesh Kumar K V
2022-04-26 9:10 ` ying.huang
2022-04-26 9:37 ` Jagdish Gediya
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=YmZtFDMJO5Kmt/0u@li-6e1fa1cc-351b-11b2-a85c-b897023bb5f3.ibm.com \
--to=jvgediya@linux.ibm.com \
--cc=akpm@linux-foundation.org \
--cc=aneesh.kumar@linux.ibm.com \
--cc=baolin.wang@linux.alibaba.com \
--cc=dan.j.williams@intel.com \
--cc=dave.hansen@linux.intel.com \
--cc=gthelen@google.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=shy828301@gmail.com \
--cc=weixugc@google.com \
--cc=ying.huang@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).