[PATCH] x86/resctrl: Add task resctrl information display

[PATCH] x86/resctrl: Add task resctrl information display

[Chen Yu]

Monitoring tools that want to find out which resctrl CTRL
and MONITOR groups a task belongs to must currently read
the "tasks" file in every group until they locate the process
ID.

Add an additional file /proc/{pid}/resctrl to provide this
information.

For example:
 cat /proc/1193/resctrl
CTRL_MON:/ctrl_grp0
MON:/ctrl_grp0/mon_groups/mon_grp0

If the resctrl filesystem has not been mounted,
reading /proc/{pid}/resctrl returns an error:
cat: /proc/1193/resctrl: No such device

Tested-by: Jinshi Chen <jinshi.chen@intel.com>
Reviewed-by: Reinette Chatre <reinette.chatre@intel.com>
Reviewed-by: Fenghua Yu <fenghua.yu@intel.com>
Reviewed-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Chen Yu <yu.c.chen@intel.com>

---
 arch/x86/include/asm/resctrl_sched.h   |  4 +++
 arch/x86/kernel/cpu/resctrl/rdtgroup.c | 46 ++++++++++++++++++++++++++
 fs/proc/base.c                         |  9 +++++
 3 files changed, 59 insertions(+)

diff --git a/arch/x86/include/asm/resctrl_sched.h b/arch/x86/include/asm/resctrl_sched.h
index f6b7fe2833cc..bba362e0e00f 100644
--- a/arch/x86/include/asm/resctrl_sched.h
+++ b/arch/x86/include/asm/resctrl_sched.h
@@ -5,6 +5,7 @@
 #ifdef CONFIG_X86_CPU_RESCTRL
 
 #include <linux/sched.h>
+#include <linux/proc_fs.h>
 #include <linux/jump_label.h>
 
 #define IA32_PQR_ASSOC	0x0c8f
@@ -84,6 +85,9 @@ static inline void resctrl_sched_in(void)
 		__resctrl_sched_in();
 }
 
+int proc_resctrl_show(struct seq_file *m, struct pid_namespace *ns,
+		      struct pid *pid, struct task_struct *tsk);
+
 #else
 
 static inline void resctrl_sched_in(void) {}
diff --git a/arch/x86/kernel/cpu/resctrl/rdtgroup.c b/arch/x86/kernel/cpu/resctrl/rdtgroup.c
index a46dee8e78db..2317174174e9 100644
--- a/arch/x86/kernel/cpu/resctrl/rdtgroup.c
+++ b/arch/x86/kernel/cpu/resctrl/rdtgroup.c
@@ -727,6 +727,52 @@ static int rdtgroup_tasks_show(struct kernfs_open_file *of,
 	return ret;
 }
 
+int proc_resctrl_show(struct seq_file *s, struct pid_namespace *ns,
+		      struct pid *pid, struct task_struct *tsk)
+{
+	struct rdtgroup *rdtg;
+	int ret = 0;
+
+	mutex_lock(&rdtgroup_mutex);
+
+	/* Make sure resctrl has been mounted. */
+	if (!static_branch_unlikely(&rdt_enable_key)) {
+		ret = -ENODEV;
+		goto unlock;
+	}
+
+	list_for_each_entry(rdtg, &rdt_all_groups, rdtgroup_list) {
+		struct rdtgroup *crg;
+
+		/*
+		 * Task information is only relevant for shareable
+		 * and exclusive groups.
+		 */
+		if (rdtg->mode != RDT_MODE_SHAREABLE &&
+		    rdtg->mode != RDT_MODE_EXCLUSIVE)
+			continue;
+
+		if (rdtg->closid == tsk->closid) {
+			seq_printf(s, "CTRL_MON:/%s\n", rdtg->kn->name);
+			list_for_each_entry(crg, &rdtg->mon.crdtgrp_list,
+					    mon.crdtgrp_list) {
+				if (tsk->rmid != crg->mon.rmid)
+					continue;
+				seq_printf(s, "MON:%s%s/mon_groups/%s\n",
+					   rdtg == &rdtgroup_default ? "" : "/",
+					   rdtg->kn->name, crg->kn->name);
+				goto unlock;
+			}
+			goto unlock;
+		}
+	}
+	ret = -ENOENT;
+unlock:
+	mutex_unlock(&rdtgroup_mutex);
+
+	return ret;
+}
+
 static int rdt_last_cmd_status_show(struct kernfs_open_file *of,
 				    struct seq_file *seq, void *v)
 {
diff --git a/fs/proc/base.c b/fs/proc/base.c
index ebea9501afb8..d8a61db78db5 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
@@ -95,6 +95,9 @@
 #include <linux/sched/stat.h>
 #include <linux/posix-timers.h>
 #include <trace/events/oom.h>
+#ifdef CONFIG_X86_CPU_RESCTRL
+#include <asm/resctrl_sched.h>
+#endif
 #include "internal.h"
 #include "fd.h"
 
@@ -3060,6 +3063,9 @@ static const struct pid_entry tgid_base_stuff[] = {
 #endif
 #ifdef CONFIG_CGROUPS
 	ONE("cgroup",  S_IRUGO, proc_cgroup_show),
+#endif
+#ifdef CONFIG_X86_CPU_RESCTRL
+	ONE("resctrl", S_IRUGO, proc_resctrl_show),
 #endif
 	ONE("oom_score",  S_IRUGO, proc_oom_score),
 	REG("oom_adj",    S_IRUGO|S_IWUSR, proc_oom_adj_operations),
@@ -3460,6 +3466,9 @@ static const struct pid_entry tid_base_stuff[] = {
 #endif
 #ifdef CONFIG_CGROUPS
 	ONE("cgroup",  S_IRUGO, proc_cgroup_show),
+#endif
+#ifdef CONFIG_X86_CPU_RESCTRL
+	ONE("resctrl", S_IRUGO, proc_resctrl_show),
 #endif
 	ONE("oom_score", S_IRUGO, proc_oom_score),
 	REG("oom_adj",   S_IRUGO|S_IWUSR, proc_oom_adj_operations),
-- 
2.17.1

Re: [PATCH] x86/resctrl: Add task resctrl information display

[Borislav Petkov]

On Thu, Nov 07, 2019 at 11:27:31AM +0800, Chen Yu wrote:
> Monitoring tools that want to find out which resctrl CTRL
> and MONITOR groups a task belongs to must currently read
> the "tasks" file in every group until they locate the process
> ID.
> 
> Add an additional file /proc/{pid}/resctrl to provide this
> information.
> 
> For example:
>  cat /proc/1193/resctrl
> CTRL_MON:/ctrl_grp0
> MON:/ctrl_grp0/mon_groups/mon_grp0
> 
> If the resctrl filesystem has not been mounted,
> reading /proc/{pid}/resctrl returns an error:
> cat: /proc/1193/resctrl: No such device
> 
> Tested-by: Jinshi Chen <jinshi.chen@intel.com>
> Reviewed-by: Reinette Chatre <reinette.chatre@intel.com>
> Reviewed-by: Fenghua Yu <fenghua.yu@intel.com>
> Reviewed-by: Tony Luck <tony.luck@intel.com>
> Signed-off-by: Chen Yu <yu.c.chen@intel.com>
> 
> ---
>  arch/x86/include/asm/resctrl_sched.h   |  4 +++
>  arch/x86/kernel/cpu/resctrl/rdtgroup.c | 46 ++++++++++++++++++++++++++
>  fs/proc/base.c                         |  9 +++++
   ^^^^^^^^^^^^^^^

So you're touching this file here and yet your Cc list doesn't have
*anyone* who might be responsible for the proc fs (yeah, you have
linux-fsdevel but I don't think that's enough). Have you heard of
scripts/get_maintainer.pl?

Use it in the future.

Cc-ing some more people for the generic /proc bits.

>  3 files changed, 59 insertions(+)
> 
> diff --git a/arch/x86/include/asm/resctrl_sched.h b/arch/x86/include/asm/resctrl_sched.h
> index f6b7fe2833cc..bba362e0e00f 100644
> --- a/arch/x86/include/asm/resctrl_sched.h
> +++ b/arch/x86/include/asm/resctrl_sched.h
> @@ -5,6 +5,7 @@
>  #ifdef CONFIG_X86_CPU_RESCTRL
>  
>  #include <linux/sched.h>
> +#include <linux/proc_fs.h>
>  #include <linux/jump_label.h>
>  
>  #define IA32_PQR_ASSOC	0x0c8f
> @@ -84,6 +85,9 @@ static inline void resctrl_sched_in(void)
>  		__resctrl_sched_in();
>  }
>  
> +int proc_resctrl_show(struct seq_file *m, struct pid_namespace *ns,
> +		      struct pid *pid, struct task_struct *tsk);
> +
>  #else
>  
>  static inline void resctrl_sched_in(void) {}
> diff --git a/arch/x86/kernel/cpu/resctrl/rdtgroup.c b/arch/x86/kernel/cpu/resctrl/rdtgroup.c
> index a46dee8e78db..2317174174e9 100644
> --- a/arch/x86/kernel/cpu/resctrl/rdtgroup.c
> +++ b/arch/x86/kernel/cpu/resctrl/rdtgroup.c
> @@ -727,6 +727,52 @@ static int rdtgroup_tasks_show(struct kernfs_open_file *of,
>  	return ret;
>  }
>  
> +int proc_resctrl_show(struct seq_file *s, struct pid_namespace *ns,
> +		      struct pid *pid, struct task_struct *tsk)
> +{
> +	struct rdtgroup *rdtg;
> +	int ret = 0;
> +
> +	mutex_lock(&rdtgroup_mutex);
> +
> +	/* Make sure resctrl has been mounted. */
> +	if (!static_branch_unlikely(&rdt_enable_key)) {
> +		ret = -ENODEV;
> +		goto unlock;
> +	}
> +
> +	list_for_each_entry(rdtg, &rdt_all_groups, rdtgroup_list) {
> +		struct rdtgroup *crg;
> +
> +		/*
> +		 * Task information is only relevant for shareable
> +		 * and exclusive groups.
> +		 */
> +		if (rdtg->mode != RDT_MODE_SHAREABLE &&
> +		    rdtg->mode != RDT_MODE_EXCLUSIVE)
> +			continue;
> +
> +		if (rdtg->closid == tsk->closid) {

Save an indentation level:

		if (rdtg->closid != tsk->closid)
			continue;

		seq_printf...

> +			seq_printf(s, "CTRL_MON:/%s\n", rdtg->kn->name);
> +			list_for_each_entry(crg, &rdtg->mon.crdtgrp_list,
> +					    mon.crdtgrp_list) {
> +				if (tsk->rmid != crg->mon.rmid)
> +					continue;
> +				seq_printf(s, "MON:%s%s/mon_groups/%s\n",
> +					   rdtg == &rdtgroup_default ? "" : "/",
> +					   rdtg->kn->name, crg->kn->name);
> +				goto unlock;
> +			}
> +			goto unlock;
> +		}
> +	}
> +	ret = -ENOENT;
> +unlock:
> +	mutex_unlock(&rdtgroup_mutex);
> +
> +	return ret;
> +}
> +
>  static int rdt_last_cmd_status_show(struct kernfs_open_file *of,
>  				    struct seq_file *seq, void *v)
>  {
> diff --git a/fs/proc/base.c b/fs/proc/base.c
> index ebea9501afb8..d8a61db78db5 100644
> --- a/fs/proc/base.c
> +++ b/fs/proc/base.c
> @@ -95,6 +95,9 @@
>  #include <linux/sched/stat.h>
>  #include <linux/posix-timers.h>
>  #include <trace/events/oom.h>
> +#ifdef CONFIG_X86_CPU_RESCTRL
> +#include <asm/resctrl_sched.h>
> +#endif

If anything, this should be abstracted nicely into an
include/linux/resctrl.h header. Other architectures would probably wanna
use it too, I hear ARM64 has something like resctrl.

>  #include "internal.h"
>  #include "fd.h"
>  
> @@ -3060,6 +3063,9 @@ static const struct pid_entry tgid_base_stuff[] = {
>  #endif
>  #ifdef CONFIG_CGROUPS
>  	ONE("cgroup",  S_IRUGO, proc_cgroup_show),
> +#endif
> +#ifdef CONFIG_X86_CPU_RESCTRL
> +	ONE("resctrl", S_IRUGO, proc_resctrl_show),
>  #endif
>  	ONE("oom_score",  S_IRUGO, proc_oom_score),
>  	REG("oom_adj",    S_IRUGO|S_IWUSR, proc_oom_adj_operations),
> @@ -3460,6 +3466,9 @@ static const struct pid_entry tid_base_stuff[] = {
>  #endif
>  #ifdef CONFIG_CGROUPS
>  	ONE("cgroup",  S_IRUGO, proc_cgroup_show),
> +#endif
> +#ifdef CONFIG_X86_CPU_RESCTRL
> +	ONE("resctrl", S_IRUGO, proc_resctrl_show),
>  #endif

Same here. proc_resctrl_show() looks like a generic function but it
is x86-only. Need to call the x86-specific one in a real generic
proc_resctrl_show() which other arches can add their code to, too.

Not like that.

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette

Re: [PATCH] x86/resctrl: Add task resctrl information display

[Yu Chen]

Hi Boris,
Thanks for looking at the patch.
On Wed, Nov 13, 2019 at 07:23:07PM +0100, Borislav Petkov wrote:
> On Thu, Nov 07, 2019 at 11:27:31AM +0800, Chen Yu wrote:
> > Monitoring tools that want to find out which resctrl CTRL
> > and MONITOR groups a task belongs to must currently read
> > the "tasks" file in every group until they locate the process
> > ID.
> > 
> > Add an additional file /proc/{pid}/resctrl to provide this
> > information.
> > 
> > For example:
> >  cat /proc/1193/resctrl
> > CTRL_MON:/ctrl_grp0
> > MON:/ctrl_grp0/mon_groups/mon_grp0
> > 
> > If the resctrl filesystem has not been mounted,
> > reading /proc/{pid}/resctrl returns an error:
> > cat: /proc/1193/resctrl: No such device
> > 
> > Tested-by: Jinshi Chen <jinshi.chen@intel.com>
> > Reviewed-by: Reinette Chatre <reinette.chatre@intel.com>
> > Reviewed-by: Fenghua Yu <fenghua.yu@intel.com>
> > Reviewed-by: Tony Luck <tony.luck@intel.com>
> > Signed-off-by: Chen Yu <yu.c.chen@intel.com>
> > 
> > ---
> >  arch/x86/include/asm/resctrl_sched.h   |  4 +++
> >  arch/x86/kernel/cpu/resctrl/rdtgroup.c | 46 ++++++++++++++++++++++++++
> >  fs/proc/base.c                         |  9 +++++
>    ^^^^^^^^^^^^^^^
> 
> So you're touching this file here and yet your Cc list doesn't have
> *anyone* who might be responsible for the proc fs (yeah, you have
> linux-fsdevel but I don't think that's enough). Have you heard of
> scripts/get_maintainer.pl?
> 
> Use it in the future.
> 
> Cc-ing some more people for the generic /proc bits.
> 
Okay, thanks.
> >  3 files changed, 59 insertions(+)
> > 
> > +
> > +		if (rdtg->closid == tsk->closid) {
> 
> Save an indentation level:
> 
> 		if (rdtg->closid != tsk->closid)
> 			continue;
> 
> 		seq_printf...
>
Okay.
> > +#ifdef CONFIG_X86_CPU_RESCTRL
> > +#include <asm/resctrl_sched.h>
> > +#endif
> 
> If anything, this should be abstracted nicely into an
> include/linux/resctrl.h header. Other architectures would probably wanna
> use it too, I hear ARM64 has something like resctrl.
>
Okay, created this header in the next version.
> > +#endif
> > +#ifdef CONFIG_X86_CPU_RESCTRL
> > +	ONE("resctrl", S_IRUGO, proc_resctrl_show),
> >  #endif
> 
> Same here. proc_resctrl_show() looks like a generic function but it
> is x86-only. Need to call the x86-specific one in a real generic
> proc_resctrl_show() which other arches can add their code to, too.
> 
> Not like that.
> 
Okay, in next version proc_resctrl_show() is declared in resctrl.h.
However since there's no common c source file for resctrl currently,
different architectures might need to implement proc_resctrl_show()
accordingly - currently only x86 needs to do that.

Thanks,
Chenyu
> -- 
> Regards/Gruss,
>     Boris.
> 
> https://people.kernel.org/tglx/notes-about-netiquette

Re: [PATCH] x86/resctrl: Add task resctrl information display

[Borislav Petkov]

On Fri, Nov 15, 2019 at 12:24:11PM +0800, Yu Chen wrote:
> Okay, in next version proc_resctrl_show() is declared in resctrl.h.
> However since there's no common c source file for resctrl currently,

And what do we do if there's no .c file in the repo?

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette