[PATCH v3 2/2] proc: add /proc/<pid>/arch_state

[Aubrey Li]

Add a /proc/<pid>/arch_state interface to expose per-task cpu specific
state values.

Exposing AVX-512 Hi16_ZMM registers usage is for the user space job
scheduler to cluster AVX-512 using tasks together, because these tasks
could cause core turbo frequency drop.

Signed-off-by: Aubrey Li <aubrey.li@xxxxxxxxxxxxxxx>
Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
Cc: Andi Kleen <ak@xxxxxxxxxxxxxxx>
Cc: Tim Chen <tim.c.chen@xxxxxxxxxxxxxxx>
Cc: Arjan van de Ven <arjan@xxxxxxxxxxxxxxx>
---
 arch/x86/kernel/fpu/xstate.c | 16 ++++++++++++++++
 fs/proc/base.c               | 13 +++++++++++++
 2 files changed, 29 insertions(+)

diff --git a/arch/x86/kernel/fpu/xstate.c b/arch/x86/kernel/fpu/xstate.c
index 87a57b7..10224ee 100644
--- a/arch/x86/kernel/fpu/xstate.c
+++ b/arch/x86/kernel/fpu/xstate.c
@@ -7,6 +7,7 @@
 #include <linux/cpu.h>
 #include <linux/mman.h>
 #include <linux/pkeys.h>
+#include <linux/seq_file.h>
 
 #include <asm/fpu/api.h>
 #include <asm/fpu/internal.h>
@@ -1245,3 +1246,18 @@ int copy_user_to_xstate(struct xregs_state *xsave, const void __user *ubuf)
 
 	return 0;
 }
+
+/*
+ * report CPU specific thread state
+ */
+void arch_thread_state(struct seq_file *m, struct task_struct *task)
+{
+	/*
+	 * Report AVX-512 Hi16_ZMM registers usage
+	 */
+	if (task->thread.fpu.hi16zmm_usage)
+		seq_putc(m, '1');
+	else
+		seq_putc(m, '0');
+	seq_putc(m, '\n');
+}
diff --git a/fs/proc/base.c b/fs/proc/base.c
index aaffc0c..efd51ec 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
@@ -2893,6 +2893,17 @@ static int proc_pid_patch_state(struct seq_file *m, struct pid_namespace *ns,
 }
 #endif /* CONFIG_LIVEPATCH */
 
+void __weak arch_thread_state(struct seq_file *m, struct task_struct *task)
+{
+}
+
+static int proc_pid_arch_state(struct seq_file *m, struct pid_namespace *ns,
+				 struct pid *pid, struct task_struct *task)
+{
+	arch_thread_state(m, task);
+	return 0;
+}
+
 /*
  * Thread groups
  */
@@ -2994,6 +3005,7 @@ static const struct pid_entry tgid_base_stuff[] = {
 #ifdef CONFIG_LIVEPATCH
 	ONE("patch_state",  S_IRUSR, proc_pid_patch_state),
 #endif
+	ONE("arch_state",  S_IRUSR, proc_pid_arch_state),
 };
 
 static int proc_tgid_base_readdir(struct file *file, struct dir_context *ctx)
@@ -3372,6 +3384,7 @@ static const struct pid_entry tid_base_stuff[] = {
 #ifdef CONFIG_LIVEPATCH
 	ONE("patch_state",  S_IRUSR, proc_pid_patch_state),
 #endif
+	ONE("arch_state",  S_IRUSR, proc_pid_arch_state),
 };
 
 static int proc_tid_base_readdir(struct file *file, struct dir_context *ctx)
-- 
2.7.4