x86, fpu: lazy allocation of FPU area - v5

Only allocate the FPU area when the application actually uses FPU, i.e., in the
first lazy FPU trap. This could save memory for non-fpu using apps.

for example: on my system after boot, there are around 300 processes, with
only 17 using FPU.

Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Cc: Arjan van de Ven <arjan@linux.intel.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

1 files changed, 30 insertions, 8 deletions

diff --git a/arch/x86/kernel/i387.c b/arch/x86/kernel/i387.c
index baf632b221d..db6839b5319 100644
--- a/arch/x86/kernel/i387.c
+++ b/arch/x86/kernel/i387.c
@@ -8,7 +8,6 @@
 #include <linux/module.h>
 #include <linux/regset.h>
 #include <linux/sched.h>
-#include <linux/bootmem.h>
 
 #include <asm/sigcontext.h>
 #include <asm/processor.h>
@@ -63,7 +62,6 @@ void __init init_thread_xstate(void)
         else
                 xstate_size = sizeof(struct i387_fsave_struct);
 #endif
-        init_task.thread.xstate = alloc_bootmem(xstate_size);
 }
 
 #ifdef CONFIG_X86_64
@@ -93,12 +91,22 @@ void __cpuinit fpu_init(void)
  * value at reset if we support XMM instructions and then
  * remeber the current task has used the FPU.
  */
-void init_fpu(struct task_struct *tsk)
+int init_fpu(struct task_struct *tsk)
 {
         if (tsk_used_math(tsk)) {
                 if (tsk == current)
                         unlazy_fpu(tsk);
-                return;
+                return 0;
+        }
+
+        /*
+         * Memory allocation at the first usage of the FPU and other state.
+         */
+        if (!tsk->thread.xstate) {
+                tsk->thread.xstate = kmem_cache_alloc(task_xstate_cachep,
+                                                      GFP_KERNEL);
+                if (!tsk->thread.xstate)
+                        return -ENOMEM;
         }
 
         if (cpu_has_fxsr) {
@@ -120,6 +128,7 @@ void init_fpu(struct task_struct *tsk)
          * Only the device not available exception or ptrace can call init_fpu.
          */
         set_stopped_child_used_math(tsk);
+        return 0;
 }
 
 int fpregs_active(struct task_struct *target, const struct user_regset *regset)
@@ -136,10 +145,14 @@ int xfpregs_get(struct task_struct *target, const struct user_regset *regset,
                 unsigned int pos, unsigned int count,
                 void *kbuf, void __user *ubuf)
 {
+        int ret;
+
         if (!cpu_has_fxsr)
                 return -ENODEV;
 
-        init_fpu(target);
+        ret = init_fpu(target);
+        if (ret)
+                return ret;
 
         return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
                                    &target->thread.xstate->fxsave, 0, -1);
@@ -154,7 +167,10 @@ int xfpregs_set(struct task_struct *target, const struct user_regset *regset,
         if (!cpu_has_fxsr)
                 return -ENODEV;
 
-        init_fpu(target);
+        ret = init_fpu(target);
+        if (ret)
+                return ret;
+
         set_stopped_child_used_math(target);
 
         ret = user_regset_copyin(&pos, &count, &kbuf, &ubuf,
@@ -312,11 +328,14 @@ int fpregs_get(struct task_struct *target, const struct user_regset *regset,
                void *kbuf, void __user *ubuf)
 {
         struct user_i387_ia32_struct env;
+        int ret;
 
         if (!HAVE_HWFP)
                 return fpregs_soft_get(target, regset, pos, count, kbuf, ubuf);
 
-        init_fpu(target);
+        ret = init_fpu(target);
+        if (ret)
+                return ret;
 
         if (!cpu_has_fxsr) {
                 return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
@@ -344,7 +363,10 @@ int fpregs_set(struct task_struct *target, const struct user_regset *regset,
         if (!HAVE_HWFP)
                 return fpregs_soft_set(target, regset, pos, count, kbuf, ubuf);
 
-        init_fpu(target);
+        ret = init_fpu(target);
+        if (ret)
+                return ret;
+
         set_stopped_child_used_math(target);
 
         if (!cpu_has_fxsr) {