fs/spinlock: replace no sched_lock() version to improve the performance

after below change merge to kernel, spin_lock() will turn off preemption by default,
but this change is not applicable to all scenarios. The locations in the kernel that
use spin_lock() extensively only require short critical sections and do not trigger
scheduling, which leads to serious performance degradation of NuttX in AMP mode.

In this PR, I try to expose similar problems and hope that each subsystem will carefully check the code coverage

apache#14578
|commit b69111d
|Author: hujun5 <[email protected]>
|Date:   Thu Jan 23 16:14:18 2025 +0800
|
|    spinlock: add sched_lock to spin_lock_irqsave
|
|    reason:
|    We aim to replace big locks with smaller ones. So we will use spin_lock_irqsave extensively to
|    replace enter_critical_section in the subsequent process. We imitate the implementation of Linux
|    by adding sched_lock to spin_lock_irqsave in order to address scenarios where sem_post occurs
|    within spin_lock_irqsave, which can lead to spinlock failures and deadlocks.
|
|    Signed-off-by: hujun5 <[email protected]>

Signed-off-by: chao an <[email protected]>

fs/inode/fs_files.c

@@ -72,9 +72,9 @@ static FAR struct file *files_fget_by_index(FAR struct filelist *list,
   FAR struct file *filep;
   irqstate_t flags;
 
-  flags = spin_lock_irqsave(&list->fl_lock);
+  flags = raw_spin_lock_irqsave(&list->fl_lock);
   filep = &list->fl_files[l1][l2];
-  spin_unlock_irqrestore(&list->fl_lock, flags);
+  raw_spin_unlock_irqrestore(&list->fl_lock, flags);
 
 #ifdef CONFIG_FS_REFCOUNT
   if (filep->f_inode != NULL)
@@ -164,7 +164,7 @@ static int files_extend(FAR struct filelist *list, size_t row)
     }
   while (++i < row);
 
-  flags = spin_lock_irqsave(&list->fl_lock);
+  flags = raw_spin_lock_irqsave(&list->fl_lock);
 
   /* To avoid race condition, if the file list is updated by other threads
    * and list rows is greater or equal than temp list,
@@ -173,7 +173,7 @@ static int files_extend(FAR struct filelist *list, size_t row)
 
   if (orig_rows != list->fl_rows && list->fl_rows >= row)
     {
-      spin_unlock_irqrestore(&list->fl_lock, flags);
+      raw_spin_unlock_irqrestore(&list->fl_lock, flags);
 
       for (j = orig_rows; j < i; j++)
         {
@@ -195,7 +195,7 @@ static int files_extend(FAR struct filelist *list, size_t row)
   list->fl_files = files;
   list->fl_rows = row;
 
-  spin_unlock_irqrestore(&list->fl_lock, flags);
+  raw_spin_unlock_irqrestore(&list->fl_lock, flags);
 
   if (tmp != NULL && tmp != &list->fl_prefile)
     {
@@ -565,21 +565,21 @@ int file_allocate_from_tcb(FAR struct tcb_s *tcb, FAR struct inode *inode,
 
   /* Find free file */
 
-  flags = spin_lock_irqsave(&list->fl_lock);
+  flags = raw_spin_lock_irqsave(&list->fl_lock);
 
   for (; ; i++, j = 0)
     {
       if (i >= list->fl_rows)
         {
-          spin_unlock_irqrestore(&list->fl_lock, flags);
+          raw_spin_unlock_irqrestore(&list->fl_lock, flags);
 
           ret = files_extend(list, i + 1);
           if (ret < 0)
             {
               return ret;
             }
 
-          flags = spin_lock_irqsave(&list->fl_lock);
+          flags = raw_spin_lock_irqsave(&list->fl_lock);
         }
 
       do
@@ -608,7 +608,7 @@ int file_allocate_from_tcb(FAR struct tcb_s *tcb, FAR struct inode *inode,
     }
 
 found:
-  spin_unlock_irqrestore(&list->fl_lock, flags);
+  raw_spin_unlock_irqrestore(&list->fl_lock, flags);
 
   if (addref)
     {