ext4: update PF_MEMALLOC handling in ext4_write_inode()

 *

 * We are called from a few places:

 *

 * - Within generic_file_write() for O_SYNC files.

 * - Within generic_file_aio_write() -> generic_write_sync() for O_SYNC files.

 *   Here, there will be no transaction running. We wait for any running

 *   transaction to commit.

 *

 * - Within sys_sync(), kupdate and such.

 *   We wait on commit, if tol to.

 * - Within flush work (sys_sync(), kupdate and such).

 *   We wait on commit, if told to.

 *

 * - Within prune_icache() (PF_MEMALLOC == true)

 *   Here we simply return.  We can't afford to block kswapd on the

 *   journal commit.

 * - Within iput_final() -> write_inode_now()

 *   We wait on commit, if told to.

 *

 * In all cases it is actually safe for us to return without doing anything,

 * because the inode has been copied into a raw inode buffer in

 * ext4_mark_inode_dirty().  This is a correctness thing for O_SYNC and for

 * knfsd.

 * ext4_mark_inode_dirty().  This is a correctness thing for WB_SYNC_ALL

 * writeback.

 *

 * Note that we are absolutely dependent upon all inode dirtiers doing the

 * right thing: they *must* call mark_inode_dirty() after dirtying info in

 *	stuff();

 *	inode->i_size = expr;

 *

 * is in error because a kswapd-driven write_inode() could occur while

 * `stuff()' is running, and the new i_size will be lost.  Plus the inode

 * will no longer be on the superblock's dirty inode list.

 * is in error because write_inode() could occur while `stuff()' is running,

 * and the new i_size will be lost.  Plus the inode will no longer be on the

 * superblock's dirty inode list.

 */

int ext4_write_inode(struct inode *inode, struct writeback_control *wbc)

{

	int err;

	if (current->flags & PF_MEMALLOC)

	if (WARN_ON_ONCE(current->flags & PF_MEMALLOC))

		return 0;

	if (EXT4_SB(inode->i_sb)->s_journal) {