ext4: update PF_MEMALLOC handling in ext4_write_inode()

The special handling of PF_MEMALLOC callers in ext4_write_inode()
shouldn't be necessary as there shouldn't be any. Warn about it. Also
update comment before the function as it seems somewhat outdated.

(Changes modeled on an ext3 patch posted by Jan Kara to the linux-ext4
mailing list on Februaryt 28, 2014, which apparently never went into
the ext3 tree.)

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Cc: Jan Kara <jack@suse.cz>
This commit is contained in:
Theodore Ts'o 2014-04-08 11:38:28 -04:00
parent ec4cb1aa2b
commit 87f7e41636

View file

@ -4427,21 +4427,20 @@ static int ext4_do_update_inode(handle_t *handle,
* *
* We are called from a few places: * 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 * Here, there will be no transaction running. We wait for any running
* transaction to commit. * transaction to commit.
* *
* - Within sys_sync(), kupdate and such. * - Within flush work (sys_sync(), kupdate and such).
* We wait on commit, if tol to. * We wait on commit, if told to.
* *
* - Within prune_icache() (PF_MEMALLOC == true) * - Within iput_final() -> write_inode_now()
* Here we simply return. We can't afford to block kswapd on the * We wait on commit, if told to.
* journal commit.
* *
* In all cases it is actually safe for us to return without doing anything, * 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 * 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 * ext4_mark_inode_dirty(). This is a correctness thing for WB_SYNC_ALL
* knfsd. * writeback.
* *
* Note that we are absolutely dependent upon all inode dirtiers doing the * Note that we are absolutely dependent upon all inode dirtiers doing the
* right thing: they *must* call mark_inode_dirty() after dirtying info in * right thing: they *must* call mark_inode_dirty() after dirtying info in
@ -4453,15 +4452,15 @@ static int ext4_do_update_inode(handle_t *handle,
* stuff(); * stuff();
* inode->i_size = expr; * inode->i_size = expr;
* *
* is in error because a kswapd-driven write_inode() could occur while * is in error because write_inode() could occur while `stuff()' is running,
* `stuff()' is running, and the new i_size will be lost. Plus the inode * and the new i_size will be lost. Plus the inode will no longer be on the
* will no longer be on the superblock's dirty inode list. * superblock's dirty inode list.
*/ */
int ext4_write_inode(struct inode *inode, struct writeback_control *wbc) int ext4_write_inode(struct inode *inode, struct writeback_control *wbc)
{ {
int err; int err;
if (current->flags & PF_MEMALLOC) if (WARN_ON_ONCE(current->flags & PF_MEMALLOC))
return 0; return 0;
if (EXT4_SB(inode->i_sb)->s_journal) { if (EXT4_SB(inode->i_sb)->s_journal) {