On Mon, Sep 3, 2018 at 7:43 AM Ryusuke Konishi
[off-list ref] wrote:

Hi Souptick,

On Thu, 30 Aug 2018 16:33:52 -0700, Andrew Morton wrote:

quoted

On Thu, 30 Aug 2018 22:55:47 +0530 Souptick Joarder wrote:

quoted

Return type for fault handlers  in ext4 and nilfs are
changed to use vm_fault_t.

Return type of block_page_mkwrite() is changed from
int to vm_fault_t. The function signature of
block_page_mkwrite() is changed to add one new parameter
int *err. This will provide a way for caller functions
to get error value along with return value and use it
further.

Return type of block_page_mkwrite_return() is also changed
to use new vm_fault_t type.

--- a/fs/nilfs2/file.c
+++ b/fs/nilfs2/file.c

@@ -51,13 +51,14 @@ int nilfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
     return err;
 }

-static int nilfs_page_mkwrite(struct vm_fault *vmf)
+static vm_fault_t nilfs_page_mkwrite(struct vm_fault *vmf)

nilfs_page_mkwrite() already has return type vm_fault_t in Linus's
kernel, due to the now-merged
fs-nilfs2-adding-new-return-type-vm_fault_t.patch.  Looks like a simple
fix.

In the first patch in this thread, this return type change was
excluded correctly for nilfs_page_mkwrite() though the changelog was
inaccurate in that sense.

Please confirm your base point of the revised patch.

Ryusuku,
Base point of revised patch is to address your comment by
removing err = 0. Nothing else was suppose to there in v2.

But unfortunately I mixed up my local branches and created v2 against
old branch which leads to confusion of v1 -> v2 delta changes.
Sorry for creating more confusions.

quoted

I'm beginning to feel vm_fault_t exhaustion.  Please remind me what
benefit we're going to get out of all this churn?

quoted

...