[RFC] Renaming page_offset() to page_pos()

[RFC] Renaming page_offset() to page_pos()

[Matthew Wilcox]

Without looking at the source, can you tell me what page_offset() does?

At least one regular contributor thought it meant the pgoff_t of this
page within the file.  It's actually the byte offset of this page into
the file.

We have a perfectly good name for byte offset into the file --
file->f_pos.  So I propose renaming it to page_pos().  To minimise
disruption to other development, I'm going to send Linus a pull request
at the end of the merge window with the results of this coccinelle script:

@@ expression a; @@
-       page_offset(a)
+       page_pos(a)

I've reviewed the output and the only slight weirdness is an extra space
in casts:

                btrfs_warn(BTRFS_I(page->mapping->host)->root->fs_info,
                           "page private not zero on page %llu",
-                          (unsigned long long)page_offset(page));
+                          (unsigned long long) page_pos(page));

Sometimes Coccinelle fixes the surrounding whitespace to be better
than it currently is:

-               ow->bv[i].bv_len = min(page_offset(ow->pages[i]) + PAGE_SIZE,
-                   ow->off + ow->len) -
-                   max(ow->off, page_offset(ow->pages[i]));
+               ow->bv[i].bv_len = min(page_pos(ow->pages[i]) + PAGE_SIZE,
+                                      ow->off + ow->len) -
+                   max(ow->off, page_pos(ow->pages[i]));

(it's still bad, but it's an improvement)

Any objections?  Anyone got a better name than page_pos()?

Re: [RFC] Renaming page_offset() to page_pos()

[William Kucharski]

page_pos seems quite reasonable to me.

Reviewed-by: William Kucharski <william.kucharski@oracle.com>


> On Apr 3, 2020, at 9:33 AM, Matthew Wilcox <willy@infradead.org> wrote:
> 
> Without looking at the source, can you tell me what page_offset() does?
> 
> At least one regular contributor thought it meant the pgoff_t of this
> page within the file.  It's actually the byte offset of this page into
> the file.
> 
> We have a perfectly good name for byte offset into the file --
> file->f_pos.  So I propose renaming it to page_pos().  To minimise
> disruption to other development, I'm going to send Linus a pull request
> at the end of the merge window with the results of this coccinelle script:
> 
> @@ expression a; @@
> -       page_offset(a)
> +       page_pos(a)
> 
> I've reviewed the output and the only slight weirdness is an extra space
> in casts:
> 
>                btrfs_warn(BTRFS_I(page->mapping->host)->root->fs_info,
>                           "page private not zero on page %llu",
> -                          (unsigned long long)page_offset(page));
> +                          (unsigned long long) page_pos(page));
> 
> Sometimes Coccinelle fixes the surrounding whitespace to be better
> than it currently is:
> 
> -               ow->bv[i].bv_len = min(page_offset(ow->pages[i]) + PAGE_SIZE,
> -                   ow->off + ow->len) -
> -                   max(ow->off, page_offset(ow->pages[i]));
> +               ow->bv[i].bv_len = min(page_pos(ow->pages[i]) + PAGE_SIZE,
> +                                      ow->off + ow->len) -
> +                   max(ow->off, page_pos(ow->pages[i]));
> 
> (it's still bad, but it's an improvement)
> 
> Any objections?  Anyone got a better name than page_pos()?
>