On Wed, Nov 09, 2022 at 01:58:29PM +0000, Matthew Wilcox wrote:
> On Tue, Nov 08, 2022 at 11:03:16PM +0000, Daniel Golle wrote:
> > + /* map first page */
> > + page = read_mapping_page(
> > + mapping, fit_start_sector >> (PAGE_SHIFT - SECTOR_SHIFT), NULL);
> > +
> > + if (IS_ERR(page))
> > + return -EFAULT;
> > +
> > + if (PageError(page))
> > + return -EFAULT;
>
> Why are you checking for PageError? You won't ever get a page with an
> error back from read_mapping_page(). And you have the real error in
> 'page', so why return -EFAUlT, which would indicate a problem copying
> from the user. Also, this is a great place to use the new folio APIs
> instead of the old page APIs. So:
>
> folio = read_mapping_folio(mapping,
> fit_start_sector >> PAGE_SECTORS_SHIFT, NULL);
> if (IS_ERR(folio))
> return PTR_ERR(folio);
>
> > + init_fit = page_address(page);
>
> init_fit = folio_address(folio) +
> offset_in_folio(folio, fit_start_sector * SECTOR_SIZE);
>
> > + if (!init_fit) {
> > + put_page(page);
> > + return -EFAULT;
> > + }
>
> page_address() or folio_address() can't ever return NULL, you should
> just drop this nonsense check.
Thank you for the pointers, I will implement your suggestions and post
v5 after the upcoming weekend.
>
> ... actually, why can't you call read_part_sector() and avoid all of
> this?
I've tried that before and the problem is that read_part_sector()
returns a pointer to one sector (typically 512 bytes) of data.
And this pointer should not be accesses beyond sector boundaries,
right? You'd have to call read_part_sector() again for the next
sector.
The FIT structure, however, usually exceeds the size of one sector,
and having a continous memory area covering the structure as a whole
is crucial for libfdt to do its job.
I could, of course, use read_part_sector() to copy all sectors
covering the FIT structure into a buffer, but that seemed strange
given that read_part_sector() actually used read_mapping_page()
(and now uses read_mapping_folio()) internally and then returns a
pointer to the offset within the page/folio. So why not read it in one
piece in first place instead of having it first split up to sectors
by read_part_sector() just to then having to reassemble it into a
continous buffer again.
