Re: [f2fs-dev] [PATCH 07/16] fsverity: don't issue readahead for non-ENOENT errors from __filemap_get_folio

2026-01-27 Thread Eric Biggers via Linux-f2fs-devel
On Tue, Jan 27, 2026 at 07:38:09AM +0100, Christoph Hellwig wrote:
> PTR_ERR(ptr) == -EFOO checks if ptr is an error pointer for the errno
> value -EFOO.

To reiterate (again): when ptr may or may not be an error pointer, it
should be written as ptr == ERR_PTR(-EFOO), as is normally done.
Otherwise an error code is being extracted from something that doesn't
have an error code, which is nonsense, even if it works by accident.

- Eric


___
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel


Re: [f2fs-dev] [PATCH 07/16] fsverity: don't issue readahead for non-ENOENT errors from __filemap_get_folio

2026-01-26 Thread Christoph Hellwig
On Mon, Jan 26, 2026 at 10:28:49PM -0800, Darrick J. Wong wrote:
> > Yes, it's really just a cast, and 'PTR_ERR(folio) == -ENOENT' actually
> > still works when folio isn't necessarily an error pointer.  But normally
> > it would be written as a pointer comparison as I suggested.
> 
> How does one know that a pointer is an error pointer?  Oughtn't there be
> some kind of obvious marker, or is IS_ERR the only tool we've got?

IS_ERR(ptr) is the interface to check is a pointer is an error pointer
or not.

PTR_ERR(ptr) == -EFOO checks if ptr is an error pointer for the errno
value -EFOO.




___
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel


Re: [f2fs-dev] [PATCH 07/16] fsverity: don't issue readahead for non-ENOENT errors from __filemap_get_folio

2026-01-26 Thread Christoph Hellwig
On Mon, Jan 26, 2026 at 10:20:55PM -0800, Eric Biggers wrote:
> > That's new to me, and I can't find anything in the documentation or
> > implementation suggesting that.  Your example code above also does
> > this as does plenty of code in the kernel elsewhere.
> 
> Not sure why this is controversial.

It wasn't controversial until you came up with that claim.

> The documentation for PTR_ERR() is
> clear that it's for error pointers:

Yes, but anything that stores an ERR_PTR is an error pointer.  There
never has been any explicit requirement to first call IS_ERR.

One very common pattern is to extract it first an then check
for errors like:

error = PTR_ERR(ptr);
if (IS_ERR(ptr)))
goto handler_error;

one could come up with arguments that this is special, because error
is not used until after the branch.  But there's plenty of other code
like:

type = alg_get_type(sa->salg_type);
if (PTR_ERR(type) == -ENOENT) {
request_module("algif-%s", sa->salg_type);
type = alg_get_type(sa->salg_type);
}

if (IS_ERR(type))
return PTR_ERR(type);

>  * PTR_ERR - Extract the error code from an error pointer.
>  * @ptr: An error pointer.
>  * Return: The error code within @ptr.
>  */
> static inline long __must_check PTR_ERR(__force const void *ptr)
> {
> return (long) ptr;
> }
> 
> Yes, it's really just a cast, and 'PTR_ERR(folio) == -ENOENT' actually
> still works when folio isn't necessarily an error pointer.  But normally
> it would be written as a pointer comparison as I suggested.

You suggestion is using PTR_ERR before checking, to quote from the
previous mail:

> Or as a diff from this series:
>
> - if (PTR_ERR(folio) == -ENOENT ||
> - !(IS_ERR(folio) && !folio_test_uptodate(folio))) {
> + if (folio == ERR_PTR(-ENOENT) ||
> + (!IS_ERR(folio) && !folio_test_uptodate(folio))) {



___
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel


Re: [f2fs-dev] [PATCH 07/16] fsverity: don't issue readahead for non-ENOENT errors from __filemap_get_folio

2026-01-26 Thread Darrick J. Wong via Linux-f2fs-devel
On Mon, Jan 26, 2026 at 10:20:55PM -0800, Eric Biggers wrote:
> On Tue, Jan 27, 2026 at 07:00:39AM +0100, Christoph Hellwig wrote:
> > > - if (PTR_ERR(folio) == -ENOENT ||
> > > - !(IS_ERR(folio) && !folio_test_uptodate(folio))) {
> > > + if (folio == ERR_PTR(-ENOENT) ||
> > > + (!IS_ERR(folio) && !folio_test_uptodate(folio))) {
> > > 
> > > (Note that PTR_ERR() shouldn't be used before it's known that the
> > > pointer is an error pointer.)
> > 
> > That's new to me, and I can't find anything in the documentation or
> > implementation suggesting that.  Your example code above also does
> > this as does plenty of code in the kernel elsewhere.
> 
> Not sure why this is controversial.  The documentation for PTR_ERR() is
> clear that it's for error pointers:
> 
> /**
>  * PTR_ERR - Extract the error code from an error pointer.
>  * @ptr: An error pointer.
>  * Return: The error code within @ptr.
>  */
> static inline long __must_check PTR_ERR(__force const void *ptr)
> {
> return (long) ptr;
> }
> 
> Yes, it's really just a cast, and 'PTR_ERR(folio) == -ENOENT' actually
> still works when folio isn't necessarily an error pointer.  But normally
> it would be written as a pointer comparison as I suggested.

How does one know that a pointer is an error pointer?  Oughtn't there be
some kind of obvious marker, or is IS_ERR the only tool we've got?

--D

> - Eric


___
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel


Re: [f2fs-dev] [PATCH 07/16] fsverity: don't issue readahead for non-ENOENT errors from __filemap_get_folio

2026-01-26 Thread Eric Biggers via Linux-f2fs-devel
On Tue, Jan 27, 2026 at 07:00:39AM +0100, Christoph Hellwig wrote:
> > -   if (PTR_ERR(folio) == -ENOENT ||
> > -   !(IS_ERR(folio) && !folio_test_uptodate(folio))) {
> > +   if (folio == ERR_PTR(-ENOENT) ||
> > +   (!IS_ERR(folio) && !folio_test_uptodate(folio))) {
> > 
> > (Note that PTR_ERR() shouldn't be used before it's known that the
> > pointer is an error pointer.)
> 
> That's new to me, and I can't find anything in the documentation or
> implementation suggesting that.  Your example code above also does
> this as does plenty of code in the kernel elsewhere.

Not sure why this is controversial.  The documentation for PTR_ERR() is
clear that it's for error pointers:

/**
 * PTR_ERR - Extract the error code from an error pointer.
 * @ptr: An error pointer.
 * Return: The error code within @ptr.
 */
static inline long __must_check PTR_ERR(__force const void *ptr)
{
return (long) ptr;
}

Yes, it's really just a cast, and 'PTR_ERR(folio) == -ENOENT' actually
still works when folio isn't necessarily an error pointer.  But normally
it would be written as a pointer comparison as I suggested.

- Eric


___
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel


Re: [f2fs-dev] [PATCH 07/16] fsverity: don't issue readahead for non-ENOENT errors from __filemap_get_folio

2026-01-26 Thread Christoph Hellwig
On Mon, Jan 26, 2026 at 12:53:01PM -0800, Eric Biggers wrote:
> Then for the final version in generic_readahead_merkle_tree(), one
> option would be:
> 
>   struct folio *folio;
> 
>   folio = __filemap_get_folio(inode->i_mapping, index, FGP_ACCESSED, 0);
>   if (folio == ERR_PTR(-ENOENT) ||
>   (!IS_ERR(folio) && !folio_test_uptodate(folio))) {
>   DEFINE_READAHEAD(ractl, NULL, NULL, inode->i_mapping, index);
> 
>   page_cache_ra_unbounded(&ractl, nr_pages, 0);
>   }
>   if (!IS_ERR(folio))
>   folio_put(folio);
> 
> Or as a diff from this series:

I ended up doing the second version (which is what I intended to do
anyway, but messed up the brace placement) in this patch.  It then
automatically carries over to the readahead split.

> 
> - if (PTR_ERR(folio) == -ENOENT ||
> - !(IS_ERR(folio) && !folio_test_uptodate(folio))) {
> + if (folio == ERR_PTR(-ENOENT) ||
> + (!IS_ERR(folio) && !folio_test_uptodate(folio))) {
> 
> (Note that PTR_ERR() shouldn't be used before it's known that the
> pointer is an error pointer.)

That's new to me, and I can't find anything in the documentation or
implementation suggesting that.  Your example code above also does
this as does plenty of code in the kernel elsewhere.



___
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel


Re: [f2fs-dev] [PATCH 07/16] fsverity: don't issue readahead for non-ENOENT errors from __filemap_get_folio

2026-01-26 Thread Eric Biggers via Linux-f2fs-devel
On Mon, Jan 26, 2026 at 11:11:02AM -0800, Darrick J. Wong wrote:
> On Mon, Jan 26, 2026 at 05:50:53AM +0100, Christoph Hellwig wrote:
> > Issuing more reads on errors is not a good idea, especially when the
> > most common error here is -ENOMEM.
> > 
> > Signed-off-by: Christoph Hellwig 
> > ---
> >  fs/verity/pagecache.c | 3 ++-
> >  1 file changed, 2 insertions(+), 1 deletion(-)
> > 
> > diff --git a/fs/verity/pagecache.c b/fs/verity/pagecache.c
> > index 1efcdde20b73..63393f0f5834 100644
> > --- a/fs/verity/pagecache.c
> > +++ b/fs/verity/pagecache.c
> > @@ -22,7 +22,8 @@ struct page *generic_read_merkle_tree_page(struct inode 
> > *inode, pgoff_t index,
> > struct folio *folio;
> >  
> > folio = __filemap_get_folio(inode->i_mapping, index, FGP_ACCESSED, 0);
> > -   if (IS_ERR(folio) || !folio_test_uptodate(folio)) {
> > +   if (PTR_ERR(folio) == -ENOENT ||
> > +   !(IS_ERR(folio) && !folio_test_uptodate(folio))) {
> 
> I don't understand this logic at all.  If @folio is actually an
> ERR_PTR, then we dereference the non-folio to see if it's not uptodate?
> 
> I think (given the previous revisions) that what you want is to initiate
> readahead if either there's no folio at all (ENOENT) or if there is a
> folio but it's not uptodate?  But not if there's some other error
> (ENOMEM, EL3HLT, EFSCORRUPTED, etc)?
> 
> So maybe you want:
> 
>   folio = __filemap_get_folio(...);
>   if (!IS_ERR(folio)) {
>   if (folio_test_uptodate(folio))
>   return folio_file_page(folio);
>   folio_put(folio);
>   } else if (PTR_ERR(folio) == -ENOENT) {
>   return ERR_CAST(folio);
>   }
> 
>   if (num_ra_pages > 1)
>   page_cache_ra_unbounded(&ractl, num_ra_pages, 0);
>   folio = read_mapping_folio(inode->i_mapping, index, NULL);
>   if (IS_ERR(folio))
>   return ERR_CAST(folio);
> 
>   return folio_file_page(folio);
> 
> 

That version is wrong too: the condition 'PTR_ERR(folio) == -ENOENT' is
backwards.

This code gets replaced later in the series anyway.  For this patch, we
could simply insert two lines:

folio = __filemap_get_folio(inode->i_mapping, index, FGP_ACCESSED, 0);
+   if (IS_ERR(folio) && folio != ERR_PTR(-ENOENT))
+   return folio;

Then for the final version in generic_readahead_merkle_tree(), one
option would be:

struct folio *folio;

folio = __filemap_get_folio(inode->i_mapping, index, FGP_ACCESSED, 0);
if (folio == ERR_PTR(-ENOENT) ||
(!IS_ERR(folio) && !folio_test_uptodate(folio))) {
DEFINE_READAHEAD(ractl, NULL, NULL, inode->i_mapping, index);

page_cache_ra_unbounded(&ractl, nr_pages, 0);
}
if (!IS_ERR(folio))
folio_put(folio);

Or as a diff from this series:

-   if (PTR_ERR(folio) == -ENOENT ||
-   !(IS_ERR(folio) && !folio_test_uptodate(folio))) {
+   if (folio == ERR_PTR(-ENOENT) ||
+   (!IS_ERR(folio) && !folio_test_uptodate(folio))) {

(Note that PTR_ERR() shouldn't be used before it's known that the
pointer is an error pointer.)

- Eric


___
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel


Re: [f2fs-dev] [PATCH 07/16] fsverity: don't issue readahead for non-ENOENT errors from __filemap_get_folio

2026-01-26 Thread Darrick J. Wong via Linux-f2fs-devel
On Mon, Jan 26, 2026 at 05:50:53AM +0100, Christoph Hellwig wrote:
> Issuing more reads on errors is not a good idea, especially when the
> most common error here is -ENOMEM.
> 
> Signed-off-by: Christoph Hellwig 
> ---
>  fs/verity/pagecache.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/fs/verity/pagecache.c b/fs/verity/pagecache.c
> index 1efcdde20b73..63393f0f5834 100644
> --- a/fs/verity/pagecache.c
> +++ b/fs/verity/pagecache.c
> @@ -22,7 +22,8 @@ struct page *generic_read_merkle_tree_page(struct inode 
> *inode, pgoff_t index,
>   struct folio *folio;
>  
>   folio = __filemap_get_folio(inode->i_mapping, index, FGP_ACCESSED, 0);
> - if (IS_ERR(folio) || !folio_test_uptodate(folio)) {
> + if (PTR_ERR(folio) == -ENOENT ||
> + !(IS_ERR(folio) && !folio_test_uptodate(folio))) {

I don't understand this logic at all.  If @folio is actually an
ERR_PTR, then we dereference the non-folio to see if it's not uptodate?

I think (given the previous revisions) that what you want is to initiate
readahead if either there's no folio at all (ENOENT) or if there is a
folio but it's not uptodate?  But not if there's some other error
(ENOMEM, EL3HLT, EFSCORRUPTED, etc)?

So maybe you want:

folio = __filemap_get_folio(...);
if (!IS_ERR(folio)) {
if (folio_test_uptodate(folio))
return folio_file_page(folio);
folio_put(folio);
} else if (PTR_ERR(folio) == -ENOENT) {
return ERR_CAST(folio);
}

if (num_ra_pages > 1)
page_cache_ra_unbounded(&ractl, num_ra_pages, 0);
folio = read_mapping_folio(inode->i_mapping, index, NULL);
if (IS_ERR(folio))
return ERR_CAST(folio);

return folio_file_page(folio);



--D

>   DEFINE_READAHEAD(ractl, NULL, NULL, inode->i_mapping, index);
>  
>   if (!IS_ERR(folio))
> -- 
> 2.47.3
> 
> 


___
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel