2024-02-21 19:43:27

by Suren Baghdasaryan

[permalink] [raw]
Subject: [PATCH v4 05/36] fs: Convert alloc_inode_sb() to a macro

From: Kent Overstreet <[email protected]>

We're introducing alloc tagging, which tracks memory allocations by
callsite. Converting alloc_inode_sb() to a macro means allocations will
be tracked by its caller, which is a bit more useful.

Signed-off-by: Kent Overstreet <[email protected]>
Signed-off-by: Suren Baghdasaryan <[email protected]>
Cc: Alexander Viro <[email protected]>
Reviewed-by: Kees Cook <[email protected]>
---
include/linux/fs.h | 6 +-----
1 file changed, 1 insertion(+), 5 deletions(-)

diff --git a/include/linux/fs.h b/include/linux/fs.h
index 023f37c60709..08d8246399c3 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -3010,11 +3010,7 @@ int setattr_should_drop_sgid(struct mnt_idmap *idmap,
* This must be used for allocating filesystems specific inodes to set
* up the inode reclaim context correctly.
*/
-static inline void *
-alloc_inode_sb(struct super_block *sb, struct kmem_cache *cache, gfp_t gfp)
-{
- return kmem_cache_alloc_lru(cache, &sb->s_inode_lru, gfp);
-}
+#define alloc_inode_sb(_sb, _cache, _gfp) kmem_cache_alloc_lru(_cache, &_sb->s_inode_lru, _gfp)

extern void __insert_inode_hash(struct inode *, unsigned long hashval);
static inline void insert_inode_hash(struct inode *inode)
--
2.44.0.rc0.258.g7320e95886-goog



2024-02-21 21:24:34

by Pasha Tatashin

[permalink] [raw]
Subject: Re: [PATCH v4 05/36] fs: Convert alloc_inode_sb() to a macro

On Wed, Feb 21, 2024 at 2:41 PM Suren Baghdasaryan <[email protected]> wrote:
>
> From: Kent Overstreet <[email protected]>
>
> We're introducing alloc tagging, which tracks memory allocations by
> callsite. Converting alloc_inode_sb() to a macro means allocations will
> be tracked by its caller, which is a bit more useful.
>
> Signed-off-by: Kent Overstreet <[email protected]>
> Signed-off-by: Suren Baghdasaryan <[email protected]>
> Cc: Alexander Viro <[email protected]>
> Reviewed-by: Kees Cook <[email protected]>

Reviewed-by: Pasha Tatashin <[email protected]>

2024-02-26 16:45:31

by Vlastimil Babka

[permalink] [raw]
Subject: Re: [PATCH v4 05/36] fs: Convert alloc_inode_sb() to a macro

On 2/21/24 20:40, Suren Baghdasaryan wrote:
> From: Kent Overstreet <[email protected]>
>
> We're introducing alloc tagging, which tracks memory allocations by
> callsite. Converting alloc_inode_sb() to a macro means allocations will
> be tracked by its caller, which is a bit more useful.
>
> Signed-off-by: Kent Overstreet <[email protected]>
> Signed-off-by: Suren Baghdasaryan <[email protected]>
> Cc: Alexander Viro <[email protected]>
> Reviewed-by: Kees Cook <[email protected]>
> ---
> include/linux/fs.h | 6 +-----
> 1 file changed, 1 insertion(+), 5 deletions(-)
>
> diff --git a/include/linux/fs.h b/include/linux/fs.h
> index 023f37c60709..08d8246399c3 100644
> --- a/include/linux/fs.h
> +++ b/include/linux/fs.h
> @@ -3010,11 +3010,7 @@ int setattr_should_drop_sgid(struct mnt_idmap *idmap,
> * This must be used for allocating filesystems specific inodes to set
> * up the inode reclaim context correctly.
> */
> -static inline void *
> -alloc_inode_sb(struct super_block *sb, struct kmem_cache *cache, gfp_t gfp)

A __always_inline wouldn't have the same effect? Just wondering.

> -{
> - return kmem_cache_alloc_lru(cache, &sb->s_inode_lru, gfp);
> -}
> +#define alloc_inode_sb(_sb, _cache, _gfp) kmem_cache_alloc_lru(_cache, &_sb->s_inode_lru, _gfp)
>
> extern void __insert_inode_hash(struct inode *, unsigned long hashval);
> static inline void insert_inode_hash(struct inode *inode)


2024-02-26 17:49:20

by Suren Baghdasaryan

[permalink] [raw]
Subject: Re: [PATCH v4 05/36] fs: Convert alloc_inode_sb() to a macro

On Mon, Feb 26, 2024 at 7:44 AM Vlastimil Babka <[email protected]> wrote:
>
> On 2/21/24 20:40, Suren Baghdasaryan wrote:
> > From: Kent Overstreet <[email protected]>
> >
> > We're introducing alloc tagging, which tracks memory allocations by
> > callsite. Converting alloc_inode_sb() to a macro means allocations will
> > be tracked by its caller, which is a bit more useful.
> >
> > Signed-off-by: Kent Overstreet <[email protected]>
> > Signed-off-by: Suren Baghdasaryan <[email protected]>
> > Cc: Alexander Viro <[email protected]>
> > Reviewed-by: Kees Cook <[email protected]>
> > ---
> > include/linux/fs.h | 6 +-----
> > 1 file changed, 1 insertion(+), 5 deletions(-)
> >
> > diff --git a/include/linux/fs.h b/include/linux/fs.h
> > index 023f37c60709..08d8246399c3 100644
> > --- a/include/linux/fs.h
> > +++ b/include/linux/fs.h
> > @@ -3010,11 +3010,7 @@ int setattr_should_drop_sgid(struct mnt_idmap *idmap,
> > * This must be used for allocating filesystems specific inodes to set
> > * up the inode reclaim context correctly.
> > */
> > -static inline void *
> > -alloc_inode_sb(struct super_block *sb, struct kmem_cache *cache, gfp_t gfp)
>
> A __always_inline wouldn't have the same effect? Just wondering.

I think inlining it would still keep __LINE__ and __FILE__ pointing to
this location in the header instead of the location where the call
happens. If we change alloc_inode_sb() to inline we will have to wrap
it with alloc_hook() and call kmem_cache_alloc_lru_noprof() inside it.
Doable but this change seems much simpler.

>
> > -{
> > - return kmem_cache_alloc_lru(cache, &sb->s_inode_lru, gfp);
> > -}
> > +#define alloc_inode_sb(_sb, _cache, _gfp) kmem_cache_alloc_lru(_cache, &_sb->s_inode_lru, _gfp)
> >
> > extern void __insert_inode_hash(struct inode *, unsigned long hashval);
> > static inline void insert_inode_hash(struct inode *inode)
>
> --
> To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
>

2024-02-26 20:50:48

by Kent Overstreet

[permalink] [raw]
Subject: Re: [PATCH v4 05/36] fs: Convert alloc_inode_sb() to a macro

On Mon, Feb 26, 2024 at 04:44:51PM +0100, Vlastimil Babka wrote:
> On 2/21/24 20:40, Suren Baghdasaryan wrote:
> > From: Kent Overstreet <[email protected]>
> >
> > We're introducing alloc tagging, which tracks memory allocations by
> > callsite. Converting alloc_inode_sb() to a macro means allocations will
> > be tracked by its caller, which is a bit more useful.
> >
> > Signed-off-by: Kent Overstreet <[email protected]>
> > Signed-off-by: Suren Baghdasaryan <[email protected]>
> > Cc: Alexander Viro <[email protected]>
> > Reviewed-by: Kees Cook <[email protected]>
> > ---
> > include/linux/fs.h | 6 +-----
> > 1 file changed, 1 insertion(+), 5 deletions(-)
> >
> > diff --git a/include/linux/fs.h b/include/linux/fs.h
> > index 023f37c60709..08d8246399c3 100644
> > --- a/include/linux/fs.h
> > +++ b/include/linux/fs.h
> > @@ -3010,11 +3010,7 @@ int setattr_should_drop_sgid(struct mnt_idmap *idmap,
> > * This must be used for allocating filesystems specific inodes to set
> > * up the inode reclaim context correctly.
> > */
> > -static inline void *
> > -alloc_inode_sb(struct super_block *sb, struct kmem_cache *cache, gfp_t gfp)
>
> A __always_inline wouldn't have the same effect? Just wondering.

nope, macro expansion within an inline happens once, and will show
__func__ and __line__ of the helper, we want it expanded in the caller