| Message ID | 20240821-work-i_state-v2-6-67244769f102@kernel.org (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | inode: turn i_state into u32 | expand |
On Aug 21, 2024, at 9:47 AM, Christian Brauner <brauner@kernel.org> wrote: > > Now that we use the wait var event mechanism make i_state a u32 and free > up 4 bytes. This means we currently have two 4 byte holes in struct > inode which we can pack. > > Signed-off-by: Christian Brauner <brauner@kernel.org> > --- > include/linux/fs.h | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/include/linux/fs.h b/include/linux/fs.h > index 8525f8bfd7b9..a673173b6896 100644 > --- a/include/linux/fs.h > +++ b/include/linux/fs.h > @@ -681,7 +681,7 @@ struct inode { > #endif > > /* Misc */ > - unsigned long i_state; > + u32 i_state; Is it worthwhile to add a comment that there is a hole here, instead of leaving it for future re-discovery? /* 32-bit hole */ > struct rw_semaphore i_rwsem; > > unsigned long dirtied_when; /* jiffies of first dirtying */ Normally this would be excess noise, but since struct inode is micro-optimized (like struct page) it probably makes sense to document this case. Cheers, Andreas
On Wed, Aug 21, 2024 at 03:03:01PM GMT, Andreas Dilger wrote: > On Aug 21, 2024, at 9:47 AM, Christian Brauner <brauner@kernel.org> wrote: > > > > Now that we use the wait var event mechanism make i_state a u32 and free > > up 4 bytes. This means we currently have two 4 byte holes in struct > > inode which we can pack. > > > > Signed-off-by: Christian Brauner <brauner@kernel.org> > > --- > > include/linux/fs.h | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/include/linux/fs.h b/include/linux/fs.h > > index 8525f8bfd7b9..a673173b6896 100644 > > --- a/include/linux/fs.h > > +++ b/include/linux/fs.h > > @@ -681,7 +681,7 @@ struct inode { > > #endif > > > > /* Misc */ > > - unsigned long i_state; > > + u32 i_state; > > Is it worthwhile to add a comment that there is a hole here, instead > of leaving it for future re-discovery? > > /* 32-bit hole */ > > > struct rw_semaphore i_rwsem; > > > > unsigned long dirtied_when; /* jiffies of first dirtying */ > > Normally this would be excess noise, but since struct inode is > micro-optimized (like struct page) it probably makes sense to > document this case. Good idea. Added now!
diff --git a/include/linux/fs.h b/include/linux/fs.h index 8525f8bfd7b9..a673173b6896 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -681,7 +681,7 @@ struct inode { #endif /* Misc */ - unsigned long i_state; + u32 i_state; struct rw_semaphore i_rwsem; unsigned long dirtied_when; /* jiffies of first dirtying */
Now that we use the wait var event mechanism make i_state a u32 and free up 4 bytes. This means we currently have two 4 byte holes in struct inode which we can pack. Signed-off-by: Christian Brauner <brauner@kernel.org> --- include/linux/fs.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)