| Message ID | 20190218100433.20048-9-chandan@linux.ibm.com |
|---|---|
| State | Superseded |
| Headers | show |
| Series | Consolidate Post read processing code | expand |
On Mon, Feb 18, 2019 at 03:34:31PM +0530, Chandan Rajendra wrote: > readpage() implmentations zero out part of the page-cache page mapping a > file hole. This commit adds a call back to "struct fsverity_operations" > which readpage() can invoke to verify the file hole. I don't see why this is needed. Since now CONFIG_FS_VERITY will be a bool, can't you just call fsverity_verify_hole() directly? > > Signed-off-by: Chandan Rajendra <chandan@linux.ibm.com> > --- > fs/ext4/super.c | 1 + > include/linux/fsverity.h | 1 + > 2 files changed, 2 insertions(+) > > diff --git a/fs/ext4/super.c b/fs/ext4/super.c > index 4493ddc357c6..94e0d4b56653 100644 > --- a/fs/ext4/super.c > +++ b/fs/ext4/super.c > @@ -1459,6 +1459,7 @@ static const struct fsverity_operations ext4_verityops = { > .read_metadata_page = ext4_read_verity_metadata_page, > .verity_required = ext4_verity_required, > .readpage_limit = ext4_readpage_limit, > + .check_hole = fsverity_check_hole, > }; > #endif /* CONFIG_FS_VERITY */ > > diff --git a/include/linux/fsverity.h b/include/linux/fsverity.h > index fc8113acbbfe..8bc28b07d550 100644 > --- a/include/linux/fsverity.h > +++ b/include/linux/fsverity.h > @@ -20,6 +20,7 @@ struct fsverity_operations { > struct page *(*read_metadata_page)(struct inode *inode, pgoff_t index); > bool (*verity_required)(struct inode *inode, pgoff_t index); > loff_t (*readpage_limit)(struct inode *inode); > + bool (*check_hole)(struct inode *inode, struct page *page); > }; > > #ifdef CONFIG_FS_VERITY > -- > 2.19.1 >
On Wednesday, February 20, 2019 6:05:07 AM IST Eric Biggers wrote: > On Mon, Feb 18, 2019 at 03:34:31PM +0530, Chandan Rajendra wrote: > > readpage() implmentations zero out part of the page-cache page mapping a > > file hole. This commit adds a call back to "struct fsverity_operations" > > which readpage() can invoke to verify the file hole. > > I don't see why this is needed. Since now CONFIG_FS_VERITY will be a bool, > can't you just call fsverity_verify_hole() directly? Yes, I will drop this patch and replace the call back invocation (provided in the next patch) with a call to fsverity_check_hole().
diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 4493ddc357c6..94e0d4b56653 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -1459,6 +1459,7 @@ static const struct fsverity_operations ext4_verityops = { .read_metadata_page = ext4_read_verity_metadata_page, .verity_required = ext4_verity_required, .readpage_limit = ext4_readpage_limit, + .check_hole = fsverity_check_hole, }; #endif /* CONFIG_FS_VERITY */ diff --git a/include/linux/fsverity.h b/include/linux/fsverity.h index fc8113acbbfe..8bc28b07d550 100644 --- a/include/linux/fsverity.h +++ b/include/linux/fsverity.h @@ -20,6 +20,7 @@ struct fsverity_operations { struct page *(*read_metadata_page)(struct inode *inode, pgoff_t index); bool (*verity_required)(struct inode *inode, pgoff_t index); loff_t (*readpage_limit)(struct inode *inode); + bool (*check_hole)(struct inode *inode, struct page *page); }; #ifdef CONFIG_FS_VERITY
readpage() implmentations zero out part of the page-cache page mapping a file hole. This commit adds a call back to "struct fsverity_operations" which readpage() can invoke to verify the file hole. Signed-off-by: Chandan Rajendra <chandan@linux.ibm.com> --- fs/ext4/super.c | 1 + include/linux/fsverity.h | 1 + 2 files changed, 2 insertions(+)