| Message ID | 20170914095047.23935-4-agruenba@redhat.com (mailing list archive) |
|---|---|
| State | Not Applicable, archived |
| Headers | show |
On Thu, Sep 14, 2017 at 11:50:46AM +0200, Andreas Gruenbacher wrote: > +int ext4_inline_data_iomap(struct inode *inode, struct iomap *iomap) > +{ > + __u64 addr; > + int error = -EAGAIN; > + struct ext4_iloc iloc; > + > + down_read(&EXT4_I(inode)->xattr_sem); > + if (!ext4_has_inline_data(inode)) > + goto out; .... > +out: > + up_read(&EXT4_I(inode)->xattr_sem); > + return error; > +} If we race with the inline data flag getting cleared, ext4_iomap_begin() will return with -EAGAIN. As near as I can tell, this will get reflected all the way up to userspace, instead of having the retry happen in the kernel. Is this intentional? It looks like a user visible change, no? - Ted -- To unsubscribe from this list: send the line "unsubscribe linux-xfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Thu, Sep 28, 2017 at 7:57 PM, Theodore Ts'o <tytso@mit.edu> wrote: > On Thu, Sep 14, 2017 at 11:50:46AM +0200, Andreas Gruenbacher wrote: >> +int ext4_inline_data_iomap(struct inode *inode, struct iomap *iomap) >> +{ >> + __u64 addr; >> + int error = -EAGAIN; >> + struct ext4_iloc iloc; >> + >> + down_read(&EXT4_I(inode)->xattr_sem); >> + if (!ext4_has_inline_data(inode)) >> + goto out; > .... >> +out: >> + up_read(&EXT4_I(inode)->xattr_sem); >> + return error; >> +} > > > If we race with the inline data flag getting cleared, > ext4_iomap_begin() will return with -EAGAIN. Actually, in that case, ext4_iomap_begin will treat the file as non-inline. It will not return -EAGAIN. > As near as I can tell, > this will get reflected all the way up to userspace, instead of having > the retry happen in the kernel. Is this intentional? > > It looks like a user visible change, no? That would be a bug. Thanks, Andreas -- To unsubscribe from this list: send the line "unsubscribe linux-xfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Mon, Oct 02, 2017 at 04:39:59PM +0200, Andreas Gruenbacher wrote: > > If we race with the inline data flag getting cleared, > > ext4_iomap_begin() will return with -EAGAIN. > > Actually, in that case, ext4_iomap_begin will treat the file as non-inline. > It will not return -EAGAIN. Ah, thanks, I missed that. - Ted -- To unsubscribe from this list: send the line "unsubscribe linux-xfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index e2abe01c8c6b..ae3e4a25821a 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -3048,6 +3048,10 @@ extern struct buffer_head *ext4_get_first_inline_block(struct inode *inode, extern int ext4_inline_data_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, int *has_inline, __u64 start, __u64 len); + +struct iomap; +extern int ext4_inline_data_iomap(struct inode *inode, struct iomap *iomap); + extern int ext4_try_to_evict_inline_data(handle_t *handle, struct inode *inode, int needed); diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c index 28c5c3abddb3..f0bbc8cb6555 100644 --- a/fs/ext4/inline.c +++ b/fs/ext4/inline.c @@ -12,6 +12,7 @@ * GNU General Public License for more details. */ +#include <linux/iomap.h> #include <linux/fiemap.h> #include "ext4_jbd2.h" @@ -1827,6 +1828,38 @@ int ext4_destroy_inline_data(handle_t *handle, struct inode *inode) return ret; } +int ext4_inline_data_iomap(struct inode *inode, struct iomap *iomap) +{ + __u64 addr; + int error = -EAGAIN; + struct ext4_iloc iloc; + + down_read(&EXT4_I(inode)->xattr_sem); + if (!ext4_has_inline_data(inode)) + goto out; + + error = ext4_get_inode_loc(inode, &iloc); + if (error) + goto out; + + addr = (__u64)iloc.bh->b_blocknr << inode->i_sb->s_blocksize_bits; + addr += (char *)ext4_raw_inode(&iloc) - iloc.bh->b_data; + addr += offsetof(struct ext4_inode, i_block); + + brelse(iloc.bh); + + iomap->addr = addr; + iomap->offset = 0; + iomap->length = min_t(loff_t, ext4_get_inline_size(inode), + i_size_read(inode)); + iomap->type = 0; + iomap->flags = IOMAP_F_DATA_INLINE; + +out: + up_read(&EXT4_I(inode)->xattr_sem); + return error; +} + int ext4_inline_data_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, int *has_inline, __u64 start, __u64 len) diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index d9e633c12aae..7755f41bdfc3 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -3404,8 +3404,20 @@ static int ext4_iomap_begin(struct inode *inode, loff_t offset, loff_t length, struct ext4_map_blocks map; int ret; - if (WARN_ON_ONCE(ext4_has_inline_data(inode))) - return -ERANGE; + + if (flags & IOMAP_REPORT) { + if (ext4_has_inline_data(inode)) { + ret = ext4_inline_data_iomap(inode, iomap); + if (ret != -EAGAIN) { + if (ret == 0 && offset >= iomap->length) + ret = -ENOENT; + return ret; + } + } + } else { + if (WARN_ON_ONCE(ext4_has_inline_data(inode))) + return -ERANGE; + } map.m_lblk = first_block; map.m_len = last_block - first_block + 1;