On 2019/7/1 14:08, Christoph Hellwig wrote: > On Sun, Jun 30, 2019 at 04:19:32PM -0700, Darrick J. Wong wrote: >> On Sat, Jun 29, 2019 at 03:30:20PM +0800, Chao Yu wrote: >>> Some filesystems like erofs/reiserfs have the ability to pack tail >>> data into metadata, however iomap framework can only support mapping >>> inline data with IOMAP_INLINE type, it restricts that: >>> - inline data should be locating at page #0. >>> - inline size should equal to .i_size >> >> Wouldn't it be easier simply to fix the meaning of IOMAP_INLINE so that >> it can be used at something other than offset 0 and length == isize? >> IOWs, make it mean "use the *inline_data pointer to read/write data >> as a direct memory access"? > > Yes. I vaguely remember Andreas pointed out some issues with a @Andreas, could you please help to explain which issue we may encounter without those limits? Thanks, > general scheme, which is why we put the limits in. If that is not just > me making things up we'll need to address them. Either way just copy > and pasting code isn't very helpful. > . >
